article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
l723 is an isolated dark cloud at a distance of 300 @xmath6 150 pc ( goldsmith et al . 1984 ) .
located in this cloud is a class 0 source , iras 19156 + 1906 , with a luminosity of @xmath03.4 l@xmath7 ( dartois et al .
the iras source is associated with a co outflow , first mapped by goldsmith et al .
the outflow shows a peculiar quadrupolar morphology , consisting of two pairs of red - blue lobes with a common center , as is clearly seen in the maps of avery et al .
the larger pair of lobes extends along a direction with a position angle ( p.a . )
@xmath8 100@xmath9 , while the smaller pair extends along a direction with a p.a .
@xmath8 30@xmath9 . over the years
, the peculiar quadrupolar morphology of the l723 outflow has been interpreted as due to the presence of either one or two driving sources ( see anglada , rodrguez & torrelles 1996 and references therein ) . in the first case , the quadrupolar morphology is attributed to limb - brightening effects in the lobes of a single bipolar outflow , to splitting of the lobes due to interaction with ambient clumps , or to precession of the outflow axis . in the second case ,
the quadrupolar structure is attributed to two independent bipolar outflows , each one driven by a different source .
ccccccrc & & & & bootstrapped & & & + & & & effective & flux density of & & + & observation & hour angle ranges & time on source & phase calibrator & hpbw & & rms noise + epoch & date & ( h ) & ( h ) & ( jy ) & ( arcsec ) & & ( @xmath10jy beam@xmath11 ) + 1995.6 & 95-aug-12 & @xmath124.4 to @xmath122.5 , 2.0 to 4.0 & 3.7 & 1.146 @xmath6 0.006 & 0.32 @xmath13 0.27 & @xmath1285 & 11 + 1997.0 & 97-jan-10 & @xmath122.0 to 2.5 & 4.0 & 0.926 @xmath6 0.004 & 0.27 @xmath13 0.25 & 15 & 11 + 1998.2 & 98-mar-27 & 1.0 to 3.5 , 4.3 to 5.8 & 3.7 & 0.94 @xmath6 0.01 & 0.34 @xmath13 0.25 & 58 & 12 + 1998.4 & 98-may-26 & 1.0 to 3.5 , 4.3 to 5.8 & 3.7 & 0.952 @xmath6 0.007 & 0.34 @xmath13 0.25 & 58 & 11 + 1999.5 & 99-jul-03 & 1.0 to 3.5 , 4.3 to 5.8 & 3.7 & 1.260 @xmath6 0.006 & 0.39 @xmath13 0.26 & 50 & 14 + 2000.9 & 00-dec-16 & @xmath125.0 to 4.5 & 8.0 & 1.65 @xmath6 0.02 & 0.28 @xmath13 0.27 & @xmath1270 & 6 ccccrc[!b ] & & bootstrapped & & & + & & flux density of & & + observation & vla & phase calibrator & hpbw & & rms noise + date & configuration & ( jy ) & ( arcsec ) & & ( mjy beam@xmath11 ) + 02-jan-05 & d & 2.38 @xmath6 0.04 & [email protected] & @xmath1222 & 0.12 + 04-feb-01 & cnb & 1.85 @xmath6 0.05 & [email protected] & @xmath1276 & 0.12 + 04-mar-26 & c & 2.22 @xmath6 0.05 & [email protected] & @xmath128 & 0.12 + 05-jul-29 & c & 1.68 @xmath6 0.02 & [email protected] & @xmath1434 & 0.12 very large array ( vla ) d - configuration observations ( angular resolution of @xmath08@xmath1 ) towards the center of the outflow revealed two 3.6 cm continuum sources , vla 1 and vla 2 , separated by 15@xmath1 ( anglada et al .
however , only vla 2 is associated with millimeter emission from circumstellar dust ( cabrit & andr 1991 ) , suggesting that it is the counterpart of iras 19156 + 1906 .
vla 2 has also been associated with a nearby ( @xmath00@xmath57 northeast from vla 2 ) h@xmath15o maser as well as with nh@xmath16 emission ( girart et al .
also , only vla 2 seems to be associated with the co outflow as shown by interferometric observations with the nobeyama millimeter array ( hirano et al .
1998 ) and with the berkeley illinois maryland association ( bima ) array ( lee et al . 2002 ) . in the 3.6 cm maps of anglada
1996 ) , made with data from the a and b configurations of the vla ( angular resolution of @xmath00@xmath53 ) , the centimeter emission of vla 2 is elongated along a direction with a p.a .
@xmath8 118@xmath9 , that is close to that of the large pair of co lobes .
in addition , h@xmath15 emission and herbig - haro ( hh ) objects have been found in the region along a direction with a p.a .
@xmath8 115@xmath9 , similar to that of the elongation of vla 2 ( vrba et al .
1986 ; hodapp 1994 ; palacios & eiroa 1999 ; lpez et al .
these results suggest that the outflow activity along the direction @xmath17-@xmath18 is driven by the young protostellar object vla 2 . however , the identification of the exciting source of the smaller pair of co lobes at p.a.@xmath830@xmath9 remained unclear . as already pointed out , one possibility is that vla 2 is the driving source of the two pairs of bipolar lobes , e.g. , driving each pair in a different epoch as a consequence of precession of the jet axis .
the second possibility is the presence of another , still undetected source in the vicinity of vla 2 , that could be driving the small pair of co lobes .
other regions harboring multipolar outflows , when studied with subarcsecond angular resolution ( equivalent to physical scales of tens to hundreds of aus for the relatively nearby regions that are better studied ) and sensitivity at the level of tens of @xmath10jy , have shown that the multipolar outflow morphology is usually associated with the presence of multiple ysos at the center of the outflows : e.g. , iras 16293@xmath122422 ( co outflow : mizuno et al . 1990 ; radio continuum : estalella et al .
1991 ) ; iras 20050 + 2720 ( co outflow : bachiller et al . 1995 ; radio continuum : anglada , rodrguez & torrelles 1998 ) ; hh111 ( co outflow : cernicharo & reipurth 1996 ; radio continuum : reipurth et al .
1999 ) ; l1634 ( h@xmath15 jet : hodapp & ladd 1995 ; radio continuum : beltrn et al .
2002 ) ; hh288 ( co outflow : gueth et al . 2001 ; radio continuum : franco - hernndez & rodrguez 2003 ) ; omc-1s ( optical jets : bally et al .
2000 ; radio continuum : zapata et al .
2004 ) .
in this paper we present sensitive , high angular resolution multi - epoch observations at 3.6 cm of the radio sources detected in the central region of the l723 outflow .
our goals were to search for changes in the structure of vla 2 that could help us to clarify the origin of the quadrupolar morphology of the outflow and to search for additional indications of multiplicity at very small scales .
we also analyze vla archive 7 mm continuum data of the region .
cccccccc & & & & spectral + component & ra & dec & error & & 3.6 cm & 7 mm & index + a & 19 17 53.673 & + 19 12 19.59 & 0@xmath501 & & 0.11 @xmath6 0.04 & 0.7 @xmath6 0.2 & + 1.1 @xmath6 0.3 + b & 19 17 53.685 & + 19 12 19.34 & 0@xmath501 & & 0.08 @xmath6 0.02 & @xmath190.37 & @xmath19 + 0.9 + c & 19 17 53.739 & + 19 12 19.77 & 0@xmath507 & & 0.05 @xmath6 0.02 & 0.7 @xmath6 0.2 & + 1.5 @xmath6 0.4 + d & 19 17 53.894 & + 19 12 17.96 & 0@xmath509 & & @xmath190.02 & 0.6 @xmath6 0.2 & @xmath20 + 2.0 the 3.6 cm continuum observations were carried out in six epochs , ranging from 1995 august 12 to 2000 december 16 .
the data from the 1995.6 observations were already reported by anglada et al .
the vla of the national radio astronomy observatory ( nrao ) was used in its a configuration , providing an angular resolution of @xmath00@xmath53 .
absolute flux calibration was achieved by observing 3c 286 , for which a flux density of 5.20 jy at 3.6 cm was adopted .
data editing and calibration were carried out using the astronomical image processing system ( aips ) package of nrao , following the standard vla procedures .
the observation dates , the bootstrapped flux densities of 1925 + 211 , the phase calibrator used in all the observations , as well as the parameters of the synthesized beams , and the rms noise of the naturally weighted maps are given in table [ tabla1 ] .
due to the weakness of the sources detected the signal - to - noise ratio per baseline was insufficient to self - calibrate the data .
for each epoch , we made cleaned maps with different values of the parameter robust of imagr ( briggs 1995 ) .
the robust parameter ranges in value from + 5 to @xmath125 , with the maps made with robust=+5 ( equivalent to natural weighting of the visibility data ) having the highest sensitivity but the lowest angular resolution and the maps made with robust=@xmath125 ( equivalent to uniform weighting ) having the highest angular resolution but the lowest sensitivity . for each value of the robust parameter
the map of each epoch was restored using a circular beam whose hpbw is the geometric mean of the major and minor axes of the synthesized beam obtained by concatenating the uv data of all the six epochs ( hpbw = 0@xmath529 for naturally weighted maps and hpbw
= 0@xmath519 for uniformly weighted maps ) .
this average restoring beam was adopted to facilitate comparison between the different epochs and to allow averaging of all the 3.6 cm images into a single image .
we consider this procedure appropriate since the major and minor axes of the synthesized beams of the individual observing sessions ( see table [ tabla1 ] ) are within @xmath010% of the adopted average hpbw .
to improve the sensitivity and uv coverage , we have concatenated the uv data from all the epochs to obtain single maps , which we will refer to as `` all - epochs '' maps .
since some of the structures persist from one epoch to another , while others change , the procedure of concatenating data could result in spurious effects in the final images .
therefore , we checked that the maps obtained by concatenating all the uv data do not show significant differences from the maps obtained by averaging all the single epoch maps .
the 7 mm continuum data were taken from the vla archive .
the data consist of four sets of observations in the d , c and cnb configurations ( see table [ obs7 mm ] ) .
absolute flux calibration was achieved by observing 3c 286 , and a model image at 7 mm for this calibrator was used to calibrate the data in amplitude . in order to improve the sensitivity of the final image , but keeping a balanced contribution from the different configurations , we have concatenated all the available uv data giving the same weight to the d and c+cnb data .
since these data come from different configurations , it is not possible to average the individual maps . from these uv data we obtained a 7 mm `` all - epochs '' naturally weighted map .
in figure [ fig1 ] we show naturally weighted 3.6 cm maps of vla 2 for each individual epoch .
we have detected radio continuum emission from vla 2 in all the epochs , with total flux densities varying in the range 0.2 - 0.5 mjy . in these maps ,
we see changes in the morphology of vla 2 , but some of the structures persist from one epoch to another . in particular , in three of the epochs vla 2 shows a clear double morphology with two components separated by @xmath00@xmath53 along a p.a . of @xmath0150@xmath9 .
this double morphology appears in the 1998.2 and 1998.4 epochs , where the uv coverage was exactly the same , and in the 2000.9 epoch , which was a full - track observation and therefore the uv coverage obtained was much better than in the other epochs ( see table [ tabla1 ] for the ha ranges and time on source for each epoch of observation ) . in the 1999.5 epoch ,
the uv coverage was also very similar to that of the 1998 epochs , but the double morphology is not so evident .
we think this could be due to the higher rms noise of the 1999.5 map and to the increase of the flux density of the northern component ( see below ) that partially masks the emission of the southern component . given these considerations ,
it is likely that vla 2 is actually a persistent double radio source , with the differences in the observed morphology being mainly due to the different uv coverage from one epoch to another .
lccl[!b ] & & + epoch & vla 2a & vla 2b & + 1995.6 & 0.08 @xmath6 0.01 & 0.06 @xmath6 0.01 & 0.30 @xmath6 0.03 + 1997.0 & 0.07 @xmath6 0.01 & 0.06 @xmath6 0.01 & 0.18 @xmath6 0.03 + 1998.2 & 0.06 @xmath6 0.01 & 0.06 @xmath6 0.01 & 0.20 @xmath6 0.02 + 1998.4 & 0.06 @xmath6 0.01 & 0.07 @xmath6 0.01 & 0.16 @xmath6 0.02 + 1999.5 & 0.15 @xmath6 0.01 & 0.05 @xmath6 0.01 & 0.21 @xmath6 0.03 + 2000.9 & 0.13 @xmath6 0.01 & 0.07 @xmath6 0.01 & 0.24 @xmath6 0.02 a way to mitigate the effects introduced by the different uv coverages would be to make uniformly weighted maps of the individual epochs . in uniformly weighted maps all the grid cells have equal weight , which compensates for the high concentration of data points in the inner region of the uv plane .
this provides a more uniform coverage of the gridded uv plane , resulting in maps having higher angular resolution than naturally weighted maps .
unfortunately , uniform weighting yields lower sensitivity , and due to the weakness of the source , it is difficult to detect it in the uniformly weighted map of a single epoch .
however , it is still possible to test our hypothesis of a double radio source with the available data . for this purpose
we have made uniformly weighted maps from two different sets of concatenated uv data : ( i ) the three epochs in which vla 2 does not show clearly a double morphology , and ( ii ) the three epochs in which vla 2 shows a clear double morphology .
this procedure allows us to compare two maps with high enough angular resolution and sensitivity , as well as a good uv coverage .
as can be seen in figure [ fig2 ] , vla 2 shows always a double morphology ( with a slight change of @xmath00@xmath51 in the position of the sources ) , even in the map made with the data of the epochs in which the double morphology was not evident .
this strongly supports that two radio continuum sources ( that we call vla 2a and vla 2b ) persist through all the epochs , covering a time span @xmath21 5 yr . since all the data seem to be consistent with the presence of a double source , we obtained an `` all - epochs '' uniformly weighted map ( fig .
[ fig3]a ) by concatenating all the uv data .
the positions and flux densities of components vla 2a and 2b , derived from gaussian ellipsoid fits to this map , are given in table [ tabla2 ] .
the angular separation of the two components is 0@xmath529 , which corresponds to a projected separation of @xmath090 au ( assuming a distance of 300 pc ) .
if these sources were tracing some kind of high - velocity outflowing features , such as knots in a jet , a significant variation in their positions in the single epoch maps would be expected ( e.g. , a velocity of @xmath0100 km s@xmath11 in the plane of the sky would result in an angular displacement of @xmath22 over the period of @xmath05 yr covered by our observations ) .
the lack of a significant displacement from one epoch to another excludes that these sources trace high - velocity material , unless the axis of the motion was very close the line of sight ( which does not seem to be the case , given that the blue and red lobes of the molecular outflow appear clearly separated on the sky ) .
this , together with the additional properties observed for these sources ( see below ) led us to interpret vla 2a and vla 2b as a close binary system of embedded ysos .
the individual epoch maps shown in figure [ fig1 ] suggest variability in the 3.6 cm flux density of the compact components of the double source . in table
[ tabla3 ] we give the peak intensities of vla 2a and vla 2b at each epoch of observation . because the angular resolution of our observations is similar to the angular separation of the two sources , these values of the peak intensity
must be taken with care .
however , given that the maps of the different epochs have been restored with the same beam , the variations of the peak intensities are expected to be representative of the variability of the individual components .
the peak intensity of vla 2b appears to be constant at all epochs , consistent with a value of 0.06 @xmath6 0.01 mjy beam@xmath11 ( see table [ tabla3 ] ) .
on the other hand , the peak intensity of vla 2a is consistent with a value of 0.07 @xmath6 0.01 mjy beam@xmath11 ( see table [ tabla3 ] ) in the four first epochs , while in the last two epochs the peak intensity of the source increased approximately by a factor of two .
similar flux density variability has been observed in several embedded ysos ( e.g. , cep a , porras et al .
2002 ; b335 , vila et al . 2001 ) .
we interpret the variability in the flux density of vla 2a as due to material recently ejected that still remains very close to the yso and it is not distinguishable from the source .
in addition , we have detected 7 mm emission from vla 2a ( see fig .
[ fig4 ] ) .
the flux densities at 3.6 cm and 7 mm of vla 2a ( see table [ tabla2 ] ) imply a spectral index @xmath23 @xmath8 [email protected] ( s@xmath24 @xmath25 @xmath26 ) , consistent with partially optically thick free - free emission that could be tracing a radio jet .
the nominal value of the spectral index obtained is slightly higher than the typical values found in ionized jets driven by ysos , 0.3 @xmath27 @xmath23 @xmath27 0.9 ( anglada 1996 and references therein ) .
this higher value of the spectral index of vla 2a could be indicating that we are partially missing the extended emission of the possible radio jet , either because of insufficient uv coverage or because part of this extended emission is confused with the emission of the nearby source vla 2b .
since the outer part of a thermal radio jet is where the emission is optically thinner , we would be observing preferentially the emission coming from the region closer to the driving source , where the optical depth is higher .
this effect would increase the value of the spectral index with respect to that of the overall emission of a radio jet .
alternatively , a high value of the spectral index could be indicative of a significant contribution from dust emission at 7 mm .
both possibilities ( free - free emission from a jet with a possible contribution of dust emission from a circumstellar disk or an envelope ) are consistent with vla 2a being a yso .
we should note that the spectral indices derived in this paper have been obtained from non - simultaneous observations and can be affected by time - variability of the sources .
however , as discussed above , the flux variations observed in our multi - epoch observations suggest modest variability that it is not expected to affect our conclusions on the nature of the emission derived from the spectral index .
also , the 3.6 cm and 7 mm observations have different uv coverages , that could introduce additional errors in the flux density measurements of extended emission .
finally , the presence of several sources in a small area makes difficult to isolate the contribution of each source .
therefore , the spectral indices reported in this paper should be considered as rough estimates .
in addition to the compact emission of the two sources vla 2a and vla 2b , in the individual 3.6 cm maps of figure [ fig1 ] we also note the presence of several weak `` knots '' that could be tracing extended emission . in order to gain sensitivity to extended structures
, we have concatenated all the 3.6 cm uv data and made a naturally weighted `` all - epochs '' map , that is shown in figure [ fig3]b .
this map shows extended emission with a p.a . of @xmath0115@xmath9 , where we identify the knots
w1 , w2 , e1 , and e2 , that can also be seen in several of the individual epoch maps ( fig .
[ fig1 ] ) .
the extended emission seems to emanate from vla 2a , the northern component of the double source , since a line running through the major axis of this extended emission is centered close to the position of vla 2a ( see fig .
3b ) , which is the strongest source along this direction . the p.a . of this extended emission ( @xmath0115@xmath9 ) coincides with that of the system of hh objects observed in the region ( lpez et al .
2006 ) , suggesting that vla 2a is a yso that is driving these hh objects . the p.a . of
the extended emission is also similar to that of the symmetry axis of the larger pair of co outflow lobes ( @xmath0100@xmath9 ) , thus suggesting that vla 2a is also associated with this lobe pair .
we interpret the extended radio emission as tracing an ionized jet that originates at the position of vla 2a , and where the knots w1 , w2 , e1 , and e2 would be tracing electron density enhancements produced by shock - ionization .
we do not detect clear proper motions associated with these knots from one epoch to another , setting an upper limit to the proper motion velocity of the knots of @xmath027 km s@xmath11 ( however , due to the coarse temporal sampling and the low signal - to - noise ratio of the knots in the individual epoch maps , they are difficult to identify from one epoch to another , and we can not discard higher proper motion velocities ) .
this suggests that these knots have been excited by the interaction of the jet of vla 2a with dense , stationary clumps in the surrounding gas .
we also note the detection of the faint knots , n1 ( in the 2000.9 image ; fig .
[ fig1]f ) and s1 ( in the 1997.0 image ; fig .
[ fig1]b ) .
we interpret these two knots as shock - ionized gas associated with episodic ejections of material .
their location , nearly aligned with source vla 2b , at a p.a .
@xmath8 20@xmath9 , is consistent with being excited by a jet coming from this source .
the p.a . of this jet is similar to that of the smaller pair of co outflow lobes ( @xmath030@xmath9 ) and therefore , we tentatively propose vla 2b as its exciting source .
cccccccc & & & & & + & & flux density & major axis & minor axis & p.a .
+ epoch & ra & dec & error & ( mjy ) & ( arcsec ) & ( arcsec ) & ( deg ) + 1995.6 & 19 17 52.922 & + 19 12 08.85 & 0@xmath501 & 0.36 @xmath6 0.02 & 0.11 & 0.10 & 179 + 1997.0 & 19 17 52.923 & + 19 12 08.85 & 0@xmath501 & 0.38 @xmath6 0.02 & 0.13 & 0.00 & 37 + 1998.2 & 19 17 52.922 & + 19 12 08.86 & 0@xmath501 & 0.34 @xmath6 0.02 & 0.12 & 0.04 & 179 + 1998.4 & 19 17 52.922 & + 19 12 08.86 & 0@xmath501 & 0.31 @xmath6 0.02 & 0.11 & 0.05 & 179 + 1999.5 & 19 17 52.923 & + 19 12 08.86 & 0@xmath501 & 0.34 @xmath6 0.03 & 0.13 & 0.07 & 179 + 2000.9 & 19 17 52.923 & + 19 12 08.85 & 0@xmath501 & 0.34 @xmath6 0.01 & 0.13 & 0.00 & 30 in summary , we detected two compact 3.6 cm sources , vla 2a and vla 2b , separated by @xmath00@xmath53 along a p.a .
of @xmath0150@xmath9 .
we also detected extended 3.6 cm emission , centered on vla 2a , and along a p.a .
of @xmath0115@xmath9 , that coincides with the p.a of the system of hh objects in the region and is similar to that of the larger pair of co outflow lobes .
the spectral index of vla 2a in the 3.6 cm to 7 mm wavelength range , is consistent with this source tracing the origin of an ionized radio jet with a possible contribution from dust .
therefore , we conclude that vla 2a is an embedded yso associated with a radio jet ; it is also the likely exciting source of the system of hh objects and of the larger pair of lobes of the co outflow observed in the region .
we identify several knots ( w1 , w2 , e1 , and e2 ) distributed along the extended 3.6 cm emission , at both sides of vla 2a .
since these knots are much weaker than vla 2a and are aligned along the axis of the extended emission we interpret them as tracing electron density enhancements in the jet from vla 2a rather than tracing the location of other ysos .
the lack of clear proper motions suggests that these knots have been excited by the interaction of the jet with dense , stationary clumps in the gas surrounding vla 2a .
the compact source vla 2b has a flux density comparable to that of vla 2a and both sources are aligned along a p.a .
of @xmath0150@xmath9 , a direction where there are no indications of outflow activity .
therefore , we interpret vla 2b as a second embedded yso , constituting a close ( projected separation of 90 au ) binary system .
neither vla 2a nor vla 2b show noticeable proper motions ; this result strongly suggests that these sources do not trace high - velocity material , and is consistent with these sources being ysos .
unfortunately , with the current data it is not possible to obtain a precise enough measure of proper motions that would allows us to derive possible orbital motions , as has been done in other sources ( e.g. , loinard et al .
2002 ) ; this kind of data could be helpful to confirm that this is a binary system . also , there are no high angular resolution data outside the centimeter wavelength range that allow us to characterize the individual spectral energy distribution of each component in order to establish the evolutionary status of the sources and to confirm their nature as ysos
. therefore , these sources are good targets for future observations with the evla and alma .
we also detect in the single epoch images two faint knots ( n1 and s1 ) nearly aligned with the source vla 2b along a p.a . of @xmath020@xmath9 , that is similar to the p.a . of the smaller pair of co outflow lobes .
we tentatively interpret these two knots as shock - ionized gas associated with episodic ejections from vla 2b , that could be the driving source of these knots as well as of this pair of co outflow lobes .
in addition to vla 2a , our 7 mm image ( fig . [ fig4 ] ) shows emission from other two sources .
one of them , that we will refer to as vla 2c , coincides with the position of the water maser emission detected by girart et al .
( 1997 ) and furuya et al .
the coincidence of this source with the water maser suggests that vla 2c is a yso . at 7 mm , the source vla
2c appears elongated , with a deconvolved size of 1@xmath52 @xmath6 0@xmath52 with a p.a .
@xmath8 @xmath28 . at 3.6 cm ,
we detect weak extended emission that coincides with the source detected at 7 mm .
the 3.6 cm emission shows two peaks , that we call vla 2ca and vla 2cb . with the current data it is unclear if these two peaks actually correspond to two objects or to a single extended source .
sensitive high angular resolution multiwavelength observations would be required to further define the nature of this emission . from the total flux density at 3.6 cm ( vla 2ca @xmath14 vla 2cb ) and 7 mm ( see table [ tabla2 ] ) , we estimate a spectral index of [email protected] for vla 2c , a value suggestive of thermal emission from a yso ( see discussion in 3.1 ) .
the third 7 mm source , vla 2d , is located @xmath29 from vla 2a and does not have a detected 3.6 cm counterpart . however , this source is located within @xmath00@xmath56 of the nominal position of a 1.35 mm source detected by girart et al .
( 2007 ) with the submillimeter array ( sma ) ( see fig . [ fig4 ] ) .
given that the beam size of the sma observations is 3@xmath53@xmath131@xmath55 ( p.a.=79@xmath9 ) , we suggest that vla 2d is the 7 mm counterpart of the 1.35 mm source . from the total flux density at 7 mm of vla 2d , and adopting a 4@xmath30 upper limit to the 3.6 cm flux density , we estimate a lower limit to the spectral index of this source of @xmath02.0 in this wavelength range , suggesting that the 7 mm emission is dominated by thermal dust emission .
additionally , the 1.35 mm source seems to be associated with an elongated sio structure , probably tracing material shocked by a high velocity flow ( girart et al .
2007 ) , suggesting that this source is associated with a yso .
in addition to the sources vla 2a , vla 2b , vla 2c , and vla 2d that appear located at the center of the l723 outflow , our 3.6 cm images detect the source vla 1 , first reported by anglada et al .
( 1991 ) , and located @xmath015@xmath1 sw of the position of vla 2a .
the source vla 1 appears always compact ( deconvolved size @xmath31 ; see table [ tabla4 ] ) with a roughly constant flux density consistent at all epochs with a value of [email protected] mjy ( see table [ tabla4 ] ) .
we have not detected this source at 7 mm up to a 4@xmath30 level of 0.32 mjy . with our 3.6 cm flux density and the 7 mm upper limit
, we estimate for vla 1 a spectral index @xmath320.1 , a value consistent with the spectral index @[email protected]@xmath60.2 measured by anglada et al .
( 1996 ) between 3.6 and 6 cm .
as discussed by anglada et al .
( 1996 ) and girart et al .
( 1997 ) , this source is most probably a radio - emitting , optically - obscured t tauri star located very near to the denser part of the molecular cloud .
our accurate astrometry indicates that this compact source does not show noticeable proper motions , with a position displacement @xmath33 ( 3@xmath30 ) over the 5.3 yr period covered by our observations ( see table [ tabla4 ] ) .
since a total proper motion displacement of @xmath34 is expected for an object embedded in the l723 cloud ( assuming a distance of 300 pc and the brand & blitz 1993 model of galactic rotation ) , this result is compatible with vla 1 belonging to the l723 cloud .
however , the possibility that vla 1 is a line - of - sight source not directly associated with the molecular cloud is not discarded .
the remaining radio sources ( vla 2a , vla 2b , vla 2c , and vla 2d ) discussed in this paper are embedded in regions of extended radio emission and are associated with ejecta , which does not allow a precise determination of their possible proper motions . in the maps of avery
et al . ( 1990 ) , the l723 co outflow shows a four - lobe morphology with a large pair of bipolar lobes along a p.a . of @xmath17 and
a smaller pair of bipolar lobes along a p.a . of @xmath35 . as discussed in the previous sections , our radio continuum data suggest that a multiple system of at least four ysos is present , within a region of @xmath36 au , at the intersection of the axes of symmetry of the two pairs of co lobes .
furthermore , in 3.1 we discussed the results that led us to propose that vla 2a is related to the excitation of the larger pair of co lobes , while vla 2b could be driving the smaller pair of co lobes . nevertheless , the interferometric maps of lee et al .
( 2002 ) ( obtained with higher angular resolution than the maps of avery et al .
1990 ) reveal additional details of the co outflow structure ( see fig .
[ fig5]a ) .
as can be seen in this figure , the smaller pair of co lobes shows a compact shape with a well defined axis at p.a .
@xmath37 . on the other hand ,
the shape of the larger pair of co lobes appears clumpy , with only a few weak co emission knots near the symmetry axis of this lobe pair ( at p.a .
@xmath38 ) , and most of the emission concentrated at both sides of this central axis , roughly at p.a . @xmath39 and p.a .
the distribution of the co emission of this larger lobe pair was interpreted by lee et al .
( 2002 ) as tracing the limb - brightened cavity walls of a bipolar outflow with a p.a .
@xmath38 . however , we note that the distribution of the co emission at both sides of the central axis ( p.a .
@xmath38 ) of this lobe pair is asymmetric .
the emission at p.a .
@xmath41 starts at the center of the outflow ( the position of vla 2 ) and is stronger than the emission at p.a .
@xmath39 , which shows a gap , with the bulk of the emission apparently starting @xmath42 away from the outflow center .
furthermore , the p.a . of @xmath43 coincides with that of the extended radio emission associated with vla 2a and with that of the system of hh objects .
these considerations led us to propose a new interpretation of the high - velocity co emission in l723 in terms of three bipolar outflows . one bipolar outflow at p.a .
@xmath37 , that could be driven by vla 2b . a second bipolar outflow at p.a .
@xmath40 , which is associated with a system of hh objects and shock - ionized extended radio continuum emission with the same position angle , that is driven by vla 2a .
finally , a third bipolar outflow at p.a .
@xmath39 , whose exciting source is uncertain .
the lobes of the proposed outflows at p.a . @xmath37 and p.a .
@xmath40 appear well `` connected '' with the outflow center , suggesting recent outflow activity in these directions .
on the contrary , the lobes of the outflow at p.a .
@xmath39 are weaker and well detached from the central region , suggesting that they resulted from activity taking place in the past . therefore , we speculate that this outflow at p.a .
@xmath8 90@xmath9 could be a `` fossil '' outflow in the sense that its exciting source has not been active in the recent past .
the motion observed will then be the result of momentum conservation of the gas accelerated when the exciting source was active .
a similar fossil molecular outflow has been proposed to exist in the molecular cloud b5 ( yu , billawala , & bally 1999 ) .
our high angular resolution 3.6 cm and 7 mm continuum observations towards the previously known source vla 2 , at the center of the l723 multipolar co outflow , have revealed a group of at least four ysos within a region of @xmath44 ( 1200 au ) in size .
the size of this region is comparable with the size of the region where girart et al .
( 1997 ) found a local heating of the high density gas .
two of our continuum sources , vla 2a and vla 2b , form a close radio binary whose components are separated by @xmath00@xmath529 ( @xmath0 90 au ) .
another source , vla 2c , is associated with a water maser , and a fourth source , vla 2d , that we only detect at 7 mm , appears to be the counterpart of a recently discovered 1.35 mm source .
we propose that the multipolar co outflow in l723 could result from the superposition of at least three independent bipolar outflows with position angles of @xmath0115@xmath9 , @xmath090@xmath9 , and @xmath030@xmath9 , driven by three different ysos .
our 3.6 cm observations suggest that vla 2a is associated with an ionized jet at a position angle of @xmath0115@xmath45 .
we propose that this is the exciting source of the system of hh objects previously detected in the region , as well as of a bipolar co outflow with the same position angle .
the presence of two 3.6 cm knots aligned with vla 2b at a p.a . of @xmath020@xmath9 suggests that vla 2b could be the exciting source of the bipolar co outflow at a p.a of @xmath030@xmath9 .
finally , the third bipolar co outflow , with a position angle of @xmath0 90@xmath45 , seems to be a `` fossil '' outflow whose exciting source has not been very active in the recent past .
_ acknowledgements .
_ c.c .- g . acknowledges support from mec ( spain ) fpu fellowship .
, c.c .- g .
, m.o . and j.m.t .
acknowledge support from mec ( spain ) grant aya 2005 - 08523-c03 ( including feder funds ) .
c.c .- g . and m.o .
acknowledge partial support from junta de andaluca ( spain ) .
l.f.r . acknowledges the support of dgapa , unam , and of conacyt ( mxico ) .
we acknowledge chin - fei lee for providing us the fits file of the co emission in l723 . | we present high angular resolution very large array multi - epoch continuum observations at 3.6 cm and 7 mm towards the core of the l723 multipolar outflow revealing a multiple system of four radio sources suspected to be ysos in a region of only @xmath04@xmath1 ( 1200 au ) in extent .
the 3.6 cm observations show that the previously detected source vla 2 contains a close ( separation @xmath2 or @xmath090 au ) radio binary , with components ( a and b ) along a position angle of @xmath3 .
the northern component ( vla 2a ) of this binary system is also detected in the 7 mm observations , with a positive spectral index between 3.6 cm and 7 mm .
in addition , the source vla 2a is associated with extended emission along a position angle of @xmath4 , that we interpret as outflowing shock - ionized gas that is exciting a system of hh objects with the same position angle . a third , weak 3.6 cm source , vla 2c , that is detected also at 7 mm , is located @xmath00@xmath57 northeast of vla 2a , and is possibly associated with the water maser emission in the region .
the 7 mm observations reveal the presence of an additional source , vla 2d , located @xmath03@xmath55 southeast of vla 2a , and with a 1.35 mm counterpart .
all these radio continuum sources have a positive spectral index , compatible with them being ysos .
we also propose that the high velocity co emission observed in the region could be the superposition of multiple outflows ( at least three independent bipolar outflows ) excited by the ysos located at the core , instead of the previous interpretations in terms of only one or two outflows . |
a holomorphic map @xmath1 of a riemann surface @xmath2 into the complex lie group @xmath3 is called _ null _ if @xmath4 vanishes identically , where @xmath5 is a local complex coordinate of @xmath2 .
we consider two projections , one into the hyperbolic @xmath6-space @xmath7 and the other into the de sitter @xmath6-space @xmath8 where the definition of @xmath9 is in appendix [ app : conj ] .
it is well - known that the projection of a holomorphic null immersion into @xmath10 by @xmath11 gives a conformal ( constant mean curvature one ) immersion ( see @xcite , @xcite , @xcite ) .
moreover , conformal immersions are always given locally in such a manner .
on the other hand , spacelike surfaces given by the projection of holomorphic null immersions into @xmath0 by @xmath12 can have singularities , and are called _
we work with this class of surfaces that is larger than the class of immersions .
in fact , the class of immersions is too small , since _ there is only one , up to congruency , complete spacelike immersion _
@xcite , which we call an _ @xmath0-horosphere_. ( we also give a simple proof of this here .
see the last remark of section [ sec : prelim ] . )
the relationship between surfaces in @xmath10 and faces in @xmath0 is analogous to that between minimal surfaces in euclidean 3-space @xmath13 and spacelike maximal surfaces with singularities in lorentz - minkowski 3-space @xmath14 ( called _ maxfaces _ @xcite ) .
note that maximal surfaces also admit a weierstrass - type representation formula ( @xcite ) .
as in the case of maxfaces ( see @xcite ) , the first author @xcite investigated the global behavior of faces in @xmath0 , in particular proving an osserman - type inequality for complete faces of finite type whose ends are all elliptic , where a complete end of a face is called _ elliptic , parabolic _ , or _ hyperbolic _ if the monodromy matrix of the holomorphic lift @xmath1 is elliptic , parabolic , or hyperbolic , respectively ( see section [ sec : prelim ] ) .
one of our main results is the following , which implies that the ellipticity or parabolicity of ends follows from completeness : [ thm : main ] a complete end of a face in @xmath15 is never hyperbolic , so must be either elliptic or parabolic .
moreover , the total curvature over a neighborhood of such an end is finite .
we remark that there exist incomplete elliptic and parabolic ends .
it is remarkable that just completeness of an end is sufficient to conclude that it has finite total curvature .
this is certainly not the case for surfaces in @xmath10 nor for minimal surfaces in @xmath13 , but is similar to the case of maximal surfaces in @xmath16 @xcite .
although the asymptotic behavior of regular elliptic ends in @xmath15 is investigated in @xcite , there do also exist complete parabolic ends , and to describe them , a much deeper analysis is needed , which we will conduct in this article . as an application of theorem [ thm : main ] , we prove the following osserman - type inequality , which improves the result of @xcite by removing the assumptions of finite type and ellipticity of ends : [ thm : fujimori ] suppose a face @xmath17 is complete .
then there exist a compact riemann surface @xmath18 and a finite number of points @xmath19 such that @xmath2 is biholomorphic to @xmath20 , and @xmath21 where @xmath22 is the hyperbolic gauss map of @xmath23 and @xmath24 is the euler characteristic of @xmath18 .
furthermore , equality holds if and only if each end is regular and properly embedded .
trinoids in @xmath0 were constructed by lee and the last author using hypergeometric functions @xcite , and those trinoids with elliptic ends are complete in the sense of @xcite , and attain equality in . however , those having other types of ends are not complete , as their singular sets are not compact .
for this reason , our goal is not only to prove the above two theorems , but also to extend the framework for surfaces to include a larger class of surfaces , relaxing the immersedness and completeness conditions . if @xmath25 is of finite topology , i.e. if @xmath25 is diffeomorphic to a compact riemann surface @xmath18 with finitely many punctures @xmath26 , and if a face @xmath27 is weakly complete , whose precise definition will be given in section [ sec : prelim ] , we say that @xmath23 is a weakly complete face of finite topology .
we shall develop the framework under this more general notion , which includes all the trinoids in @xcite . in section [ sec : prelim
] , we recall definitions and basic results . in section [ sec
: projective ] , we investigate the monodromy of the hyperbolic metrics on a punctured disk around an end . as an application , we prove theorem [ thm : main ] in section [ sec : ends ] . in section [ sec : lgm ] , we give a geometric interpretation of the hyperbolic gauss map . in section [ sec : fujimori ] , we investigate the asymptotic behavior of regular parabolic ends , and prove theorem [ thm : fujimori ] . in appendix [ app : q - mero ] , we prove meromorphicity of the hopf differential for complete faces . in appendix
[ app : conj ] , we explain the conjugacy classes of @xmath9 .
generic singularities of faces are classified in @xcite .
a face is called _ embedded _ ( in the wider sense ) if it is embedded outside of some compact set of @xmath0 .
examples of complete embedded faces are given in @xcite as deformations of maxfaces in lorentz - minkowski 3-space @xmath14 .
let @xmath28 be the lorentz - minkowski space of dimension @xmath29 , with the lorentz metric @xmath30 then de sitter @xmath6-space is @xmath31 with metric induced from @xmath28 , which is a simply - connected lorentzian @xmath6-manifold with constant sectional curvature @xmath32 .
we identify @xmath28 with the set of @xmath33 hermitian matrices @xmath34 ( @xmath35 ) by @xmath36 where @xmath37 and @xmath38 .
then @xmath0 is @xmath39 with the metric @xmath40 the projection @xmath41 mentioned in the introduction is written explicitly as @xmath42 .
note that the hyperbolic @xmath6-space @xmath10 is given by @xmath43 and the projection is @xmath44 .
an immersion into @xmath45 is called _ spacelike _ if the induced metric on the immersed surface is positive definite . the complex lie
group @xmath3 acts isometrically on @xmath46 , as well as @xmath0 , by @xmath47 in fact , @xmath48 is isomorphic to the identity component @xmath49 of the isometry group @xmath50 of @xmath0 . note that each element of @xmath49 corresponds to an orientation preserving and orthochronous ( i.e. , time orientation preserving ) isometry .
the group of orientation preserving isometries of @xmath0 is generated by @xmath51 and the map @xmath52 aiyama - akutagawa @xcite gave a weierstrass - type representation formula in terms of holomorphic data for spacelike immersions in @xmath0 .
the first author @xcite extended the notion of surfaces as follows , like as for the case of maximal surfaces in the minkowski space @xcite .
let @xmath2 be a @xmath53-manifold .
a @xmath54-map @xmath55 is called a _ face _ if 1 .
there exists an open dense subset @xmath56 such that @xmath57 is a spacelike immersion , 2 . for any singular point ( that is , a point where the induced metric degenerates ) @xmath58
, there exists a @xmath59-differentiable function @xmath60 , defined on the intersection of neighborhood @xmath61 of @xmath58 with @xmath62 , such that @xmath63 extends to a @xmath59-differentiable riemannian metric on @xmath61 , where @xmath64 is the first fundamental form , i.e. , the pull - back of the metric of @xmath0 by @xmath23 , and 3 .
@xmath65 for any @xmath66 .
[ rmk : ori ] though the original definition of faces in @xcite assumed the orientability of the source manifold , our definition here does not . however , this difference is not of an essential nature .
in fact , for any face @xmath67 , @xmath2 is automatically orientable .
( see @xcite . )
[ rmk : causal ] a @xmath54-map @xmath67 is called a _ frontal _ if @xmath23 lifts to a @xmath54-map @xmath68 such that @xmath69 lies in the canonical contact plane - field on @xmath70 .
moreover , @xmath23 is called a _ wave front _ or a _ front _
if @xmath71 is an immersion , that is , @xmath72 is a legendrian submanifold .
if a frontal @xmath71 can lift up to a smooth map into @xmath73 , @xmath23 is called _ co - orientable _ , and otherwise it is called _ non - co - orientable_. wave fronts are a canonical class for investigating flat surfaces in the hyperbolic 3-space @xmath10 .
in fact , like for faces ( see theorem [ thm : fujimori ] in the introduction ) , an osserman - type inequality holds for flat fronts in @xmath10 ( see @xcite . )
although our faces belong to a special class of horospherical linear weingarten surfaces ( cf .
@xcite ) , they may not be ( wave ) fronts in general , but are co - orientable frontals . in particular , there is a globally defined non - vanishing normal vector field @xmath74 on the whole of @xmath2 for a given face @xmath67 .
it should be remarked that the limiting tangent plane at each singular point contains a lightlike direction , that is , a face is not spacelike on the singular set .
an oriented @xmath53-manifold @xmath2 on which a face @xmath75 is defined always has a complex structure ( see @xcite ) . since faces are all orientable and co - orientable ( cf .
@xcite ) , from now on , we will treat @xmath2 as a riemann surface , and we can assume the existence of a globally defined non - vanishing normal vector field .
the representation formula in @xcite can be extended for faces as follows : [ th : aa - rep ] let @xmath76 be a simply connected riemann surface .
let @xmath77 be a meromorphic function and @xmath78 a holomorphic @xmath32-form on @xmath76 such that @xmath79 is a riemannian metric on @xmath76 and @xmath80 is not identically @xmath32 . take a holomorphic immersion
@xmath81 satisfying @xmath82 then @xmath83 defined by @xmath84 is a face which is conformal away from its singularities .
the induced metric @xmath64 on @xmath85 , the second fundamental form @xmath86 , and the hopf differential @xmath87 of @xmath23 are given as follows : @xmath88 the singularities of the face occur at points where @xmath89 .
conversely , for any face @xmath83 , there exist a meromorphic function @xmath77 ( with @xmath80 not identically @xmath32 ) and a holomorphic @xmath32-form @xmath78 on @xmath76 so that @xmath90 is a riemannian metric on @xmath76 and holds , where @xmath91 is an immersion satisfying .
[ rmk : line ] by definition , faces have dense regular sets
. however , the projection of null holomorphic immersions might not have dense regular sets , in general .
such an example has been given in ( * ? ? ?
* remark 1.8 ) .
fortunately , we can explicitly classify such degenerate examples , as follows : let @xmath2 be a connected riemann surface and @xmath92 be a null _ immersion_. we assume that the set of singular points of the corresponding map @xmath93 has an interior point . then the secondary gauss map @xmath77 is constant on @xmath2 and @xmath89 . without loss of generality , we may assume @xmath94 .
since @xmath95 is an immersion , @xmath96 is positive definite .
then @xmath97 everywhere .
hence for each @xmath66 , one can take a complex coordinate @xmath5 such that @xmath98 .
then @xmath95 is a solution of @xmath99 without loss of generality , we may assume that @xmath100
. then we have @xmath101 and the corresponding map @xmath23 is computed as @xmath102 whose image is a lightlike line in @xmath0 .
thus , we have shown that the image of any degenerate surface is a part of a lightlike line .
[ lem : bryant ] theorem [ th : aa - rep ] is an analogue of the bryant representation for surfaces in @xmath10 , which explains why surfaces in both @xmath10 and @xmath0 are characterized by the projections @xmath103 and @xmath104 .
the surfaces in @xmath10 and @xmath0 are both typical examples in the class of linear weingarten surfaces .
a bryant - type representation formula for linear weingarten surfaces was recently given by j. glvez , a. martnez and f. miln @xcite .
[ lem : data ] following the terminology of @xcite , @xmath77 is called a _ secondary _ gauss map of @xmath23 .
the pair @xmath105 is called _ weierstrass data _ of @xmath23 , and @xmath95 is called a _ holomorphic null lift _ of @xmath23 .
the holomorphic @xmath53-differential @xmath87 as in is called the _ hopf differential _ of @xmath23 . in analogy with the theory of surfaces in @xmath10 , the meromorphic function @xmath106
is called the _ hyperbolic gauss map_. a geometric meaning for the hyperbolic gauss map is given in section [ sec : lgm ] .
[ rem : minimal - maximal ] corresponding to theorem [ th : aa - rep ] , a weierstrass - type representation formula is known for spacelike maximal surfaces in @xmath16 ( @xcite ) .
in fact , the weierstrass data @xmath105 as in theorem [ th : aa - rep ] defines null curves in @xmath107 by @xmath108 any maxface ( see @xcite for the definition ) is locally obtained as the real part of some @xmath109 .
moreover , their first fundamental forms and hopf differentials are given by .
the meromorphic function @xmath77 can be identified with the lorentzian gauss map . in this case
, we call the pair @xmath105 the _ weierstrass data _ of the maxface .
[ rem : small ] let @xmath110 be meromorphic functions on a riemann surface . set @xmath111
\dfrac{da}{dg } & \dfrac{db}{dg } \end{pmatrix},\qquad a=\sqrt{\frac{dg}{dg}},\quad b =- ga.\ ] ] then @xmath95 is a meromorphic null map with hyperbolic and secondary gauss maps @xmath22 and @xmath77 .
formula is called small s formula ( @xcite , @xcite ) .
[ rem : ambiguity ] the holomorphic null lift @xmath95 of a face @xmath23 is unique up to right - multiplication by matrices in @xmath9 , that is , for each @xmath112 , the projection of @xmath113 is also @xmath23 . under the transformation @xmath114 , the secondary gauss map @xmath77 changes by a mbius transformation : @xmath115 the conditions @xmath89 , @xmath116 , @xmath117 are invariant under this transformation . in particular ,
let @xmath118 be a face of a ( not necessarily simply connected ) riemann surface @xmath2 .
then the holomorphic null lift @xmath95 is defined only on the universal cover @xmath76 of @xmath2 .
take a deck transformation @xmath119 in @xmath85 .
since @xmath120 , there exists a @xmath121 such that @xmath122 the representation @xmath123 is called the _ monodromy representation _ , which induces a @xmath124-representation @xmath125 satisfying @xmath126 [ rem : action - null ] the action @xmath127 , @xmath128 , induces a rigid motion @xmath129 in @xmath0 , and the isometric motion @xmath130 as in corresponds to @xmath131 the secondary gauss map of @xmath132 is @xmath133 .
[ rem : gauss ] let @xmath134 be the gaussian curvature of @xmath64 on the set of regular points of @xmath23 . then @xmath135 is a pseudometric of constant curvature @xmath136 , which degenerates at isolated umbilic points .
we have @xmath137 [ rem : lift - metric ] the metric @xmath138 is induced from the canonical hermitian metric of @xmath3 via @xmath139 . when the face is defined on @xmath2 , @xmath22 and @xmath87
are as well , so @xmath140 is well - defined on @xmath2 , and is called the _ lift metric_. it is nothing but the _ dual metric _ of the surface @xmath141 in @xmath10 , see @xcite .
we now define two different notions of completeness for faces as follows : [ def : c ] we say a face @xmath142 is _ complete _ if there exists a symmetric @xmath53-tensor field @xmath143 which vanishes outside a compact subset @xmath144 such that the sum @xmath145 is a complete riemannian metric on @xmath2 .
see @xcite , with similar definitions in @xcite for flat fronts in @xmath10 and in @xcite for maxfaces .
[ def : wc ] we say that @xmath23 is _ weakly complete _ if it is congruent to an @xmath0-horosphere or if the lift metric is a complete riemannian metric on @xmath2 . here
, the _ @xmath0-horosphere _ is the totally umbilic surface , which is also the only complete immersed surface ( see remark [ rem : complete ] ) .
it has the weierstrass data @xmath146 @xmath147 and @xmath98 .
the metric @xmath148 of an @xmath0-horosphere can not be defined by as @xmath22 is constant and @xmath87 is identically @xmath149 , but can still be defined as the metric induced by @xmath150 , and is a complete flat metric on @xmath151 .
[ def : ft ] we say that @xmath23 is of _ finite type _ if there exists a compact set @xmath152 of @xmath2 such that the first fundamental form @xmath64 is positive definite and has finite total ( absolute ) curvature on @xmath153 .
let @xmath118 be a face of finite topology , that is , @xmath2 is diffeomorphic to a compact riemann surface @xmath154 with a finite number of points @xmath155 excluded . we can take a punctured neighborhood @xmath156 of @xmath157 which is biholomorphic to either the punctured unit disk @xmath158 or an annular domain , and @xmath157 is called a _ puncture - type end _ or an _ annular end _
, respectively .
[ prop : complete ] let @xmath67 be a face . if @xmath23 is complete , then 1 .
[ item : complete:0 ] the singular set of @xmath23 is compact , 2 . [
item : complete:1 ] @xmath23 is weakly complete , 3 . [
item : complete:2 ] @xmath2 has finite topology and each end is of puncture - type .
[ item : complete:0 ] is obvious .
if @xmath23 is totally umbilic , it is congruent to an @xmath0-horosphere and the assertion is obvious .
so we assume the hopf differential @xmath87 does not vanish identically . since the gaussian curvature of @xmath23 is nonnegative , completeness implies [ item : complete:2 ] by the appendix of @xcite .
so we shall now prove that completeness implies weak completeness : fix an end @xmath157 of @xmath23 . by an appropriate choice of a coordinate @xmath5 , the restriction of @xmath23 to a neighborhood of @xmath157 is @xmath159 .
we denote by @xmath160 the induced metric of the corresponding surface @xmath161 into hyperbolic @xmath6-space .
take a path @xmath162 such that @xmath163 as @xmath164 . then by and , @xmath165 holds , and hence completeness of @xmath23 implies that each lift @xmath166 of @xmath167 has infinite length with respect to @xmath160 . here ,
@xmath160 and @xmath168 are the pull - backs of the hermitian metric of @xmath3 by @xmath95 and @xmath150 , respectively .
yu @xcite showed that completeness of these two metrics are equivalent .
hence , @xmath169 has infinite length with respect to the metric @xmath170 .
since @xmath170 is well - defined on @xmath171 , @xmath167 also has infinite length with respect to @xmath170 , that is , the metric @xmath170 on @xmath171 is complete at @xmath149 .
thus , @xmath172 is a weakly complete end . for further properties of complete ends ,
see theorems [ thm : completeend ] and [ thm : labeledbyyang3 ] .
our definition of weak completeness of faces is somewhat more technical than that of maxfaces @xcite , but it is the correctly corresponding concept in @xmath15 : for data @xmath105 , weak completeness of the associated maxface in @xmath14 is equivalent to that of the face in @xmath0 .
the trinoids in @xmath0 constructed in @xcite are all weakly complete ( sometimes complete as well ) and all ends are @xmath77-regular , see section [ sec : ends ] .
the hopf differential @xmath87 of a complete face @xmath67 is meromorphic on its compactification @xmath18 , even without assuming that all ends of @xmath23 are regular .
see appendix [ app : q - mero ] .
it should be remarked that for surfaces in hyperbolic @xmath6-space , finiteness of total curvature is needed to show the meromorphicity of @xmath87 ( see @xcite ) . for any real number @xmath173
, we set @xmath174 a matrix in @xmath9 is called 1 .
_ elliptic _ if it is conjugate to @xmath175 ( @xmath176 $ ] ) in @xmath9 , 2 .
_ parabolic _ if it is conjugate to @xmath177 ( @xmath178 ) in @xmath9 , and 3 .
_ hyperbolic _ if it is conjugate to @xmath179 ( @xmath180 ) in @xmath9 .
any matrix in @xmath9 is of one of these three types , see appendix [ app : conj ] .
note that the parabolic matrices @xmath181 and @xmath182 are conjugate in @xmath9 if and only if @xmath183 .
though the set of conjugate classes of parabolic matrices is fully represented by @xmath184 , we may use various values of @xmath173 in this paper for the sake of simplicity . let @xmath142 be a weakly complete face of finite topology , where @xmath2 is diffeomorphic to a compact riemann surface @xmath154 with finitely many punctures @xmath185 .
any puncture @xmath157 , or occasionally a small neighborhood @xmath186 of @xmath157 , is called an _ end _ of @xmath23 .
an end is called _ elliptic _ , _ parabolic _ or _ hyperbolic _ when the monodromy matrix @xmath121 is elliptic , parabolic or hyperbolic , respectively , where @xmath187 is as in remark [ rem : ambiguity ] and @xmath119 is the deck transformation corresponding to the counterclockwise loop about @xmath157 .
let @xmath188 be a local complex coordinate of a riemann surface @xmath2 , and @xmath189 a meromorphic function on @xmath61 .
then @xmath190 is the _
schwarzian derivative _ of @xmath191 with respect to the coordinate @xmath5 . if @xmath192 at @xmath193 @xmath194 , where @xmath195 denotes higher order terms , then the positive integer @xmath196 is called the _ ( ramification ) order _ of @xmath189 , and we have @xmath197 we write @xmath198 , which we also call the _ schwarzian derivative .
the schwarzian derivative depends on the choice of local coordinates , but the difference does not , that is , @xmath199 is a well - defined holomorphic @xmath53-differential .
_ the schwarzian derivative is invariant under mbius transformations : @xmath200 holds for @xmath201 , where @xmath202 denotes the mbius transformation as in .
conversely , if @xmath203 , there exists an @xmath204 such that @xmath205 .
let @xmath118 be a face with the hyperbolic gauss map @xmath22 , a secondary gauss map @xmath77 and the hopf differential @xmath87
. then @xmath206 [ rem : complete ] here we give a proof that the only complete immersion is the totally umbilic one , that is , the @xmath0-horosphere , which is simpler than the original proofs in @xcite .
( the proof is essentially the same as for the case of maximal surfaces in @xmath16 given in ( * ? ? ?
* remark 1.2 ) . )
let @xmath67 be a complete immersion .
without loss of generality , we may assume that @xmath2 is both connected and simply connected .
then the weierstrass data @xmath105 as in theorem [ th : aa - rep ] is single - valued on @xmath2 .
since @xmath23 has no singular points , we may assume that @xmath117 holds on @xmath2 .
since @xmath207 , the metric @xmath208 is a complete flat metric on @xmath2 .
then the uniformization theorem yields that @xmath2 is bi - holomorphic to @xmath151 , and @xmath77 must be a constant function , which implies that the image of @xmath23 must be totally umbilic .
by remark [ rem : ambiguity ] , the monodromy of a holomorphic null immersion @xmath95 is elliptic , parabolic or hyperbolic if and only if the monodromy of its secondary gauss map @xmath77 is elliptic , parabolic or hyperbolic , respectively . in this section , in an abstract setting , we give results needed for investigating the behavior of @xmath77 at a puncture - type end , in terms of the monodromy of @xmath77 .
let @xmath209 be the punctured unit disk and @xmath210 a holomorphic @xmath53-differential on @xmath171 .
then there exists a holomorphic developing map @xmath211 such that @xmath212 , where @xmath213 is the universal cover of @xmath171 . for any other holomorphic function @xmath191 such that @xmath214 , there exists an @xmath215 so that @xmath216 .
thus there exists a matrix @xmath217 such that @xmath218 where @xmath219 is the generator of @xmath220 corresponding to a counterclockwise loop about the origin .
we call @xmath143 the monodromy matrix of @xmath221 . if there exists a @xmath222 so that @xmath223 , @xmath224 is called a _ @xmath124-projective connection _ on @xmath171 and @xmath222 is called a @xmath124-lift of @xmath224
. a @xmath124-projective connection on @xmath171 has a removable singularity , a pole or an essential singularity at @xmath149 , and is said to have a _
regular singularity _ at @xmath149 if it has at most a pole of order 2 at @xmath149 .
( the general definition of projective connections is given in @xcite and @xcite . there
exist holomorphic @xmath53-differentials on @xmath171 which are not @xmath124-projective connections . ) when @xmath225 , it is conjugate to one of the matrices in .
the @xmath124-projective connection @xmath224 is then called _
elliptic _ , _ parabolic _ or _ hyperbolic
_ when @xmath143 is _ elliptic _ , _ parabolic _ or _ hyperbolic _ , respectively .
this terminology is independent of the choice of @xmath222 . by the property
, the schwarzian derivative @xmath226 of the secondary gauss map @xmath77 of a face is an example of a @xmath124-projective connection .
note that a @xmath124-lift @xmath222 has the @xmath124 ambiguity @xmath227 for @xmath228 .
the property that @xmath229 ( resp .
@xmath230 ) is independent of this ambiguity .
[ rem : inverse ] let @xmath221 be a @xmath124-lift of a @xmath124-projective connection @xmath224 .
then @xmath231 is also a @xmath124-lift of @xmath224 , because @xmath232 for any @xmath228 .
however , @xmath233 , and one can show that there is no matrix @xmath234 such that @xmath235 , that is , @xmath236 is not @xmath124-equivalent to @xmath221 . in the rest of this article , as well as in the following proposition
, we use @xmath237 which is motivated by an isomorphism between @xmath238 and @xmath9 . see appendix [ app : conj ] .
[ n ] let @xmath224 be a @xmath124-projective connection on @xmath171 . then the following assertions hold : 1 . [
item : elliptic ] suppose that @xmath224 is elliptic .
then , 1 . [ item : elliptic:1 ] there exist a real number @xmath239 and a single - valued meromorphic function @xmath189 on @xmath171 such that @xmath240 is a @xmath124-lift of @xmath224 .
[ item : elliptic:2 ] @xmath224 has a regular singularity at @xmath241 if and only if @xmath189 has at most a pole at @xmath241 .
[ item : parabolic ] suppose that @xmath224 is parabolic and take an arbitrary positive number @xmath173 .
then , 1 . [ item : parabolic:1 ] for each @xmath242 , there exists a single - valued meromorphic function @xmath189 on @xmath171 such that @xmath243 is a @xmath124-lift of @xmath224 .
[ item : parabolic:2 ] the function @xmath189 has at most a pole at @xmath241 if and only if @xmath224 has a pole of order exactly @xmath53 at @xmath241 .
[ item : parabolic:3 ] @xmath189 is holomorphic at @xmath241 if and only if @xmath244 has at most a pole of order @xmath32 at @xmath241 .
[ item : parabolic:4 ] when @xmath189 is holomorphic at @xmath241 , @xmath245 ( resp .
@xmath246 ) holds for sufficiently small @xmath247 if and only if @xmath248 ( resp .
@xmath249 ) .
[ item : hyperbolic ] suppose that @xmath224 is hyperbolic .
then , 1 . [ item : hyperbolic:1 ] there exist a positive number @xmath239 and a single - valued meromorphic function @xmath189 on @xmath171 such that @xmath250 is a @xmath124-lift of @xmath224 .
[ item : hyperbolic:2 ] @xmath189 has at most a pole at @xmath241 if and only if @xmath224 has a pole of order exactly @xmath53 at @xmath241 . in the statements of proposition [ n ] , the function @xmath251 @xmath252 is defined by @xmath253 where @xmath254 is considered as a function defined on the universal cover @xmath213 of @xmath171 . to prove this , we consider the following ordinary differential equation @xmath255 if we assume @xmath256 has a regular singularity at @xmath241 , then @xmath257 for some @xmath258 and has the fundamental system of solutions @xmath259 where @xmath260 @xmath261 are holomorphic functions on @xmath262 such that @xmath263 @xmath261 . the constant @xmath264 is called the _ log - term coefficient _ and @xmath265 are the solutions of the indicial equation @xmath266 if @xmath267 , then @xmath268 vanishes .
( see @xcite or the appendix of @xcite ) .
the following lemma is easy to show : [ lem : ode ] in the above setting , @xmath269 if @xmath270 .
take the matrix @xmath143 as in .
we first prove the elliptic case .
since @xmath224 is elliptic , there exist a @xmath271 and an @xmath112 such that @xmath272 .
so @xmath273 , and @xmath274 is single - valued on @xmath171 , proving the first part of [ item : elliptic ] .
if the origin @xmath149 is at most a pole of @xmath191 , a direct calculation shows that @xmath224 has a regular singularity . to show the converse , we set @xmath275 , with @xmath276 as in .
then by lemma [ lem : ode ] we have @xmath269 .
the monodromy matrix @xmath277 of @xmath278 is conjugate to @xmath279 \begin{pmatrix } 1 & -2\pi i k \\ 0 & 1 \end{pmatrix } & \qquad\mbox{(if $ k\ne 0$)}. \end{cases}\ ] ] since @xmath224 is elliptic , the log - term coefficient @xmath280 and @xmath281 .
thus @xmath282 since @xmath283 , there exists a @xmath284 so that @xmath285
. then @xmath286
so @xmath287 . if @xmath288 , then @xmath289 is meromorphic , proving [ item : elliptic ] .
otherwise , @xmath290 for some @xmath291 , and [ item : elliptic ] follows from @xmath292 respectively .
next , we assume @xmath224 is parabolic and take a positive number @xmath173 and @xmath242 .
then by theorem [ thm : canonical - su ] and remark [ rem : parab - conj ] in appendix [ app : conj ] , there exists a matrix @xmath112 such that @xmath293 is one of @xmath294 , @xmath295 , @xmath296 , @xmath297 .
note that @xmath294 and @xmath296 are not conjugate in @xmath124 . replacing @xmath77 with @xmath133
if @xmath298 ( see remark [ rem : inverse ] ) , we can choose a @xmath124-lift @xmath77 such that @xmath299 then , @xmath300 . here
the @xmath301-ambiguity of @xmath293 does not affect the @xmath202-action .
thus , @xmath302 hence @xmath303 is a single - valued meromorphic function on @xmath171 , proving the first part of [ item : parabolic ] .
if @xmath189 has at most a pole at @xmath241 , then a direct computation shows that @xmath224 has a pole of order exactly @xmath53 .
therefore , it suffices to show that @xmath189 has at most a pole at @xmath241 when @xmath224 has a regular singularity .
we now show this : we set @xmath275 .
since @xmath224 is parabolic , yields that the log - term coefficient @xmath304 and @xmath305 is non - positive .
hence @xmath306 here @xmath307 is single - valued on @xmath171 and has at most a pole at @xmath241 .
take a matrix @xmath284 such that @xmath308 .
then @xmath309 replacing @xmath310 by @xmath311 and renaming @xmath312 to @xmath313 , @xmath314 becomes @xmath315 , and we have @xmath316 and there is no @xmath301-ambiguity in the above equation , as the eigenvalues of the left hand matrix must have the same sign as those of the right hand matrix .
thus we can choose @xmath317 for some @xmath318 , which proves the second part of [ item : parabolic ] .
it is easy to see that @xmath189 is holomorphic at @xmath241 if and only if @xmath319 , that is , @xmath320 in is @xmath321 , which proves the third part of [ item : parabolic ] .
assume @xmath191 is holomorphic on @xmath322 . since the mbius transformation @xmath323 maps the disk @xmath322 onto the upper - half plane @xmath324 , the condition @xmath116 ( equivalently @xmath325 ) is equivalent to @xmath326 for all @xmath112 . and
since @xmath327 is bounded , this is equivalent to @xmath328 .
thus we obtain the last part of [ item : parabolic ] .
next , we assume @xmath224 is hyperbolic . by theorem [ thm : canonical - su ] in appendix [ app : conj ] , there are a matrix @xmath112 and @xmath329 such that @xmath330 or @xmath331 .
then @xmath332 , which implies @xmath333 .
so @xmath334 is a single - valued meromorphic function in @xmath171 .
this proves the first part of [ item : hyperbolic ] . to prove the second part of [ item : hyperbolic ] , analogous to the parabolic case , we only need to prove one direction .
suppose that @xmath224 has a regular singularity .
we set @xmath275 . since @xmath224 is hyperbolic , yields that @xmath280 and @xmath335 .
thus @xmath336 exchanging @xmath337 and @xmath310 if necessary , we may assume @xmath338 without loss of generality .
take a @xmath284 such that @xmath339 .
then we have @xmath340 so @xmath341 that is , @xmath342 .
as we have assumed that @xmath180 and @xmath338 , we have @xmath343 , and then @xmath344 must be diagonal . hence @xmath345 proving the assertion .
we consider a conformal metric @xmath346 on @xmath171 of constant gaussian curvature @xmath136 , called a _ punctured hyperbolic metric_. then there exists a meromorphic function @xmath347 such that @xmath348 which is called the _ developing map _ of @xmath346 . since @xmath346 is a well - defined hyperbolic metric on @xmath171 , either @xmath117 or @xmath116 holds on @xmath171 .
we remark that , for a immersion @xmath349 , the metric @xmath346 as in is an example of a hyperbolic metric , and the secondary gauss map is a developing map of it .
the developing map @xmath77 is not unique , and the set of all developing maps of @xmath346 coincides with @xmath350 set @xmath351 where @xmath352 , that is , @xmath353 .
we call the projective connection @xmath354 the _ schwarzian derivative _ of @xmath346 .
since the metric @xmath346 is well - defined on @xmath171 , the developing map @xmath77 is a @xmath124-lift of the @xmath124-projective connection @xmath354 . if @xmath77 is a developing map of @xmath355 in , then @xmath356 if the matrix @xmath143 is _ elliptic _ ( resp .
_ parabolic _ , _ hyperbolic _ ) , the metric @xmath346 is said to have elliptic ( resp .
parabolic , hyperbolic ) monodromy .
[ def : reg - sing ] we say that a hyperbolic punctured metric @xmath346 has a _ regular singularity _ at the origin if @xmath354 has a regular singularity at the origin , that is , it has at most a pole of order @xmath53 .
[ thm : key ] any conformal hyperbolic metric on @xmath171 has a regular singularity at @xmath241 . let @xmath77 be a developing map of a conformal hyperbolic metric @xmath346 on @xmath171 .
suppose @xmath346 has elliptic monodromy . since @xmath346 has no singular points on @xmath171 , @xmath117 or @xmath116 holds on @xmath171 .
since @xmath133 is also a developing map of @xmath346 , we may assume that @xmath117 .
by proposition
[ n ] , there exists a real number @xmath239 such that @xmath357 is a single - valued function on @xmath171 .
multiplying @xmath189 by @xmath358 ( @xmath359 ) , we may assume that @xmath360 without loss of generality .
thus @xmath361 and @xmath189 has more than two exceptional values , so has at most a pole at @xmath241 , by the great picard theorem .
then by [ item : elliptic:2 ] in proposition [ n ] , @xmath226 has a regular singularity at the origin .
suppose @xmath346 has parabolic monodromy .
applying proposition [ n ] for the @xmath124-projective connection @xmath354 with @xmath249 and @xmath362 , we can take a @xmath124-lift @xmath77 such that @xmath363 is a single - valued meromorphic function on @xmath171 , where @xmath364 is the matrix in .
since @xmath346 has no singular points on @xmath171 , @xmath116 or @xmath117 holds .
in particular , because @xmath365 maps the unit disk into the upper - half plane , we have @xmath366 ( resp .
@xmath367 ) if @xmath117 ( resp .
@xmath116 ) . here
, it holds that @xmath368 thus , @xmath369 thus by the great picard theorem , there exist an integer @xmath196 and a holomorphic function @xmath370 with @xmath371 such that @xmath372 , that is , @xmath373 since @xmath189 is single - valued , @xmath196 must be @xmath149 .
therefore , @xmath189 can be extended to be holomorphic at @xmath241 , and then by [ item : parabolic:2 ] of proposition [ n ] , the origin is a regular singularity of @xmath354 . to prove the hyperbolic case , we need the following [ montel ]
if a family of holomorphic functions @xmath374 defined on a domain @xmath375 have two exceptional values in common , then they are a normal family , that is , there is a subsequence @xmath376 such that either @xmath376 or @xmath377 converges uniformly on every compact set in @xmath378 .
the proof for the hyperbolic case is parallel to the proofs of propositions 4 and 5 in @xcite .
suppose @xmath346 has hyperbolic monodromy .
again , we may assume that @xmath117 without loss of generality .
by proposition
[ n ] , again replacing @xmath77 by @xmath289 for some @xmath379 if necessary , there exists a positive real number @xmath239 such that @xmath380 is a single - valued meromorphic function on @xmath171 .
the function @xmath381 has neither zeros nor poles in @xmath171 , and @xmath366 .
we now define a set @xmath382 and analytic functions @xmath383 and @xmath384 for @xmath385 from @xmath386 to @xmath151 by @xmath387 and @xmath388 .
then @xmath389 is a family of holomorphic functions on @xmath386 . since @xmath390
, we have @xmath391 .
thus @xmath392 is a normal family by montel s theorem .
two possible cases arise .
[ [ case-1 ] ] * case 1 : * + + + + + + + + + first we consider the case that a subsequence @xmath392 converges to a holomorphic function uniformly on any compact subset of @xmath386 . since @xmath393 for a positive integer
@xmath394 is a compact subset of @xmath386 , there exist a positive number @xmath395 and an @xmath396 such that @xmath397 holds on @xmath398 for @xmath399 .
this implies that @xmath400 on @xmath401 for @xmath399 . then by the maximum principle
, we have @xmath402 for each @xmath403 .
thus we have @xmath404 on the other hand , since @xmath405 for @xmath406 and @xmath407 , the function @xmath189 is bounded in a punctured neighborhood of @xmath241 and has a removable singularity there .
[ [ case-2 ] ] * case 2 : * + + + + + + + + + next we consider the case that a subsequence @xmath408 converges to a holomorphic function @xmath23 .
then we can conclude that @xmath409 is bounded on @xmath171 . in this case
@xmath189 has at most a pole at the origin . in both cases , @xmath410 has at most a pole of order two at @xmath241 .
[ rem : d - sigma ] in corollary [ h - monodromy ] , we shall show that in fact the monodromy of @xmath346 can never be hyperbolic .
let @xmath118 be a weakly complete face of finite topology , and let @xmath154 be a compact riemann surface such that @xmath2 is diffeomorphic to @xmath411 .
[ def : reg ] a puncture - type end @xmath157 of @xmath23 is called _ regular _ if the hyperbolic gauss map @xmath22 has at most a pole at @xmath157 .
[ def : g - reg ] on the other hand , we say a puncture - type end @xmath157 is _
@xmath77-regular _ if the schwarzian derivative @xmath226 of the secondary gauss map @xmath77 has at most a pole of order @xmath53 at @xmath157 , that is , the pseudometric @xmath412 has a regular singularity at @xmath157 ( cf .
definition [ def : reg - sing ] ) .
when @xmath77 is single - valued , @xmath77-regularity implies that @xmath77 has at most a pole at the end .
when the hopf - differential has at most a pole of order @xmath53 , regularity and @xmath77-regularity are equivalent , by .
theorem [ thm : key ] can now be stated in terms of faces as follows : [ lem : key2 ] all ends of a complete face are @xmath77-regular . by proposition [ prop : complete ] , all ends are of puncture - type .
so we can set @xmath413 , where @xmath154 is a compact riemann surface .
let @xmath105 be a weierstrass data for @xmath23 .
since the singular set is compact , the metric @xmath346 as in is a punctured hyperbolic metric in a punctured neighborhood of @xmath157 .
then @xmath346 has a regular singularity by theorem [ thm : key ] , and hence @xmath23 is @xmath77-regular at @xmath157 .
[ def : labeledbyyang2 ] an elliptic end of a face is _ integral _ if the monodromy of the secondary gauss map is the identity , and _ non - integral _ otherwise .
let @xmath349 be a @xmath77-regular non - integral elliptic end .
then the singular set does not accumulate at the end @xmath149 .
one can take the secondary gauss map @xmath77 to be @xmath414 on a neighborhood of the end , where @xmath415 and @xmath189 is holomorphic at the end @xmath241 with @xmath416 .
since @xmath417 , the singular set @xmath418 can not accumulate at the origin . on the other hand , an integral elliptic end might or might not be complete : [ ex : integral - elliptic - ends ] for non - zero integers @xmath196 and @xmath419 with @xmath420 , we set @xmath421 and @xmath422 . setting @xmath423 and @xmath424
, we see that gives a riemannian metric on @xmath425 .
so using small s formula , we have a face with integral elliptic ends at @xmath426 .
the singular set is @xmath427 ( see figure [ fg : sing - e ] for the case @xmath428 ) .
thus the singular set accumulates at @xmath241 but not at @xmath429 .
thus @xmath429 is a complete integral elliptic end , but @xmath241 is an incomplete integral elliptic end . to state the behavior of an incomplete ( integral ) elliptic end , we introduce a notation : for a positive integer @xmath196 , an @xmath430 and a @xmath431 $ ] , we define the open subset ( which is a union of sectors , see figure [ fg : sing - e ] ) @xmath432 for the case @xmath428 .
the shaded parts indicate the set @xmath433 . ]
suppose @xmath434 is a @xmath77-regular integral elliptic end .
if the singular set accumulates at the end , then there are an @xmath435 and a @xmath436 such that , for any @xmath437 , there exists an @xmath438 so that the singular set of @xmath23 in @xmath439 lies in @xmath440 .
the assertion of the lemma does not depend on a choice of the complex coordinate at the origin .
since the singular set accumulates at @xmath149 , we have @xmath441 . then by proposition [ n ] [ item : elliptic ] ,
@xmath442 is holomorphic at @xmath241 .
moreover , we may set @xmath443 .
then @xmath444 is well - defined on a neighborhood of @xmath241 and @xmath445 . using the weierstrass preparation theorem
, we may further assume without loss of generality that @xmath446 for some positive integer @xmath196 .
here , @xmath447 is equivalent to @xmath448 .
thus , the singular set is expressed as @xmath449 , where @xmath450 .
[ def : first - second ] a parabolic end of a face is of the _ first kind _ if @xmath451 has at most a pole of order @xmath32 . otherwise , it is of the _
second kind_. let @xmath434 be a @xmath77-regular parabolic end .
if the end is of the first kind , the singular set does not accumulate at the end .
if the end is of the second kind , then the singular set does accumulate at the end . in this case
, there exist an @xmath435 and a @xmath452 @xmath453)$ ] such that , for all @xmath454 , there exists an @xmath438 so that the singular set of @xmath23 in @xmath455 lies in @xmath440 .
let @xmath77 be the secondary gauss map .
since the end is parabolic , the schwarzian derivative @xmath456 determines a @xmath124-projective connection of parabolic monodromy .
then by [ item : parabolic ] in proposition [ n ] for @xmath249 and @xmath362 , there exists a @xmath124 lift @xmath278 such that @xmath457 is a meromorphic function on @xmath171 . here , there exists a matrix @xmath112 such that @xmath458 or @xmath459 holds .
thus , by the @xmath9-ambiguity of the secondary gauss map , we may assume @xmath460 or @xmath461 .
moreover , replacing @xmath23 with @xmath462 if necessary ( see remark [ rem : action - null ] ) , we may assume @xmath460 without loss of generality .
since the end is @xmath77-regular , [ item : parabolic:2 ] of proposition [ n ] yields that @xmath189 is meromorphic at @xmath241 .
thus , we can write @xmath463 where @xmath464 is a holomorphic function on a neighborhood of the origin .
then there exist an @xmath465 and a @xmath466 such that @xmath467 here , the singular set @xmath418 is written as @xmath468 .
if the end is of the first kind , then @xmath469 by [ item : parabolic:3 ] in proposition
[ n ] . therefore , for each fixed @xmath470 , the right - hand side approaches @xmath471 as @xmath472 , which implies that the singular points do not accumulate at the end .
if the end is of the second kind , then @xmath473 . therefore , for each fixed @xmath470 , the right hand side approaches @xmath471 if @xmath474 and @xmath475 if @xmath476 as @xmath477 , giving solutions of @xmath478 for sufficiently small @xmath479 near the lines @xmath480 .
this implies the second assertion .
let @xmath434 be a @xmath77-regular hyperbolic end .
then any ray in @xmath171 emanating from the origin meets the singular set infinitely many times .
( see figure @xmath481 . )
this intersection property does not depend on the choice of a complex coordinate for a punctured neighborhood of the end . by proposition [ n ] and an appropriate choice of @xmath77 ,
if we set @xmath482 , then there is a @xmath483 such that @xmath484 is a meromorphic function on @xmath171 .
since @xmath23 is @xmath77-regular , [ item : hyperbolic:2 ] of proposition [ n ] implies @xmath191 has at most a pole at the origin , and @xmath226 has a pole of order exactly @xmath53 at @xmath241 .
thus we can rewrite @xmath485 where @xmath370 is a single - valued holomorphic function on @xmath486 .
now , we set @xmath487 which gives a new coordinate @xmath488 around the end , now at @xmath489 .
then @xmath490 .
since @xmath491 , setting @xmath492 , the singular set is @xmath493 that is , @xmath494 , @xmath495 , which is a log - spiral when @xmath496 .
if @xmath497 , the singular set is a union of infinitely many disjoint circles . in any case , the singular set meets any ray based at @xmath489 infinitely many times .
( see figure @xmath481 , left - hand side for the case @xmath497 and right - hand side for the case @xmath496 . ) [ cols="^,^ " , ] [ h - monodromy ] the monodromy of a hyperbolic metric on @xmath171 is either elliptic or parabolic .
that is , hyperbolic monodromy never occurs .
suppose that a hyperbolic metric @xmath346 on @xmath171 has hyperbolic monodromy .
let @xmath77 be a developing map for @xmath346 .
the data @xmath498 produces an @xmath95 as in , and then the immersion @xmath499 is without singularities , since @xmath346 is nonsingular . by theorem [ thm :
key ] , @xmath23 is @xmath77-regular at @xmath241 .
then by lemma h , the singularities accumulate at the end , a contradiction .
lemmas e1 , e2 , p and h imply : a @xmath77-regular end @xmath434 of a face is hyperbolic if and only if every ray in @xmath171 emanating from the origin meets the singular set infinitely many times .
we now give two theorems on complete faces .
[ thm : completeend ] any complete end of a face is either @xmath77-regular elliptic or @xmath77-regular parabolic of the first kind . by proposition [ prop : complete ] ,
the end is of puncture - type .
moreover , theorem [ thm : key ] implies the end is @xmath77-regular .
thus the theorem follows from corollary [ h - monodromy ] and lemma p. [ thm : labeledbyyang3 ] any complete face is of finite type .
@xmath500the definition of finite type is given in definition [ def : ft].@xmath501 let @xmath67 be a complete face . then by proposition [ prop : complete ] , there is a compact riemann surface @xmath18 such that @xmath25 is biholomorphic to @xmath502 .
we fix any end @xmath157 , and take a small coordinate neighborhood @xmath188 with @xmath241 at @xmath157 .
we may assume that there are no singular points on @xmath503 , and thus we may also assume that @xmath117 on @xmath503 for a secondary gauss map @xmath77 . we know from theorem [ thm : completeend ] that the end is a @xmath77-regular elliptic end or a @xmath77-regular parabolic end of the first kind .
first , we consider the elliptic case .
since @xmath117 , we may assume that there exist some @xmath504 and a holomorphic function @xmath189 on @xmath61 with @xmath505 such that @xmath506 .
if @xmath441 , then @xmath507 and @xmath508 for some @xmath509 , @xmath510 and @xmath511 , which contradicts the fact that @xmath117 on @xmath512 . hence @xmath513
therefore , there exist a neighborhood @xmath514 of @xmath157 and an @xmath454 such that @xmath515 on @xmath516 .
so , on @xmath516 , @xmath517 where @xmath160 is the metric as in , which is defined on @xmath518 because @xmath519 . since @xmath157 is a regular singularity of the punctured spherical metric @xmath520 @xmath521 has finite area , so @xmath64 has finite total curvature on @xmath518 , by .
next we consider the parabolic case . by theorem [ thm : completeend ] , the end is parabolic of the first kind . then without loss of generality
, we may assume there exists a holomorphic function @xmath189 on @xmath322 such that ( we set @xmath362 and replace @xmath189 by @xmath522 in the proof of proposition [ n ] ) @xmath523 if we set @xmath524 , we have @xmath525 , @xmath526 and @xmath527
so we have @xmath528 we set @xmath529 and @xmath530 .
since @xmath531 as @xmath532 , we may assume @xmath533 .
then @xmath534 @xmath535 where @xmath536 .
since @xmath537 the area of a sufficiently small punctured neighborhood of @xmath241 with respect to @xmath346 is finite , which proves the assertion .
theorem [ thm : main ] in the introduction follows from theorems [ thm : completeend ] and [ thm : labeledbyyang3 ] .
let @xmath538 be the light - cone of @xmath28 , with future and past light cones @xmath539 the multiplicative group @xmath540 of the positive real numbers acts on @xmath541 by scalar multiplication .
the _ ideal boundary _ @xmath542 of @xmath0 consists of two ( future and past ) components @xmath543 each of which are identified with @xmath544 by the projection @xmath545\longmapsto \frac{1}{v_0-v_3}(v_1+i v_2)\in { \boldsymbol{c}}\cup\{\infty\}.\ ] ] the isometries of @xmath28 induce mbius transformations on @xmath544 .
the boundary @xmath542 is identified with the set of equivalence classes of oriented time - like geodesics in @xmath0 .
in particular , for a space - like immersion @xmath118 with the ( time - like ) unit normal vector field @xmath74 , the equivalence class @xmath546 $ ] determines a point in @xmath542 for each @xmath66 .
hence we have the _ light - cone gauss map _ @xmath547\colon{}m^2 \longrightarrow \partial s^3_1.\ ] ] let @xmath118 be a face , and @xmath66 a regular point , that is , @xmath23 is an immersion in a neighborhood of @xmath58 . under the identification of @xmath28 and @xmath548 as in
, we can compute that the unit normal vector @xmath74 is @xmath549 where @xmath95 is the holomorphic lift of @xmath23 and @xmath77 is the secondary gauss map . hence @xmath550 = { \operatorname{sgn}}(|g|^2 - 1)\!\ ! \left [ f \begin{pmatrix } |g|^2 & g \\
\bar g & 1 \end{pmatrix } f^ * \right],\ ] ] where @xmath551 is the sign of the function @xmath552 .
thus , we have : [ prop : lgm ] the light - cone gauss map @xmath553 of a face takes values in @xmath554 ( resp .
@xmath555 ) if @xmath116 ( resp .
@xmath117 ) .
moreover , its projection @xmath556 is the hyperbolic gauss map @xmath22 as in , which extends to the singular set . by , the @xmath557-component of @xmath558 is @xmath559 where @xmath560 . here , @xmath561 holds because @xmath562 , implying the first part of the proposition . by , , and , we have @xmath563= \frac{gf_{11}+f_{12}}{gf_{21}+f_{22 } } = \frac{df_{11}}{df_{21}}=g,\ ] ] and this completes the proof .
next we give a criterion for when a complete regular end approaches @xmath554 or @xmath564 : [ prop : fatend ] let @xmath434 be a complete regular end at @xmath241 and let @xmath77 be a secondary gauss map of @xmath23 .
then the image of @xmath23 converges to a point in @xmath554 ( resp .
@xmath564 ) at the end if and only if @xmath117 ( resp .
@xmath116 ) near the end . we can change the holomorphic null lift @xmath95 to @xmath132 as in , so that @xmath23 and its secondary gauss map @xmath77 change to @xmath462 and @xmath133 .
the end of @xmath23 approaches @xmath565 if and only if @xmath462 approaches @xmath566 , so it is sufficient to prove this result under the assumption @xmath117 on @xmath171 . by theorem [ thm : completeend ] , the end is either elliptic or parabolic .
first we assume the end is elliptic . replacing @xmath95 by @xmath567 ( @xmath568 , @xmath569 ) and using the weierstrass preparation theorem if necessary
, we may assume without loss of generality that the hyperbolic and secondary gauss maps are @xmath570 where @xmath191 is a holomorphic function on @xmath322 with @xmath416 .
here @xmath338 because @xmath117 .
if @xmath571 , small s formula implies that @xmath572 since @xmath196 and @xmath239 are positive , the first component @xmath557 is @xmath573 where @xmath450
. other components of @xmath574 are expressed as @xmath575 we now consider the stereographic projection given in @xcite : @xmath576 which is a diffeomorphism . then @xmath577 is expressed as @xmath578 thus , @xmath577 approaches @xmath579 . when @xmath580 , by again , @xmath581 , @xmath582 and @xmath583 are bounded on a neighborhood of @xmath149 , and these components can be extended to become holomorphic on a neighborhood of @xmath149 . if @xmath584 is bounded , @xmath95 must be holomorphic and then the induced metric is bounded , which contradicts the weak completeness of the end .
hence @xmath584 has a pole at @xmath149 .
so the @xmath557-component of @xmath23 is @xmath585 moreover , since @xmath586 we have @xmath587 as @xmath588 .
next we assume the end is parabolic .
again we may set @xmath589 , @xmath511 .
applying proposition [ n ] for @xmath590 and @xmath249 , there exists a @xmath124-lift @xmath278 of @xmath226 such that @xmath591 is a single - valued meromorphic function on @xmath171 , and the secondary gauss map @xmath77 satisfies @xmath592 or @xmath459 for some @xmath112 . by completeness ,
lemma p implies that the end is of first kind .
hence by [ item : parabolic:3 ] of proposition [ n ] , @xmath189 is holomorphic on a neighborhood of the origin .
moreover , by the assumption @xmath117 , [ item : parabolic:4 ] of proposition [ n ] yields that @xmath592 for some @xmath112 .
thus , without loss of generality , we may set @xmath593 here we set @xmath594 . for a holomorphic null lift @xmath95 of @xmath23 with the secondary gauss map @xmath77 , set @xmath595
then @xmath596 is a holomorphic null immersion whose hyperbolic gauss map @xmath597 and secondary gauss map @xmath598 are given by @xmath599 so applying for this @xmath600 , the components of @xmath596 are written as @xmath601 where @xmath602 , @xmath603 , @xmath604 and @xmath605 are holomorphic functions defined on a neighborhood of the origin such that @xmath606 since @xmath607 , @xmath608 satisfies @xmath609 hence the components of @xmath23 are expressed as @xmath610 where @xmath611 . here , @xmath612
( @xmath613 ) and @xmath614 ( @xmath613 ) ( resp . @xmath615 and @xmath616 ) are real - valued ( resp .
complex - valued ) differentiable functions defined on a neighborhood of the origin , such that @xmath617 ( @xmath618 ) .
the equations yield that @xmath619 and @xmath620 as @xmath588 .
here we prove theorem [ thm : fujimori ] stated in the introduction .
first we prepare : [ lem : complete - p ] the hopf differential of a face has a pole of order @xmath53 at any complete regular parabolic end .
let @xmath434 be a complete regular parabolic end at @xmath241 . by theorem [ thm : completeend ] ,
the end is of the first kind .
then @xmath621 since @xmath22 is meromorphic at @xmath241 , we may assume that @xmath622 , where @xmath196 is a positive integer and @xmath464 is a holomorphic function on a neighborhood of @xmath149 with @xmath623 .
applying to @xmath624 , it follows that @xmath87 has a pole of order @xmath53 at @xmath241 .
the next lemma improves a result in ( * ? ? ?
* proposition 4.4 ) : [ lem : osserman - ineq ] let @xmath434 be a complete regular end at @xmath241 of a face with hopf differential @xmath87 and hyperbolic gauss map @xmath22
. then the ramification order @xmath196 of @xmath625 at @xmath241 satisfies @xmath626 ( for the definition of the ramification order , see the subsection about the schwarzian derivative in section [ sec : prelim ] . ) here , @xmath627 denotes the order of @xmath87 at the origin , that is , @xmath628 if @xmath629 , where @xmath464 is holomorphic at @xmath241 and @xmath623 . by theorem [ thm : completeend ] , a complete end is either elliptic or parabolic of the first kind .
the elliptic case has been proved in @xcite .
assume that the end is parabolic .
then by lemma [ lem : complete - p ] , @xmath87 must have a pole of order @xmath53 at @xmath241 , which proves the inequality since @xmath630 .
it should be remarked that the order of the metric @xmath631 at @xmath149 is equal to @xmath632 , where @xmath196 is the ramification order of @xmath22 .
using lemma [ lem : osserman - ineq ] instead of ( * ? ? ?
* proposition 4.4 ) , the inequality in theorem [ thm : fujimori ] is proved in the same way as @xcite , @xcite .
the condition for equality in in theorem [ thm : fujimori ] for elliptic ends was completely analyzed in @xcite .
so , it suffices to show the following theorem for parabolic ends .
note that @xmath633 for complete regular parabolic ends , hence the equality in holds if and only if @xmath22 does not branch at @xmath58 ( see @xcite for details ) . [ thm : equality ] a complete regular parabolic end of a face
is properly embedded if and only if the hyperbolic gauss map @xmath22 does not branch at the end .
let @xmath434 be a complete regular parabolic end at @xmath241 .
taking @xmath462 instead of @xmath23 if necessary , we may assume that @xmath117 in a neighborhood of the end , and that @xmath634 and @xmath635 , as in the proof of proposition [ prop : fatend ] .
then @xmath23 is represented as in and . by proposition [ prop : fatend ]
, the image of @xmath23 tends to a point in @xmath636 .
so we may assume that @xmath637 on @xmath171 , and @xmath638 is well - defined , where @xmath639 is the projection in . here , by , @xmath640
since @xmath641 and @xmath642 are differentiable functions defined on a neighborhood of @xmath149 , we have @xmath643 now we suppose that the ramification order @xmath196 of the hyperbolic gauss map at @xmath241 is @xmath32 , that is @xmath644 .
as seen in the proof of proposition [ prop : fatend ] , @xmath577 converges to @xmath645 .
then @xmath646 and yield that @xmath647 which implies that the correspondence @xmath648 is bijective near the origin , and the end is properly embedded .
conversely , suppose that the end is properly embedded .
we have already seen that @xmath649 as @xmath532 .
moreover @xmath650 implies that for any sufficiently small @xmath437 , the image of the end @xmath651 does not meet the @xmath652-axis and is diffeomorphic to a cylinder .
then the image of @xmath653 by the orthogonal projection @xmath654 is an embedded closed curve with the winding number @xmath196 with respect to the origin .
so @xmath644 . in proposition 4.4 of @xcite
, the first author showed the equality condition in theorem [ thm : fujimori ] for elliptic ends using the expression of the solution of the ordinary differential equation .
here we proved the equality condition in theorem [ thm : fujimori ] for parabolic ends by using small s formula .
it is also possible to prove the result in @xcite more directly by using .
we give here three important examples : we set @xmath655 and @xmath656 then gives a face @xmath67 with hyperbolic and secondary gauss maps @xmath22 and @xmath77 , and hopf differential @xmath657 since the lift metric @xmath658 is complete on @xmath2 , @xmath23 is weakly complete .
the end @xmath429 is complete and elliptic , and @xmath659 are parabolic ends of the second kind . hence @xmath659 are incomplete ends . since @xmath660 ,
@xmath23 does not satisfy .
this implies that completeness is an essential assumption in theorem [ thm : fujimori ] in the introduction .
we set @xmath661 then @xmath662 has two parabolic regular ends .
the hyperbolic gauss map @xmath22 , the secondary gauss map @xmath77 and the hopf differential @xmath87 are computed as follows : @xmath663 since @xmath664 , the singular set is compact , and hence @xmath23 is complete .
any genus zero face with two parabolic regular ends and with degree @xmath32 hyperbolic gauss map is congruent to this @xmath23 .
we call this face the
_ parabolic catenoid_. on the other hand , the face with @xmath665 , @xmath666 ( @xmath667 ) given in ( * ? ? ?
* example 5.4 ) is called the _
elliptic catenoid_. [ ex : elliptic - integral ] ( a complete 4-noid with 4 integral elliptic ends satisfying the equality in ) since @xmath3 can be identified with the complex hyperquadric @xmath668 of @xmath669 , the null ( meromorphic ) curves in @xmath3 can be identified with those in @xmath668 .
the null curve in @xmath3 with @xmath670 belongs to the moduli space @xmath671 in the classification list of null curves in @xmath668 in bryant @xcite , which has four integral elliptic ends at the roots of @xmath672 and @xmath429 .
since @xmath22 is of degree @xmath6 and @xmath673 , the corresponding face attains equality in of theorem [ thm : fujimori ] .
( for the definition of an integral elliptic end , see definition [ def : labeledbyyang2 ] . )
we can deform an elliptic catenoid to a parabolic catenoid .
let @xmath674 be an elliptic catenoid with the hyperbolic gauss map @xmath665 and the secondary gauss map @xmath675 , where @xmath338 .
then the hyperbolic metric corresponding to @xmath674 is @xmath676 it can be easily checked that @xmath677 which is the hyperbolic metric of a parabolic catenoid with @xmath678 see . on the other hand , by small s formula
there exists a unique smooth @xmath32-parameter family of faces @xmath679 ( @xmath680 ) with hyperbolic gauss map @xmath665 and associated hyperbolic metric @xmath681 .
then @xmath679 is congruent to @xmath674 , and @xmath682 gives a parabolic catenoid . as a consequence of remark [ rem : complete ] ,
we know that there are no compact immersed surfaces in @xmath0 . here
we give an alternative proof of this : let @xmath2 be a compact riemann surface without boundary , and suppose there exists a compact face @xmath67 which has no singular points .
let @xmath95 be a holomorphic lift of @xmath23 .
we may assume that @xmath117 since there are no singular points .
then , by , we have @xmath683 where @xmath5 is a local complex coordinate and @xmath684 .
thus , @xmath685 is a nonconstant subharmonic function , which is a contradiction to the maximum principle .
this proof does not apply to compact faces , leading us to the following open problem : is there a compact face ?
if such a face exists , the genus @xmath167 must be greater than or equal to @xmath6 , since equality in in the introduction holds in this case and the degree of the hyperbolic gauss map must be @xmath686 .
in this appendix , we shall give a proof of the following [ thm : q - mero ] let @xmath18 be a compact riemann surface .
then the hopf differential @xmath87 of a complete face @xmath687 is meromorphic on @xmath154 .
it is sufficient to show the meromorphicity of @xmath87 at a complete end @xmath688 at the origin .
we write the hopf differential @xmath87 as @xmath689 where @xmath690 is a holomorphic function on @xmath171 . by theorem [ thm :
completeend ] , a complete end @xmath349 is either a @xmath77-regular elliptic end or a @xmath77-regular parabolic end of the first kind . (
the definition of @xmath77-regularity is given in definition [ def : g - reg ] . ) first , we consider the case that @xmath23 is elliptic . by @xmath77-regularity
, the secondary gauss map is written in the form @xmath691 where @xmath239 is a real number .
since @xmath692 by completeness , we may set @xmath693 , or @xmath471 , because of the @xmath9-ambiguity of @xmath77 . moreover , replacing @xmath23 by @xmath462 if necessary , we may assume @xmath338 without loss of generality . in this case ,
the corresponding hyperbolic metric @xmath346 is written as @xmath694 since @xmath695 and @xmath696 tend to @xmath149 as @xmath588 and @xmath189 is bounded near the origin , we have that @xmath697 holds on a neighborhood of the origin , where @xmath698 is the smallest integer such that @xmath699 and @xmath700 is a positive constant .
then , by , we have @xmath701 since @xmath702 is complete at @xmath149 , we have meromorphicity of the one - form @xmath703 at the origin , because of ( * ? ? ?
* lemma 9.6 , page 83 ) .
next , we consider the case that @xmath23 is parabolic . since the end is @xmath77-regular parabolic of the first kind
, one can choose the secondary gauss map @xmath77 as in : @xmath704 where @xmath189 is a holomorphic function on @xmath705 .
hence @xmath346 is written as in : @xmath706 since @xmath707 as @xmath588 and @xmath191 is bounded on a neighborhood of the origin , @xmath708 holds on a neighborhood of the origin , where @xmath700 is a positive constant .
thus , @xmath709 hence , by the same argument as in the elliptic case , we have meromorphicity of @xmath87 at the origin .
the lie group @xmath9 is the set of matrices @xmath710 satisfying @xmath711 . two matrices @xmath712 are called _ conjugate in _
@xmath3 if there exists a matrix @xmath713 such that @xmath714 , and are called _ conjugate in _
@xmath9 if @xmath714 for some @xmath715 .
[ rem : conj - neg ] though the matrices @xmath718 and @xmath722 are conjugate in @xmath3 , they are not conjugate in @xmath9 if @xmath723 .
that is , for any elliptic matrix @xmath112 , there exists a _
unique _ real number @xmath724 $ ] such that @xmath725 and @xmath718 are conjugate in @xmath9 .
[ rem : parab - conj ] on the other hand , @xmath181 and @xmath182 ( @xmath726 ) are conjugate in @xmath9 if and only if @xmath183 .
in fact , if @xmath727 , @xmath728 ( @xmath715 ) holds if and only if @xmath729 where @xmath730 and @xmath731 . in particular , the sign of @xmath173 in @xmath732 is invariant under such a conjugation .
though @xmath732 ( @xmath178 ) is conjugate with @xmath733 or @xmath734 , we choose various values of @xmath173 in this paper for the sake of convenience . by assumption , there is a @xmath744 with @xmath745 .
we set @xmath746 , for real matrices @xmath61 , @xmath747 .
then @xmath748 , @xmath749 and @xmath750 if @xmath751 vanishes identically for all @xmath271 , holomorphicity of @xmath752 yields that @xmath753 , a contradiction .
thus for some @xmath754 , @xmath755 , and then @xmath756 . if @xmath757 , we set @xmath758 , giving @xmath759 . if @xmath760 , we set @xmath761 , giving @xmath762 . if the eigenvalues of @xmath725 are not real numbers , they are written as @xmath764 , where @xmath765 . in this case
, @xmath766 is conjugate in @xmath767 to @xmath768 .
hence by lemma [ key - l ] , @xmath766 is conjugate in @xmath743 to @xmath769 or @xmath770 .
thus , @xmath771 is conjugate in @xmath9 to @xmath772 or @xmath773 . if the eigenvalues of @xmath725 are @xmath774 ( @xmath775 ) and @xmath776 , @xmath766 is conjugate in @xmath767 to @xmath777 for any @xmath778 .
hence @xmath766 is conjugate in @xmath743 to either @xmath779 or @xmath780 .
thus , @xmath725 is conjugate in @xmath9 to @xmath781 or @xmath782 .
as mentioned in remark [ rem : parab - conj ] , @xmath783 for @xmath784 is conjugate in @xmath9 to @xmath785 . if the eigenvalues of @xmath725 are two distinct real numbers , they are represented as @xmath786 , where @xmath778 and @xmath242 .
thus , @xmath766 is conjugate in @xmath3 to the diagonal matrix @xmath787 .
hence , similarly to the first case , @xmath725 is conjugate in @xmath9 to @xmath788 .
krsuy r. aiyama and k. akutagawa , _ kenmotsu - bryant type representation formulas for constant mean curvature surfaces in @xmath789 and @xmath790 _ , ann .
global anal .
( 1 ) , * 17 * ( 1998 ) , 4975 .
k. akutagawa , _ on spacelike hypersurfaces with constant mean curvature in the de sitter space _ , math .
z. , * 196 * ( 1987 ) , 1319 .
r. bryant , _ surfaces of mean curvature one in hyperbolic space _ ,
astrisque , * 154155 * ( 1987 ) , 321347 . , _ surfaces in conformal geometry _ , proceedings of symposia in pure mathematics , * 48 * ( 1988 ) , 227240 .
e. a. coddington and n. levinson , theory of ordinary differential equations , mcgraw - hill , 1955 .
p. collin , l. hauswirth and h. rosenberg , _ the geometry of finite topology bryant surfaces _ , ann . of math .
( 2 ) * 153 * ( 2001 ) , 623659 .
s. fujimori , _ spacelike cmc @xmath32 surfaces with elliptic ends in de sitter @xmath6-space _ , hokkaido math .
j. , * 35 * ( 2006 ) , 289320 .
s. fujimori , w. rossman , m. umehara , k. yamada and s .- d .
yang , _ new maximal surfaces in minkowski 3-space with arbitrary genus and their cmc-1 cousins in de sitter @xmath6-space _ , preprint .
s. fujimori , k. saji , m. umehara and k. yamada , _ singularities of maximal surfaces _ , math . z.
* 259 * ( 2008 ) , 827848 .
j. glvez , a. martnez and f. miln , _ complete linear weingarten surfaces of bryant type .
a plateau problem at infinity _ , trans .
* 356 * ( 2004 ) , 34053428 .
y. w. kim and s .- d .
yang , _ a family of maximal surfaces in lorentz - minkowski three - space _ , proc .
* 134 * ( 2006 ) , 33793390 .
o. kobayashi , _
maximal surfaces in the 3-dimensional minkowski space @xmath791 _ , tokyo j. math .
* 6 * ( 1983 ) , 297309 . m. kokubu , w. rossman , k. saji , m. umehara and k. yamada , _ singularities of flat fronts in hyperbolic @xmath6-space _ , pacific j. math . , * 221 * ( 2005 ) , 303351 .
m. kokubu and m. umehara , _ global properties of linear weingarten surfaces of bryant type in @xmath10 _ , preprint .
m. kokubu , m. umehara and k. yamada , _ an elementary proof of small s formula for null curves in @xmath51 and an analogue for legendrian curves in @xmath51 _ , osaka j. math . ,
* 40 * ( 2003 ) , 697715 . ,
_ flat fronts in hyperbolic @xmath6-space _ , pacific j. math .
, * 216 * ( 2004 ) , 149175 . s. lee and s .- d .
yang , _ spacelike constant mean curvature @xmath32 trinoids in de sitter three - space _ ,
osaka j. math .
, * 43 * ( 2006 ) , 641663 .
p. li , _ complete surfaces of at most quadratic area growth _ , comment .
, * 72 * ( 1997 ) , 6771 . l. lima and p. roitman , _ constant mean curvature one surfaces in hyperbolic @xmath6-space using the bianchi - cal method _ ,
* 74 * ( 2002 ) , 1924 .
h. l. liu and g. l. liu , _ weingarten rotation surfaces in 3-dimensional de sitter space _
, j. of geometry , * 79 * ( 2004 ) , 156168 .
r. osserman , a survey of minimal surfaces , dover publications , 1986 .
j. ramanathan , _ complete spacelike hypersurfaces of constant mean curvature in de sitter space _ , indiana univ . math . j. ( 2 ) , * 36 * ( 1987 ) , 349359 . w. rossman , m. umehara and k. yamada , _ irreducible constant mean curvature 1 surfaces in hyperbolic space with positive genus _ , thoku math .
j. * 49 * ( 1997 ) , 449484 . , _
mean curvature @xmath32 surfaces in hyperbolic @xmath6-space with low total curvature i _ , hiroshima math .
j. , * 34 * ( 2004 ) , 2156 .
k. saji , m. umehara and k. yamada , _ the geometry of fronts _ , ann . of math . ,
* 169 * ( 2009 ) , 491529 .
a. j. small , _ surface of constant mean curvature in @xmath10 and algebraic curves on a quadric _ , proc .
* 122 * ( 1994 ) , 12111220 .
m. troyanov , _ metric of constant curvature on a sphere with two conical singularities _ , in `` differential geometry'' , lect .
notes in math .
* 1410 * ( 1989 ) , pp .
296308 . m. umehara and k. yamada , _ complete surfaces of constant mean curvature one in the hyperbolic @xmath6-space _ , ann . of math .
* 137 * ( 1993 ) , 611638 . ,
ann . , * 304 * ( 1996 ) , 203224 . , _ a duality on cmc-@xmath32 surfaces in hyperbolic space , and a hyperbolic analogue of the osserman inequality ,
_ tsukuba j. math .
, * 21 * ( 1997 ) , 229237 . ,
_ metrics of constant curvature one with three conical singularities on the @xmath53-sphere _ , illionois j. math .
, * 44 * ( 2000 ) , 7294 . , _
maximal surfaces with singularities in minkowski space _ , hokkaido math .
j. , * 35 * ( 2006 ) , 1340 .
yu , _ value distribution of hyperbolic gauss map _ , proc .
soc . , * 125 * ( 1997 ) , 29973001 . | the first author studied spacelike constant mean curvature one ( ) surfaces in de sitter 3-space @xmath0 when the surfaces have no singularities except within some compact subset and are of finite total curvature on the complement of this compact subset .
however , there are many surfaces whose singular sets are not compact . in fact , such examples have already appeared in the construction of trinoids given by lee and the last author via hypergeometric functions . in this paper , we improve the osserman - type inequality given by the first author .
moreover , we shall develop a fundamental framework that allows the singular set to be non - compact , and then will use it to investigate the global behavior of surfaces . |
the critical behavior of systems with quenched disorder is of considerable theoretical and experimental interest .
a typical example is obtained by mixing an ( anti)-ferromagnetic material with a non - magnetic one , obtaining the so - called dilute magnets .
these materials are usually described in terms of a lattice short - range hamiltonian of the form @xmath14 where @xmath15 is an @xmath1-component spin and the sum is extended over all nearest - neighbor sites .
the quantities @xmath16 are uncorrelated random variables , which are equal to one with probability @xmath17 ( the spin concentration ) and zero with probability @xmath18 ( the impurity concentration ) .
the pure system corresponds to @xmath19 .
one considers quenched disorder , since the relaxation time associated to the diffusion of the impurities is much larger than all other typical time scales , so that , for all practical purposes , one can consider the position of the impurities fixed . for sufficiently low spin dilution @xmath18 , i.e. as long as one is above the percolation threshold of the magnetic atoms .
the system described by the hamiltonian @xmath20 undergoes a second - order phase transition at @xmath21 ( see e.g. ref .
@xcite for a review ) .
the relevant question in the study of this class of systems is the effect of the disorder on the critical behavior . the harris criterion @xcite states that the addition of impurities to a system which undergoes a second - order phase transition does not change the critical behavior if the specific - heat critical exponent @xmath22 of the pure system is negative . if @xmath22 is positive , the transition is altered
indeed , in disordered systems the exponent @xmath23 satisfies the inequality @xmath24 @xcite this fact has been questioned however in refs .
@xcite and therefore , by hyperscaling , @xmath25 is negative .
thus , if @xmath22 is positive , @xmath25 differs from @xmath22 , so that the pure system and the dilute one have a different critical behavior . in pure @xmath1-vector models with @xmath26 , the specific - heat exponent @xmath22 is negative ; therefore , according to the harris criterion , no change in the critical asymptotic behavior is expected in the presence of weak quenched disorder .
this means that in these systems disorder leads only to irrelevant scaling corrections .
three - dimensional ising systems are more interesting , since @xmath22 is positive . in this case
, the presence of quenched impurities leads to a new universality class . theoretical investigations , using approaches based on the renormalization group @xcite , and numerical monte carlo simulations @xcite , support the existence of a new random ising fixed point describing the critical behavior along the @xmath27 line : the critical exponents are dilution independent ( for sufficiently low dilution ) and different from those of the pure ising model . experiments confirm this picture .
cristalline mixtures of an ising - like uniaxial antiferromagnet with short - range interactions ( e.g. fef@xmath28 , mnf@xmath28 ) with a nonmagnetic material ( e.g. znf@xmath28 ) provide a typical realization of the random ising model ( rim ) ( see e.g. refs .
some experimental results are reported in table [ experiments ] . this is not a complete list , but it gives an overview of the experimental state of the art .
other experimental results can be found in refs .
the experimental estimates are definitely different from the values of the critical exponents for pure ising systems , which are ( see ref . @xcite and references therein ) @xmath29 , @xmath30 , @xmath31 , and @xmath32 .
moreover , they appear to be independent of the concentration .
we mention that in the presence of an external magnetic field along the uniaxial direction , dilute ising systems present a different critical behavior , equivalent to that of the random - field ising model @xcite .
this is also the object of intensive theoretical and experimental investigations ( see e.g. refs .
@xcite ) .
several experiments also tested the effect of disorder on the @xmath33-transition of @xmath34he that belongs to the @xmath35 universality class , corresponding to @xmath36 @xcite .
they studied the critical behaviour of @xmath34he completely filling the pores of porous gold or vycor glass .
the results indicate that the transition is in the same universality class of the @xmath33-transition of the pure system in agreement with the harris criterion @xcite .
the final results for some values of @xmath1 are reported in table [ esponenteomega_mvettore ] , together with estimates of the theoretical prediction @xmath37 . for @xmath38
these results clearly indicate that the @xmath3-symmetric point is stable .
the results are somewhat lower than the theoretical prediction , especially if we consider the high - temperature estimates of the critical exponents of ref .
this is not surprising : indeed the estimates of the subleading exponents @xmath39 show in many cases discrepancies with estimates obtained by using other methods .
this is probably connected to the non - analyticity of the @xmath40-function at the fixed point @xcite . a similar discrepancy ,
although still well within a combined error bar , is observed for @xmath36 . in this case , we obtain @xmath41 , indicating that the fixed point is stable .
the error however does not allow to exclude the opposite case .
a. b. harris , j. phys .
c * 7 * , 1671 ( 1974 ) .
j. t. chayes , l. chayes , d. s. fisher , and t. spenser , phys .
lett . * 57 * , 2999 ( 1986 ) .
a. aharony , a. b. harris , and s. wiseman , phys .
81 * , 252 ( 1998 ) .
r. r. p. singh and m. e. fisher , phys .
lett . * 60 * , 548 ( 1988 ) .
f. pzmndi , r. t. scalettar , and g. t. zimnyi , phys .
* 79 * , 5130 ( 1997 ) .
a. b. harris and t. c. lubensky , phys .
* 33 * , 1540 ( 1974 ) . v. j. emery , phys . rev .
b * 11 * , 239 ( 1975 ) .
t. c. lubensky , phys .
b * 11 * , 3573 ( 1975 )
. d. e. khmelnitskii , zh .
* 68 * , 1960 ( 1975 ) [ sov .
jetp * 41 * , 981 ( 1975 ) ] .
a. aharony , y. imry , and s. k. ma , phys .
b * 13 * , 466 ( 1976 ) .
g. grinstein and a. luther , phys .
b * 13 * , 1329 ( 1976 ) .
a. aharony , in _ phase transitions and critical phenomena _ , edited by c. domb and j. lebowitz ( academic press , new york , 1976 ) , vol . 6 , p. 357 .
g. grinstein , s. k. ma , and g. f. mazenko , phys .
b * 15 * , 258 ( 1977 ) . c. jayaprakash and h. j. katz , phys .
b * 16 * , 3987 ( 1977 ) .
b. n. shalaev , zh .
* 73 * , 2301 ( 1977 ) [ sov .
jetp * 46 * , 1204 ( 1977 ) ] .
k. e. newman and e. k. riedel , phys .
b * 25 * , 264 ( 1982 ) .
j. jug , phys .
b * 27 * , 609 ( 1983 ) .
s. a. newlove , j. phys .
c * 16 * , l423 ( 1983 ) . i. o. mayer and a. i. sokolov , fiz .
tela * 26 * , 3454 ( 1984 ) [ sov .
solid state * 26 * , 2076 ( 1984 ) ] .
r. a. pelcovits and a. aharony , phys .
b * 31 * , 350 ( 1985 ) . k. de bell and d. j. w. geldart , phys .
b * 32 * , 4763 ( 1985 ) . n. a. shpot , phys .
a * 133 * , 125 ( 1988 ) ; a * 142 * , 474 ( 1989 ) .
i. o. mayer , a. i. sokolov , and b. n. shalaev , ferroelectrics * 95 * , 93 ( 1989 ) .
i. o. mayer , j. phys .
a * 22 * , 2815 ( 1989 ) .
n. a. shpot , zk .
fiz . * 98 * , 1762 ( 1990 ) [ sov .
jetp * 71 * , 989 ( 1990 ) ] . yu .
holovatch and m. shpot , j. stat . phys .
* 66 * , 867 ( 1992 ) . c. bervillier and m. shpot , phys .
b * 46 * , 955 ( 1992 ) .
h. k. janssen , k. oerding , and e. sengespeick , j. phys .
a * 28 * , 6073 ( 1995 ) .
b. n. shalaev , s. a. antonenko , and a. i. sokolov , phys .
a * 230 * , 105 ( 1997 ) .
r. folk , yu . holovatch , and t. yavorskii , j. phys
* 2 * , 213 ( 1998 ) .
holovatch and t. yavorskii , j. stat .
* 92 * , 785 ( 1998 ) .
i. o. mayer , physica a * 252 * , 450 ( 1998 ) .
r. folk , yu . holovatch , and t. yavorskii , pisma v zh .
teor . fiz . *
69 * , 698 ( 1999 ) [ jetp lett .
* 69 * , 747 ( 1999 ) ] .
k. b. varnashev , `` stability of a cubic fixed point in three dimensions .
critical exponents for generic n '' , e - print cond - mat/9909087 , phys .
b * 61 * , ( 2000 ) , in press . r. folk , yu . holovatch , and t. yavorskii , `` effective and asymptotic critical exponents of weakly diluted quenched ising model : 3d approach versus @xmath42-expansion . '' , e - print cond - mat/9909121 . d. v. pakhnin and a. i. sokolov , `` five - loop renormalization - group expansions for the three - dimensional @xmath43-vector cubic model and critical exponents for impure ising systems '' , e - print cond - mat/9912071
. d. p. landau , phys . rev .
b * 22 * , 2450 ( 1980 ) .
j. marro , a. labarta , and j. tejada , phys .
b * 34 * , 347 ( 1986 ) .
d. chowdhury and d. stauffer , j. stat .
* 44 * , 203 ( 1986 ) .
braun and m. fhnle , j. stat . phys . * 52 * , 775 ( 1988 ) .
p. braun , u. staaden , t. holey , and m. fhnle , int .
b * 3 * , 1343 ( 1989 ) .
wang and d. chowdhury , j. phys .
( france ) * 50 * , 2905 ( 1989 ) .
wang , m. whlert , h. mhlenbein , and d. chowdhury , physica a * 166 * , 173 ( 1990 ) . t. holey and m. fhnle , phys .
b * 41 * , 11709 ( 1990 ) .
heuer , phys .
b * 42 * , 6476 ( 1990 ) .
a. weyersberg , t. holey and m. fhnle , j. stat .
* 66 * , 133 ( 1992 ) .
heuer , j. phys .
a * 26 * , l333 ( 1993 ) .
m. hennecke , phys .
b * 48 * , 6271 ( 1993 ) .
s. wiseman and e. domany , phys .
lett . * 81 * , 22 ( 1998 ) ; phys .
e * 58 * , 2938 ( 1998 ) .
h. g. ballesteros , l. a. fernndez , v. martn - mayor , a. muoz sudupe , g. parisi , and j. j. ruiz - lorenzo , phys .
b * 58 * , 2740 ( 1998 ) .
k. hukushima , random fixed point of three - dimensional random - bond ising models " , e - print cond - mat/0002089 .
r. a. dunlap and a. m. gottlieb , phys .
b * 23 * , 6106 ( 1981 ) .
r. j. birgeneau , r. a. cowley , g. shirane , h. joshizawa , d. p. belanger , a. r. king , and v. jaccarino , phys .
b * 27 * , 6747 ( 1983 ) .
j. m. hastings , l. m. corliss , and w. kunnmann , phys .
b * 31 * , 2902 ( 1985 ) .
d. p. belanger , a. r. king , and v. jaccarino , phys .
b * 34 * , 452 ( 1986 ) . p. h. barrett , phys .
b * 34 * , 3513 ( 1986 ) . p. w.
mitchell , r. a. cowley , h. yoshizawa , p. bni , y. j. uemura , and r. j. birgeneau , phys .
b * 34 * , 4719 ( 1986 ) .
d. p. belanger , a. r. king , i. b. ferreira , and v. jaccarino , phys .
b * 37 * , 226 ( 1988 ) .
n. rosov , a. kleinhammes , p. lidbjrk , c. hohenemser , and m. eibschtz , phys .
b * 37 * , 3265 ( 1988 ) . c. a. ramos , a. r. king , and v. jaccarino , phys . rev .
b * 37 * , 5483 ( 1988 ) .
t. r. thurston , c. j. peters , r. j. birgeneau , and p. m. horn , phys .
b * 37 * , 9559 ( 1988 ) .
i. b. ferreira , a. r. king , and v. jaccarino , phys .
b * 43 * , 10797 ( 1991 ) .
d. p. belanger , j. wang , z. slani , s .- j .
han , r. m. nicklow , m. lui , c. a. ramos , and d. lederman , journ .
magn . mater . *
140 - 144 * , 1549 ( 1995 ) .
d. p. belanger , j. wang , z. slani , s .- j .
han , r. m. nicklow , m. lui , c. a. ramos , and d. lederman , phys .
b * 54 * , 3420 ( 1996 ) . j. p. hill , q. feng , q. j. harris , r. j. birgeneau , a. p. ramirez , and a. cassanho , phys . rev .
b * 55 * , 356 ( 1997 ) .
z. slani , d. p. belanger , and j. a. fernandez - baca , j. magn .
magn . mater . *
177 - 181 * , 171 ( 1998 ) .
z. slani and d. p. belanger , j. magn .
magn . mater . *
186 * , 65 ( 1998 ) .
z. slani , d. p. belanger , and j. a. fernandez - baca , phys . rev
* 82 * , 426 ( 1999 ) . m. campostrini , a. pelissetto , p. rossi , and e. vicari , phys . rev .
e * 60 * , 3526 ( 1999 ) . c. w. kieweit , h. e. hall , and j. d. reppy , phys .
lett . * 35 * , 1286 ( 1975 ) .
d. finotello , k. a. gillis , a. wong , and m. h. w. chan , phys .
lett . * 61 * , 1954 ( 1988 ) .
m. h. w. chan , k. i. blum , s. q. murphy , g. k. s. wong , and j. d. reppy , phys .
lett . * 61 * , 1950 ( 1988 ) .
a. tyler , h. a. cho , and j. d. reppy , j. low temp .
phys . * 89 * , 57 ( 1992 ) .
j. yoon and m. h. w. chan , phys .
lett . * 78 * , 4801 ( 1997 ) .
g. m. zassenhaus and j. d. reppy , phys .
lett . * 83 * , 4800 ( 1999 ) . ref .
@xcite reports @xmath44 and ref .
@xcite finds that the exponent @xmath23 is compatible with 2/3 .
these estimates agree with the best experimental result for the pure system @xmath45 ( obtained from the estimate of @xmath46 using the hyperscaling relation @xmath47 ) @xcite and with the theoretical prediction @xmath48 @xcite .
there esist also experiments for @xmath34he in aerogels , which , however , find larger values for the exponent @xmath23 @xcite .
the current explanation of these results is that , in aerogels , the silica network is correlated to long distances , and therefore , the harris criterion and the model studied in this paper do not apply .
a simple model describing these materials has been studied in ref .
@xcite .
j. a. lipa , d. r. swanson , j. nissen , t. c. p. chui , and u. e. israelson , phys .
lett . * 76 * , 944 ( 1996 ) .
m. campostrini , a. pelissetto , p. rossi , and e. vicari , `` high - precision determination of the critical exponents for the @xmath33-transition of @xmath49he by improved high - temperature expansion '' , e - print cond - mat/9905395 , phys .
b * 61 * ( 2000 ) , in press . n. mulders , r. mehrotra , l. s. goldner , and g. ahlers , phys . rev . lett . * 67 * , 695 ( 1991 ) .
m. chan , n. mulders , and j. reppy , phys . today * 49 * , 30 ( 1996 ) .
y .- h . li and s. teitel , phys .
b * 41 * , 11388 ( 1990 ) .
v. s. dotsenko , b. harris , d. sherrington , and r. stinchcombe , j. phys .
a * 28 * , 3093 ( 1995 ) . v. s. dotsenko and d. e. feldman , j. phys .
a * 28 * , 5183 ( 1995 ) . v. s. dotsenko , j. phys .
a * 32 * , 2949 ( 1999 ) .
j. sak , phys .
b * 10 * , 3957 ( 1974 ) .
r. guida and j. zinn - justin , j. phys .
a * 31 * , 8103 ( 1998 ) .
m. hasenbusch and t. trk , j. phys .
a * 32 * , 6361 ( 1999 ) .
j. carmona , a. pelissetto , and e. vicari , `` the @xmath2-component ginzburg - landau hamiltonian with cubic anisotropy : a six - loop study '' , e - print cond - mat/9912115 , phys .
b * 61 * ( 2000 ) in press .
k. g. wilson and m. e. fisher , phys .
* 28 * ( 1972 ) 240 .
g. parisi , cargse lectures ( 1973 ) , j. stat .
* 23 * , 49 ( 1980 ) .
j. p. eckmann , j. magnen , and r. snor , comm .
phys . * 39 * , 251 ( 1975 ) ; j. magnen and r. snor , comm .
phys . * 56 * , 237 ( 1977 ) ; j. s. feldman and k. osterwalder , ann .
( ny ) * 97 * , 80 ( 1976 ) ; j. magnen and v. rivasseau , comm .
phys . * 102 * , 59 ( 1985 ) .
l. n. lipatov , zh .
. fiz . * 72 * , 411 ( 1977 ) [ sov .
jetp * 45 * , 216 ( 1977 ) ] .
e. brzin , j. c. le guillou , and j. zinn - justin , phys .
d * 15 * , 1544 , 1588 ( 1977 ) . g. a. baker , jr .
, b. g. nickel , m. s. green , and d. i. meiron , phys . rev .
lett . * 36 * , 1351 ( 1976 ) ; g. a. baker , jr . , b. g. nickel , and d. i. meiron , phys . rev .
b * 17 * , 1365 ( 1978 ) . j. c. le guillou and j. zinn - justin , phys .
lett . * 39 * , 95 ( 1977 ) ; phys .
b * 21 * , 3976 ( 1980 ) .
j. zinn - justin , _ quantum field theory and critical phenomena _
, third edition ( clarendon press , oxford , 1996 ) .
a. j. bray , t. mccarthy , m. a. moore , j. d. reger , and a. p. young , phys .
b * 36 * , 2212 ( 1987 )
. a. j. mckane , phys .
b * 49 * , 12003 ( 1994 ) .
g. lvarez , v. martn - mayor , and j. j. ruiz - lorenzo , j. phys . a * 33 * , 841 ( 2000 ) .
h. kleinert and v. schulte - frohlinde , phys .
b * 342 * , 284 ( 1995 ) .
r. schloms and v. dohm , nucl .
b * 328 * , 639 ( 1989 ) .
b. g. nickel , d. i. meiron , and g. a. baker , jr .
, `` compilation of 2-pt and 4-pt graphs for continuum spin models '' , university guelph report , 1977 , unpublished . b. r. heap , j. math . phys .
* 7 * , 1582 ( 1966 ) . s. a. antonenko and a. i. sokolov , phys .
e * 51 * , 1894 ( 1995 ) .
d. b. murray and b. g. nickel , `` revised estimates for critical exponents for the continuum @xmath43-vector model in 3 dimensions '' , guelph university report , 1991 , unpublished .
a. pelissetto and e. vicari , in preparation .
p. butera and m. comi , phys .
b * 56 * , 8212 ( 1997 ) . b. g. nickel , in _ phase transitions _ , m. lvy , j. c. le guillou , and j. zinn - justin eds .
, ( plenum , new york and london , 1982 ) p. 291324 . b. g. nickel , physica a * 117 * , 189 ( 1991 ) .
a. pelissetto and e. vicari , nucl .
b * 519 * , 626 ( 1998 ) ; nucl .
b ( proc . suppl . ) * 73 * , 775 ( 1999 ) , e - print hep - lat/9809041 . | we consider the ginzburg - landau @xmath0-model that describes @xmath1 @xmath2-vector cubic models with @xmath3-symmetric couplings .
we compute the renormalization - group functions to six - loop order in @xmath4 .
we focus on the limit @xmath5 which describes the critical behaviour of an @xmath1-vector model in the presence of weak quenched disorder .
we perform a detailed analysis of the perturbative series for the random ising model @xmath6 .
we obtain for the critical exponents : @xmath7 , @xmath8 , @xmath9 , @xmath10 , @xmath11 , @xmath12 .
for @xmath13 we show that the @xmath3 fixed point is stable , in agreement with general non - perturbative arguments , and that no random fixed point exists .
-0.4 in 0.4 in 9.0 in |
this work was partially supported by the french renatech network for the nanofabrication of the samples . | neuromimetic systems are systems mimicking the functionalities or architecture of biological neurons and may present an alternative path for efficient computing and information processing .
we demonstrate here experimentally temporal summation in a neuromimetic micropillar laser with integrated saturable absorber .
temporal summation is the property of neurons to integrate delayed input stimuli and to respond by an all - or - none kind of response if the inputs arrive in a sufficiently small time window .
our system alone may act as a fast optical coincidence detector and paves the way to fast photonic spike processing networks .
neuromimetic photonic systems are optical systems that mimic the functionalities or the architecture of biological neurons , and can represent an alternative path for computing and processing information very efficiently both in terms of energy , speed , and robustness versus noise @xcite . from a functional and basic point of view
, a biological neuron can be thought of as a system than can integrate information from various stimuli , and respond in an all - or - none fashion if the integrated input stimuli exceed a certain threshold @xcite .
this latter property is called excitability and has been experimentally demonstrated already in many nonlinear semiconductor optical systems , like active semiconductor cavities with feedback @xcite , with optical injection @xcite or with saturable absorber @xcite .
the former property is called temporal summation .
temporal summation refers to the ability of the system to integrate different , potentially delayed , presynaptic stimuli and to emit a spike if the integration of the inputs exceeds the excitable threshold . since in that case the system acts as an integrator
, it has a time constant and the summation only takes place if the presynaptic stimuli arrive within in a given time window . while most of the cortical neurons are integrators @xcite , note that there also exists so called resonator neurons for which summation rules depend more on the phase of the input stimuli with respect to their subthreshold oscillation frequency @xcite . in optical systems ,
the response is in the form of an intensity spike with a characteristic shape and can be well under the nanosecond timescale @xcite . following the demonstration of excitability in a planar semiconductor laser with integrated saturable absorber @xcite , it has been suggested @xcite that this kind of system could act as a leaky integrate - and - fire neuron , a model of neuron widespread in neuroscience @xcite and optically implemented with telecom components in @xcite .
the excitable response of micropillar lasers with integrated saturable absorber has been investigated in @xcite demonstrating the absolute and relative refractory periods , in complete analogy to biological systems . in this letter
we investigate the response of a micropillar laser with integrated saturable absorber to sub - threshold stimuli and show that the system can integrate the stimuli and emit an excitable spike if the stimuli are close enough in time .
this demonstrates temporal summation in this system and its ability to act as a fast optical coincidence detector @xcite , a property usable for optical pattern recognition tasks .
this property is also at the base of some neurocomputational models for e.g. sound localization in the barn owl auditory system @xcite .
it is also of paramount importance to fabricate optical neuromimetic circuits by e.g. coupling several micropillar units since , together with excitability , it demonstrates a fast and compact leaky integrate - and - fire optical neuron .
the experiment consists of a micropillar laser with integrated saturable absorber of original design described in @xcite .
the micropillar is optically pumped by a laser diode array emitting around 800 nm and emits light close to cavity resonance designed at 980 nm .
the active medium consists of two ingaas / algaas quantum wells while the saturable absorber medium has only one quantum well .
optical perturbations are sent to the system thanks to a ti : sa model - locked laser emitting @xmath0ps duration pulses with a @xmath1mhz repetition rate ( see figure [ fig : setup ] )
. the rate can be down - sampled thanks to a pulse picker .
a tunable delay line is inserted in the perturbation path allowing consecutive perturbations with a delay of several hundreds of picoseconds to several nanoseconds .
the pump and perturbation beams are coupled into the micropillar thanks to a dichroic mirror and a microscope objective .
two fast avalanche photodetectors ( @xmath2 and @xmath3 ) with 10ghz and 4ghz bandwidth respectively record the input pulses and the response from the micropillar .
the signals are analyzed with a 6ghz oscilloscope . ,
na=0,85 ) . ti : sa , mode - locked laser ( 80ps pulse duration ) ; aom , pulse picker ; apd@xmath4 , avalanche photodiodes ; cam , cmos camera . ]
the micropillar is driven in the excitable regime , 9% below the self - pulsing threshold , thus no laser light is emitted .
two consecutive sub - threshold perturbations are sent onto the micropillar at 797.5 nm , close to the pump wavelength , with a variable delay @xmath5 . the first and second perturbation are set to 74% and 80% of the excitable threshold so that none of each perturbation is able by itself alone to trigger an excitable response .
since noise is present in the system ( either internal , spontaneous emission noise or external noise sources such as pump noise or noise in the perturbation amplitude ) , we record the response of the system after sending 10000 identical perturbations .
the results are shown on fig.[fig : traces - sum - resp ] . for a large perturbation delay ( @xmath6 , fig.[fig :
traces - sum - resp]f ) , the perturbations rarely adds - up to trigger an excitable response while for shorter ones , a clear , large amplitude response @xmath7 is visible .
this means that the system integrates the perturbations which produces an above - threshold stimulus able to trigger a large , excitable response .
note that the excitable response , when present , has always the shape depicted on fig.[fig : traces - sum - resp]a ) when a single event is plotted ( green curve ) .
the average response has a dispersion in time because of the dynamical delay induced by the noise , which can be mostly attributed to pump and perturbation noise . in order to quantify the response we plot in fig.[fig : ampli]a the median of the response amplitude @xmath7 versus delay @xmath5 .
a clear transition is visible in the amplitude : for delays below 610ps the perturbations trigger an excitable response and are thus temporally summed . for larger delays
the summation does not occur anymore .
this behavior is in stark contrast to the case of a gain - switched laser that would respond linearly .
the transition depends on the excitable threshold value and thus on the bias pump @xcite , and on the amplitude of the incoming perturbations .
summation also occurs for a resonant perturbation at the cavity resonance close to 980 nm , as can be seen on fig.[fig : temp - sum - coherent ] . in that case
, the excitation wavelength is @xmath8 nm . the bias pump is set to 90% of the self - pulsing threshold in the excitable regime , as can be seen in the inset .
the two perturbations are set 44% and 66% of the excitability threshold for the direct and delayed pulses respectively .
they are clearly visible in the recorded traces and are marked by arrows .
temporal summation occurs for short delays ( a ) , 220ps and to a lesser extent b ) 350ps ) and is absent for the other delays ( c e ) 450 , 540 and 630ps ) .
contrarily to the incoherent excitation case , the excitable curve in inset shows a marked plateau for stimuli above the excitable threshold .
this behavior is important in view of cascading several excitable units .
the fact that temporal summation is possible for incoherent ( perturbation on the gain carrier density ) and `` coherent '' ( perturbation on intensity at cavity resonance ) adds some flexibility to the system and has no counterpart in biological systems .
note also that there is a delay of nonlinear origin in the response that depends on the summed perturbations ( fig.[fig : ampli]a ) .
this mechanism is interesting from a neuromimetic point of view since it provides a natural time - coding mechanism for the amplitudes : for a large input the delay is short , while it is long for a small ( supra - threshold ) summed input . :
a ) 210 , b ) 320 , c ) 420 , d ) 520 , e ) 610 and f ) 700ps .
the plots show the statistical density of points ( in log scale ) for 10000 different realizations . on the plots a typical response pulse
is shown in green .
inset : excitable response to a single perturbation .
red stars are the detected response maxima .
orange is a plot of the median in a sliding window with 500 points . ]
nm .
the bias pumping is set to 90% of the self - pulsing threshold .
the input stimuli are set to 44% and 66% of the excitability threshold for the direct and delayed pulses respectively .
the plots show the statistical density of points ( in log scale ) for 10000 different realizations of the input perturbations with delays ( a e ) : 220ps , 350ps , 450ps , 540ps and 630ps . on the plots ,
a typical response pulse is shown in green .
the intensity perturbations are indicated by arrows . in inset : excitable response to a single perturbation median - averaged over 500 points . ]
( a)median of the maximum response @xmath7 ( blue ) from experimental data in fig.[fig : traces - sum - resp ] for two consecutive stimuli separated by the delay @xmath5 .
the response to each perturbation alone is plotted in light blue ( empty diamond , empty circle respectively ) .
the response delay ( relative to the shortest response delay ) @xmath9 is shown in red .
( b ) same for model with parameters : @xmath10 , @xmath11 , @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 .
the response delay @xmath19 is relative to the second perturbation . ]
the experimental results are compared to numerical simulations of the yamada model with spontaneous emission @xcite .
this model reads : @xmath20 .
the dynamical variables are the intracavity intensity @xmath21 , the gain @xmath22 and the absorption @xmath23 .
recombination rates of carriers in the gain and saturable absorber sections are respectively @xmath24 and @xmath25 .
other parameters are @xmath26 the saturation parameter , @xmath27 the linear losses and @xmath28 the spontaneous emission factor .
the term @xmath29 is directly proportional to the carrier density in the gain section .
the @xmath28 parameter is small here and steady state solutions can be expended in power series of @xmath28 @xcite .
let @xmath30 be a steady state solution of eqs.[eq : yamada ] . a simple approach to get
an analytic insight into the dynamics is to consider that the intensity is small ( @xmath31 ) and almost constant as long as a pulse has not been triggered .
suppose at time @xmath32 a first delta - like sub - threshold perturbation is sent e.g. on the gain followed by a second perturbation on at time @xmath33 ( in units of the photon lifetime in the cavity ) such that @xmath34 , where @xmath35 is the dirac delta function .
we can then solve for @xmath36 such that @xmath37+\nonumber \\ & & \pi(t-\delta)\mu_{2}\exp\left[\frac{-\gamma_{g}\left(t-\delta\right)}{1+i_{ss}}\right]\label{eq : gain - net}\end{aligned}\ ] ] with @xmath38 the unit - step function .
as noted in @xcite , in a first approximation the net gain @xmath39 governs the triggering of an excitable pulse : a pulse can only be excited if @xmath40 exceeds zero for a sufficiently large amount of time .
hence the final state of the system is controlled by the net gain immediately after the second perturbation at @xmath41 ( always considering a first , sub - threshold stimulus ) .
numerical simulations are shown on fig.[fig : gainnet ] .
the net gain is plotted together with the response to the consecutive stimuli .
simulations of the full system ( eqs.[eq : yamada ] ) and of the approximate solution ( eqs.[eq : gain - net ] ) for the net - gain are shown .
the parameters are similar to the one used in @xcite except the recombinations rates that have been tuned to match better the experimental results . with the estimated photon lifetime of the empty cavity being 3.25ps
, a reasonable qualitative agreement is met between the model and the experiment .
the critical delay is found to be @xmath42 , that is to say 708ps in physical units with the estimated photon lifetime .
this is in reasonable agreement with the experimental value in fig.[fig : ampli]a between 600 and 700ps and for incoherent perturbation conditions similar to those of the experiment : namely , consecutive perturbations of amplitude 74% of the single perturbation excitable threshold @xmath43 with the pump at 91% of the self - pulsing threshold ( corresponding here to the numerical value of 2.88 ) .
the sub - threshold dynamics of the net gain is very well reproduced by the linear approximation solution until it completely fails when a pulse is triggered , as expected . for the longest delay ( @xmath44 ) ,
when the summed perturbations are sub - threshold , the approximate net gain deviates from the full solution after the second perturbation because it passes close to 0 and hence the intensity dynamics , even failing to trigger an excitable response , has a non - negligible influence on it .
the response and the delay in the response versus delay between the sub - threshold stimuli are shown in fig.[fig : ampli]b .
a good agreement is found with the experimental results .
the response abruptly switches to zero when the delay @xmath5 is larger than a critical delay @xmath42 . at this point ,
the response delay @xmath19 diverges as one expects .
note that there is a point for the response delay @xmath19 in the experimental curve ( fig.[fig : ampli ] ) even after the critical delay @xmath45 because of noise that is able to trigger a few excitable pulses even for stimuli on average below the excitable threshold .
these events are taken into account for the determination of @xmath19 .
the divergence of the delay for perturbations close to the excitable threshold is also responsible for the large dispersion in response times clearly visible in fig.[fig : traces - sum - resp]d - f ) because a larger number of stimuli will be brought just above the excitable threshold in these cases .
it is interesting to notice that the final state of the system is different for asymmetric perturbations : for a balanced perturbation such that @xmath46 is constant , the highest net gain at time @xmath5 is obtained for asymmetric perturbations with @xmath47 .
this slight effect means that the temporal summation considered here is a non - commutative operation .
showing the response to two sub - threshold stimuli on the gain ( @xmath48 ) with a variable delay @xmath5 : net gain ( blue ) , intensity ( red ) and linear solution ( eqs[eq : gain - net ] , orange dashed line ) .
parameters are the same as in fig.[fig : ampli ] .
the excitable threshold is @xmath49 , and @xmath50 .
the delays are @xmath51 corresponding respectively to physical delays of 200 , 400 , 600 and 800ps . ]
the critical delay is computed for two input stimuli of equal amplitudes @xmath52 in fig.[fig : critical - delay ] . for an amplitude @xmath53 ,
temporal summation never occurs since the summed stimuli are not large enough to cross the excitable threshold even for zero delay . on the contrary , for larger stimuli
the maximum possible delay giving rise to an excitable response increases first almost linearly with the stimulus strength . when a single stimulus is almost able to trigger an excitable response obviously the maximum delay increases and diverges .
critical delay @xmath45 versus amplitude of the two stimuli for @xmath54 .
dashed lines correspond to the case of fig.[fig : ampli](b ) and @xmath55 . ] in conclusion we have studied the effect of consecutive sub - threshold stimuli on the response of a semiconductor micropillar neuromimetic system .
we have shown that the system can integrate stimuli with a time constant of the order of several hundreds of picoseconds which depends on the recombination times of carriers in the active medium and respond with a macroscopic , excitable pulse if the time delay between the stimuli is shorter than a critical delay .
the critical delay depends on the excitable threshold which is controllable by the amount of pumping of the system .
this system thus simulates the behavior of a leaky integrate - and - fire neuron and is particularity suited for building more advanced processing functionalities .
incoherent and coherent perturbation schemes have been demonstrated , the latter being important to demonstrate cascadability of such excitable systems . as was demonstrated in @xcite ,
a sufficient number of excitable processing units coupled appropriately is capable of universal computation .
it is however important to keep in mind that even a single unit has also processing capabilities @xcite . in the case of two stimuli
, this system behaves as a fast optical coincidence detector , a functionality that may be of importance in many applications like pattern recognition in optical signals , in analogy to biological systems using these property for sound localization @xcite . |
master equations , or evolution equations for reduced density operators are of fundamental importance in the treatment of open systems in quantum theory .
they are particularly relevant in the field of quantum optics @xcite , where generally one has interest in studying the evolution of confined subsystems ( a single mode of the electromagnetic field , for instance ) `` coupled '' to external reservoirs .
this may model a number of ( phase - sensitive or not ) processes , such as the decay of a single mode of the field confined in a lossy cavity @xcite , as well as amplification processes @xcite .
interesting applications of phase - sensitive amplification , for example , are the reduction of noise in lasers through the injection of squeezed vacuum @xcite , as well as noise - free amplification via the two - photon correlated - emission - laser @xcite .
nevertheless , complete operator solutions of master equations are not usually presented . only simple cases such as the decay of a single mode of the field into a vacuum state reservoir ( @xmath0 ) or stationary regime cases are the ones normally treated @xcite .
instead , master equations are normally transformed into c - number equations either in the number or coherent state representations , for instance .
it would be therefore interesting to obtain a direct solution of the master equation while in its operator form .
in this paper we show how to solve exactly master equations of the form @xmath1 where @xmath2 and @xmath3 being @xmath4 and @xmath5 the bosonic annihilation and creation operators .
the @xmath6s are in general complex parameters which may represent gain and decay . for the density operator to be hermitean ,
it is necessary that @xmath7 we may now apply the unitary transformation @xmath8 where @xmath9 is the squeeze operator @xmath10 with @xmath11 . if we now substitute eq.([eq : transfo ] ) into eq.([eq : mastereq ] ) , we obtain an equation for the transformed density operator @xmath12 , or @xmath13 where @xmath14 with @xmath15 we proceed in rewriting eq.([eq : mastereqt ] ) in terms of the original creation and annihilation operators as @xmath16 we note that one can choose @xmath17 and @xmath18 in such a way that the last two terms in eq.([eq : eqtrans ] ) become zero .
the appropriate choices for these parameters are @xmath19 and @xmath20 with the above choice of parameters , then we can write eq.([eq : eqtrans ] ) as @xmath21\hat{\rho}',\label{eq : mmaeq}\ ] ] where @xmath22 and @xmath23 it follows from the relations above that @xmath24 and @xmath25 we now write the new master equation ( [ eq : mmaeq ] ) in a more convenient form @xcite @xmath26 where we have defined the following super - operators @xmath27 @xmath28 and @xmath29 from the equations above , we can write the formal solution of eq.([eq : smaeq ] ) as @xmath30\hat{\rho}'(0).\label{eq : forsol}\ ] ] it is not difficult to show that the superoperators @xmath31 , @xmath32 and @xmath33 obey the following commutation relations @xcite @xmath34\hat{\rho}'=\left[\frac{4\tilde{\gamma}_1 \tilde{\gamma}_2}{\tilde{\gamma}_1 + \tilde{\gamma}_2}\hat{j}_3 - 4\tilde{\gamma}_1\tilde{\gamma}_2\right]\hat{\rho}',\ ] ] @xmath35\hat{\rho}'=-2\left(\tilde{\gamma}_1+\tilde{\gamma}_2 \right ) \hat{j}_1\hat{\rho}',\ ] ] and @xmath36\hat{\rho}'=2\left(\tilde{\gamma}_1+\tilde{\gamma}_2 \right ) \hat{j}_2\hat{\rho}'.\ ] ] in order to disentangle the exponential in eq.([eq : forsol ] ) , we propose the ansatz @xmath37\exp[f_2(t)\hat{j}_2 ] \exp[f_0(t)\hat{j}_3]\exp[f_1(t)\hat{j}_1(t)]\hat{\rho}'(0)\label{eq : ansat}\ ] ] by inserting @xmath38 in eq.([eq : ansat ] ) above into equation in eq.([eq : smaeq ] ) , we obtain the following system of differential equations for the functions @xmath39 : @xmath40 @xmath41=1,\ ] ] @xmath42 and @xmath43 although the system is a non - linear one , its solution is rather straightforward . in order to have the condition @xmath44 satisfied
, we should have @xmath45 ( @xmath46 ) as initial conditions for the set of eqs.([eq : eqfo])-([eq : eqf3 ] ) .
the result is @xmath47 @xmath48 and @xmath49 where @xmath50)\tilde{\gamma}_2 $ ] . by finally defining the dimensionless superoperators @xmath51 where the superoperators @xmath52 , @xmath53 and @xmath54 obey the commutation relations @xmath55\hat{\rho}=\hat{l}_3\hat{\rho}$ ] and @xmath56\hat{\rho}=\pm 2\hat{l}_{\pm}\hat{\rho}$ ] , we can write the solution of ( [ eq : mastereq ] ) in the form , @xmath57^{\hat{l}_3 } e^{\frac{\tilde{\gamma}_1}{\tilde{\gamma}_2 } \frac{\gamma ( t)}{\kappa+\gamma(t)}\hat{l}_- } \label{finalrho } \hat{\rho}(0),\ ] ] where in ( [ btilde ] ) we have defined @xmath58 which should be compared with eq .
( [ bdef ] ) .
note that if in eq .
( [ finalrho ] ) we set the parameters @xmath59 and @xmath60 we recover the usual solution for a dissipative cavity at zero temperature as well as for the phase insensitive case .
we have therefore obtained the full solution of the master equation for @xmath12 [ see eq.([eq : smaeq ] ) ] , given by ( [ finalrho ] ) .
we remind that we still have to apply the transformation in eq.([eq : transfo ] ) in order to recover the original density operator @xmath61
. it would be appropriate now to show which interesting cases could be easily treated by employing our solution .
for instance , if we make @xmath62 , @xmath63 , @xmath64 and @xmath65 , we will obtain the standard equation describing the decay of a bosonic field into a phase - sensitive reservoir @xcite . if on the other hand @xmath66 and @xmath67 , but keeping @xmath68 and @xmath69 , we will have the standard equation describing phase - sensitive amplification @xcite .
the constants @xmath6 and @xmath70 represent decay and gain , respectively , and @xmath71 ( @xmath72 ) are connected with phase - sensitive ( or not ) reservoir fluctuations .
each of this two cases are relevant in the problems of field decay and in the reduction of noise in lasers @xcite .
we remark that our operator solution turns unnecessary the convertion of the master equation into sometimes cumbersome c - number equations .
let us finally consider as an example an initial density matrix of the form @xmath73 or a squeezed vacuum state .
after inserting it into ( [ finalrho ] ) it is easy to show that the density operator of the field at a time @xmath74 may be written as @xmath75^m @xmath76 are the squeezed number states @xcite .
we may now to follow the state s evolution in phase space , using for instance the @xmath77-function @xcite , defined as @xmath78 where @xmath79 is a coherent state . in our case
the @xmath77-function is given by @xmath80^m|\langle\beta|m,\xi\rangle|^2 , \label{qfuncomp}\ ] ] with @xmath81 and where @xcite @xmath82 in figure 1 it is illustrated the @xmath77-function in ( [ qfuncomp ] ) for different times . at @xmath83 we have the @xmath77 function of the initial squeezed vacuum state .
as time goes on , thermal fluctuations of the reservoir cause a `` spread '' of the @xmath77 function , associated to the increase in the quadrature noise .
nevertheless , due to the phase - sensitive properties of the reservoir , the noise assymmetry characteristic of squeezed states is somehow preserved , as it may be seen in figure 1 .
to summarize , we have presented an alternative way of treating problems involving certain types of master equations .
our operator solution may be useful to retrieve any sort of information related to the evolution of sub - systems having specific initial conditions .
this contrasts with the usual approaches , where only partial information , such as mean values of amplitudes and/or diffusion coefficients is normally obtained .
_ note added in proof : _ while preparing the answer to the referee , we became aware of the paper by dung and knll @xcite where they also solve eq .
( [ eq : mastereq ] ) but using fokker - planck equations instead .
we would like to stress that our method of solving the master equation ( [ eq : mastereq ] ) is considerably more straightforward and simple . | we present a method of solving master equations which may describe , in their most general form , phase sensitive processes such as decay and amplification .
we make use of the superoperator technique . |
the earth orientation is generally considered as ( i ) earth rotation axis movements in space ( precession - nutation ) , ( ii ) earth rotation axis movements in the earth ( polar motion ) , or ( iii ) earth rotation speed variations ( exces in the length of the day ) .
these movements come from earth inside masses distributions .
the earth gravity field can give us information about this distribution of masses because nowadays we can determine the variations of the earth gravity field by space geodetic techniques .
hence , there is a link between the variations of the earth gravity field and the variations of the earth orientation parameters . and
the high accuracy now reached in the vlbi ( very long baseline interferometry ) earth orientation parameters ( eop ) determination requires looking further at the various geophysical contributions to variations in eop .
so we investigate here if this variable gravity field can be valuable for the improving the modelisation of the earth rotation .
the fundamental equations for the rotation of the earth in an inertial frame are euler s dynamical equations , based on the conservation of the angular momentum @xmath0 of the earth under an external torque @xmath1 ( lambeck 1980 ) : @xmath2 for a non - rigid earth , these equations in a rotating frame become : @xmath3 + \vec \omega \wedge \left [ i(t)~\vec \omega + \vec h(t)\right ] = \vec l\ ] ] where the inertia tensor @xmath4 is time dependent , as well as the relative angular momentum @xmath5 , and @xmath6 is the earth instantaneous rotation vector which direction is the one of the rotation axis and which norm is the rotation speed .
it depends on the earth orientation parameters ( eop ) .
the inertia tensor , which is symetric , can be written as : @xmath7 = \left [ \begin{array}{ccc } a+c_{11 } & c_{12 } & c_{13 } \\ c_{12 } & b+c_{22 } & c_{23 } \\ c_{13 } & c_{23 } & c+c_{33 } \end{array } \right]\ ] ] with @xmath8 the constant part and @xmath9 ( @xmath10 ) the variable part of the inertia tensor .
the earth gravity field of the earth devived from the external gravitational potential @xmath11 which is expressed in a spherical harmonic expansion as ( lambeck 1980 ) : @xmath12\ ] ] where @xmath13 is the geocentric distance , @xmath14 the latitude and @xmath15 the longitude of the point at which @xmath11 is detremined .
@xmath16 is the gravitational constant , @xmath17 and @xmath18 are the mass and the equatorial radius of the earth , respectively . @xmath19 and @xmath20 are the stokes coefficients of degree @xmath21 and order @xmath22 , and @xmath23 are the legendre polynomials . hence the second - degree stokes coefficients can be directely related to the inertia tensor components ( lambeck , 1988 ) : @xmath24
we just have shown that the earth rotation ( with @xmath6 and the eop ) could be related to the earth gravity field ( with the degree 2 stokes coefficients ) .
then , we investigate now how we can link each eop with these coefficients .
the exces in the length of the day @xmath25 ( with respect to a mean lod ) can be related to ( i ) the third component @xmath26 of the variable part of the inertia tensor and ( ii ) the third component @xmath27 of the relative angular momentum of the earth , ignoring the external torques : @xmath28 moreover , with the help of eq .
( [ eq : coeffs_c_inertie ] ) , we can write : @xmath29 where @xmath30 is the variation in time of the sum of the diagonal elements of the inertia tensor .
we can consider that it is equal to zero ( rochester & smylie 1974 ) .
then , we can obtain : @xmath31 where the coefficient @xmath32 accounts for the loading effects and @xmath33 is the third moment of inertia of the earth s mantle ( barnes et al .
then we have compared the @xmath25 obtained with eq .
( [ eq : lod_fin ] ) and the @xmath34 data in fig .
[ fig : bourda_fig2 ] with the one usually used but corrected from zonal tides , atmopheric wind effects ( @xmath27 ) and long terms ( see fig . [
fig : bourda_fig1 ] ) .
the study of the earth precession nutation angles variations influenced by the temporal variations of the @xmath35 coefficients of the geopotential is developped in the article of bourda & capitaine ( 2004 ) .
it is based on the works of williams ( 1994 ) and capitaine et al .
( 2003 ) which considered secular terms for the @xmath35 variations , whereas we consider also annual and semi - annual ones . the polar motion @xmath36 , where @xmath37 and @xmath38 are the components of the rotation axis in space can be theoretically related to the degree 2 and order 1 coefficients of the earth gravity field : @xmath39 where @xmath40 is related to @xmath41 and @xmath42 with eq .
( [ eq : coeffs_c_inertie ] ) , and @xmath43 .
the part of the length of the day obtained with the @xmath34 data corresponds to gravitational terms .
then we have compared @xmath25 corrected from the movements terms ( as atmospheric ones ) , the zonal tides and the decadal terms ( from magnetic effects in the core - mantle boundary ) .
but the residual term has an amplitude of the order of @xmath44 @xmath45s , whereas the better precision on these lod data is of the order of @xmath46 @xmath45s .
barnes , r. t. h. , hide , r. , white , a. a. , & wilson , c. a. 1983 , proc .
lond . , a 387 , 31 bourda , g. , & capitaine , n. 2004 , a&a , in press capitaine , n. , wallace , p. t. , & chapront , j. 2003 , a & a , 412 , 567 lambeck , k. 1980 , the earth s variable rotation ( cambridge : cambridge univ . press ) lambeck , k. 1988 , geophysical geodesy : the slow deformations of the earth ( oxford : oxford science publications ) rochester , m. g. , & smylie , d. e. 1974 , j. geophys .
res . , 79 , 4948 williams , j. g. 1994 , astron . j. , 108(2 ) , 711 | the determination of the earth gravity field from space geodetic techniques now allows us to obtain the temporal variations of the low degree coefficients of the geopotential , combining the orbitography of several satellites ( e.g. lageos1 , lageos2 , starlette ) .
these temporal variations of the earth gravity field can be related to the earth orientation parameters ( eop ) through the inertia tensor .
this paper shows these relations and discusses how such geodetic data can contribute to the understanding of the variations in eop . |
solid state oxides containing transition metal ions with unfilled @xmath12 shell electrons have been at the centre of attention in condensed matter and materials physics for the last two decades because of the variety and novelty of electronic phenomena in them and the possibility of new applications .
the @xmath12 electrons in solids are not as extended as @xmath13 electrons or as localized as @xmath14 electrons .
their motion in the system is highly constrained by the large local ( on site ) @xmath15 repulsion or mott - hubbard correlation @xmath10 .
the central question is how this strong correlation , and other factors such as orbital degeneracy and electron lattice coupling , lead to electronic behaviour qualitatively different from that of conventional solids which are successfully described as degenerate fermi liquids with well defined interacting electronic quasi - particles . while the best known examples of these are the high @xmath16 cuprate superconductors @xcite , during the last decade another group of oxides , namely the manganites @xmath0 ( where @xmath17 etc . and @xmath18 etc
. ) has become a focus of major activity .
the initial interest was sparked by the discovery @xcite that their electrical resistivity changes enormously with the application of a magnetic field , the change ( colossal magnetoresistance or cmr ) being two or more orders of magnitude larger than the normal cyclotron orbital effect characterized by the dimensionless parameter @xmath19 .
subsequent work has shown a bewildering variety of phases , phase transitions and phenomena @xcite depending on the doping @xmath8 , temperature @xmath9 , and ionic species @xmath20 and @xmath21 as well as external perturbations .
an example is the phase diagram of @xmath22 ( fig .
[ phdiag ] ) which shows an insulating , @xmath23 bond ( jahn - teller ) distorted but structurally ordered phase for small @xmath8 , the transition of this at low @xmath9 from a ferromagnetic insulator to a ferromagnetic metal at @xmath24 , and thence to a charge ordered insulating phase for @xmath25 . for @xmath26
this becomes a paramagnetic insulator above a @xmath27 , with cmr near @xmath16 .
the phase diagram varies considerably with the ionic species .
for example , @xmath28 has a paramagnetic metallic phase for @xmath29 and shows no charge / orbital order , while @xmath30 has no metallic phases ( even ferromagnetic)@xcite .
two other general characteristics are the following .
first , physical properties are extremely sensitive to small perturbations ; examples being the cmr itself , the unusually large strain and ion size effects @xcite , the melting of charge / orbital ordering for anomalously small magnetic fields and a metal to insulator transition induced by the electronically benign substitution of @xmath31 by @xmath32 @xcite . secondly , over a wide range of @xmath8 and @xmath9 , two very different types of regions , one insulating and locally lattice distorted and the other metallic and undistorted , coexist @xcite .
the regions can be static @xcite or dynamic @xcite ; their size can vary from @xmath33 @xcite to @xmath34 @xcite .
all these observations suggest that metallic and insulating phases are always very close in free energy .
these phenomena can not be described satisfactorily in terms of well known models for electrons in solids .
for example , one might expect generally that doping @xmath35 ( a mott insulator ) with @xmath36 introduces mobile holes in the @xmath37 band so that the system ought to be metallic at least in the high symmetry paramagnetic phase ; but it is not till @xmath38 .
the observed properties of manganites need to be understood in terms of the active degrees of freedom which are believed to be the twofold degenerate @xmath3 electrons and the @xmath39 core - spins of @xmath4 , and jahn - teller(jt ) optical phonon modes of the oxygen octahedra and the interactions between them .
there are three strong interactions present , namely the large on site @xmath15 repulsion @xmath40 @xcite amongst the @xmath3 electrons , strong jt mode - @xmath3 electron coupling @xmath41 which splits the doubly degenerate @xmath3 level by about @xmath42 @xcite , and the large ferromagnetic hund s rule coupling @xmath43 between @xmath3 and @xmath39 spins ( @xmath44 1 to 2 @xmath45 @xcite ) . for comparison , the @xmath46 electron kinetic energy scale or band width
@xmath47 is @xmath48 @xcite .
the presence of a large jt coupling is clearly indicated by the two ` phase ' coexistence@xcite ( with one phase being lattice distorted and insulating , and the other metallic and undistorted ) as well as by the sensitivity of properties to lattice strain or local disorder , and giant isotope effect@xcite alluded to above .
indeed , large jt distortions , while long ranged in @xmath35 , persist locally well into the metallic regime@xcite ( @xmath49 , and @xmath50 ) , where they are short ranged .
understanding the observed consequences of these interactions is one of the major challenges in the physics of strongly correlated electrons .
most existing theories @xcite for manganites neglect one or the other of these strong interactions , make further approximations , and are inadequate qualitatively and quantitatively .
the earliest theoretical approaches @xcite , commonly referred to as double exchange theories , consider solely the hund s rule exchange @xmath43 .
however , only a metallic state is possible in this case .
a theory due to millis , mueller and shraiman @xcite additionally includes the effect of the coupling @xmath41 , but treats the local jt lattice distortion classically , as annealed static disorder , and neglects @xmath10 .
a polaronic insulating phase also occurs in that case for large enough @xmath41 , but the predicted results do not resemble experiments ; for example at @xmath51 , one finds only metal - metal or insulator - insulator curie transitions , unlike the commonly observed metal - insulator transition .
the magnetoresistance is not colossal and there is no isotope effect .
dagotto and coworkers ( see eg .
@xcite ) have done extensive numerical simulations of several models , but on rather small lattices , and seen lots of instances of `` phase separation '' .
the early simulations explored the competition between ( double exchange induced ) ferromagnetism and anti - ferromagnetic ( super)exchange , amplification and generation of small scale phase co - existence by disorder , and identified the two kinds of magnetic domains with metallic and insulating regions .
later simulations explored similar issues including static jt couplings .
based on these , they suggest@xcite that transport in manganites should be pictured in terms of a random resistor network arising from tunnelling between misaligned ferromagnetic metallic domains across insulating regions , and that the cmr arises from enhanced tunnelling due to field induced alignment of the magnetic domains .
there are many other models designed to address specific effects , but no theoretical ideas which explain the novel general features of manganites cohesively .
we propose here a new approach to the physics of manganites which includes all the three strong interactions , the double degeneracy of @xmath3 orbitals as well as the quantum dynamics of phonons .
it leads to a simple physical description from which the properties of manganites follow naturally .
calculations of ground state and transport properties at finite temperatures are also presented .
we believe that the ideas are of wide relevance to a large number of solid state and molecular systems with similar ingredients .
our ideas are principally based on two facts ; namely the two fold degeneracy of the @xmath3 orbital and the strong ( degeneracy lifting ) jt interaction characterized by the large dimensionless ratio @xmath52 , ( @xmath53 for manganites , see ref.@xcite ) where @xmath54 is the jt optical phonon energy @xmath55 @xcite .
this coupling leads ( as shown in greater detail below ) to a large local jt distortion @xmath56 being associated with one linear combination of the orbitals ( jt polaron , labelled @xmath5 by us ) when it occupied by a single electron . if the jt modes are not approximated as static displacements @xcite but are treated quantum - dynamically , the inter - site hopping of the jt polaron is reduced by the exponential huang rhys @xcite factor @xmath57 for @xmath58 and @xmath59 .
this is the _
antiadiabatic limit _ overlap between the initial and final jt phonon wave functions ( centred around @xmath60 and @xmath61 ) when the @xmath5 electron moves in or out of a site .
the corresponding effective @xmath5-bandwidth , @xmath62 , is thus very small , so the @xmath5 states are easily localized by any disorder present ( eg . in cation site energy ) .
they can hence be regarded , to a first approximation ( which may be inaccurate for @xmath63 because of inter - site coherence , see section 8) , as non dispersive localized levels .
the @xmath5 polaron is probably close to this limit over a wide range of @xmath8 and @xmath9 , where large local @xmath64 distortion without long range order is seen @xcite .
since the @xmath3 orbital at each site is doubly degenerate initially , there is another orthogonal set of states which we label @xmath6 , which have their largest amplitudes at the fraction @xmath8 of hole sites where the polaron is not present ; their occupancy on the polaron site costs a large extra energy @xmath65 .
the ( bare ) hopping @xmath66 amongst these @xmath6 states is not reduced and they form a broad band ( of bare width @xmath67 ev ) whose properties are strongly affected , eg .
their effective bandwidth @xmath68 is renormalized to smaller values , by the other two strong interactions present in the system , namely the repulsive scattering from the @xmath5 polarons ( @xmath69 ) and the coupling to the @xmath39 spins @xmath70 depending on @xmath8 and @xmath9 .
roughly , @xmath68 increases with @xmath8 as well as with @xmath71 and @xmath72 , because the inhibition of @xmath6 hopping due to large @xmath69 is reduced when there are more hole - sites , and that due to large @xmath43 is reduced when the @xmath39 spin order is enhanced .
we believe that the unique feature of manganites is this necessary coexistence of antiadiabatic , jt distorted , localized @xmath73 states and adiabatic , undistorted , broad band @xmath74 states , arising from a _ spontaneous reorganization of the doubly degenerate @xmath3 states due to the large jt coupling and quantum phonon dynamics_. this dynamically generated coexistence is qualitatively different from that of localized @xmath14 and extended @xmath75 electrons in rare earth solids , which has an atomic origin .
we show below that for a wide range of @xmath8 and @xmath9 , both sets of states are partially occupied .
this , we believe , underlies the ubiquitous two ` phase ' coexistence observed @xcite .
the high sensitivity of the physical properties of manganites to small perturbations arises because the @xmath73 states , being localized and lattice distorted , are strongly influenced by local perturbations , which then affects the delicate relative stability of @xmath5 and @xmath6 states .
finally , ( again as discussed in more detail below , ) the existence of the localized polaronic @xmath5 states _ in the presence of large @xmath10 and @xmath43 _ give rise to a _
new , major , doping dependent ferromagnetic nearest neighbour exchange coupling _
@xmath76 between the @xmath39 core spins .
this comes about due to _ virtual , fast ( adiabatic ) _ hopping processes of the @xmath5 electrons to neighbouring empty sites and back , leading to a @xmath76 roughly of order @xmath77 , the intermediate state energy due to the unrelaxed lattice distortion being @xmath78 .
we get @xmath79 for @xmath80 .
our calculations suggest that this _ virtual , correlated double exchange due to the localized @xmath5 electrons _ , and not conventional double exchange due to mobile @xmath3 electrons as hitherto believed , is the dominant source of ferromagnetism and of the ferromagnetic transition temperature @xmath16 in the hole doped manganites .
based on the above ideas , and for the purposes of making quantitative calculations and predictions , we have proposed@xcite a new model hamiltonian for manganites given by @xmath81 here @xmath82 creates the jt polaronic state of energy @xmath83 and spin @xmath84 , localized at site @xmath85 ( @xmath86 is the corresponding destruction operator ) .
the broad band electron , created by the operator @xmath87 , has mean energy zero , and nearest neighbour effective hopping amplitude @xmath66 .
the two repel each other on site with energy @xmath10 .
the common chemical potential @xmath88 is determined by the constraint that the filling , i.e. , the average number of @xmath3 electrons per site must be determined by the doping @xmath8 according to @xmath89 the term @xmath90 in eq .
[ eqhlb ] models the spin dependent interactions , and is given by @xmath91 it includes the strong ferromagnetic hund s rule coupling @xmath43 between the @xmath3 spins @xmath92 and the ( spin @xmath93 ) @xmath39 spins @xmath94 , the net effective ferromagnetic exchange coupling @xmath76 between these @xmath39 spins ( from the mechanism alluded to above , and discussed in more detail below ) , and the interaction of the latter with an external magnetic field @xmath95 @xcite . for simplicity , in what follows we further approximate the @xmath39 spins as classical fixed length spins whose directions fluctuate , i.e. write @xmath96 where @xmath97 is a unit vector .
we now discuss how the above model hamiltonian can be motivated from the microscopic hamiltonian for the manganites in the limit of strong jt interaction . a reasonably realistic microscopic model hamiltonian or energy operator @xmath72 of @xmath12 electrons ( @xmath3 and @xmath39 ) in the manganites has three types of contributions , namely those involving only the @xmath3 electrons ( @xmath98 ) , the coupling of these electrons with the jt lattice modes as well as the energy of these lattice modes themselves ( @xmath99 ) , and the part @xmath90 which involves the @xmath3 and @xmath39 spins .
thus @xmath100 where @xmath101 @xmath102 here the operators @xmath103 respectively add and remove an electron at site @xmath85 , in the spin state @xmath84 which can take two values , and in the two @xmath3 orbital states with labels @xmath104 corresponding to @xmath105 and @xmath106 corresponding to @xmath107 ; i.e. , @xmath108 is a short hand for @xmath109 .
the number operator @xmath110 is the usual sum @xmath111 .
the @xmath3 electron has site energy @xmath112 , equal to 0 for a clean system .
the ( anisotropic and orbital dependent ) nearest neighbour hopping _ matrix _ is @xmath113 and @xmath10 is the repulsion between electrons at @xmath85 in different states .
the first set of terms in @xmath114 describes the coupling of @xmath3 electrons with the two jt lattice modes @xmath115 and @xmath116 with strength @xmath41 , ( @xmath117 being the usual pauli matrices in orbital space ) the former displacement leading to a splitting and the latter to a mixing of the two @xmath3 states .
the last two terms in @xmath118 are the potential and kinetic energies of the modes neglecting anharmonic and intersite terms .
@xmath90 is the same magnetic hamiltonian as in eqn .
[ eqhs ] ( except that , in the context of the microscopic hamiltonian , the @xmath3 spin @xmath119 @xmath120 where @xmath121 is the pauli spin operator ) . in manganites
the jt coupling @xmath41 is large compared to @xmath122 ( where @xmath56 is the typical size of the jt distortion ) , and one can first concentrate on the single site , coupled electron - phonon problem .
the schrodinger equation corresponding to @xmath123 can be solved exactly at a single site .
@xmath124 is like a local pseudo - magnetic field splitting the pseudo - spin @xmath125 orbital levels 1 and 2 , hence @xmath126 has eigenvalues @xmath127 , where @xmath128 is the magnitude of the jt distortion .
the diagonalization of @xmath126 can be done explicitly using the operators @xmath129 and @xmath130 where @xmath131 , i.e. , it determines the orientation of the jt distortion . when an electron is present at site @xmath85 in the @xmath132 state , the effective lattice potential energy at site @xmath85 in @xmath123 has the form @xmath133 this has a minimum at @xmath134 ( the jt polaronic distortion ) with an energy lowering of @xmath135 .5 ev independent of @xmath136 @xcite .
the jt polaronic @xmath5 state that we referred to above is this @xmath137 state _ together with the jt distortion_. when an electron is present in the other ` anti ' jt , @xmath6 state , the effective lattice potential energy is @xmath138 with a minimum at zero displacement so that no distortion is associated with it .
thus , when an @xmath5 polaron is present at a site , low energy lattice configurations around the deep minimum of @xmath139 involve a shift of @xmath140 by an amount @xmath56 at each site .
this is formally achieved by a lang - firsov transformation @xcite @xmath141 where @xmath142 is the radial momentum operator conjugate to @xmath140 . in other words ,
formally , one has @xmath143 .
the huang - rhys reduction factor@xcite @xmath144 we have alluded to earlier , which multiplies the hopping involving the @xmath5 states , is just the ground state expectation value of @xmath145 , corresponding to the anti - adiabatic limit overlap of the distorted phonon wavefunction at a site with the undistorted phonon wavefunction .
( we ignore fluctuations on the basis that @xmath146 @xcite ) .
the coherent intersite hopping of the @xmath5 electrons is thus @xmath147 , very small , and can be neglected to a first approximation .
the ` anti ' jahn teller state @xmath6 has no associated lattice distortion , so its hopping amplitude @xmath148 is not reduced , but depends on the angles @xmath136 and @xmath149 at sites @xmath85 and @xmath150 .
we assume a nearest neighbour hopping amplitude @xmath151 averaged over angles @xmath136 either statistically ( and classically ) or by quantum fluctuations .
such an _ orbital liquid _
approximation is reasonable for the phases with no long range orbital order , i.e. , for @xmath152 in most manganites , but poor for those other values of @xmath8 for which one has strong orbital correlations or long range orbital order .
the localized polaronic @xmath5 and the broad band @xmath6 states coexist in manganites since the conditions for anti - adiabaticity of the former , namely @xmath153 and the adiabaticity of the latter , namely @xmath154 , are _ both _ fulfilled .
the @xmath6 electrons are repelled strongly by the @xmath5 electrons , with energy @xmath155 , and hence will have the largest amplitudes at sites where the @xmath5 electrons are not present , i.e. at the hole sites .
they hop quickly ( time scale @xmath156 ) among such sites , avoiding strongly the sites with static @xmath5 electrons , where they hence have small amplitudes@xcite .
finally we discuss the result alluded to above namely that the existence of localized , jt distorted @xmath5 states in the presence of large @xmath43 and large @xmath10 can give rise to a new mechanism of ferromagnetic exchange between the @xmath39 core spins , which we refer to as `` virtual double exchange '' .
suppose that a localized @xmath5 electron is present at a site , and that site is distorted .
the @xmath5 electron can take part in fast ( adiabatic ) virtual hopping processes to neighbouring sites , i.e. , leaving the local lattice distortion unrelaxed , by paying an energy cost of @xmath78 in the intermediate state . for large @xmath10
this can happen only if the neighboring state is empty , and for large @xmath43 only if the @xmath39 spins on the two sites are parallel .
( otherwise the energy of the intermediate state will increase by @xmath10 and @xmath43 respectively . ) clearly , from second order perturbation theory this process will give rise to a _
new , @xmath5 occupancy dependent , ferromagnetic exchange coupling _ between the @xmath39 core spins of the form @xmath157\ ] ] ( where we have again ignored the dependence on the angles @xmath136 etc )
. the @xmath158 factor comes from large @xmath43 , and the occupancy dependent terms from large @xmath69 . within the homogeneous orbital liquid approximation we are using in this paper
, this translates to an effective ferromagnetic interaction proportional to @xmath159 as stated earlier .
the normal super - exchange coupling between @xmath39 core spins on neighboring sites is of order @xmath160 , and its sign depends on the nearest neighbour orbital correlations @xcite . for example , in @xmath35 it is anti - ferromagnetic along the c axis , and ferromagnetic in plane .
conventional double exchange can , in principle , explain metallic ferromagnetism in the manganites , but would be unable to account for ferromagnetism in their insulating phases in the absence of orbital correlations .
the above mechanism yields a ferromagnetic coupling which is clearly much larger than and dominates the super - exchange for all intermediate doping , operates even in the insulating states and even in the orbital liquid phase , and as we discuss later , makes the dominant contribution to the ferromagnetic @xmath16 even in the metallic case .
putting all this together , expressing the on site coulomb and exchange interactions in the @xmath5 and @xmath6 basis , and including the new exchange mechanism , leads us to the suggestion that in the large @xmath161 ) limit , and for describing low energy electronic properties the microscopic model hamiltonian in eq .
[ eqhmicro ] can be approximated by the much simpler , two fermion species , hamiltonian @xmath162 given in eq .
[ eqhlb ] . the relatively small @xmath163 term and other coulomb repulsion terms ( between @xmath5 and @xmath6 electrons of different spins )
have been neglected in @xmath164 because of the large @xmath43 ( @xmath165 ) limit of relevance here ( whence the spin of the @xmath5 and @xmath6 electrons is forced to be parallel to the @xmath39 spin ) , and we have relabelled the effective repulsion between the two types of electrons as @xmath10 .
apart from the constraint equation eq .
[ eqnconstr ] and the spin dependent @xmath90 , @xmath162 is essentially the well known falicov - kimball model ( fkm)@xcite for non hybridizing @xmath14 electrons in a broad band metal with correlations .
indeed at @xmath166 , in the ferromagnetic phase and for large @xmath43 , the spin degrees of freedom are completely frozen , whence it reduces to the ( spin - less ) fkm .
more generally , it describes the dynamics of the @xmath6 electrons moving in an _ annealed disordered _ background of immobile @xmath5 electrons and @xmath39 spins in the presence of strong on site repulsion @xmath10 and hund s coupling @xmath43 , the annealed disorder distribution being _ thermodynamically and self - consistently determined_. this strong correlation problem can not be solved in general , but can be solved within the framework of dynamical mean field theory ( dmft ) @xcite which is exact at @xmath167 , and is quite accurate for three dimensions .
we have carried out these dmft calculations for thermodynamic properties ( the occupancies @xmath168 , @xmath169 , the magnetization @xmath170 , the specific heat , magnetic susceptibility , etc . ) , spectral behaviour ( eg . the @xmath6 band self - energy , propagator and density of states(dos ) ) and transport ( resistivity and magnetoresistance ) .
while the specific results we discuss below come from such calculations , we believe that many of the physical mechanisms that underlie the results have a larger sphere of validity .
the properties of @xmath162 are determined largely by the dynamics of the @xmath6 electrons , eg .
, their propagator or green s function @xmath171 , as affected by the ( annealed disordered ) distribution of @xmath172 and @xmath173 due to strong @xmath10 and @xmath43 . in the dmft framework , the @xmath6 electron self energy @xmath174 due to these interactions is site local , i.e. @xmath175 , and is determined self consistently .
the general approach is well known @xcite .
essentially , the problem reduces to that of one site , with its local degrees of freedom and interactions , immersed in an effective medium ( comprising the other sites ) .
we make the simplest reasonable approximation corresponding to a homogeneous annealed system , namely that @xmath176 , @xmath177 and @xmath178 are the same at every site @xmath85 , and that there are no correlations between these quantities at different sites@xcite .
consistent with this , we approximate the ferromagnetic exchange interactions in @xmath90 in terms of a homogeneous molecular field in the standard way and replace @xmath179 here @xmath180 , @xmath181 where @xmath182 is the co - ordination number .
the quantity @xmath183 , proportional to the magnetization , and @xmath184 , the average occupancy of the polaronic states , are also to be determined self - consistently .
thus , in the present context , the effective medium is a homogeneous electron bath with which the local @xmath6 electron hybridizes , the molecular field @xmath185 in eq .
[ eqhsmofld ] , and the local distributions of @xmath186 and @xmath187 .
the on - site @xmath6 electron propagator due to the medium , but without the local interactions , is @xmath188 where the ( matsubara ) frequency variable and the spin label are suppressed . the single site problem can be solved exactly , just as in the falicov - kimball case @xcite , to determine ( 1 ) the ( homogeneous ) annealed distributions @xmath189 and @xmath190 as functionals of @xmath188 , @xmath191 and the model parameters and ( 2 ) the local @xmath6 electron propagator @xmath192 as a functional of @xmath188 , @xmath191 , the model parameters and these annealed distributions .
there are two other formal relations among @xmath192 , @xmath193 and @xmath194 , namely @xmath195 and @xmath196 $ ] . by iterating these equations to self - consistency
, one can determine @xmath194 , @xmath192 , @xmath188 , @xmath189 and @xmath190 ( and hence @xmath197 , @xmath198 and @xmath191 ) explicitly for any chosen values of the model parameters @xmath10 , @xmath199 , @xmath66 and @xmath200 at a fixed @xmath9 and @xmath88 ( the latter being chosen such that @xmath201 ) .
thermal , spectral and transport properties can be calculated in a straightforward way @xcite from the @xmath6 electron propagator .
for example , the current - current correlation function ( the kubo formula ) which determines the electrical ( and optical ) conductivity , can be expressed entirely in terms of @xmath171 , vertex corrections being negligible . the resulting calculations , while straightforward , still involve extensive numerics .
the calculations whose results are described below have been done in the large @xmath43 limit , i.e. assuming that the spin of the @xmath5 and @xmath6 electrons are forced to be aligned along @xmath39 spin @xmath202 , whence there is considerable simplification @xcite .
we have done such calculations both for a realistic model @xmath203 and for a model semicircular density of states .
the two results are not very different , and here we mostly discuss results for the latter case , which has the advantage that some exact analytical results can be obtained in the @xmath204 , @xmath205 limit , and even the finite @xmath10 , non - zero @xmath9 calculations are much simpler .
the details are described in refs .
@xcite and other papers we will publish elsewhere .
to start with , we show how the ideas above lead to a simple physical understanding of insulator metal transitions ( imt ) in manganites , based on the variation of the occupancies and relative energies of the @xmath5 and @xmath6 states with @xmath206 etc .
[ figt0dos ] shows the density of states for @xmath207 , @xmath208 , @xmath209 ( and @xmath210 ) for several values of @xmath8 at @xmath211 . at @xmath166 , because of @xmath76 the ground state is ferromagnetic , and the spin degrees of freedom are completely frozen . for large @xmath10
, the renormalized or effective @xmath6 bandwidth @xmath212 self consistently goes to zero as @xmath213 , whence for small @xmath8 the bottom of the @xmath6 band will be above the @xmath5 level ( of energy -@xmath214 ) and all electrons are in the latter states , as in fig .
2a . for , the low energy @xmath6 band states
are constrained to reside mostly on the small fraction @xmath8 of empty , undistorted sites , and have a small amplitude at the lattice distorted sites ( where @xmath5 electrons are present ) from which they are repelled with energy @xmath10 .
the system is hence an insulator .
this can not happen in a pure double exchange model , where ferromagnetism and metallicity go together .
as @xmath8 increases , so does @xmath212 , and there is a critical @xmath215 (= 0.25 within dmft for the parameters used in fig.2 ) at which the @xmath6 band bottom touches the @xmath5 level energy as in fig.2b . for @xmath216 , the @xmath6 band bottom goes below the @xmath5 level , and some of the @xmath5 electrons are transferred to the @xmath6 band until all @xmath6 levels up to the ( now renormalized ) @xmath5 level are occupied as in fig .
the system is then a ferromagnetic metal , but most of the electrons ( 0.6 per site out of the 0.7 present ) are in the polaronic @xmath5 state though there is no long range jt order .
thus there is an imt , not at @xmath217 as in a naive doped mott insulator picture , but at @xmath218 .
this shift is due to the strong electron lattice coupling @xmath41 which makes the jt polaronic @xmath5 level possible and stabilizes it by an energy @xmath214 , and large @xmath10 , which makes the @xmath6 electron bandwidth vanish as @xmath213 . for @xmath219 , within dmft we can show@xcite that @xmath220 whence we obtain @xmath215 analytically as @xmath221 where @xmath222 is the bare half - width of the b - band .
the decrease of @xmath215 with increasing @xmath222 and its numerical value for reasonable choices of @xmath214 and @xmath222 are in accord with experiments ( for example , as one goes from @xmath223 to @xmath224 which has a larger bandwidth ) .
finite @xmath10 corrections to @xmath215 , which we have calculated , are small for @xmath225 which is appropriate to these systems .
a more detailed discussion of these and other zero temperature results is presented in @xcite .
fig.[figchap4sr5 ] shows the evolution of the spectral functions with temperature , for parameter values that correspond to @xmath226 doping , for a fixed doping of @xmath227 for this system , and for the temperature values @xmath228100k , 180k , 208k , 300k . for @xmath216 ,
the system is metallic so that both @xmath6 and @xmath5 electronic states occupied . on increasing the temperature , the @xmath39 spin order decreases and the @xmath6 band narrows again , since due to double exchange @xcite the effective hopping amplitude of the @xmath6 electrons decreases .
the spectral weight for the down spin polarization also turns on , reflecting the reduction in the magnetization from its zero temperature value .
the sequence of figures 3a to 3c illustrates this progression .
eventually , both up and down spin spectral functions become equal , and the @xmath6 band bottom crosses the @xmath5 level and moves up as @xmath9 rises beyond @xmath16 (= 208 k in the present context , see fig .
3c ) where ferromagnetism disappears , and the system becomes a paramagnetic insulator as in fig .
3d , with @xmath6 states occupied thermally across a gap .
we thus have a simple picture of the thermal ferromagnetic - metal to paramagnetic - insulator transition .
there are very few calculations @xcite of the transport properties of manganites though this is one of their main unusual features . within the dmft for our model
, the electrical resistivity @xmath229 can be calculated in a straightforward way@xcite from the @xmath6 electron propagator or green s function .
4 shows some of our results , for @xmath230 0.3 and parameters representative of @xmath231 and @xmath232 @xcite .
experimental values are also given for comparison @xcite . .
[ fig - res - cmr ] we see that the resistivity of the paramagnetic insulating state is fairly well described by the theory . in this phase ,
the conducting @xmath6 band is only thermally occupied . the effective electrical gap @xmath233 ( the experimental value is @xmath234@xcite ) . the positive feedback between the thermally excited , strongly @xmath9 dependent number @xmath235 of @xmath6 electrons , and the double exchange broadening of the @xmath6 band as @xmath9 decreases , leads to a very rapid decrease and closure of the electrical gap just below @xmath16 and the consequent rapid decrease in the resistivity just below @xmath16 as in experiment to a rather large value @xmath236 .
but we note that the calculated resistivity does not decrease much thereafter down to @xmath166 , unlike experiments , where @xmath237 decreases to residual values of @xmath238 below about @xmath239 .
we believe that the latter is due to inter - site @xmath5 coherence , neglected here ( see later ) .
we also show in the fig .
[ fig - res - cmr ] the resistivity in a field of @xmath240 for @xmath241 .
we clearly see cmr .
the external magnetic field polarizes the @xmath39 spins , and via @xmath242 increases the @xmath6 band width which reduces the electrical gap .
cmr arises because resistivity depends exponentially on the gap .
the effect is largest near @xmath16 as the change in the magnetization due to a given magnetic field is the largest there .
as our mean field approximation for the curie transition neglects short range magnetic order above @xmath16 , @xmath229 is somewhat overestimated and the cmr somewhat underestimated .
the properties of manganites vary strongly and characteristically with the ionic species @xmath20 and @xmath21 .
the material systematics of such variation , eg . , in the thermal imt as well as the cmr depend in our model largely and sensitively on the ratio @xmath243 and somewhat on the @xmath222 and @xmath76 . a small increase in @xmath244
reduces the high temperature resistivity @xmath245 enormously , as is clear in fig .
[ fig - res - cmr ] where we have increased @xmath222 from 1.05 to 1.15 @xmath45 and @xmath246 from 52.7 to 60.2 @xmath247 to fit @xmath16 and @xmath248 for @xmath249 .
broadly this is because the density of current carrying @xmath6 electrons and hence @xmath250 depends exponentially on @xmath251 .
this is also the general reason for the observed large variation of @xmath252 and @xmath16 with strain and pressure . making the reasonable assumption that @xmath214 , a single octahedron quantity , is roughly unchanged across materials , but that @xmath246 , the effective ferromagnetic exchange , scales as @xmath253 , we have calculated properties as a function of @xmath222 for @xmath254 .
we find that @xmath16 is typically fairly close to the ( curie - weiss ) value ( @xmath255 ) ( and hence increases as @xmath256 as @xmath222 increases ) , with a slight enhancement due to a nonzero @xmath257 and double exchange . with increasing @xmath222 , the resistive transition on lowering @xmath9 through @xmath16 changes from insulator - insulator to insulator - metal , and then to metal - metal .
these trends are seen experimentally @xcite as a function of increasing cation radius @xmath258 , which is known to roughly track the bandwidth @xmath222 .
we also find that the relative cmr , defined as @xmath259/r(h)$ ] , depends exponentially on @xmath16 ( increasing as @xmath16 decreases ) .
this is indeed the behaviour extracted from measurements on several systems @xcite .
the observed unusually large dependence of @xmath16 and of the resistivity at @xmath16 on pressure @xcite in @xmath260 is reproduced if we assume that @xmath222 increases with pressure at a realistic rate of about .01 @xmath261 .
a more detailed discussion of these finite temperature results and material trends is presented in @xcite .
a key feature of our theory is the result mentioned above that the dominant contribution to @xmath16 comes from virtual double exchange of the localized @xmath5 polarons and not from conventional double exchange , since the mobile carrier density ( @xmath262 ) is very small .
interestingly , the precipitous , universal and nearly linear drop in @xmath16 with local ion size variance reported by attfield@xcite in a variety of multi - component manganites is much larger than what is expected from a simple double exchange picture . in the double exchange mechanism
, @xmath16 is proportional to @xmath148 which depends on the @xmath263 bond angle @xmath264 , which is strongly affected by the local ion - sizes , as @xmath265 .
the variance in @xmath264 , namely @xmath266 is at most 10 degrees , so that one expects that the resulting reduction in @xmath16 is @xmath267 .
the observed reduction is an order of magnitude larger !
we believe that such a large reduction is however understandable within our theory if we extend the calculations described above to include the effect of ion size variance on @xmath214 , and @xmath268 hybridization effects as discussed in section 8 .
we now illustrate through three examples how other unusual properties of manganites can be understood in our approach .
specifically , we consider the anomalously low carrier density as inferred from optical conductivity in the metallic phase @xcite , the small electronic specific heat @xcite and the electron hole asymmetry @xcite . in the metallic regime @xmath269 , one can estimate @xmath270 , the effective density of carriers , from the the drude weight i.e. the frequency integral of the real part of the optical conductivity @xmath271 . in a free electron like limit , this integral is @xmath272 where @xmath273 is the density of carriers , and @xmath274 is the band optical mass .
since there is no evidence for enhancement of @xmath274 with respect to the band mass ( eg . see ref.@xcite ) , results for @xmath275@xcite lead to @xmath276 for @xmath277 and @xmath166 .
the expected value ( for a doped mott insulator ) is @xmath278 .
although there are several uncertainties in inferring @xmath270 precisely from @xmath279 , its reduction by nearly an order of magnitude , and even more so its observed strong decrease with temperature on the scale of @xmath16 which is much less than the ( normal ) fermi temperature , are puzzling .
but the results are easily understood in our model , since the low frequency @xmath280 is due to @xmath6 electrons . as seen from fig .
( 2c ) at @xmath166 there are indeed very few electrons in the @xmath6 band , all near its bottom , below the @xmath5 ( or fermi ) level .
the dmft calculation with parameters appropriate for @xmath281 at @xmath282 , namely @xmath283 ( and consequently @xmath284 ) , leads to @xmath285 , and a decrease with temperature which is very similar to what is seen in experiment . the low @xmath9 electronic specific heat in manganites ( see ref .
@xcite for a review ) goes as @xmath286 with a @xmath287 nearly independent of @xmath8 , unlike other doped mott insulators(eg ref .
@xcite ) and has a value expected for a tight binding band with bandwidth @xmath288 . in polaron theories with @xmath289 polarons per site @xmath290 in the classical gas limit , and @xmath291 , where @xmath292 is the effective bandwidth , for @xmath293 . neither of these
is observed and this is often argued to be evidence against small polarons in manganites . in our model , although the electrons are mostly jt polarons , we expect that they will make very little contribution to the low temperature specific heat .
for , being localized they will strongly couple to disorder potentials and affected by coulomb interactions , and we expect that their entropy will be frozen out at a large temperature scale . we find that @xmath287 ( calculated in the paramagnetic metallic phase to avoid large magnetic contributions to the specific heat ) is about @xmath294 with the parameters used above for @xmath295 . the value for a band of width @xmath296 is about @xmath297 .
the physical reason for the smaller @xmath287 is that the few @xmath6 electrons which are present occupy states near the bottom of the @xmath6 band with energy close to @xmath83 , and the density of states there is relatively smaller(fig .
2c ) . the electronic and magnetic properties of hole doped @xmath298 and `` electron doped '' ( @xmath299 ) manganites are very different .
for example , the latter are much more metallic than the former in the paramagnetic phase @xcite ( although the af or charge ordered phases can be insulating for other reasons ) .
this behaviour is unexpected in a one orbital strong coupling antiadiabatic polaron model since for low @xmath3 carrier density the dilute small polaron assembly will form an insulator , as also in a model with only @xmath242 , where there is electron hole symmetry .
it is often stated that polaronic effects are not seen for @xmath3 orbitals in the @xmath300 doped regime because the jt coupling becomes weaker as @xmath301 .
this seems unlikely given the local nature of the jt interaction and the stability of the ( @xmath302 ) octahedron .
the metallic nature of the @xmath300 doped regime which has low @xmath3 density is natural in our theory because for large @xmath8 the effectively uncorrelated @xmath6 electrons form a wide band whose bottom is occupied by the small number of @xmath3 electrons , and the @xmath5 states are unoccupied , as depicted in fig .
this fact also explains why pure band models are fairly successful@xcite in describing the magnetic ground states in the the @xmath300 doped limit .
in summary , we have presented here a new model of coexisting localized jt polarons and broad band electrons for manganites , argued that it arises inevitably in the presence of orbital degeneracy and strong jt coupling , and shown that it explains a wide variety of characteristic properties of manganites . in this concluding section
we compare our theory with some of the other theories , discuss some of the inadequacies of the present model and extensions to overcome these .
we believe that a more general treatment of the model eq . [ eqhlb ] with some extensions can lead to a complete description of manganites .
two examples , namely inclusion of spatial correlations and of intersite @xmath5 state coherence are discussed below . to begin with , we note that the picture developed here is very different from that in other polaronic models , which either neglect @xmath3 orbital degeneracy @xcite or work in the adiabatic approximation @xcite . in the former , for example , at high temperatures transport is due to the activated hopping of localized small polarons . in the latter ( adiabatic ) models ,
the polarons also form a broad band , whence it is difficult to obtain a paramagnetic insulator to ferromagnetic metal transition say at @xmath254 . furthermore
, small polarons disappear below @xmath16 even for large @xmath41 , the drude weight is not small , and there is no isotope effect , all in disagreement with experiment . in both , it is argued that small polarons are likely only at high temperatures , and that small and large polarons coexist only in a narrow range of @xmath8 and @xmath9 determined by the effective electron phonon coupling and that there are no small polarons at low @xmath9 . in our theory ,
the carriers are broad band electrons thermally promoted out of localized jt polaronic states .
small jt polarons ( in the anti - adiabatic limit , with negligible bandwidth ) and band states(in the adiabatic limit ) necessarily coexist over a wide range of @xmath8 @xmath303 say ) _ and for all _ @xmath9 .
in contrast , in single orbital polaron models , there is a crossover with increasing dimensionless electron phonon coupling @xmath304 from large to small polarons , so that the two can occur together only over a narrow range of @xmath8 and @xmath9 . _
an unusual feature of the polaron level in our model is that it has no prominent thermodynamic or spectroscopic ( sharp level ) manifestations . _
firstly as pointed out above , it does not lead to a large electronic specific heat as we expect that their entropy will be frozen out at energy scales determined by coulomb and impurity interactions not explicitly included in our model .
the remaining linear specific heat for the @xmath6 band metal , is roughly of the size seen experimentally@xcite .
secondly , no sharp @xmath5 like spectral feature is to be expected .
rather , the @xmath5 excitation spectrum will be an incoherent continuum , starting from a weak ( weight @xmath305 ) anti - adiabatic low energy part building up to adiabatic ( franck - condon ) higher energy features , as the fast removal of an @xmath5 electron leads to highly excited lattice states ( energy @xmath306 ) . indeed , optical conductivity data in manganites do show@xcite such an incoherent continuum which is largely independent of temperature , and we believe that this is their origin . in our theory
presented earlier we made the simple approximation that the local @xmath64 distortion @xmath307 is independent of @xmath8 and @xmath9 , and that its large value is determined entirely by on - site properties , namely the electron @xmath64 phonon coupling @xmath41 and the phonon mode force constant @xmath308 .
this is probably valid for high temperatures not too far below @xmath16 and for @xmath309 .
however , jt polaronic effects can become dynamical and unobservable at low @xmath9 for two reasons .
firstly , the local lattice kinetic energy has a term @xmath310 where @xmath307 is the magnitude , and @xmath136 the direction or orientation , of the local jt distortion .
hence @xmath136 , which also determines the orbital mixing amplitudes of the @xmath5 and @xmath6 states , rotates due to quantum fluctuations , although due to anharmonicity and crystallinity , the rotation could be hindered and slow . a second ,
perhaps more important , source of this dynamical @xmath311 effect is the quantum mechanical intersite @xmath5 state coherence which makes the @xmath136 ill defined .
the energy scale for the latter , @xmath312 , is consistent with the widely observed @xcite reduction or disappearance of static or of long time scale local @xmath64 distortion in metallic manganites as their temperature decreases well below 125k .
finally , intersite @xmath5 coherence , especially hybridization with energetically degenerate broad band @xmath6 states in the metallic regime , will reduce @xmath307 and broaden the @xmath5 band .
the two effects feed back on each other , and it is in principle possible that small polarons may weaken and disappear altogether at low temperatures in the metallic phase rather than merely becoming dynamic . we have not explored this possibility here .
intersite @xmath5 state coherence can be included in our model by adding to eq.(1 ) an @xmath268 hybridization term @xmath313 $ ] where @xmath314 , and @xmath315 .
we expect that its inclusion , and the consequent development of long range intersite @xmath5 state coherence at low temperatures will , in addition , lead to the observed smooth decrease @xcite in the electrical resistivity of clean metallic manganites from about @xmath316 just below @xmath16 to a small value , @xmath317 , as @xmath318 past a characteristic crossover temperature of about 100k , of the same scale as @xmath319 . in the dmft calculations presented above ,
the residual resistivity at @xmath166 due to random scattering from the incoherent @xmath5 sites is too large , @xmath320 ( fig.3 ) .
if the @xmath5 states become coherent , the @xmath6 electron scattering and the consequent resistivity would then vanish at @xmath166 , and is nonzero only if static disorder is present .
this can lead to a metallic state with a small residual resistivity or to an anderson localized insulating state depending on the amount of disorder .
we believe that the giant isotope effect observed in manganites @xcite is another dramatic consequence of the @xmath268 hybridization scale @xmath321 and its exponential dependence on the inverse square root of the isotopic mass . for it can add , via ( conventional ) double exchange , to the ferromagnetic @xmath16 an amount roughly given by @xmath322 for @xmath323 where @xmath324 is less than one . because of the huang rhys factor in @xmath319 , which depends exponentially on square root of the isotopic mass , we find that @xmath325 so that for @xmath326 the difference @xmath327 , close to the observed value@xcite .
in addition to the neglect of intersite coherence effects discussed above , we have also neglected in our theory spatial correlations between the local electron densities @xmath328 and @xmath329 , as well as between the local angles @xmath330 .
this may be adequate for the phases which correspond to _ homogenous orbital liquids_. however , many phenomena in manganites such as short range order , long range charge / orbital order ( or _ orbital solidification _ ) , various types of antiferromagnetism @xcite and mesoscale structures @xcite depend on spatial correlations . to treat these we need to add to eq .
( 1 ) a number of more complicated longer range coulomb , anharmonic , steric , elastic , magneto - elastic , etc .
interactions that couple spin , orbital and lattice degrees of freedom to each other , and to strain , ion size mismatch , disorder , etc . as appropriate .
a self consistent determination of long or short range order in @xmath331 ( a new ` internal ' variable)@xcite , @xmath136 and @xmath332 , and their effects on the @xmath6 electron dynamics for different @xmath8 and @xmath9 , in the presence of such interactions can lead to a complete description of manganites including the above phenomena , and others such as first order transitions , two phase co - existence , etc .. we hope to discuss these elsewhere . a significant general question raised by our work is that of the detailed nature of adiabatic to non adiabatic crossover as @xmath41 increases , and the conditions for the _ coexistence of adiabatic and antiadiabatic states with exponentially separated dynamical timescales_. we have shown that the latter leads to new phenomena , argued that this happens in manganites because of the large @xmath41 , and have developed the consequences in a simple model assuming this separation .
the results closely correspond with a wide variety of observations .
it would be of great interest to explore such a crossover and time - scale or energy - scale separation experimentally and theoretically in the many systems such as organic solids , transition metal oxides and chemical ( molecular ) systems , which have degenerate orbitals and strong symmetry breaking jt couplings .
we would like to acknowledge support from several agencies .
hrk has been supported in part by grant no .
2404 - 1 of the indo french centre for promoting advanced research , tvr in part by a us - india project onr n 00014 - 97 - 0988 , and srh and gvp in part by the council of scientific and industrial research , india .
99 see for example , p. w. anderson , _ the theory of superconductivity in the high-@xmath16 cuprates _
( princeton univ . press , princeton , 1997 ) .
s. jin , et al .
_ science _ * 264 * , 413 ( 1994 ) .
y. tokura , ed .
_ colossal magnetoresistance oxides_(gordon and breach , new york , 2000 ) .
m. b. salamon , and m. jaime , _ rev .
phys . _ * 73 * , 583 ( 2001 ) .
n. d. mathur and p. b. littlewood , _ solid state communications _ * 119 * , 271 ( 2001 ) .
these could arise from the mismatch of @xmath20 and @xmath21 ion size with respect to the value for the ideal close packed @xmath333 structure , see for example : h. y. hwang , et .
lett . _ * 75 * , 914 ( 1995 ) .
j. p. attfield , _ chemistry of materials _ * 10 * , 3239 ( 1998 ) .
n. a. babushkina , et al , _ nature _ * 391 * , 159 ( 1998 ) .
e. dagotto , t. hotta , and a. moreo , _ physics reports .
_ * 344 * , 1 ( 2001 ) .
see also , e. dagotto , _ nanoscale phase separation and colossal magnetoresistance _ ( springer , berlin , 2002 ) m. uehara , s. mori , c. m. chen , and s. w. cheong , _ nature _ * 399 * , 560 ( 1999 ) .
d. louca , et al.,_phys .
b. _ * 56 * , r8475 ( 1997 ) ; t. egami in _ structure and bonding _
j. b. goodenough ) vol .
98 , p115 ( 2001 ) c meneghini , et al ,
. cond . mat .
_ * 14 * , 1967 ( 2002 ) .
r. h. heffner , et .
_ * 85 * , 3285 ( 2000 ) .
m. fth , et al .
_ science _ * 285 * , 1540 ( 1999 ) .
s. satpathy , z. s. popovic and f. r. vukajlovic , _ phys .
lett . _ * 76 * , 960 ( 1996 ) .
d. d. sarma , et .
_ * 75 * , 1126 ( 1995 ) .
a. j. millis , _ phil . trans .
a _ * 356 * , 1473 ( 1998 ) .
large jt splitting is the traditional basis of our understanding of @xmath334 structure , see @xmath335 j. kanamori,_j .
phys . _ * 31 * , 14s ( 1960 ) .
g. m. zhao , et .
b _ * 63 * , 060402 ( 2001 ) .
n. furukawa , _ journal of the physical society of japan _ * 64 * , 2734 ( 1995 ) .
a. j. millis , r. mueller , and b. i. shraiman , _ phys .
b _ * 54 * , 5405 ( 1996 ) .
a. c. m. green and d. m. edwards , _ jl .
mater . _ * 226 * , 886 ( 2001 ) .
h. roder , jun zang , and a. r. bishop , _ phys .
* 76 * , 1356 ( 1996 ) .
m. uehara , k. h. kim and s. w. cheong , private communication . n. m. iliev and v. m. abrashev , _ j. raman spectrosc .
_ * 32 * , 805 ( 2001 ) k. huang and f. rhys , _ proc .
london ser .
a _ * 204 * , 406 ( 1950 ) t. v. ramakrishnan , h. r. krishnamurthy , s. r. hassan and g. v. pai cond - matt/0308376 , _ submitted for publication _ g. venketeswera pai , s. r. hassan , h. r. krishnamurthy and t. v. ramakrishnan _ submitted for publication _ the relatively small @xmath336 term in @xmath90 and other coulomb repulsion terms in @xmath164 have been neglected because of the large @xmath43 ( @xmath165 ) limit , of relevance here . for a discussion of such vibronic states with jt coupling , see m. d. sturge , in _ solid state physics _
( eds . f. sietz , d. turnbull and h. ehrenreich , academic press , new york . ) * 20 * , 91 ( 1967 ) .
g. i. lang and a. yu .
soviet physics jetp _ * 43 * , 1843 ( 1962 ) .
eg . , see t. v. ramakrishnan and g. v. pai , _ jl .
low temp.phys._ * 126 * , 1055 ( 2002 ) for a discussion of such issues in a simpler one orbital spinless polaron model . in the limit that this amplitude is zero
, we have two classical regions , one with small polarons and another with broad band states ; this is the percolative regime discussed in the literature ( eg .
t. mizokawa , d. i. khomskii and g. a. sawatzky , _ phys .
rev . b_*63 * , 024403 ( 2001 ) ; see also a. j. millis , _ phys .
b. _ * 55 * , 6405 ( 1997 ) l. m. falicov and j. c. kimball , _ phys .
* 22 * , 997 ( 1969 ) .
a. georges , g. kotliar , w. krauth and m. j. rozenberg , _ rev .
* 68 * , 13 ( 1996 ) .
we note that the approach of millis , muller and shraiman @xcite which we have commented on earlier is similar , except that they set @xmath337 and treat the jt distortions purely classically .
s. r. hassan , _ dynamical mean - field theory of falicov - kimball and related models and their application to cmr phenomana in manganites _ , ph .
d. thesis ( indian institute of science , bangalore ) 2003 ( unpublished ) .
p. w. anderson and h. hasegawa , _ phy .
rev . _ * 100 * , 675 ( 1955 ) .
we choose @xmath338 and @xmath339 , for both the systems .
@xmath340 is @xmath341 and @xmath342 , and @xmath343 is @xmath344 and @xmath345 respectively .
these choices are made such that the observed @xmath16 as well as @xmath346 are reproduced .
the electrical activation energies above @xmath16 , @xmath347 , are in reasonable agreement with @xmath348 respectively .
further the steep drop in @xmath229 below @xmath16 is well reproduced .
p. dai , et .
_ phys . rev .
b _ * 61 * , 9553 ( 2000 ) . k. khazeni et .
lett . _ * 76 * , 295 ( 1996 ) .
j. j. neumeier , m. f. hundley , j. d. thompson and r. h. heffner _ phys .
b _ * 52 * , r7006 ( 1995 ) .
y. okimoto , et.al .
b _ * 55 * , 4206 ( 1997 )
. a. maignan , c. martin , f. damay and b. raveau , _ phys .
b _ * 58 * , 2758 ( 1998 ) .
g. v. pai , _ phys .
b _ * 63 * , 064431 ( 2001 ) m. s. laad , l. craco and e. muller - hartmann , _ phys . rev . b _ * 63 * , 214419 ( 2001 ) .
variations in the local charge density @xmath349 are likely to be small and of short range because of the large coulomb energy cost .
( a coulomb energy term needs to be added to eq .
[ eqhlb ] to describe this ) . | the electronic properties of many transition metal oxide systems require new ideas concerning the behaviour of electrons in solids for their explanation .
a recent example , subsequent to that of cuprate superconductors , is of rare earth manganites doped with alkaline earths , namely @xmath0 , which exhibit colossal magnetoresistance , metal insulator transition and many other poorly understood phenomena .
here we show that the strong jahn teller coupling between the twofold degenerate ( @xmath1 and @xmath2 ) @xmath3 orbitals of @xmath4 and lattice modes of vibration ( of the oxygen octahedra surrounding the @xmath4 ions ) dynamically reorganizes the former into a set of states ( which we label @xmath5 ) which are localized with large local lattice distortion and exponentially small intersite overlap , and another set ( labelled @xmath6 ) which form a broad band .
this hitherto unsuspected but microscopically inevitable @xmath7 of radically different @xmath5 and @xmath6 states , and their relative energies and occupation as influenced by doping @xmath8 , temperature @xmath9 , local coulomb repulsion @xmath10 etc . , underlies the unique effects seen in manganites .
we present results from strong correlation calculations using the dynamical mean - field theory which accord with a variety of observations in the orbital liquid regime ( say , for @xmath11).we outline extensions to include intersite @xmath5 coherence and spatial correlations / long range order . |
in the one - dimensional _ asymmetric simple exclusion process _ ( asep ) , particles perform continuous - time random walks on @xmath4 with rate @xmath0 $ ] of jumping to the right and @xmath5 of jumping to the left .
particles interact by exclusion ; attempted jumps to occupied sites are suppressed .
the resulting process @xmath6 is markov in the state space @xmath7{0.2}{\circle}}}\}^{\mathbb{z}}$ ] , where for a site @xmath8 , @xmath9{0.2}{\circle}}}$ ] indicates that the site @xmath10 is occupied at time @xmath11 , and @xmath12 indicates that it is empty .
if @xmath3 then all jumps are to the right and the process is known as the _ totally asymmetric simple exclusion process _ ( ) .
a convenient way to construct the process is using a family of independent poisson processes @xmath13 , where for each @xmath10 , @xmath14 has rate @xmath15 and @xmath16 has rate @xmath17 . at the time of a point in the process @xmath18 , if there is a particle at site @xmath10 , it attempts to jump to site @xmath19 ; the jump succeeds if site @xmath19 is empty .
two versions of the process started from different initial conditions can be jointly constructed under the _ basic coupling _ by using the same @xmath18 for both versions .
a _ second - class particle _ , denoted @xmath20 , is a particle that interacts with particles like a hole , and interacts with holes like a particle .
that is , a second - class particle jumps right at rate @xmath15 if there is a hole on its right , and jumps left at rate @xmath17 if there is a hole on its left ; if the site on its left contains a particle , they exchange positions at rate @xmath15 , and if the site on its right contains a particle , they exchange positions at rate @xmath17 . a second - class particle can be regarded as a discrepancy between two systems evolving together under the basic coupling . if the systems start from two initial conditions which coincide except at a single site , where one has a particle and the other has a hole , then there will be exactly one discrepancy between the two systems for all time , and this discrepancy moves as a second - class particle .
consider an initial state in which every negative site is occupied by a particle , every positive site has a hole , and there is a single second - class particle at the origin .
let @xmath21 be the position of the second - class particle at time @xmath11 .
for the , ferrari and kipnis @xcite proved that @xmath22 converges in distribution as @xmath23 to a uniform random variable on the interval @xmath24 $ ] .
the extension of the method of @xcite to the asep is straightforward , but for completeness we give the result and its proof ( indeed it can be extended much further , to more general asymmetric exclusion processes and to a larger class of initial distributions ) .
for the particular case of the , almost sure convergence has been proved by mountford and guiol @xcite , ferrari and pimentel @xcite , and ferrari , martin and pimentel @xcite .
we then consider systems with two second - class particles .
we will be particularly interested in the event that one of the second - class particles attempts a jump onto the site of the other one .
various conventions are possible for what happens when such a jump is attempted ; the particles may annihilate , or coalesce , or one may have priority over the other , or no change at all may occur .
( see also section [ growthproofsec ] for discussion of another relevant possibility ) .
the difference between these conventions will not be important for our purposes , since we concentrate on the probability that such a jump attempt happens at all .
for convenience we will refer to such an event as the _ collision _ of the two second - class particles .
we start the process from a deterministic configuration , denoted @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{{{\circledast}}}{\hat{{\circle}}}$ ] , that has all negative sites occupied by ( first - class ) particles , the origin and site 1 occupied by second - class particles , and all sites to the right of site 1 empty ( in this notation the origin is underlined and @xmath26{0.2}{\circle}}}$ ] and @xmath27 denote the semi - infinite sequences @xmath28{0.2}{\circle}}}{{\textcolor[gray]{0.2}{\circle}}}{{\textcolor[gray]{0.2}{\circle}}}$ ] and @xmath29 respectively ) .
we show that the probability that the two second - class particles collide is @xmath30 .
if instead the particles start at distance two apart , from the configuration @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{{{\circle}}}{{{\circledast}}}{\hat{{\circle}}}$ ] , then the corresponding probability is @xmath31 .
these probabilities can also be interpreted as `` coupling probabilities '' for systems with only one type of particle .
if two one - type aseps are started from the initial conditions @xmath25{0.2}{\circle}}}}{{\underline { { \textcolor[gray]{0.2}{\circle}}}}}{{{\circle}}}{\hat{{\circle}}}$ ] and @xmath26{0.2}{\circle}}}{{\underline { { \circle}}}}{{{\textcolor[gray]{0.2}{\circle}}}}{\hat{{\circle}}}$ ] and evolve together under the basic coupling , then the probability that they eventually reach the same state is @xmath30 .
the second probability @xmath31 applies to the case where the initial conditions of the two systems are @xmath25{0.2}{\circle}}}}{{\underline { { \textcolor[gray]{0.2}{\circle}}}}}{{{\circle}}}{{{\circle}}}{\hat{{\circle}}}$ ] and @xmath25{0.2}{\circle}}}}{{\underline { { \circle}}}}{{{\circle}}}{{{\textcolor[gray]{0.2}{\circle}}}}{\hat{{\circle}}}$ ] .
in the particular case of the , where jumps only occur in one direction , various interesting alternative representations are possible .
consider a multi - type system in which the initial condition has a particle labelled @xmath32 at site @xmath32 , for all @xmath33 .
lower - numbered particles have priority over higher - numbered particles ; that is , the particles start out in reverse order of priority and every jump consists of a particle labelled @xmath32 exchanging places with a particle labelled @xmath34 to its right , for some @xmath35 . whenever the particle at site @xmath10 has priority over the particle at site @xmath36 , such an exchange between sites @xmath10 and @xmath36 occurs at rate 1 . from the coupling / collision probabilities described above
, one can deduce that the probability that at some point particle @xmath32 overtakes particle @xmath37 is @xmath38 , while the probability that particle @xmath32 overtakes both particle @xmath37 and particle @xmath39 is @xmath40 .
the with two second - class particles can also be mapped onto a two - type version of the `` corner growth model '' associated to directed last - passage percolation .
this model has also recently been studied by coupier and heinrich @xcite .
the event that the two particles stay separated for ever in the corresponds to the event that both clusters in the growth process grow unboundedly , without one surrounding the other . in @xcite it is proved that this event has positive probability . using the correspondence with the , we obtain that the probability is in fact 1/3 ; this gives perhaps the first example of a model where such a `` coexistence probability '' can be precisely calculated . in section [ s2 ]
we define notation and state our main results for the general @xmath41 , along with some discussion . in section [ tasepsec ]
we discuss the particular case of the @xmath42 and explain the representation as a multi - type particle system , and in section [ growthsec ] we describe the two - type competition growth model .
section [ s3 ] collects together some known results on hydrodynamics and couplings .
proofs of our results are given in sections [ ss4 ] , [ ss5 ] and [ ss6 ] .
let @xmath0 $ ] and write @xmath5 .
we define the two - type asep as follows . for @xmath43{0.2}{\circle}}}\}$ ] , we say that @xmath32 has priority over @xmath34 and write @xmath44 if @xmath45{0.2}{\circle } } } , j\in\{{{\circle}},{{\circledast}}\}$ ] or if @xmath46 .
let @xmath47{0.2}{\circle}}}\}^{\mathbb z}$ ] and consider the generator given on cylinder functions @xmath48 by : @xmath49 + q{\mathbf 1 } _ { \{\eta(x)\prec\eta(x-1)\ } } ] \,[f(\eta^{x , x-1})-f(\eta ) ] \big\ } , \label{a30gen}\ ] ] where @xmath50 if the initial configuration @xmath51 belongs to @xmath52{0.2}{\circle}}}\}^{\mathbb z}$ ] , we recover the asep with a single type of particle .
if @xmath3 the attempted jumps are only to the right , and we have the . the existence of this process can be proven using liggett s construction @xcite or using the harris graphical construction as in @xcite .
see also the comments in section [ s3 ] about the basic coupling .
our results concern systems in which all sites to the left of a certain point contain first - class particles , and all sites to the right of a certain point contain holes .
( this property is preserved by the dynamics ) .
as mentioned in the introduction , we use a compact notation for such states : @xmath26{0.2}{\circle}}}=\dots{{{\textcolor[gray]{0.2}{\circle}}}}{{{\textcolor[gray]{0.2}{\circle}}}}{{{\textcolor[gray]{0.2}{\circle}}}}$ ] denotes a semi - infinite string of first - class particles , @xmath53 denotes a semi - infinite string of holes , and an underline indicates the position of the origin .
for example , @xmath54{0.2}{\circle}}}}}{{\underline { { \circledast}}}}{{{\circledast}}}{\hat{{\circle}}}$ ] denotes the configuration defined by @xmath55{0.2}{\circle}}}}}{{\underline { { \circledast}}}}{{{\circledast}}}{\hat{{\circle}}})(x):= \left\ { \begin{array}{ll } { { \textcolor[gray]{0.2}{\circle}}},&\hbox{if } x<0;\\ { { \circledast}},&\hbox{if } x\in\{0,1\};\\ { { \circle}},&\hbox{if } x>1\ , .
\end{array } \right.\ ] ] ( in fact the absolute position of the origin is rarely important for the results , and sometimes we may omit the underline , but it is often convenient to fix it ) .
the following result extends to the asep a result proved for the by ferrari and kipnis @xcite : [ 2cptheorem ] consider the asep with @xmath56 $ ] , starting from the state @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{\hat{{\circle}}}$ ] .
let @xmath21 be the position of the second - class particle at time @xmath11 .
then @xmath57 in distribution as @xmath58 , where @xmath59 is uniformly distributed on the interval @xmath60 $ ] .
we note that in fact the same method of proof can be used to prove a much more general result . already in @xcite
, the initial condition can be taken to be the product measure with density @xmath61 on negative sites and density @xmath62 on positive sites , with @xmath63 .
in addition one can extend to more general asymmetric exclusion processes , in which the particles ( still interacting via exclusion ) carry out more general continuous - time random walks with some drift @xmath64 .
the convergence is then to a uniform distribution on the interval @xmath65 $ ] . in the particular case of the
, the convergence has been shown to hold almost surely @xcite , @xcite , @xcite .
we now turn to systems with a pair of second - class particles .
the following theorem again extends to the asep results proved for the in @xcite .
the proofs are analogous ; we include them for completeness .
[ t2 ] consider the asep with @xmath66 $ ] , starting from @xmath26{0.2}{\circle}}}{{\underline { { \circledast}}}}{{{\circledast}}}{\hat{{\circle}}}$ ] .
let @xmath67 be the first time that a jump is attempted by one of the second - class particles onto the site of the other ( which may be infinity ) .
for @xmath68 , let @xmath21 and @xmath69 be the positions of the two second - class particles at time @xmath11 , with @xmath70 .
* for any @xmath71 $ ] , @xmath72 * @xmath73=\frac{(p - q)^2}{3p}.\ ] ] in display ( 3.7 ) of @xcite , the limit was stated for the case @xmath3 , but it has the extra term `` @xmath74 '' which does not belong there .
the following is our main result : [ t1 ] consider an asep with @xmath66 $ ] , containing two second - class particles .
let @xmath67 be the first time that a jump is attempted by one of the second - class particles onto the site of the other ( which may be infinity ) . * if the initial condition is @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{{{\circledast}}}{\hat{{\circle}}}$ ] , then @xmath75 * if instead the initial condition is @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{{{\circle}}}{{{\circledast}}}{\hat{{\circle}}}$ ] , then @xmath76 in the ( @xmath3 ) these probabilities are @xmath38 in part ( a ) ( an upper bound of @xmath77 was obtained by @xcite in this case ) and @xmath40 in part ( b ) . the value in part ( b ) is obtained from that in part ( a ) using a short argument based on conditioning on the first jump and some symmetries of the process under this initial configuration .
these probabilities also represents a coupling probability in the context of the asep with only one class of particle .
consider two processes started from the states @xmath25{0.2}{\circle}}}}{{\underline { { \textcolor[gray]{0.2}{\circle}}}}}{{{\circle}}}{\hat{{\circle}}}$ ] and @xmath25{0.2}{\circle}}}}{{\underline { { \circle}}}}{{{\textcolor[gray]{0.2}{\circle}}}}{\hat{{\circle}}}$ ] , and coupled using the _ basic coupling _ , under which jumps are attempted at the same sites at the same times in both processes ( see section [ s3 ] for details ) . at time 0 there are discrepancies between the two processes at sites 0 and 1 ; these discrepancies move around , each behaving as a second - class particles , and at time @xmath67 , the first time that one jumps onto the site of the other , the two processes reach the same state and stay coupled thereafter .
hence the probability that they ever reach the same state is @xmath30 , or @xmath38 for the case of the . if instead the initial conditions are @xmath25{0.2}{\circle}}}}{{\underline { { \textcolor[gray]{0.2}{\circle}}}}}{{{\circle}}}{{{\circle}}}{\hat{{\circle}}}$ ] and @xmath25{0.2}{\circle}}}}{{\underline { { \circle}}}}{{{\circle}}}{{{\textcolor[gray]{0.2}{\circle}}}}{\hat{{\circle}}}$ ] then the two discrepancies start at sites 0 and 2 , and the coupling probability is @xmath31 , giving @xmath40 for the case of the .
our proof of theorem [ t1 ] involves estimating the expected number of particles which are to the right of the second - class particle at time @xmath11 in a process started from the initial condition @xmath26{0.2}{\circle}}}{{\underline { { \circledast}}}}{\hat{{\circle}}}$ ] . as a corollary of these methods
, we can deduce the following result , which relates to the convergence of the environment around a second - class particle : [ 2cpcorollary ] let @xmath78 be the asep with @xmath66 $ ] started from the state @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{\hat{{\circle}}}$ ] .
let @xmath21 be the position of the second - class particle at time @xmath11
. then @xmath79
when @xmath3 , the system is `` totally asymmetric '' and various interesting connections can be drawn .
we consider a system with infinitely many classes of particle .
initially , site @xmath32 contains a particle labelled @xmath32 , for all @xmath33 .
a particle with label @xmath32 has priority over a particle with label @xmath34 if @xmath35 . if the particle at site @xmath10 has priority over the particle at site @xmath36 , the two exchange places with rate 1 .
since the process is totally asymmetric , once two particles have exchanged places , they can never swap back again , so the higher - priority particle stays ahead of the lower - priority particle thereafter .
let @xmath80 be the position at time @xmath11 of the particle with label @xmath32 ( so @xmath81 for all @xmath32 ) .
consider @xmath82 for example .
particle @xmath83 has lower priority than all the particles starting to its left , and higher priority than all the particles starting to its right .
hence @xmath84 behaves just as the position @xmath21 of the second - class particle starting from the configuration @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{\hat{{\circle}}}$ ] . the same is true for any other @xmath32 ( shifted by a constant distance @xmath32 to the right ) . in particular , from the strong law of large numbers mentioned after theeorem [ 2cptheorem ] , we know that the limits @xmath85 exist almost surely for all @xmath32 , and all have the uniform distribution on @xmath24 $ ] .
now consider the particles labelled @xmath83 and @xmath86 .
again they see stronger particles to their left and weaker particles to their right , so their behaviour corresponds to that of the pair of second - class particles starting from the configuration @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{{{\circledast}}}{\hat{{\circle}}}$ ] .
in particular , we know from theorem [ t1](a ) that the probability of a jump attempt between these two particles is @xmath38 .
since the process is totally asymmetric , the only way for this to happen is that particle @xmath83 attempts to jump right onto particle @xmath86 at some point . in this case the two particles exchange places .
the same applies for any pair of particles @xmath32 and @xmath37 .
so we obtain the following result giving the probability that a particle eventually overtakes its original right - hand neighbour : for any @xmath32 , @xmath87 . from theorem [ t2](a ) , we can obtain further information about the joint distibution of the `` asymptotic speeds '' @xmath88 and @xmath89 of neighbouring particles @xmath32 and @xmath37 , as defined at ( [ ddef ] ) .
we get that @xmath90 for all @xmath91 $ ]
. this expression would be consistent with @xmath88 and @xmath89 being independent ; however , since we know that particle @xmath32 overtakes particle @xmath37 with probability @xmath38 , such independence does not in fact hold .
from the second part of theorem [ t1 ] we can make the following deduction about the probability that a particle overtakes both of the two particles starting just to its right : [ distancetwocorr ] for any @xmath32 , @xmath92 . to deduce corollary [ distancetwocorr ] from theorem [ t2](b )
, one can compare the multi - type of this section with the two - type of theorem [ t2](b ) , started from the initial condition @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{{{\circle}}}{{{\circledast}}}{\hat{{\circle}}}$ ] , as shown in figure [ initialfig ] .
aligned initial conditions for the multi - type ( above ) and the two - type ( below ) ] let the two processes evolve together under the basic coupling ; potential jumps occur at the same site at the same times in the two processes . folllowing the joint evolution of the two processes
, one can observe that the site occupied by particle 0 ( in the multi - type process ) always corresponds to the site of a second - class particle ( in the two - type process ) .
in addition , the other second - class particle is at the site either of particle 1 or of particle 2 , whichever is further to the right at the time .
so the first time that one second - class particle tries to jump onto the site of the other is precisely the time that particle 0 jumps onto the site of whichever is further right of particle 1 and particle 2 , i.e. the first moment that particle 0 has overtaken both particle 1 and particle 2 .
a similar argument can be used to show the following ; the probability that particle @xmath32 overtakes all of particles @xmath93 in the multi - type process is the same as the probability that the two second - class particles collide in a two - type process starting from the state @xmath25{0.2}{\circle}}}}{{\underline { { \circledast}}}}{{{\circle}}\dots{{\circle}}}{{{\circledast}}}{\hat{{\circle}}}$ ] , where there is a string of @xmath94 holes between the two second - class particles .
we have seen that this probability is @xmath38 for @xmath95 and @xmath40 for @xmath96 .
simulations suggest that the probability is in fact @xmath97 for all @xmath98 .
in this section we describe the two - type growth model which corresponds to the with two second - class particles .
the connection between the and the corner growth model dates back to rost @xcite . in the simplest case , the growth takes place in the positive integer quadrant @xmath99 ( where we write @xmath100 ) . at time 0
all the sites of this quadrant are empty .
a site @xmath101 may become occupied if both sites @xmath102 and @xmath103 are occupied ; once this condition holds , the time until site @xmath101 becomes occupied is an exponential random variable with rate 1 ( independently for each site ) . as a boundary condition one assumes that the sites @xmath104 and @xmath105 are always occupied , for each @xmath106 .
we will consider an extension of this model , where each occupied site has one of two colours , red and blue .
a site takes its colour at the moment it is occupied and never changes thereafter . in the initial condition ,
the sites @xmath107 , @xmath108 and @xmath109 , and all other sites are unoccupied .
( the colour assigned to the three already - occupied points is not important ) . when a site becomes occupied , its colour is chosen as follows : *
any site @xmath110 or @xmath111 , for @xmath112 , becomes blue .
* the site @xmath113 becomes red .
* a site @xmath101 , @xmath112 , @xmath114 , @xmath115 , takes the colour of whichever of its parents became occupied _
more recently_. ( since the delay times for occupation of each site are independent continuous random variables , with probability one all times of occupation are different and so there are no ties to break in the third case above )
. there are three possibilities for the first event of the growth process : @xmath116 becomes occupied and blue , or @xmath117 becomes occupied and blue , or @xmath113 becomes occupied and red .
each of these occurs with probability @xmath118 . to the left an example showing a growth process after its first three jumps .
we can see that @xmath119 , since site @xmath120 took its colour from @xmath117 rather than @xmath113 .
also clearly @xmath121 ( since @xmath116 is not yet occupied ) . as a result site @xmath122 will take its colour from @xmath116 and so will also be blue ; the red cluster will be surrounded and unable to grow further .
to the right , an example showing a growth process after its first nine jumps . here
one can deduce , for example , that the first site to be occupied was @xmath116 .
given that @xmath116 is the first to be occupied , the conditional probability that the red cluster survives is 1/2 ( see the remark at the end of this section ) . ] to the left an example showing a growth process after its first three jumps .
we can see that @xmath119 , since site @xmath120 took its colour from @xmath117 rather than @xmath113 .
also clearly @xmath121 ( since @xmath116 is not yet occupied ) . as a result site @xmath122 will take its colour from @xmath116 and so will also be blue ; the red cluster will be surrounded and unable to grow further .
to the right , an example showing a growth process after its first nine jumps . here
one can deduce , for example , that the first site to be occupied was @xmath116 .
given that @xmath116 is the first to be occupied , the conditional probability that the red cluster survives is 1/2 ( see the remark at the end of this section ) . ]
note that there will certainly be infinitely many points which take the colour blue ( for example , all the points on the axes @xmath111 and @xmath110 ) . however , there may or may not be infinitely many red points .
for example , suppose that the first event in the process is that @xmath113 is occupied ( red ) . in this case , both @xmath116 and @xmath117 are occupied later than @xmath113 . as a result
, @xmath122 will take its colour from @xmath116 ( its more recently occupied parent ) and similarly @xmath120 will take its colour from @xmath117 .
so both @xmath117 and @xmath116 become blue , and the red cluster has been `` surrounded '' ; there will be no further red sites . if @xmath113 is not the first site to become occupied , the red cluster may nevertheless become surrounded later ; alternatively , it may grow unboundedly see figures [ examplefig ] and [ meetingfig ] for some examples .
[ growththm ] the probability that the red cluster grows unboundedly is @xmath118 . two simulations of the growth process up to time 300 .
the left simulation is a rare example where the red cluster grows for a long time ( here up until time 271.8 ) but then becomes surrounded . on the right
, the red cluster looks likely to survive for ever .
] two simulations of the growth process up to time 300 .
the left simulation is a rare example where the red cluster grows for a long time ( here up until time 271.8 ) but then becomes surrounded . on the right , the red cluster looks likely to survive for ever . ] coupier and heinrich @xcite recently showed that the probability in theorem [ growththm ] was strictly positive , without using any representation , by proving that , in the associated directed last - passage percolation model , there exists with positive probability an infinite geodesic starting from the point @xmath107 and passing through the point @xmath113 .
theorem [ growththm ] is again deduced from theorem [ t1](a ) .
we give the details of the proof in section [ growthproofsec ] .
the main element in the proof of theorem [ growththm ] is the representation introduced by ferrari and pimentel @xcite in which a site containing a second - class particle particle @xmath20 is replaced by a pair of sites in the state @xmath123{0.2}{\circle}}}]$ ] .
ferrari and pimentel treated systems with just one second - class particle , so that the two - type system with initial condition @xmath25{0.2}{\circle}}}}{{\circledast}}{\hat{{\circle}}}$ ] was replaced by a one - type system with initial condition @xmath25{0.2}{\circle}}}}{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}{\hat{{\circle}}}$ ] . in rost
s correspondence , this equates to a growth process starting from the state where site @xmath107 alone is occupied .
analogously , the initial condition @xmath25{0.2}{\circle}}}}{{\circledast}}{{\circledast}}{\hat{{\circle}}}$ ] from theorem [ t1](a ) corresponds to the initial condition @xmath25{0.2}{\circle}}}}{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}{\hat{{\circle}}}$ ] , or to a growth model starting with sites @xmath107 , @xmath109 and @xmath108 occupied .
this hole - particle representation of two second - classs particles survives only until the moment when one second - class particle attempts to jump onto another ; but this is enough for our purposes , since we will be concerned only with the probability of such a collision event occurring . in @xcite ,
the path of one second - class particle corresponded to the interface between two clusters . here
we have two interfaces , between the red cluster in the middle and the blue clusters on either side of it .
the event that the two second - class particles meet will correspond to the event the two interfaces meet .
if this happens , the red cluster between the two interfaces is surrounded and is unable to grow any further .
this will happen with probability 2/3 , as in theorem [ t1](a ) .
see section [ growthproofsec ] for full details .
[ [ remark ] ] _ remark _
+ + + + + + + + note that the first site to be occupied in the growth process is equally likely to be @xmath116 , @xmath113 or @xmath117 , with probability @xmath118 each . as observed above , if site @xmath113 is the first to be occupied , the red cluster can not survive .
we know that the overall probability that the red cluster survives is @xmath118 . hence , given that the first site to be occupied is _ not _
@xmath113 , the conditional probability that the red cluster survives is @xmath40 . this statement corresponds to part ( b ) of theorem [ t1 ] , in the case @xmath3 , ( while theorem [ growththm ] corresponds to part ( a ) of theorem [ t1 ] ) . by symmetry
, it does not matter whether @xmath117 or @xmath116 is occupied first . to emphasise the connection , note that if we start from the configuration where @xmath116 has been occupied in addition to @xmath107 , @xmath108 and @xmath109 , this corresponds to starting the from the configuration @xmath124{0.2}{\circle}}}}}{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}{{\circle}}{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle}}}}$ ] , which , under the hole - particle representation of the second - class particle , corresponds in turn to the initial condition @xmath124{0.2}{\circle}}}}}{{\circledast}}{{\circle}}{{\circledast}}{{\hat{{\circle}}}}$ ] of theorem [ t1](b ) .
again , see section [ growthproofsec ] for details of the correspondence between states of the and states of the growth model .
in this section we state some known results on hydrodynamics and on coupling that will be used in the proofs .
[ [ hydrodynamics ] ] * hydrodynamics * + + + + + + + + + + + + + + + define the product measures @xmath125 under which every site is occupied by a particle with probability @xmath61 and a hole with probability @xmath126 , independently for different sites .
any translation - invariant stationary distribution of the one - type asep is a mixture of these product measures @xcite .
suppose the initial distribution puts probability 1 on configurations with asymptotic density 1 to the left and 0 to the right .
( in particular this is the case whenever all sites to the left of some point contain particles and all sites to the right of some point contain holes , i.e.the initial configuration has the form @xmath26{0.2}{\circle}}}\ , a\ , \hat{{\circle}}$ ] for some finite string @xmath127 ) .
then the process converges to local equilibrium : @xmath128 for any cylinder function @xmath129 , where @xmath130 is the translation by @xmath10 , @xmath131 is the semigroup of the asep with generator given by ( [ a30gen ] ) , @xmath132 is the integer part and for @xmath133 , @xmath134 , @xmath135 is the entropy solution of the inviscid burgers equation @xmath136 which is explicitly given by @xmath137 [ [ basic - coupling ] ] * basic coupling * + + + + + + + + + + + + + + + + the harris construction of the process makes use of a family of independent poisson processes @xmath138 , where @xmath14 has rate @xmath15 and @xmath16 has rate @xmath17 .
these live in a probability space @xmath139 . for the process with first - class and second - class particles , we write @xmath47{0.2}{\circle}}}\}^{\mathbb z}$ ] as above , and we define a function @xmath140 inductively in time as follows ( informally ) : fix an initial configuration @xmath51 and call @xmath141 .
then @xmath142 that is , the particles attempt to jump from @xmath10 to @xmath19 at the epochs of the poisson process @xmath18 and do so if their classes admit .
of course this definition is incomplete , as the epochs of the poisson process are not well ordered , but a simple argument shows that for all @xmath11 and almost all realizations @xmath143 , @xmath144 is partitioned into finite intervals that do not interact with each other in the interval @xmath145 $ ] ( see @xcite and @xcite for details ) .
the process so defined is markov and has generator @xmath146 given in .
the basic coupling between @xmath147 versions of the process with initial configurations @xmath148 is the process which uses the same @xmath143 for all marginals : @xmath149 that is , in the basic coupling the particles attemp to jump from @xmath10 to @xmath19 at the same epochs in each marginal , always respecting the classes ( at the marginal ) .
* proof of theorem [ 2cptheorem]*. as in the proof of theorem 1 of @xcite , the main ingredient is a coupling argument and the convergence to local equilibrium ( [ localequilibrium ] ) .
one wants to show that for @xmath150 $ ] , @xmath151 we will consider asep processes @xmath152 and @xmath153 started from the initial states @xmath154{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{\hat{{\circle}}}$ ] . and
@xmath155{0.2}{\circle}}}}{\underline { { \circle}}}{\hat{{\circle}}}$ ] .
let @xmath156 be the number of particles of @xmath152 which are to the right of @xmath157 , and similarly for @xmath158
. we will compute @xmath159 in two different ways . the initial states @xmath160 and @xmath161 differ only at the origin , where @xmath160 has an extra particle .
if we couple the processes using the basic coupling , then there is exactly one discrepancy at all times , and it behaves as the single second - class particle of a process started in the state @xmath25{0.2}{\circle}}}}{\underline { { \circledast}}}{\hat{{\circle}}}$ ] . hence @xmath162 on the other hand , by shifting the origin by 1 , we can see that the number of particles to the right of @xmath157 in @xmath153 has the same distribution as the number of particles to the right of @xmath163 in @xmath152 .
hence @xmath164 which , by the convergence to local equilibrium in ( [ localequilibrium ] ) , is @xmath165 , so that ( [ way1 ] ) and ( [ way2 ] ) together give ( [ form ] ) as desired .
in this section we give the proof of theorem [ t2 ] . for @xmath91 $ ]
let @xmath167 be the number of particles at the right of @xmath157 at time @xmath11 ; for convenience , we consider that if there is a particle at @xmath168 a fraction @xmath169 of it is to the right of @xmath157 .
[ [ proof - of - theorem - t2 ] ] * proof of theorem [ t2 ] * + + + + + + + + + + + + + + + + + + + + + + + consider an asep @xmath170 started from the initial condition @xmath171{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{\hat{{\circle } } } = { \hat{{\textcolor[gray]{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circle}}{\hat{{\circle}}}$ ] . from @xmath172 ,
the first jump of the process is to the state @xmath173{0.2}{\circle}}}{\hat{{\circle}}}$ ] and this jump occurs at rate @xmath15 . hence using the definition of @xmath174 in ( [ c1 ] ) and the kolmogorov backward equation , we have that @xmath175 is continuous and is differentiable everywhere except possibly at points @xmath11 where @xmath176 , with derivative @xmath177{0.2}{\circle}}}}{\underline { { \circle}}}{{\textcolor[gray]{0.2}{\circle}}}{\hat{{\circle } } }
\big ) -e(j_{t}^{r}\big ( { \hat{{\textcolor[gray]{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circle}}{\hat{{\circle } } } \big)\big)\big ] -re(\eta_t(\lfloor rt\rfloor)).\ ] ] we couple two versions of the asep starting from @xmath25{0.2}{\circle}}}}{\underline { { \circle}}}{{\textcolor[gray]{0.2}{\circle}}}{\hat{{\circle}}}$ ] and from @xmath25{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circle}}{\hat{{\circle}}}$ ] , aligned as follows : @xmath178{0.2}{\circle}}}}{\underline { { \circle}}}{{\textcolor[gray]{0.2}{\circle}}}{\hat{{\circle } } }
\\ { \hat{{\textcolor[gray]{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circle}}{\hat{{\circle } } } \end{matrix}\ ] ] at time zero there are two discrepancies , located at sites 0 and 1 . if a jump is attempted from the site of one the discrepancies into the other , the discrepancies disappear and the two configurations coincide from that instant on . until that time , denoted by @xmath67 , the discrepancies behave as the positions of two second - class particles , @xmath21 and @xmath69 .
the difference of the number of particles to the right of @xmath157 for the two coupled processes can be expressed in terms of the positions of @xmath21 and @xmath69 , as follows : @xmath179{0.2}{\circle}}}}{\underline { { \circle}}}{{\textcolor[gray]{0.2}{\circle}}}{\hat{{\circle } } } ) -j^r_{t } ( { \hat{{\textcolor[gray]{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circle}}{\hat{{\circle } } } \big ) = \begin{cases } 1 , & \hbox{if $ \tau > t$ and $ x(t)<\lfloor rt\rfloor < y(t)$ ; } \\
rt-\lfloor rt\rfloor , & \hbox{if $ \tau > t$ and $ x(t)=\lfloor rt\rfloor < y(t)$ ; } \\
1-rt+\lfloor rt\rfloor , & \hbox{if $ \tau > t$ and $ x(t)<\lfloor rt\rfloor = y(t)$ ; } \\ 0 , & \hbox{otherwise . }
\end{cases}\ ] ] inserting in ( [ current2 ] ) , @xmath180 - re(\eta_t(\lfloor rt\rfloor))\end{gathered}\ ] ] now we compute ( [ current2 ] ) in a different way by associating to @xmath181 the following martingale : @xmath182 where , for a fixed configuration @xmath51 , @xmath183 denotes the instantaneous current through the bond @xmath184 $ ] , namely : @xmath185 this expression together with the convergence to local equilibrium ( [ localequilibrium ] ) imply @xmath186 the same convergence to local equilibrium also gives that @xmath187 note that @xmath188 ( for example , @xmath21 represents the discrepancy between two coupled processes starting from the states @xmath25{0.2}{\circle}}}}{\underline { { \circle}}}{\hat{{\circle}}}$ ] and @xmath25{0.2}{\circle}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{\hat{{\circle}}}$ ] ; the probability of finding this discrepancy precisely at the site @xmath168 at time @xmath11 goes to 0 as @xmath58 since , again by the convergence to local equilibrium in ( [ localequilibrium ] ) , the probability of site @xmath168 being occupied at time @xmath11 converges to the same limit @xmath189 for both processes . ) putting equations ( [ a37])-([fracsimp ] ) together gives @xmath190.\ ] ] the form of the function @xmath191 from then gives ( [ separationprob ] ) .
the fact that @xmath70 whenever @xmath192 , and the relation @xmath193= \sum_y
p\big(x(t)\leq y < y(t ) , \tau > t\big),\ ] ] then give @xmath194= \frac{1}{4p(p - q)}\int_{-(p - q)}^{p - q}\left[(p - q)^2-r^2\right]dr,\ ] ] giving as required.@xmath166
[ [ proof - of - theorem - t1a ] ] * proof of theorem [ t1](a ) * + + + + + + + + + + + + + + + + + + + + + + + + + + denote by @xmath124{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle}}}}$ ] the configuration that has a second - class particle at the origin , while the negative sites are occupied by first - class particles and the rest is empty .
denote by @xmath124{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle}}}}$ ] the configuration that has a second - class particle at site 1 , while the non - positive sites are occupied by first - class particles and the rest is empty .
we couple both processes and prove later that the positions of the discrepancies initially at sites @xmath83 and @xmath86 behave as @xmath21 and @xmath69 , the two second - class particles of the theorem , with @xmath195 and @xmath196 . for a configuration @xmath51 with a unique second - class particle and a finite number of first - class particles to the right of it , denote by @xmath197 the position of the second - class particle at time @xmath11 and by @xmath198 the number of first - class particles to the right of @xmath197 at time @xmath11 : @xmath199 where @xmath152 is the configuration at time @xmath11 for the process with initial configuration @xmath200 .
define the configuration @xmath201{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\hat{{\circle}}}}$ ] . for @xmath202 ,
the kolmogorov backwards equation gives @xmath203{0.2}{\circle}}}}}{{\circledast}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\hat{{\circle}}}})\big ) + p e\big(j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle}}}})\big ) -2pe\big(j_{t}^{(2)}({\tilde\eta})\big)\big).\ ] ] in @xmath124{0.2}{\circle}}}}}{{\circledast}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\hat{{\circle}}}}$ ] is attained when the rightmost first - class particle at site @xmath204 jumps to @xmath83 , the site occupied by the second - class particle , and @xmath124{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle}}}}$ ] arises when the second - class particle jumps from @xmath83 to @xmath86 , site occupied by the leftmost hole .
the law of @xmath205 does not depend on the actual location of @xmath206 but on the relative positions of the other particles with respect to @xmath206 .
so , we are free to change the origin .
in particular @xmath207{0.2}{\circle}}}}}{{\circledast}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\hat{{\circle}}}})\big)= e\big(j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle}}}})\big)$ ] and @xmath208{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\hat{{\circle}}}})\big)= e\big(j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle}}}})\big)+e\big(j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle}}}})\big)$ ] . using this , for any ( four - marginals ) coupling , we can rewrite as @xmath209{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle } } } } ) -j_{t}^{(2 ) } ( { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle } } } } ) + j_{t}^{(2 ) } ( { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle } } } } ) -j_{t}^{(2 ) } ( { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle } } } } ) \big],\ ] ] so that we are dealing with four initial configurations that are aligned as follows : @xmath210{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle}}}}\\ { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle}}}}\\ { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle}}}}\\ { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle}}}}\end{matrix}\ ] ] we perform the basic coupling , where recall the jumps are attempted at the same time at the same sites for the four marginals .
the first and third lines contribute with a plus sign to while the others do so with a minus sign .
we start with two discrepancies at the sites 0 and 1 .
as before , under the basic coupling each of these discrepancies moves as a second - class particle , up until the moment when a jump is attempted from the site of one onto the site of the other .
let @xmath67 denote this time , which could be infinity . the states of the four marginals at various times look as follows : @xmath211{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circledast}}b{{\textcolor[gray]{0.2}{\circle}}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circledast}}{{\textcolor[gray]{0.2}{\circle}}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circledast}}{{\textcolor[gray]{0.2}{\circle}}}c{{\hat{{\circle}}}}&&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\textcolor[gray]{0.2}{\circle}}}{{\circledast}}c{{\hat{{\circle}}}}\\ { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circledast}}b{{\circle}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circledast}}{{\circle}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circle}}{{\circledast}}c{{\hat{{\circle}}}}&\textbf{or}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a { { \circledast}}{{\circle}}c{{\hat{{\circle}}}}\\ { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circle}}b{{\circledast}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circle}}{{\circledast}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circle}}{{\circledast}}c{{\hat{{\circle}}}}&&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circledast}}{{\circle}}c{{\hat{{\circle}}}}\\ { { \hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\textcolor[gray]{0.2}{\circle}}}b{{\circledast}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\textcolor[gray]{0.2}{\circle}}}{{\circledast}}c{{\hat{{\circle}}}}&{\qquad}&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\circledast}}{{\textcolor[gray]{0.2}{\circle}}}c{{\hat{{\circle}}}}&&{{\hat{{\textcolor[gray]{0.2}{\circle}}}}}a{{\textcolor[gray]{0.2}{\circle}}}{{\circledast}}c{{\hat{{\circle}}}}\end{matrix}\ ] ] here @xmath127 , @xmath212 and @xmath213 are some finite sequences of holes and first - class particles . ( they are the same within any given column , though their values may of course differ between columns ) .
the two possibilities corresponding to time @xmath67 reflect the fact that the jump at time @xmath67 may be from the discrepancy on the left onto the discrepancy on the right , or vice versa .
if the jump is from left to right , one gets the left column ; if the jump is from right to left , one gets the right column .
note that before time @xmath67 , the first row has one more @xmath214{0.2}{\circle}}}$ ] to the right of the @xmath20 than the second row does . meanwhile the third and the fourth rows have the same number of particles to the right of the @xmath20 .
hence for any @xmath68 , @xmath215{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle } } } } ) -j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle } } } } ) + j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle } } } } ) -j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle } } } } ) = 1.\ ] ] at time @xmath67 , the first and the fourth rows become equal , and ( since they are coupled ) will still be equal at all later times .
the same is true of the second and third rows .
hence for any @xmath216 , @xmath215{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle } } } } ) -j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle } } } } ) + j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle } } } } ) -j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle } } } } ) = 0.\ ] ] thus we have , for any @xmath11 , @xmath215{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\textcolor[gray]{0.2}{\circle}}}{{\hat{{\circle } } } } ) -j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\hat{{\circle } } } } ) + j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \circle}}}{{\circledast}}{{\hat{{\circle } } } } ) -j_{t}^{(2)}({{\hat{{\textcolor[gray]{0.2}{\circle}}}}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle } } } } ) = i(t<\tau).\ ] ] substituting into , we get @xmath217 since @xmath218 are decreasing events in @xmath219 , the following limits exist : @xmath220 by theorem [ 2cptheorem ] , @xmath221 converges in distribution as @xmath23 to a uniform random variable with support on @xmath222 $ ] denoted again by @xmath223 . together with the convergence to equilibrium in ( [ a38 ] ) , this allows us to conclude that @xmath224 \,\frac{1}{p - q}\big(\frac{(p - q)-\mathcal{u}}{2}\big)^2\ ] ] in distribution .
in addition , @xmath225 is non - negative and dominated by a poisson process of rate 1 . hence @xmath226
= \frac{p - q}3.\ ] ] taking the limit as @xmath58 in ( [ a32 ] ) gives @xmath227 as required .
@xmath166 *
proof of corollary [ 2cpcorollary]*. using and the forward equations , we have @xmath228 since @xmath229 increases by one whenever the second - class particle swaps with a particle on its left , and decreases by one whenever the second - class particle swaps with a particle on its right .
on the other hand , by the symmetry of the system under exchange of hole and particle and of left and right , @xmath230 has the same law as @xmath231 .
hence , @xmath232 as @xmath58 , the left - hand side of converges to @xmath233 which is @xmath234 by theorem [ t1 ] .
putting and together then gives and as desired .
* proof of theorem [ t1](b)*. consider the process started from the state @xmath124{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circledast}}{{\hat{{\circle}}}}$ ] . from theorem [ t1](a ) , we know that the probability of collision of the two second - class particles is @xmath2 .
we condition on the first jump .
there are three possibilities . at rate 1 ,
the two second - class particles collide ( either because of a jump right from site 0 or because of a jump left from site 1 ) . at rate @xmath15
there is a jump right by the first - class particle at site @xmath204 , displacing the second - class particle at site 0 and leading to the state @xmath124{0.2}{\circle}}}}}{{\circledast}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle}}}}$ ] . also at rate @xmath15
there is a jump right by the second - class particle at site 1 , leading to the state @xmath124{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\circledast}}{{\hat{{\circle}}}}$ ] .
so with probability @xmath235 the first jump leads to immediate collision , and with probability @xmath236 the first jump is to one of the states @xmath124{0.2}{\circle}}}}}{{\circledast}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle}}}}$ ] or @xmath124{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\circledast}}{{\hat{{\circle}}}}$ ] .
notice that the dynamics of the system are invariant under the operation which reverses left and right and also exchanges the roles of first - class particles and holes . from this one
can see that the probability of collision of the two second - class particles is the same from the two states @xmath124{0.2}{\circle}}}}}{{\circledast}}{\underline { { \textcolor[gray]{0.2}{\circle}}}}{{\circledast}}{{\hat{{\circle}}}}$ ] and @xmath124{0.2}{\circle}}}}}{\underline { { \circledast}}}{{\circle}}{{\circledast}}{{\hat{{\circle}}}}$ ] .
let this probability be @xmath237 .
conditioning on the first jump , we obtain @xmath238 solving this we obtain @xmath239 as required .
in this section we explain the correspondence between the growth model with three types and the with two second - class particles , which leads to theorem [ growththm ] .
this is a natural extension of the correspondence between a two - type growth model and the with _
one _ second - class particle , developed by ferrari and pimentel in @xcite .
we start by giving an account of that correspondence ; for a more detailed description , the reader may like to refer to @xcite itself .
first we make explicit the correspondence between the ( one - type ) corner growth model and the one - type , as described by rost in @xcite . in the corner growth model ,
all sites of @xmath99 start off empty at time 0 ; each site @xmath240 waits until its two neighbours below and to the left , @xmath241 and @xmath242 , have been occupied ; it then waits for a random further amount of time which is exponential with mean 1 , and then becomes occupied itself .
all sites @xmath104 and @xmath105 are already occupied at time 0 .
write @xmath243 for the time that site @xmath240 becomes occupied in the corner growth model .
we can write a system of recurrences @xmath244 , for @xmath245 , where we set the boundary conditions @xmath246 whenever @xmath82 or @xmath247 . here
the @xmath248 are i.i.d .
exponential random variables with mean 1 . to represent the in this way , suppose that the particles are labelled p1 , p2 , p3 , @xmath249 from right to left , and the holes are labelled h1 , h2 , h3 from left to right , as shown in figure [ sequencefig ] .
initial condition for the , with particles and holes labelled sequentially .
the first event , at time @xmath250 , is the exchange of p1 and h1 . ]
( these labels are not to be confused with priorities or with the types of particle in section [ growthsec ] ; the sequence of particles and the sequence of holes will both stay in their initial orderings throughout , although particles and holes will exchange places ) . for @xmath251 ,
let @xmath252 be the time at which particle p@xmath32 overtakes hole h@xmath34 .
then exactly the same recurrences hold as in the previous paragraph .
once particle p@xmath32 has overtaken hole h@xmath253 , and hole h@xmath34 has been overtaken by particle p@xmath254 , particle p@xmath32 finds itself immediately on the left of hole h@xmath34 ; now an exponential amount of time with rate 1 passes before p@xmath32 overtakes h@xmath34 .
this gives a one - to - one correspondence between a set of states of the ( in which all sites far enough left contain particles and all sites far enough right contain holes ) and a set of states of the growth process ( with finitely many sites of @xmath99 occupied and satisfying the condition that if @xmath240 is occupied then so are @xmath241 and @xmath242 ) .
@xmath255 means that particle p@xmath32 is to the right of hole h@xmath34 at time @xmath11 , or correspondingly that site @xmath256 is already occupied at time @xmath11 .
ferrari and pimentel @xcite extended this correspondence to represent a with a single second - class particle in terms of two - type growth model .
the first element in this representation was to replace the site containing the second - class particle , @xmath20 , by a _
pair _ of sites , containing a hole and a particle in that order , @xmath257{0.2}{\circle}}}}]$ ] .
this pair is called the @xmath258pair , consisting of the @xmath258holeand the @xmath258particle .
when it is immediately to the left of a hole , the @xmath258particle jumps onto the hole to its right at rate 1 ; when this happens , we consider that the @xmath258pair itself moves one unit to the right , and @xmath259{0.2}{\circle}}}]{{\circle}}$ ] becomes @xmath260{0.2}{\circle}}}]$ ] .
similarly when a particle on the left of the @xmath258pairjumps to the right , the @xmath258pairitself moves left , so that @xmath214{0.2}{\circle}}}[{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}]$ ] becomes @xmath259{0.2}{\circle}}}]{{\textcolor[gray]{0.2}{\circle}}}$ ] .
hence the @xmath258pair , considered as a single unit , behaves precisely as a second - class particle .
hence we may represent the initial condition @xmath26{0.2}{\circle}}}{{\circledast}}\hat{{\circle}}$ ] , of a system with a single second - class particle , by the initial condition @xmath26{0.2}{\circle}}}[{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}]\hat{{\circle}}$ ] of a one - type system .
again we label holes and particles sequentially , and we let @xmath252 be the first time that particle @xmath32 is to the right of hole @xmath34 .
the boundary conditions are as before , except that now @xmath261 ( since we start with particle p1 already to the right of hole h1 ) . initial condition corresponding to the system with one second - class particle , and one of the two possible first jumps .
the jump shown occurs at time @xmath262 , if @xmath263 . ]
we keep track of the labels of the @xmath258holeand @xmath258particleat time @xmath11 .
let these labels be @xmath264 and @xmath265 .
we have @xmath266 , since initially the @xmath258paircovers hole h1 and particle p1 .
thereafter , whenever the @xmath258pairjumps right ( as in figure [ jumpexamplefig ] for example ) , @xmath264 increases by one while @xmath265 remains the same ; when the @xmath258pairjumps left , @xmath264 remains the same and @xmath265 increases by one .
in particular , the position @xmath21 of the second - class particle at time @xmath11 corresponds to @xmath267 ( the number of its jumps to the right minus the number of its jumps to the left ) .
see figure [ jumpexamplefig ] for an example . the evolution of the process @xmath268 can be described as follows . at time @xmath11 , the hole with label @xmath264 is immediately to the left of the particle with label @xmath265 .
hence @xmath269 , but @xmath270 and @xmath271 .
the next jump in the process @xmath272 occurs either when particle with label @xmath265 jumps right , or when the hole with label @xmath264 moves left , whichever happens sooner .
let @xmath273 be the @xmath147th term in the jump chain of the process @xmath272 .
then @xmath274 and we have the following recurrence for the sequence @xmath273 : @xmath275 so in terms of the growth model , @xmath273 behaves as follows .
we start from the initial condition in which only site @xmath107 is occupied , and @xmath274 . then at each step ,
@xmath273 moves from its current position to whichever of its neighbours above and to the right is the first to be occupied . we now describe how the sequence @xmath273 corresponds to a _ competition interface _ in the same growth model when occupied sites take one of two colours .
the rules are : * any site @xmath111 , @xmath112 takes the colour blue .
* any site @xmath110 , @xmath276 takes the colour red . * for @xmath277 , the site @xmath101 takes the same colour as whichever of its neighbours @xmath102 and @xmath103 became occupied most recently .
the colour of @xmath107 is not important .
let @xmath278 be the set of sites that take the colour blue ( when they are eventually occupied ) and let @xmath279 be the set of set of sites that take the colour red .
so @xmath280 . using a simple induction argument based on the growth rules above and on ( [ phirec ] )
, we get the following characterisation which justifies the name `` competition interface '' : the set @xmath281 consists precisely of those points @xmath240 such that @xmath282 while @xmath283 . in addition , if @xmath101 is in the competition interface ( i.e. if @xmath284 for some @xmath147 , or equivalently if @xmath285 for some @xmath11 ) then all the points @xmath286 to its right are in @xmath278 and all the points @xmath287 above it are in @xmath279 . see figure [ oneinterfacefig ] for an example of the competition interface .
an example of the evolution of the growth model with one competition interface .
the asterisks indicate those already occupied sites which form part of the competition interface .
for example , since the competition interface passes from @xmath107 to @xmath108 , @xmath116 and @xmath122 , we see that @xmath263 , that @xmath288 and that @xmath289 . ]
now we are finally ready to turn to the with _ two _ second - class particles , and the corresponding coexistence question in theorem [ growththm ] .
in the same way that the system with one second - class particle corresponded to a growth model with one interface , the system with two second - class particles will correspond to the growth model with two interfaces .
consider the starting from the state @xmath26{0.2}{\circle}}}{\underline { { \circledast}}}{{\circledast}}\hat{{\circle}}$ ] .
we can again replace the sites containing second - class particles by pairs of sites , each containing a hole and a particle in that order .
this representation works just as it did in the system with a single second - class particle , _ until the moment when one second - class particle tries to jump onto the other_. at this point we move from @xmath259{0.2}{\circle}}}][{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}]$ ] to @xmath290{0.2}{\circle}}}{{\textcolor[gray]{0.2}{\circle}}}$ ] and we can no longer maintain the two @xmath259{0.2}{\circle}}}]$ ] pairs ( although in fact one can identify a * pair in the configuration @xmath260{0.2}{\circle}}}]{{\textcolor[gray]{0.2}{\circle}}}$ ] and show that this * pair is in correspondence with a competition interface see the remark below ) . in any case , up until the collision moment , the two second - class particle representation is valid , and in particular theorem [ t1](a ) tells us that the probability that such a collision is ever attempted is @xmath38 .
we will again label the particles and holes sequentially .
the initial state is : in this state particle p1 has overtaken holes h1 and h2 , and particle p2 has overtaken hole h1 .
hence we put @xmath291 .
we will now keep track of both the @xmath258pairs : let @xmath292 and @xmath293 be respectively the labels of the hole and the particle covered by the left - hand @xmath258pair , and let @xmath294 and @xmath295 be those covered by the right - hand @xmath258pair .
we have @xmath296 and @xmath297 .
more generally , if at time @xmath11 the two @xmath258pairsare next to each other at time @xmath11 if and only if @xmath298 the local evolution both of @xmath299 and of @xmath300 depends on the values @xmath243 in just the same way as in the paragraph around ( [ phirec ] ) .
if we define @xmath301 and @xmath302 to be the jump chains of these two processes , then each of them separately obeys the same recurrence as the process @xmath303 at ( [ phirec ] ) .
now recall the rules of the growth model with two interfaces from section [ growthsec ] . in the initial condition ,
the sites @xmath107 , @xmath108 and @xmath109 , and all other sites are unoccupied .
when a new site becomes occupied , its colour is chosen as follows : * any site @xmath110 becomes dark blue .
* any site @xmath111 becomes light blue .
* the site @xmath113 becomes red .
* a site @xmath101 , @xmath112 , @xmath114 , @xmath115 , takes the colour of whichever of its parents became occupied _
more recently_. [ [ remark-1 ] ] _ remark _ + +
+ + + + + + in fact , we have elaborated slightly by introducing both dark and light blue colours at this point for convenience , to distinguish the clusters on either side of the red cluster .
the difference is convenient for the argument below , but it is not significant for the result , since we only ask whether or not the red cluster is unbounded , and we are free to identify the two blue clusters as before whenever we like .
in addition , once we make the distinction between the two blue clusters , we can consider the interface between those clusters _
after _ the red cluster disappears .
this interface is identified by the * pair resulting from the coalescence of the * pairs related to the light - blue / red and red / dark - blue interfaces , if we make the jump from @xmath259{0.2}{\circle}}}][{{\circle}}{{\textcolor[gray]{0.2}{\circle}}}]$ ] to @xmath260{0.2}{\circle}}}]{{\textcolor[gray]{0.2}{\circle}}}$ ] suggested above . as before ,
the processes @xmath304 and @xmath305 are competition interfaces , with the following properties .
let @xmath306 and @xmath307 .
then any point @xmath308 to the right of the lower / right interface is ( light ) blue , and any point @xmath309 which is above that interface is either red or ( dark ) blue .
similarly any point @xmath310 which is above the upper / left interface is ( dark ) blue , and any point @xmath311 which is to the right of that interface is either red or ( light ) blue .
an example of the evolution of the growth model with two interfaces .
here we see that @xmath312 , since the lower - right interface ( indicated by asterisks ) moved to @xmath117 rather than @xmath113 but the upper - left interface ( indicated by stars ) moved to @xmath113 rather than @xmath116 . in the state
shown , the two interfaces are at @xmath313 and @xmath314 .
if @xmath315 is smaller than both @xmath316 and @xmath317 then both interfaces will move next to the point @xmath318 . in this case
the point @xmath318 will be red but no further points will ever become red ; the red cluster will be surrounded .
if , instead , one of @xmath319 or @xmath320 is occupied before @xmath318 , the red cluster will grow further , perhaps unboundedly . ] from ( [ nexteq ] ) , we see that the moment ( if it ever happens ) when the two interfaces coincide for the first time is the first moment when the left - hand second - class particle attempts to jump onto the right - hand second - class particle .
once the interfaces meet in this way , the red cluster is `` surrounded '' and can not grow any further .
specifically , the last paragraph tells us that all the points above the upper - left interface will be ( dark ) blue and all those to the right of the lower - right interface will be ( light ) blue , so that no further points will be able to take on the colour red .
see figure [ meetingfig ] for an example .
on the other hand , suppose that the second - class particles have not yet collided , so that the interfaces have not yet coincided .
even if the interfaces are already adjacent , it is still possible for them to move apart again , and any sites between the two interfaces must take the colour red ; so certainly the red cluster is not yet surrounded .
see figure [ meetingfig ] for an illustrative example .
hence the event that the red cluster grows unboundedly corresponds to the event that the two second - class particles never meet . from theorem
[ t1](a ) , we know that that event has probability @xmath118 .
we thank soledad torres and karine bertin for nice discussions about their simulation results .
we are most grateful to a referee who pointed out an important error in an earlier version of the paper . | we consider the one - dimensional asymmetric simple exclusion process ( asep ) in which particles jump to the right at rate @xmath0 $ ] and to the left at rate @xmath1 , interacting by exclusion . in the initial state there is a finite region such that to the left of this region all sites are occupied and to the right of it all sites are empty . under this initial state ,
the hydrodynamical limit of the process converges to the rarefaction fan of the associated burgers equation .
in particular suppose that the initial state has first - class particles to the left of the origin , second - class particles at sites 0 and 1 , and holes to the right of site 1 .
we show that the probability that the two second - class particles eventually collide is @xmath2 , where a _ collision _ occurs when one of the particles attempts to jump over the other .
this also corresponds to the probability that two asepprocesses , started from appropriate initial states and coupled using the so - called `` basic coupling '' , eventually reach the same state .
we give various other results about the behaviour of second - class particles in the asep . in the totally asymmetric case ( @xmath3 )
we explain a further representation in terms of a multi - type particle system , and also use the collision result to derive the probability of coexistence of both clusters in a two - type version of the corner growth model . |
here we present standard techniques for working with gaussian fermions @xcite , and use them to show that decoherence is independent of bath temperature .
relaxation of gaussian fermionic open systems has be studied in detail ( see e.g. refs .
@xcite , but these did not include a model of the bath as also composed of gaussian fermions . the first step is to map @xmath4 dirac fermions ( e.g. electrons ) with annihilation and creation operators @xmath5 into @xmath6 majorana operators @xmath7 they are still fermionic in anti - commutation @xmath8 , but differ from dirac fermions in that they satisfy @xmath9 and @xmath10 . for any quantum state @xmath11 ,
the covariance matrix has real elements composed of second moments @xmath12 .\ ] ] due to fermion anticommutation , the covariance matrix is skew - symmetric @xmath13 . because of conservation of fermion parity , first moments always vanish @xmath14 . for gaussian states , expectation values of higher moments
are determined by the covariance matrix via wick s theorem .
likewise a quadratic hamiltonian @xmath15 is described by a matrix @xmath16 so that @xmath17 again , @xmath16 must be real for the hamiltonian to be hermitian , and furthermore @xmath16 can always be chosen skew - symmetric @xmath18 . for a closed quantum system evolving unitarily
, the covariance matrix evolves according to @xmath19,\ ] ] where @xmath20 $ ] is the commutator . for a time
independent hamiltonian , this results in @xmath21 a joint system - bath covariance matrix has the form @xmath22 where @xmath23 and @xmath24 represent , respectively , the system and bath covariance matrices , and @xmath25 records system - bath correlations . in other words , given a state @xmath11 with covariance matrix @xmath26 , tracing out the bath gives a reduced density matrix @xmath27 with covariance matrix @xmath28 .
we define @xmath29 to denote this process of reducing the covariance matrix , so @xmath30 . in general
, the reduced covariance matrix of an open quantum system will be @xmath31 for an uncorrelated system @xmath32 , the covariance matrix has a direct sum form @xmath33 .
the direct sum is linear , so that uncorrelated states have the form @xmath34 , where @xmath35 denotes an all zero matrices .
the covariance reduction @xmath29 is also linear , and so uncorrelated states evolve to @xmath36 notice that the first term is independent of the bath variables such as temperature , and can be more compactly written as @xmath37 where @xmath38 .
we are interested in the rate of decoherence .
how quickly will two states become indistinguishable ?
consider two different initial covariance matrices @xmath39 and @xmath40 , e.g. describing logical encodings of qubit states .
the time evolved difference between these covariance matrices is @xmath41 we observe that this is entirely independent of the bath temperature . as @xmath42 ,
the states becomes indistinguishable .
using @xmath43 to denote the operator norm ( the largest singular value ) of a matrix , it is straightforward to show @xmath44 \leq || \delta(t ) || , \ ] ] where @xmath45 and @xmath46 are any pair of anti - commuting majorana operators .
therefore , small @xmath47 entails low probability of distinguishing @xmath11 and @xmath48 through a single gaussian measurement .
we show later that this statement can be extended to completely general measurements .
the operator norm is submultiplicative and transpose invariant so that @xmath49 with smaller @xmath50 entailing more decoherence .
we have used @xmath51 to present an upperbound that is also independent of the initial state . without system - bath interactions
@xmath52 is unitary so that @xmath53 , but interactions lead to dissipation and @xmath54 . under very general conditions
we have determined that gaussian decoherence occurs , quite remarkably , independently of the bath temperature . in some instances @xmath50 may decrease with time , only to revive later . however , for a sufficiently complex bath we expect markovian behavior lead to exponentially fast decoherence @xmath55 for some @xmath56 .
next we introduce the formalism of gaussian fermionic master equations , and then proceed to perform a microscopic derivation assuming weak - coupling . in such derivations various approximations are made , yet we find they respect temperature invariance .
here we review gaussian fermionic master equations following refs .
@xcite . the dynamical equation for such
a system is @xmath57 where , for dissipative dynamics , @xmath58 is not necessarily skew - symmetric . in general , @xmath59 where @xmath16 represents the unitary component ( and is skew symmetric ) and @xmath60 represents the dissipative part ( and is symmetric and positive ) .
typically , dissipative systems will have one steady state @xmath61 that satisfies @xmath62 , and then @xmath63 these very general dynamical equations are the covariance matrix representation of a class of lindblad master equations of the form @xmath64 + \sum_{\mu } ( 2 l_{\mu } \rho(t ) l_{\mu}^{\dagger } - \ { l_{\mu}^{\dagger } l_{\mu } , \rho(t ) \ } ) , \ ] ] where the hamiltonian @xmath16 must be of the quadratic form introduced in eq .
( [ eqquadham ] ) and the lindblad operators are linear in majorana operators @xmath65 one can prove @xcite that such a master equation gives rise to a gaussian quantum channel with @xmath66 where @xmath67 has elements @xmath68 .
the matrices @xmath58 and @xmath69 will always be real so that @xmath70 and @xmath71 . in the literature , there is some discussion of how such systems decohere majorana qubits .
however , prior work has centered on only the kitave wire and has not considered how variables in lindblad equations depend on the underlying system parameters .
we quantify the decoherence by the operator norm of @xmath72 . in general , @xmath73 and from the previous section
we see that @xmath74 .
theorem ix.3.1 of bhatia @xcite shows that for any @xmath58 , we have @xmath75 as previously employed in this setting by bravyi - knig @xcite .
using the shorthand @xmath76 , this entails @xmath77 we know @xmath78 is non - negative , so provided @xmath60 does not have any zero eigenvalues there will be exponentially rapid decoherence .
it is well known that decoherence rates are governed by the spectral properties of @xmath60 .
our contribution is investigation of the dependence of this spectrum on microscopic factors .
the above statement holds for any pair of gaussian states and any gaussian channel .
we are also interested in the more specific scenario where the gaussian states lie in the degenerate groundspace of a physical hamiltonian .
eigenstates of this hamiltonian can be simultaneously diagonalised , and similarly their covariance matrices can be simultaneously brought into williamson normal form .
there exists a orthogonal matrix @xmath79 , such that for every pure gaussian @xmath80 that is a eigenstate of the hamiltonian we have @xmath81 where @xmath82 distinguish different eigenstates , and @xmath83 for some values of @xmath84 , the numbers @xmath85 tell us whether there is an excitation present .
however , for degenerate hamiltonians there are a set of degeneracy indices @xmath86 , such that @xmath85 can vary without creating excitations for all @xmath87 .
therefore , the covariance matrices of groundstates break up into two blocks @xmath88o^t$ ] where @xmath89 is the same for all groundstates , and @xmath90 let us consider two encoded ground states @xmath91 and @xmath92 .
we deduce that @xmath93 o^t$ ] where @xmath35 is a zero matrix .
the matrix representing hamiltonian dynamics has the same block structure as the covariance matrices , so that @xmath94 o^t$ ] .
when such a system is exposed to an environment , its dynamics are dictated by some matrix @xmath58 . in many situations ( including weakly coupled markovian systems )
, @xmath58 will obtain the same block structure as @xmath16 , so that @xmath95 o^t$ ] .
it follows that @xmath96o^t .\ ] ] defining @xmath97 , and using the same arguments as earlier we find @xmath98 therefore , the decoherence of encoded ground states is governed by the spectrum of @xmath99 .
the above arguments tell us that two initial gaussian states undergoing markovian dynamics will exponentially converge towards having identical covariance matrices .
therefore , the probability of distinguishing these states through a single gaussian measurement decreases exponentially in time .
however , a non - gaussian measurement or multiple gaussian measurements could prove more successful . in general , any strategy for distinguishing two states
can always be captured by an observable @xmath67 with eigenvalues @xmath100 , with an average success probability @xmath101).\ ] ] it is well known that @xmath102 where the trace norm is @xmath103 $ ] . therefore , we aim to show convergence in 1-norm .
we again compare two initial pure states encoding a qubit into 4 majorana modes , and find that the time evolved density matrices states @xmath104 and @xmath105 satisfies @xmath106 this follows quickly from eq .
( [ deltacon ] ) as we show in app .
[ apptrnorm ] .
in this section we present a weak - coupling derivation of a gaussian master equation in the covariance matrix formalism .
many of the steps directly mirror those made in a textbook density matrix derivation ( see e.g. chap 3 of ref .
we assume both the system , the heat bath and their interaction is entirely gaussian .
in addition , we make the usual assumptions involved in deducing master equations , notably that system - bath coupling is weak and that the system - bath are effectively uncorrelated at all times .
the whole system - bath dynamics are described by a hamiltonian with block matrix structure @xmath107 where @xmath108 and @xmath109 represent , respectively , the system and bath hamiltonians and satisfy @xmath110 for @xmath111 .
the interaction is represented by @xmath112 a real - valued , not necessarily square , matrix .
the initial ( @xmath113 ) system - bath convariance matrix has the form @xmath114 before proceeding we shift to an interaction picture , defining @xmath115 .
it follows that @xmath116 , \ ] ] where @xmath117 this simplifies to @xmath118 where @xmath119 .
once in the interaction picture , we integrate over time to find @xmath120 ds .\ ] ] therefore the time derivative is @xmath121 \\
\nonumber & + \int_{0}^{t } [ [ \gamma_{\mathrm{int}}(s ) , h_{\mathrm{int}}(s ) ] , h_{\mathrm{int}}(t ) ] ds . \end{split}\ ] ] we are only interested in the system covariance matrix , which is the covariance reduction @xmath122 of the above expression .
it is straightforward to verify @xmath123 \right\}_{b}=0 $ ] , so @xmath124,h_{\mathrm{int}}(t ) ] \}_{b } ds .\ ] ] next , we assume the coupling is weak and that the system stays uncorrelated from the bath . formally , this entails that @xmath125 , and also that @xmath126 and the integral is extended to infinity .
such assumptions directly mirror those made on the level of hilbert spaces and result in the expression @xmath127,h_{\mathrm{int}}(t ) ] \}_{b } ds , \ ] ] where have made use of the shorthand @xmath128 .
next we may use our knowledge of the block structure of the covariance matrices to evaluate the commutator , and find @xmath129 , h_{\mathrm{int}}(t ) ] \}_{b } = \\ -h_{i}^{t}(t)h_{i}(t - s)\tilde{\gamma}(t ) - \tilde{\gamma}(t)h_{i}^{t}(t - s ) h_{i}(t ) \\
+ h_{i}^{t}(t)\gamma_{b } h_{i}(t - s ) + h_{i}^{t}(t - s)\gamma_{b } h_{i}(t ) .\end{gathered}\ ] ] combing this with eq .
( [ eqdiffint ] ) , and collecting terms to match eq .
( [ eqgenquantumchannel ] ) we have @xmath130 where @xmath131 we have succeeded in deriving a form of a gaussian quantum channel . though to make these equations meaningful we require that the integrals converge to finite values . for finite size matrices
the integrands will be periodic functions and typically do not converge to a finite value . whereas , in the limit of infinite matrices the eigenvalue spectrum may become continuous and the integrand may vanish in the large @xmath132 limit .
furthermore , to yield markovian dynamics the resulting @xmath58 and @xmath69 must be time independent . before proceeding we
can already observe that all @xmath133 dependence has vanished from @xmath58 .
presently , the matrix @xmath58 still carries an overt time dependence , which can be removed by making the secular approximation ( sa ) .
first we recall the explicit time dependence , @xmath134 so that @xmath135 we proceed by noting that @xmath108 is real and skew - hermitian , so it has imaginary eigenvalues @xmath136 , eigenvectors @xmath137 , and a diagonal form @xmath138 this entails @xmath139 where @xmath140 the sa asserts that terms with rapidly oscillating phases @xmath141 can be neglected , except of course when @xmath142 .
this is valid at times longer than the reciprocal of the smallest energy gaps , @xmath143^{-1 } $ ] . for
now , we assume this to be true , but later we will show that a much weaker energy gap condition entails many of the same features .
making the sa leads to : @xmath144 we see that sa has removed any dependence on @xmath145 making time evolution markovian .
furthermore , the sa forces @xmath58 to commute with @xmath108 , and so @xmath58 has the same block - diagonal structure as @xmath146 .
all matrices can be decomposed into @xmath59 where @xmath147 and @xmath148 .
performing just such a decomposition of @xmath58 we can show , via bochner s theorem , that the hermitian part @xmath60 has eigenvalues that are real and nonnegative ( see app .
[ appbochner ] ) .
furthermore , both matrices have real - valued elements , so @xmath149 and @xmath150 , which entails that @xmath16 has purely imaginary eigenvalues , just as expected . in sec .
[ sec_convergence ] we saw that decoherence rates are governed by @xmath60 .
recall that @xmath151 is the restriction of @xmath60 to the kernel of @xmath146 ( equivalently the groundspace of the associated hamiltonian , assuming @xmath152 ) , and the decoherence rates between groundstates are governed by the spectrum of @xmath151 .
furthermore , this restricted @xmath151 matrix naturally emerges when one considers a relaxed sa assumption .
recall that the validity of sa required that all energy gaps are large compared to a relevant time scale .
many interesting topological systems have a degenerate groundspace with a large gap to the first excited state , but then the spectrum of excitations will be dense or even a continuum in the large system limit .
this means that sa can not be used to eliminate transitions between different excited states .
however , provided the groundspace is gapped from excitations , we will have a limited application of sa that decouples @xmath58 into @xmath153 and with @xmath154 describing the dynamics within the ground space .
although sa will not apply to the dynamics @xmath155 of the excitations , we saw in sec .
[ sec_convergence ] that decoherence in the groundspace is governed by only @xmath154 .
in particular , it is dictated by the largest eigenvalue of @xmath151 .
the bath temperature only influences what state we converge to , and not how quickly we get there .
this conclusion is quite remarkable .
so much so , that naively it seems to violate some basic tenet of physics .
two candidates are the arrhenius law and detailed balance .
the arrhenius law is an empirical rule of thumb that has been successful in modeling chemical reactions .
recently , it has been suggested that it may also apply to quantum memories , though violations have been observed in various settings @xcite .
the arrhenius law predicts that decoherence times scale as @xmath156 where @xmath0 is the system gap .
we have a gapped degenerate ground space , but quasiparticles from the environment can poison the system without creating an excitation . from this perspective
perhaps we should consider @xmath157 , and our temperature independence to be consistent with the arrhenius law .
however , although we have focused on ground space decoherence our observation apply also to the dynamics of excitations with a discrete spectrum .
that is , the rate at which excitation populations equilibrate is also temperature independent !
indeed , the arrhenius law is not a universal law and we conclude that it is absolutely violated in the domain of gaussian fermions .
detailed balance is a form of microscopic reversibility .
it states that at thermal equilibrium the population transfer is symmetric for each process .
consequently , the rate of transitions must depend on the temperature . at first
this seems to imply that decoherence rates must be temperature dependent .
indeed , detailed balance has been used to study decoherence times of spin ( qubit ) systems with the toric code hamiltonian @xcite and cubic code hamiltonian @xcite .
these results revealed an exponential temperature dependence , and so one may expect this feature to be generic .
though these are highly non - gaussian systems . in the gaussian
setting , temme _ et .
@xcite have rigorously proved very general bounds on thermalization rates .
their analysis appears to show an explicit temperature dependence , but closer inspection reveals variables that depend on the specific features of the bath .
these variables can be set , whilst still respecting detailed balance , to precisely cancel all temperature dependence .
we have seen that when the whole system - bath is gaussian , this is indeed what happens .
to further illustrate that temperature independence is consistent with detailed balance we present in app .
[ appdb ] a very simple classical markov process where convergence rates decouple from temperature .
detailed balance has an esteemed history going back to boltzmann , who proclaimed it a key axiom of statistical mechanics and used it to great effect .
however , there has been a recent surge of interest in non - equilibrium statistical mechanics , where detailed balance is violated , both in quantum @xcite and classical settings @xcite .
mazza _ et .
@xcite studied the effect of various noise models on the 1-dimensional kitaev chain .
their main result is that various hamiltonian perturbations will decohere the system , but this decoherence is suppressed by increasing the length of the kitaev chain .
conclude their paper by also discussing a more destructive noise model , open systems dynamics ( see pg.4 of ref .
this model is a master equation with the hamiltonian of the standard kitaev wire ( with chemical potential set to zero ) , and lindblad operators @xmath158 which allows for fermions hopping to the environment and where we use @xmath159 to parameterize the strength of the hopping .
they give numerical plots for decoherence of this model , but it can also be understood analytically .
we proceed by casting these master equations in the language of covariance matrices , and find@xmath160 composing @xmath161 will cancel the imaginary parts , giving @xmath162 note that the eigenvalues of @xmath16 are purely imaginary so that @xmath163 is unitary , but the dissipative component adds a constant real negative component .
we have that @xmath164 and @xmath165
. therefore , decoherence occurs at the rate @xmath166 . here
it is clear that system size and hamiltonian gap play no role , provided @xmath159 is not a function of these variables .
al . _ do not discuss how @xmath159 itself might depend on the energy gap or on temperature .
budich _ et .
@xcite made similar observations .
they considered two models where only the end of a kitaev wire couples to the environment . in both models
the system - environment coupling has the form @xmath167 where @xmath168 is some operator acting on the bath .
they also considered the standard kitaev chain with zero chemical potential , so that the interaction commuted with the system hamiltonian .
they observed rapid decoherence of majorana edge modes , with no dependence on the energy gap .
these toy models are excellent ways to illustrate a serious deficit in prior claims to effectiveness of topological protection of majorana edge modes .
however , they tell us little about what to expect when the interaction and hamiltonian do not commute . furthermore , one may also wish to consider much more exotic models , such as majorana fermions in 2d systems or even higher dimensions .
these gaps in previous work are now filled by the more general insights presented here .
we have seen that decoherence of gaussian fermionic systems can not be reduced by cooling . for markovian dynamics
, we provided a microscopic derivation of the master equation in the weak coupling limit , leading to exponentially fast decoherence .
therefore , to use gaussian systems as a quantum memory they must be either highly non - markovian or have minimal tunneling with any nearby gaussian heat baths .
eliminating tunneling is potentially challenging when in any of the many popular proposals for acquiring topological order through the proximity effect @xcite . in these proposals ,
electron hopping with an external @xmath132-wave superconductor is the mechanism by which topological robustness is acquired .
both electron hopping and the superconducting hamiltonian are gaussian , and so this opens the door to temperature invariant decoherence .
for such systems it is urgent that we acquire a better understanding of the proximity effect from an open systems perspective . after completing this work ,
the author forwarded the manuscript to leonardo mazza who in return kindly shared several unpublished yet interesting results @xcite .
these tackle related problems of interactions with fermionic baths , including various numerical simulations of small fermionic ( and bosonic ) baths and numerous analytic insights .
particularly relevant is sec .
7 of his phd thesis @xcite , where mazza makes several observations also made here , although does he not remark on the temperature independence of decoherence rates .
we thank pieter kok and keith burnett for interesting discussions on majorana fermions that lead to this research .
we thank michael kastoryano , tomaz prozen , jens eisert and leonardo mazza for comments on the manuscript .
we assume initial states of the form @xmath169 and @xmath170 encoding different qubit states .
these initial states differ only on 4-majorana modes within the groundspace , and the evolution of the covariance matrix shows that this property holds at later times so that @xmath171 therefore , @xmath172 where we have used @xmath173 and @xmath174 .
the hilbert space of 4 majorana modes supports one qubit in the even parity subspace and one qubit in the odd parity subspace . in other words , @xmath175 and similarly @xmath176 . using @xmath177 we have @xmath178 for a single qubit , we have @xmath179 $ ] where the maximum is over all single qubit hermitian unitary operators , such as the pauli spin operators . in a majorana encoding the pauli spin operators , indeed all single qubit hermitian unitary operators , are quadratic observables .
these expectation values never exceed the operator norm of the @xmath180 . therefore , @xmath181 finally , we make use of eq .
( [ deltacon ] ) to arrive at eq .
( [ trnormcon ] ) .
here we describe the concept of detailed balance for 2-state systems .
we show in this simple setting the concept is consistent with temperature invariant decoherence rates .
furthermore , we show that gaussian 2-mode markov processes always obey this principle .
consider , a classical system with two possible states , with probabilities described by a markov chain @xmath182 for simplicity we consider time to be in discrete steps , with a transition matrix @xmath183 so that @xmath184 .
conserving flow of probability requires @xmath185 for @xmath186 .
we say @xmath187 is a stationary state of @xmath60 , if @xmath188 , and where @xmath189 .
the process @xmath60 satisfies detailed balance if @xmath190 one can think of @xmath187 as a thermal distribution , so @xmath191 and @xmath192 , where @xmath193 is the partition function . as usual
, @xmath194 is inverse temperature . in this thermal language
, detailed balance entails that @xmath195 where @xmath0 is the energy gap @xmath196 .
it appears that the ( ratio of ) transition rates depend on the temperature of the steady state , and so one might be tempted to conclude that convergence rates likewise depend on temperature .
the conservation of probability and detailed balance give 3 independent linear constraints on @xmath60 , out of the 4 parameters of the matrix .
therefore , the space of valid matrices is 1-dimensional and includes @xmath197 both matrices satisfy eq .
( [ eqdetbal ] ) .
furthermore , for all markov chains @xmath198 we have @xmath199 and @xmath200 .
the whole set of suitable matrices is contained in the span of these matrices , @xmath201 with @xmath202 \leq 2 $ ] to ensure @xmath203 $ ] . within these limits
, @xmath159 is a free parameter .
we consider a general initial probability distribution it always has the form @xmath204 for some value @xmath205 , where @xmath206 .
it is easy to confirm @xmath207 and @xmath208 .
therefore , @xmath209 and for @xmath145 time steps this extends to @xmath210 this clearly shows exponentially rapid convergence to the equilibrium state at a speed governed by @xmath159 .
more precisely , using any norm @xmath211 to measure distance we have @xmath212 the convergence speed is entirely independent of temperature , and only depends on the free parameter @xmath159 .
the only temperature dependence lies in @xmath213 $ ] , since @xmath214 depends on temperature .
however , @xmath215 $ ] and so the range @xmath216 $ ] is valid at all temperatures .
therefore we can consider a family of markov process with varying temperature and constant @xmath216 $ ] .
this family is consistent with detailed balance , but has a convergence rate independent of temperature .
the convergence rate could vary with temperature if @xmath159 is a non - constant function of temperature .
although , there is no reason _ a piori _ to favour one function for @xmath159 over another . certainly , many possible temperature dependencies are consistent with detailed balance . unless , one has a physical model of the encompassing system and can perform a microscopic derivation of the markov process , and so derive @xmath159 .
this is exactly what we have performed for the case of gaussian fermions , showing the analogous result of temperature independent @xmath159 .
lastly , we remark that this entire discussion can be recast in continuous time by considering @xmath60 to be the generator of a markov process with transition matrix @xmath217 .
it is still interesting to ask if gaussian markov processes obey detailed balance .
let us just consider a pair of modes , with a 2-by-2 covariance matrix @xmath218 the physical system is in one of two states ( superpositions are disallowed by fermion parity superselection ) , with probability @xmath219 and @xmath220 .
a markov process maps @xmath221 where @xmath69 is skew - symmetric . under this process , we find some @xmath222 such that @xmath223
. therefore , @xmath224 and the probability transition matrix has the form @xmath225 in the steady state @xmath226 and so that @xmath227 therefore , @xmath228 so that detailed balance is satisfied for all @xmath222 .
here we show that the hermitian part of @xmath58 has strictly negative eigenvalues .
the proof makes use of bochner s theorem , which relates properties of functions to their fourier transform . to introduce this theorem
, we first define the concept of functions of positive - type returning to the problem at hand , @xmath234 is hermitian by construction . from eq .
( [ eq_x_rwaed ] ) we already have an expression for @xmath58 , and considering @xmath235 we observe that @xmath236 where we have made the change of variables @xmath237 . switching the order of integration , @xmath238 next
, we use that @xmath239 , which can be seen from @xmath240 and using @xmath241 we have the result . applying this to our expression for @xmath235 , and switching the dummy variables @xmath242 gives @xmath243 this differs from @xmath58 in only the domain of the integral and so @xmath244 for each set of variables , @xmath245 , the integral is a fourier transform of @xmath246 evaluated at @xmath247 , so @xmath248 notice , we have denoted fourier transforms with a tilde .
next , we show that @xmath246 is of positive - type . for all @xmath249 @xmath250 where @xmath251 since @xmath252 , we can apply bochner s theorem and conclude that all @xmath253 are positive and real .
if @xmath108 is a nondegenerate matrix , then there would be no multiplicity of eigenvectors with the same eigenvalue and @xmath254 would represent the real - negative eigenvalues of @xmath255 .
however , for degenerate matrices there is a freedom of choice in the basis @xmath256 , but we can always set this to be the eigenbasis of @xmath255 .
therefore , @xmath255 has real negative eigenvalues .
40ifxundefined [ 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 * * , ( ) http://stacks.iop.org/0034-4885/75/i=7/a=076501 [ * * , ( ) ] http://stacks.iop.org/1063-7869/44/i=10s/a=s29 [ * * , ( ) ] link:\doibase 10.1126/science.1222360 [ * * , ( ) ] @noop * * , ( ) http://dx.doi.org/10.1038/nnano.2013.33 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://stacks.iop.org/1742-5468/2010/i=07/a=p07020 [ * * , ( ) ] http://stacks.iop.org/1367-2630/12/i=2/a=025016 [ * * , ( ) ] @noop ( ) @noop ( ) http://stacks.iop.org/1742-5468/2015/i=2/a=p02008 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.121405 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.205142 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.205109 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.075431 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.174533 [ * * , ( ) ] @noop ( ) @noop ( ) link:\doibase 10.1103/physrevb.72.024526 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.127002 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/1751-8121/47/i=40/a=405303 [ * * , ( ) ] @noop _ _ , vol .
( , ) @noop _ _ ( , ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.100.096407 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.094522 [ * * , ( ) ] _ _ , @noop ph.d . thesis , ( ) _ _ , @noop master s thesis , ( ) @noop | coupling to a thermal bath leads to decoherence of stored quantum information . for a system of gaussian fermions , the fermionic analog of linear or gaussian optics ,
these dynamics can be elegantly and efficiently described by evolution of the system s covariance matrix . taking both system and bath to be gaussian fermionic
, we observe that decoherence occurs at a rate that is independent of the bath temperature .
furthermore , we also consider a weak coupling regime where the dynamics are markovian .
we present a microscopic derivation of markovian master equations entirely in the language of covariance matrices , where temperature independence remains manifest .
this is radically different from behaviour seen in other scenarios , such as when fermions interact with a bosonic bath .
our analysis applies to many majorana fermion systems that have been heralded as very robust , topologically protected , qubits . in these systems
, it has been claimed that thermal decoherence can be exponentially suppressed by reducing temperature , but we find gaussian decoherence can not be cooled away .
thermalization through interaction with an external bath is one of the principal mechanisms by which quantum systems lose information . in quantum technologies ,
rapid thermalisation destroys their advantage over classical counterparts . by better understanding these processes , one hopes to identify and engineer physical systems that act as more robust stores of quantum information . in topologically ordered systems , information is stored non - locally within the degenerate ground space of some large many - body system .
the primary benefit of topology is robustness against random adiabatic fluctuations in the system hamiltonian .
damage from such noise is exponentially suppressed with system size .
topological systems also have an energy gap @xmath0 between the degenerate ground space and excited states , and are said to be protected by the gap against thermal excitations . a common claim
@xcite is that thermal processes occur at a rate @xmath1 , which is sometimes called the arrhenius law .
the bold conclusion is that topology can exponentially eliminate noise merely by increasing system size and decreasing temperature . of all topological systems ,
majorana zero modes have attracted the most attention .
it was theorized that a so - called kitaev wire supports majorana zero modes at edges , which could be realised in simple solid state hetrostructures @xcite , for example a nanowire coupled to a conventional s - wave superconductor @xcite .
this drove a series of experiments , eventually leading to observations of majorana edge modes @xcite . beyond topological robustness
, majorana zero - modes also possess the braiding statistics of non - abelian ising anyons .
though insufficient for direct quantum computation , braiding ising anyons can demonstrate nonlocality , teleportation and superdense coding @xcite . furthermore ,
ising anyon braiding can be promoted to full quantum computing when supplemented with some nontopological ( noisy ) operations @xcite .
the physics of these majorana systems is especially tractable as their hamiltonians are quadratic in fermion creation and annihilation operators .
we say such a system is gaussian , or quasifree fermionic , in analogy with gaussian linear optics .
gaussian states can be described purely in terms of the expectation value of quadratic observables , which are captured by a covariance matrix .
furthermore , some dissipative processes can be described within this powerful covariance matrix formalism @xcite , and allow single fermions to hop between system and bath via @xmath2 .
single fermion hopping violates conservation of fermion parity in the system , which is otherwise respected by unitary evolution .
it is a toxic process that can cause errors without creating excitations , circumventing arguments that energy penalties suppress thermal processes to a rate @xmath1 .
in particular , majorana modes in the kitaev wire ( see fig . [ figwirebath ] ) have been shown to decohere due to fermion hopping at rates independently of system size or the system gap @xcite , and we will review these results in detail .
this article considers all gaussian fermionic systems , not just the kitave wire , and how they decohere as a function of temperature .
a single fermion appearing in the system will have a partner appear in the environment , and so perhaps there is hope that a gapped bath hamiltonian will provide an energy penalty inhibiting these processes .
we present a very general , yet simple , argument that decoherence is independent of temperature , assuming only that the system - bath is governed by a gaussian hamiltonian .
we extend this argument by providing a microscopic derivation of a master equation in the weak coupling regime , and again observe temperature independent decoherence .
when fermions couple to bosonic baths , or through quartic fermion - fermion interactions @xmath3 , one would instead find a non - trivial temperature dependence .
any real physical system will experience noise from multiple sources , mostly with temperature dependent rates .
however , fermionic hopping presents a constant background noise that can not be suppressed through cooling .
this adds to a growing body of work @xcite that shows the outlook for majorana fermions makes them less promising quantum memories than initially supposed .
our conclusions can be avoided by going beyond gaussian fermions , for instance by making use of complex many - body interactions used in the passive quantum memories reviewed in ref .
@xcite .
we discuss how our results demonstrate a break down of the arrhenius law , whilst still satisfying a notion of detailed balance .
decoherence of two - level systems , such as spins , has been studied when they couple to spin or fermion baths @xcite where a temperature dependence is observed but at low temperatures relaxation rates plateau , similarly breaking the arrhenius law . throughout , we use the phrase thermalize as synonymous with equilibrate or approach steady state .
the reader should not infer that the steady state is the thermal gibbs distribution with respect the system hamiltonian and ambient temperature , as may not be the case . |
reduced decompositions of permutations are classical objects in combinatorics that appear throughout the literature . following the work of rodica simion and frank schmidt in @xcite , the study of permutation patterns , particularly pattern avoidance , has become a frequently studied field as well . in @xcite , sara billey , william jockusch , and richard stanley
relate these two concepts , possibly for the first time .
there they show that @xmath0-avoiding permutations are exactly those permutations where the subsequence @xmath1 never occurs in a reduced decomposition .
relatedly , victor reiner shows in @xcite that the number of @xmath2 occurrences in reduced decompositions of the longest element in the symmetric group , which has the maximal number of occurrences of @xmath0 , is equal to the number of such reduced decompositions .
stanley had previously shown that this is the number of standard young tableaux of a staircase shape in @xcite . inspired by these results , and more generally by the relationship they suggest between the two aspects of permutations , this paper studies elements of the symmetric group from the combined perspectives of their reduced decompositions and their patterns .
while these aspects of a permutation appear extensively in combinatorial literature , they are not often treated together .
this paper strives to remedy that fact , addressing several questions where reduced decompositions and permutation patterns together lead to interesting results . after introducing basic terminology and notation in section [ section defs ] , section [ section vexillary ]
generalizes the result of billey , jockusch , and stanley , via a new characterization of vexillary permutations in theorem [ vexthm ] .
this characterization is based on the reduced decompositions of the permutations _ containing _ the permutation in question , and is strikingly different from all previous equivalent characterizations .
in addition to requiring that each of the permutations containing the vexillary permutation has a certain kind of reduced decomposition , the proof of theorem [ vexthm ] explicitly constructs such a reduced decomposition .
there are three algorithms which appear in this paper , the first of which occurs in the proof of theorem [ vexthm ] .
it should be noted that these are not deterministic , and include a certain amount of choice .
for instance , example [ vex ex ] describes only one possible route that the algorithm may take on a particular input .
there is an equivalence relation , sometimes known as the commutation relation , on the set of reduced decompositions of a particular permutation .
this and an associated graph are discussed in section [ section eq ] .
theorem [ 1 lbm ] and corollary [ 1 ulbm ] characterize permutations with graphs and commutation classes having certain properties .
these results are strengthened in theorem [ chainthm ] . the results in sections [ section polygon ] and [ section poset ] discuss permutation patterns with respect to a polygon defined by serge elnitsky in @xcite .
the rhombic tilings of this polygon are in bijection with the commutation classes of a permutation .
new results include that the number of commutation classes of a permutation is monotonically increasing with respect to pattern containment ( theorem [ monotone ] ) , and several results pertaining to a poset associated with tilings of the polygon . finally , section [ section fb ] completely describes this poset in the case of a freely braided permutation , as defined by richard green and jozsef losonczy in @xcite and @xcite .
the main definitions and notation that appear throughout the paper are discussed below . for more information about these objects , including proofs of elementary facts , see @xcite and @xcite .
let @xmath3 denote the symmetric group on @xmath4 elements .
an element @xmath5 permutes @xmath6 by mapping @xmath7 .
this permutation will be written in one - line notation @xmath8 .
@xmath9 maps @xmath10 to @xmath11 , @xmath12 to itself , @xmath13 to @xmath10 , and @xmath11 to @xmath13 . for @xmath14 , the map @xmath15 transposes @xmath16 and @xmath17 , and fixes all other elements in a permutation .
the symmetric group @xmath3 is the coxeter group of type @xmath18 , and it is generated by the adjacent transpositions @xmath19 .
the adjacent transpositions satisfy the coxeter relations : @xmath20 equation is called the _ short braid relation _ , and equation is the _ long braid relation_. a map is written to the left of its input , so @xmath21 interchanges the positions of values @xmath16 and @xmath17 in the permutation @xmath22 , while @xmath23 interchanges the values in positions @xmath16 and @xmath17 in @xmath22 .
if @xmath24 , then @xmath25 . because the symmetric group is generated by adjacent transpositions , any permutation @xmath5 can be written as @xmath26 for some @xmath27 .
the least such @xmath28 is the _ length _ of @xmath22 , denoted @xmath29
. an _ inversion _ in @xmath22 is a pair @xmath30 where @xmath31 and @xmath32 .
inversion set _ is @xmath33 . since @xmath34 \times [ 1,n]$ ] , the inversion set can also be viewed as an array .
the number of inversions in @xmath22 is equal to @xmath29 ( see @xcite ) . for obvious reasons ,
the permutation @xmath35 is called the _
longest element _ in @xmath3 . for a permutation @xmath22 with @xmath36 , a string @xmath37 such that @xmath26 is a _ reduced decomposition _ of @xmath22 .
( some sources call this a _ reduced word_. ) the set @xmath38 consists of all reduced decompositions of @xmath22 .
a _ factor _ is a consecutive substring of a reduced decomposition .
similar to the coxeter relations , a factor @xmath39 in a reduced decomposition will be called a _ short braid move _ if @xmath40 , and a factor @xmath41 will be called a _
long braid move_. the set @xmath38 has been studied in various contexts , notably by stanley in @xcite . there , stanley computes @xmath42 for several classes of permutations in terms of the number of standard young tableaux of certain shapes .
in the case of a vexillary permutation , this is equal to the number of standard young tableaux of a single shape @xmath43 ( see also exercise 7.22 of @xcite ) .
the definition of vexillary permutations is postponed until section [ section vexillary ] , where they will be discussed in depth .
let @xmath24 and @xmath44 for @xmath45 .
the permutation @xmath22 _ contains the pattern @xmath46 _ if there exist @xmath47 such that @xmath48 is in the same relative order as @xmath49 .
that is , @xmath50 if and only if @xmath51 .
if @xmath22 does not contain @xmath46 , then @xmath22 _ avoids _
@xmath46 , or is _
@xmath46-avoiding_. suppose that @xmath22 contains the pattern @xmath46 , with @xmath52 as defined above .
then @xmath48 is an _ occurrence _ of @xmath46 in @xmath22 .
the notation @xmath53 will denote the value @xmath54 . if @xmath55 , then @xmath56 .
let @xmath57 , @xmath58 , and @xmath59
. then @xmath60 is an occurrence of @xmath46 , with @xmath61 , @xmath62 , @xmath63 , and @xmath64 .
also , @xmath65 .
the permutation @xmath22 is @xmath66-avoiding .
let @xmath22 contain the pattern @xmath46 , and let @xmath67 be a particular occurrence of @xmath46 .
if @xmath68 , then @xmath69 is a _ pattern entry _ in @xmath22 . otherwise @xmath69 is a _ non - pattern entry_. if a non - pattern entry lies between two pattern entries in the one - line notation for @xmath22 , then it is _ inside _ the pattern .
otherwise it is _ outside _ the pattern .
`` inside '' and `` outside '' are only defined for non - pattern entries .
[ obstruction ] let @xmath67 be an occurrence of @xmath70 in @xmath22 .
suppose that @xmath71 is inside the pattern , that @xmath72 for some @xmath73 $ ] , and that the values @xmath74 appear in increasing order in the one - line notation for @xmath22 .
let @xmath75 be maximal so that the values @xmath76 appear in increasing order in the one - line notation for @xmath22 .
the entry @xmath71 is _ obstructed to the left _ if a pattern entry smaller than @xmath77 appears between @xmath77 and @xmath71 in @xmath22 .
likewise , @xmath71 is _ obstructed to the right _ if a pattern entry larger than @xmath78 appears between @xmath71 and @xmath78 in @xmath22 .
let @xmath79 and @xmath80
. then @xmath81 and @xmath82 are both occurrences of @xmath46 in @xmath22 .
obstruction is only defined for the latter , with @xmath83 and @xmath84 .
then @xmath85 , and @xmath11 is obstructed to the left and not to the right .
[ nonvex obst ] let @xmath86 and @xmath87 .
then @xmath88 is an occurrence of @xmath46 in @xmath22 . using @xmath89 , @xmath90 in definition [ obstruction ]
shows that @xmath91 , and @xmath13 is obstructed both to the left and to the right .
vexillary permutations first appeared in @xcite and subsequent publications by alain lascoux and marcel - paul schtzenberger .
they were also independently found by stanley in @xcite .
there have since emerged several equivalent definitions of these permutations , and a thorough discussion of these occurs in @xcite .
the original definition of lascoux and schtzenberger , and the one of most relevance to this discussion , is the following .
[ 2143vex ] a permutation is _ vexillary _ if it is @xmath92-avoiding .
the permutation @xmath93 is vexillary , but @xmath94 is not vexillary because @xmath95 is an occurrence of @xmath92 in the latter .
the following proposition is key to proving one direction of theorem [ vexthm ] .
[ vex not obst ] let @xmath22 contain the pattern @xmath46 .
let @xmath71 be inside the pattern , with @xmath72 and the values @xmath74 appearing in increasing order in @xmath22 .
if @xmath46 is vexillary then @xmath71 can not be obstructed both to the left and to the right .
such obstructions would create a @xmath92-pattern in @xmath46 .
example [ nonvex obst ] illustrates a non - vexillary permutation which has an element @xmath71 that is obstructed on both sides .
equivalent characterizations of vexillarity concern the inversion set @xmath96 or the following objects .
the _ diagram _ of a permutation @xmath22 is @xmath97 \times [ 1,n]$ ] where @xmath98 the _ code _ of @xmath22 is the vector @xmath99 where @xmath100 is the number of elements in row @xmath16 of @xmath96 .
the shape @xmath43 is the partition formed by writing the entries of the code in non - increasing order .
[ vex defs ] the following are equivalent definitions of vexillarity for a permutation @xmath22 : 1 .
[ vex1 ] @xmath22 is @xmath92-avoiding ; 2 .
[ vex2 ] the set of rows of @xmath96 is totally ordered by inclusion ; 3 .
[ vex3 ] the set of columns of @xmath96 is totally ordered by inclusion ; 4 .
[ vex4 ] the set of rows of @xmath101
is totally ordered by inclusion ; 5 .
[ vex5 ] the set of columns of @xmath101 is totally ordered by inclusion ; 6 .
[ vex6 ] @xmath102 , where @xmath103 is the transpose of @xmath43 .
see @xcite .
this section proves a new characterization of vexillary permutations , quite different from those in proposition [ vex defs ] . a partial ordering
can be placed on the set of all permutations @xmath104 , where @xmath105 if @xmath106 contains the pattern @xmath107 .
definition [ 2143vex ] determines vexillarity by a condition on the principal order ideal of a permutation .
the new characterization , theorem [ vexthm ] , depends on a particular condition holding for the principal _ dual _ order ideal .
let @xmath108 be a reduced decomposition of @xmath24 . for @xmath109 , the _ shift of @xmath110 by @xmath111
_ is @xmath112 [ vexthm ] the permutation @xmath46 is vexillary if and only if , for every permutation @xmath22 containing a @xmath46-pattern , there exists a reduced decomposition @xmath113 containing some shift of an element @xmath114 as a factor .
first suppose that @xmath70 is vexillary .
let @xmath5 contain a @xmath46-pattern .
assume for the moment that there is a @xmath115 such that 1 .
@xmath116 ; 2 .
@xmath117 has a @xmath46-pattern in positions @xmath118 for some @xmath119 $ ] .
choose a reduced decomposition @xmath114 .
let @xmath120 be the permutation obtained from @xmath117 by placing the values @xmath121 in increasing order and leaving all other entries unchanged .
choose any @xmath122 .
then @xmath123 it remains only to find a @xmath124 satisfying ( r@xmath10 ) and ( r@xmath12 ) .
this will be done by an algorithm that takes as input a permutation @xmath5 containing a @xmath46-pattern and outputs the desired permutation @xmath124 .
because the details of this algorithm can be cumbersome , a brief description precedes each of the major steps . 1 . + set @xmath125 } : = w$ ] and @xmath126 .
2 . [ alg halt ]
+ if @xmath127}$ ] has no entries inside the pattern , then . otherwise , choose @xmath128}$ ] inside the pattern .
[ alg big ] + if @xmath128 } > \langle k \rangle$ ] , then 1 .
let @xmath129})= \{y \ge x_{[i ] } : y \text { is inside the pattern}\}$ ] .
2 . consider the elements of @xmath129})$ ] in decreasing order .
multiply @xmath127}$ ] on the right by adjacent transpositions ( changing _ positions _ in the one - line notation ) to move each element immediately to the right of @xmath67 .
3 . let @xmath130}$ ] be the resulting permutation .
set @xmath131 and .
4 . [ alg small ] + if @xmath128 } < \langle 1 \rangle$ ] , then 1 .
let @xmath132 } ) = \{y \le x_{[i ] } : y \text { is inside the pattern}\}$ ] .
2 . consider the elements of @xmath132})$ ] in increasing order .
multiply @xmath127}$ ] on the right by adjacent transpositions to move each element immediately to the left of @xmath67 .
3 . let @xmath130}$ ] be the resulting permutation .
set @xmath131 and .
[ alg bound ] + let @xmath73 $ ] be the unique value such that @xmath133 } < \langle m+1 \rangle$ ] . 6 .
[ alg between ] + if the values @xmath134 } , \langle m+1 \rangle\}$ ] appear in increasing order in @xmath127}$ ] , then define @xmath135 and @xmath136 as in definition [ obstruction ] and 1 .
[ alg unob right ] if @xmath128}$ ] is unobstructed to the right , then 1 .
let @xmath137})$ ] be the set of non - pattern entries at least as large as @xmath128}$ ] and lying between @xmath128}$ ] and @xmath78 in the one - line notation of @xmath127}$ ] .
2 . consider the elements of @xmath137})$ ] in decreasing order .
for each @xmath138})$ ] , multiply on the right by adjacent transpositions until @xmath139 is immediately to the right of @xmath78 , or the right neighbor of @xmath139 is @xmath140 . in the latter case ,
@xmath141 for some @xmath142 $ ] because all larger non - pattern entries are already to the right of @xmath78 .
interchange the roles of @xmath139 and @xmath143 , and move this new @xmath139 to the right in the same manner , until it is to the right of ( the redefined ) @xmath78 .
3 . let @xmath130}$ ] be the resulting permutation , with @xmath67 redefined as indicated .
let @xmath144}$ ] be the non - pattern entry in the final move after any interchange of roles .
this is greater than @xmath128}$ ] and the newly redefined @xmath78 , and occurs to the right of the new @xmath78 . if @xmath144}$ ] is outside of the pattern , with @xmath131 . otherwise with @xmath131 .
[ alg unob left ] the entry @xmath128}$ ] is unobstructed to the left ( proposition [ vex not obst ] ) .
1 . let @xmath145})$ ] be the set of non - pattern entries at most as large as @xmath128}$ ] and lying between @xmath77 and @xmath128}$ ] in the one - line notation of @xmath127}$ ] .
2 . consider the elements of @xmath145})$ ] in increasing order .
for each @xmath146})$ ] , multiply on the right by adjacent transpositions until @xmath139 is immediately to the left of @xmath77 , or the left neighbor of @xmath139 is @xmath147 . in the latter case ,
@xmath148 for some @xmath149 $ ] because all smaller non - pattern entries are already to the left of @xmath77 .
interchange the roles of @xmath139 and @xmath150 , and move this new @xmath139 to the left in the same manner , until it is to the left of ( the redefined ) @xmath77 .
3 . let @xmath130}$ ] be the resulting permutation , with @xmath67 redefined as indicated .
let @xmath144}$ ] be the non - pattern entry in the final move after any interchange of roles . this is less than @xmath128}$ ] and the newly redefined @xmath77 , and occurs to the left of the new @xmath77 . if @xmath144}$ ] is outside of the pattern , .
otherwise with @xmath131 .
[ alg big left ] + if @xmath127}(s ) = \langle m+1 \rangle$ ] and @xmath127}(t ) = x_{[i]}$ ] with @xmath151 , multiply @xmath127}$ ] on the left by adjacent transpositions ( changing _ values _ in the one - line notation ) to obtain @xmath130}$ ] with the values @xmath152 } , \langle m+1 \rangle]$ ] in increasing order .
then @xmath130}(s)$ ] is in the half - open interval @xmath152 } , \langle m+1 \rangle)$ ] , and @xmath130}(t)$ ] is in the half - open interval @xmath153 } , \langle m+1 \rangle]$ ] . with @xmath144}:=w_{[i+1]}(t)$ ] , the pattern redefined so that @xmath154}(s)$ ] , and @xmath131 .
[ alg small right ] + if @xmath127}(s ) = \langle m \rangle$ ] and @xmath127}(t ) = x_{[i]}$ ] with @xmath155 , multiply @xmath127}$ ] on the left by adjacent transpositions to obtain @xmath130}$ ] with the values @xmath156}]$ ] in increasing order .
then @xmath130}(s)$ ] is in the half - open interval @xmath157}]$ ] , and @xmath130}(t)$ ] is in the half - open interval @xmath156})$ ] . with @xmath144}:=w_{[i+1]}(t)$ ] , the pattern redefined so that @xmath158}(s)$ ] , and @xmath131 .
each subsequent visit to step [ alg halt ] involves a permutation with strictly fewer entries inside the pattern than on the previous visit .
each multiplication by an adjacent transposition indicated in the algorithm removes an inversion , and so decreases the length of the permutation .
this is crucial because of requirement ( r@xmath10 ) .
consider the progression of : * step [ alg halt ] @xmath159 or begin a pass through ; * step [ alg big ] @xmath159 step [ alg halt ] ; * step [ alg small ] @xmath159 step [ alg halt ] ; * step [ alg unob right ] @xmath159 steps [ alg big ] or [ alg big left ] ; * step [ alg unob left ] @xmath159 steps [ alg small ] or [ alg small right ] ; * step [ alg big left ] @xmath159 steps [ alg big ] , [ alg between ] , or [ alg big left ] ; * step [ alg small right ] @xmath159 steps [ alg small ] , [ alg between ] , or [ alg small right ] .
step [ alg unob right ] concludes with @xmath144}$ ] to the left of its lower pattern bound , and smaller pattern elements lying between @xmath144}$ ] and this bound .
therefore , no matter how often step [ alg big left ] is next called , the algorithm will never subsequently go to step [ alg unob left ] before going to step [ alg halt ] . likewise
, a visit to step [ alg unob left ] means that step [ alg unob right ] can never be visited until step [ alg halt ] is visited and a new entry inside the pattern is chosen .
steps [ alg big ] and [ alg small ] do not change the relative positions of @xmath67 .
steps [ alg unob right ] and [ alg big left ] imply @xmath144 } > x_{[i]}$ ] , while @xmath144 } < x_{[i]}$ ] after steps [ alg unob left ] and [ alg small right ] .
let @xmath160 be as in step [ alg bound ] .
until revisiting step [ alg halt ] , the values @xmath161 , for @xmath162 , do not increase if @xmath144 } > x_{[i]}$ ] . nor do the values @xmath161 , for @xmath163 , decrease if @xmath144 } < x_{[i]}$ ] .
the other pattern values are unchanged .
the definition of @xmath160 means that the reordering of values in steps [ alg big left ] and [ alg small right ] does not change the positions in which the pattern @xmath46 occurs .
additionally , these steps change the value of the entry inside the pattern ( that is , @xmath144 } \neq x_{[i]}$ ] ) , but not its position .
these observations indicate not only that terminates , but that it outputs @xmath124 as in equation satisfying ( r1 ) and ( r2 ) .
this completes one direction of the proof .
now suppose @xmath70 is not vexillary .
there is an occurrence @xmath164 such that @xmath165 define @xmath166 to be the index such that @xmath167 .
define @xmath168 by @xmath169 for example , if @xmath87 , then @xmath86 .
if there is a reduced decomposition @xmath113 such that @xmath170 for @xmath114 and @xmath109 , then there is a @xmath171 as in equation satisfying ( r1 ) and ( r2 ) . keeping the values @xmath172 , @xmath173 , @xmath174 , and @xmath175
as defined above , the permutation @xmath22 was constructed so that @xmath176 one of the values in the consecutive subsequence @xmath177 must move to get a consecutive @xmath46-pattern in @xmath117 .
however , the values @xmath178 appear in increasing order in @xmath22 , and the consecutive subsequence @xmath179 in @xmath22 is increasing .
therefore , there is no way to multiply @xmath22 by adjacent transpositions , always eliminating an inversion , to obtain a consecutive @xmath46-pattern .
hence , if @xmath46 is not vexillary then there exists a permutation @xmath22 containing a @xmath46-pattern such that no reduced decomposition of @xmath22 contains a shift of a reduced decomposition of @xmath46 as a factor .
[ vex ex ] if @xmath180 and @xmath181 , with the chosen occurrence @xmath67 in bold , the algorithm may proceed as follows .
* @xmath125 } : = { \mbox{\boldmath $ 3$}}14{\mbox{\boldmath $ 6$}}5{\mbox{\boldmath $ 2$}}$ ] . * step [ alg halt ] : @xmath182 } : = 1 $ ] . * step [ alg small ] : @xmath125 } \mapsto w_{[0 ] } s_1 = 1{\mbox{\boldmath $ 3$}}4{\mbox{\boldmath $ 6$}}5{\mbox{\boldmath $ 2 $ } } = : w_{[1]}$ ] . *
step [ alg halt ] : @xmath183 } : = 5 $ ] .
* step [ alg big left ] : @xmath184 } \mapsto s_5 w_{[1 ] } = 1{\mbox{\boldmath $ 3$}}4{\mbox{\boldmath $ 5$}}6{\mbox{\boldmath $ 2 $ } } = : w_{[2]}$ ] ; @xmath185 } : = 6 $ ] . * step [ alg big ] : @xmath186 } \mapsto w_{[2 ] } s_5 = 1{\mbox{\boldmath $ 3$}}4{\mbox{\boldmath
$ 52$}}6 = : w_{[3]}$ ] . * step [ alg halt ] : @xmath187 } : = 4 $ ] .
* step [ alg unob right ] : @xmath188 } \mapsto w_{[3 ] } = 1{\mbox{\boldmath $ 34$}}5{\mbox{\boldmath $ 2$}}6 = : w_{[4]}$ ] ; @xmath189 } : = 5 $ ] . *
step [ alg big ] : @xmath190 } \mapsto w_{[4]}s_4 = 1{\mbox{\boldmath $ 342$}}56 = : w_{[5]}$ ] . *
step [ alg halt ] : output @xmath191 .
therefore @xmath192 , and @xmath193 . keeping the notation of equation , @xmath194 and @xmath195 . the unique reduced decomposition of @xmath196
is @xmath197 , and indeed @xmath198 let @xmath86 and @xmath87 .
no element of @xmath199 contains a shift of any element of @xmath200 as a factor .
suppose that @xmath113 contains a shift of @xmath114 as a factor , @xmath201 then @xmath114 can be replaced by any @xmath202 in equation .
some care must be taken regarding factors in reduced decompositions .
this is clarified in the following definition and lemma , the proof of which is straightforward .
let @xmath5 and @xmath203 .
write @xmath204 , where @xmath205 and @xmath206 .
suppose that @xmath207 contains only letters in @xmath208 .
if no element of @xmath209 has an element of @xmath210 as its rightmost character and no element of @xmath211 has an element of @xmath210 as its leftmost character , then @xmath207 is _ isolated _ in @xmath110 .
equivalently , the values @xmath118 must appear in increasing order in @xmath106 , and the positions @xmath118 must comprise an increasing sequence in @xmath107 .
if @xmath212 and a shift of @xmath207 appears as a factor in a reduced decomposition of some permutation , then @xmath207 is necessarily isolated .
this is because @xmath207 has maximal reduced length in the letters @xmath213 , so any factor of length greater than @xmath214 in the letters @xmath215 is not reduced .
[ isolated ] if a reduced decomposition of @xmath22 contains an isolated shift of a reduced decomposition of @xmath46 , then @xmath22 contains the pattern @xmath46 .
the converse to lemma [ isolated ] holds if @xmath46 is vexillary .
the characterization of vexillary in theorem [ vexthm ] differs substantially from those in proposition [ vex defs ] .
there is not an obvious way to prove equivalence with any of the definitions ( v[vex2])-(v[vex6 ] ) , except via ( v[vex1 ] ) .
this raises the question of whether more may be understood about vexillary permutations ( or perhaps other types , such as grassmannian or dominant permutations ) by studying their reduced decompositions or the permutations that contain those in question as patterns .
theorem [ vexthm ] has a number of consequences , and will be used often in the subsequent sections of this paper .
most immediately , notice that it generalizes the result of billey , jockusch , and stanley mentioned earlier : @xmath0-avoiding permutations are exactly those whose reduced decompositions contain no long braid moves , and observe that @xmath216 .
recall the definition of short and long braid moves in a reduced decomposition , as well as the short and long braid relations described in equations and .
it is well known that any element of @xmath38 can be transformed into any other element of @xmath38 by successive applications of the braid relations . because the short braid relation represents the commutativity of particular pairs of adjacent transpositions , the following equivalence relation is known as the _ commutation relation_. [ commutation ] for a permutation @xmath22 and @xmath217 , write @xmath218 if @xmath110 can be obtained from @xmath219 by a sequence of short braid moves .
let @xmath220 be the set of commutation classes of reduced decompositions of @xmath22 , as defined by @xmath221 .
the commutation classes of @xmath222 are @xmath223 , @xmath224 , and @xmath225 .
[ graph ] for a permutation @xmath22 , the graph @xmath226 has vertex set equal to @xmath220 , and two vertices share an edge if there exist representatives of the two classes that differ by a long braid move .
elnitsky gives a very elegant representation of this graph in @xcite , which will be discussed in depth in section [ section polygon ] .
a consequence of his description , although not difficult to prove independent of his work , is the following .
[ graph facts ] the graph @xmath226 is connected and bipartite .
see @xcite . despite proposition [ graph facts ] , much remains to be understood about the graph @xmath226 . for example , even the size of the graph for @xmath227 ( that is , the number of commutation classes for the longest element ) is unknown .
billey , jockusch , and stanley characterize all permutations with a single commutation class , and hence whose graphs are a single vertex , as @xmath0-avoiding permutations . a logical question to ask next is : for what permutations does each reduced decomposition contain at most one long braid move ?
more restrictively : what if this long braid move is required to be a specific shift of @xmath228 or @xmath229 ?
moreover , what are the graphs in these cases ?
let @xmath230 .
[ 1 lbm ] @xmath231 is the set of permutations such that every @xmath0-pattern in @xmath22 has the same maximal element and the same minimal element .
assume @xmath22 has a @xmath0-pattern .
suppose that every occurrence of @xmath0 in @xmath22 has @xmath232 and @xmath233 .
suppose that @xmath113 has at least one long braid move .
choose @xmath234 so that @xmath235 is the first such .
each adjacent transposition in a reduced decomposition increases the length of the product . then by the supposition ,
@xmath236 is @xmath0-avoiding , so @xmath237 has no long braid moves .
it remains only to consider when @xmath238 is also a long braid move .
the only possible reduced configurations for such a factor @xmath239 are shifts of @xmath240 and @xmath241 . if either of these is not isolated , then it is part of a shift of @xmath242 , @xmath243 , @xmath244 , or @xmath245 .
notice that * @xmath246 ; * @xmath247 ; * @xmath248 .
if @xmath239 is isolated in @xmath219 then @xmath22 contains a @xmath249- or @xmath250-pattern by lemma [ isolated ] .
otherwise , @xmath22 contains a @xmath251-pattern . however , every @xmath0-pattern in @xmath22 has @xmath232 and @xmath233 .
therefore @xmath252 is the only long braid move in @xmath219 , so @xmath253 .
now let @xmath22 be an element of @xmath231 . if @xmath22 has two @xmath0-patterns that do not have the same maximal element and the same minimal element , then they intersect at most once or they create a @xmath251- , @xmath249- , or @xmath250-pattern .
these three patterns are vexillary .
thus by theorem [ vexthm ] and the examples above , containing one of these patterns would imply that some element of @xmath38 has more than one long braid move . if the two @xmath0-patterns intersect at most once , their union may be a non - vexillary pattern , so theorem [ vexthm ] does not necessarily apply .
however , a case analysis shows that it is possible to shorten @xmath22 by adjacent transpositions and make one @xmath0-pattern increasing ( via a long braid move ) without destroying the other @xmath0-pattern .
thus an element of @xmath38 would have more than one long braid move , contradicting @xmath253 .
let @xmath254 consist of permutations with some @xmath0-pattern , where every long braid move that occurs must be @xmath255 or @xmath256 .
[ 1 ulbm ] @xmath257 has a unique @xmath0-pattern and @xmath258 . if @xmath22 has a unique @xmath0-pattern , then @xmath259 .
a unique @xmath0-pattern implies that @xmath260 all appear to the left of @xmath173 in @xmath261 , and @xmath262 all appear to the right of @xmath173 , so the second statement follows .
consider the long braid moves that may appear for elements of @xmath263 .
let @xmath253 have @xmath234 distinct @xmath0-patterns . by theorem
[ 1 lbm ] , these form a pattern @xmath264 in @xmath22 .
the permutation @xmath46 is vexillary , so there exists @xmath109 and a reduced decomposition @xmath265 for each @xmath114 .
there are elements in @xmath266 with long braid moves @xmath267 for each @xmath268 $ ] .
for example , @xmath269 .
therefore , if @xmath270 , then @xmath271 , so @xmath22 has a unique @xmath0-pattern .
suppose that @xmath22 has a unique @xmath0-pattern . because @xmath259 , the only possible long braid moves in reduced decompositions of @xmath22 are @xmath272 or @xmath273 . [ u_n graph ] if @xmath253 and @xmath22 has @xmath234 distinct @xmath0-patterns , then @xmath274 and the graph @xmath226 is a path of @xmath275 vertices connected by @xmath234 edges . because @xmath22 contains the pattern @xmath264
, there is a subgraph of @xmath226 that is a path of @xmath275 vertices connected by @xmath234 edges . since @xmath46 accounts for all of the @xmath0-patterns in @xmath22 , this is all of @xmath226 .
[ u_n(j ) graph ] if @xmath270 , then @xmath276 and the graph @xmath226 is a pair of vertices connected by an edge .
in his doctoral thesis and in @xcite , elnitsky developed a bijection between commutation classes of reduced decompositions of @xmath5 and rhombic tilings of a particular @xmath277-gon @xmath278 .
this bijection leads to a number of interesting questions about tilings of @xmath278 and their relations to the permutation @xmath22 itself .
a number of these ideas are studied in this and the following section .
[ polygon ] for @xmath5 , let @xmath278 be the @xmath277-gon with all sides of unit length such that 1 .
sides of @xmath278 are labeled @xmath279 in order ; 2 . the portion labeled @xmath280 is convex ; and 3 .
sides with the same label are parallel .
orient the polygon so that the edge labeled @xmath10 lies to the left of the top vertex and the edge labeled @xmath281 lies to its right .
polygon_. for @xmath282 , the polygon @xmath283 is a centrally symmetric @xmath277-gon .
the hexagon @xmath284 can be tiled by rhombi with sides of unit length in exactly two ways .
each of these tilings is the _ flip _ of the other .
let @xmath285 be the set of tilings of @xmath278 by rhombi with sides of unit length .
define a graph @xmath286 with vertex set @xmath285 , and connect two tilings by an edge if they differ by a flip of the tiling of a single sub - hexagon .
unless otherwise indicated , the term _ tiling _ refers to an element of @xmath285 .
[ elthm ] the graphs @xmath226 and @xmath286 are isomorphic .
henceforth , both graphs will be denoted @xmath226 . before discussing new results related to this polygon
, it is important to understand elnitsky s bijection , outlined in the following algorithm .
a more thorough treatment appears in @xcite . 1
. set the polygon @xmath287 } : = x(w)$ ] , the string @xmath288}$ } } : = \emptyset$ ] , and @xmath289 .
[ elnstart ] if @xmath290}$ ] has no area , then . 3 .
[ elnchoice ] there is at least one tile @xmath291 that shares two edges with the right side of @xmath290}$ ] .
if @xmath291 includes the @xmath292 and @xmath293 edges from the top along the right side of @xmath290}$ ] , set @xmath294}$ } } : = j{\mbox{\boldmath $ j_{[i]}$}}$ ] .
5 . let @xmath295}$ ] be @xmath290}$ ] with the tile @xmath291 removed .
set @xmath131 and .
yields the entire commutation class because of the choice of tile in step [ elnchoice ] .
the tiling in figure [ tilingfig ] corresponds to the equivalence class consisting solely of the reduced decomposition @xmath296 .
[ subgraph ] if @xmath46 is vexillary and @xmath22 contains a @xmath46-pattern , then @xmath297 is a subgraph of @xmath226 .
this follows from theorems [ vexthm ] and [ elthm ] .
elnitsky s correspondence , described in , combined with theorem [ 1 lbm ] and corollary [ 1 ulbm ] , indicates that any tiling of @xmath278 for @xmath253 has at most one sub - hexagon ( every tiling has exactly one sub - hexagon if @xmath22 is not @xmath0-avoiding ) .
moreover , the sub - hexagon has the same vertical position for all elements of @xmath254 . under certain circumstances ,
the polygon @xmath278 for @xmath5 can be rotated or reflected to give a polygon @xmath298 for another @xmath299 .
[ reflect ] let @xmath24 and @xmath300 . then @xmath301 and @xmath302 .
[ rotate ] let @xmath24 .
if @xmath303 , then @xmath304and @xmath305 where @xmath306 and all entries are modulo @xmath4 . likewise ,
if @xmath307 , then @xmath308 and @xmath309 where @xmath310 and all entries are modulo @xmath4 .
elnitsky s result interprets the commutation classes of @xmath38 as rhombic tilings of @xmath278 , with long braid moves represented by flipping sub - hexagons .
the following theorem utilizes this interpretation , and demonstrates that the number of commutation classes of a permutation is monotonically increasing with respect to pattern containment , thus generalizing one aspect of corollary [ subgraph ] .
note that @xmath46 is not required to be vexillary in theorem [ monotone ] , unlike in theorem [ vexthm ] .
[ monotone ] if @xmath22 contains the pattern @xmath46 , then @xmath311 .
consider a tiling @xmath312 . this represents a commutation class of @xmath266 .
for an ordering of the tiles in @xmath313 as defined by , label the tile @xmath314 by @xmath315 , the tile @xmath316 by @xmath317 , and so on .
if the tile with label @xmath318 corresponds to the adjacent transposition @xmath319 , then @xmath320 . 1 .
set @xmath125}:=w$ ] , @xmath321}:=p$ ] , @xmath322}:=t$ ] , @xmath323}:=\emptyset$ ] , and @xmath289 .
[ monostart ] if @xmath324}$ ] is the identity permutation , then define @xmath325}$ ] to be the tiles of @xmath326}$ ] together with any tiling of @xmath327})$ ] . .
3 . let @xmath328}$ ] be such that the tile labeled @xmath329 includes edges @xmath324}(j_{[i]})$ ] and @xmath324}(j_{[i]}+1)$ ] .
note that @xmath324}(j_{[i ] } ) > p_{[i]}(j_{[i]}+1)$ ] .
4 . define @xmath330 so that @xmath127}(r ) = \langle p_{[i]}(j_{[i ] } ) \rangle$ ] and @xmath127}(s ) = \langle p_{[i]}(j_{[i]}+1 ) \rangle$ ] . note that @xmath127}(t)$ ] is a non - pattern entry for @xmath331 .
5 . let @xmath332}$ ] be the permutation defined by @xmath333}(t ) \mapsto \left\ { \begin{array } { c@{\quad:\quad}l } w_{[i]}(t ) & t < r \text { or } t >
s\\ \widetilde{w}_{[i]}(t ) & r \le t \le s \end{array } \right.\ ] ] where @xmath334}(r ) , \ldots , \widetilde{w}_{[i]}(s))$ ] is @xmath335}(r ) , \ldots , w_{[i]}(s)\}$ ] in increasing order .
set @xmath130 } : = w_{[i]}v_{[i]}$ ] , and notice that @xmath336 } ) = \ell(w_{[i ] } ) - \ell(v_{[i]})$ ] . 7 .
[ monochoice ] the right boundaries of @xmath337})$ ] and @xmath327})$ ] differ only in the @xmath338 edges , and the left side of this difference ( part of the boundary of @xmath337})$ ] ) is convex .
therefore , this difference has a rhombic tiling @xmath339}$ ] .
define @xmath325}$ ] to be the tiles in @xmath326}$ ] together with the tiles in @xmath339}$ ] .
set @xmath131 and .
the algorithm takes a tiling @xmath312 and outputs one of possibly several tilings @xmath340 due to the choice in steps [ monostart ] and [ monochoice ] .
a tiling @xmath340 so obtained can only come from this @xmath313 , although possibly with more than one labeling of the tiles . however , this labeling of the tiles merely reflects the choice of a representative from the commutation class , so indeed @xmath341 , and @xmath311 .
let @xmath342 and @xmath343 .
the pattern @xmath46 occurs in @xmath22 as @xmath344 .
figure [ mono ex ] depicts the output of , given the two tilings of @xmath345 .
@xmath159 + @xmath159
elnitsky s bijection considers the rhombic tilings of the polygon @xmath278 .
rhombi are a special case of a more general class of objects known as zonotopes .
a polytope is a _
@xmath346-zonotope _ if it is the projection of a regular @xmath4-cube onto a @xmath346-dimensional subspace .
centrally symmetric convex polygons are exactly the @xmath12-zonotopes .
these necessarily have an even number of sides .
a _ zonotopal tiling _ of a polygon is a tiling by centrally symmetric convex polygons .
let @xmath347 be the set of zonotopal tilings of elnitsky s polygon .
rhombi are centrally symmetric , so @xmath348 .
[ 2k - gon ] there is a tiling in @xmath347 containing a @xmath349-gon with sides parallel to the sides labeled @xmath47 if and only if @xmath350 is an occurrence of @xmath351 in @xmath22 .
since the tiles are convex , a @xmath349-gon in the tiling with sides as described has right side labeled @xmath352 from top to bottom and left side labeled @xmath353 from top to bottom .
therefore elnitsky s bijection shows that this tile ( or rather , any decomposition of it into rhombi ) transforms the sequence @xmath354 into @xmath355 .
reduced decompositions have minimal length , so no inversions can be `` undone '' by subsequent adjacent transpositions .
therefore @xmath350 must be an occurrence of @xmath351 in @xmath22 .
conversely , suppose that @xmath350 is an occurrence of the vexillary pattern @xmath351 in @xmath22 .
for a decreasing pattern , the algorithm can be modified slightly to produce @xmath117 as in equation , where the consecutive occurrence @xmath356 is @xmath350 .
let @xmath357 and @xmath358 for @xmath122 .
removing the rhombi that correspond to @xmath359 yields the polygon @xmath360 , and the rhombi that correspond to @xmath361 form a sub-@xmath349-gon with sides parallel to the sides labeled @xmath52 in @xmath278 .
less specifically , theorem [ 2k - gon ] states that a tiling in @xmath347 can contain a @xmath349-gon if and only if @xmath22 has a decreasing subsequence of length @xmath234 .
using a group theoretic argument , pasechnik and shapiro showed in @xcite that no element of @xmath362 consists entirely of hexagons for @xmath363 .
their result states that at least one rhombus must be present in a hexagonal / rhombic tiling .
kelly and rottenberg had previously obtained a better bound in terms of arrangements of pseudolines in @xcite . working with reduced decompositions and elnitsky s polygons yields a different proof that no element of @xmath362 can consist of entirely hexagonal tiles for @xmath363 , and generalizes the result to other types of tiles .
theorem [ 2k tiles ] is , in a sense , a counterpart to theorem [ 2k - gon ] .
[ 2k tiles ] let @xmath227 be the longest element in @xmath3 .
there is a tiling @xmath364 consisting entirely of @xmath349-gons if and only if one of the following is true : 1 .
@xmath365 ; or 2 .
@xmath366 .
if @xmath367 , the result holds for all @xmath4 : every @xmath283 can be tiled by rhombi with unit side length . for
the remainder of the proof , assume that @xmath368 .
suppose that there exists @xmath364 consisting entirely of @xmath349-gons .
then there exists @xmath369 , for some @xmath370 .
because @xmath371 is the longest element in @xmath372 , the factor @xmath373 can have any of @xmath374 at either end .
thus , if @xmath375 for @xmath376 , then @xmath377 .
because @xmath378 $ ] , there is at most one shift by @xmath379 and at most one shift by @xmath380 .
in fact , there is exactly one of each of these shifts , since all of @xmath283 must be tiled and the shifts correspond to the vertical placement of the @xmath349-gons .
consider the tile corresponding to @xmath381 .
this @xmath349-gon is as high vertically as possible and it is the only tile placed so high .
thus it shares the top vertex and its incident sides with @xmath283 .
that is , two of the tile s sides are labeled @xmath10 and @xmath4 .
similarly , the tile for @xmath382 also has two sides labeled @xmath10 and @xmath4 . by theorem [ 2k - gon ]
, each of these tiles corresponds to a @xmath371-pattern with @xmath383 and @xmath61 .
however , once the value @xmath10 is to the right of the value @xmath4 , it remains to the right as adjacent positions are transposed to lengthen the permutation .
therefore the two patterns , the tiles , and their shifts must be equal : @xmath384 .
indeed , there is always a tiling @xmath364 consisting of a single @xmath277-gon .
there is a poset @xmath385 that arises naturally when studying @xmath347 . for a permutation @xmath22 ,
let the poset @xmath385 have elements equal to the zonotopal tilings @xmath347 , partially ordered by reverse edge inclusion .
in the poset @xmath386 , the tiling in figure [ tilingfig ] is less than the tiling in figure [ ztilingfig ] .
[ p(w_0 ) max ] for the longest element @xmath282 , the poset @xmath387 has a maximal element equal to the tiling in @xmath388 that consists of a single @xmath277-gon .
[ posetgraph ] the minimal elements of @xmath385 are the rhombic tilings , which are the vertices of the graph @xmath226 .
moreover , edges in the graph @xmath226 correspond to flipping a single sub - hexagon in the tiling .
therefore these edges correspond to the elements of @xmath385 that cover the minimal elements .
the relationships in remark [ posetgraph ] are immediately apparent .
another relationship is not as obvious .
this follows from a result of boris shapiro , michael shapiro , and alek vainshtein in @xcite .
[ ssvlem ] the set of all @xmath11- and @xmath389-cycles in @xmath226 form a system of generators for the first homology group @xmath390 .
additionally , anders bjrner noted that gluing @xmath12-cells into those @xmath11- and @xmath389-cycles yields a simply connected complex ( @xcite ) . in @xcite ,
lemma [ ssvlem ] is stated only for @xmath391 .
however , the proof easily generalizes to all @xmath5 .
a straightforward argument demonstrates that a @xmath11-cycle in @xmath226 corresponds to @xmath392 with rhombi and two hexagons , and an @xmath389-cycle corresponds to @xmath392 with rhombi and an octagon .
these are exactly the elements of @xmath385 which cover those that correspond to edges of @xmath226 .
[ level2 ] the elements of @xmath385 that cover the elements ( corresponding to edges of @xmath226 ) covering the minimal elements ( corresponding to vertices of @xmath226 ) correspond to a system of generators for the first homology group @xmath390 .
little is known about the structure of the graph @xmath226 for arbitrary @xmath22 . however , in some cases a description can be given via theorem [ 2k - gon ] and lemma [ ssvlem ] .
[ chainthm ] the following statements are equivalent for a permutation @xmath22 : 1 .
[ treecon ] @xmath226 is a tree ; 2 .
[ chaincon ] @xmath226 is a path ( that is , no vertex has more than two incident edges ) ; 3 .
[ posetcon ] the maximal elements of @xmath385 cover the minimal elements ; 4 .
[ patterncon ] @xmath22 is @xmath251-avoiding and any two @xmath0-patterns intersect at least twice .
( [ treecon ] ) @xmath393 ( [ posetcon ] ) by corollary [ level2 ] . from theorem
[ 2k - gon ] and the discussion preceding corollary [ level2 ] , an @xmath389-cycle in the graph is equivalent to having a @xmath251-pattern . similarly , a @xmath11-cycle is equivalent to two sub - hexagons whose intersection has zero area , so some reduced decomposition has two disjoint long braid moves .
this implies that two @xmath0-patterns intersect in at most one position
. therefore ( [ treecon ] ) @xmath393 ( [ patterncon ] ) .
finally , suppose that @xmath226 is a tree and a vertex has three incident edges .
the corresponding tiling has at least three sub - hexagons .
however , it is impossible for every pair of these to overlap
. this contradicts ( [ treecon ] ) @xmath393 ( [ posetcon ] ) @xmath393 ( [ patterncon ] ) , so ( [ treecon ] ) @xmath393 ( [ chaincon ] ) . if @xmath394 is the set of all @xmath5 for which @xmath226 is a path , then @xmath395 by corollaries [ u_n graph ] and [ u_n(j ) graph ] .
following convention , the unique maximal element in a poset , if it exists , is denoted @xmath396 .
remark [ p(w_0 ) max ] noted that the poset @xmath387 has a @xmath396 .
in fact , there are other @xmath22 for which @xmath385 has a @xmath396 , as described below .
[ max elt ] the poset @xmath385 has a @xmath396 if and only if @xmath22 is @xmath397- , @xmath249- , and @xmath250-avoiding . the definition of the poset @xmath385 and theorem [ 2k - gon ] indicate that @xmath385 has a @xmath396 if and only if the union of any two decreasing subsequences that intersect at least twice is itself a decreasing subsequence .
suppose there are decreasing subsequences in @xmath22 of lengths @xmath398 that intersect @xmath399 times , for @xmath400 .
let @xmath401 , and choose a @xmath275 element subsequence of @xmath402 that includes @xmath403 and one more element from each descending subsequence .
let @xmath404 be the resulting pattern .
no @xmath396 in @xmath385 is equivalent to there being subsequences so that @xmath405 for some @xmath406 $ ] .
there are two ways to place a @xmath349-gon in a zonotopal tiling of @xmath345 , but these overlapping @xmath349-gons do not both lie in any larger centrally symmetric polygon .
the permutation @xmath46 is always vexillary , so theorem [ vexthm ] implies that @xmath385 will not have a @xmath396 if @xmath22 contains such a @xmath46 .
therefore , considering the permutation @xmath46 for each possible @xmath407 , the poset @xmath385 has a @xmath396 if and only if @xmath22 is @xmath397- , @xmath249- , and @xmath250-avoiding .
the permutations for which @xmath385 has a @xmath396 have recently been enumerated by toufik mansour in @xcite .
although the graph @xmath226 and poset @xmath385 are not known in general , there is a class of permutations for which these objects can be completely described .
this paper concludes with a study of this special case . in @xcite and @xcite ,
green and losonczy introduce and study `` freely braided '' elements in simply laced coxeter groups . in the case of type @xmath408 , these are as follows . a permutation @xmath22 is _ freely braided _ if every pair of distinct @xmath0-patterns in @xmath22 intersects at most once .
equivalently , @xmath22 is freely braided if and only if @xmath22 is @xmath251- , @xmath397- , @xmath249- , and @xmath250-avoiding .
the poset of a freely braided permutation has a unique maximal element by theorem [ max elt ] .
the permutation @xmath409 is not freely braided because @xmath0 and @xmath410 are both occurrences of the pattern @xmath0 , and they intersect twice .
the permutation @xmath411 is freely braided .
mansour enumerates freely braided permutations in @xcite . in @xcite ,
green and losonczy show that a freely braided @xmath22 with @xmath234 distinct @xmath0-patterns has latexmath:[\[\label{fb size } any simply laced coxeter group , here stated only for type @xmath408 .
if a permutation @xmath22 is freely braided with @xmath234 distinct @xmath0-patterns , then there exists @xmath203 with @xmath234 disjoint long braid moves .
[ fb tiles ] this means that there is a tiling of @xmath278 with @xmath234 sub - hexagons , none of which overlap .
furthermore , equation implies that flipping any sequence of these sub - hexagons does not yield any new sub - hexagons .
hence _ every _ tiling of @xmath278 has exactly @xmath234 sub - hexagons , none of which overlap , and @xmath396 in @xmath385 corresponds to the zonotopal tiling with rhombi and @xmath234 hexagons . from remark [ fb tiles ] , the structures of the graph @xmath226 and the poset @xmath385 are clear for a freely braided permutation @xmath5 .
let @xmath22 be freely braided with @xmath234 distinct @xmath0-patterns .
the graph @xmath226 is the graph of the @xmath234-cube , and the poset @xmath385 is isomorphic to the face lattice of the @xmath234-cube without its minimal element .
the permutation @xmath413 is freely braided .
its three @xmath0-patterns are @xmath414 , @xmath415 , and @xmath416 .
figures [ 3cubefig ] and [ 3posetfig ] depict its graph and poset .
particular thanks are due to richard stanley for his continued guidance and for the suggestion to study reduced decompositions .
anders bjrner provided helpful advice and discussion , and together with richard stanley coordinated the semester on algebraic combinatorics at the institut mittag - leffler , during which much of the research for this paper occurred .
thanks are also owed to john stembridge for his referral to the work of elnitsky and green and losonczy .
finally , the thoughtful suggestions of two anonymous referees have been greatly appreciated . l. m. kelly and r. rottenberg , simple points in pseudoline arrangements , _ pacific j. math . _
* 40 * ( 1972 ) , 617 - 622 .
a. lascoux and m. p. schtzenberger , polynmes de schubert , _ c. r. acad .
paris , srie i _ * 294 * ( 1982 ) , 447 - 450 .
d. v. pasechnik and b. shapiro , in search of higher permutahedra , preprint . v. reiner , note on the expected number of yang - baxter moves applicable to reduced decompositions , _ europ .
j. combin .
_ * 26 * ( 2005 ) , 1019 - 1021 . | billey , jockusch , and stanley characterized @xmath0-avoiding permutations by a property of their reduced decompositions .
this paper generalizes that result with a detailed study of permutations via their reduced decompositions and the notion of pattern containment .
these techniques are used to prove a new characterization of vexillary permutations in terms of their principal dual order ideals in a particular poset .
additionally , the combined frameworks yield several new results about the commutation classes of a permutation . in particular ,
these describe structural aspects of the corresponding graph of the classes and the zonotopal tilings of a polygon defined by elnitsky that is associated with the permutation . |
the study of the properties of hadrons in hot and dense matter is receiving a lot of attention over the last years .
a special effort has been invested to understand the properties of antikaons in the medium , especially after the speculation of the possible existence of an antikaon condensed phase in neutron stars @xcite .
heavy - ion collisions at energies around 1 - 2 agev offer the possibility of studying hadrons under extreme conditions @xcite .
in particular , production and propagation of kaons and antikaons have been investigated with the kaon spectrometer ( kaos ) of the sis heavy - ion synchroton at gsi ( darmstadt ) .
one first observation in c@xmath3c and ni@xmath3ni collisions @xcite is that , as a function of the energy difference @xmath4-@xmath5 , where @xmath5 is the energy for the particle production ( 2.5 gev for @xmath6 via @xmath7 and 2.9 gev for @xmath1 via @xmath8 ) , the number of @xmath1 balanced the number of @xmath6 although in @xmath9 collisions the production cross - sections close to threshold are 2 - 3 orders of magnitude different .
this could be interpreted as an indication of an attractive @xmath1 optical potential in the medium , although a complementary explanation in terms of in - medium enhanced @xmath10 production has also been suggested @xcite .
it has also been observed that the @xmath0 ratio remains almost constant for c@xmath3c , ni@xmath3ni and au@xmath3au collisions for 1.5 agev @xcite .
this could indicate that the absorption of @xmath1 via @xmath11 is suppressed and/or an enhanced @xmath1 production is obtained because of an attractive @xmath1 optical potential .
finally , a centrality independence for the @xmath0 ratio has been noticed in au@xmath3au and pb@xmath3pb reactions at 1.5 agev @xcite .
a recent interpretation claims that this centrality independence is a consequence of the strong correlation between the @xmath6 and @xmath1 yields @xcite .
in fact , the centrality independence of the @xmath0 ratio has often been advocated as signalling the lack of in - medium effects in the framework of statistical models as the volume cancels out exactly in the ratio @xcite .
however , brown et al . introduced the concept of `` broad - band equilibration '' @xcite according to which the independence on centrality of the @xmath0 ratio can be explained including medium effects .
this paper is devoted to investigate the influence of dressing the antikaons on the @xmath0 ratio with particular emphasis of bringing new insight into the role of in - medium effects in heavy - ion collisions at gsi energies .
in this section we present a calculation of the @xmath0 ratio in the framework of the statistical model .
the basic hypothesis is to assume that the particle ratios in relativistic heavy - ion collisions can be described by two parameters , the baryonic chemical potential @xmath12 and the temperature @xmath13 @xcite .
the fact that the number of strange particles in the final state is small at gsi energies requires an exact treatment of strangeness conservation , while the baryonic and charge conservation laws can be satisfied on average .
therefore , using statistical mechanics , one obtains for the @xmath0 ratio @xcite @xmath14 where @xmath15 @xmath16 is the one - particle partition function for @xmath17 , and @xmath18 @xmath19 indicate the sum of one - particle partition functions for baryons ( mesons ) with @xmath20 .
it is interesting to note that the @xmath0 ratio is independent of the volume and , hence , the same in the canonical and grandcanonical scheme used for strangeness conservation in contrast to the @xmath1 and @xmath6 multiplicities alone @xcite .
our objective is to study how the in - medium modifications of the properties of the hadrons at finite temperature affect the value of the @xmath1/@xmath6 ratio , focusing our attention on the dressing of antikaons . for consistency with previous works ,
we prefer to compute the inverse ratio @xmath6/@xmath1 @xmath21 this expression is equivalent to eq .
( [ eq : largevolume ] ) but takes into account only the most relevant contributions . for balancing the number of @xmath6 , the main contribution in the @xmath22 sector comes from @xmath23 and @xmath24 hyperons and , in a smaller proportion , from @xmath1 mesons and @xmath25 resonances .
the number of @xmath1 is balanced by the presence of @xmath6 mesons .
then , the particles involved in the calculation should be dressed according to their properties in the hot and dense medium in which they are embedded . for the @xmath23 and @xmath24
, the partition function @xmath26 is built using a mean - field dispersion relation for the single - particle energies ( see ref .
@xcite for the dressing of @xmath23 and ref .
@xcite for @xmath24 ) .
the resonance @xmath25 is described by a breit - wigner shape @xmath27 with @xmath28 and @xmath29 .
finally , two different prescriptions for the single - particle energy of the antikaons have been used ( see fig.[fig : antikaon ] ) .
first , we use the on - shell or mean - field approximation for the @xmath1 potential in which the partition function reads @xmath30 being @xmath31 the @xmath1 single - particle potential in the brueckner - hartree - fock approach @xcite ( see l.h.s . of fig .
[ fig : antikaon ] ) .
the second approach incorporates the @xmath1 spectral density , @xmath32 using the s - wave component of the jlich interaction and adding the p - wave contributions as done in ref .
@xcite ( see r.h.s . of fig .
[ fig : antikaon ] ) .
this p - wave components come from the coupling of the @xmath1 meson to hyperon - hole states ( @xmath33 ) .
in this section we discuss the effects of dressing the @xmath1 mesons in hot and dense matter on the @xmath0 ratio using an experimental value of 0.031 @xmath34 0.005 as reported in @xcite for ni@xmath3ni collisions at 1.93 agev .
optical potential at zero momentum for different densities as a function of temperature .
right : the boltzmann factor ( dotted line ) and the @xmath1 spectral function , including s - wave ( dashed line ) or s- and p - wave ( solid line ) components of the @xmath35 interaction , as functions of the energy , for a momentum @xmath36 mev at saturation density and temperature @xmath37 mev . , title="fig:",scaledwidth=49.5%,scaledwidth=53.0% ] optical potential at zero momentum for different densities as a function of temperature .
right : the boltzmann factor ( dotted line ) and the @xmath1 spectral function , including s - wave ( dashed line ) or s- and p - wave ( solid line ) components of the @xmath35 interaction , as functions of the energy , for a momentum @xmath36 mev at saturation density and temperature @xmath37 mev .
, title="fig:",scaledwidth=49.5%,scaledwidth=53.0% ] the inverse ratio , @xmath38 , is shown in fig .
[ fig : ratio - dens ] as a function of density at two given temperatures ( @xmath39 mev and @xmath40 mev ) using different approaches for the dressing of the @xmath1 meson : free gas ( dot - dashed lines ) , the on - shell approach ( dotted lines ) and using the @xmath1 spectral density including s - waves ( long - dashed lines ) or both s- and p - waves ( solid lines ) . since the baryonic chemical potential @xmath12 grows with density , we note that the ratio grows with @xmath41 with increasing density .
the curves representing the @xmath38 ratio tend to bend down after the initial increase when the in - medium @xmath1 properties are included .
this effect is particularly notorious when the s- and p - wave contributions of the @xmath1 self - energy are taken into account in the spectral density . actually , the low energy components of the @xmath1 spectral density related to @xmath33 excitations are responsible for this behaviour ( see the overlap of the boltzmann factor with the low energy region of the @xmath1 spectral density in fig . [
fig : antikaon ] ) .
these results are in qualitative agreement with the `` broad - band equilibration '' notion introduced by brown et al.@xcite .
however , this behaviour was found using a mean - field model , through a compensation of the increased attraction of the mean - field @xmath1 potential with the increase in the baryon chemical potential as density grows .
in contrast , our mean - field approach does not achieve this `` broad - band '' behaviour . for @xmath42 within different approaches .
right : @xmath43 for different values of the @xmath38 ratio using the full @xmath1 spectral density.,title="fig:",scaledwidth=49.5%,scaledwidth=55.0% ] for @xmath42 within different approaches .
right : @xmath43 for different values of the @xmath38 ratio using the full @xmath1 spectral density.,title="fig:",scaledwidth=49.5%,scaledwidth=55.0% ] in the framework of the statistical model , one obtains a relation between the temperature and the baryonic chemical potential by fixing the value of the @xmath0 ratio .
the l.h.s of fig .
[ fig : t - m ] shows the values of temperature and chemical potential compatible with a value of the inverse ratio @xmath38 of around 30 for the approaches discussed above .
the dot - dashed line stands for a free gas , similar to the calculations of ref .
the on - shell approach ( dotted line ) does not represent the broad - band effect , as already mentioned .
but due to the enhanced attraction felt by the @xmath1 mesons for higher densities , the chemical potential @xmath12 compatible with the value of the experimental ratio measured also increases for a given temperature .
when the @xmath1 spectral density containing s - wave components is used ( dashed line ) , two possible solutions that are compatible with the ratio emerge .
finally , a band of chemical potentials @xmath12 up to 850 mev at a temperature of @xmath44 mev appears , when both , s- and p - wave contributions are considered ( solid line ) . however , in the latter case , the temperature is too low to be compatible with the measured temperature and the corresponding freeze - out densities are too small ( up to @xmath45 only ) , so we can hardly speak of a `` broad band '' feature in the sense of that of brown et al . in the r.h.s . of fig .
[ fig : t - m ] we represent the temperature and chemical potential for different values of the ratio when the full @xmath1 spectral density is used .
we observe that the ratio is substantially lower at the more plausible temperature t @xmath46 70 mev , being more likely around 15 for a large region of baryonic chemical potential .
note that this reduced ratio translates into an overall enhanced production of @xmath1 by a factor of 2 compared to the experimental value .
this effect is a consequence of the additional strength of the @xmath1 spectral density at low energies .
the boltzman factor amplifies the contribution of the low energy region of the spectral function so that this additional strength is becoming the main reason for the overall enhanced production of the @xmath1 in the medium .
the influence of a hot and dense medium on the properties of the hadrons involved in the determination of the @xmath1/@xmath6 ratio has been studied .
we have focused our attention on incorporating the effects of the antikaon properties in the medium at finite temperature .
it is found that the temperature and chemical potential compatible with a given ratio depend very strongly on the approach used for the in - medium properties of the @xmath1 meson ( @xmath44 mev and @xmath12 up to @xmath47 mev for @xmath6/@xmath48 with the full @xmath1 spectral density ) .
the `` broad - band equilibration '' advocated by brown , rho and song is not achieved in the on - shell approach .
this behaviour is only observed when @xmath1 is described by the full spectral density due to the coupling of the @xmath1 meson to @xmath33 states .
however , the @xmath0 ratio is in excess by a factor of 2 with respect to the experimental one .
what needs to be clarified is how the particles get on - shell at the freeze - out and this has to be addressed in dynamical non - equilibrium effects .
however , this study is left for forthcoming work .
this work is partially supported by dgicyt project bfm2001 - 01868 , by the generalitat de catalunya project 2001sgr00064 and by nsf grant phy-03 - 11859 .
l.t . also wishes to acknowledge support from the alexander von humboldt - foundation .
j. cleymans , d. elliot , a. kernen , and e. suhonen , phys . rev .
* c 57 * , 3319 ( 1998 ) ; j. cleymans , h. oeschler , and k. redlich , phys . rev .
* c 59 * , 1663 ( 1999 ) ; j. cleymans , and k. redlich , phys . rev .
* c 60 * , 054908 ( 1999 ) . | the @xmath0 ratio in heavy - ion collisions at gsi is studied including the properties of the participating hadrons in hot and dense matter , paying a particular attention to the in - medium properties of antikaons at finite temperature .
the determination of the temperature and chemical potential at freeze - out conditions compatible with the ratio @xmath0 is very delicate , and depends on the approach adopted for the antikaon self - energy .
different approaches for the @xmath1 self - energy have been considered : non - interacting @xmath1 , on - shell self - energy and single - particle spectral density .
we observe that the use of an energy dependent @xmath2 spectral density lowers considerably the freeze - out temperature with respect to the on - shell approach .
we also conclude that the full off - shell approach gives rise to the broad - band equilibration " advocated by brown , rho and song . |
a previous attempt to derive light - wimp limits by the xenon100 collaboration @xcite was received with criticism @xcite pointing out the unphysical behavior of the quenching factor in the production of direct scintillation by nuclear recoils ( @xmath2 ) employed . an additional critique
@xcite emphasized traceable mistakes made in the data analysis of previous @xmath2 measurements performed by the xenon10 and xenon100 collaborations .
a new recent @xmath2 measurement by the xenon100 collaboration ( plante _ et al . _
@xcite ) addresses the concerns in @xcite , specifically the systematic effect introduced by erroneously normalizing simulated recoil rates to their corresponding measured values , and the sub - optimal design of several earlier calibration detectors , prone to multiple scattering involving inert materials .
the new detector used in @xcite features a compact design that bypasses this concern and maximizes light collection from the active liquid xenon ( lxe ) volume .
not surprisingly , the monotonically decreasing behavior of @xmath2 towards zero energy predicted in @xcite is now observed by plante _
et al . _
while great strides towards a better understanding of @xmath2 have been made in @xcite , significant room for improvement remains : @xmath3=0.em an unnecessary degree of freedom in the fits comparing lxe scintillation measurements and simulations has been introduced by plante _
et al . _ , namely the energy resolution as a function of recoil energy , which is a predictable quantity , and not independent of @xmath2 , as implicitly postulated in @xcite .
this is in contrast to an earlier measurement by manzur _
@xcite , also correctly pointing at a decreasing @xmath2 , where the resolution was determined by measurements at energies well - above any threshold effects , and for all lower energies unambiguously defined according to its expected dependence on photoelectron yield ( a function of @xmath2 ) .
the introduction of this gratuitous degree of freedom can reinstate the deleterious effects described in @xcite , by substantially biasing @xmath2 towards artificially large values and reducing uncertainty .
this concern is particularly important below @xmath46.5 kev@xmath5 , where threshold effects become dominant in @xcite .
the approach taken in @xcite does not necessarily have to constitute an issue , as long as the obtained best - fit resolution follows the expected behavior@xmath6 .
no mention of this comparison is made in @xcite .
as described in sec .
ii below , the extrapolated behavior of @xmath2 to zero energy critically determines lxe sensitivity to light - wimps , making attention to such details very important . a discussion of this comparison between expected and best - fit energy resolution would considerably improve the credibility of the lowest - energy @xmath2 values obtained by plante _
et al._. measurements in @xcite were performed in single - phase mode , i.e. , in the absence of the electric drift field present during the operation of the xenon100 detector .
this field is expected to suppress electron - ion recombination , reducing the scintillation yield . while this effect was found to be small by manzur _
@xcite , the @xmath2 values by plante _
_ should be considered an upper limit to the actual nuclear recoil scintillation yield in the xenon100 detector .
this consideration as an upper limit is revisited in sec .
ii within a different ( instrumental ) context .
it must be kept in mind that the definition of @xmath2 used by the lxe detector community differs from the traditional one for a quenching factor , by relativizing the scintillation yield from low - energy nuclear recoils to that from electron recoils at a relatively high ionization energy ( 122 kev ) .
the more conventional definition uses the ratio of scintillation yield from nuclear and electron recoils of identical energy .
this may seem like a moot point , until the large non - proportionality typically observed in heavy scintillators @xcite , including lxe @xcite , is examined : a large increase in scintillation yield for electron recoils ( the denominator in the traditional definition of quenching factor ) is typically observed below few hundred of kev down to few kev .
compton scattering measurements complementary to those in @xcite are clearly advisable .
the analysis of a 100 day exposure from the xenon100 detector @xcite has resulted in a claim of sizeable improvement in light - wimp sensitivity with respect to a previous shorter ( 11 day ) run @xcite .
a discussion of the strong assumptions implicitly made to arrive to this conclusion and of the uncertainties neglected in @xcite is provided below . by plante _
@xcite , dashed ( blue ) to that by manzur
_ et al . _
@xcite , and dash - dotted ( red ) to a most recent monte carlo - independent @xmath2 determination by zeplin - iii @xcite .
the notation used to describe each case lists number of irreducible recoil events accepted , @xmath2 considered ( central value of logarithmic extrapolation or its lower c.l .
boundary ) , and statistics used ( see text ) : a ) 1 event , central , poisson ; b ) 1 ev .
, 2@xmath7 , poisson ; c ) 4 evs . , 2@xmath7 , poisson ; d ) 4 evs . , 2@xmath7 , binomial
; e ) 1 ev . , central , poisson ; f ) 1 ev . , central , binomial ;
, 1@xmath7 , poisson ; h ) 4 evs . , 1@xmath7 , binomial ; i ) 1 ev . , central , poisson ; j ) 1 ev . , central , binomial ; k ) 4 evs . , central , binomial ;
l ) 4 evs . , 1@xmath7 , binomial .
additional instrumental uncertainties not reflected in this figure are listed in the text .
, width=321 ] @xmath3=0.em contour `` a '' in fig .
1 is similar to the exclusion curve in @xcite .
it is obtained by assuming the logarithmic extrapolation of the @xmath2 from plante _ et al .
_ , as proposed in @xcite , and that only the lowest in energy of the three accepted nuclear - recoil events in @xcite could be due to a light - wimp .
contour `` b '' in the same curve represents the exclusion obtained when the 2@xmath7 c.l .
uncertainty band in this @xmath2 is adopted instead .
the resulting change in sensitivity is much larger than what is indicated by the very narrow uncertainty bands in fig . 5 of @xcite .
this issue can be confirmed by performing a self - consistency test between the xenon100 exclusion curves in @xcite and @xcite : the two values of @xmath2 contemplated in @xcite ( fig
. 1 there ) generated exclusion curves diverging by a very large factor for light wimp masses ( fig . 5 in @xcite ) .
those two values of @xmath2 are coincidentally not very different from the central and 2@xmath7 c.l . boundaries of the @xmath2 from plante _ et al . _
( fig . 1 in @xcite ) .
however , in fig .
5 of @xcite , the new xenon100 analysis assigns an insignificant impact on the exclusions to this large spread in @xmath2 . the origin for this lack of self - consistency must be addressed@xmath1 .
in a departure from the blind analysis initially intended by the xenon100 collaboration , three events next to threshold were rejected immediately following unblinding @xcite .
these events have been ascribed to photomultiplier ( pmt ) noise affecting the s1 ( direct scintillation ) channel .
post - unblinding corrective actions are often required and no judgment on this decision should be passed until more details become available . however , it is worth remembering that this type of pmt noise was already present in xenon10 data @xcite ( and not rejected a posteriori ) and is also ubiquitous in a xenon100 example event catalogued as `` good '' @xcite , indicating that data cuts originally deemed as adequate must have been in place against it@xmath8 .
details about the post - unblinding criteria developed to reject these events while in the presence of robust s2 ( ionization ) signals will be of special interest .
contours labelled `` c '' , `` d '' and `` h '' in fig .
1 display the non - negligible effect of including these three rejected events into the calculation of xenon100 exclusions . dashed lines in fig .
1 represent xenon100 exclusions using a logarithmic extrapolation of the alternative @xmath2 obtained by manzur _
. _ @xcite using another optimally - designed lxe chamber .
differences in data treatment and mode of operation between @xcite and @xcite are described above .
criticisms concerning the data treatment in @xcite are given in @xcite .
contours labelled `` g '' and `` h '' use the logarithmic extrapolation of the lower 1@xmath7 c.l .
boundary rather than the central @xmath2 value ( contours `` e '' and `` f '' ) .
a lingering critical question is to what extent a determination of @xmath2 performed using highly - optimized compact calibration detectors like those in @xcite can be applied with confidence to a much larger device like the xenon100 detector , featuring a small s1 light - detection efficiency ( just @xmath46% @xcite ) , different hardware trigger configuration , data processing , etc .
for instance , simulations like those used within xenon100 to obtain a cumulative cut acceptance near threshold can only be regarded as best - effort estimates .
their limited meaning and tendency to significantly overestimate near - threshold efficiency has been recently encountered by plante _
@xcite , even for the near - ideal conditions of their small calibration chamber ( @xmath418 c.c . of lxe , with 4@xmath9 pmt coverage ) .
another example of instrumental constraints is the negligibly small low - energy _ effective _ @xmath2 derived by the 12 kg zeplin lxe dark matter detector @xcite when its measurement is attempted _ in situ _
( if adopted , this @xmath2 generates essentially no lxe sensitivity to wimp masses below 10 gev / c@xmath1 @xcite ) .
going back some time , a dramatic deficit in observed neutron - induced recoil rates compared to few - kev@xmath5 expectations was observed with the xenon10 detector @xcite .
such comparisons should be revisited within the context of the existing xenon100 neutron calibrations : if the expected response to few - kev@xmath5 recoils is still absent due to instrumental limitations , light - wimp limits should not be distilled from sheer wishful thinking . as discussed next ,
light - wimp limits are obtained by xenon100 under the strong assumption of positive poissonian fluctuations in scintillation light from wimp - induced recoils well - below the detector energy threshold .
the agreement between expected and observed neutron - induced recoil rates should therefore be demonstrated into that energy range .
the xenon100 collaboration is invited to produce this much needed validation of their claims@xmath10 .
the mentioned assumption of poissonian statistics governing the microscopic processes@xmath11 of light and charge generation by few kev@xmath5 nuclear recoils in lxe is not only presently unwarranted @xcite , but seemingly counter to the scarce available experimental information : the very small value of the fano factor in lxe is for instance indicative of sub - poissonian statistics ruling those processes @xcite .
similarly , the electron emission statistics by few - kev , heavy - mass ions during surface collisions is known to be better described by binomial rather than poissonian statistics @xcite .
contours `` d '' , `` f '' and `` h '' in fig . 1 display the effect of a small deviation from the poisson assumption ( a binomial distribution of same mean , taking a probability of s1 photon detection of 6% @xcite ) .
these contours should be considered as illustrative ansatzes for information - carrier statistics that could generate even less light production . in this respect ,
it is worth emphasizing that a 7 gev / c@xmath1 light - wimp is expected to impart a mean recoil energy in lxe of just @xmath40.6 kev@xmath5 , with an absolute maximum ( occurring with infinitesimal probability ) of @xmath44.1 kev@xmath5 . at this very endpoint ,
the probability of surpassing the xenon100 four - photoelectron ( @xmath48.5 kev@xmath5 ) threshold is never larger than @xmath410% , even when the logarithmic extrapolation of the @xmath2 by plante _
et al . _ and
poisson fluctuations are adopted .
not all light - wimp detecting media fare as poorly from this point of view of generation of information carriers : the same 7 gev / c@xmath1 wimp at its spectral endpoint ( @xmath41.4 kev ionization ) in cogent germanium diodes would generate a readily detectable @xmath4470 electron - hole pairs .
1 includes the present uncertainty in the quenching factor for sodium recoils in dama / libra @xcite , a subject of discussion avoided by the xenon100 collaboration in @xcite .
this uncertainty extends the dama / libra region to considerably lower wimp masses than what is represented in @xcite .
a recent reanalysis of xenon10 data uses strictly the ionization channel in that detector to impose limits on light - mass wimps @xcite .
this s2 light emitted via electroluminescence from charge drifted into the gas phase of the device is , when examined alone , sufficient to extend the sensitivity of lxe detectors to recoils of o(1 ) kev@xmath5 .
not including the information from s1 ( direct scintillation ) allows a reduction in threshold at the expense of losing the ability to distinguish nuclear from electron recoils .
while this approach is promising , the few kev@xmath5 nuclear recoil energy scale corresponding to this s2 channel is presently hopelessly ill - defined .
this is a result of the inadequate `` best - fit monte carlo '' method @xcite employed to arrive at it .
an extensive critique of this method , ignored thus far by xenon10 authors , can be found in @xcite . in a troubling case of double - standards ,
the gist of this critique ( that with this method all uncertainties are absorbed into the energy scale ) has been recently echoed by the xenon100 collaboration ( sec .
i in @xcite ) , when rebuking indirect measurements of @xmath2 using the same methodology . in lieu of a reiteration of the criticisms in @xcite ,
the reader is invited to inspect fig . 2 :
each of the colored energy scales shown there , generated by the `` best - fit monte carlo '' method , has been claimed to be the correct one by the xenon10 collaboration over the brief span of the last two years .
the scale is observed to change as rapidly as from workshop presentation to its published proceedings .
its monotonic evolution has been towards the black curves , held in @xcite to correspond to the most plausible energy scale ( one derived from an earlier method laid out by xenon10 authors : see pertinent discussion in @xcite ) . a critically - minded reader would ( rightly ) argue that none of these can be presently assigned any credibility at few kev@xmath5 .
however , reasons have been provided in @xcite to support the solid black curve , representing the lindhard theory modified below a @xmath440 kev@xmath5 kinematic threshold @xcite by an example of adiabatic correction , as in @xcite : @xmath3=0.em this energy scale generates a similar quenching for the ionization yield and the @xmath2 observed by plante _
et al . _ or manzur _ et al . _ ,
i.e. , a monotonic decrease in the generation of information carriers ( free charge , direct scintillation photons ) below kinematic threshold @xcite , having an effective cutoff at @xmath41 kev@xmath5 . that both processes should decrease hand - in - hand
can be argued based on the dominant role of ionization as the main precursor to direct scintillation for low energy recoils in lxe @xcite .
the energy scales postulated thus far by xenon10 ( color curves in fig .
2 ) overestimate , by several orders of magnitude , the very small average charge yields observed in impact ionization experiments involving few kev and sub - kev xenon ions .
the relevance of these measurements and examples from the literature are discussed in @xcite , where it is also emphasized that xenon10 workers are not without a reference on what to expect at these low energies .
the introduction of the adiabatic correction proposed in @xcite resolves this disagreement .
contour in @xcite , as per @xcite ; dashed blue @xcite .
dark lines correspond to the predicted energy scales in @xcite ( see text ) .
inset : alternative analysis of the few - electron s2 background in xenon10 @xcite ( see text ) . ,
width=283 ] of special interest is a population of xenon10 low energy ionization signals described in @xcite as single electrons@xmath12 .
this definition is both surprising and misleading . as mentioned in @xcite , the large amplification gain provided by the electroluminescence provides a good resolution in the multiplicity of drifted charge in lxe .
these events clearly include a multi - electron component and their population has been described as such by the contact author of @xcite , as recently as in @xcite .
their origin is unknown , and hard to ascribe to minimum ionizing particles in an efficient self - shielding medium such as lxe @xcite . while none of this is discussed in @xcite , their accumulation towards low values of s2 pulse - width is to be expected from the effect of charge multiplicity on this variable , and does not have to correspond to a radioactive contamination close to the @xmath13 detector coordinate ( this is unlikely , given that roughly the same number of pmts , major sources of internal activity in the xenon10 detector , are placed at both extrema of @xmath14 ) .
the inset in fig . 2 represents the differential rate of few - electron s2 events in xenon10 , obtained by applying the same five background cuts as in @xcite , extending the analysis down to the s2 = 1e@xmath15 boundary , and adopting the energy scale described by the black solid line in the same figure .
this differential rate seems to be also compatible with fig . 2 in @xcite , once the adopted energy scale is included .
it offers a good match to the expected signal from a light wimp in the region of interest of other searches ( dama / libra , cogent , cresst ) .
the outcome of this exercise , performed here strictly for the sake of argument , should be strongly de - emphasized at this time , given the present lack of knowledge about this energy scale evidenced in fig . 2
. the reader should remember instead that a further evolution by a mere @xmath41 kev@xmath5 in the ever - changing s2 energy scales postulated by xenon10 can transform the severe constraints " of @xcite into a signal in principle compatible with a light - wimp . in conclusion ,
the claims in @xcite are clearly presently untenable .
awaiting clarification of the several pending issues pointed out in sec .
i and ii , light - wimp limits obtained through a more conventional analysis of xenon100 data @xcite can only be assigned the very limited meaning illustrated by fig . 1 .
the xenon100 collaboration is congratulated for the recent advancement in their understanding of @xmath2 , encouraged to develop improved methods of characterization of the s2 energy scale leading to a reliable exploration of light - wimp candidates , and urged to employ transparency in the discussion of uncertainties and assumptions underlaying their results , in view of the very limited performance of lxe as a light - wimp detection medium . finally , while several interesting phenomenological routes to alleviate tension between lxe constraints and other light - wimp searches have been put forward recently @xcite
, these deviations from arguably more conventional assumptions do not seem to be mandatory at this time .
a new measurement of @xmath2 by the zeplin - iii collaboration appeared coincident with the release of this preprint @xcite .
fig . 1 now reflects xenon100 exclusions obtained with it .
zeplin - iii derives a @xmath2 decreasing below 40 kev@xmath5 and vanishing at few kev@xmath5 , in tight agreement with the lxe kinematic threshold described in @xcite .
barbeau , j.i . collar and p.m. whaley , nucl .
a574 * ( 2007 ) 385 ; p.s .
barbeau , j.i .
collar and o. tench , jcap 0709:009,2007 ; p.s .
barbeau , phd dissertation , university of chicago , 2009 .
fox , j. liu and n. weiner , arxiv:1011.1915v1 ; a.l .
fitzpatrick and k.m .
zurek , arxiv:1007.5325 ; j.l .
et al . _ , arxiv:1102.4331 ; m. frandsen _
et al . _ , arxiv:1105.3734 ; c. arina
_ , arxiv:1105.5121 ; e. del nobile _
_ , arxiv:1105.5431 . | the underlaying assumptions and uncertainties involved in the derivation of wimp exclusion limits from xenon10 and xenon100 detectors are examined . in view of these
, recent claims of sensitivity to light - mass weakly interacting massive particles ( wimps ) are shown to be overstated .
specifically , bounds constraining regions of interest in wimp parameter space from the dama / libra , cogent and cresst experiments can be assigned only a very limited meaning .
the possibility of relatively - light ( @xmath010 gev / c@xmath1 ) weakly interacting massive particles ( wimps ) being the source of the annual modulation effect @xcite observed in the dama / libra experiment @xcite was first proposed in @xcite and later revisited in @xcite .
since then , it has gained in popularity as constraints from other searches have depleted alternative dark matter scenarios , and several particle phenomenologies have generated plausible candidates in this mass range .
the situation has recently gained in complexity with the observations from cogent @xcite and cresst @xcite , which may point at a light - wimp parameter space compatible with dama / libra .
studies of detector sensitivity to such light dark matter particles must be regarded with caution @xcite . for presently - available technologies
, light - wimp signals would fall uncomfortably close to detector thresholds , a region where systematic effects can lead to rushed claims of exclusion or detection .
experimentalists should attempt not to aggravate what is a naturally murky area of study , by describing the assumptions made to arrive to their conclusions , and by including an assessment of the uncertainties involved in their analyses .
recent work by the xenon10 @xcite and xenon100 @xcite collaborations is lacking in these respects .
the goal of this report is to provide this missing information as succinctly as possible . |
brown dwarfs ( bds ) are an ubiquitous outcome of the star formation process .
all young regions investigated so far with sufficient depth host a population of bds with masses down to 0.01@xmath2 or even below . the mechanism that governs their formation
, however , remains unknown .
it is clear that additional physics needs to be included in the models for the cloud fragmentation and subsequent evolution , to allow for the formation of a sizable number of bds @xcite .
plausible options for these processes include fragmentation driven by turbulence , dynamical ejection of embryonic bds from multiple systems , fragmentation of filaments falling into a cluster potential , or fragmentation of protoplanetary disks , again combined with ejection @xcite .
young brown dwarfs are a critical population to test the relevance of these processes .
the standard diagnostic to distinguish between theoretical scenarios is the distribution of stellar and substellar masses after star formation is finished , or the initial mass function ( imf ) . in the literature several parameterisations for the imf are used , for example a series of power laws @xcite or a lognormal form @xcite . for our goal of determining the abundance of brown dwarfs ,
an often used parameterisation of the imf is the star / bd ratio @xmath0 , the ratio of the number of objects in the two mass bins from 0.08 to 1.0@xmath2 and from 0.03 to 0.08@xmath2 , where the low mass cutoff as 0.03@xmath2 is chosen to assure completeness .
the upper mass limit for the stars of 1.0@xmath2 is to some extent an arbitrary definition , but the relatively small number of higher - mass stars in the nearby star forming region assures that this particular choice does not affect the result much .
the star / bd ratio as a metric has the advantage of maximising the sample size in the substellar regime and thus minimizing the statistical errors .
because the stellar side of the imf is well - determined for the nearby star forming regions and shows , in the overwhelming majority of regions , no evidence for environmental differences , any variation in the star / bd ratio from one region to another would indicate a change in the bd abundance . measuring the substellar mass function and the star /
bd ratio is a challenging task .
it needs a consistent survey procedure and a careful analysis of possible incompleteness , but the core problem is to estimate the masses .
this requires one to make assumptions about the distance to and age of the region , as well as the extinction law used to deredden the photometry or spectra .
furthermore , the conversion from observed quantities to masses can only be done in the framework of a given theoretical isochrone and thus depends on the status of the evolutionary models for young stars and brown dwarfs . in previous work within the project sonyc ( substellar objects in nearby young clusters )
we have presented tentative evidence for regional differences in the star / bd ratio .
in particular , for the young cluster ngc1333 in the perseus star forming complex we find @xmath3 based on a very deep survey with comprehensive spectroscopic follow - up ( @xcite , hereafter sonyc - i ; @xcite , hereafter sonyc - iv ) . for other regions
we and other groups have published r - values ranging from 2 to 8 ( see sonyc - iv ) .
one of the most extreme cases in the literature is the cluster ic348 , the second embedded cluster in perseus and a slightly older sibling to ngc1333 @xcite , with a star / bd ratio of @xmath4 @xcite . at face value , this indicates changes in the bd abundance by a factor of 4 occuring within the same star forming association .
so far , however , the uncertainties for the star / bd ratio have not been assessed accurately .
the goal of this paper is to carry out a benchmark test for the two extreme cases ngc1333 and ic348 to verify their discrepant star / bd ratios .
the core idea of this paper is to determine the mass distribution and the star / bd ratio for two young clusters in perseus and to assess the associated uncertainties
. we will start with consistently selected samples for the two clusters , i.e. samples which have been put together in a homogeneous way , to minimize the influence of selection biases and incompleteness .
for these samples , we will then define a consistent way of estimating object masses by comparing photometry with a given model isochrone .
in addition to the choice of the isochrone , the distance to the cluster , the extinction law , and the age of the region enter as free parameters into this procedure .
we will define a set of scenarios with plausible choices for these parameters and different isochrones .
we will then estimate object masses and calculate the star / bd ratio and other indicators of the mass distribution for each of these scenarios .
this will yield a useful dataset to discuss the uncertainties in these indicators and assess whether there is evidence for regional differences in the substellar imf between ngc1333 and ic348 .
we use the spitzer - selected sample of young stellar / substellar objects presented by @xcite which includes our two target regions . for each of the regions , a region of @xmath5 centered on the core of the cluster was observed .
the selection in @xcite is based on colours and magnitudes in spitzer and 2mass bands from 1 to 24@xmath6 .
their multi - colour selection process uses a series of criteria designed to exclude background stars , non - stellar emission features and extragalactic objects . according to their analysis
, this process yields only minimal contamination ( a few percent ) .
an earlier version of this process was used in @xcite in a survey of ngc1333 .
assuming that the distances to the clusters are similar , the identical selection procedure also ensures that the depth and completeness in terms of magnitudes in the two samples is comparable .
this removes a major obstacle for an accurate comparison of the bd abundance .
since the primary selection criterion is excess emission in the infrared , this sample only contains objects with emission from circumstellar material , i.e. either disks or envelopes .
it does not include disk - less young stars and brown dwarfs ( class iii objects ) .
therefore , to infer star / bd ratios , we have to make the assumption that the fraction of objects with disks does not change with object mass in the low - mass regime .
as recently shown in @xcite , this is a plausible assumption for many star forming regions , including ic348 . for ngc1333 , there might be a slight mass dependence , as the disk fraction in the total spitzer - selected sample is found to be @xmath7% @xcite , whereas the value for the very low mass objects is only 55 - 66% ( sonyc - iv ) .
this indicates that the star / bd ratio in this cluster could be slightly overestimated .
the entire perseus cloud , including the two target clusters have also been observed as part of the spitzer legacy program from cores to disks ( c2d , pi : n. evans ) . their perseus yso catalogue derived from irac photometry
is discussed in detail in @xcite .
we prefer to use the @xcite selection , because it is slightly deeper , which is beneficial for our purposes .
the downside of the @xcite sample is the limited spatial coverage . here
the c2d catalogue is useful to check the spatial completeness of our samples . + in fig .
[ f0 ] we show the spatial distribution of the selected objects in ngc1333 and ic348 . with red squares we plot the samples from @xcite , with black crosses the c2d sample that covers the entire perseus region .
the figure shows that the number of yso candidates outside the region covered by @xcite is small compared with the total population .
this is particularly true for ngc1333 which shows a very compact profile and can be considered to be spatially complete ( see also sonyc - iv ) . for ic348
, there is a tail of a yso population towards the south - west , which represents the transition region to another densely populated area in the perseus star forming region .
in addition , there are about 10 objects outside the coverage of the @xcite survey . however , there is also a large number of additional yso candidates only contained in the c2d sample within the cluster core .
we checked the objects only in c2d and found that they do not show an obvious magnitude or extinction bias with respect to the sample we are using , thus , even in case we are missing members in the outskirts of the cluster , this is not going to affect our analysis in any significant way . while @xcite do find a difference in the imf between the core and the halo in ic348 , both regions ( in their definitions ) are within the survey area of @xcite .
we conclude that the samples we are using are not affected by a spatial bias .
[ f0 ] also illustrates one major difference between ic348 and ngc1333 .
the core of ic348 has about twice the diameter of the core of ngc1333 ( 2.1 vs. 1.2pc , * ? ? ?
* ) , i.e. the cluster volume in ic348 is about 8 times larger than in ngc1333 . on the other hand
, ic348 has only about 20% more ysos than ngc1333 , according to the spitzer surveys @xcite .
the fraction of diskless class iii objects is higher in ic348 , taken that into account the total the yso population in ic348 could be up to twice as large as in ngc1333 .
this still implies that the object density in ngc1333 is 4 - 7 times higher than in ngc1333 .
given the age , size , and number of members in these clusters , this difference is likely to be primordial , and not caused by dynamical evolution ( see fig . 1 in * ? ? ?
hence , these two clusters constitute an excellent test case to probe the effects of dynamical interactions and cluster potential on the formation of bds . for the overwhelming majority of young objects , masses can only be estimated indirectly by comparing an observed quantity with predictions from theoretical isochrones for a given age .
for the observed quantity , there are two options , either the effective temperature or the luminosity ( or a photometric proxy ) .
the luminosity has the problem that model derivations are sensitive to the age for pre - main sequence objects that are still contracting .
in addition , measurements can be affected by extinction as well as excess emission from disk and/or accretion .
the effective temperature is problematic for other reasons ; it depends on atmosphere models and can be altered by magnetic activity @xcite .
this can lead us to underestimate object masses by up to a factor of two . for our chosen samples ,
accurate multi - band photometry is available , while the spectroscopic follow - up is not complete .
therefore , we will rely on photometry in the optical and near - infrared to estimate masses .
we complement the 2mass photometry provided by @xcite with optical photometry from @xcite and @xcite for ic348 ( landolt r- and i - band ) and from sonyc - i for ngc1333 ( sloan i- and z - band ) .
for some objects without 2mass near - infrared magnitudes , we were able to complement the dataset using the photometry from @xcite for ic348 and sonyc - i for ngc1333 . in total , the samples contain 142 ( for ic348 ) and 95 ( for ngc1333 ) objects with photometry in jhk , with smaller subsets of 86 ( ic348 ) and 23 ( ngc1333 ) with additional optical magnitudes available . for the objects with 2mass photometry , we also obtain the error as listed in the database ( mostly between 0.03 and 0.05mag ) .
similarly , the photometry from @xcite provides errors for all measurements . for the remaining photometry , errors for individual objects
are not reported in the literature .
for the optical magnitudes in ic348 , we adopt a generic and conservative uncertainty of 0.1mag . for the sonyc magnitudes in ngc1333
, we adopt errors of 0.1mag for the optical bands and 0.05mag for the near - infrared bands . based on the information given in the papers listed above , the error values adopted for
the optical photometry should be typical for the samples .
however , some objects might be affected by additional uncertainties introduced to calibration imperfections .
since the z- and i - bands are located at the long - wavelength edge of the sensitivity of the optical ccds , they are highly susceptible to colour terms in the calibration , which are difficult to measure with the usual photometric standard stars .
this can introduce errors larger than 0.1mag in individual sources , which can not be quantified accurately .
this issue is a particular problem for very red sources , since most of their optical flux is emitted in the part of the spectrum where the ccd sensitivity declines .
we derive masses using three different sets of isochrones from the lyon group , bt - settl ( with agss2009 opacities ) , bt - dusty ( with agss2009 opacities ) , and bt - nextgen ( with gnc93 opacities ) .
the latter two are updated versions from the standard ames - dusty and nextgen models .
the main difference between the three sets is the treatment of dust .
in contrast to nextgen , dusty includes dust opacities .
settl includes a full dust cloud model . for more information on the isochrones ,
see @xcite .
note that atmospheric dust becomes a major source of opacity for @xmath8k @xcite , corresponding to @xmath9 for young brown dwarfs , which is the low - mass limit in our analysis , thus , the treatment of dust should not have a major effect on our results .
the isochrones predict absolute magnitudes as a function of object mass in all photometric bands for which observations are available .
they cover the range from 0.02@xmath2 or below to 1.4@xmath10 and are available for ages starting from 1myr , which is adequate for our purposes . to compare the observed with the predicted magnitudes , we first shift the isochrones from absolute magnitude to the distance of our target regions , which enters here as a free parameter ( see sect .
[ s23 ] ) .
we also re - bin the isochrones to a uniform stepsize of 0.01@xmath2 , using a linear interpolation over a small portion of the isochrone .
we then calculate a series of reddened isochrones for @xmath11mag in steps of 1mag . for this step ,
the choice of the extinction law is important ( see sect . [ s23 ] ) .
the upper limit is chosen to be 20mag , for two reasons .
first , independent studies indicate that only a small fraction in the perseus star forming complex exceeds this extinction value .
second , beyond this value the samples are biased towards bright sources , thus , incompleteness becomes an issue . after these preparations , the best fit for mass and @xmath12 is determined with a @xmath13 minimization .
the number of degrees of freedom in this process is the number of photometric bands for which data is available ( @xmath14 to @xmath15 ) minus the number of free parameters ( mass and @xmath12 , i.e. 2 ) . for each object
, we saved the combination of mass and @xmath12 that results in the minimum value for @xmath13 , the corresponding reduced @xmath13 ( @xmath16 , i.e. @xmath13 divided by the number of degrees of freedom ) and the number of available bands @xmath17 .
objects with best fit value of @xmath18 are discarded from the analysis since this is the upper limit in our grid of isochrones , their mass estimate is not reliable .
a typical example of the resulting @xmath16 plotted vs. the best mass estimate is shown in fig .
[ f2 ] , left panel .
this figure reveals that the procedure produces similar fitting results for high- and low - mass objects and for objects with and without optical photometry .
we also show a typical @xmath12 vs. mass plot from this procedure ( fig .
[ f2 ] , right panel ) .
the upper limit in @xmath12 is not changing significantly with mass , i.e. there is no evidence for an extinction bias in these samples ( apart from the @xmath19 cutoff ) .
the distribution of reduced @xmath13 can in principle be used to assess the goodness - of fit for our mass estimates . for a good fit
, we expect @xmath16 to have an average of 1.0 and a standard deviation of @xmath20 . however , as shown in fig .
[ f2 ] , our procedure yields significantly higher values in @xmath16 .
this indicates that either the model does not reflect the data well or that the errors are underestimated . in our case , the high values for @xmath16 are mostly explained by the fact that our model is discretely sampled in mass-@xmath12 space .
the stepsize in mass and @xmath12 results in magnitude steps that are often larger than the typical photometric error .
in addition , in some cases the errors of the optical photometry may be underestimated , see above .
therefore , we only use the procedure to select the best fit solution . in our estimation of masses from photometry , the distance , age , as well as the extinction law , expressed in the quantity @xmath21 ,
are considered free parameters .
in addition , we have to choose the theoretical isochrone .
what we call scenarios in the following are combinations of isochrone , distance , age , and extinction law for which we estimate masses .
these scenarios have been chosen to cover the plausible range of these parameters and to give insight into the impact of the specific choice of a parameter or an isochrone on the mass estimates . in the following ,
we justify the choice of the range of the parameters . for the extinction , we use the parameterised law by @xcite , with @xmath22 , the canonical value used for the ism . this law yields extinction offsets that are consistent with the often - used extinction values published by @xcite . in reality , @xmath23 depends on the grain properties and is not the same for every line of sight ; @xcite report values ranging from 2.6 to 5.6 , with the overwhelming majority ( 22 out of 27 cases ) below 4.5 .
we therefore use @xmath24 as an alternative value to be able to assess the impact of the choice of @xmath23 on the mass estimates .
the distances to the two clusters are not well constrained .
the entire perseus cloud is usually assumed to have an average distance of @xmath25pc , which we use as a default value .
based on the hipparcos parallaxes for the early - type stars @xcite estimate @xmath26pc for the cloud .
based on a kinematical analysis of a much larger sample of a stars , infer 300pc ( 270 - 330pc ) . however , there are indications that ngc1333 is located at a shorter distance .
@xcite report a distance of 235pc for ngc1333 based on interferometry of the maser emission from a source that may be associated with the cluster .
this is also consistent with an earlier photometric estimate of the distance of ngc1333 ( 220pc ) by .
in addition , suggest , based on extinction map analysis , that the northern part of the perseus region ( where ic348 is located ) is slightly more distant than the southern part ( where ngc1333 is located ) . to take this into account
, we use a distance of 230pc as an alternative value , noting that this is only a viable option for ngc1333 .
the ages that are typically quoted are 2 - 4myr for ic348 and 1 - 3myr for ngc1333 ( see * ? ? ?
* and references therein ) .
judging from model - independent indicators of evolutionary state ( fraction of objects with disks , fraction of objects in class i stage , luminosity function ) , ngc1333 is definitely younger than ic348 , and both are clearly younger than star forming regions with established ages of 5 - 10myr like upper scorpius and the tw hydrae association ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the context of our study ,
the relevant quantity is not the age of the cluster , but the average age of the objects contained in our samples , which may not include the youngest , embedded population because we require a near - infrared detection . in fact , we showed in sonyc - iv that most of the very low mass objects in ngc1333 are consistent with an age of 1 - 5myr , based on their position in the hertzsprung - russell diagram .
therefore , we use a default value of 3myr , which is plausible for both clusters . to assess the influence of the age on the mass estimates , we additionally estimate masses for an age of 1myr . to evaluate the impact of the choice of the parameters , we define 6 scenarios for which we estimate object masses .
these scenarios are listed in table [ t1 ] .
the default scenario # 1 uses a distance of 300pc , an age of 3myr , @xmath23 of 3.1 , and the bt - settl model , which has the most recent opacities and the most advanced treatment of dust . in scenarios # 2 to # 4
we vary the cluster parameters .
in scenario # 2 we use the younger age of 1myr , in # 3 the alternative distance of 230pc , and in # 4 the alternative value for @xmath23 of 4.5 . in scenarios # 5 and # 6 we switch to the nextgen and dusty isochrones .
lccclclcccc ic348 & 1 & 300 & 3 & bt - settl & 3.1 & 3.6 ( 96/27 ) & 2.8 - 4.4 & 2.6 - 4.1 & 0.72 & 0.27 + ic348 & 2 & 300 & 1 & bt - settl & 3.1 & 2.1 ( 88/42 ) & 1.7 - 2.5 & 1.5 - 2.9 & 0.95 & 0.13 + ic348 & 3 & 230 & 3 & bt - settl & 3.1 & 1.9 ( 84/45 ) & 1.5 - 2.2 & 1.3 - 2.1 & 1.00 & 0.16 + ic348 & 4 & 300 & 3 & bt - settl & 4.5 & 3.2 ( 94/29 ) & 2.6 - 4.0 & 2.7 - 4.3 & 0.76 & 0.26 + ic348 & 5 & 300 & 3 & bt - nextgen & 3.1 & 4.0 ( 99/25 ) & 3.1 - 4.9 & 3.4 - 5.5 & 0.67 & 0.19 + ic348 & 6 & 300 & 3 & bt - dusty & 3.1 & 2.9 ( 92/32 ) & 2.3 - 3.5 & 2.4 - 3.4 & 0.81 & 0.22 + n1333 & 1 & 300 & 3 & bt - settl & 3.1 & 2.2 ( 43/20 ) & 1.6 - 2.8 & 1.9 - 2.5 & 0.94 & 0.27 + n1333 & 2 & 300 & 1 & bt - settl & 3.1 & 2.1 ( 47/22 ) & 1.6 - 2.7 & 2.0 - 2.5 & 0.94 & 0.13 + n1333 & 3 & 230 & 3 & bt - settl & 3.1 & 2.4 ( 47/20 ) & 1.8 - 3.0 & 2.2 - 2.4 & 0.90 & 0.18 + n1333 & 4 & 300 & 3 & bt - settl & 4.5 & 2.0 ( 43/21 ) & 1.6 - 2.6 & 1.9 - 2.5 & 0.96 & 0.30 + n1333 & 5 & 300 & 3 & bt - nextgen & 3.1 & 2.2 ( 44/20 ) & 1.7 - 2.8 & 1.9 - 2.6 & 0.93 & 0.18 + n1333 & 6 & 300 & 3 & bt - dusty & 3.1 & 1.9 ( 42/22 ) & 1.5 - 2.5 & 1.8 - 2.0 & 0.99 & 0.22 + from the resulting mass distribution in a given scenario , we derive the cumulative distribution of object masses , i.e. the fraction of objects below a given mass , as a function of mass .
these plots are shown in fig .
for each scenario , we determine the number of objects with @xmath27 ( bds ) and with @xmath28 ( stars ) and calculate the star / bd ratio @xmath0 . to be able to compare with the literature , we also calculate the slope of the mass function @xmath29 for the power - law parameterisation @xmath30 , directly from the star / bd ratios , i.e. using two bins in mass , one for bds and one for stars . in addition , we derive the median mass @xmath31 for each scenario .
the resulting parameters for the 6 scenarios are given in table [ t1 ] .
+ + + an important part of this work is to evaluate the errors in the derived parameters . in our chosen experiment ,
five factors contribute to the uncertainties : * 1 ) sample size : * the part that is easiest to quantify is the statistical uncertainty , which is purely determined by the sample size . for this paper
, we use the same approach as in sonyc - iv , which is based on the idl scripts presented in @xcite . in short , we calculate bayesian confidence intervals from the beta distribution . we note that for large samples this procedure gives results that are very similar to binomial confidence intervals .
the resulting values are listed in table [ t1 ] , column 8 .
typically , the sample size introduces an error in @xmath0 of about @xmath320.3 - 0.9 for ic348 and @xmath320.5 - 0.6 for ngc1333 . *
2 ) models : * masses are only defined in relation to evolutionary models , and at young ages the deficiencies of the available tracks are well documented . however , at the moment no tracks with self - consistent treatment of the collapse and infall are available to the community .
based on a dynamical mass estimate for the very low mass pre - main sequence object ab dor c , which is older than our target regions , @xcite claim that the existing mass - luminosity relations underestimate masses by a factor of about two for young objects , but this claim has been questioned @xcite . at young ages and very low masses ,
the only direct benchmark test for these tracks is the eclipsing brown dwarf binary 2 m j05352184 - 0546085 @xcite .
the baraffe et al .
isochrones fail to reproduce the surprising temperature reversal in this object ( i.e. the more massive object is cooler than the secondary ) , an effect that is likely related to the presence of strong magnetic fields on the primary @xcite .
the luminosities of the two components , however , are consistent with the isochrones . @xcite have discovered and analysed a very young eclipsing binary with component masses around @xmath33 ; that system confirms the isochrones within the errorbars as well .
thus , while more work is required to calibrate the isochrones , some preliminary trust in their validity seems warranted . *
3 ) cluster parameters : * as discussed in sect .
[ s22 ] , several properties of the clusters affect the mass estimates , in particular the distance , the age , and the extinction law . from our set of scenarios documented in table [ t1 ] ( scenarios 1 - 4 ) we can assess how the uncertainties in these parameters propagate through the procedure . in general , changes in age and distance cause significant changes in the estimated mass distribution , while a change in the extinction does not . for ngc1333 , the induced variations in the star / bd ratios are small ; @xmath0 varies from 2.0 to 2.4 , smaller than the statistical uncertainties . for ic348 ,
the scatter is larger , from 1.9 to 3.6 , but excluding the implausible scenarios with age of 1myr and distance of 230pc this range shrinks to 3.2 - 3.6 .
we note that the star / bd ratio increases somewhat with assumed age .
for example , for ic348 and an ( unlikely ) age of 5myr we obtain @xmath34 .
this is easy to understand as the objects evolve , they become fainter i.e. the same magnitude corresponds to a larger mass . as a result ,
objects move from the substellar to the stellar domain , and the star / bd ratio increases .
* 4 ) completeness : * the samples we are using have the same depth and completeness , in terms of magnitudes , which eliminates one major source of uncertainty .
however , in terms of object masses the depth of the survey depends on the assumed distance and age .
assuming a shorter distance , as well as a younger age , will produce lower masses for the same magnitudes , i.e. the entire mass distribution would be shifted to lower masses , including the limits for completeness . for our default scenario with age of 3myr and distance of 300pc , the magnitude limit of the samples
correponds to @xmath35 .
the alternative distance of 230pc ( scenario # 3 ) implies a magnitude shift by 0.6mag . according to bt - settl isochrones
, this translates to a mass limit of 0.015@xmath2 .
the alternative age of 1myr ( scenario # 2 ) yields a new mass limit of 0.012 - 0.015@xmath2 .
thus , in these scenarios we would be sensitive to slightly lower mass objects , but since the object density is very low in this mass domain , this has only a minuscule effect on the mass distibution .
furthermore , it is not going to affect the star / bd ratios .
* 5 ) degeneracy : * in many cases there are multiple mass/@xmath12 combinations that fit the data with a similar @xmath13 . in particular , for an object with a best mass estimate around the hydrogen burning limit ( 0.07 - 0.09@xmath2 ) this method is unable to distinguish between a star and a bd . as an estimate of
the introduced uncertainty in @xmath0 we selected all objects in this mass regime and included them first in the bd count ( for the lower limit of the star / bd ratio ) and then in the star count ( for the upper limit ) .
the resulting ranges in @xmath0 are listed in table [ t1 ] , column 9 .
these intervals are @xmath320.1 - 0.3 for ngc1333 and @xmath320.5 - 1.0 for ic348 .
we note that the degeneracy is less of a problem for the two other mass thresholds involved in the calculation of @xmath0 ( 0.03 and 1.0@xmath2 ) , simply because the number of objects around these limits is low . combining these error sources , the values of @xmath0
are affected by an uncertainty of approximately @xmath36 for the two samples studied here .
currently this is about the best that can be done in terms of estimating this indicator for young star forming regions .
this translates to an uncertainty of about @xmath37 in the power - law slope @xmath29 .
the median of the mass function can be estimated with an accuracy of @xmath38 for nearby star forming regions .
looking ahead , there are obvious ways to lower these uncertainties .
first , future evolutionary tracks need to include realistic initial conditions and require more detailed calibration ( e.g. , with eclipsing binaries ) .
second , the extinction parameters need to be studied in more detail for individual regions .
third , independent estimates for the ( relative or absolute ) ages of young clusters are needed .
fourth , the accuracy in the distances of these clusters ( and many other well - studied star forming regions ) should be improved . in this respect
, the gaia satellite will be a major opportunity , as it is anticipated to be able to measure distances with 1% accuracy for open clusters within 1kpc @xcite , a huge improvement over the current estimates .
fifth , it is a worthwhile goal to obtain comprehensive sets of multi - filter photometry for star forming regions . and sixth , additional survey work in rich star forming regions with significantly more members ( factor 10 or more ) than the well - studied nearby regions can be used to minimize the statistical uncertainties .
however , since these regions are only found at large distances of @xmath39kpc , such studies have to be postponed until larger facilities , namely jwst or elts , are available .
our procedure yields for each scenario and for each cluster a distribution of masses . prior to calculating parameters for the imf
, we examine these distributions directly . in fig .
[ f1 ] we show for all 6 scenarios the cumulative distribution of object masses , i.e. the fraction of objects below a given mass , as a function of mass . we compare the 36 combinations of the functions shown in fig . [ f1 ] ( 6 for ic348 and 6 for ngc1333 ) with a kolmogorov - smirnov ( ks ) test to search for differences in the mass distribution . by doing that , we test the null hypothesis that two given distributions are drawn from the same parent distribution . in table [ t2 ]
we list the probabilities that this hypothesis is valid .
lccccccc ic348 - 1 : default & 0.26 & * 0.02 * & * 0.04 * & 0.50 & * 0.001 * & 0.17 + ic348 - 2 : @xmath40myr & * 0.002 * & 0.49 & 0.10 & * 0.003 * & * 0.01 * & * 0.002 * + ic348 - 3 : @xmath41pc & 0.32 & 0.33 & 0.32 & 0.39 & * 0.03 * & 0.32 + ic348 - 4 : @xmath24 & 0.26 & * 0.003 * & 0.05 & 0.13 & * 0.001 * & 0.17 + ic348 - 5 : nextgen & 0.80 & * 0.016 * & 0.30 & 0.76 & * 0.02 * & 0.54 + ic348 - 6 : dusty & 0.47 & * 0.02 * & 0.08 & 0.74 & * 0.004 * & 0.32 + for 14 out of 36 combinations , this probability is @xmath42% , i.e. the null hypothesis should be rejected .
6 of them are combinations that include scenario # 5 for ngc1333 , i.e. the one that uses the nextgen isochrone .
this scenario produces an unusual large number of low - mass bds ( 0.02 - 0.05@xmath2 ) for ngc1333 , which causes the discrepancy with other distributions .
this shows that apparent differences in the mass distribution of young clusters can be introduced simply by the choice of the isochrone .
we caution against comparing mass distributions derived with inconsistent isochrones .
7 further combinations of scenarios with significant differences in the mass distribution include scenario # 2 for one of the clusters , i.e. the scenario with an age of 1myr . as an example , we show in fig .
[ f3 ] ( left panel ) a comparison between the default scenario for ic348 and scenario # 2 for ngc1333 . from this figure
the origin of the difference is clear the mass distribution for ngc1333 shows a pronounced knee at 0.15@xmath2 .
however , the mass distribution in ic348 gives almost exactly the same knee when assuming an age of 1myr ( upper left panel in fig .
this effect is best explained by the knee in the mass - luminosity relation at this age , which is predicted to become weaker with age .
the resulting differences in the mass distribution can not be attributed to environmental differences .
the combination of the default scenario for ic348 and scenario # 3 for ngc1333 produces a significant difference as well ( see fig .
[ f3 ] , right panel ) .
this combination assumes the shorter distance of 230pc for ngc1333 .
we note that three further combinations with this assumption give marginally significant differences with probabilities between 5 - 10% . as explained in sect .
[ s23 ] , multiple independent studies indicate that ngc1333 is located at a shorter distance than ic348 , thus , this scenario is plausible .
+ in summary , his comparison shows that the choice of the cluster parameters and the choice of the isochrone can have noticable effects on the estimated distribution of object masses .
we find that there could be significant differences in the mass distributions between the two clusters , if ngc1333 is at a shorter distance than ic348 . in this case
, our analysis indicates a larger proportion of very low mass objects with masses @xmath43 in ngc1333 . from the mass distributions in the 6 scenarios we calculated the star / bd ratio ,
see table [ t1 ] for the results . for the cluster ngc1333 ,
our 6 scenarios give star / bd ratios of 1.9 - 2.4 , which is a range comparable to the statistical uncertainties ( see sect .
[ s33 ] ) .
this means that our previous estimate for this cluster from sonyc - iv ( @xmath44 ) is confirmed . for ic348
, we find a wider range of values between 1.9 and 4.0 .
these values imply that star / bd ratios reported in the literature for ic348 of @xmath45 @xcite are overestimated .
a possible reason for these large values is survey incompleteness or low number statistics .
as explained in sect .
[ s33 ] , the star / bd ratio depends on age , therefore these large star / bd ratios in the literature would become viable if ic348 is in fact significantly older than 3myr .
this is unlikely , as age - dependent observable quantities like the disk fraction are comparable to other 2 - 3myr old clusters @xcite . for ic348 an age of 1myr and a distance of 230pc
are not plausible ( sect .
[ s23 ] ) . excluding the scenarios using these parameters
gives a range for @xmath0 between 2.9 and 4.0 , which is our best estimate for this cluster .
these values are somewhat larger than in ngc1333 , although still within the margin of error . for example ,
for the default case the star / bd ratio is 3.6 for ic348 with a lower limit of 2.6 , and 2.2 for ngc1333 with an upper limit of 2.5 .
this is before taking into account the statistical uncertainties .
thus , based on the star / bd ratio the evidence for regional differences in the mass distribution of ic348 and ngc1333 is tentative . in table
[ t1 ] we also report two other quantities that are used in the literature to describe the imf .
the power - law slope of the mass function @xmath29 is directly determined from the star / bd ratio and thus reflects the same trends reported in sect .
[ s34 ] . for ngc1333
, @xmath29 is 0.9 - 1.0 ; for ic349 0.7 - 1.0 , or 0.7 - 0.8 after excluding the implausible scenarios .
note that the slope is an average value for the mass range 0.03 to 1.0@xmath2 ; therefore it is not unexpected to find values somewhere between the tyipcal slope of 1.3 in the regime of low - mass stars @xcite and the typical value of @xmath46 in the very low mass regime ( see sonyc - iv ) .
independent from the star / bd ratio , we determine the median mass for each scenario .
these values vary between 0.13 and 0.30@xmath2 , again indicating that the choice of the cluster parameters and the choice of the isochrone affect the results considerably . for a given scenario ,
the two clusters have very similar median masses ( maximum difference is 0.04@xmath2 for scenario 4 ) . increasing age and distance will also increase the median mass .
note that have recently determined the mass function for ic348 from a different survey and find a characteristic mass ( in the lognormal mass function ) of 0.21 - 0.22@xmath2 consistent with the values derived here .
in sect . [ s3 ] we established that differences in the mass distributions of the two clusters are significant , if ngc1333 is closer than ic348 . in addition , there is tentative evidence that the star / bd ratio in ic348 is slightly larger than in ngc1333 . given that our two target regions differ in object density by a factor of 4 - 7 ( sect .
[ s22 ] ) , this finding can in principle be used to put constraints on theories for bd formation in which the stellar density is a critical parameter for the yield .
one popular scenario to form bds is as part of dynamical cluster formation . here
, very low mass objects are removed from their accretion reservoir by dynamical ejections and thus stop their growth ; the final mass is set by the competition between accretion and ejection @xcite . in this model the efficiency of bd formation is partly controlled by the likelihood for dynamical encounters which is related to the object density .
the most recent radiation - hydrodynamical simulations by bate ( see their table 1 ) are comparable to the clusters studied here in terms of initial cloud mass and number of objects produced .
the simulations yield star / bd ratios ( @xmath47 , @xmath48 ) and median masses ( 0.21 , 0.24@xmath2 ) that are consistent with our empirical results . however , the impact of object density is difficult to judge , since the simulations have only been carried out for a very limited set of initial conditions .
another way to form bds that has been suggested in the literature is gravitational fragmentation of infalling gas into a stellar cluster @xcite . here
the potential well , and thus the object density in the cluster , is a critical parameter for the efficiency of bd formation .
the bds and very low mass objects are expected to be formed preferentially in regions with high stellar density .
qualitatively the predictions from this scenario are confirmed by our analysis : under plausible assumptions , the denser cluster ngc1333 has indeed a larger fraction of very low mass stars and brown dwarfs ( see sect .
[ s34 ] ) .
their figure 7 shows the bd fraction as a function of object density from their simulations .
the two regions investigated in the current paper are both at the low end of the considered densities ( 1 - 100pc@xmath49 ) . for these densities ,
the predicted bd fractions are between 7 and 13% . in their paper the bd fraction is calculated as the number of bds divided by the total number of objects . from our mass distributions , this quantity is @xmath50% in ic348 and @xmath51% in ngc1333 , i.e. the predicted values are lower than the observed one .
if this formation mechanism plays a role and the predictions are realistic , it could only contribute about one third to half of the bds in the clusters . other mechanisms , for example disk fragmentation followed by the ejection of embryonic or proto - brown dwarfs @xcite , could contribute to the final tally of substellar objects in the young clusters
. judged by their figure 7 , an increase in the object density by one order of magnitudes would result in an increase in the bd fraction by a factor of about 2 .
this is consistent with the observed difference between ic348 and ngc1333 , although such a difference is , as explained in sect .
[ s34 ] , still within the uncertainties .
therefore , the scenario remains viable , but can not be rigorously verified with the current surveys .
an important test for the theory would be the measurement of the bd fraction in a cluster that is significantly denser than ngc1333 , such as rcw38 @xcite or the orion nebula cluster .
so far , the survey results in the onc give inconsistent answers regarding the frequency of very low mass objects . the aforementioned scenario for brown dwarf formation via disk fragmentation could also result in a star / bd ratio that depends on stellar density , if some of the fragmentation processes are driven by stellar encounters @xcite or disk - disk collisions @xcite .
stellar encounters could also facilitate the ejection of bound brown dwarf companions from their host stars . with these additional mechanisms
, disk fragmentation models would again produce more brown dwarfs in a region with higher stellar density , which is qualitatively what we find to be the case .
however , the expected magnitude of this effect has not been estimated yet . as pointed out in sect .
[ s21 ] , the current stellar densities in ic348 and ngc1333 are probably representative of their primordial densities @xcite .
with constant star formation rate , these should scale with the gas density in the original cloud . under these assumptions ,
we can also put limits on scenarios for brown dwarf formation through turbulent fragmentation . according to the model presented by @xcite ,
a factor of 5 in density enhancement should amount to a very large increase ( about an order of magnitude , see their fig .
1 ) in the number of brown dwarfs . in the gravoturbulent picture @xcite
the effect seems to be similar .
qualitatively the result is as seen in the perseus clusters ( i.e. the denser cluster produces more brown dwarfs ) , but the magnitude of the effect is much larger than what we derive .
however , differences in other cluster parameters , for example in the mach number , could partially erase the predicted effect .
since their predictions depend heavily on initial conditions , it is doubtful whether empirically derived imfs can provide a meaningful test for these models .
an important caveat in our analysis is the fact that what we derive is a snapshot of the mass distribution , which may not necessarily represent the imf .
this is particularly relevant because ngc1333 is at an earlier evolutionary state than ic348 and might become as rich as its sibling at the other side of the perseus star forming complex @xcite . in the typical picture of cluster formation , however , lower mass objects form later ( e.g. * ? ? ? * their fig .
8) , thus , if additional formation processes in ngc1333 have any effect on the mass distribution , they are expected to amplify the observed discrepancy with ic348 . for the comparison with the models quoted above , which typically only predict a mass distribution of cores , not an imf ,
this issue is not of practical relevance .
we present a systematic study of the mass distribution in the two young open clusters ic348 and ngc1333 , with specific emphasis on the substellar regime .
these two regions are of specific interest because ngc1333 has a higher spatial density ( by a factor of 4 - 7 ) . in the following we list our most important findings . 1 . the mass distribution as well as the parameters derived from it , e.g. , the star / bd ratio @xmath0 or the median mass , is significantly affected by the choice of the isochrone used to estimated masses and the choice of the cluster parameters .
therefore , we caution against comparing imf parameters derived using different assumptions .
if ngc1333 is in fact closer to the sun than ic348 , as indicated by several independent studies , there is a significant difference in the mass distributions of these two clusters , in the sense that ngc1333 harbours a larger fraction of very low mass stars and brown dwarfs .
the star / bd ratio in ngc1333 is 1.9 - 2.4 in ngc1333 , consistent with previous estimates , and 2.9 - 4.0 in ic348 , significantly lower than in previous estimates .
the combined uncertainty in these values is approximately @xmath36 , but can be lowered with more accurate distance estimates and age estimates .
if confirmed , these values would point to a larger fraction of brown dwarfs in ngc1333 .
these results ( 2 and 3 ) indicate that the relative number of very low mass objects in a star forming regions may depend on the stellar density , in the sense that regions with higher density ( such as ngc1333 ) produce more very low mass objects . at this point , this conclusion is only based on two clusters and needs to be verified in other regions .
we thank nickolas moeckel , gilles chabrier , and francesco palla for helpful suggestions regarding topics discussed in this paper .
as acknowledges financial support through the grant 10/rfp / ast2780 from the science foundation ireland .
additional support for this work came from grants to rj from the natural sciences and engineering research council of canada .
, f. , homeier , d. , & freytag , b. 2011 , in astronomical society of the pacific conference series , vol .
448 , 16th cambridge workshop on cool stars , stellar systems , and the sun , ed . c. johns - krull , m. k. browning , & a. a. west , 91 | the abundance of brown dwarfs ( bds ) in young clusters is a diagnostic of star formation theory . here
we revisit the issue of determining the substellar initial mass function ( imf ) , based on a comparison between ngc1333 and ic348 , two clusters in the perseus star - forming region .
we derive their mass distributions for a range of model isochrones , varying distances , extinction laws and ages , with comprehensive assessments of the uncertainties .
we find that the choice of isochrone and other parameters have significant effects on the results , thus we caution against comparing imfs obtained using different approaches .
for ngc1333 , we find that the star / bd ratio @xmath0 is between 1.9 and 2.4 , for all plausible scenarios , consistent with our previous work . for ic348 , @xmath0 is found to be between 2.9 and 4.0 , suggesting that previous studies have overestimated this value .
thus , the star forming process generates about 2.5 - 5 substellar objects per 10 stars .
the derived star / bd ratios correspond to a slope of the power - law mass function of @xmath1 for the 0.03 - 1.0@xmath2 mass range .
the median mass in these clusters the typical stellar mass is between 0.13 - 0.30@xmath2 . assuming that ngc1333 is at a shorter distance than ic348
, we find a significant difference in the cumulative distribution of masses between the two clusters , resulting from an overabundance of very low mass objects in ngc1333 .
gaia astrometry will constrain the cluster distances better and will lead to a more definitive conclusion . furthermore , the star / bd ratio is somewhat larger in ic348 compared with ngc1333 , although this difference is still within the margins of error .
our results indicate that environments with higher object density may produce a larger fraction of very low mass objects , in line with predictions for brown dwarf formation through gravitational fragmentation of filaments falling into a cluster potential . |
the momentum scale dependence of the nucleon distribution amplitude is given by the brodsky - lepage evolution equation @xcite @xmath2 = \frac{c_b}{\beta } \int_{0}^{1 } [ dy ] v[x_i , y_i ] \tilde{\phi}(y_i,\xi ) , \label{evgl}\ ] ] with @xmath3 , @xmath4 and @xmath5 the bosonic and fermionic casimir - operators of @xmath6 , respectively ; @xmath7 being the gell - mann and low function .
the asymptotic solution of this equation is @xmath8 . to leading order in @xmath9 , the interaction kernel between quark pairs @xmath10
is given in ref . .
factorization of eq .
( [ evgl ] ) leads to @xmath11 v[x_i , y_i ] \tilde{\phi}(y_i,\xi)\ \ \text{and}\ \ x_1 x_2 x_3 \left [ \frac{\partial}{\partial \xi } \tilde{\phi}(x_i,\xi)\ , + \ , \gamma\tilde{\phi}(x_i,\xi ) \right ] = 0\ ] ] with @xmath12 and the integration measure is defined by @xmath13 \equiv \int_0 ^ 1 dx_1 \int_0^{1-x_1 } dx_2 $ ] @xmath14 .
[ in the following we use the operator @xmath15\ , v[x_i , y_i ] \mapsto \hat{v}$ ] , which commutes with the permutation operator @xmath16 . ]
the evolution behavior of the eigenfunctions comes from the sensitivity on the transverse momentum integration @xcite arising from the gluon exchange kernels .
it can be expressed in the form @xmath17 .
the representation of the evolution kernel is conveniently described in terms of appell polynomials , @xmath18 , which constitute an orthogonal basis with weight @xcite @xmath19 .
it was shown in ref . that in this basis , @xmath20 becomes block - diagonal with respect to different polynomial orders @xmath21 . because of @xmath22=0 $ ] , it is useful @xcite to define a symmetrized basis of appell polynomials @xmath23 the particular importance of this basis lies in the fact that @xmath20 is block - diagonal within a particular order for different symmetry classes ( @xmath24 ) with respect to @xmath25 . as a result , it is possible to analytically diagonalize @xmath20 up to order @xmath26 ( cf . tab .
[ tpoly ] ) [ up to order @xmath27 this is done in refs .
the eigenvalues of any order @xmath28 and for a specific symmetry class @xmath29 form a `` multiplet''-like set ( see fig .
[ fig1 ] ) , which follows an approximate power - law : @xmath30 that differs from that of scalar ( s=0 ) and vector ( s=1 ) mesons @xcite .
this observation is in contrast to the claims of ref . .
the eigenfunctions of the evolution equation have properties of a commutative group @xmath31 and obey the above triangle relation .
the structure coefficients @xcite @xmath32 of the group are calculated by @xmath33\ , x_1x_3(1-x_1-x_3)\,\tilde{\phi}_k(x_i ) \tilde{\phi}_n(x_i ) \tilde{\phi}_l(x_i)$ ] , with the particularly important case @xmath34 .
the evolution effect of the nucleon distribution amplitude is important for the calculation of various form factors at intermediate values of the momentum transfer @xmath35 ( see ref . ) . in this range
, @xmath36 tends to diverge whereas the @xmath37 evolution of the distribution amplitude , given by @xmath38 significantly reduces the value of the form factors by more than @xmath39 . in the above representation of @xmath40 ,
contributions of eigenfunctions up to polynomial order 3 have been studied @xcite . by incorporating on the rhs of eq .
( [ ansz ] ) ( c.f . ref . )
the factor @xmath41 , eigenfunctions of higher orders can be taken into account having recourse to a brodsky - huang - lepage type of ansatz .
this extended ansatz for @xmath42 and @xmath43 has been studied in ref . .
in order to determine its evolution behavior , one has to project @xcite @xmath40 on the eigenfunctions @xmath44 . for this purpose
, @xmath45 can be expanded in the basis of eigenfunctions of the nucleon evolution equation @xmath46\,x_1x_2x_3\ , f(x_i,\lambda_j ) \tilde{\phi}_k(x_i).\ ] ] this expansion and the properties of products of eigenfunctions given by the structure coefficients @xmath47 lead to @xmath48 in the last equation the evolution of the distribution amplitude is fully determined by the scale - dependence of the expansion coefficients @xmath49 , which are a complicated mixture of the original expansion coefficients @xmath50 and the projection coefficients @xmath51 of the extended ansatz .
in contrast to the assumption@xcite @xmath52 the evolution of the distribution amplitude is _ not _ found to be determined by the scale - dependence @xmath53 .
the deviation of the approximation @xcite from the correct evolution behavior can be expressed by the q - dependent ratio @xmath54 .
orthogonal eigenfunctions @xmath55 of the nucleon evolution equation ( represented by the coefficient matrix @xmath56 with @xmath57 ; @xmath58 for all n ) .
the normalization is given by @xmath59\ , x_1x_2x_3\ , \tilde{\phi}_k(x_i ) \tilde{\phi}_n(x_i ) = ( n_n)^{-1}\ , \delta_{kn}$ ] .
[ cols=">,^,>,^,^,^",options="header " , ] this deviation at intermediate @xmath37 is found to be of order @xmath60 for @xmath61 and increases exponentially for higher orders . since in the ansatz of ref .
@xmath62 , one can approximate @xmath63 using the power - law behavior for @xmath64 , fig .
[ fig2 ] shows the exponetially increasing deviation of the approximation of ref .
compared to the @xmath37-scaling behavior based on the renormalization group equation .
this comparison shows that it is important to project the nucleon distribution amplitude @xmath40 , as in the case of mesons @xcite , on the eigenfunctions of the evolution equation . for this purpose ,
powerful tools have been developed and the basis of eigenfunctions @xmath44 has been extended up to polynomial order 9 . using analytical and numerical algorithms developed in refs .
the calculation of higher order eigenfunctions up to any desired precision is shown to be possible .
1 g. p. lepage und s. j. brodsky , phys .
* d22 * ( 1980 ) 2157 .
k. tesima , phys . lett .
* 110b * ( 1982 ) 319 .
m. bergmann , ph.d .
thesis , bochum university ( 1993 ) .
m. bergmann und n. g. stefanis , bochum report rub - tpii-45/93 ( 1993 ) .
r. eckardt , j. hansper , und m. f. gari , z. phys .
* a343 * ( 1992 ) 443 ; + _ ibid . _ * a341 * ( 1992 ) 339 .
a. schfer , phys . lett .
* b217 * ( 1989 ) 545 .
t. huang und q .- x .
shen , z. phys .
* d50 * ( 1991 ) 139 .
n. g. stefanis and m. bergmann , these proceedings . | we study the brodsky - lepage evolution equation for the nucleon and construct an eigenfunction basis by including contributions of up to polynomial order @xmath0 . by exployting the permutation symmetry @xmath1 of these eigenfunctions , a basis of symmetrized appell polynomials
can be constructed in which the diagonalization of the evolution kernel is considerably simplified .
the anomalous dimensions are calculated and found to follow a power - law behavior . as an application
, we consider the brodsky - huang - lepage ansatz .
an algorithm is developed to properly incorporate such higher order contributions in a systematic way .
= latex=6 true in = 8.5 true in 0.2 cm -1.5 cm 21.5 cm 15.5 cm 1.0 cm 2.6ex = 0 # 1#2#3#4#5#6 # 1 # 1 > # 1 < # 1 # 1#1^ # 1#2 ( c # 1 + # 2 ) /#10=1=0>1 0 - 0to 0.1em11 - 1to 10.1em = cmcsc10 scaled1 = cmbx10 scaled1 |
it is almost twenty years to search the exotic hadrons such as the glueballs and hybrids .
there are some special states which are regarded as the candidates of hybrids , especially the @xmath10 and @xmath11 have been studied widely , but no confirmation has been made so far .
recently , these two special states arouse great interest again .
the e852 collaboration at bnl@xcite has reported a @xmath12 isovector resonance @xmath13 in the reaction @xmath14 .
they also reported the mass @xmath15 mev and width @xmath16 mev .
the crystal barrel collaboration has also claimed to find an evidence in @xmath17 annihilation which may be resonance with a mass of @xmath18 mev and a width of @xmath19 mev@xcite .
the confirmation of these states will provide some evidence for the existence of hybrids . at present , all the experiences specialize on the light quark hybrids for some reasons , but it s necessary to extend the energy region to hybrids including b or c quark , which is also possible in the b or @xmath20 factory . theoretically , the spectrum and decay width of hybrids have been calculated widely with many methods such as bag model@xcite , flux - tube model@xcite , qcd sum rules@xcite , lattice @xcite and other models@xcite . however , there are few works about the spectrum and decay width of heavy - light hybrids@xcite .
hqet has led to much progress in the theoretical understanding of the properties of hadrons@xcite , one may ask a question whether it is suitable in the heavy - light hybrids . for the heavy - light hybrids ,
the sum rules calculation in full qcd theory in ref.@xcite shows that the component of gluon gives a contribution more than @xmath21 to the mass , so the `` light freedom '' seems too heavy to keep the @xmath9 expansion available .
the calculation of the spectrum of hybrids in hqet will give an answer to the question .
our results show that the spectrum of heavy - light hybrids including @xmath22 or @xmath23 quark are close to those in full qcd theory , it is suitable to deal with these hybrids within hqet . besides , compared with @xmath24 and @xmath25 hybrid , the spectrum and decay width of the heavy - light hybrid is easier to be calculated in hqet .
it is interesting for the experimentalists to search the exotic @xmath2 heavy - light hybrids , so theoretical determination of the property of these states is necessary and urgent . in full qcd theory
, the estimation in ref.@xcite showed that the sum rule for the mass of @xmath2 heavy - light hybrids had no platform at all .
the masses of @xmath2 hybrids were given under the main assumption that the contribution of gluon condensate is less than @xmath26 of the bare loop .
in hqet , the ambiguous situation has been improved greatly .
the surface of the @xmath27 versus borel variable @xmath28 varies little in a large region , which gives a good platform and determines the masses of the hybrids . according to the mit bag model@xcite
, the hybrids with the same @xmath29 have different internal interactions between the patrons which indicate that they are different states , that is to say , the gluon in hybrid can be in different tm(@xmath8 ) or te(@xmath3 ) mode . in order to predict the properties of them , we should choose suitable generating currents corresponding to these states for the calculation . in the case of light quark hybrids@xcite , these different @xmath0 states were found to have different masses . in heavy - light hybrids case
, the calculation shows that the mass split of the @xmath0 heavy - light hybrids with gluon in different mode is not large in the @xmath30 limit .
the mass of @xmath0 hybrids from two different currents , @xmath4 with gluon in tm(@xmath8 ) mode and @xmath6 with gluon in te(@xmath3 ) mode , is found similar .
though the mass of these two different @xmath0 hybrids is similar , their decay widths to @xmath7 final states are found different : the decay widths from current @xmath4 are about 86 mev or 16 mev corresponding to @xmath31 or @xmath32 final states , respectively , while the ones from current @xmath6 are 11 mev or 2.6 mev .
the decay widths of the @xmath0 and @xmath2 hybrids to @xmath7 have been calculated in ref.@xcite , where the decay constants of the hybrids were obtained from the formulae in full theory@xcite .
both the strong couplings and the decay constants are calculated in hqet in this paper , and the strong coupling of the @xmath0 hybrid obtained here is much larger than that in ref.@xcite .
the large difference is from the values of the decay constant calculated in two ways . in the calculation of the strong couplings ,
two - point function between the pion and vacuum is taken use of insteadly to avoid the ambiguity resulting from the double borel transformation for the three - point correlation function and the infrared problem in soft pion limit . for convenience ,
the calculation is kept in the leading order of @xmath9 expansion .
the paper is organized as follows .
the analytic formalism of hqet sum rules for the spectrum of hybrid is given in sec .
2 . in sec . 3
, we give the numerical results of the spectrum and decay constants of hybrid , the comparison of the spectrum with that in full qcd theory is given too . in sec .
4 , with the help of two - point correlation function between the pion and vacuum , the analytic formalism of hqet sum rules for the strong couplings of hybrids is derived and the numerical results of some decay widths were obtained . in the last section , we give the conclusion and discussion .
as we know in ref.@xcite , the gluon in hybrid can be in different mode , tm(@xmath8 ) or te(@xmath3 ) mode . to analyze these different hybrids ,
we must choose suitable generating current . for the spectrum of the @xmath0 and @xmath2 heavy - light hybrids with the gluon in tm(@xmath8 ) and te(@xmath3 ) mode ,
respectively , the interpolated current in hqet is chosen as @xmath33 where @xmath34 is the light quark field and @xmath35 is the heavy quark effective field , @xmath36 is the velocity of the heavy quark .
then , we construct the correlation function as @xmath37 where @xmath38 since the free heavy quark propagator in hqet is @xmath39 and the interaction of the heavy quark with the gluon field @xmath40 in the leading order of @xmath9 expansion is @xmath41 . then under the fixed - point gauge
@xmath42(which will be used throughout this paper ) , the full propagator of the heavy quark @xmath43 in the leading order of @xmath9 expansion is the same as the free one .
the freedom of heavy quark can be extracted out of the matrix element as a delta function , which facilitates the calculation . in the operator product expansion , we keep the perturbative term , two gluon condensate , three gluon condensate and two quark condensate .
the contribution of mixing condensate and higher dimension operators are negligible since their smallness.the feynman diagrams are shown in fig .
1 , where the double line represents the propagator of the heavy quark . after twice suitable borel transformation
, we obtain the @xmath44 and @xmath45 corresponding to the scalar and vector contribution , respectively @xmath46 where the light quark mass corrections are considered also in these formulae . as for the @xmath1 and @xmath3 hybrids ,
the current was chosen as @xmath47 the correlation function expanded is similar to the vector current case except for the opposite sign for the contribution of the quark condensates and their spectrum will be determined in a similar way . for the @xmath0 hybrid with the gluon in te(@xmath3 ) mode , the following current should be used to predict the mass @xmath48 the ope of the @xmath49 for this current has been carried out as @xmath50 in the chiral limit , the two gluon and two quark condensates vanish .
on the phenomenal side , the decay constants of hybrids , @xmath51 , are defined as @xmath52 where @xmath53 represent the masses of hybrids , @xmath54 is the polarization vector and the two different @xmath0 hybrids with gluon in tm(@xmath8 ) or te(@xmath3 ) mode is represented as @xmath55 and @xmath56 , respectively . so the correlation function read @xmath57 where the first term of the right side is the pole term resulting from lowest lying resonance contribution and the second term represents the contribution of the continuum state and higher resonances , @xmath58 is the continuum threshold . taking use of the dispersion relations for the correlation function to equate the both sides , we obtain @xmath59 after the borel transformation@xcite
, they are turned into @xmath60 where @xmath61 is the borel transformation variable .
so the @xmath27 can be determined as @xmath62 after the @xmath27 has been calculated , the decay constant can be carried out according to ( [ spect ] ) .
in this content , we will give the numerical results of the spectrum and decay constants of the hybrids . to proceed the process ,
the mass of the b and c quark are chosen as 4.7 gev and 1.3 gev , respectively , the condensates are chosen as @xmath63 and the scale of running coupling is set at the borel parameter @xmath61 . the continuum threshold is chosen as below in the calculation : @xmath64 gev for the @xmath1 and two @xmath0 hybrids from current @xmath65 and @xmath66 , and @xmath67 gev for the @xmath3 and @xmath2 hybrids .
we display our results in hqet and those in full qcd theory in table .
1 . in this table ,
the right two columns represent the mass of heavy - light hybrids calculated in full theory , the left represent the estimation in hqet and the bottom of this table represents the @xmath0 hybrids with the gluon in te(@xmath3 ) mode .
.[tab1 ] _ masses of heavy - light hybrids with different @xmath29 _ ( gev ) . [ cols="^,^,^,^,^,^,^ " , ] + to the processes ( [ processa ] ) and ( [ processb ] ) , the decay widths are given by the following formulae @xmath68 then in the @xmath69 , the numerical results of the decay widths read @xmath70 though the decay of hybrids appears to follow the @xmath71 selection rule , which means that the decay of hybrids to two s - wave mesons are suppressed @xcite , the selection rule is not absolute . in the flux tube and constituent glue models , it can be broken by wave function and relativistic effects , and the bag model predict that it is also possible that the excited quark loses its angular momentum to orbital angular momentum @xcite , the results obtained here support this idea .
the decay widths of the processes to @xmath7 final states for @xmath55 with gluon in tm(@xmath8 ) mode is much larger than those for @xmath72 with gluon in te(@xmath3 ) mode , the reason is that the final states in the later channels are in the p wave .
besides , the decay width of the @xmath73 obtained here is much larger than that we got in ref.@xcite , the difference is from the decay constant .
the decay constant of the @xmath0 hybrid we got there in full theory is much smaller than the one calculated above in hqet .
though the mass of these two @xmath0 hybrids is almost the same , the strong couplings to pion of them are different , so the decay widths of these two different @xmath0 hybrids are different .
the decay width of the @xmath74 are smaller than those corresponding to @xmath75 .
the physical reason about the difference of decay width of these two different @xmath0 hybrids is unknown to us yet , but the difference between these states provides a nice evidence that the decay property of these two @xmath0 hybrids with the gluon in different mode is different .
we calculate the spectrum of the @xmath0 , @xmath1 , @xmath2 and @xmath3 heavy - light hybrids with different currents in hqet , the results from current @xmath4 and @xmath5 are compatible to those in full qcd theory .
the calculation shows that the light freedom in heavy - light hybrids is not heavy enough to break down the @xmath9 expansion and it is suitable to apply hqet to heavy - light hybrid systems .
the sum rules for the masses of @xmath2 heavy - light hybrids have no platform at all in full theory , so the masses of them were given under some assumptions .
the ambiguity of these sum rules have been improved in hqet , which suggests the reasonableness of the assumptions in ref.@xcite in another way . in the calculation ,
the leading order @xmath9 expansion approximation is used and only the first two terms in ope are kept in our estimate , which will bring some errors .
besides , the decay constants will bring in large deviation too , it is necessary to determine them more precisely .
since the gluon in hybrids can be in different mode , the hybrids with the same @xmath29 but different gluon mode are in fact different states . in the case of light quark hybrids ,
the two different @xmath0 states have different masses definitely . in the heavy - light hybrids case ,
the masses of these two different states are found similar in the @xmath69 limit , however , the calculation shows that the decay widths in the processes of these two hybrids to @xmath7 final states are different .
the decay width of @xmath75 is found about 86(16 ) mev , while the decay width of @xmath74 is only 11(2.6 ) mev .
the strong couplings and decay widths of the heavy - light hybrids in the processes of @xmath76 are calculated too .
the large difference of the decay widths calculated here from those calculated in ref.@xcite lies on the large difference between the decay constants calculated in two different ways .
+ 20 d. r. thompson
_ et al_. ( e852 collab . ) , phys . rev
* 79 * ( 1997 ) 1630 .
_ et al_. ( crystal barrel collab . ) , phys .
lett . b*423 * ( 1998 ) 175 .
t. barnes , f. e. close and f.de viron , nucl .
phys . b*224 * ( 1983 ) 241 .
+ m. chanowitz and s. sharpe , nucl .
phys . b*222 * ( 1983 ) 211 . n. isgur , r. kokoski and j. paton , phys . rev
* 54 * ( 1985 ) 869 .
+ f. e. close and p. r. page , nucl .
phys . b*443 * ( 1995 ) 233 .
+ f. e. close and p. r. page , phys .
rev . d*52 * ( 1995 ) 1706 .
+ a. donnachie and p. r. page , phys .
rev . d*58 * ( 1998 ) : 114012 .
i. i. balitsky , d. i. dyakonov and a. v. yung , phys .
lett . b*112 * ( 1982 ) 71 .
+ j. i. latorre , p. pascual and s. narison , z. phys .
c*34 * ( 1987 ) 347 .
+ j. govaerts , l. j. reinders , p. francken , x. gonze and j. weyers , nucl .
b*284 * ( 1987 ) 674 .
+ tao huang , hongying and ailin zhang , eur .
j. c*8 * ( 1999 ) 465 . c. michael _ et al_. , nucl
. phys . b*347 * ( 1990 ) 854 ; phys .
lett . b*142 * ( 1984 ) 291 ; phys .
b*129 * ( 1983 ) 351 . + liverpool univ .
report lth-286(1992 ) ; proc . of the workshop on qcd : 20 years later(aachen , 9 - 13 june 1992 ) . t. barnes , f. e. close and e. s. swanson , phys . rev . d*52 * ( 1995 ) 5242 .
j. govaerts , l. j. reinders and j. weyers , nucl .
b*262 * ( 1985 ) 575 .
e. eichten and b. hill , phys .
lett . b*234 * ( 1990 ) 511 ; b*243 * ( 1990 ) 427 .
+ h. georgi , phys .
lett . b*240 * ( 1990 ) 447 .
+ n. isgur and m. b. wise , phys .
lett . b*232 * ( 1989 ) 113 ; b*237 * ( 1990 ) 527 .
+ a. f. falk , h. georgi , b. grinstein and m. b. wise , nucl .
b*343 * ( 1990 ) 1
. t. barnes , f. e. close and f.de viron , nucl .
b*224*(1983 ) 241 .
+ m. chanowitz and s. sharpe , nucl .
b*222*(1983 ) 211 .
tao huang , hongying and ailin zhang , hep - ph/9812430 , submitted to phys .
* d*. m. a. shifman , a. i. vainshtein and v. i. zakharov , nucl .
phys . b*147 * ( 1979 ) 385 .
+ m. neubert , phys .
rev . d*45 * ( 1992 ) 2451 ; phys .
rev . d*47 * ( 1993 ) 4063 . c. caso
_ it al_(particle data group ) , eur .
j. c*3 * ( 1998 ) 1 .
r. rckl , hep - ph/9810338 .
p. r. page , phys .
lett . b*401 * ( 1997 ) 313 ; phys .
lett . b*402 * ( 1997 ) 183 .
s. godfrey , j. napolitano , hep - ph/9811410 , submitted to rev . | the spectrum of the @xmath0 , @xmath1 , @xmath2 and @xmath3 heavy - light hybrids have been calculated in hqet .
the interpolated current of the hybrid is chosen as @xmath4 , @xmath5 and @xmath6 . some strong couplings and decay widths of the heavy - light hybrids to @xmath7 are calculated by using the qcd sum rules .
the mass of @xmath0 hybrid with gluon in tm(@xmath8 ) or te(@xmath3 ) mode is found similar , while their decay widths are different .
a two - point correlation function between the pion and vacuum is employed and the leading order of @xmath9 expansion is kept in our calculation . |
proton and @xmath0-decay @xmath10-value measurements provide important information on the nuclear mass surface far from the valley of @xmath11 stability .
there are many examples of long chains of @xmath0 decays between nuclear ground states which , if connected to a nuclide of known mass excess , would allow the mass excesses of nuclei beyond the proton drip line to be determined . often these nuclei are too short - lived or too weakly produced to be measured directly by precision methods such as schottky mass spectrometry or penning trap mass spectrometry .
ground state of @xmath12tl and the @xmath2 isomeric state in @xmath12tl .
the nuclides written in blue bold italic font have had their mass excesses measured directly @xcite .
the decay q - values , half - lives and branching ratios are indicated .
data not measured as a result of the present study are taken from refs .
those decays which have been observed or deduced for the first time or have been reassigned as a result of the present work are indicated by wider red arrows.,scaledwidth=46.0% ] a representative example of direct relevance to the present study is the work of poli _ et al . _
@xcite , who measured proton and @xmath0-particle emission from the @xmath1 ground state and @xmath2 isomeric state in @xmath12tl . as illustrated in fig .
[ chains ] , the daughter of @xmath12tl proton decays is @xmath13hg .
the @xmath0-decay @xmath10-values of the decay chain of @xmath13hg are known down to @xmath14er , which decays to @xmath15dy the mass of which is known from penning trap measurements @xcite .
the experimental data therefore allowed the mass excesses of the @xmath16 ground states and @xmath2 isomeric states of @xmath12tl , @xmath3au , @xmath17ir and @xmath7re to be deduced .
the mass excesses of @xmath14er , @xmath18yb and @xmath19hf were subsequently measured directly using schottky mass spectrometry @xcite and the values obtained were consistent with those deduced from the mass excess of @xmath15dy and the relevant @xmath0-decay @xmath10-values . in the present work , the known @xmath0 decay of @xmath5ta is shown to be correlated with the @xmath0 decays of the @xmath2 states in @xmath3au , @xmath17ir and @xmath7re , indicating that it also originates from a @xmath2-based state , rather than a @xmath1-based state as had been previously assumed @xcite .
this establishes the complete @xmath0-decay chain from the @xmath2 isomeric state in the proton - unbound @xmath12tl down to the corresponding state in @xmath20ho , which constitutes the ground state of this nuclide @xcite .
in addition , a previously unobserved radioactivity has been correlated with the @xmath0 decay of the ground states of @xmath3au and @xmath17ir indicating that this new decay originates from the @xmath1-based ground state of @xmath7re .
as the masses of both @xmath15dy @xcite and @xmath20ho @xcite have been measured , this study not only provides a cross check of the mass excesses deduced in ref .
@xcite but , perhaps more importantly , allows the @xmath10-value of the @xmath0 decay of the @xmath1-based state in @xmath5ta and the energy difference between this state and the @xmath2-based state to be deduced .
the @xmath3au nuclei were produced via fusion - evaporation reactions induced by the bombardment of a 0.5 mg/@xmath21 @xmath22mo target of 97 % isotopic enrichment with a beam of @xmath23sr@xmath24 ions .
the sr beam was provided by the k130 cyclotron of the accelerator laboratory of the university of jyvskyl with an energy of 392 mev for approximately 140 hours and 400 mev for approximately 145 hours with an average beam current of 150 ena .
the change in beam energy was performed with an aim to increase the production of @xmath3hg @xcite nuclei which was the primary aim of the experiment . at the target position ,
34 high - purity ge detectors , ten of eurogam phase 1 type @xcite and 24 clover detectors @xcite , were positioned to facilitate the detection of prompt @xmath6 radiation .
the recoiling fusion - evaporation residues ( recoils ) were separated from the unreacted beam using the ritu he - filled magnetic separator @xcite and were transported to the ritu focal plane where the great spectrometer @xcite was located . here
the recoils traversed an isobutane - filled multi - wire proportional chamber ( mwpc ) before implanting into one of two 300 @xmath25m - thick double - sided si strip detectors ( dssd ) .
the average rate of recoil implantations was found to be @xmath26 150 hz across both of the dssds .
in addition to the mwpc and the dssds , the great spectrometer comprised a box of 28 si pin diode detectors , a planar ge detector and four clover - type ge detectors .
the mwpc provided energy loss and ( in conjunction with the dssds ) time - of - flight information , which was used to separate the recoils from any residual scattered beam .
the signals from all detectors were passed to the total data readout aquisition system @xcite where they were time stamped with a precision of 10 ns to facilitate temporal correlations between the implantation of recoils in the dssds and their subsequent radioactive decays .
the data were analysed using the grain software package @xcite .
there are considerable experimental challenges in establishing decay correlations with low-@xmath27 members of @xmath0-decay chains .
firstly , the relatively long half - lives mean that correlations can become obscured by interfering radioactivities produced as a result of competing reaction channels .
secondly , low @xmath0-decay branching ratios mean that large samples of parent nuclei are typically needed to facilitate such studies .
however , these are generally produced with rather lower cross sections than their descendents thus further compounding the difficulties of studying long @xmath0-decay chains . in the present study , approximately 190,000 recoil-@xmath0(@xmath3au@xmath4 ) events
were produced in order to provide a sufficient sample of correlated @xmath17ir@xmath4 , @xmath7re@xmath4 and @xmath5ta@xmath4 @xmath0 decays .
the decay chain was analysed by selecting @xmath3au@xmath4 @xmath0 decays occurring within 100 ms of a recoil being implanted into the same dssd pixel .
all decays fulfilling this criterion are shown in fig .
[ isomer_alphas](a ) .
the only other conditions imposed on the search for subsequent decays were that they all had to have been observed in the same pixel and within 15 s of the implantation .
the energy of the @xmath0 particles corresponding to the decay of @xmath3au@xmath4 was found to be 6739(15 ) kev which is in good agreement with the error - weighted mean of previous measurements of this activity @xcite .
it is worth noting at this point that the dssds were calibrated using @xmath0 lines at 5407 , 6038 and 6315 kev corresponding to the known activities of @xmath28os @xcite ( not visible in fig .
[ isomer_alphas](a ) ) , @xmath29pt @xcite and @xmath30pt @xcite , respectively .
in addition to the energy measurement the difference in time between each recoil implantation and the detection of @xmath3au@xmath4 @xmath0 decays was recorded and the distribution was fitted using the least - squares method .
the half - life of this radioactivity was measured to be 12.2(1 ) ms which is consistent with previous measurements @xcite .
au@xmath4 @xmath0 decay occurring within 100 ms of a recoil implantation in the same pixel ; ( c ) all decays following the observation of both @xmath3au@xmath4 and @xmath17ir@xmath4 @xmath0 decays such that the former were detected within 100 ms of a recoil implantation in the same pixel ; ( d ) all fourth - generation @xmath0 decays following the observation of an initial @xmath0 decay within 100 ms of recoil implantation and the subsequent detection of the @xmath0 decays of both @xmath17ir@xmath4 and @xmath7re@xmath4.,scaledwidth=46.0% ] the first decays observed following the @xmath3au@xmath4 @xmath0 decays are shown in fig .
[ isomer_alphas](b ) .
the dominant peak in this spectrum has an energy of 6120(14 ) kev which is consistent with previous measurements of the @xmath0 decay of an isomeric state in @xmath17ir .
the half - life of this decay was measured using a similar method to the @xmath3au@xmath4 activity except in this case the difference in time between the @xmath3au@xmath4 and @xmath17ir@xmath4 @xmath0 decays were recorded .
a least - squares fit to this distribution yielded a half - life of 280(1 ) ms which is in good agreement with previous studies @xcite .
a branching ratio of 78(6 ) % was measured for the @xmath0 decay of this state which agrees well with the value of 84(8 ) % reported by poli _
et al . _
in addition to the main peak , a further two , relatively weak , peaks can be seen in fig .
[ isomer_alphas](b ) which have been identified as resulting from the @xmath0 decay of the ground states of @xmath3pt and @xmath30pt .
these nuclei represent the most prominent @xmath0 decays observed in this study and their presence in the spectra of fig .
[ isomer_alphas](b ) can be understood as a consequence of the incorrect correlation of @xmath3au@xmath4 @xmath0 decays with implanted @xmath3pt and @xmath30pt nuclei .
this also explains the appearance of a peak associated with the @xmath0 decay of @xmath3pt in fig.s [ isomer_alphas](c ) and ( d ) . shown in fig . [ isomer_alphas](c )
are those decays observed following the detection of a @xmath3au@xmath4 @xmath0 decay occurring within 100 ms of an implanted recoil and the subsequent detection of an @xmath17ir@xmath4 @xmath0 decay , all in the same pixel .
the dominant peak in this spectrum has a measured energy of 5520(6 ) kev which is in good agreement with the previous observations of the decay of @xmath7re@xmath4 @xcite . a half - life of 1740(60 )
ms was measured for this activity which is in good agreement with earlier work @xcite .
in addition , a branching ratio of 13(1 ) % was measured which agrees well with the value of 13(3 ) % established in earlier work @xcite .
[ isomer_alphas](d ) shows all decays which were preceded by an @xmath0 decay within 100 ms of the implantation of a recoil followed by the sequential observation of the @xmath0 decays of @xmath17ir@xmath4 and @xmath7re@xmath4 .
the dominant peak in this spectrum has an energy of 5142(6 ) kev .
the energy of this activity is in good agreement with the previously reported value for the @xmath0 decay of @xmath5ta ( 5148(5 ) kev @xcite ) .
the maximum - likelihood method @xcite was used to fit the time distribution between the detection of @xmath7re@xmath4 and @xmath5ta @xmath0 decays and has yielded a half - life of 4.5(11 ) s. this value is consistent with previous measurements @xcite . the branching ratio for the @xmath0 decay of this state
was found to be 7(3 ) % .
this represents the first occasion that this ratio has been reported as a result of an experimental measurement .
the measured value is in good agreement with the estimated branching ratio of 5 % @xcite .
previous measurements of the energy of the @xmath0 decay of the ground state of @xmath3au have yielded an error - weighted mean value of 6683(9 ) kev @xcite .
the proximity of this line to that corresponding to the decay of the @xmath3au isomeric state means that isolating the @xmath0-decay chain associated with the @xmath3au ground state is problematic . to aid
the analysis , prompt @xmath6 rays feeding the @xmath3au ground state were investigated .
[ gammas](a ) shows a background - subtracted spectrum of prompt @xmath6 rays observed at the target position in delayed coincidence with @xmath0 particles having an energy consistent with the decay of the ground state of @xmath3au .
the @xmath6-recoil-@xmath0(@xmath3au@xmath8 ) events were only considered if the @xmath0 particle was detected within 150 ms and in the same pixel as the implanted recoil .
three @xmath6 rays , of energies 207 , 327 and 726 kev , have been identified for the first time as feeding the ground state of @xmath3au .
a @xmath3au@xmath8 @xmath0-tagged @xmath31 matrix was also constructed and fig .
[ gammas](b ) shows those @xmath6 rays observed in coincidence with the 327 kev transition .
although this spectrum demonstrates that the 327 and 726 kev transitions are mutually coincident , it has not been possible on the basis of the present data to construct a level scheme
. -ray spectrum in delayed coincidence with @xmath3au@xmath8 @xmath0 decay .
( b ) projection of a @xmath3au@xmath8 @xmath0-tagged @xmath31 matrix gated on a 327 kev transition.,scaledwidth=46.0% ] fig . [ ground_alphas](a ) shows decays occurring within 100 ms of the implantation of a recoil with the added condition that they were also in delayed coincidence with either a 327 or 726 kev @xmath6 ray detected at the target position .
the relatively low @xmath6-ray detection efficiency ( @xmath32 ) at the target position , combined with the background introduced by compton scattering , ensures that this gating technique is not sufficiently selective to isolate the @xmath3au@xmath8 @xmath0 decays but has the effect of enhancing them ( compare figs .
[ isomer_alphas](a ) and [ ground_alphas](a ) ) .
the @xmath3au@xmath8 @xmath0 decays have been measured to have an energy of 6688(14 ) kev and using the least - squares method the half - life of the activity was found to be 26.3(12 ) ms .
both the energy and half - life values measured here agree well with previous measurements @xcite .
rays of 327 or 726 kev ; ( b ) all second - generation decays following the detection of a @xmath3au@xmath8 @xmath0 decay occurring within 100 ms of a recoil implantation in the same pixel and with the same @xmath6-ray conditions as ( a ) ; ( c ) third - generation decays , subject to the same conditions as ( b ) but which were also preceded by an @xmath0 decay having an energy @xmath33 kev.,scaledwidth=46.0% ] fig .
[ ground_alphas](b ) shows those decays observed to follow @xmath3au@xmath8 @xmath0 decays which occurred in the same pixel and within 100 ms of the implantation of a recoil .
the two peaks in this spectrum correspond to @xmath0 decays with energies of 6019(14 ) and 6120(14 ) kev .
these values are consistent with the previously reported energies for the decay of the ground state and an isomeric state of @xmath17ir , respectively .
a low - energy tail on the @xmath3au@xmath4 peak , arising from a combination of escaping @xmath0 particles and the effects of radiation damage on the si detectors , results in a spectrum which contains both @xmath17ir@xmath8 and @xmath17ir@xmath4 @xmath0 decays .
the number of @xmath17ir@xmath4 @xmath0 decays present in fig .
[ ground_alphas](b ) is consistent with the number of @xmath3au@xmath4 @xmath0 decays included in the gate used to identify the @xmath0 decay of the ground state of @xmath3au .
the half - life of the radioactivity associated with the decay of the @xmath3au ground state was measured to be 570(30 ) ms is in good agreement with previous studies @xcite .
the branching ratio for the @xmath0 decay of this state was found to be 57(9 ) % , which is in good agreement with the value of 45(15 ) % obtained in previous studies @xcite .
third - generation @xmath0 decays which follow the decay of both @xmath3au@xmath8 and @xmath17ir@xmath8 and are in delayed coincidence with 327 or 726 kev prompt @xmath6 rays , are shown in fig .
[ ground_alphas](c ) .
the two peaks visible in this spectrum correspond to @xmath0 decays of energies 5520(6 ) and 5556(6 ) kev .
the former is in good agreement with the previously reported energies of the decay of @xmath7re@xmath4 .
similarly to the situation discussed above , the number of @xmath7re@xmath4 @xmath0 decays observed in fig .
[ ground_alphas](c ) is consistent with the number of @xmath17ir@xmath4 included in the gate used to identify the ground state decays of @xmath17ir .
the second peak is a previously unreported acitivity .
the clear correlation of this new activity with the decays of both @xmath3au@xmath8 and @xmath17ir@xmath8 leads to its assignment as the @xmath0 decay of the ground state of @xmath7re . to determine the half - life of this new activity , the @xmath34ir@xmath8)-@xmath35re@xmath8 ) time spectrum
was fitted using the maximum - likelihood method @xcite and this yielded a value of 1.6(6 ) s. a branching ratio of 14(8 ) % has been measured for the @xmath0 decay of this state .
prior to the undertaking of this work , the known @xmath0 decay of @xmath5ta was assumed to be the result of the decay of the ground state @xcite .
however , from the results presented above it is apparent that the 5142(6 ) kev @xmath0 decay is in fact the decay of the high - spin isomeric state in @xmath5ta . the deduced @xmath9 value of 5273(6 ) kev is plotted in fig .
[ q - values](a ) and can be seen to continue the near - linear trend of the decreasing @xmath9 values with increasing neutron number .
-decay @xmath10 values of odd-@xmath27 , even-@xmath36 nuclei relevant to the discussion of the present work . in panel
( a ) is shown the @xmath9 values for high - spin states ( associated with @xmath2 orbital ) while in ( b ) the @xmath9 values for the low - spin states ( @xmath1 ) are shown .
the symbols representing the @xmath9 values determined as a result of the present work have been colored red and enlarged .
data not measured as a result of the present work are taken from refs .
@xcite.,scaledwidth=46.0% ] the @xmath9-value for the previously unreported decay of the ground state of @xmath7re is plotted in fig .
[ q - values](b ) .
this value appears to fit very well with the linear trend already established by the neighboring re isotopes . in combining this new measurement with the @xmath0-decay @xmath10 values of fig .
[ chains ] and the mass excess of @xmath18yb reported by litvinov _
( - 53283(28 ) kev @xcite ) , it is possible to determine the mass excess for the ground state of @xmath5ta : - 38816(40 ) kev . as a result of the observation of the @xmath0 decay of @xmath7re@xmath8 in the present work , it is possible to determine the excitation energy of the high - spin state in @xmath5ta which is given by : @xmath37 using the @xmath0-decay energies reported here and the excitation energy of the @xmath0-decaying isomeric state in @xmath3au , 214(23 ) kev as reported by poli _
et al . _
@xcite , it has been determined that the high - spin state in @xmath5ta has an excitation energy of 95(38 ) kev . taking this analysis one step further
, the knowledge of the energy difference of the two @xmath0-decaying states in @xmath38lu , 26(7 ) kev @xcite , allows the @xmath9-value of the unobserved decay of the ground state of @xmath5ta to be determined : @xmath9(@xmath5ta@xmath8 ) = 5204(39 ) kev .
this new value is plotted in fig .
[ q - values](b ) where once more it fits well with the trend established by the neighboring ta isotopes . using the measured mass excess for @xmath18yb and the @xmath0-decay @xmath10 values of fig .
[ chains ] , the mass excesses of the ground and isomeric states in @xmath20ho can be deduced .
the deduced mass excess of the high - spin state in @xmath20ho is -61648(40 ) kev which agrees remarkably well with the directly measured value of litvinov _ et al . _
@xcite of -61646(31 ) kev .
the mass excess deduced for the low - spin state of @xmath20ho was found to be -61582(58 ) kev which is in - line with expectations based on the previously known 49 kev excitation energy of the @xmath1-based isomer in @xmath20ho @xcite .
the mass excesses deduced in the present work are compared with the values reported in the most recent atomic mass evaluation ( ame2012 ) @xcite in fig .
[ masses ] .
overall , there is very good agreement between the values obtained in this study and those in the evaluation with the deduced mass of @xmath5ta being the notable exception .
this discrepancy is possibly the result of the inclusion of the incorrectly assigned @xmath0-decay of @xmath5ta . indeed ,
if the 69 kev energy difference between the @xmath9 values of @xmath5ta@xmath4 and @xmath5ta@xmath8 is taken into account then the difference between the mass reported here and the ame2012 value is similar to those found for the other five nuclides plotted in fig . [ masses ] .
the consistency in the mass measurements indicated by the agreement between the deduced masses of the ground and isomeric states of @xmath20ho and the masses measured in ref .
@xcite suggests that all of the @xmath0 decays proceed between ground states with no electromagnetic decays occurring at any points in the decay chain between @xmath12tl and @xmath20ho .
this is indicative that the single - particle configurations , established as @xmath2 and @xmath1 in the heavier members of the decay chain , are also consistent down the entire decay chain .
this conclusion is supported by the reduced width measurements , calculated using the rasmussen formalism @xcite and assuming @xmath39-wave emisssion , which are listed in table [ reduced_widths ] .
the reduced widths measured in the present work have been compared to the value corresponding to the @xmath0 decay of the ground state of @xmath40po .
these hindrance factors , also listed in table [ reduced_widths ] , are consistent with unhindered @xmath0 decays .
.reduced widths , @xmath41 , and hindrance factors , hf , for the @xmath0 decay of nuclei measured in the present work .
the hindrance factors have been measured relative to the ground state to ground state @xmath0 decay of @xmath40po . [
cols="^,^,^ " , ] in ref .
@xcite an extensive level scheme of excited states in @xmath5ta built upon a proposed @xmath42 state was reported . however , in that work it was not possible to establish whether this level or a @xmath43 level was the lowest - lying @xmath2 state .
the separation energy of the @xmath44 and the @xmath45 states in the neutron - deficient ta isotopes is observed to decrease from 99 kev in @xmath46ta @xcite , 71 kev in @xmath7ta @xcite to 45 kev in @xmath47ta @xcite .
extrapolating to @xmath5ta suggests the separation could be as low as @xmath26 20 kev in this nuclide .
this would be accommodated within the 40 kev uncertainty on the deduced mass excess for the high - spin state in @xmath20ho meaning that the question regarding the spin and parity of the @xmath2-based state in @xmath5ta can not be resolved by the present study .
indeed , it remains unclear whether the @xmath0-decaying isomer in @xmath5ta has @xmath44 or @xmath45 .
in summary , fusion - evaporation reactions have been used to populate states in @xmath3au .
gamma - ray transitions populating the ground state of @xmath3au have been identified .
in addition , the @xmath0-decay chains originating from the isomeric @xmath2 state and the @xmath1 ground state have been studied , culminating in the observation of the @xmath0 decay of @xmath5ta@xmath4 and @xmath7re@xmath8 , respectively .
as well as reporting a new activity in the decay of @xmath7re@xmath8 and confirming that the known @xmath0 decay of @xmath5ta is associated with the high - spin isomer , this work has enabled the relative energies of the @xmath0-decaying states in @xmath5ta to be established . in combining these new measurements with the information already available on @xmath38lu it has also been possible to deduce the @xmath9-value for the decay of the ground state of @xmath5ta . as a result of the present work @xmath48-values of -129(24 ) kev and -37(21 ) kev have been determined for the ground and isomeric states of @xmath5ta , respectively , indicating that these states are only just bound with respect to proton emission .
financial support for this work has been provided by the uk science and technology facilities council ( stfc ) and by the eu 7th framework programme integrating activities - transnational access " , project number : 262010(ensar ) and by the academy of finland under the finnish centre of excellence programme 2012 - 2017(nuclear and accelerator based physics research at jyfl ) .
tg acknowledges the support of the academy of finland , contract number 131665 .
the authors would like to express their gratitude to the technical staff of the accelerator laboratory at the university of jyvskyl for their support .
the authors would also like to thank charles reich for stimulating discussions . | the @xmath0-decay chains originating from the @xmath1 and @xmath2 states in @xmath3au have been investigated following fusion - evaporation reactions .
four generations of @xmath0 radioactivities have been correlated with @xmath3au@xmath4 leading to a measurement of the @xmath0 decay of @xmath5ta@xmath4 .
it has been found that the known @xmath0 decay of @xmath5ta , which was previously associated with the decay of the ground state , is in fact the decay of an isomeric state .
this work also reports on the first observation of prompt @xmath6 rays feeding the ground state of @xmath3au .
this prompt @xmath6 radiation was used to aid the study of the @xmath0-decay chain originating from the @xmath1 state in @xmath3au .
three generations of @xmath0 decays have been correlated with this state leading to the observation of a previously unreported activity which is assigned as the decay of @xmath7re@xmath8 .
this work also reports the excitation energy of an @xmath0-decaying isomer in @xmath5ta and the @xmath9-value of the decay of @xmath5ta@xmath8 . |
after introducing the supersymmetric quantum mechanics as a toy model for studying the supersymmetry breaking mechanism @xcite , it was applied for solving many problems in theoretical and mathematical physics @xcite .
the most recent applications of the supersymmetric quantum mechanics can be found in the dynamics of d - branes and black holes @xcite , in m - theory and matrix models @xcite , in the theory of integrable systems @xcite and fluid mechanics @xcite .
some time ago it was observed that the pure parabosonic @xcite ( and parafermionic @xcite ) systems possess the supersymmetry characterized by the nonlinear superalgebra .
such nonlinear supersymmetry takes place in the simple quantum mechanical system generalizing the usual superoscillator @xcite , and earlier it was revealed in a similar but particular form in the fermion - monopole system @xcite , and in the @xmath1-invariant systems of planar fermions @xcite and chern - simons fields @xcite .
the algebraic structure of the nonlinear supersymmetry resembles the structure of the finite @xmath2-algebras @xcite for which the commutator of any two generating elements is proportional to a finite order polynomial in them . as it was noted in ref .
@xcite , in a generic case under attempt of constructing the quantum analogue of the classical systems possessing the nonlinear supersymmetry one faces the problem of the quantum anomaly @xcite . to resolve this problem , in the present paper we investigate the nonlinear supersymmetry of one - dimensional systems at the classical and quantum levels .
this will allow us to reveal the unexpected very close relation of the nonlinear supersymmetry with associated quantum anomaly problem to the quasy - exactly solvable systems @xcite being related , in turn , to the conformal field theory @xcite .
the results will give also a new perspective on the known quantum supersymmetry characterized by the second order supercharges @xcite .
the paper is organized as follows .
section [ classics ] is devoted to the detailed investigation of the various aspects of the classical one - dimensional systems possessing the nonlinear supersymmetry of the most general form . in section 3
we consider their quantization , and reveal the two classes of anomaly - free quantum systems possessing the nonlinear supersymmetry of arbitrary order @xmath3 .
one of them turns out to be closely related to the quasy - exactly solvable systems and this aspect of the nonlinear supersymmetry is investigated in section 4 .
section 5 discusses the general quantum case of the @xmath4 supersymmetry in the context of anomaly - free quantization of the class of @xmath4 calogero - like supersymmetric systems found and analyzed in section 2 .
we show also how this @xmath4 supersymmetry can be used for constructing new exactly solvable systems . in section 6
the brief summary of the obtained results is presented and some open problems to be interesting for further investigation are discussed . in particular , we point out how the nonlinear supersymmetry can be generalized to the case of the two - dimensional classical and quantum systems .
in this section we investigate the classical supersymmetry of the most general form realizable in one - dimensional boson - fermion system .
we shall show that in generic case the supersymmetry is characterized by a nonlinear poisson algebra and includes the usual supersymmetry as a particular case . analysing the structure of the supersymmetry from the viewpoint of canonical transformations
, we shall observe the existence of the three essentially different classes : in the first class the parameter characterizing the order of superalgebra is subject to the classical quantization , in the second ( intermediate ) class the supercharges poisson bracket can be equal to any real nonnegative degree of the hamiltonian , whereas the systems of the third class allow ones the complete classical decoupling of the fermion from the boson degrees of freedom .
following ref .
@xcite , let us consider a non - relativistic particle in one dimension described by the lagrangian @xmath5 where @xmath6 are the grassmann variables , @xmath7 , @xmath8 , and @xmath9 and @xmath10 are two real functions .
the nontrivial poisson - dirac brackets for the system are @xmath11 and @xmath12 , and the hamiltonian is @xmath13 the latter generates the equations of motion @xmath14 the hamiltonian @xmath15 and the nilpotent quantity @xmath16 are the even integrals of motion for any choice of the functions @xmath9 , @xmath10 , whereas the odd quantities @xmath17 are the integrals of motion when the differential equations @xmath18 have the solutions being regular functions in the corresponding domain of the phase space defined by @xmath9 and @xmath10 .
it is obvious that such odd integrals can exist only for a special choice of the functions @xmath9 and @xmath10 .
let us investigate this question in detail and restrict ourselves to the physically interesting class of the systems given by the potential @xmath9 bounded from below .
such a potential can generally be represented in terms of a superpotential @xmath19 and real constant @xmath20 : @xmath21 the condition of regularity of @xmath22 at @xmath23 leads to the relation @xmath24 with some function @xmath25 .
having in mind that for the functions @xmath26 the substitution @xmath27 is equivalent to the complex conjugation , one can represent them as @xmath28 . then the hamiltonian and eq .
( [ tq=0 ] ) take the form @xmath29 @xmath30 where @xmath31 . in terms of the complex variables
@xmath32 eq .
( [ beqw ] ) is represented as @xmath33 general solution to eq .
( [ ez ] ) is given by @xmath34 where @xmath35 is an arbitrary function and @xmath36 .
the simplest case @xmath37 with @xmath38 corresponds to the holomorphic solution of eq .
( [ ez ] ) , @xmath39 , whereas the case @xmath40 with @xmath41 gives the antiholomorphic solution @xmath42 , both regular at @xmath43 .
if the superpotential @xmath19 is the unbounded function , one of the functions @xmath39 or @xmath44 is well defined on the whole complex plane only for @xmath45 , i.e. we have here some sort of classical quantization @xcite .
this simplest solution with @xmath46 corresponds to the nonlinear supersymmetry investigated in ref .
@xcite , and includes the usual linear supersymmetry with @xmath47 as a particular case . according to the definition ( [ qd ] ) , the functions @xmath26 are defined up to an additive nilpotent term proportional to , @xmath48 etc .
do not vanish and may be essential for preserving the supersymmetry . ]
this allows ones to represent the supercharges in the equivalent form @xmath49 we suppose here that like in the case @xmath50 , the function @xmath51 is chosen in the simplest form compatible with the requirement of regularity of the supercharges ( see below ) .
the supercharges ( [ qgen ] ) , the hamiltonian ( [ hwgen ] ) and the nilpotent integral @xmath16 form generally the nonlinear superalgebra @xmath52 in the case of unbounded superpotential @xmath19 , the dynamics of the system projected on the unity of the grassmann algebra is defined on the whole complex plane @xmath53 . in order to have the supercharges ( [ qgen ] ) to be well defined single - valued observables on @xmath53 , we have to impose the condition @xmath54 the function @xmath55 is supposed to be regular and can be represented in the form @xmath56 where @xmath57 is an arbitrary function , @xmath58 . as a consequence , the supercharge can be written as @xmath59 here we suppose that @xmath60 is nonnegative ; in the case of negative @xmath60 equation ( [ qkz ] ) with substitutions @xmath61 , @xmath62 gives the supercharge @xmath63 .
the exponential factor in eq .
( [ qkz ] ) can be removed by the transformation @xmath64 that is the canonical transformation with generating function @xmath65 obeying the differentiability condition @xmath66 in the case of general superpotential @xmath19 , the transformation ( [ ct ] ) can be used to reduce the classical hamiltonian ( [ hwgen ] ) to the most simple form , e.g. , to the form with @xmath67 , @xmath68 .
this gives the following equation for the function @xmath65 : @xmath69 though the equation ( [ geqw ] ) has exactly the form of that for @xmath70 with shifted @xmath55 ( see eq .
( [ beqw ] ) ) , there is a difference : we require for @xmath65 to be regular and single - valued function on the whole physical domain , whereas the same condition of regularity is imposed on the function @xmath71 , but not on @xmath70 . the general solution to eq .
( [ geqw ] ) is @xmath72 and its behaviour depends on the physical domain for @xmath73 defined , in turn , by the properties of the superpotential @xmath19 .
here it is necessary to separate the three different cases and the results can be summarized as follows . 1 .
the physical domain in terms of @xmath73 includes the origin ( @xmath74 , @xmath75 , @xmath76 ) .
this , in particular , corresponds to the case of unbounded superpotential with @xmath77 , @xmath78 . from the regularity of @xmath65 in such a domain
it follows that @xmath79 .
we assume that the function @xmath80 can be decomposed into the taylor series at @xmath81 . from the regularity of @xmath71
we arrive at the classical `` quantization '' condition @xmath82 , @xmath83 , and at the restriction ( [ tphi ] ) on the function @xmath80 .
thus , the most general form of the hamiltonian admitting the nonlinear supersymmetry is @xmath84n,\qquad k\in \mathbb z,\ ] ] whose associated supercharges have been described above . by the canonical transformation ( [ ct ] ) with @xmath85 ( supplemented by the transformation @xmath86 in the case @xmath87 ) we can always reduce the system with this hamiltonian and the supercharge ( [ qkz ] ) to the form of the supersymmetric system possessing the holomorphic supercharge : @xmath88 where @xmath89 .
the presence of the `` quantized '' , integer number @xmath60 in the hamiltonian ( [ hg ] ) means that the instant frequencies of the oscillator - like odd , @xmath90 , and even , @xmath91 , @xmath92 , variables are commensurable . only in this case
the regular odd integrals of motion can be constructed , and the factor @xmath93 in the supercharge
@xmath94 corresponds to the @xmath60-fold conformal mapping of the complex plane ( or the strip @xmath95 ) on itself ( or on the corresponding region in @xmath53 ) .
the physical domain is defined by the condition @xmath96 ( or @xmath97 ) and also includes the origin of the complex plane .
but unlike the previous case , there are no closed contours around @xmath98 . as a consequence , though the regularity of @xmath65 results in the same relation @xmath99 , no `` quantization '' condition appears from the regularity of @xmath71 .
the most general form of the hamiltonian and the supercharge is @xmath100n,\quad q^+=z^\alpha e^{i\int _ { \varphi _ 0}^\varphi r(\rho\cos\lambda)\,d\lambda}\theta^+,\ ] ] where we assume that @xmath101 , and the function @xmath102 is analytical at @xmath81 and .
the singularity in @xmath94 at the origin @xmath98 for @xmath103 is not physical and can be removed multiplying @xmath94 by @xmath104 , that results in changing the holomorphic function @xmath105 for the antiholomorphic function @xmath106 to be regular at @xmath107 .
after the canonical transformation ( [ ct ] ) with the function @xmath108 ( supplemented by the transformation @xmath86 for @xmath103 ) , the hamiltonian and the supercharge can be reduced to the form @xmath109 3 .
the physical domain is defined by the condition @xmath110 ( or @xmath111 ) , i.e. the origin of the complex plane is not included . in this case @xmath112 and @xmath113 are not related since the function @xmath65 admits in such a domain the terms proportional to @xmath114 . therefore , though the general form of the hamiltonian and the supercharge is @xmath115 by the canonical transformations ( [ ct ] ) with the function @xmath116 , one can reduce the hamiltonian to the form @xmath117 with trivial dynamics for the grassmann variables @xmath90 , @xmath118 , playing the role of the supercharges . thus , _ classically
_ the supersymmetry of any system with bounded non - vanishing superpotential has a `` fictive '' nature .
the obtained classification of the classical supersymmetric systems emerged from the aim to present the hamiltonian in the most simple form when the superpotential @xmath19 has the definite behaviour defining the type of the physical domain in the complex plane . on the other hand
, one can consider the supersymmetry from the viewpoint of the functional dependence of the hamiltonian on @xmath19 without specifying the superpotential s type .
then the hamiltonian and the supercharges @xmath119 give the intersection of the described three classes of the systems , and can be treated as the representatives of the more broad classes of the supersymmetric systems ( [ h1 ] ) , ( [ h2 ] ) and ( [ h4 ] ) , with which ( [ hg0 ] ) is related by the corresponding symplectomorphism . for ( [ hg0 ] )
the constant @xmath60 characterizes the degree of nonlinearity of the associated superalgebra and one can refer to it as to the system with @xmath60-supersymmetry . at the same time
, it is necessary to bare in mind that in the case of the superpotential of the third class all the systems ( [ hg0 ] ) with different @xmath120 are symplectomorphic to the system with @xmath121 .
the three types of supersymmetric hamiltonians ( [ h1 ] ) , ( [ h2 ] ) , ( [ h3 ] ) are defined up to the additive constant @xmath20 .
this arbitrariness could be used to present the potential @xmath9 in terms of other superpotential via the relation @xmath122 generally , the superpotentials @xmath19 and @xmath123 can correspond to different types of nonlinear supersymmetry .
then the natural question is : can such a transition change the type of supersymmetry when the form of the nilpotent term of the hamiltonian has been already fixed ?
in other words , the question is if the described classification of _ classical _ supersymmetric hamiltonians has an invariant sense .
the invariance of the classification is demonstrated in appendix . at the same time
, it is necessary to stress that the systems related by the canonical transformation have not to be equivalent on the quantum level due to the ordering problem and because the canonical transformations are , as a rule , nonpolynomial in momenta . for the sake of completness , let us discuss the lagrangian formulation for the @xmath60-supersymmetric system ( [ hg0 ] ) .
its lagrangian is @xmath124 in the hamiltonian formulation the supertransformations of the variables @xmath125 and @xmath6 are generated canonically by the supercharges : @xmath126,\\ \delta\theta^+&=\{\theta^+,\,q^-\}\eta^+ = { } - iz^k\eta^+,\quad \delta\theta^- = -\{\theta^-,\,q^+\}\eta^- = { } i\bar{z}{}^k\eta^-.\end{aligned}\ ] ] using the equations of the motion , we obtain the corresponding supertransformations at the lagrangian level : @xmath127 , \qquad \delta\theta^\pm={}\mp i\left(a^\pm\right)^k\eta^\pm,\ ] ] where @xmath128 .
the lagrangian ( [ l1 ] ) is quasi - invariant under these supertransformations : @xmath129.\ ] ] it is worth noting that on shell the commutator of the two supertransformations for the physical variables is proportional to the translation in time : @xmath130{\cal x}=-i2^k k e^{k-1}(x,\dot x , \theta^\pm)(\eta^-_1 \eta^+_2-\eta^-_2\eta^+_1 ) \frac
d{dt}{\cal x},\ ] ] where @xmath131 is @xmath125 or @xmath6 , and @xmath132 is the energy function of the system .
therefore , the supertransformations form an open algebra with the structure functions depending on the physical variables .
we have arrived at the three types of hamiltonians ( [ h1 ] ) , ( [ h2 ] ) , ( [ h3 ] ) , which after appropriate canonical transformations can be reduced to the more simple form with the associated supercharges represented in the holomorphic or antiholomorphic form . on the other hand ,
as it was mentioned above , the quantization of canonically equivalent classical systems can give in some cases the quantum systems with different types of supersymmetry , even of different order @xmath60 .
therefore , the search for other special representations for the hamiltonian and associated supercharges is important .
based on this remark , let us look for the representation in which the function @xmath133 defining the supercharges is the polynomial in @xmath134 of the degree @xmath60 , i.e. @xmath135 substituting this into eq .
( [ tq=0 ] ) , we obtain the recurrent equation @xmath136 where @xmath137 for @xmath138 and @xmath139 is assumed . due to the equation @xmath140
, one can fix @xmath141 .
then the part of the equations ( [ bn ] ) can be solved giving for @xmath10 and @xmath9 the relations @xmath142 to simplify the notation , we put @xmath143 and realize the change of the variables @xmath144 if the function inverse to @xmath145 does not exist globally , we can perform this transformation separately on each interval where the function @xmath145 is monotonic . under the transformation ( [ x(y ) ] ) , eq .
( [ bn ] ) acquires the form of the system of differential equations in the variable @xmath146 : @xmath147 a general solution to this system has @xmath148 real parameters .
if we know solution to the system ( [ phn ] ) , we could find the form of the functions @xmath149 , at least implicitly .
this means that the general solution to eq .
( [ bn ] ) depends on arbitrary function @xmath150 but not on its derivatives , and which can be called the superpotential .
one notes also that the system ( [ phn ] ) is invariant under the transformation @xmath151 if the functions @xmath152 obey the relation @xmath153 the solution corresponding to the holomorphic case is of this type .
the simplest case @xmath154 corresponds to the usual linear supersymmetry , and we turn to the case @xmath4 . for @xmath4 , from ( [ phn ] )
we obtain the equation for @xmath155 : @xmath156 this equation has the solution @xmath157 where @xmath158 is an arbitrary real constant . considering @xmath159 as a superpotential , one arrives at the supercharges of the form @xmath160\theta^\pm,\ ] ] which together with the hamiltonian @xmath161 + 2w'(x)n+v\ ] ] form the nonlinear superalgebra @xmath162 note that the hamiltonian ( [ calogeroh ] ) has the calogero - like form : at @xmath163 its projection to the unit of grassmann algebra takes the form of the hamiltonian of the two - particle calogero system .
the functions @xmath164 and hamiltonian ( [ calogeroh ] ) form the nonlinear poisson algebra @xmath165 which does not depend on the constant @xmath158 , and is reduced to the @xmath166 algebra at @xmath163 .
for @xmath167 the obtained @xmath4 supersymmetric system ( [ calogeroq ] ) , ( [ calogeroh ] ) is reduced to the @xmath4 supersymmetric system of the form ( [ hg0 ] ) characterized by the holomorphic form of the supercharge .
let us investigate the relationship of the system ( [ calogeroq ] ) , ( [ calogeroh ] ) with @xmath60-supersymmetry ( [ hg0 ] ) in the case @xmath168 .
first we show that for @xmath169 ( @xmath170 , @xmath171 ) the system ( [ calogeroh ] ) , ( [ calogeroq ] ) can be reduced to the linear ( @xmath154 ) supersymmetric system of the form ( [ hg0 ] ) with the corresponding holomorphic supercharge multiplied by some function of the hamiltonian . to show this we represent the potential of the hamiltonian ( [ calogeroh ] ) in the form @xmath172 without loss of generality we can consider @xmath173 and get the relation @xmath174 therefore , in terms of @xmath175 the function @xmath10 is represented as @xmath176 , \quad m(\tilde w{}^2(x))= \left ( \tilde w^2(x)+4\gamma\right)^{-1/2}.\ ] ] comparing this with the factor at the nilpotent term in hamiltonians ( [ h1 ] ) , ( [ h2 ] ) and ( [ h3 ] )
, we find that our system corresponds to the system ( [ hg0 ] ) with @xmath154 since the term with function @xmath177 can be removed by the appropriate canonical transformation .
therefore , we conclude that the system ( [ calogeroh ] ) is canonically equivalent to the @xmath154 supersymmetric system ( [ hg0 ] ) .
analogously , applying the canonical transformation ( [ ct ] ) to the system ( [ calogeroh ] ) with @xmath178 ( @xmath179 , @xmath180 ) , one can reduce it to the @xmath121 supersymmetric system with the hamiltonian of the form ( [ decouple ] ) . indeed , let us define the new superpotential @xmath175 by the relation @xmath181 where without loss of generality we assume that @xmath173 , and the constant @xmath182 has to be chosen to provide the inequality @xmath183 . as a result
, one can represent @xmath19 in terms of the new superpotential as @xmath184 where @xmath185 . expressing @xmath10 in terms of @xmath186
, we obtain @xmath187 , with the function @xmath188^{-1/2}]\cdot w^{-\frac 12}(\tilde w(x))$ ] to be regular due to the inequality @xmath173 .
therefore the nilpotent term can be removed from the hamiltonian by the canonical transformation ( [ ct ] ) since the generating function @xmath189 is regular in this case .
let us turn now to the next , @xmath190 case given by the system of equations @xmath191 substituting @xmath192 from the second equation into the first one , we arrive at the nonlinear differential equation of the second order @xmath193 which can equivalently be represented as the first order nonlinear differential equation @xmath194 when the integration constant @xmath195 , the solution to the equation is a root of the 4th order algebraic equation for @xmath196 : @xmath197 at @xmath198 , its solutions have a simple form : @xmath199 , @xmath200 .
the first solution corresponds to the holomorphic case with @xmath190 , while the second one does to the case of @xmath154 supersymmetry with the supercharge multiplied by the hamiltonian . even in the case
@xmath195 , @xmath201 , the corresponding solutions @xmath202 have a complicated form of solutions in radicals of the 4th order equation , whereas for @xmath203 we have not succeeded in finding any analytical solution of the nontrivial nature .
the same complications appear with finding the nontrivial solutions to the system ( [ phn ] ) for @xmath204 .
as we have seen in the previous section , the supersymmetry in classical one - dimensional system is defined by the arbitrary function @xmath19 and in general case the supercharges together with the hamiltonian form a nonlinear superalgebra . according to the results of ref .
@xcite on the supersymmetry in pure parabosonic systems , a priori one can not exclude the situation characterized by the supercharges to be the nonlocal operators represented in the form of some infinite series in the operator @xmath205 .
since such nonlocal supercharges have to anticommute for some function of the hamiltonian being a usual local differential operator of the second order , they have to possess a very peculiar structure - supersymmetry both the supercharges and hamiltonian have a nonlocal structure @xcite . ] . due to this reason , we restrict ourselves by the discussion of the supersymmetric systems with the supercharges being the differential operators of order @xmath60 .
classically this corresponds to the system ( [ hg0 ] ) with the holomorphic supercharges or to the systems discussed in section [ c2 ] . in ref .
@xcite it was observed that just in the simplest case of the superoscillator possessing the nonlinear @xmath60-supersymmetry and characterized by the holomorphic supercharges of the form ( [ hg0 ] ) with the simplest superpotential @xmath163 , the form of the classical superalgebra @xmath206 is changed for @xmath207 due to the quantum noncommutativity .
moreover , it was also observed that for @xmath208 in generic case we have a global quantum anomaly @xcite : the direct quantum analogue of the superoscillators loose the property of the conservation , @xmath209\neq 0 $ ] .
therefore , we arrive at the problem of looking for the classes of superpotentials and corresponding quantization prescriptions leading to the quantum @xmath60-supersymmetric systems without quantum anomaly .
this and next two sections are devoted to the solution of such a problem . in this and next subsections we consider the quantization of the nonlinear supersymmetry characterized by the holomorphic form of the supercharges ( [ hg0 ] ) .
as we shall see , the straightforward quantization without special quantum corrections leads to the rigid restrictions on the form of the superpotential @xmath19 .
let us fix the quantum supercharges in the holomorphic form corresponding to the classical @xmath60-supersymmetry , @xmath210 with @xmath211 with realization @xmath212 the operator @xmath16 has a sense of the fermionic quantum number and being normalized for @xmath213 is the projector onto the fermionic subspace , whereas the complimentary projector onto the bosonic subspace is @xmath214 . then choosing the quantum hamiltonian in the form ( [ hgen ] ) , from the requirement of conservation of the supercharges , @xmath215=0 $ ]
, we arrive at the equations @xmath216 therefore , the quantum system given by the hamiltonian @xmath217 possesses the nonlinear supersymmetry of order @xmath218 characterized by the holomorphic supercharges ( [ qf ] ) only when @xmath219 for any other form of the superpotential the nilpotent operators ( [ qf ] ) are not conserved that can be treated as a quantum anomaly .
below we shall see that the quantum anomaly for the system ( [ hq1 ] ) can be cured " for some superpotentials if to modify appropriately ( by @xmath213-dependent terms ) the supercharges .
it is necessary to stress that the relation ( [ hw ] ) fixing the form of the superpotential for @xmath218 has a purely quantum nature .
one also notes that with the prescription ( [ qn ] ) , @xmath9 given by eq .
( [ quadro ] ) plays the role of the total potential for the bosonic sector , @xmath220 , whereas the potential for the fermionic sector is @xmath221 .
the appearance in @xmath9 of the term proportional to @xmath213 can obviously be associated with the ordering ambiguity under construction of the operator @xmath16 : under the choice @xmath222 $ ] instead of ( [ qn ] ) , we again arrive at the hamiltonian ( [ hq1 ] ) , but the term linear in @xmath213 dissappears from @xmath9 . in
what follows we shall have in mind the quantum prescription ( [ qn ] ) .
the anticommutator of supercharges @xmath94 and @xmath223 gives the polynomial of order @xmath60 in @xmath15 , e.g. , for the simplest cases @xmath4 , @xmath190 and @xmath224 we have @xmath225 where for the sake of simplicity we have put @xmath226 and @xmath227 .
the family of supersymmetric systems ( [ w2 ] ) is reduced to the superoscillator at @xmath228 with the associated exact @xmath60-supersymmetry @xcite .
for @xmath229 the @xmath60-supersymmetry is realized always in the spontaneously broken phase since in this case the supercharges ( [ qf ] ) have no zero modes ( normalized eigenfunctions of zero eigenvalue ) .
it is worth noting that the nonlinear supersymmetry with quadratic superpotential ( 3.8 ) was found earlier in ref .
@xcite as a by - product in the context of the discussion of the non - renormalization theorem for supersymmetric theories .
let us take the supercharges in the form of polynomials of order @xmath60 in the oscillator - like variables @xmath230 ( [ a ] ) : @xmath231 where @xmath232 are real parameters have to be fixed .
if we treat the supercharges ( [ qp ] ) classically , then the condition of their conservation by the hamiltonian of the general form ( [ hgen ] ) results in @xmath233 i.e. we arrive again at the @xmath60-supersymmetric system ( [ hg0 ] ) with the arbitrary function @xmath234 playing the role of the superpotential .
however , as we shall see , quantum mechanically ansatz ( [ qp ] ) gives us a nontrivial family of @xmath60-supersymmetric systems related to the so called quasi - exactly solvable problems @xcite .
first one notes that the parameter @xmath235 can be removed by a simple shift of the superpotential both on the classical and quantum levels and we can put @xmath236 .
then , as in the case of the supercharges ( [ qf ] ) , the requirement of conservation of ( [ qp ] ) results in the hamiltonian ( [ hq1 ] ) as well as in some @xmath237 algebraic equations fixing the parameters @xmath238 , @xmath239 , @xmath240 , @xmath241 in terms of @xmath242 , whereas the condition ( [ hw ] ) for @xmath218 is changed now for @xmath243 this means that the supercharges ( [ qp ] ) contain only the one - parameter arbitrariness and the case of quadratic superpotential ( [ w2 ] ) is included here as a particular case corresponding to @xmath244 . for @xmath245
the solution of eq .
( [ ewexp ] ) acquires the exponential form @xmath246 where all the parameters @xmath247 are real , while the parameter @xmath248 is real or pure imaginary depending on the sign of @xmath249 , and for the sake of simplicity we put @xmath227 . in the limit
@xmath250 this superpotential is reduced to the quadratic form ( [ w2 ] ) via the appropriate rescaling of the parameters @xmath247 . for @xmath4
the anticommutator of the supercharges has the form @xmath251 where @xmath252 for @xmath218 the superalgebra for the system with superpotential ( [ wexp ] ) can be represented as @xmath253 in principle , the coefficients @xmath254 can be found explicitly since the system of linear equations arises for them .
the values of energy of the supercharges singlets are the roots of the polynomial on the right hand side of eq .
( [ p(h ) ] ) and to find them one has to solve the algebraic equation of the corresponding order .
in next section we analyse in detail the class of @xmath60- supersymmetric systems given by the superpotential ( [ wexp ] ) in the context of the partial algebraization of the spectral problem .
in the late eighties a new class of spectral problems was discovered @xcite .
it occupies the intermediate position between the exactly solvable problems and all others . according to ref .
@xcite , the quantum - mechanical system is quasi - exactly solvable or admits partial algebraization of the spectrum if its potential depends explicitly on the natural parameter @xmath255 in such a way that exactly @xmath255 levels in the spectrum can be found algebraically .
the unique nature of such systems is based on the hidden symmetry of the hamiltonian . the part of the spectrum that can be found algebraically is related to finite - dimensional representations of the corresponding lee group ( algebra ) . for one - dimensional systems the quasi - exact solvability
is associated with non - unitary finite - dimensional representations of the @xmath256 .
such representations are characterized by a parameter @xmath257 referred to as a `` spin '' , which can take integer and half - integer values .
the number of the eigenstates of the hamiltonian that can be found algebraically is equal to @xmath258 .
here we argue in favour of existence of the intimate relation between nonlinear supersymmetry and the partial algebraization scheme @xcite .
for example , if in a given system with the nonlinear supersymmetry of the order @xmath60 , @xmath259 there are @xmath60 singlets in the bosonic or fermionic sectors , i.e. @xmath60 zero modes of @xmath94 or @xmath260 , then the eigenvalues of the corresponding states are equal to @xmath261 .
then it is quite obvious that if such a system is not exactly - solvable , it admits the partial algebraization of its spectrum .
having in mind these preliminary comments , let us show that the supersymmetric system with the superpotential ( [ wexp ] ) can be related to some known families of quasi - exactly solvable problems .
let us put @xmath262 and consider the potentials @xmath263 admitting the partial algebraization of the spectrum @xcite . the potential ( [ quadro ] ) with the superpotential of the from ( [ wexp ] ) coincides with the potential ( [ e^x ] ) when @xmath264 or with the potential ( [ cos x ] ) when @xmath265 or with the potential ( [ cosh x ] ) when @xmath266 or with the potential ( [ cosh x ] ) when @xmath267 in the cases ( [ oddk1 ] ) and ( [ oddk2 ] ) , there is a complete correspondence with the quasi - exactly solvable potentials ( [ cos x ] ) and ( [ cosh x ] ) for odd @xmath60 only
. having in mind this observation , we first analyse in detail the correspondence between the nonlinear supersymmetry and the partial algebraization scheme for the case of quasi - exactly solvable potential ( [ cosh x ] ) . writing down this potential as @xmath268 we see that it has exactly the form of the potential ( [ quadro ] ) of the nonlinear supersymmetry system with the corresponding superpotential @xmath269 since the potential of the form ( [ cosh ] ) corresponds to the potential ( [ cosh x ] ) for odd @xmath60 only ,
let us investigate the @xmath60-supersymmetric systems with even @xmath60 and start from the case @xmath4 .
formally this corresponds to the spin " @xmath270 in the partial algebraization scheme @xcite for the potential ( [ cosh x ] ) . in this case @xmath271 and the supercharges read as @xmath272 where @xmath230 are defined in ( [ a ] ) .
zero modes of the supercharge @xmath94 belong to the bosonic sector and have the form @xmath273 where for definiteness we have supposed that @xmath274 .
the anticommutator of the supercharges is @xmath275 and we find that the linear combinations @xmath276 are the eigenfunctions of the bosonic hamiltonian , @xmath277 corresponding to the eigenvalues @xmath278 the fermionic hamiltonian @xmath279 is characterized by the potential @xmath280 which is the superpartner of the potential ( [ cosh ] ) . since this differs from ( [ cosh ] ) in the sign before the second term , the corresponding solutions in the fermionic sector in the case of @xmath75 can be obtained from the bosonic solutions with @xmath274 by a simple change @xmath281 .
therefore , we see that there are two bound states in the bosonic ( fermionic ) sector for @xmath274 ( @xmath75 ) , that means that the potential ( [ cosh ] ) is quasi - exactly solvable for @xmath4 as well .
as we shall see , the same conclusion is also true for any even @xmath60 .
let us consider the case @xmath190 corresponding to the spin @xmath282 . in this case
the supercharges @xmath283 acquire the form @xmath284 the zero modes of the operator @xmath94 are @xmath285 looking for the solutions to the equation ( [ h_b ] ) in the form of the linear combination of the zero modes , @xmath286 we obtain the following three ( not normalized ) eigenfunctions @xmath287 where @xmath288 .
the energies of these states are @xmath289 and the anticommutator of the supercharges is @xmath290 it is necessary to stress that though in the case @xmath190 the direct application of the partial algebraization scheme to the potential allows ones to find only two eigenstates and corresponding eigenvalues ( @xmath291 , @xmath292 ) , the concept of nonlinear supersymmetry gives the information on one more exact eigenstate ( @xmath293 ) of the hamiltonian .
the same is also valid for any odd @xmath60 : the algebraization scheme gives for the potential ( [ cosh ] ) the information on @xmath294 eigenvalues corresponding to the even ( in @xmath125 ) eigenstates , whereas the nonlinear supersymmetry allows ones to find in addition @xmath295 eigenvalues corresponding to the odd eigenfunctions .
the eigenfunctions and eigenvalues in the fermionic sector for @xmath75 can be obtained via the formal change @xmath296 , and we conclude that in the case @xmath190 there are three bound states in bosonic ( fermionic ) sector for @xmath274 ( @xmath75 ) .
now let us generalize the @xmath60-supersymmetry of the system given by the superpotential ( [ wsh ] ) for the case of arbitrary @xmath60 . using the explicit form of the supercharges for @xmath297
, one can suppose that the supercharges @xmath298 obey the recurrent relation @xmath299 let us prove the conservation of the supercharges ( [ qrec ] ) by the method of mathematical induction using this conjecture .
as the first step we suppose that for some given @xmath60 the supercharge is conserved .
thus , calculating @xmath300 $ ] , we obtain the relation @xmath301+kwa^-q_k^++\frac 14 \left(k-1\right)kw'q_{k-2}^+=0 .
\label{baserec}\ ] ] the hamiltonian @xmath302 is related to the @xmath303 as @xmath304 then we have @xmath305=\left[w',\,q_{k+2}^+ \right]+\left(k+2\right)wa^-q_{k+2}^++\frac 14 \left(k+1\right)\left(k+2\right)w'q_k^+.\ ] ] using the relations ( [ qrec ] ) and ( [ baserec ] ) , after rather cumbersome algebraic manipulations one can reveal that the expression vanishes . since the corresponding supercharges are conserved for @xmath306 , we conclude that the supercharges @xmath298 are conserved for any @xmath60 as well .
this also proves the conjecture ( [ qrec ] ) .
the relation ( [ qrec ] ) gives us the following form of the supercharges for arbitrary @xmath60 : @xmath307 using this representation , we find all the zero modes of the supercharge @xmath308 : @xmath309 since these modes belong to the bosonic sector , we can build @xmath60 eigenfunctions of the bosonic part of the hamiltonian as a linear combination of them : @xmath310 \displaystyle\sum_{n =- m}^{m-1}c_ne^{\left ( n+\frac 12\right)x},&\quad k=2 m .
\end{array } \right.\ ] ] if we put this into the corresponding stationary schrdinger equation , we arrive at the recurrent system of algebraic equations on energy @xmath311 and the coefficients @xmath254 @xmath312 where @xmath313 for @xmath314 , and @xmath315 where @xmath313 for @xmath316 or @xmath317 .
thus , we have demonstrated how the concept of nonlinear supersymmetry allows us to reduce the problem of finding the part of the spectrum for the potential ( [ cosh ] ) to the pure algebraic problem .
let us discuss shortly the potential ( [ cosh4 ] ) , to which the superpotential @xmath318 corresponds within the framework of the nonlinear supersymmetry . from the form of the superpotential one can immediately conclude that the potential ( [ cosh4 ] ) can be reduced to the form ( [ cosh ] ) just by rescaling the argument and the parameter @xmath319 . indeed , with the help of the relation @xmath320 after rescaling @xmath321 , @xmath322 , we arrive at the form ( [ cosh x ] ) .
but in the case of the potential ( [ cosh4 ] ) , there is a difference in comparison with ( [ cosh x ] ) . as we have mentioned above , according to ref .
@xcite , only the even eigenstates of hamiltonian with the potential ( [ cosh x ] ) can be found following the @xmath256 partial algebraization scheme , whereas for the potential ( [ cosh4 ] ) all the lowest @xmath258 states can be found within the framework of the very scheme .
but we see that though the potentials ( [ cosh x ] ) and ( [ cosh4 ] ) have different representations in the algebraization scheme , in reality they correspond to the same physical system .
moreover , we see that all the potentials ( [ e^x])-([cosh4 ] ) are particular cases of the potential ( [ quadro ] ) with the exponential superpotential ( [ wexp ] ) .
for the potential ( [ quadro ] ) the first @xmath60 states can be found algebraically with the help of the associated nonlinear supersymmetry . in other words ,
the nonlinear supersymmetry provides us with a universal point of view on the quasi - exactly solvable potentials ( [ e^x])-([cosh4 ] ) .
now let us discuss the superpotential ( [ wexp ] ) of the general form , which , as we have seen , allows ones to unify all the cases ( [ e^x])-([cosh4 ] ) .
the potential ( [ quadro ] ) with this superpotential has the form @xmath323 the superpartner of this potential can be obtained by the formal substitution @xmath324 .
consider in detail the simplest case @xmath4 .
the corresponding supercharge @xmath325 has the zero modes @xmath326 and for the equation @xmath327 one finds the two eigenfunctions @xmath328 with the energy @xmath329 in the fermionic sector , the zero modes of the supercharge @xmath223 are @xmath330 and all the corresponding formulas can by reproduced via the changes @xmath331 , @xmath332 . as it follows from ( [ ph ] ) and ( [ psi ] ) , in the case @xmath333
there are two bound states in bosonic ( fermionic ) sector when @xmath334 ( @xmath335 ) , whereas for @xmath336 there are no such bound states at all .
the corresponding @xmath4 superalgebra is written as @xmath337 using the obtained results for the potential ( [ cosh ] ) and realizing the substitution @xmath338 , we arrive at the supercharges obeying the recurrent relation @xmath339 from the consideration of the nonlinear supersymmetry with the superpotential ( [ wexp ] ) , it follows that for any @xmath60 the supercharges depend on the parameter @xmath340 and do not depend explicitly on other parameters of the superpotential .
this means that the supercharges are conserved and obey the same recurrent relation if we rescale the parameter @xmath249 to yield the relation @xmath341 .
then in the same way as for the potential ( [ cosh ] ) , one can look for the wave functions of the @xmath60 singlets in the form @xmath342
\displaystyle\sum_{n =- m}^{m-1}c_ne^{\left ( n+\frac 12\right)\omega x } , & \quad k=2 m .
\end{array } \right.\ ] ] the corresponding stationary schrdinger equation gives us the recurrent system of algebraic equations on the energy and on the coefficients @xmath254 : @xmath343 for odd @xmath60 , where @xmath313 for @xmath344 , and @xmath345 for even @xmath60 , where @xmath313 for @xmath316 or @xmath317 . excluding the coefficients @xmath254 from the recurrent systems , one can obtain the algebraic equation for @xmath311 of the order @xmath60 .
normalizing the corresponding polynomial to the form @xmath346 and substituting @xmath347 , one can get the exact form of the polynomial specifying the @xmath60-supersymmetry via the anticommutator of the supercharges @xmath283 . similarly to the case @xmath4 , one can find that if @xmath333 , then there are @xmath60 bound states in bosonic ( fermionic ) sector when @xmath334 ( @xmath335 ) .
if @xmath336 , such bound singlets do not exist at all . to conclude this section , we note that the choice of the parameters in the superpotential ( [ wexp ] ) @xmath348 leads to the exactly solvable morse potential @xcite @xmath349 where @xmath350 , @xmath71 are real positive parameters .
as it is well known , from the morse potential a vast number of famous exactly solvable potentials can be reproduced by the point canonical transformations and suitable limiting procedures @xcite .
this illustrates an intimate though indirect relation of the nonlinear supersymmetry not only to the quasi - exactly solvable systems but to the exactly solvable models as well .
in the subsection [ polynomcase ] , we have analysed the classical supersymmetry from the viewpoint of the polynomial in momentum structure of the supercharges . since for the second order supercharges
the solution was found by us in the general form , let us analyse this case at the quantum level too . on the other hand , there is no sense to analyse in such a context the orders higher then 2 since we do not know the general solution even at classical level .
different aspects of supersymmetry with the general second order supercharges were considered in ref .
@xcite . here
we discuss the general quantum case of the @xmath4 supersymmetry in the context of curing " the quantum anomaly problem and from the viewpoint of all the possible types of spectra .
then , as an example of application of the @xmath4 supersymmetry , we construct a new exactly solvable nontrivial quantum model associated with the infinitely deep well .
let us consider the second order supercharge of the general form @xmath351 and the most general quantum hamiltonian @xmath352 the condition @xmath353=0 $ ] partially determines the unknown functions in @xmath354 and @xmath15 : @xmath355 where @xmath158 and @xmath20 are real constants .
the function ( [ vk=2 ] ) plays the role of the potential for the bosonic sector . taking into account eqs .
( [ l2q])([b2q ] ) and identifying @xmath356 , we obtain the following most general form for the hamiltonian and the supercharge of the quantum @xmath4 supersymmetry : @xmath357 \label{hd}\\ q^+&=&\frac{1}{2 } \left[\left(\hbar\frac{d}{dx}+w\right)^2+\frac{c}{w^2 } -\delta(w)\right]\theta^+,\label{qd}\\ \delta&=&\frac{\hbar^2}{4w^2 } \left(2w''w - w'{}^2\right ) .
\label{wd}\end{aligned}\ ] ] looking at the quantum hamiltonian ( [ hd ] ) and supercharge ( [ qd ] ) , we see that their form is different from the direct quantum analogue constructed proceeding from the classical quantities ( [ calogeroh ] ) and ( [ calogeroq ] ) via the quantization prescription @xmath222 $ ] : the presence of quadratic in @xmath0 term ( [ wd ] ) in both operators @xmath15 and @xmath94 is crucial for preserving the supersymmetry at the quantum level .
therefore , one can say that the quantum correction ( [ wd ] ) cures the problem of the quantum anomaly since without it the supercharge would not be the integral of motion .
it is interesting to note that the quantum term of the form ( [ wd ] ) appears also in the method of constructing new solvable potentials from the old ones via the operator transformations @xcite . the supercharges @xmath94 and @xmath358 satisfy the relation @xmath359 being the exact quantum analogue of the first classical relation from eq . ( [ qqh ] ) .
( 280,100 ) ( -10,-10)(1,0)310 ( -10,110)(1,0)310 ( -10,-10)(0,1)120 ( 300,-10)(0,1)120 ( 0,10)(1,0)50 ( 10,0)(0,1)100 ( 25,-5)a ) ( 20,30 ) 4 ( 30,30 ) 4 ( 20,50 ) 4 ( 30,50 ) 4 ( 20,70 ) 4 ( 30,70 ) 4 ( 20,90 ) 4 ( 30,90 ) 4 ( 105,-5)b ) ( 80,10)(1,0)50 ( 90,0)(0,1)100 ( 100,10)4 ( 100,30 ) 4 ( 110,30 ) 4 ( 100,50 ) 4 ( 110,50 ) 4 ( 100,70 ) 4 ( 110,70 ) 4 ( 100,90 ) 4 ( 110,90 ) 4 ( 185,-5)c ) ( 160,10)(1,0)50 ( 170,0)(0,1)100 ( 180,10)4 ( 190,10)4 ( 180,30 ) 4 ( 190,30 ) 4 ( 180,50 ) 4 ( 190,50 ) 4 ( 180,70 ) 4 ( 190,70 ) 4 ( 180,90 ) 4 ( 190,90 ) 4 ( 265,-5)d ) ( 240,10)(1,0)50 ( 250,0)(0,1)100 ( 260,10)4 ( 260,30)4 ( 260,50 ) 4 ( 270,50 ) 4 ( 260,70 ) 4 ( 270,70 ) 4 ( 260,90 ) 4 ( 270,90 ) 4 this superalgebra defines the form of the spectrum of bounded states .
for the sake of definiteness , let us put the constant @xmath226 .
then in general the spectra of the @xmath4 supersymmetric system can be of the four types presented on fig .
[ k2sp ] . for @xmath178 , there are no singlets in the system and we have completely broken @xmath4 susy , see fig .
[ k2sp]@xmath360 .
this case corresponds , in particular , to the @xmath4 susy given by the quadratic superpotential discussed in section [ q1 ] . in the case
@xmath167 , there are three possibilities : * ( i ) * the completely broken phase ( there are no singlet states ) , see fig .
[ k2sp]@xmath360 ; * ( ii ) * there is one singlet state in either bosonic or fermionic sector , see fig .
[ k2sp]@xmath361 ; * ( iii ) * there are two singlet states with equal energy , one in bosonic and another in fermionic sector , see fig .
[ k2sp]@xmath362 .
the types of the spectra @xmath361 and @xmath362 are represented by the new exactly solvable model in the next subsection . in the case
@xmath169 , all the three mentioned types of the spectra , @xmath360 , @xmath361 and @xmath362 , can be realized and , in addition , another situation with two singlet states in one of the two ( bosonic or fermionic ) sectors can exist , see fig .
[ k2sp]@xmath363 .
examples of the type @xmath363 were represented above exhaustively by the supersymmetric systems associated with the exponential superpotential ( [ wexp ] ) . via the appropriate choice of the superpotential
, one can construct the @xmath4 supersymmetric systems associated with the exactly solvable potentials .
e.g. , the simplest choice of the superpotential @xmath163 leads to the famous ( two - particle ) calogero model .
the morse potential ( [ calogerop ] ) can also be reproduced and there are two possibilities to realize it here :
@xmath364 it is well - known that the usual linear supersymmetry allows ones to obtain new exactly solvable potentials from the given ones .
here we show that the @xmath4 supersymmetry can be exploited in the same way .
for example , one can find that in the case @xmath365 the potential ( [ vk=2 ] ) has exactly the form of the potential of the linear supersymmetry with the superpotential @xmath366 here and in what follows we put @xmath227 .
if the superpotential @xmath367 is given , we can consider this relation as differential equation . solving this equation , we obtain the one - parametric solution for @xmath4 superpotential @xmath356 . though the initial potential is the same for linear ( @xmath154 ) and @xmath4 supersymmetric systems , the corresponding superpartners are different : @xmath368 , while @xmath369 .
therefore , in the case @xmath4 one can construct the one - parametric family of isospectral potentials as superpartner to @xmath9 .
we will illustrate this by the example of the infinite square potential well . without loss of generality
, we can assume that the width of the potential is equal to 1 .
this problem is equivalent to the equation @xmath370 with the boundary conditions @xmath371 and we can think that @xmath125 runs over the segment @xmath372 $ ] only .
the eigenfunctions have the form @xmath373 with the energy @xmath374 where the energy of the ground state has been chosen to be equal to 0 .
let us consider the infinite square potential well as a bosonic potential of the system with the @xmath4 susy .
in order to obtain the superpartner of the potential , we have to find the superpotential @xmath375 .
for simplicity we put @xmath167 . then to find the superpotential
, we have to solve the equation @xmath376 the general solution has the form @xmath377 in order the superpotential to be well defined function on @xmath378 , we have to assume that the real constant @xmath379 can take any value in @xmath380 except the interval @xmath381 .
let us note that the superpartner in the framework of the usual ( @xmath154 ) supersymmetry is proportional to @xmath382 and has a pure trigonometric nature , while this is not the case for the @xmath4 susy superpartner .
the superpartner of the potential is defined as @xmath383 , and in this case it acquires the form @xmath384 acting by the supercharge @xmath94 of the form ( [ qd2 ] ) with the superpotential ( [ fdeep ] ) on the wave functions ( [ deepes ] ) of the square potential well , we obtain the eigenfunctions of the potential ( [ deepvf ] ) , @xmath385 given here up to a normalization constant @xmath386 .
one can verify that @xmath387 is identically equal to zero .
it is worth noting that the function @xmath388 has @xmath255 nodes when @xmath389 and @xmath390 nodes when @xmath391 or @xmath392 .
this means that in the latter case @xmath393 is the ground state of the potential ( [ deepvf ] ) , but this is not valid for the former case .
this situation is illustrated on figs .
[ pic12 ] , [ pic34 ] by the example of the function @xmath394 .
the arrows indicate direction of the node motion when @xmath379 goes to @xmath395 ( fig .
[ pic12 ] ) or to @xmath396 ( fig .
[ pic34 ] ) . for the case @xmath397
, the ground state can be found by calculating the limit @xmath398 .
having done this , one can verify that the function obtained in such a way , @xmath399 is indeed the ground state of the potential .
the function vanishes when @xmath391 or @xmath400 that corresponds to the statement above .
therefore , by a choice of the parameter @xmath379 we can obtain the unbroken @xmath4 susy of the two types , namely , with one singlet for @xmath391 or @xmath400 , and with two singlets with equal energies for other admissible values of @xmath379 .
thus , we have illustrated the special case of the @xmath4 supersymmetric system by the example of the infinite square potential well .
we have also shown that in the case @xmath167 , the two types of the spectra represented on fig .
[ k2sp]@xmath361 and @xmath362 can be realized .
the spectrum of the first type has exactly the form of that of a system with the usual supersymmetry in the exact phase .
the spectrum of the second type has rather unusual properties .
it has the form characteristic to a system with @xmath154 broken supersymmetry ( with coinciding two lowest energy levels ) , but here the two lowest states with equal energies are the true supersymmetric singlets and , therefore , the @xmath4 supersymmetry is unbroken .
let us summarize briefly the obtained results and then discuss some open problems that deserve further attention .
* classical supersymmetry is characterized by the poisson algebra being nonlinear in the hamiltonian , and includes the usual linear supersymmetry as a particular case . *
any supersymmetric system is symplectomorphic to the supersymmetric system of the canonical form with the holomorphic supercharges . *
the canonical supersymmetric systems are separated into the three classes defined by the behaviour of the superpotential . in the first class
the parameter @xmath112 characterizing the order of the superalgebra is subject to a classical quantization : @xmath401 ; in the second class it can take any non - negative value : @xmath402 ; in the systems of the third class the fermion degrees of freedom can be decoupled completely by a canonical transformation and , hence , their supersymmetry has a rather fictive nature .
* we have investigated the nonlinear supersymmetry with supercharges being polynomial of order @xmath60 in the momentum , and for @xmath4 have found the most general one - parametric solution of the calogero - like form ( [ calogeroh ] ) .
depending on the value of the parameter @xmath158 , classically the calogero - like @xmath4 supersymmetric system can be reduced to the @xmath121 , @xmath154 or @xmath4 supersymmetric system of a canonical form .
the quantization of a system with nonlinear supersymmetry is a nontrivial problem due to the quantum anomaly taking place in general case .
we have shown that * the anomaly - free quantization of the classical @xmath60-supersymmetric system with the holomorphic supercharges is possible for the superpotential of the quadratic ( [ w2 ] ) and exponential forms ( [ wexp ] ) only ; * the @xmath60-supersymmetric systems with the exponential superpotential are closely related to the well - known families of the quasi - exactly solvable systems @xcite ; * the @xmath4 supersymmetric calogero - like systems can be quantized in the anomaly - free way .
the problem of the quantum anomaly is cured " here by the specific superpotential - dependent term of order @xmath0 .
such a quantum term appeared earlier in the operator transformations method of constructing new solvable potentials from the known ones @xcite ; * the general @xmath4 supersymmetry associated with the calogero - like systems can be used for producing new exactly solvable potentials .
the supersymmetric systems ( [ hg0 ] ) given by the superpotentials of the third type are canonically equivalent to the system ( [ decouple ] ) with the completely decoupled fermion degrees of freedom . on the other hand ,
the @xmath154 quantum analogues of ( [ hg0 ] ) with superpotential of the third type are the systems with spontaneously broken linear supersymmetry . in this case
the nilpotent operators @xmath403 are well defined and their anticommutator is equal to @xmath404 , i.e. @xmath405 look like @xmath121 quantum supercharges .
however , these operators are not decoupled from the even operators @xmath125 and @xmath134 .
therefore , the question is whether the quantum analogue of the above mentioned canonical transformations exists .
if so , the quantum fermion degrees of freedom could be completely decoupled and the initial @xmath154 supersymmetric system would be reduced to the @xmath121 supersymmetric system in correspondence with the classical picture . on the other hand ,
if such a transformation does not exist ( at least for some superpotentials @xmath19 ) , we face a sort of quantum transmutation .
the classical equivalence of the calogero - like @xmath4 supersymmetric systems to the @xmath121 ( for @xmath178 ) and to the @xmath154 ( for @xmath169 ) supersymmetric systems of the canonical form ( [ hg0 ] ) is also based on the existence of the corresponding canonical transformations .
if for the quantum calogero - like @xmath4 supersymmetric systems ( [ hd])([wd ] ) the quantum analogues of the above mentioned canonical transformations do not exist , we again have a quantum transmutation .
the supersymmetric system ( [ hg0 ] ) represents the whole class of the symplectomorphic systems ( [ h1 ] ) .
however , the quantization apparently breaks the equivalence of the systems ( [ h1 ] ) with different functions @xmath57 .
therefore the quantization of the systems ( [ h1 ] ) may lead to different nontrivial quantum systems with @xmath60-supersymmetry .
most likely , the anomaly - free quantization could be possible only for some special cases of the function @xmath57 and the superpotential @xmath19 . in this context
it is necessary to stress that the quantum anomaly is specific not only for the nonlinear supersymmetry .
the quantization of the general system ( [ h1 ] ) with @xmath154 and @xmath406 gives rise to the quantum anomaly .
from this view - point one can say that the system of the form ( [ hg0 ] ) underlies the quantum linear supersymmetry since the quantization of such a system with arbitrary superpotential does not reveal any anomaly . however , it would be interesting to find at least one example of the quantum - mechanical system of the non - standard form ( [ h1 ] ) with the linear supersymmetry .
since for the nontrivial cases the quantum analogue of the transformation ( [ ct ] ) is non - local , the corresponding system has to operate with non - local objects .
it is possible that the supersymmetry of the pure parabosonic systems @xcite can be arrived at in this way .
we hope that further investigations will shed light on the relation between the classical and quantum supersymmetries in the context on the described hypothetical quantum transmutations and the quantum anomaly problem .
though we have considered the concept of the nonlinear supersymmetry for one - dimensional systems , it can be extended to the higher dimensional systems as well .
for example , the two - dimensional system describing the motion of a charged particle in external magnetic field admits the nonlinear supersymmetries .
the pauli hamiltonian of such a system is given by ( @xmath407 ) @xmath408 , where @xmath409 , and @xmath410 , @xmath411 , form the 2d vector gauge potential .
for the gyromagnetic ratio @xmath412 this system reveals the usual linear supersymmetry both at the classical and quantum levels @xcite .
it turns out that for @xmath413 the system possesses the @xmath60-supersymmetry . to show this
let us introduce the complex variables @xmath414 . in terms of these variables
the hamiltonian can be rewritten as @xmath415 . but
this is exactly the form of the hamiltonian ( [ hg0 ] ) with the variables @xmath73 , @xmath416 changed for @xmath417 , @xmath418 .
therefore , the hamiltonian commutes with the supercharges @xmath419 and , as a consequence , the system possesses the classical @xmath60-supersymmetry .
the quantization of this system reveals the properties similar to those for the @xmath60-supersymmetric system with the holomorphic supercharges : the requirement of conservation of the nonlinear supersymmetry leads to the restrictions on the form of the magnetic field @xmath420 .
one can show that as in the one - dimensional case , the form of the magnetic field turns out to be restricted by the quadratic case and by the sum of exponential functions in the variables @xmath421 @xcite .
we suppose that such systems are also closely related to the two - dimensional quasi - exactly solvable problems . to conclude
, we note that our investigation of the nonlinear supersymmetry was motivated by the existence of nonlinear supersymmetry in pure parabosonic systems @xcite , where the reflection ( parity ) operator @xmath422 plays the role of the @xmath423-grading operator .
such a special role of the reflection operator was the main idea in the construction of the minimally bosonized @xmath154 supersymmetric quantum mechanics @xcite , where the role of the superpotential is played by the arbitrary odd function .
therefore , the construction of the bosonized nonlinear supersymmetry is an attractive problem which we hope to consider elsewhere @xcite . when this paper was finished , the interesting papers @xcite devoted to the development of the @xmath424 nonlinear supersymmetry of ref . @xcite
have appeared .
we note that the generalization of the @xmath60-supersymmetry with the holomorphic supercharges found in ref .
@xcite and possessing the typical calogero - like structure of the @xmath4 supersymmetry from section 5 could also be treated in the context of the quasi - exactly solvable systems discussed here .
0.5 cm * acknowledgements * the work was supported by the grants 1980619 , 1010073 and 3000006 from fondecyt ( chile ) and by dycit ( usach ) .
here we show that the arbitrariness in the definition of the superpotential , @xmath425 can not be used to change the type of the classical supersymmetry of the given system , i.e. we demonstrate the invariance of the classification obtained in section [ c2 ] . for the purpose ,
it is enough to prove that the redefinition of the superpotential can not provoke a transition between the adjacent classes of supersymmetry .
let us start by analysing the possibility of the transition from the supersymmetry of the first type with the superpotential unbounded from below to the supersymmetry of the second type .
for the sake of simplicity we assume that the potential @xmath9 has finite number of minima , the lowest one is at the origin and is equal to zero .
for such a potential , the superpotentials of the first and second types obey the relation @xmath426 .
suppose that the bounded and unbounded superpotentials @xmath427 and @xmath428 with the associated supersymmetries of the first and second types correspond to the potential .
the superpotential @xmath428 is a continuously differentiable function at @xmath429 only when the potential has the local behaviour @xmath430 . indeed , if @xmath431 , the bounded superpotential locally is @xmath432 , and as a consequence , its derivative is not regular at the origin .
the case of the potential admitting the regular superpotential @xmath428 is presented on the fig .
[ pic ] . in this case
the bounded and unbounded superpotentials @xmath428 and @xmath427 coincide for positive values of @xmath125 and have different sign for negative @xmath125 : @xmath433 if we have the nonlinear supersymmetry of the first type , i.e. @xmath434 , @xmath435 , then rewriting this function in terms of @xmath436 we obtain @xmath437 .
the function @xmath10 is not good in terms of @xmath436 in the sense that there exist no canonical transformation that could reduce this system to that with the supersymmetry of the second type given by eq .
( [ h3 ] ) . the potential with several minima can be considered in a similar way .
the case of supersymmetry of the first type with a bounded superpotential merits the special analysis . after the transition ( [ trans ] ) to the superpotential @xmath186 of the second type
, the function @xmath10 can be rewritten as @xmath438 , @xmath439 . the function @xmath440 has singularities at the points where the initial superpotential vanishes .
therefore , the type of the supersymmetry can not be changed either .
now let us consider the possibility of transition from the supersymmetry of the second type to that of the third type by means of the change ( [ trans ] ) .
we suppose that the superpotential @xmath19 of a given system has a minimum equal to zero at the origin of coordinates .
the superpotential of the third type defined by @xmath441 has the minimum equal to @xmath442 at the origin . rewriting @xmath10 in terms of the new superpotential
, we obtain @xmath443 .
there exists no canonical transformation ( [ ct ] ) that could remove the corresponding nilpotent part of the hamiltonian since the function @xmath444 is singular at the origin .
the case of the superpotential with several minima can be treated in a similar way .
this completes our proof of the invariance of the obtained classification of the classical supersymmetries .
e. witten , _ nucl .
_ * b188 * ( 1981 ) 513 .
e. witten , _ nucl .
* b202 * ( 1982 ) 253 . l.e .
gendenshtein and i.v .
krive , _ sov .
_ * 28 * ( 1985 ) 645 .
f. cooper , a. khare and u. sukhatme , _ phys .
* 251 * ( 1995 ) 267 , hep - th/9405029 .
p. claus , m. derix , r. kallosh , j. kumar , p. k. townsend , a. van proeyen , + _ phys .
* 81 * ( 1998 ) 4553 , hep - th/9804177 .
j. michelson , a. strominger _ jhep _ * 9909 * ( 1999 ) 005 , hep - th/9908044 .
t. banks , w. fischler , s. h. shenker and l. susskind , _ phys .
* d55 * ( 1997 ) 5112 , hep - th/9610043 .
r. jackiw , a particle field theorist s lectures on supersymmetric , non - abelian fluid mechanics and d - branes , physics/0010042 .
r. jackiw and a. p. polychronakos , _ phys .
rev . _ * d62 * ( 2000 ) 085019 . m. plyushchay , _ int . j. mod .
* a23 * ( 2000 ) 3679 , hep - th/9903130 .
s. klishevich and m. plyushchay , _ mod .
_ * a14 * ( 1999 ) 2739 , hep - th/9905149 .
f. de jonghe , a. j. macfarlane , k. peeters and j. w. van holten , _ phys . lett . _ * b359 * ( 1995 ) 114 , hep - th/9507046 . m. s. plyushchay , _ phys .
_ * b485 * ( 2000 ) 187 , hep - th/0005122 .
g. grignani , m. plyushchay and p. sodano , _ nucl .
_ * b464 * ( 1996 ) 189 , hep - th/9511072 .
kh . s. nirov and m. s. plyushchay , _ nucl .
* b512 * ( 1998 ) 295 , hep - th/9803221 .
j. de boer , f. harmsze , t. tjin , _ phys .
* 272 * ( 1996 ) 139 , hep - th/9503161 .
e. witten , _ phys .
_ * b117 * ( 1982 ) 324 .
s. elitzur , y. frishman , e. rabinovici and a. schwimmer , _ nucl .
phys . _ * b273 * ( 1986 ) 93 . s. b. treiman , e. witten , r. jackiw and b. zumino , current algebra and anomalies ( singapore , world scientific , 1985 ) .
a. turbiner , _ commun math .
_ * 118 * ( 1988 ) 467 .
shifman , _ int . j. mod
. phys . _ * a4 * ( 1989 ) 2897 .
a. ushveridze , quasi - exactly solvable models in quantum mechanics , ( iop publishing , bristol , 1994 ) .
f. finkel , a. gonzalez - lopez , n. kamran , p. j. olver and m. a. rodriguez , lie algebras of differential operators and partial integrability , hep - th/9603139 . c. m. bender , g. v. dunne and m. moshe , _ phys . rev .
_ * a55 * ( 1997 ) 2625 , hep - th/9609193 .
a. y. morozov , a. m. perelomov , a. a. roslyi , m. a. shifman and a. v. turbiner , _ int .
phys . _ * a5 * ( 1990 ) 803 .
gorsky , _ pisma zh .
* 54 * ( 1991 ) 296 .
v. v. bazhanov , s. l. lukyanov and a. b. zamolodchikov , spectral determinants for schroedinger equation and q - operators of conformal field theory , hep - th/9812247 .
p. e. dorey , c. dunning and r. tateo , ordinary differential equations and integrable models , hep - th/0010148 .
j. suzuki , functional relations in stokes multipliers - fun with @xmath445 potential , quant - ph/0003066 .
andrianov , m.v .
ioffe and v.p .
spiridonov , _ phys .
lett . _ * a174 * ( 1993 ) 273 , hep - th/9303005 .
m. plyushchay , lecture notes in physics , vol . *
524 * , p. 270
( springer - verlag , berlin , 1999 ) , hep - th/9808130 . k. s. nirov and m. plyushchay , _ phys .
lett . _ * b405 * ( 1997 ) 114 , hep - th/9707070 .
h. aoyama , h. kikuchi , i. okouchi , m. sato and s. wada , _ nucl .
phys . _ * b553 * ( 1999 ) 644 , hep - th/9808034 .
shifman , quasi - exactly solvable problems , preprint umn - th-1701 - 98 .
m. s. plyushchay , _ mod .
_ * a11 * ( 1996 ) 397 , hep - th/9601141 .
j. gamboa , m. plyushchay , and j. zanelli , _ nucl.phys . _ * b543 * ( 1999 ) 447 , + hep - th/9808062 . h. aoyama , m. sato , t. tanaka and m. yamamoto , _ phys . lett .
_ * b498 * ( 2001 ) 117 , quant - ph/0011009 . h. aoyama , m. sato and t. tanaka , quant - ph/0012065 . | the nonlinear supersymmetry of one - dimensional systems is investigated in the context of the quantum anomaly problem .
any classical supersymmetric system characterized by the nonlinear in the hamiltonian superalgebra is symplectomorphic to a supersymmetric canonical system with the holomorphic form of the supercharges .
depending on the behaviour of the superpotential , the canonical supersymmetric systems are separated into the three classes . in one of them the parameter specifying the supersymmetry order is subject to some sort of classical quantization , whereas the supersymmetry of another extreme class has a rather fictive nature since its fermion degrees of freedom are decoupled completely by a canonical transformation .
the nonlinear supersymmetry with polynomial in momentum supercharges is analysed , and the most general one - parametric calogero - like solution with the second order supercharges is found .
quantization of the systems of the canonical form reveals the two anomaly - free classes , one of which gives rise naturally to the quasi - exactly solvable systems .
the quantum anomaly problem for the calogero - like models is `` cured '' by the specific superpotential - dependent term of order @xmath0 .
the nonlinear supersymmetry admits the generalization to the case of two - dimensional systems . -1.0 |
in this paper we investigate the potential for the use of incoherent scattered data for 2d reconstruction in x - ray scanning applications .
the use of scattered data for image reconstruction is considered in the literature , typically for applications in gamma ray imaging , where the photon source is monochromatic @xcite . however , in many applications ( e.g security screening of baggage ) a type of x - ray tube is often used that generates a polychromatic spectrum of initial photon energies ( see section [ phys ] for an example spectrum ) .
there has been recent interest in the use of energy sensitive detectors in tomography @xcite , and in the present paper their application is key to the ideas presented .
our main goal is to show that the electron density may be reconstructed analytically using the incoherent scattered data and to lay the foundations for a practical reconstruction method based on our theory .
we apply our method to a machine configuration commonly used in x - ray ct . in addition , by use of the reconstructed density values in conjunction with an attenuation coefficient reconstruction , we show under the right assumptions that the atomic number of the target is uniquely determined . for a photon incident upon an electron compton ( incoherently ) scattering at an angle @xmath0 with initial energy @xmath1 , the scattered energy @xmath2
is given by the equation : @xmath3 where @xmath4kev is the electron rest energy .
equation ( [ equ1 ] ) implies that @xmath0 remains fixed for any given @xmath2 and @xmath1 .
so in the case of a monochromatic source , assuming only single scatter events , for every fixed measured energy @xmath2 ( possible to measure if the detectors are energy - resolved ) the locus of scattering points is a circular arc intersecting the source and detector in question .
for example , refer to @xcite . in an x - ray tube a cathode is negatively charged and electrons are accelerated by a large voltage ( @xmath5 kv ) towards a positively charged target material ( e.g tungsten ) . a small proportion of the initial electron energy ( @xmath6 ) is converted to produce photons . due to energy conservation ,
the resulting photon energies are no more than @xmath5 kev .
so in the polychromatic source case , again assuming only single scatter events , for each given data set ( photon intensity recorded with energy @xmath2 ) , the set of scatterers lie on a collection of circular arcs intersecting the source and detector points . together
these form a toric section in which the photons scatter , with a maximum scattering angle @xmath7 given by : @xmath8 see figure [ fig1 ] below : ( -cos(30),0.5 ) coordinate ( s ) ; ( cos(30 ) , 0.5 ) coordinate ( d ) ; ( -0.5,0.86 ) coordinate ( w ) ; ( -0.13,1.22 ) coordinate ( a ) ; plot ( cos ( ) , sin ( ) ) ; ( s ) ( d ) ; ( s ) ( w ) ; ( w ) ( d ) ; ( w)(a ) ; at ( -0.95,0.48 ) @xmath9 ; at ( 0.95,0.5 ) @xmath10 ; at ( -0.25,0.88 ) @xmath7 ; at ( -0.3,0.4 ) t ; plot ( , -sqrt(1-pow(,2))+1 ) ; + in the present paper we consider a setup consisting of a ring of fixed energy sensitive detectors and a single rotating fan beam polychromatic source
. see figure [ fig3 ] . with this setup we can measure photon intensity in the dark field .
we image an electron density @xmath11 compactly supported within the detector ring ( the blue and green circle in figure [ fig3 ] ) , with @xmath12 .
( 0,0 ) circle [ radius=0.7 ] ; ; ( -0.3,-0.2 ) rectangle ( 0.3,0.2 ) ; ( 0.651,-0.546 ) circle [ radius=0.8 ] ; at ( -0.05,-1.08 ) @xmath9 ; at ( 0.7,0.34 ) @xmath10 ; at ( 0.651 + 0.8,0 ) @xmath13 ; at ( -0.15,0.1 ) @xmath14 ; at ( 0.2,-0.4 ) r ; ( 0,-1)(0,0.7 ) ; ( 0,-1)(0.1,0.6928 ) ; ( 0,-1)(-0.1,0.6928 ) ; ( 0,-1)(0.2,0.6708 ) ; ( 0,-1)(-0.2,0.6708 ) ; ( 0,-1)(0.3,0.6324 ) ; ( 0,-1)(-0.3,0.6324 ) ; ( 0,-1)(0.4,0.5744 ) ; ( 0,-1)(-0.4,0.5744 ) ; ( 0.9,0.9)(0.3,0.6324 ) ; at ( 1.1,1 ) light field ; ( -0.9,-0.9)(-0.3,-0.6324 ) ; at ( -1.1,-1 ) dark field ; ( 0.03,-0.1)->(0.65,0.255 ) ; ( -0.05,-0.4)->(0.65,0.255 ) ; plot ( , sqrt(0.49-pow(,2 ) ) ) ; plot ( , -sqrt(1-pow(,2 ) ) ) ; if we assume an equal scattering probability throughout the region @xmath15 leaving only the electron density to vary , and if we assume that the majority of scattering events occur within r , then in this case the integral of @xmath14 over @xmath13 is approximately determined by the scattered intensity recorded at the detector @xmath10 with some fixed energy @xmath2 .
see the appendix for an example application where these approximations are valid . with these assumptions and with suitable restrictions on the support of @xmath14
, we aim to reconstruct @xmath14 from its integrals over discs whose boundaries intersect a fixed point , namely the source at a given position along its scanning path . in section [ sec1 ]
, we present a disc transform and go on to prove our main theorem ( theorem [ th1 ] ) , which explains the relationship between our transform and the straight line radon transform . as a corollary to this theorem , with known results on the radon transform
, we show that a unique solution exists on the domain of smooth functions compactly supported on an annulus centred at the origin . here based on the criterion of natterer in @xcite and using the theory of sobolev space estimates , we determine a measure for the ill posedness of our problem . in section [ phys ] , we discuss a possible means to approximate the physical processes such as to allow for the proposed reconstruction method .
here we also present a least squares fit for the total cross section ( scattering plus absorbtion ) in terms of @xmath16 ( the atomic number ) . from this , we show that @xmath16 is uniquely determined by the attenuation coefficient and electron density .
in section [ res ] we apply our reconstruction formulae to simulated data sets , with varying levels of added pseudo random noise .
this is applied to the given machine configuration .
we recover a simple water bottle cross section image ( a circular region of uniform density 1 ) and reconstruct the atomic number in each case using the curve fit presented in section [ phys ] . to give an example reconstruction of a target not of uniform density , we also present reconstructions of a simulated hollow tube cross section .
in this section we aim to recover a smooth function compactly supported on an annulus centred at the origin @xmath17 from its integrals over discs whose boundaries intersect @xmath17 ( the given source position ) .
let @xmath18 denote the set of points on the disc whose boundary intersects the origin , with centre given in polar coordinates as @xmath19 .
see figure [ figure8 ] .
let @xmath20 be the set of smooth functions on @xmath21 and let @xmath22 denote the set of smooth functions compactly supported on @xmath23 .
let @xmath24 and for a function in the plane @xmath25 , let @xmath26 be defined as @xmath27 .
then we define the disc transform @xmath28 as : @xmath29 ( -0.5,0)(1.5,0 ) ; ( 0.5,0.5 ) circle [ radius=0.707 ] ; ( 0,0)(1,1 ) ; at ( 0.6,0.5 ) @xmath30 ; at ( 0.17,0.08 ) @xmath31 ; at ( -0.1,-0.1 ) @xmath17 ; at ( 1.3,0.8 ) @xmath32 ; [ figure8 ] after making the change of variables : @xmath33 in equation ( [ equdef1 ] ) , we have : @xmath34 we now present further definitions which will be important in the following subsection ( section [ sob ] ) , where we provide our sobolev space estimates .
let @xmath35 denote the unit cylinder in @xmath36
. then we define @xmath37 as follows : @xmath38 which is piecewise continuous as a function of @xmath30 .
we can remove this discontinuity by adding the function : @xmath39 where @xmath40 .
we define @xmath41 as : @xmath42 let @xmath43 be the set of points on a line
. then we define the radon transform @xmath44 as : @xmath45 we are now in a position to prove our main theorem , where we give the explicit relation between @xmath46 and the radon transform @xmath47 for smooth functions on an annulus .
[ th1 ] let @xmath48 be the annulus centred on @xmath17 with inner radius @xmath49 and outer radius @xmath50 .
let @xmath51 for some @xmath52 and let @xmath53 be defined as @xmath54
. then @xmath55 .
let @xmath56 and @xmath57 be defined as @xmath58 and @xmath59 . then from our definition of @xmath60
, we have @xmath61 .
now we have : @xmath62 and hence @xmath63 .
so the partial derivative of @xmath64 with respect to @xmath30 exists and is continuous for all @xmath65 , and @xmath55 .
we now aim to prove injectivity of the disc transform @xmath46 on the domain of smooth functions compactly supported on an annulus .
first we state helgason s support theorem @xcite .
let @xmath66 be a compact convex set in @xmath67 and let @xmath14 be continuous on @xmath68 . if @xmath69 for all @xmath30 and @xmath31 such that @xmath70 and @xmath14 is rapidly decreasing , in the sense that : @xmath71 then @xmath72 for all @xmath73 . [ cor1 ]
let @xmath74 for some @xmath52 , and let @xmath75\}$ ] .
then @xmath14 is uniquely determined by @xmath64 known for all @xmath76 .
let : @xmath77 and let @xmath60 be defined as in theorem [ th1 ] .
then by theorem [ th1 ] , we have : @xmath78 and hence @xmath60 is rapidly decreasing .
let @xmath79 .
then @xmath66 is clearly compact and convex . by ( [ unique ] ) , @xmath80 for all @xmath30 and @xmath31 such that @xmath81 .
so by helgason s support theorem , we have that @xmath82 for all @xmath73
. the result follows . for the proposed machine configuration
, we can define the set of points within the detector ring formally as @xmath83 , where @xmath52 depends on the machine specifications ( i.e the detector ring radius and the source path radius ) .
we now have : [ cor2 ] let @xmath84 .
let @xmath85 denote the boundary of @xmath86 , and let @xmath87 .
then the values of @xmath88 for @xmath30 and @xmath31 such that : @xmath89 determine @xmath14 uniquely .
we consider two cases . if @xmath90 then @xmath91 , which is known as condition ( [ equcor2 ] ) is satisfied for @xmath92 and @xmath93 . if @xmath94 , then @xmath95 . in any other case
, @xmath88 is known by our assumption .
hence we have a full data set for @xmath88 and hence for @xmath64 .
the result follows from corollary [ cor1 ] .
so for the proposed application , we see from the above corollaries that for any given source position , the incoherent scattered data is sufficient to reconstruct the target density uniquely . in this section
we provide sobolev space estimates for the disc transform @xmath46 . from these we obtain an upper bound for the least squares error in our solution @xmath14 in terms of @xmath96 , where @xmath96 is an upper bound for the least squares error in our measurements .
first we define our sobolev spaces and the norms which will be used in our estimates .
let @xmath97 be an arbitrary domain and let @xmath98 denote the set of square integrable functions on @xmath23 .
we define the fourier transform of a function @xmath99 as : @xmath100 then we can define sobolev spaces @xmath101 of real degree @xmath102 as : @xmath103 with the norm : @xmath104 for functions on the cylinder @xmath105 , we have the norm : @xmath106 where the fourier transform of @xmath107 is taken with respect to the variable @xmath65 .
we now state some preliminary results on sobolev spaces and the radon transform which will be used in our theorems ( * ? ? ? * and 203 ) . for @xmath108 , where @xmath109 is the schwartz space on @xmath110
, we have : @xmath111 where @xmath112 and the fourier transform of @xmath107 is taken with respect to the @xmath30 variable only .
let @xmath113 be some multi index and let @xmath114 , where the @xmath115 are defined in the weak sense .
let @xmath116 be a positive integer and let @xmath117 .
then for @xmath118 , the norm ( [ norm2 ] ) is equivalent to the norm : @xmath119 when @xmath120 .
we now prove a slice theorem for the disc transform @xmath46 .
let @xmath121 for some @xmath52 and let @xmath60 be defined as in theorem [ th1 ] .
then we have : @xmath122 where @xmath112 and the fourier transform of @xmath64 is taken with respect to the @xmath30 variable .
let @xmath123 and @xmath124 be as defined in section [ sec1 ] .
then we have : @xmath125\\ & = -\left(2\pi\right)^{-1/2}\big[\int_{-\infty}^{0}\frac{\partial}{\partial p}\mathcal{d}_2f\left(p,\phi\right)e^{-ip\sigma}\mathrm{d}p+\int_{0}^{\infty}\frac{\partial}{\partial p}\mathcal{d}_2f\left(p,\phi\right)e^{-ip\sigma}\mathrm{d}p\big]\\ & = -\left(2\pi\right)^{-1/2}\left(\mathcal{d}_2f\left(0^{-},\phi\right)-\mathcal{d}_2f\left(0^{+},\phi\right)\right)-\frac{i\sigma}{\left(2\pi\right)^{1/2}}\int_{-\infty}^{\infty}\mathcal{d}_2f\left(p,\phi\right)e^{-ip\sigma}\mathrm{d}p\\ & = -\frac{2c\left(\phi\right)}{\left(2\pi\right)^{1/2}}-i\sigma\hat{\mathcal{d}_2f}\left(\sigma,\phi\right)\\ & = -i\sigma\big[\frac{2c\left(\phi\right)}{i\sigma\left(2\pi\right)^{1/2}}+\hat{\mathcal{d}_2f}\left(\sigma,\phi\right)\big]\\ & = -i\sigma \hat{\mathcal{d}f}\left(\sigma.\phi\right ) \end{split}\ ] ] the result follows from the fourier slice theorem
. in @xcite natterer explains why it is reasonable to consider picture densities as functions @xmath14 of compact support in @xmath101 with @xmath126 .
he then gives a bound for the least squares error in his reconstruction from plane integral data in terms of @xmath127 , where @xmath128 . with this in mind
we will show that the map @xmath129 is bounded and has a bounded inverse from @xmath130 for @xmath131 .
first , from @xcite , we have the lemma : [ lemma2 ] let @xmath132 and let @xmath133 .
then the map @xmath134 is bounded in @xmath135 for any @xmath136 .
now we have our result : let @xmath86 be as defined in section [ sec1 ] .
let @xmath84 for some @xmath52 and let @xmath60 be defined as in theorem [ th1 ] .
then there exist constants @xmath137 and @xmath138 such that : @xmath139 for any @xmath131 .
let @xmath140 be @xmath141 on @xmath86 and let @xmath142 .
then by lemma [ lemma2 ] , we have : @xmath143 for @xmath131 , this proves the left hand inequality .
the right hand inequality can be proven in a similar way . before proving the main theorem of this section we state the interpolation inequality for sobolev spaces on @xmath67 @xcite .
let @xmath144 .
then we have : @xmath145 for any @xmath146 .
now we have our main theorem for this section : [ the2 ] let @xmath84 for some @xmath52 .
then we have : @xmath147\times s^1\right)}\ ] ] for any @xmath148 with @xmath149 .
let @xmath60 be defined as in theorem [ th1 ] and let @xmath112 .
then we have : @xmath150\times s^1\right)}&=2\int_{s^1}\int_{-1}^{1}|\mathcal{d}f\left(p,\phi \right)|^2\mathrm{d}p\mathrm{d}\phi\\ & \geq \int_{s^1}\int_{-1}^{1}|\mathcal{d}f\left(p,\phi \right)+\mathcal{d}f\left(-p,\phi \right)|^2\mathrm{d}p\mathrm{d}\phi\\ & = \int_{s^1}\int_{-\infty}^{\infty}|\hat{\mathcal{d}f}\left(\sigma,\phi \right)+\hat{\mathcal{d}f}\left(-\sigma,\phi \right)|^2\mathrm{d}\sigma\mathrm{d}\phi\\ & = 2\pi\int_{s^1}\int_{-\infty}^{\infty}\frac{1}{\sigma^2}|\hat{\tilde{f}}\left(\sigma\phi\right)-\hat{\tilde{f}}\left(-\sigma\phi\right)|^2\mathrm{d}\sigma\mathrm{d}\phi\\ & \geq 2\pi\int_{s^1}\int_{-\infty}^{\infty}\left(1+\sigma^2\right)^{-1}|\hat{\tilde{f}}\left(\sigma\phi\right)-\hat{\tilde{f}}\left(-\sigma\phi\right)|^2\mathrm{d}\sigma\mathrm{d}\phi\\ & = 4\pi\int_{s^1}\int_{0}^{\infty}\left(1+\sigma^2\right)^{-1}|\hat{\tilde{f}}\left(\sigma\phi\right)-\hat{\tilde{f}}\left(-\sigma\phi\right)|^2\mathrm{d}\sigma\mathrm{d}\phi \end{split}\ ] ] after making the substitution @xmath151 , we have : @xmath152\times s^1\right)}&\geq \int_{\mathbb{r}^2}|\xi|^{-1}\left(1+|\xi|^2\right)^{-1}|\hat{\tilde{f}}\left(\xi\right)-\hat{\tilde{f}}\left(-\xi\right)|^2\mathrm{d}\xi\\ & \geq \|\tilde{f}-\tilde{f}^{-}\|^2_{h^{-\frac{3}{2}}\left(\mathbb{r}^2\right ) } \end{split}\ ] ] where @xmath153 . applying the interpolation inequality with @xmath154 and @xmath155 , yields : @xmath156\times s^1\right)}\|\tilde{f}\|^{\frac{3/2}{\beta+3/2}}_{h^{\beta}\left(\mathbb{r}^2\right)}\\ & \leq c\left(\beta\right)\|\mathcal{d}f\|^{\frac{\beta}{\beta+3/2}}_{l^2\left([-1,1]\times s^1\right)}\|f\|^{\frac{3/2}{\beta+3/2}}_{h^{\beta}\left(\mathbb{r}^2\right)}\\ & \leq c\left(\beta\right)\rho^{\frac{3/2}{\beta+3/2 } } \|\mathcal{d}f\|^{\frac{\beta}{\beta+3/2}}_{l^2\left([-1,1]\times s^1\right ) } \end{split}\ ] ] for any @xmath148 with @xmath149 .
let @xmath84 for some @xmath52 and let @xmath157 .
let @xmath158\times s^1\right)$ ] be such that @xmath159\times s^1\right)}<\epsilon$ ] .
then for any @xmath160 which satisfy @xmath161\times s^1\right)}<\epsilon$ ] , we have : @xmath162 for any @xmath148 with @xmath163 for @xmath164 .
we can interpret this last corollary to mean that given some erroneous data @xmath165 which differs in the least squares sense from @xmath64 absolutely by @xmath96 , the least squares error in our solution is bounded above by @xmath166 for some constant @xmath167 with the a - priori knowledge that @xmath149 .
in @xcite natterer uses the value @xmath168 as a measure for the ill posedness of his problem and gives his criteria for a linear inverse problem to be modestly , mildly or severely ill posed .
if we set @xmath169 close to @xmath170 , then based on these criteria the above arguments would suggest that our problem is mildly ill posed , but more ill posed than the inverse radon transform , which we would expect given that the disc transform @xmath46 is a degree smoother than @xmath47 .
( 0,0 ) circle [ radius=0.7 ] ; ( -1.3,-1)(1.3,-1 ) ; at ( -0.1,-1.1 ) @xmath17 ; ( 0,-1)(0,1 ) ; at ( 0.1,1 ) @xmath171 ; at ( 1.3,-1.1 ) @xmath172 ; ( -1.3,0)(1.3,0 ) ; ( 0.04,-1)(0.04,-0.7 ) ; ( -0.04,-1)(-0.04,0.7 ) ; at ( -0.1,-0.1 ) @xmath173 ; at ( 0.1,-0.85 ) @xmath141 ; at ( 0.6,-0.6 ) @xmath85 ; ( 0,0)(0.3,0.632 ) ; at ( 0.13,0.08 ) @xmath174 ; at ( 0.4,0.7 ) @xmath175 ; another source of error in our solution can be due to limited sampling of the data . in practice
the number of detectors will be finite .
let us parameterize the set of points on the detector ring @xmath176 in terms of a polar angle @xmath177 , and let the finite set of polar angles @xmath178 determine a finite set of detector positions @xmath179 .
see figure [ fig0 ] .
then for every @xmath180 $ ] we can sample @xmath181 for : @xmath182 where @xmath183 is such that @xmath184 for @xmath185 . from ( * ? ? ?
* and 42 ) we have : [ lemma10 ] let @xmath186 be bounded and sufficiently regular . for @xmath187 let @xmath188 be a finite subset of @xmath23 such that @xmath189 , where @xmath10 is the is the hausdorff distance metric between sets
let @xmath190 where @xmath118 for some integer @xmath116 and @xmath191 , and for @xmath192 define the seminorm : @xmath193 then there is a constant @xmath194 such that : @xmath195 for every @xmath192 which is zero on @xmath188 .
let @xmath196 be the unit ball in @xmath67 .
for every @xmath197 there exist positive constants @xmath198 and @xmath199 such that for @xmath200 @xmath201 from these we have the theorem : for each @xmath180 $ ] let @xmath202 $ ] be a finite subset of the unit interval .
let : @xmath203\right)\ ] ] where @xmath10 is the hausdorff distance metric .
let @xmath84 and let @xmath204 with @xmath131 . if @xmath205 is zero on @xmath206 for every @xmath207 $ ] , then there exists a constant @xmath137 such that : @xmath208 let @xmath60 be defined as in theorem [ th1 ] and let @xmath209 be the seminorm defined in lemma [ lemma10 ] .
let @xmath210 for some integer @xmath116 and @xmath211 .
then we have : @xmath212\times s^1\right)}&=\int_{s^1}\|\mathcal{d}f_{\phi}\|^2_{l^2\left([-1,1]\right ) } \mathrm{d}\phi\\ & \leq c^2 h^{2\alpha+3}\int_{s^1}|\mathcal{d}f_{\phi}|^2_{h^{\alpha+3/2}\left([-1,1]\right ) } \mathrm{d}\phi\\ & \leq c^2 h^{2\alpha+3}\int_{s^1}\iint_{[-1,1]\times [ -1,1]}\frac{|\frac{\partial^m}{\partial p^m } \mathcal{d}f_{\phi}-\frac{\partial^m}{\partial p^m } \mathcal{d}f_{\phi}|^2}{|x - y|^{n+2\sigma}}\mathrm{d}x\mathrm{d}y \mathrm{d}\phi\\ & = c^2 h^{2\alpha+3}\int_{s^1}\iint_{[-1,1]\times [ -1,1]}\frac{|\frac{\partial^{m-1}}{\partial p^{m-1 } } r\tilde{f}_{\phi}-\frac{\partial^{m-1}}{\partial p^{m-1 } } r\tilde{f}_{\phi}|^2}{|x - y|^{n+2\sigma}}\mathrm{d}x\mathrm{d}y \mathrm{d}\phi\\ & = c^2 h^{2\alpha+3}\int_{s^1}|r\tilde{f}_{\phi}|^2_{h^{\alpha+1/2}\left([-1,1]\right ) } \mathrm{d}\phi\\ & \leq c^2 h^{2\alpha+3}\int_{s^1}\|r\tilde{f}_{\phi}\|^2_{h^{\alpha+1/2}\left([-1,1]\right ) } \mathrm{d}\phi\\ & = c^2 h^{2\alpha+3}\|r\tilde{f}\|^2_{h^{\alpha+1/2}\left(z\right)}\\ & \leq c_1\left(\alpha\right)^2 h^{2\alpha+3}\|\tilde{f}\|^2_{h^{\alpha}\left(\mathbb{r}^2\right)}\\ & \leq c_2\left(\alpha\right)^2 h^{2\alpha+3}\|f\|^2_{h^{\alpha}\left(\mathbb{r}^2\right)}\\ & \leq c_2\left(\alpha\right)^2 h^{2\alpha+3}\rho^2\\ \end{split}\ ] ] for @xmath131 with @xmath128 . applying theorem [ the2 ]
, we have : @xmath213\times s^1\right)}\\ & \leq c\left(\alpha\right)h^{\alpha}\rho \end{split}\ ] ] which completes the proof .
this last result tells us that given a finite set of detectors with a disc diameter sampling determined by equation ( [ sample ] ) and with @xmath214 being a measure of the uniformity of the sample , the least squares error in our solution is bounded above by @xmath215 with the a - priori knowledge that @xmath128 for some @xmath131 .
in this section we present an accurate physical model and a possible approximate model which allows for the proposed reconstruction method .
we consider an intensity of photons scattering from a point @xmath216 as illustrated in figure [ figurep ] .
( -1,-1)(0,0 ) ; ( 0,0)(1,1 ) ; ( 0,0)(1.3,-0.2 ) ; at ( 0.25,0.1 ) @xmath0 ; at ( -1.1,-1.1 ) @xmath9 ; at ( 1.4,-0.3 ) @xmath10 ; at ( -0.6,-0.3 ) @xmath1 ; at ( 1.3,0 ) @xmath2 ; at ( -0.1,0.1 ) @xmath216 ; at ( 1.15,1.1 ) @xmath217 ; the intensity of photons scattered from @xmath216 to @xmath10 with energy @xmath2 is : @xmath218 where @xmath219 is the initial intensity , which depends on the energy @xmath1 ( see figure [ figspec ] for an example polychromatic spectrum ) .
@xmath220 is the linear attenuation coefficient , which is dependant on the energy @xmath221 and the atomic number of the target material . here
@xmath222 is the number of electrons in a volume @xmath223 around the scattering point @xmath216 .
so @xmath224 ( number of electrons per unit volume ) is the quantity to be reconstructed .
@xmath225 and @xmath226 are the line segments connecting @xmath9 to @xmath216 and @xmath216 to @xmath10 respectively .
the klein - nishina differential cross section @xmath227 , is defined by:@xmath228 where @xmath229 is the classical electron radius .
this predicts the scattering distribution for a photon off a free electron at rest .
given that the atomic electrons typically are neither free nor at rest , a correction factor is included , namely the incoherent scattering function @xmath230 . here
@xmath231 is the momentum transferred by a photon with initial energy : @xmath232 scattering at an angle @xmath0 , where @xmath214 is planck s constant and @xmath194 is the speed of light .
the scattering function @xmath233 also depends on the atomic number @xmath16 , so we set @xmath234 to some average atomic number as an approximation . for @xmath235 ( rhodium )
we have the expression : @xmath236 to acquire equation ( [ equ16 ] ) we have extended the least squares fit given in @xcite to the values of @xmath230 given in @xcite .
the solid angle subtended by @xmath216 and @xmath10 is defined : @xmath237 where @xmath238 , @xmath239 is the detector area and @xmath240 is the unit vector normal to the detector surface . given our machine geometry and proposed reconstruction method , it is difficult to include the more accurate model stated above as an additional weighting to our integral equations ( as in done in @xcite for example ) while allowing for the same inversion formulae .
so we average equation ( [ equ17 ] ) over the scattering region @xmath87 , for each @xmath30 and @xmath31 .
here @xmath18 and @xmath86 are as defined in section [ sec1 ] , where @xmath173 is fixed depending on the machine specifications .
let @xmath241 . here
@xmath242 depends on the scattering point @xmath216 and @xmath30 and @xmath31 as defined in section [ sec1 ] .
when @xmath243 , @xmath30 and @xmath31 determine the detector position @xmath10 and the measured energy @xmath2 .
we have : @xmath244 which gives the average of @xmath245 over @xmath246 . here
@xmath247 denotes the area of @xmath246 .
let @xmath25 be an example density with support contained in @xmath86 , and let : @xmath248 be the scattered intensity measured for a constant density @xmath249 over @xmath246 , where @xmath9 is the ( constant ) slice thickness .
then if we assume that the scattering probability is constant and equal to @xmath250 throughout each scattering region @xmath251 , the absolute error in our approximation would satisfy : @xmath252 for all @xmath253 . here
@xmath254 is the intensity of photons we measure .
so provided that the range of the density values is small over the majority of scattering regions considered , the averaged model given above will have a similar level of accuracy to the more precise model given in equation ( [ equ17 ] ) . if the linear attenuation coefficient @xmath255 is known a - priori , then the exponential terms of equation ( [ equ17 ] ) may be included in @xmath245 .
otherwise we may approximate : @xmath256 where @xmath257 is the line segment from @xmath9 to the detector in the forward direction @xmath217 ( see figure [ figurep ] ) .
this is the approximation made in @xcite . by the beer - lambert law
, we have : @xmath258 where @xmath259 is the recorded straight through intensity . to account for the physical modelling
, we would divide the data by @xmath260 to calculate approximate values for @xmath88 and hence for @xmath64 . with the proposed machine configuration
, we can show that the data collected in the light field determines the linear attenuation coefficient @xmath255 uniquely ( this is the standard 2d reconstruction problem ) . with the additional information provided by our theory , we show under the right assumptions that the atomic number of the target is determined uniquely by the full data ( light plus dark field ) .
the electron density @xmath224 and the linear attenuation coefficient @xmath255 are related via the formula : @xmath261 where @xmath262 is the total cross section per electron .
the cross section @xmath262 is continuous and monotone increasing as a function of @xmath16 on @xmath263 $ ] , where @xmath264 .
for example , when @xmath265kev we can calculate @xmath266 . with this in mind
, we fit a smooth curve to known values of @xmath262 given in the atomic data tables @xcite .
this allows us to calculate values of @xmath262 for non integer @xmath16
. see figure [ f6 ] .
for @xmath265kev up to @xmath267 where a sudden dip in the @xmath262 values occurs .
the tabulated values of @xmath262 given in @xcite are shown alongside the fitted curve . ]
the formula for the fit presented for a fixed energy @xmath265kev is : @xmath268 this was obtained via a combination of the formula for @xmath269 ( the total scattering cross section ) presented by jackson and hawkes in @xcite and the suggested fit for @xmath270 ( the photoelectric cross section ) given in @xcite .
fits for energies other than @xmath265kev are also possible via the same fitting method . from our theory
we know that the data determines @xmath271 and @xmath224 uniquely , where @xmath272 .
if we assume that atomic numbers @xmath273 are not present in the target material , then it is clear from the above arguments that the atomic number of the target is uniquely determined . without this assumption the atomic number would be limited to a range of values .
if we reconstruct both @xmath271 and @xmath224 for a suitably high energy @xmath221 , we can then calculate values for @xmath16 from our curve fit for @xmath262 .
we will test this additional method in our results also .
to test our reconstruction methods , let us consider the water bottle cross section @xmath14 and the corresponding function @xmath60 represented in figure [ figure5 ] .
we calculate values of @xmath64 for @xmath30 in the range @xmath274 $ ] and for @xmath275 .
these were calculated using the exact formula for the area of intersection of two discs .
we approximate the derivative of @xmath64 with respect to @xmath30 as the finite difference : @xmath276 for a chosen step size @xmath214 . to reconstruct @xmath60
we apply the matlab function iradon " , which filters ( choosing from a selection of filters pre - coded by matlab ) and backprojects the projection data @xmath277 to recover @xmath60 .
we then make the necessary change in coordinates to produce our density image @xmath14 . in the absence of noise we find our results to be satisfactory .
see figure [ figure7 ] .
let us now perturb the calculated values of @xmath64 slightly such as to simulate random noise .
we multiply each exact value of @xmath64 by a pseudo random number in the range @xmath278 $ ] ( we use the c++ function rand " to generate random numbers ) , where @xmath279 is the desired amount of percentage error to be added . in this case , even with a relatively small amount of added noise , the data must be smoothed sufficiently before applying approximation ( [ app1 ] ) . to smooth the data
, we apply a simple moving average filter and calculate any intermediate values via a shape preserving cubic interpolation method ( pchip " interpolation in matlab ) .
we expect this interpolation method to preserve the monotonicity of the data ( monotone decreasing ) as a function of @xmath30 . to illustrate this technique we refer to figure [ figure9 ] .
we have presented our reconstructions after smoothing with @xmath280 , @xmath281 and @xmath282 added noise in figures [ figure11 ] , [ figure13 ] and [ figure14 ] . here , we have reconstructed @xmath14 from a single view point , using data collected from a single source projection . with the proposed machine configuration however , there are a number of views from which @xmath14 may be reconstructed .
so we take an average over 360 views ( for source positions at equal @xmath283 intervals over the range @xmath284 $ ] ) .
our results are presented in figures [ figure16 ] and [ figure15 ] . here
we see an improvement in the signal - to - noise - ratio . the rotational symmetry of @xmath14 about the centre of the circular region of @xmath14 s support is also recovered .
let @xmath285 be the average of the non zero pixel values shown in the left hand image of figure [ figure16 ] and let @xmath286 be the effective atomic number for water .
then we can calculate @xmath287 and using equation ( [ equ30 ] ) we can calculate the total cross section to be : @xmath288 for @xmath265kev assuming no additional error . based on our curve fit for @xmath289 , this would yield a reconstructed atomic number of @xmath290 , which differs from the accepted value by @xmath291 . for the remaining averaged density reconstructions the @xmath285 and @xmath16 values
are given in the figure caption .
we have presented reconstructions of a density which is homogeneous where it is not known to be zero . to give an inhomogeneous example ,
we have presented reconstructions with varying levels of added noise of a simulated hollow tube cross section in figures [ figure17 ] and [ figure18 ] .
we can summarize our method as follows : 1 .
measure the scattered intensity energy @xmath2 and divide by @xmath292 and the slice thickness to calculate values for @xmath64 .
2 . smooth the data sufficiently and apply approximation ( [ app1 ] ) to calculate values for @xmath293 .
3 . reconstruct @xmath60 by filtered backprojection and recover @xmath14 from the definition given in theorem [ th1 ] .
4 . average over a number of source views to improve the image quality and set @xmath14 to @xmath294 outside its support . is simulated as a circular region of uniform density @xmath141 on the left .
the function @xmath60 as defined in theorem [ th1 ] is shown on the right.,title="fig : " ] is simulated as a circular region of uniform density @xmath141 on the left .
the function @xmath60 as defined in theorem [ th1 ] is shown on the right.,title="fig : " ] in the absence of added noise is shown on the left .
we have backprojected from 180 views with the default ram - lak cropped filter .
the corresponding pixel values of @xmath14 are presented on the right .
both @xmath14 and @xmath60 are set to @xmath294 outside of their support .
, title="fig : " ] in the absence of added noise is shown on the left .
we have backprojected from 180 views with the default ram - lak cropped filter .
the corresponding pixel values of @xmath14 are presented on the right .
both @xmath14 and @xmath60 are set to @xmath294 outside of their support .
, title="fig : " ] for @xmath295 with @xmath281 random noise added . on the right
we have applied a simple moving average filter to the simulated data and taken a subsample of the smoothed data before interpolating as specified earlier .
the exact values are presented alongside the fitted values in the right hand figure.,title="fig : " ] for @xmath295 with @xmath281 random noise added . on the right we have applied a simple moving average filter to the simulated data and taken a subsample of the smoothed data before interpolating as specified earlier .
the exact values are presented alongside the fitted values in the right hand figure.,title="fig : " ] after smoothing with @xmath280 added noise .
we have again backprojected from 180 views , although here we have multiplied the standard ramp filter by a hamming window to reduce the high frequency noise .
the corresponding pixel values for @xmath14 are presented on the right.,title="fig : " ] after smoothing with @xmath280 added noise .
we have again backprojected from 180 views , although here we have multiplied the standard ramp filter by a hamming window to reduce the high frequency noise .
the corresponding pixel values for @xmath14 are presented on the right.,title="fig : " ] after smoothing with @xmath281 added noise .
we have multiplied the ramp filter by a hamming window and backprojected from 180 views .
the corresponding pixel values for @xmath14 are displayed on the right.,title="fig : " ] after smoothing with @xmath281 added noise .
we have multiplied the ramp filter by a hamming window and backprojected from 180 views .
the corresponding pixel values for @xmath14 are displayed on the right.,title="fig : " ] after smoothing with @xmath282 added noise is shown on the left .
we have multiplied the ramp filter by a hamming window and backprojected from 180 views .
the corresponding pixel values for @xmath14 are displayed on the right.,title="fig : " ] after smoothing with @xmath282 added noise is shown on the left .
we have multiplied the ramp filter by a hamming window and backprojected from 180 views .
the corresponding pixel values for @xmath14 are displayed on the right.,title="fig : " ] is shown with no noise added to each dataset before reconstruction .
for the right hand image @xmath280 random noise was added to each dataset before reconstruction . in this case @xmath296 which gives an atomic number value of @xmath297.,title="fig : " ] is shown with no noise added to each dataset before reconstruction .
for the right hand image @xmath280 random noise was added to each dataset before reconstruction . in this case @xmath296 which gives an atomic number value of @xmath297.,title="fig : " ] is shown with @xmath281 noise added to each dataset before reconstruction . here
@xmath298 which gives an atomic number value of @xmath299 .
for the right hand image @xmath282 random noise was added to each dataset before reconstruction . in this case @xmath300 which gives a reconstructed atomic number value of @xmath301.,title="fig : " ]
is shown with @xmath281 noise added to each dataset before reconstruction . here
@xmath298 which gives an atomic number value of @xmath299 .
for the right hand image @xmath282 random noise was added to each dataset before reconstruction . in this case @xmath300 which gives a reconstructed atomic number value of @xmath301.,title="fig : " ] . on the right
is an averaged reconstruction of @xmath302 with no noise added to each dataset.,title="fig : " ] . on the right
is an averaged reconstruction of @xmath302 with no noise added to each dataset.,title="fig : " ] with @xmath281 and @xmath282 added noise in the left and right hand images respectively.,title="fig : " ] with @xmath281 and @xmath282 added noise in the left and right hand images respectively.,title="fig : " ]
we have proposed a new fast method to determine the electron density in x - ray scanning applications , with a fixed energy sensitive detector machine configuration where it is possible to measure photon intensity in the dark field .
we have shown that the density may be reconstructed analytically using the compton scattered intensity .
this method does not require the photon source to be monochromatic as is the case in recent literature , which is important from a practical standpoint as it may not be reasonable to assume a monochromatic source in some applications .
also if the source is monochromatic we can not gain any insight into the energy dependence of the attenuation coefficient , which would rule out the recent advances in image rendering @xcite , where a combination of multivariate and cluster analysis can be used to render a colour x - ray image . using sobolev space estimates , we have determined an upper bound for the least squares error in our solution in terms of the least squares error in our data .
this work is based on the approach taken by natterer in @xcite .
we have shown , under the right assumptions , that the atomic number of the target is determined uniquely by the full data . with this theory in place we intend to pursue a more practical means to reconstruct the atomic number @xmath16 , as the graph reading method used in the present paper was ineffective in giving an accurate reconstruction for @xmath16 . we summarize our method to recover the density image in section [ res ] and we reconstruct a simulated water bottle cross section via a possible practical implementation of this method . in this simple case the smoothing method ( simple moving average ) applied was effective and we were able to reconstruct a circular cross section of approximately uniform density .
although in the presence of noise the pixel values of our reconstructed density image on average differed from the original values by as much as @xmath303 .
we have also provided reconstructions of a simulated hollow tube cross section . in this case
the inner edge of the tube cross section appeared quite blurred in the reconstruction when noise was added to the simulated data .
we performed a number of trial reconstructions with different randomly generated datasets .
the results presented in this paper are typical of our trial results .
we hope also to test our methods through experiment .
for example , if we were to take an existing x - ray machine of a similar configuration to that discussed in the present paper , and attach energy sensitive detectors alongside the existing detectors or if we were to replace them , then we could see how closely our forward problem models the intensity of photons measured in the dark field in practice .
i would like to thank my ph.d .
supervisor prof william lionheart for his guidance and inspiration .
the author is also grateful to prof robert cernik for his helpful comments and discussion regarding energy sensitive detectors , and to dr ed morton and dr tim coker of rapiscan systems for information on baggage scanning .
this work has been funded jointly by the epsrc and rapiscan systems .
the rtt80 ( real time tomography ) x - ray scanner is a switched source , offset detector ct machine designed with the aim to scan objects in real time . developed by rapiscan systems , the rtt80 is currently used in airport security screening of baggage .
the rtt80 consists of a single fixed ring of polychromatic x - ray sources and multiple offset rings of detectors , with a conveyor belt and scanning tunnel ( within which the scanned object would be placed ) passing through the centre of both sets of rings .
see figure [ figrtt ] .
if the detectors are energy sensitive , then in this case we have the problem of reconstructing a density slice supported within the scanning tunnel from its integrals over toric sections , with tips at the source and detector locations .
we wish to check whether it is reasonable to approximate a set of toric section integrals as integrals over discs whose boundaries intersect a given source point , as then we can apply our proposed reconstruction method to reconstruct the density slice analytically .
let us refer to figure [ figrtt1 ] and let @xmath18 be defined as in section [ sec1 ] .
we define the toric sections @xmath304 , @xmath305 , @xmath306 and @xmath307 .
let @xmath308 denote the area of a set @xmath309 and let @xmath310 denote the set of points within our roi ( region of interest , i.e the scanning tunnel ) . for a large sample of discs , we will check for every disc @xmath18 in the sample , whether @xmath311 such that @xmath312 .
let @xmath86 be defined as in section [ sec1 ] .
then if we consider the machine specifications for the rtt80 , we can calculate @xmath313 and the difference in radius between the detector ring and the scanning tunnel to be @xmath314 .
see figure [ figrtt ] . for our test
, we consider a sample of 36000 discs with diameters @xmath315 for @xmath316 and @xmath317 for @xmath318 .
we have chosen @xmath319 $ ] and @xmath320 $ ] values in a range sufficient to determine a unique density slice image for densities supported on @xmath321 .
refer to corollary [ cor1 ] . for each of our chosen @xmath30 and @xmath31 value pairs ,
the difference : @xmath322 was found to be negligible .
let @xmath25 be an example density slice with support contained in @xmath321 .
then for any disc @xmath18 in our sample we have : @xmath323 which holds for some @xmath324 .
so , the integral of @xmath14 over @xmath18 is equal to at least one of four toric section integrals over @xmath14 . assuming also that there is little error implied by our physical approximations ( these are discussed in detail in section [ phys ] ) , the integral ( [ intd ] )
would be determined approximately by at least one of four data sets , namely the photon intensity measured for two possible energy levels at two possible detector locations ( @xmath325 or @xmath326 ) .
thus , given that the inverse disc transform is only mildly ill posed ( this was determined to be the case in section [ sob ] , based on the criteria given by natterer in @xcite ) , it seems that we should be able obtain a satisfactory density image reconstruction in this application . in airport baggage screening , we are interested in identifying a given material as either a threat or non - threat .
let @xmath224 be the electron density and let @xmath16 denote the effective atomic number .
we define the threat space to be the set of materials with @xmath327 , where @xmath328 $ ] is the class of threat @xmath329 pairs . for a given suspect material
, we can apply the methods presented in this paper to reconstruct @xmath224 and @xmath16 .
then if @xmath330 , we can identify the suspect material as a potential threat .
we note that although we failed to obtain an accurate @xmath16 reconstruction in the present paper , we aim to show that a more precise determination of @xmath16 is possible in future work . also , the reconstruction methods we have presented should be fast to implement as they are largely based on the filtered back - projection algorithm .
this is important in an application such as airport baggage screening , as we require the threat detection method we apply to not only be accurate in threat identification , but to also be an efficient process .
( 0,0 ) circle [ radius=0.7 ] ; ( -0.3,-0.2 ) rectangle ( 0.3,0.2 ) ; ( 0,-1)(0,0.7 ) ; ( 0,-1)(0.1,0.6928 ) ; ( 0,-1)(-0.1,0.6928 ) ; ( 0,-1)(0.2,0.6708 ) ; ( 0,-1)(-0.2,0.6708 ) ; ( 0,-1)(0.3,0.6324 ) ; ( 0,-1)(-0.3,0.6324 ) ; ( 0,-1)(0.4,0.5744 ) ; ( 0,-1)(-0.4,0.5744 ) ; ( 0.9,0.9)(0.3,0.6324 ) ; at ( 1.1,1 ) detector ring ; ( 0,0 ) circle [ radius=1 ] ; ( -1.1,-0.9)(-0.7,-0.71 ) ; at ( -1.1,-1 ) source ring ; ( -0.4,-0.3)(0.4,-0.3 ) ; ( -0.9,0.9)(-0.3,0 ) ; at ( -0.9,1 ) scanned object ; ( 1.1,-0.9)(0.3,-0.3 ) ; at ( 1.1,-1 ) conveyor belt ; circle [ radius=0.57 ] ; ( 1.3,0.3)(0.57,0 ) ; at ( 1.4,0.4 ) scanning tunnel ; ( -1.8,-1)(-1.8,-0.7 ) ; ( -1.8,-0.7)(-1.8,-0.57 ) ; ( -1.8,-0.57)(-1.8,0.57 ) ; at ( -1.9,-0.85 ) 1 ; at ( -2,-0.635 ) 0.375 ; at ( -1.9,0 ) 5 ; ( 0,0 ) circle [ radius=0.7 ] ; ; ( 0.651,-0.546 ) circle [ radius=0.8 ] ; at ( -0.05,-1.08 ) @xmath17 ; at ( 0.72,0.34 ) @xmath325 ; at ( 0.651 + 0.8,0 ) @xmath18 ; at ( -0.18,0.2 ) @xmath331 ; circle [ radius=0.57 ] ; ( -1.3,-1)(1.3,-1 ) ; at ( -0.2,-0.75 ) @xmath326 ; ( 0,-0.206 ) circle [ radius=0.8 ] ; ( -0.78,-01.14 ) circle [ radius=0.8 ] ; ( -1.1,0.4)(-0.78,0 ) ; ( -1.3,-0.1)(-0.9,-0.35 ) ; at ( -1.1,0.5 ) @xmath332 ; at ( -1.3,0 ) @xmath333 ; ( 0,0 ) circle [ radius=1 ] ; ( 0,-1)(1.302,-0.092 ) ; at ( 0.651,-0.5 ) @xmath30 ; at ( 0.18,-0.93 ) @xmath31 ; 10 v.p .
palamodov , an analytic reconstruction for the compton scattering tomography in a plane " inverse problems 27 ( 2011 ) 125004 ( 8pp ) . v. maxim , m. frandes , r. prost , analytical inversion of the compton transform using the full set of available projections " inverse problems 25 ( 2009 ) 095001 ( 21pp ) .
compton scattering tomography " j. appl .
76 200715 ( 1994 ) . q. xu , h. yu , j. bennett , image reconstruction for hybrid true - color micro ct " ieee trans biomed eng .
59(6 ) 17111719 ( 2012 ) .
egan , s.d.m .
jacques , r.j .
cernik , multivariate analysis of hyperspectral hard x - ray images " wiley 42 151157 ( 2012 ) .
helgason , s. gropes and geometric analysis " academic press , orlando - san diego - san francisco - new york - london - toronto - montreal - sydney - tokyo - sao paulo ( 1984 ) .
palinkas , g. analytic approximations for the incoherent x - ray intensities of the atoms from ca to am " acta cryst .
a29 , 10 ( 1973 ) .
hubbell , j. h. et .
al atomic form factors , incoherent scattering functions and photon scattering cross sections " j. phys .
chem . ref .
data , vol .
3 , ( 1975 ) . wm .
j. veigele , photon cross sections from 0.1kev to 1mev for elements @xmath334 to @xmath335 " atomic data tables , 5 , 51 - 111 ( 1973 ) .
d. f. jackson , d. j. hawkes an accurate parametrisation of the x - ray attenuation coefficient " phys
biol . 25 1167 ( 1980 ) .
wadeson , n. , modelling and correction of scatter in a switched source multi ring detector ct machine " phd thesis , university of manchester uk , ( 2012 ) .
f. natterer the mathematics of computerized tomography " siam ( 2001 ) . | we lay the foundations for a new fast method to reconstruct the electron density in x - ray scanning applications using measurements in the dark field .
this approach is applied to a type of machine configuration with fixed energy sensitive ( or resolving ) detectors , and where the x - ray source is polychromatic .
we consider the case where the measurements in the dark field are dominated by the compton scattering process .
this leads us to a 2d inverse problem where we aim to reconstruct an electron density slice from its integrals over discs whose boundaries intersect the given source point .
we show that a unique solution exists for smooth densities compactly supported on an annulus centred at the source point . using sobolev space estimates
we determine a measure for the ill posedness of our problem based on the criterion given by natterer in @xcite .
in addition , with a combination of our method and the more common attenuation coefficient reconstruction , we show under certain assumptions that the atomic number of the target is uniquely determined .
we test our method on simulated data sets with varying levels of added pseudo random noise . |
a common assumption in work on collaborative problem solving is that interaction should be efficient . when language is the mode of interaction , the measure of efficiency has been , in the main , the number of utterances required to complete the dialogue @xcite .
one problem with this efficiency measure is that it ignores the cognitive effort required by resource limited agents in collaborative problem solving .
another problem is that an utterance - based efficiency measure shows no sensitivity to the required quality and robustness of the problem solution .
cognitive effort is involved in processes such as making inferences and swapping items from long term memory into working memory .
when agents have limited working memory , then only a limited number of items can be salient , i.e. accessible in working memory .
since other processes , e.g. inference , operate on salient items , an inference process may require the cognitive effort involved with retrieving items from long term memory , in addition to the effort involved with reasoning itself .
the required quality and robustness of the problem solution often determines exactly how much cognitive effort is required .
this means that a resource - limited agent may do well on some tasks but not on others @xcite .
for example , consider constraint - based tasks where it is difficult for an agent to simultaneously keep all the constraints in mind , or inference - based tasks that require a long deductive chain or the retrieval of multiple premises , where an agent may not be able to simultaneously access all of the required premises .
furthermore , contrary to the efficiency hypothesis , analyses of problem - solving dialogues shows that human agents in dialogue engage in apparently inefficient conversational behavior .
for example , naturally - occurring dialogues often include utterances that realize facts that are already mutually believed , or that would be mutually believed if agents were logically omniscient @xcite .
consider -26a , which repeats information given in -20 @xmath0 -23 : utterances such as -26a , that repeat , paraphrase or make inferences explicit , are collectively called informationally redundant utterances , irus . in , the utterances that originally added the belief that _ they will get 400 dollars _ to the context are in _ italics _ and the iru is given in caps . about 12% of the utterances in a corpus of 55 naturally - occurring problem - solving dialogues were irus @xcite , but the occurrence of irus contradicts fundamental assumptions of many theories of communication @xcite , _
inter alia_. the hypothesis that is investigated in this paper is that irus such as -26a are related to agents limited attentional and inferential capacity and reflect the fact that beliefs must be salient to be used in deliberation and inference .
hence apparently redundant information serves an important cognitive function . in order to test the hypothesized relationship of communicative strategies to agents resource limits we developed a test - bed environment , design - world , in which we vary task requirements , agents resources and communicative strategies .
our artificial agents are based on a cognitive model of attention and memory .
our experimental results show that communicative strategies that incorporate irus can help resource - limited cognitive agents coordinate , limit processing , and improve the quality and robustness of the problem solution .
we will show that the task determines whether a communicative strategy is beneficial , depending on how the task is defined in terms of fault intolerance and the level of belief coordination required .
the design - world task consists of two agents who carry out a dialogue in order to come to an agreement on a furniture layout design for a two room house @xcite .
figure [ final - state - fig ] shows a potential final plan constructed as a result of a dialogue .
the agents shared intention is to design the house , which requires two subparts of designing room-1 ( the study ) and designing room-2 ( the living room ) .
a room design consists of four intentions to put a furniture item into the room .
each furniture item has a color and point value , which provides the basis for calculating the utility of a put - act involving that furniture item .
agents start with private beliefs about the furniture items they have and their colors .
beliefs about which furniture items exist and how many points they are worth are mutual .
the agent architecture for deliberation and means - end reasoning is based on the irma architecture , also used in the tileworld simulation environment @xcite , with the addition of a model of limited attention / working memory , awm .
see figure [ irma - fig ] .
the attention / working memory model , awm , is adapted from @xcite .
while the awm model is extremely simple , landauer showed that it could be parameterized to fit many empirical results on human memory and learning @xcite .
awm consists of a three dimensional space in which propositions acquired from perceiving the world are stored in chronological sequence according to the location of a moving memory pointer .
the sequence of memory loci used for storage constitutes a random walk through memory with each loci a short distance from the previous one . if items are encountered multiple times , they are stored multiple times @xcite .
when an agent retrieves items from memory , search starts from the current pointer location and spreads out in a spherical fashion .
search is restricted to a particular search radius : radius is defined in hamming distance .
for example if the current memory pointer loci is ( 0 0 0 ) , the loci distance 1 away would be ( 0 1 0 ) ( 0 -1 0 ) ( 0 0 1 ) ( 0 0 -1 ) ( -1 0 0 ) ( 1 0 0 ) .
the actual locations are calculated modulo the memory size .
the limit on the search radius defines the capacity of attention / working memory and hence defines which stored beliefs and intentions are salient .
the radius of the search sphere in the awm model is used as the parameter for design - world agents resource - bound on attentional capacity . in the experiments below , memory is 16x16x16 and the radius parameter varies between 1 and 16 , where awm of 1 gives severely attention limited agents and awm of 16 means that everything an agent knows is salient .
the advantages of the awm model is that it was shown to reproduce , in simulation , many results on human memory and learning . because search starts from the current pointer location ,
items that have been stored most recently are more likely to be retrieved , predicting recency effects @xcite .
because items that are stored in multiple locations are more likely to be retrieved , the model predicts frequency effects @xcite .
because items are stored in chronological sequence , the model produces natural associativity effects @xcite .
because deliberation and means - end reasoning can only operate on salient beliefs , limited attention produces a concomitant inferential limitation , i.e. if a belief is not salient it can not be used in deliberation or means - end - reasoning .
this means that mistakes that agents make in their planning process have a plausible cognitive basis .
agents can both fail to access a belief that would allow them to produce an optimal plan , as well as make a mistake in planning if a belief about how the world has changed as a result of planning is not salient .
a communicative strategy is a strategy for communicating with another agent , which varies according to the agents initiative , amount of information about the task , degree of resource - bounds , and communication style @xcite .
design - world agents communicate with an artificial language whose primitive communicative acts are propose , accept , reject , say .
these primitive acts can be composed to produce higher level discourse acts such as proposals , acceptances , rejections , openings and closings @xcite .
see figure [ dial - stat - fig ] .
a discourse act may be left implicit , or may be varied to consist of one or more communicative acts .
discourse acts are different from actions on the environment because they are actions whose intended effect is a change in the other agent s mental state . because the other agent is an active intelligence , it is possible for it to supplement an underspecified discourse action with its own processing .
the variation in the degree of explicitness of a discourse act is the basis of agents communicative strategies . here
we will compare three communicative strategies : ( 1 ) all - implicit ; ( 2 ) close - consequence ; and ( 3 ) explicit - warrant .
the all - implicit strategy is a ` bare bones ' strategy , exemplified by the partial dialogue in . in each utterance
is shown both as a gloss in _ italics _ , and in the artificial language that the agents communicate with . in design - world , unlike tileworld , an option that is generated via means - end reasoning or from proposals of other agents only becomes an intention if it is accepted by both agents .
see figure [ dial - stat - fig ] . in dialogue
, bill makes a proposal in 1 , and then kim implicitly accepts this proposal with a new proposal in 2 . in -3 bill rejects kim s proposal and makes a counter - proposal .
the content of communicative acts are beliefs and ( potential ) intentions .
dialogue illustrates part of the cycle for achieving a design - house plan : ( 1 ) individual agents means - end reason about options in the domain ; ( 2 ) individual agents deliberate about which options are preferable ; ( 3 ) then agents make proposals to other agents , based on the options identified in a reasoning cycle , about actions that contribute to the satisfaction of their intentions ; ( 4 ) then these proposals are accepted or rejected by the other agent , or acceptance / rejection is postponed by asking for more information
. see figure [ irma - fig ] .
deliberating whether to accept or reject a proposal is based on beliefs about the proposed action s utility @xcite .
agents parameterized with the all - implicit strategy do not include irus in any discourse act or produce any discourse acts labelled as potentially implicit in figure [ dial - stat - fig ] .
agents parameterized with the close - consequence and explicit - warrant strategies include irus at dialogue segment closings and in proposals .
in dialogue agent clc uses the close - consequence strategy .
clc makes explicit closing statements , such as -2 , on the completion of the intention associated with a discourse segment .
clc s closing discourse act also includes irus as in -3 ; clc makes the inference explicit that since they have agreed on putting the green rug in the study , bill no longer has the green rug ( act - effect inference ) .
the close - consequence strategy of making inferences explicit at the close of a segment is intended to parallel the naturally occurring example in [ so-400-examp ] . in both cases
an inference is made explicit that follows from what has just been said , and the inference is sequentially located at the close of a discourse segment . the explicit - warrant strategy varies the proposal discourse act by including warrant irus in each proposal . in general
a warrant for an intention is a reason for adopting the intention , and here warrants are the score propositions that give the utility of the proposal , which are mutually believed at the outset of the dialogues . in , the warrant iru is in caps .
since warrants are used by the other agent in deliberation , the explicit - warrant strategy can save the other agent the processing involved with determining which facts are relevant for deliberation and retrieving them from memory .
the explicit - warrant strategy also occurs in natural dialogues @xcite .
design - world supports the parameterization of the task so that it can be made more difficult to perform by making greater processing demands on the agents .
these task variations will be shown to interact with variations in communicative strategies and attentional capacity in section [ results - sec ] .
[ task - eval - fig ] the standard task is defined so that the raw score that agents achieve for a design - house plan , constructed via the dialogue , is the sum of all the furniture items for each valid step in their plan .
the point values for invalid steps in the plan are simply subtracted from the score so that agents are not heavily penalized for making mistakes .
the zero - invalids task is a fault - intolerant version of the task in which any invalid intention invalidates the whole plan . in general
, the effect of making a mistake in a plan depends on how interdependent different subparts of the problem solution are .
figure [ task - eval - inval - fig ] shows the choices for the effect of invalid steps for the design - world task .
the score for invalid steps ( mistakes ) can just be subtracted out ; this is how the standard task is defined .
alternately , invalid steps can propagate up so that an invalid put - act means that the design - room plan is invalid .
finally , mistakes can completely propagate so that the design - house plan is invalid if one step is invalid , as in the zero - invalids task .
the zero - nonmatching - beliefs task is designed to investigate the effect of the level of agreement that agents must achieve .
figure [ task - eval - fig ] illustrates different degrees of agreeing in a collaborative task , e.g. agents may agree on the actions to be done , but not agree on the * reasons * for intending that action .
the zero - nonmatching - beliefs task is defined so that a warrant w , a reason for doing p , must be mutually supposed .
we wish to evaluate the relative benefits of the communicative strategies in various tasks for a range of resource limits . in section [ task - def - effect - sec ]
we defined an objective performance measure for the design - house plan for each task variation .
we must also take cognitive costs into account . because cognitive effort can vary according to the communication situation and the agent architecture ,
performance evaluation introduces three additional parameters : ( 1 ) commcost : cost of sending a message ; ( 2 ) infcost : cost of inference ; and ( 3 ) retcost : cost of retrieval from memory : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [ cols= " < " , ] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ we simulate 100 dialogues at each parameter setting and calculate the normalized performance distributions for each sample run . in the results to follow , commcost , infcost and retcost are fixed at 1,1 , .01 respectively , and the parameters that are varied are ( 1 ) communication strategy ; ( 2 ) task definition ; and ( 3 ) awm settings .
differences in the performance distributions for each set of parameters are evaluated for significance over the 100 dialogues using the kolmogorov - smirnov ( ks ) two sample test @xcite . a strategy a is defined to be beneficial as compared to a strategy b , for a set of fixed parameter settings ,
if the difference in distributions using the kolmogorov - smirnov two sample test is significant at p @xmath1 .05 , in the positive direction , for two or more awm settings .
a strategy is detrimental if the differences go in the negative direction .
strategies may be neither beneficial or detrimental , since there may be no difference between two strategies .
a difference plot such as that in figure [ cost - clc - kim - fig ] will be used to summarize a comparison of strategy 1 and strategy 2 . in the comparisons below , strategy 1
is either close - consequence or explicit - warrant and strategy 2 is the all - implicit strategy .
* differences * in performance between two strategies are plotted on the y - axis against awm parameter settings on the x - axis .
each point in the plot represents the difference in the means of 100 runs of each strategy at a particular awm setting .
these plots summarize the information from 18 performance distributions ( 1800 simulated dialogues ) .
every simulation run varies the awm radius from 1 to 16 to test whether a strategy only has an effect at particular awm settings .
if the plot is above the dotted line for 2 or more awm settings , then strategy 1 may be beneficial , depending on whether the differences are significant . in the reminder of this section , we first compare within strategy , for each task definition and
show that whether or not a strategy is beneficial depends on the task .
then we compare across strategies for a particular task , showing that the interaction of the strategy and task varies according to the strategy .
the comparisons will show that what counts as a good collaborative strategy depends on cognitive limits on attention and the definition of success for the task .
the difference plot in figure [ cost - clc - kim - fig ] shows that close - consequence is detrimental in the standard task at awm of 1 @xmath0 5 ( ks @xmath2 0.19 , p @xmath1 .05 ) .
in contrast , if the task is the fault - intolerant zero - invalids task , then the close - consequence strategy is beneficial .
figure [ clc - inval - fig ] demonstrates that strategies which include consequence irus can increase the robustness of the planning process by decreasing the frequency with which agents make mistakes ( ks for awm of 3 to 6 @xmath2 .19 , p @xmath1 .05 ) .
this is a direct result of * rehearsing * the act - effect inferences , making it unlikely that attention - limited agents will forget that they have already used a furniture item .
figure [ clc - nmb - fig ] shows that the close - consequence strategy is detrimental when the task requires agents to achieve matching beliefs on the warrants for their intentions ( ks 1,3 ) @xmath2 0.3 , p @xmath1 .01 ) .
this is because irus displace other facts from awm . in this case
agents forget the scores of furniture pieces under consideration , which are the warrants for their intentions .
thus here , as elsewhere , we see that irus can be detrimental by making agents forget critical information .
figure [ ret - iei - fig ] shows that explicit - warrant is beneficial in the standard task at awm values of 3 and above . here
, the scores improve because the beliefs necessary for deliberating the proposal are made available in the current context with each proposal ( ks for awm of 3 and above @xmath2 .23 , p @xmath1 .01 ) , so that agents do nt have to search memory for them . at awm
parameter settings of 16 , where agents can search a huge belief space for beliefs to be used as warrants , the saving in processing time is substantial . when the task is zero - invalid ( no figure due to space ) , the benefits of the explicit - warrant strategy are dampened from the benefits of the standard task , because explicit - warrant does nothing to address the reasons for agents making mistakes .
in comparison with the all - implicit strategy , it is detrimental at awm of 1 and 2 , but is still beneficial at awm of 5,6,7 , and 11 .
in contrast to close - consequence , the explicit - warrant strategy is highly beneficial when the task is zero - nonmatching - beliefs , see figure [ iei - nmb - fig ] ( ks @xmath2 .23 for awm from 2 to 11 , p @xmath1 .01 ) .
when agents must agree on the warrants underlying their intentions , including these warrants with proposals is a good strategy even if the agent already knows the warrants .
this is due to agents resource limits , which means that retrieval is indeterminate and that there are costs associated with retrieving warrants from memory . at high
awm the differences between the two strategies are small .
design - world was inspired by the tileworld simulation environment : a rapidly changing robot world in which an artificial agent attempts to optimize reasoning and planning @xcite .
tileworld is a single agent world in which the agent interacts with its environment , rather than with another agent .
design - world uses similar methods to test a theory of the effect of resource limits on communicative behavior between two agents .
the belief reasoning mechanism of design - world agents was informed by the theory of belief revision and the multi - agent simulation environment developed in the automated librarian project @xcite . the communicative acts and
discourse acts used by design - world agents are similar to those used in @xcite .
design - world is also based on the method used in carletta s jam simulation for the edinburgh map - task @xcite .
jam is based on the map - task dialogue corpus , where the goal of the task is for the planning agent , the instructor , to instruct the reactive agent , the instructee , how to get from one place to another on the map .
jam focuses on efficient strategies for recovery from error and parametrizes agents according to their communicative and error recovery strategies .
given good error recovery strategies , carletta argues that ` high risk ' strategies are more efficient , where efficiency is a measure of the number of utterances in the dialogue .
while the focus here is different , we have shown that that the number of utterances is just one parameter for evaluating performance , and that the task definition determines when strategies are effective .
in this paper we showed that collaborative communicative behavior can not be defined in the abstract : what counts as collaborative depends on the task , and the definition of success in the task .
we used two empirical methods to support our argument : corpus based analysis and experimentation in design - world .
the methods and the focus of this work are novel ; previous work on resource limited agents has not examined the role of communicative strategies in multi - agent interaction whereas work on communication has not considered the effects of resource limits .
we showed that strategies that are inefficient under assumptions of perfect reasoners with unlimited attention and retrieval are effective with resource limited agents .
furthermore , different tasks make different cognitive demands , and place different requirements on agents collaborative behavior .
tasks which require a high level of belief coordination can benefit from communicative strategies that include redundancy .
fault intolerant tasks benefit from redundancy for rehearsing the effects of actions .
because the communicative strategies that we tested were based on a corpus analysis of human human financial advice dialogues and because variations in the design - world task were parametrized , we believe the results presented here may be domain independent , though clearly more research is needed . here
we fixed the parameters for the cost of communication , inference and retrieval , only discussed a few of the implemented discourse strategies , and did nt discuss design - world parameters that increase the inferential complexity of the task and that limit inferential processing .
elsewhere we show that : ( 1 ) when retrieval is free or when communication cost is high , that the explicit - warrant strategy is detrimental at low awm @xcite ; ( 2 ) some iru strategies are only beneficial when inferential complexity is higher than in the standard task @xcite ; ( 3 ) irus that make inferences explicit can help inference limited agents perform as well as logically omniscient ones @xcite .
one ramification of the results presented here is that experimental environments for testing agent architectures should support task variation @xcite .
furthermore the task variation should test aspects of the interaction of the agents involved .
these results also inform the design of multi - agent problem solving systems and for systems for teaching , advice and explanation .
a. chapanis , r.b .
ochsman , r.n .
parrish , and g.d . weeks .
studies in interactive communication : the effects of four communication modes on the behavior of teams during cooperative problem - solving . , 14:487509 , 1972 . marilyn walker . informational redundancy and resource bounds in dialogue . in _
aaai spring symposium on reasoning about mental states _ , 1993 .
also available as ircs techreport ircs-93 - 20 , university of pennsylvania . | _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ effective problem solving among multiple agents requires a better understanding of the role of communication in collaboration . in this paper
we show that there are communicative strategies that greatly improve the performance of resource - bounded agents , but that these strategies are highly sensitive to the task requirements , situation parameters and agents resource limitations .
we base our argument on two sources of evidence : ( 1 ) an analysis of a corpus of 55 problem solving dialogues , and ( 2 ) experimental simulations of collaborative problem solving dialogues in an experimental world , design - world , where we parameterize task requirements , agents resources and communicative strategies .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
* hamiltonian . *
we investigate the bilayer system where the top and bottom ferromagnetic ( fm ) layers are coupled antiferromagnetically by the exchange interaction , as illustrated in fig.[fig1]a .
the hamiltonian for each layer reads @xmath3+h_{\text{ddi } } , \label{hamilaf}\end{aligned}\ ] ] where @xmath4 is the layer index ( @xmath5 t , b ) , @xmath6 represents the local magnetic moment orientation normalized as @xmath7 at the site @xmath8 , and @xmath9 runs over all the nearest neighbor sites in each layer .
the first term represents the fm exchange interaction with the fm exchange stiffness @xmath10 .
the second term represents the dmi with the dmi vector @xmath11 .
the third term represents the pma with the anisotropic constant @xmath12 .
@xmath13 represents the dipole - dipole interaction .
there exists an afm coupling between the top and bottom layers , @xmath14 the sign of @xmath15 is negative reflecting that the interlayer interaction is antiferromagnetic .
we assume that the spins in the top layer are pointing upward .
then the spins in the bottom layer are pointing downward due to the interlayer afm couplings
. * topological number . *
the classical field @xmath16 is introduced for the spin texture in the fm system by considering the zero limit of the the lattice constant , @xmath17 .
the ground - state spin textures are @xmath18 and @xmath19 .
a magnetic skyrmion is a spin texture which has a quantized topological number .
spins swirl continuously around the core and approach the ground - state value asymptotically .
the skyrmion is characterized by the topological number @xmath20 in each layer , @xmath21 we obtain @xmath22 for a skyrmion in a sufficiently large area .
we also call @xmath20 the skyrmion number . even if the skyrmion spin texture is deformed , its skyrmion number does not change , as far as the boundary condition is not modified
. it can be neither destroyed nor separated into pieces , _
i.e. _ , a skyrmion is topologically protected .
the spins in the top and bottom layers are tightly bounded due to the afm coupling . accordingly ,
if one skyrmion is created in the top layer , a second skyrmion is also created in the bottom layer ( fig.[fig2]a ) simultaneously .
the topological number of the bottom layer is opposite to that of the top layer since all the spins are inverted , @xmath23 .
see fig.[fig2]b how these topological numbers evolve after the creation of a skyrmion in the top layer .
* excitation energy .
* we may determine numerically the spin profile of each skyrmion and estimate the excitation energy based on the hamiltonian @xmath24 .
we compared the energy of one bilayer - skyrmion state and the energy of the homogeneous state in fig.[fig2]c as a function of the dmi .
similar result is obtained for any value of the interlayer afm coupling strength .
this is because the spin directions between the top and bottom layers are always opposite .
the energy of the bilayer - skyrmion becomes lower than that of the uniform ground state for @xmath25 mj m@xmath26 , which is the threshold value .
consequently , the bilayer - skyrmion excitation is energetically favorable when the dmi strength is larger than the threshold value .
however , spontaneous generation of skyrmions does not occur due to the topological protection .
nevertheless , the topological protection can be violated in condensed matter physics because of the existence of the lattice structure and the boundary of the sample .
we may create skyrmions by leveraging these properties .
* skyrmion hall effect . *
it is well understood@xcite that the center - of - mass motion of a skyrmion is determined by the lagrangian , @xmath27 where @xmath28 is the berry phase term of the spin texture , @xmath29 and @xmath30 is the potential . here , @xmath31 is the gyromagnetic coupling constant representing the magnus force with @xmath32 , and @xmath33 is the center - of - mass coordinate of the skyrmion .
the euler - lagrange equation yields @xmath34 where @xmath35 , and @xmath36 is the force acting on the skyrmion .
consequently , a moving skyrmion feels the magnus force and bends toward the direction perpendicular to the force .
it is called the skyrmion hall effect .
the direction of the magnus force is opposite when the sign of the skyrmion number @xmath37 is opposite .
for instance , if a skyrmion is driven by the current along the @xmath38 direction it will be bended toward the @xmath39 ( @xmath40 ) direction in the top ( bottom ) layer .
we may apply a cpp spin - polarized current injection from magnetic tunnel junction ( mtj ) or the spin hall effect in heavy metal layer@xcite .
we numerically solve the landau - lifshitz - gilbert - slonczewski ( llgs ) equation , which governs the dynamics of the magnetization @xmath41 at the lattice site @xmath8 . by suppressing the layer index
, it reads @xmath42where @xmath43 is the effective magnetic field induced by the hamiltonian @xmath24 , @xmath44 is the gilbert gyromagnetic ratio , @xmath45 is the gilbert - damping coefficient originating from spin relaxation , @xmath46 is the slonczewski stt coefficient , @xmath47 is the out - of - plane stt coefficient , and @xmath48 represents the electron polarization direction . here , @xmath49 with @xmath50 the vacuum magnetic permittivity , @xmath51 the film thickness , @xmath52 the saturation magnetization , and @xmath53 the current density .
we take @xmath54 direction for creating the skyrmion , while @xmath55 direction for moving the skyrmion .
the stt is induced either by injection from a magnetic tunnel junction polarizer or by the spin hall effect@xcite .
alternatively we may apply a cip injection@xcite to move skyrmions .
we numerically solve the llg equation , @xmath56,\end{aligned}\ ] ] where @xmath57 is the strength of the non - adiabatic torque and @xmath58 is the lattice constant .
* creation of a bilayer - skyrmion by vertical spin current .
* we employ a cpp injection with a circular geometry in a nanodisk .
the spin - polarized current ( polarized along @xmath59 is injected into the top layer in the central circle region , as illustrated in fig.[fig1]a .
we demonstrate how the spin textures develop in fig.[fig2]a .
the spins start to flip in both layers following the spin current injection only in the top layer . when there is no interlayer afm coupling , a skyrmion is formed only in the top layer ( ) . by contrast , a skyrmion is formed also in the bottom layer upon the current injection in the presence of the interlayer afm coupling ( ) .
we show the evolution of the skyrmion number in fig.[fig2]b .
it oscillates at the initial stage for @xmath60 ns , and rapidly increases to @xmath61 .
the skyrmion remains stable even when the current is switched off , demonstrating that it is topologically protected . during this process ,
the spins in the top and bottom layers are always anti - parallel .
a comment is in order .
the saturated skyrmion number is not exactly @xmath62 but @xmath63 .
this is due to the fact that there is a background skyrmion number which originates from the tilting the edge spins .
it is @xmath64 .
accordingly , the calibrated skyrmion number is @xmath65 , which is almost unity .
we present a nucleation phase diagram of a bilayer - skyrmion pair as a function of the current density and the interlayer afm coupling in fig.[fig2]d and fig.[fig2]e . when the magnitude of the injected current is strong enough
, the bilayer - skyrmion is created .
this is due to the fact that spin flip costs a certain energy . on the other hand , if the interlayer afm coupling is too strong , the bilayer - skyrmion is suppressed due to the fact that the nucleation field and the coercivity increases with the interlayer afm exchange , leading to a larger current density for nucleation .
* creation of a bilayer - skyrmion from a bilayer dw pair . *
a magnetic skyrmion can be created from a dw pair by using a junction geometry@xcite . in this scenario
we first make a dw pair into a nanotrack of the top layer through the local cpp injection with @xmath54 direction .
we show how the spins start to flip in the top layer and subsequently in the bottom layer driven by the afm exchange force in fig.[fig3]a .
then , the bilayer dw pair is shifted by applying cpp current , as shown in the process from @xmath66 ps to @xmath67 ps in fig.[fig3]b .
here we consider the vertical injection of a spin current towards @xmath68 and polarized along @xmath55 in the bottom layer .
the cpp injection moves the bilayer dw in the rightward direction .
when the bilayer dw arrives at the junction interface ( @xmath69 ps ) , both the end spins of the dw are pinned at the junction , whereas the central part of the dw continues to move due to stt in the wide part of the nanotrack .
therefore , the structure is deformed into a curved shape and a bilayer - skyrmion texture forms at @xmath70 ps ( ) .
* current - driven motion of a bilayer - skyrmion in a nanotrack . *
the magnetic bilayer - skyrmion can be displaced by the vertical spin - polarized current as in the case of the magnetic skyrmion .
we may employ the cpp injection or the cip injection to drive a bilayer - skyrmion . in general , a moving skyrmion
is easily destroyed by the sample edges due to the skyrmion hall effect .
therefore , the maximum velocity of skyrmion in fm nanotrack is typically much less than 10@xmath71 m / s , limited by the edge confining force@xcite of @xmath72 .
the skyrmion in the top layer follows the motion of the skyrmion in the bottom layer even when the current is not injected into the top layer .
this is because that two skyrmions are bounded by the interlayer afm coupling .
there is no skyrmion hall effect for a magnetic bilayer - skyrmion .
this can be explained as follows .
if there is no interlayer afm coupling , a skyrmion in the top layer moves left - handed and the skyrmion in the bottom layer moves right - handed .
however , when the afm coupling is strong enough , two skyrmions are tightly bounded and the magnus forces acting on the skyrmions between the top and bottom layers are exactly cancelled .
accordingly the bilayer - skyrmion will move straightly .
this mechanism works both for the cpp and cip cases ( ) .
we show the relation between the magnitude of the injected current and the velocity in fig.[fig4 ] .
the velocity is proportional to the injected current density . for strong enough current
, the bilayer - skyrmion is destroyed and split into two independent skyrmions .
this is because that the skyrmion hall effect increases as the current increases , which acts as the repulsive force between two skyrmions ( ) . with strong interlayer afm exchange coupling ,
the coupled skyrmions move along the central line of the nanotrack at a high speed of a few hundred meters per second , without any transverse motion .
however , with small interlayer afm exchange coupling , the skyrmions in the top and bottom layers will be decoupled due to the fast motion of the skyrmion in the bottom layer driven by large current .
once the skyrmions in the top and bottom layers are decoupled , the skyrmion hall effect becomes active , leading to the destruction of skyrmions in the top and/or bottom layer by touching edge . at the same time , in the cpp case , the skyrmion in the top layer will stop motion .
it can be seen that the coupled skyrmions driven by cip current does nt require a fine tuning of damping and non - adiabatic torque coefficients .
it is also worth noting that the interlayer afm coupling does not produce a mass of the bilayer - skyrmion .
when the driving current is suddenly turned off , the bilayer - skyrmion stops high - speed motion immediately ( see supplementary movie 8) . on the other hand
, we also investigated the case where the dmi constant @xmath73 is different between the top and bottom layers .
it is found that the results of current - driven motion of the bilayer - skyrmion do not change much since the dmi only changes the radius of the skyrmion ( see supplementary movies 9 - 10 ) .
the massless property and its robustness make the bilayer - skyrmion an ideal candidate for practical applications .
we have presented a novel solution of inhibiting the hall effect of skyrmions without affecting their topological properties , by exploring a new device made of antiferromagnetically exchange - coupled bilayer nanodisks and nanotracks . compared with the mostly investigated skyrmion in the fm monolayer system ,
the bilayer - skyrmion exhibits entirely distinct characteristics with regard to the current - transport behaviour and robustness .
first , it can move strictly along the direction of the spin current flow , which makes it more appealing for motions in nanowires for ultradense memory applications .
this is in high contrast with the case of monolayer skyrmion , where the skyrmion information carrier can be easily destroyed by the edges of the nanotracks .
second , it will be immune to magnetic field perturbations which might be generated externally or internally within the device circuitry since the net magnetic moment is zero .
third , by introducing the bilayer degree of freedom , the transport properties of the device can be engineered to achieve desired performance .
for example , the skyrmion hall effect can be intentionally suppressed or enhanced by tuning the magnetic properties of individual layers .
this newly proposed solution of transporting skyrmion information carrier for arbitrarily long distances at much enhanced velocity may be very appealing for versatile applications such as ultradense memory and information processing .
similar ideas can be extended to multilayer or superlattice where the skyrmions are strongly coupled to realize a better manipulation of skyrmions in nanotrack or extended thin films .
* modeling and simulation . * the micromagnetic simulations are performed using the object oriented micromagnetic framework ( oommf ) including the dzyaloshinskii - moriya interaction ( dmi ) extended module @xcite .
the time - dependent magnetization dynamics is governed by the landau - lifshitz - gilbert ( llg ) equation including spin torque@xcite .
the average energy density @xmath74 is a function of @xmath75 , which contains the intralayer exchange , the interlayer exchange , the anisotropy , the applied field ( zeeman ) , the demagnetization and the dmi energy terms . for micromagnetic simulations ,
the intrinsic magnetic parameters are adopted from refs.@xcite : gilbert damping coefficient @xmath76 and the value for gilbert gyromagnetic ratio is -2.211@xmath77 10@xmath78 m a@xmath79 s@xmath79 .
saturation magnetization @xmath80 = 580 ka m@xmath79 , intralayer exchange stiffness @xmath81 = 15 pj m@xmath79 , dmi constant @xmath73 = 0 @xmath82 6 mj m@xmath26 and perpendicular magnetic anisotropy ( pma ) @xmath12 = 0.8 mj m@xmath83 unless otherwise specified .
the interlayer exchange coefficient @xmath84 is set from 0 to -10 pj m@xmath79 , whereas the corresponding interface exchange coefficient @xmath85 equals from 0 to -10 mj m@xmath26 ( @xmath86 = @xmath15 / 1 nm ) , where `` - '' denotes that the interface is antiferromagnetically coupled . the field - like out - of - plane stt coefficient @xmath47 is set to zero .
all samples are discretized into cells of 2 nm @xmath77 2 nm @xmath77 1 nm in the simulation , which is sufficiently smaller than the typical exchange length ( @xmath82 4.3 nm ) and the skyrmion size to ensure the numerical accuracy . for all simulation of current - driven skyrmions
reported throughout this paper , the skyrmions are firstly created at the designed spot of the nanotrack ( @xmath87 = 100 nm ) by a local spin current perpendicular to the plane of the top layer . then the system is relaxed to an energy minimum state without applying any current .
next , we start the timer and the spin current ( @xmath88 = 0.4 ) is injected into the nanotrack with the geometry of current - in - plane ( cip ) or current - out - of - plane ( cpp ) as shown in fig.[fig1 ] . in the configuration of cip , the electrons flow toward the right in both the top and bottom layers , @xmath89 , the currents flow toward the left , while in the configuration of cpp , the electrons flow toward the top only in the bottom layer . 99 s. mhlbauer et .
science * 323 * , 915 ( 2009 ) .
x. z. yu et .
nature * 465 * , 901 ( 2010 ) .
s. heinze et .
nature phys .
* 7 * , 713 ( 2011 ) .
n. nagaosa and y. tokura , nat .
* 8 * , 899 ( 2013 ) .
a. fert , v. cros , and j. sampaio , nat . nanotech .
* 8 * , 152 ( 2013 ) .
j. sampaio , v. cros , s. rohart , a. thiaville , and a. fert , nat . nanotech .
* 8 * , 839 ( 2013 ) .
y. tchoe and j. h. han , phys .
b * 85 * , 174416 ( 2012 ) .
y. zhou , e. iacocca , a. awad , r. k. dumas , f. c. zhang , h. b. braun , and j. akerman , cond - mat / arxiv:1404.3281 .
m. finazzi , m. savoini , a. r. khorsand , a. tsukamoto , a. itoh , l. duo , a. kirilyuk , th .
rasing , and m. ezawa , phys .
rev . lett .
* 110 * , 177205 ( 2013 ) .
x. c. zhang , m. ezawa , and y. zhou , sci
. rep . * 5 * , 9400 ( 2015 ) .
x. c. zhang , g. p. zhao , h. fangohr , j. p. liu , w. x. xia , j. xia , and f. j. morvan , sci
. rep . * 5 * , 7643 ( 2015 ) . c. moreau - luchaire , c. moutafis , n. reyren , j. sampaio , n. van horne , c.a.f .
vaz , k. bouzehouane , k. garcia , c. deranlot , p. warnicke , p. wohlhter , j.m .
george , j. raabe , v. cros , and a. fert , arxiv:1502.07853 ( 2015 ) .
s. woo , k. litzius , b. krger , m. y. i m , l. caretta , k. richter , m. mann , a. krone , r. reeve , m. weigand , p. agrawal , p. fischer , m. klui , g. s. d. beach , arxiv:1502.07376 ( 2015 ) .
s. parkin , and s. yang , nat .
* 10 * , 195 ( 2015 ) .
n. bogdanov and d. a. yablonskii , sov .
jetp * 68 * , 101 ( 1989 ) . n. bogdanov and a. j. hubert , magn .
magn . mater . *
138 * , 255 ( 1994 ) .
u. k. roessler , n. bogdanov , and c. pfleiderer , nature * 442 * , 797 ( 2006 ) .
j. iwasaki , w. koshibae , and n. nagaosa , nano lett .
, * 14 * , 4432 ( 2014 ) .
j. iwasaki , m. mochizuki , and n. nagaosa , nat .
* 8 * , 742 ( 2013 ) .
j. iwasaki , m. mochizuki , and n. nagaosa , nat .
commun . * 4
* , 1463 ( 2012 ) .
s. h. yang , k. s. ryu , and s. parkin , nat .
* 10 * , 221 ( 2015 ) .
m. stone , phys .
b * 53 * , 16573 ( 1996 ) . a.v .
khvalkovskiy , v. cros , d. apalkov , v. nikitin , m. krounbi , k.a .
zvezdin , a. anane , j. grollier , and a. fert , phys .
b * 87 * , 020402(r ) ( 2013 ) .
r. tomasello , e. martinez , r. zivieri , l. torres , m. carpentieri , and g. finocchio , sci .
rep . * 4 * , 6784 ( 2014 ) .
y. zhou and m. ezawa , nat . com . * 5 * , 4652 ( 2014 ) .
m. j. donahue and d. g. porter , national institute of standards and technology , interagency report nistir , * 6376 * ( 1999 oommf user s guide , version 1.0 . ) .
s. rohart and a. thiaville , phys .
b * 88 * , 184422 ( 2013 ) . w. f. brown j. , micromagnetics ( krieger , new york , 1978 ) .
t. l. gilbert , phys .
100 * , 1243 ( 1955 ) .
l. landau and e. lifshitz , physik .
z. sowjetunion * 8 * , 153 ( 1935 ) .
a. thiaville , s. rohart , e. jue , v. cros , and a. fert , europhys . lett . * 100 * , 57002 ( 2012 ) .
a. thiaville , y. nakatani , j. miltat , and y. suzuki , europhys . lett .
* 69 * , 990 ( 2005 ) .
y.z . thanks the support by the seed funding program for basic research and seed funding program for applied research from the hku , itf tier 3 funding ( its/171/13 ) , the rgc - grf under grant hku 17210014 , and university grants committee of hong kong ( contract no .
aoe / p-04/08 ) .
thanks the support by the grants - in - aid for scientific research from the ministry of education , science , sports and culture , no . 25400317 .
is very much grateful to n. nagaosa for many helpful discussions on the subject .
m.e . conceived the idea and designed the project .
m.e . and y.z .
coordinated the project .
x.z . performed the numerical simulations supervised by y.z .
all authors discussed the results and wrote the manuscript .
supplementary information is available .
the authors declare no competing financial interests . | * arising from emergent electromagnetic field of magnetic skyrmions due to their nontrivial topology , the skyrmion hall effect might be a roadblock for practical applications since any longitudinal motions of skyrmions in nanotrack is accompanied by a transverse motion .
a direct consequence of such an effect is easy destruction of skyrmions at the nanotrack edges during their fast motions along the nanotrack , despite their topological protection .
here we propose an entirely novel solution of completely inhibiting such skyrmion hall effect without affecting its topological properties based on a antiferromagnetic - coupling bilayer system .
we show that a pair of magnetic skyrmions can be nucleated in such a bilayer system through vertical current injection or converted from a current - driven domain - wall pair .
once nucleated , the skyrmion pair can be displaced through current - induced spin torque either from a vertical injected current or in - plane current .
the skyrmion hall effect is completely suppressed due to the cancellation of back - action forces acting on each individual skyrmion , resulting in a straight and fast motion of skyrmions along the current direction .
this proposal will be of fundamental interests by introducing the bilayer degree of freedom into the system .
moreover , it provides an easy way to engineer the transport properties of the skyrmionic devices to achieve desired performance , making it highly promising for practical applications such as ultradense memory and information - processing devices based on skyrmions . * since
the first experimental observations of magnetic skyrmion lattices in bulk non - centrosymmetric magnets@xcite and films@xcite , there has been tremendous interest in these topologically protected spin configurations with a quantized topological number@xcite .
however , the creation and transmission of isolated magnetic skyrmion in magnetic thin films are required for any practical applications such as encoding information in individual skyrmion to allow entirely novel devices and circuitry@xcite .
significant efforts and progress have been made towards realizing such ultrathin film based on perpendicularly magnetized magnetic layer@xmath0heavy metal structure to host a rich variety of chiral spin textures including skyrmions@xcite .
the strong spin orbit coupling at the interfaces between the magnetic layer with perpendicular magnetic anisotropy ( pma ) and the underlying heavy metal layer provides a sizeable dzyaloshinsky - moriya interaction ( dmi ) to stabilize skyrmions@xcite . to move a skyrmion in nanotrack for information - processing applications ,
a convenient and efficient way is by means of spin current which can transfer the angular momentum from itinerant conduction electrons to the magnetic moments of the skyrmion@xcite . however , one major roadblock to the manipulation and transmission of skyrmions in nanotrack is the skyrmion hall effect , _
i.e. _ , skyrmions exhibit the hall effect driven by spin currents due to the presence of the magnus force which in turn originates from its nontrivial topology@xcite .
thus a skyrmion will not move parallel with the direction of the current .
instead it will gain a transverse velocity with a magnitude proportional to the spin current density which displaces the skyrmion towards the edge of the nanotrack .
therefore a skyrmion will be easily destroyed for a distance of much less than 1@xmath1 m for a nanotrack made of typical magnetic layer@xmath0heavy metal system@xcite . since both the skyrmion hall effect and
its topological protection arise from the same mechanism , _
i.e_. , its nontrivial topological number of @xmath2 , it seems to be impossible to inhibit the skyrmion hall effect without breaking its topological protection . in this work
, we propose a novel solution of two perpendicularly magnetized sublayers strongly coupled via the antiferromagnetic ( afm ) exchange interaction with a heavy metal layer beneath the bottom magnetic layer@xcite ( fig.[fig1]a , b , c ) .
when a skyrmion is created in one of the sublayers , another skyrmion is simultaneously created in the other sublayer under certain conditions .
we refer to such a pair of magnetic skyrmions as a magnetic bilayer - skyrmion ( fig.[fig1]d , e ) .
moreover , we show a bilayer - skyrmion can be displaced over arbitrarily long distances driven by spin currents without touching the edges due to the absence of the skyrmion hall effect , which is distinct from a skyrmion in the monolayer thin - film structure .
in addition to address the above - mentioned dilemma , bilayer - skyrmions have many distinct characteristics from conventional skyrmions in monolayer thin film , which allow for versatile and multifunctional ultradense and ultrafast information processing and logic applications .
these features are true for both the current - in - plane ( cip ) and current - perpendicular - to - plane ( cpp ) cases .
the cip implementation has been extensively studied in the past ten years for displacing domain walls or switching magnetization .
recently , the spin hall effect has been demonstrated to be a more efficient means of manipulating magnetization .
the current for the cip case is along the nanowire axial direction .
by contrast , the spin current is perpendicular to the heavy metal / ferromagnetic interface for the cpp case . in our nomenclature for cpp ,
the spin current is perpendicular - to - film plane whereas the charge current can be either in the film plane ( such as the spin hall effect scenario ) or perpendicular to the film plane ( such as the perpendicular mram where a perpendicularly magnetized polarizer is incorporated ) .
the cpp scheme has been proven to be much more efficient to move the domain wall or skyrmion than the cip method@xcite . in this work
, we first describe the nucleation process of an isolated skyrmion within the top layer by utilizing the cpp injection into a disk composed of two afm - coupled magnetic sublayers . owing to the interlayer afm exchange coupling
, another skyrmion will automatically emerge in the bottom layer .
in so doing we explore various properties of a bilayer - skyrmion .
furthermore , we move a bilayer - skyrmion in the nanotracks either by the cpp or cip .
we also study the generation of a domain - wall ( dw ) pair and the conversion process from a dw pair to a skyrmion in the bilayer nanotracks . |
the study of extreme eigenvalue statistics in random matrix theory ( rmt ) has attracted much attention during the last twenty years . in particular , the tracy - widom ( tw ) distributions @xcite describing the largest eigenvalue @xmath9 ( as well as the smallest one , @xmath10 ) in the classical gaussian ensembles , orthogonal ( goe , @xmath11 ) , unitary ( gue , @xmath12 ) and symplectic ( gse , @xmath13 ) where @xmath14 is the dyson index have become cornerstones of the theory of extreme value statistics of strongly correlated variables .
quite remarkably , it was shown that the tw distributions , denoted by @xmath15 in the following , appear in a wide variety of problems @xcite , a priori not directly related to rmt , ranging from the longest increasing sequence of random permutations of integers @xcite , stochastic growth and related directed polymer models in the kardar - parisi - zhang @xcite universality class @xcite and sequence alignment problems @xcite to non - intersecting interfaces and brownian motions @xcite as well as in finance @xcite . while the tw distribution describes the typical fluctuations of @xmath9 and @xmath10 , a large body of work has also been devoted to the study of large deviations of extreme eigenvalues in gaussian ensembles @xcite .
another interesting ensemble of random matrices , which we focus on in this paper , is the so - called wishart - laguerre ensemble here we focus on the case of complex matrices ( @xmath16 ) .
let @xmath1 be a @xmath2 rectangular matrix with i.i.d .
complex gaussian entries and @xmath17 .
the wishart - laguerre matrix @xmath18 is defined as @xmath19 which is thus a @xmath20 hermitian matrix , having @xmath4 real and positive eigenvalues @xmath21 .
the joint probability density function ( pdf ) of these @xmath4 eigenvalues is given by @xcite @xmath22 where @xmath23 , @xmath24 ... @xmath25 are positive and where @xmath26 is a normalization constant , depending on @xmath27 .
these wishart - laguerre matrices play an important role in statistics , in particular in principal component analysis , where the matrix @xmath18 is a covariance matrix .
hence in this case both @xmath5 and @xmath4 , and thus @xmath28 , are positive integers . however , this joint pdf in eq .
( [ pjoint ] ) is well defined for any real value of @xmath29 , which , for some non - integer value of @xmath27 , may have some physical applications .
an interesting example is the case of non - intersecting brownian excursions @xcite ,
i.e. , @xmath4 non - colliding positive brownian paths @xmath30 on the unit time interval @xmath31 $ ] constrained to start and end at the origin @xmath32 . the joint pdf of the positions of the @xmath4 walkers , at a given time @xmath33 , can indeed be written as @xcite @xmath34 where @xmath35 and @xmath36 is a normalization constant . from this expression ( [ pdf_watermelons ] ) we obtain that the scaled variables @xmath37 behave statistically like the eigenvalues of random matrices from the wishart - laguerre ensemble ( [ pjoint ] ) with a non - integer parameter @xmath38 . hence it is physically relevant to study the joint distribution in eq .
( [ pjoint ] ) for any real @xmath29 .
a first important characteristic associated to this ensemble ( [ pjoint ] ) is the mean density of eigenvalues , @xmath39 , defined by @xmath40 where the overline denotes an average over the different realizations of the random variables @xmath41 s according to the joint pdf in eq .
( [ pjoint ] ) . in the large @xmath4 limit
, it is well known that @xmath39 is given by the marchenko - pastur ( mp ) distribution ( see fig .
[ fig : wishart_soft_vs_hard ] ) @xmath42 where @xmath43 are the right and left edges of the support , with @xmath44 .
the case where @xmath45 corresponds to the case where @xmath46 . here
, we mainly focus on the case @xmath47 which corresponds instead to the case where @xmath27 is finite , while both @xmath4 and @xmath5 are large . in this case , which we will mainly focus on in this paper , the mp distribution takes the particular form ( see the left panel of fig . [
fig : wishart_soft_vs_hard ] ) @xmath48 at the right edge , near @xmath49 , @xmath50 vanishes as a square - root , @xmath51 and therefore the fluctuations near this _ soft _ edge are governed , for large @xmath4 , by the airy kernel .
in particular , the distribution of the largest eigenvalue @xmath9 , appropriately shifted and scaled , converges , when @xmath52 , to the tw distribution @xmath53 mentioned above ( the same as for gue ) @xcite .
it is now well known that this distribution can be expressed in terms of a special solution of a painlev ii equation . while this connection to painlev transcendents was initially obtained by tracy and widom using fredholm operator techniques , nadal and majumdar @xcite provided , more recently , a derivation of this result ( for @xmath54 ) using semi - classical orthogonal polynomials ( ops ) , see also ref .
this method is at the heart of the present paper .
on the other hand , at the left edge , @xmath55 , @xmath50 has a square - root singularity , @xmath56 ( see the left panel of fig .
[ fig : wishart_soft_vs_hard ] ) .
what about the fluctuations of the smallest eigenvalue @xmath10 in this case ?
one can estimate the typical scale of @xmath10 , for large @xmath4 , by considering that there is typically one eigenvalue in the interval @xmath57 $ ] , i.e. , @xmath58 which implies that @xmath59 . within this scale @xmath60 ,
the fluctuations are governed by the bessel kernel @xcite .
furthermore , it has been shown that the distribution of @xmath61 converges to a limiting form which ( i ) is different from the standard tw distribution @xmath53 and depends continuously on the exponent @xmath27 in ( [ pjoint ] ) and ( ii ) can be written in terms of a special solution of a painlev iii equation @xcite . if one introduces @xmath62 then , one has indeed @xmath63 where @xmath64 is the unique solution of a painlev iii @xcite : @xmath65 satisfying @xmath66 this result ( [ eq : result_tw ] ) was shown by tracy and widom using fredholm operator techniques @xcite . note that for integer values of @xmath27 , the limiting distribution @xmath67 can be written as an @xmath68 determinant whose entries are expressed in terms of bessel functions ( see eq .
( [ fdetbessel ] ) below ) a result which can be obtained by clever manipulations of determinants @xcite .
in particular for @xmath69 the result is extremely simple as @xmath70 for all @xmath4 , implying @xmath71 , which is obviously solution of eq .
( [ piii ] ) with the boundary condition ( [ asympt_piii ] ) .
the limiting distribution of @xmath10 for complex wishart matrices , in the limit @xmath52 , is thus well known ( [ eq : result_tw ] ) we refer the reader to ref .
@xcite to a recent work on the smallest eigenvalue for real wishart matrices in the hard edge limit . what about the finite @xmath4 corrections to this asymptotic form ?
such a question is quite natural for practical applications of extreme value statistics ( evs ) , where one always deals with finite samples here matrices of finite size .
this issue was recently revisited for evs of independent and identically distributed random variables using a renormalization group approach @xcite . for evs of strongly correlated variables , there are actually few cases where these corrections have been worked out explicitly , including random walks @xcite , the _
largest _ eigenvalue of random matrices belonging to various ensembles @xcite , non - intersecting brownian motions @xcite or ( poissonized ) random matchings @xcite . for _ real _ wishart matrices , the first corrections to the limiting distribution of the _ smallest _ eigenvalue in the soft edge limit were studied in ref .
@xcite where it was shown that corrections to the limiting distribution of @xmath10 and @xmath9 are quite different , although the limiting distributions for both observables are actually the same , namely the tw distribution for goe , @xmath72 . what about the corrections to the limiting distribution @xmath73 in eq . ( [ eq : result_tw ] ) of @xmath10 for complex wishart matrices in the hard edge limit ?
this question was recently raised by edelman , guionnet et pch @xcite in their study of finite size covariance matrices with non - gaussian entries .
based on the large @xmath4 expansion of the exact formulas obtained in ref .
@xcite for small integer values of @xmath27 , they conjectured the following form of the first @xmath7-correction @xmath74 note that this first @xmath7 correction in eq .
( [ conjecture_finite_n ] ) can be interpreted as a correction to the width , i.e. , @xmath75 it is interesting to notice @xcite that for most of the cases which have been studied in rmt @xcite , it was actually found that the first order correction to the limiting distribution of extreme eigenvalue actually corresponds to a correction of the scaling variable , as in eq .
( [ eq : correction ] ) .
one exception concerns the smallest eigenvalue of real wishart matrices in the soft - edge limit , where the first correction has a more complicated structure @xcite .
the main goal of this paper is to provide an explicit computation of this first correction in the hard edge limit and we will show that it has indeed the conjectured form given above in eq .
( [ conjecture_finite_n ] ) . to perform this computation
, we will use a method relying on semi - classical ops , in the spirit of refs . @xcite and @xcite . as we will see , our method does not only allow us to compute explicitly the first @xmath7 corrections but provides also a rather straightforward derivation of the expression for the limiting distribution @xmath67 in terms of the solution of a painlev iii equation , without using fredholm operators theory but relying instead only on the recurrence relations associated to the ( semi - classical ) ops system .
finally , we will also study the first finite @xmath4 corrections to the limiting distribution of @xmath10 at the soft edge .
note that after the results obtained in the present paper were presented in a conference @xcite , another independent proof of the conjecture in eq .
( [ conjecture_finite_n ] ) was achieved in ref .
@xcite , using operator theoretic techniques .
more recently , yet another independent proof of this conjecture was given in ref .
the distribution of the smallest eigenvalue @xmath76 is given by @xmath77 in this paper , we will compute @xmath78 using semi - classical ops @xmath79 which are polynomials of the variable @xmath80 while @xmath33 and @xmath27 are parameters ( for the sake of clarity in the notations , this dependence is omitted here ) : @xmath81 the cumulative distribution @xmath78 can be expressed , using standard manipulations , in terms of the norms @xmath82 s as @xmath83 as we will see , the norms @xmath82 s can be computed from the three term recurrence relation satisfied by the ops @xmath84 from which we deduce the following important relations : @xmath85 note that the starting point of our analysis is very similar to the one of basor and chen in ref . @xcite but the analysis of the recursion relations is different . in particular , we do not make use of ladder operators techniques , which are heavily used in ref . @xcite . in addition , we provide an asymptotic analysis of this op system ( [ eq : def_op ] ) for large @xmath4 , beyond the leading order . in section [ section : op ] , we will study the variables @xmath86 , @xmath87 , @xmath82 and @xmath88 .
in particular , we will show that @xmath86 and @xmath87 satisfy a coupled set of equations , named schlesinger equations in the literature on ops @xmath89 this system of equations , together with the initial condition given in eq .
( [ ci ] ) below , determines uniquely the values of @xmath87 and @xmath86 for all values of @xmath90 , as they can be computed by induction .
it is however quite difficult to analyze the large @xmath4 behavior of @xmath91 and @xmath92 using only this set of equations ( [ schlesinger_intro ] ) . to circumvent this difficulty ,
it is customary to use another set of relations , called the laguerre - freud equations , which we derive here using the method based on turn determinants , as developed in ref .
they read @xmath93 some details of this derivation have been relegated in [ appendix : turan ] . in section [ section : pv ] , by manipulating these two sets of equations ( [ schlesinger_intro ] ) and ( [ lf_intro ] ) , we show that @xmath78 , for finite @xmath4 , is related to a special solution of a painlev v equation [ see eqs .
( [ def : hn ] , [ painleve5 ] ) ] , thus recovering a previous result of tracy and widom .
section [ section : piii ] is devoted to the hard edge scaling limit : @xmath94 it is known that the limiting distribution @xmath67 can be obtained by analyzing the large @xmath4 limit of this painlev v equation in eqs .
( [ def : hn ] , [ painleve5 ] ) , obtained for finite @xmath4 , leading to ( [ eq : result_tw ] ) and ( [ piii ] ) . as we show in section 5 , the finite @xmath4 corrections
are then easily obtained from the schlesinger equation ( [ schlesinger_intro ] ) , from which we obtain eq .
( [ conjecture_finite_n ] ) . finally , section [ section : pii ] is devoted to the soft edge scaling limit , when @xmath46 , for @xmath4 large .
to study this op system ( [ eq : def_op ] ) it is useful @xcite to introduce a deformation parameter @xmath95 and study the following op system @xmath96 such that the norms @xmath82 s are defined by @xmath97 as we show here , some useful relations can be obtained by varying @xmath95 .
eventually , we will of course set @xmath98 ( [ eq : def_op ] ) .
the first polynomials can be computed from ( [ eq : op_deformed ] ) to obtain @xmath99 where @xmath100 is the incomplete gamma function .
obviously , the expression of the ops @xmath101 becomes more and more complicated as @xmath90 grows .
the polynomials @xmath101 being ops , they satisfy a three - term recursion relation , which can be obtained as follows . as @xmath102 is a polynomial of degree @xmath103 , we can expand it on the basis of these ops .
because @xmath104 if @xmath105 , we can write the three term recurrence relation @xcite : @xmath106 where , by definition @xmath107 from ( [ first_polynomials0 ] ) and ( [ first_polynomials1 ] ) , we can compute the first terms @xmath108 in this section , we derive a couple of recursion relations called the schlesinger equations that couple @xmath87 and @xmath86 .
we first write @xmath109 therefore , using eq .
( [ 3term_rec ] ) with the substitution @xmath110 , we have @xmath111 and finally we obtain the standard relation @xmath112 on the other hand , using the definition of the scalar product in ( [ eq : op_deformed ] ) , we have @xmath113 from which we deduce the relation between @xmath86 and @xmath82 @xmath114 by combining eq .
( [ s_en_h ] ) and eq .
( [ r_en_h ] ) , we obtain straightforwardly @xmath115 we now study the coefficient @xmath88 of the term of degree @xmath116 in the polynomial @xmath101 [ see eq .
( [ eq : def_op ] ) ] : @xmath117 taking the derivative of this equation with respect to ( w.r.t . ) @xmath95 we obtain @xmath118 multiplying both sides of eq .
( [ eq : deriv_zeta ] ) by @xmath80 and projecting on @xmath101 yields @xmath119 besides , by looking at the term of degree @xmath90 , i.e. @xmath120 , in eq .
( [ 3term_rec ] ) we find a recursion relation between @xmath86 and @xmath88 @xmath121 which can be solved for @xmath88 , using the initial condition ( [ ci0 ] ) , to get @xmath122 furthermore , by differentiating eq .
( [ snhn ] ) w.r.t .
to @xmath95 we have @xmath123 where , in the second line , we have simply used the definition of the scalar product in ( [ eq : def_op ] ) . using the three - term recurrence relation ( [ 3term_rec ] ) to rewrite the first term and eq .
( [ eq : deriv_zeta2 ] ) to rewrite the second one , we have @xmath124 we can also write @xmath125 by replacing the left hand side of eq .
( [ deriv_prod ] ) by eq .
( [ id_prod ] ) , dividing the resulting equation by @xmath82 and using eqs .
( [ r_en_h ] ) and ( [ s_en_h ] ) , we obtain a last recursion relation between @xmath87 , @xmath86 and @xmath88 @xmath126 combining eq .
( [ zeta_s ] ) and ( [ 2_recu ] ) , we have @xmath127 and finally , with the initial condition @xmath128 in eq .
( [ ci0 ] ) we obtain @xmath129 therefore substituting this relation in eq .
( [ 2_recu ] ) we finally obtain a closed system of two coupled recursion relations between @xmath86 and @xmath87 @xmath130 where the second equation was previously obtained in eq .
( [ 1_recu ] ) .
our goal now is to relate @xmath95-derivatives to @xmath33-derivatives , i.e. , find a relation between @xmath131 and @xmath132 . differentiating eq .
( [ def_hn ] ) w.r.t .
@xmath33 , using that @xmath133 as @xmath134 is a polynomial of degree @xmath116 ( [ eq : def_op ] ) , we have @xmath135 we now start from the expression of @xmath131 given in eq .
( [ snhn ] ) and use integration by parts to obtain @xmath136^{\lambda=\infty}_{\lambda = t}\\ & + & \frac{1}{\alpha}\int_t^\infty{\rm d}\lambda e^{-\alpha\lambda}\lambda^{a } \left((a+1 ) \pi_k^2(\lambda ) + 2 \lambda \pi_k(\lambda ) \partial_\lambda\pi_k(\lambda)\right),\nonumber \\ & = & \frac{1}{\alpha } ( -t\partial_t h_k+(a+1)h_k+2{\ensuremath{\langle \pi_k | \lambda\partial_\lambda\pi_{k}\rangle}\xspace } ) \;,\end{aligned}\ ] ] where we have used eq .
( [ dot_hn ] ) .
we can easily calculate the last scalar product @xmath137 to obtain the desired relation between @xmath131 and @xmath132 @xmath138 with this relation ( [ deriv_alpha_t_h ] ) , it is then straightforward to relate @xmath139 and @xmath140 to @xmath141 and @xmath142 , using ( [ r_en_h ] ) and ( [ s_en_h ] ) . note that from now on , we set @xmath98 @xmath143 we finally obtain , using eq .
( [ recu_derivalpha ] ) , the so - called schlesinger equations @xcite @xmath144 note that , for @xmath98 , the initial condition ( [ ci0 ] ) reads @xmath145 using the schlesinger equations ( [ recurrence_dots ] ) and this initial condition ( [ ci ] ) , we can compute step by step all the terms for arbitrary @xmath90 .
we end up this section by providing a useful relation between @xmath88 [ see eq .
( [ def_zeta ] ) ] and the cumulative distribution of the smallest eigenvalue @xmath146 .
one has @xmath147 where the last equality is obtained by using the classical tricks of replacing the vandermonde determinant by the determinant built from the ops @xmath101 s ( [ eq : def_op ] ) and then use the cauchy - binet formula @xcite . using eq .
( [ zeta_k_s ] ) with ( [ s_en_h_sans_alpha ] ) as well as ( [ deriv_alpha_t_h ] ) , we can write @xmath148 and therefore , from the expression of @xmath78 given in eq .
( [ f_k_fonction_de_hn ] ) , we have @xmath149 this expression thus provides a link between @xmath150 , which is the first non trivial coefficient of the op s [ see eq .
( [ def_zeta ] ) ] and the cumulative distribution of the smallest eigenvalue . in this section
, we derive another set of recursion relations between the coefficients @xmath87 s and @xmath86 s , the so - called laguerre - freud equations , following the procedure used by belmehdi and ronveaux in @xcite . to derive these equations , we start by searching two functions @xmath151 and @xmath152 , which are polynomials in @xmath80 , satisfying for any polynomial @xmath153 @xmath154 where the polynomials @xmath152 and @xmath151 may depend explicitly on the parameters @xmath33 and @xmath27 and @xmath155 . denoting by @xmath156 the weight associate to the scalar product in eq .
( [ eq : def_op ] ) and using an integration by parts , this relation ( [ def_phi_psi ] ) is satisfied provided that @xmath157 we find that the simplest non trivial solution to this equation is given by @xmath158 then , using this property ( [ def_phi_psi ] ) for @xmath159 and for @xmath160 , we can write after expansion , using the three - term recursion relation ( [ 3term_rec ] ) @xmath161 where we have introduced the four family of integrals @xmath162 these integrals are all calculated ( for @xmath163 and @xmath164 ) in [ appendix : turan ] .
using these expressions together with eq .
( [ r_en_h ] ) , we find the two relations @xmath165 we can rewrite these equations by subtracting rank @xmath90 to the rank @xmath103 and find the two laguerre - freud recurrence equations ( which are here of order 2 ) : @xmath166 as we show below these two sets of equations ( [ recurrence_dots ] ) and ( [ laguerre - freud ] ) allow us to ( i ) derive the connection to the painlev equation and ( ii ) perform the asymptotic analysis of these coefficients for large @xmath4 .
in this section , we proceed to the derivation of the painlev v equation , following the method of ref .
first , it is useful to introduce the quantities @xcite @xmath167 manipulating the laguerre - freud equations ( [ laguerre - freud ] ) we can prove ( see [ appendix : lf ] ) @xmath168 which is a simple algebraic relation ( and not a recursion relation ) between the different variables @xmath169 and @xmath88 . using the previous relations with the index @xmath170
, we find from eq .
( [ lien_zeta_f ] ) , @xmath171 summing up the first schlesinger equation ( [ recurrence_dots ] ) from @xmath172 to @xmath173 , we find @xmath174 which can be simplified using ( [ zeta_k_s ] ) , ( [ ci0 ] ) and ( [ def_omega_txt ] ) to yield @xmath175 on the other hand , using ( [ def : hn ] ) and ( [ zeta : omegan ] ) , we have @xmath176 taking a derivative of this equation w.r.t .
@xmath33 , we find @xmath177 from the second schlesinger equation ( [ recurrence_dots ] ) , we find @xmath178 where we have used ( [ def_omega_txt ] ) .
using the relation derived in [ appendix : lf ] in eq .
( [ theta_km ] ) , we can rewrite this recursion equation as a simple algebraic relation @xmath179 therefore , combining ( [ partialrn : h ] ) and ( [ partialr_k : omegarnthetan ] ) , we find @xmath180 on the other hand , from our eq .
( [ lf02_bonne_limite ] ) in which we inject the definition of @xmath181 ( [ def_omega_txt ] ) to eliminate @xmath150 @xmath182 summing and subtracting the last two equations ( [ h:omegarnthetan ] ) and ( [ lf_modif_kozeta ] ) , we obtain @xmath183 multiplying these two equations ( [ multi1 ] ) and ( [ multi2 ] ) together , we find @xmath184 finally , by using ( [ hn_to_rn ] ) , we eliminate @xmath181 and @xmath91 ( by expressing them in terms of @xmath185 and @xmath186 ) and find the equation @xmath187 which is a painlev v equation in the jimbo - miwa - okamoto @xmath188 form @xcite .
note that this equation coincides exactly with the equation first found by tracy and widom in @xcite .
in this section , we study the behavior of the quantities @xmath189 , @xmath91 , @xmath92 and @xmath150 in the hard edge limit ( [ def_hard_edge ] ) , which , in the language of ops , corresponds to a double scaling limit . of course
, as we are eventually interested in the study of the cumulative distribution of the smallest eigenvalue @xmath78 , we could perform this asymptotic analysis directly on the painlev v equation ( [ painleve5 ] ) , as done by tracy and widom in their original study of gue @xcite .
but it turns out to be much more convenient , especially to extract the @xmath7 corrections , to perform this asymptotic analysis on the schlesinger ( [ recurrence_dots ] ) and laguerre - freud ( [ laguerre - freud ] ) equations . to understand the structure of the coefficients @xmath189 , @xmath91 , @xmath92 and @xmath150 in this double scaling limit ( [ def_hard_edge ] ) , it is useful to study their behavior for small @xmath33 ( keeping @xmath4 fixed ) . for @xmath190 , the ops @xmath101 s in eq .
( [ eq : def_op ] ) can be expressed in terms of the generalized laguerre polynomials @xmath191 hence , thanks to standard properties of laguerre polynomials , we easily obtain these coefficients , for @xmath190 , using ( [ def_hn ] ) , ( [ r_en_h ] ) and ( [ s_en_h_sans_alpha ] ) as : @xmath192 we now take the index @xmath170 . when @xmath33 is closed to 0 , we can use eq .
( [ dot_hn ] ) and find @xmath193 ^ 2}{[\gamma(a+1)]^2 } + o(t^a ) \;.\ ] ] using ( [ polym0 ] ) , we can integrate and find the first correction @xmath194 ^ 2}{\gamma(a+2 ) \gamma(a+1)}+o(t^{a+1}).\end{aligned}\ ] ] we are interested in the scaling behavior when @xmath4 goes to infinity , @xmath33 goes to @xmath195 with @xmath196 finite ( [ def_hard_edge ] ) . in this double scaling regime
, the expansion above reads @xmath197.\end{aligned}\ ] ] this suggests the ansatz @xmath198,\end{aligned}\ ] ] where the function @xmath199 has thus the small @xmath200 expansion , read from eq .
( [ hn_small_x ] ) @xmath201 we can introduce the ansatz ( [ ansatz ] ) in the relations ( [ r_en_h ] ) and ( [ s_en_h_sans_alpha ] ) which give @xmath91 and @xmath92 in terms of the function @xmath199 .
we obtain @xmath202 the first non trivial terms in eq .
( [ ansatz_r_s ] ) @xmath203 for @xmath91 and @xmath204 for @xmath92 are necessary to compute the leading order of the cumulative distribution , @xmath205 . to compute the first @xmath7 correction to the limiting distribution
, we need to expand @xmath91 and @xmath92 in eq .
( [ ansatz_r_s ] ) to the next order in @xmath7 .
one actually expects the following expansion @xmath206 where the first term in this expansion are given in ( [ ansatz_r_s ] ) , i.e. , @xmath207 and @xmath208 . besides
, using the exact relation @xmath209 [ see eq .
( [ eq : dot_h ] ) ] , one can show that @xmath210 as well as @xmath211 when @xmath212 .
one can check , in principle , the validity of this expansion ( [ r_s_expansion ] ) order by order in powers of @xmath7 by injecting it in the schlesinger equations ( [ recurrence_dots ] ) .
proving it rigorously to arbitrary order @xmath213 is however a hard task .
however , here , we only need this asymptotic expansion up to order @xmath214 , which can be obtained explicitly as follows . to compute the second correction
, we truncate the expansion ( [ r_s_expansion ] ) up to the @xmath215 order and write @xmath216 in the hard edge limit @xmath52 , @xmath217 , keeping @xmath218 fixed , we can obtain the expansion of @xmath219 and @xmath220 at the same order by replacing @xmath4 by @xmath221 and using @xmath222 in ( [ r_s_expansion2 ] ) @xmath223 by injecting these two expansions into the schlesinger equation ( [ recurrence_dots ] ) , we find ( using @xmath224 ) at the first non trivial order ( @xmath225 for the first schlesinger equation and at @xmath226 for the second ) @xmath227 which can be solved as follows .
first , by subtracting the first equation of ( [ r_s_expansion2 ] ) to the second one , one obtains @xmath228 by injecting this relation ( [ rel_s2_r1 ] ) in the first equation of ( [ r_s_expansion2 ] ) , one finds that @xmath229 satisfies the following equation @xmath230 which can be solved , using that @xmath231 as @xmath212 , yielding @xmath232 consequently , from eq .
( [ rel_s2_r1 ] ) one has @xmath233 we finally find the two first terms of the expansion of @xmath91 and @xmath92 , for large @xmath4 , as @xmath234 from the expansion of @xmath92 in eq .
( [ expansion_second_order ] ) , we compute @xmath150 in the double scaling limit ( [ def_hard_edge ] ) , using eq .
( [ zeta_s ] ) , up to the second non - trivial order for large @xmath4 @xmath235 finally , from the expression of @xmath150 we obtain the large @xmath4 expansion of @xmath78 in the hard edge limit ( [ def_hard_edge ] ) , using eq . ( [ lien_zeta_f ] ) , which is given by @xmath236 using the initial condition @xmath237 and ( [ f_dvpt0 ] ) we can rewrite this equation as @xmath238 expanding eq .
( [ dvpt_f ] ) up to first order in @xmath7 , one obtains finally @xmath239 where @xmath67 is given by @xmath240 hence we easily obtain the functional form of the @xmath7 correction as given in eq .
( [ finfty2 ] ) , fully consistent with the conjecture in ( [ conjecture_finite_n ] ) made in @xcite .
however , at this stage , we still need to find the equation satisfied by the function @xmath199 , which enters in the definition of @xmath67 in eq .
( [ finfty(x ) ] ) . to obtain this equation , we expand @xmath241 and @xmath181 in eq .
( [ def_omega_txt ] ) , which is easily done from the expansions of @xmath91 , @xmath92 and @xmath150 in ( [ ansatz_r_s ] ) and ( [ ansatz_zeta ] ) to yield @xmath242 finally , by injecting these expansions ( [ ansatz_zeta ] ) and ( [ dvpt_omega_theta ] ) in ( [ lf02_bonne_limite ] ) with @xmath243 , we obtain , by canceling the first term , of order @xmath225 , in eq .
( [ lf02_bonne_limite ] ) that @xmath199 satisfies a painlev iii equation @xmath244 with the small argument behavior in eq .
( [ f_dvpt0 ] ) .
this result coincides with the one obtained , by a quite different method , by tracy and widom @xcite ( note the correspondence @xmath245 , where @xmath246 is the notation used in @xcite ) .
note that for integer values of @xmath27 , @xmath64 can be written explicitly in terms bessel functions @xcite [ see eq .
( [ fdetbessel ] ) ] .
these results in eqs .
( [ finfty2 ] ) , ( [ finfty(x ) ] ) and ( [ painleveiii ] ) yield the results announced in the introduction in eqs .
( [ eq : result_tw ] ) , ( [ piii ] ) and ( [ conjecture_finite_n ] ) . finally , in [ appendix : numeric ] we present a comparison between numerical simulations of wishart matrices of size @xmath247 and the asymptotic formula in eq .
( [ finfty2 ] ) describing the first @xmath7 correction .
we now turn to the analysis of the pdf of the smallest eigenvalue @xmath10 in the case where @xmath46 , and we set @xmath248 . in this case
, the density of eigenvalues has a single support on @xmath249 $ ] [ see eq . ( [ density_mp ] ) ] , with @xmath250 with a soft edge at both extremities ( see fig . [
fig : wishart_soft_vs_hard ] ) .
therefore , one expects that @xmath10 will be close to @xmath251 , while its fluctuations , of order @xmath252 , are governed by the tracy - widom distribution for @xmath16 . in the soft edge limit , the large @xmath4 analysis of @xmath78 in eq .
( [ def_fn ] ) is more conveniently done directly on the painlev v equation ( [ painleve5 ] ) @xcite . following this route
, one can indeed show @xcite that , for large @xmath4 @xmath253 where @xmath254 is given by @xcite @xmath255 and where @xmath256 is distributed according to the tracy - widom distribution @xmath53 , i.e. @xmath257 = { \cal f}_2(s)$ ] where @xmath258 is given by @xcite @xmath259 here @xmath260 is the so - called hatings - mcleod solution of the painlev ii equation @xmath261 where @xmath262 is the airy function .
the result in eq .
( [ baker_forrester ] ) can be equivalently written as @xmath263 which implies @xmath264 with @xmath265 and where @xmath266 , and where the tilde refers to the soft edge scaling limit . by injecting this form ( [ hn_0 ] ) into the painlev v equation satisfied by @xmath267 ( [ painleve5 ] )
one finds that @xmath268 satisfies the following equation @xmath269=0 \;.\end{aligned}\ ] ] using the painlev ii equation ( [ pii ] ) , one can indeed check that @xmath270 is solution of this equation ( [ pii_bis ] ) . note that to check this , it is useful to use the identity @xmath271 what is the first correction to the limiting form in eq .
( [ tilde_f0 ] ) ? unfortunately , in the soft edge limit , it turns out the schlesinger equations ( [ recurrence_dots ] ) do not allow to determine easily this first correction while they were very helpful in the hard edge scaling limit .
an alternative way to compute this first correction is to analyze directly the painlev v equation in ( [ painleve5 ] ) . by inspection of this equation ( [ painleve5 ] )
, we conjecture that the first correction to eq .
( [ tilde_f0 ] ) is of the form @xmath272 by inserting this expansion ( [ hn_kext ] ) in eq .
( [ painleve5 ] ) we obtain that @xmath273 satisfies the following linear differential equation @xmath274 where @xmath275 is given in eq .
( [ identity ] ) . of course , we know that @xmath276 when @xmath277 ( [ hn_kext ] ) .
in addition , from eq .
( [ eq_h1 ] ) one can show , using eq .
( [ identity ] ) and the large argument behavior in ( [ pii ] ) @xmath278 that @xmath279 .
but , of course , the equation ( [ eq_h1 ] ) being linear , the amplitude @xmath280 can not be determined from this analysis .
one way to determine it would be to analyze the op system ( [ eq : def_op ] ) in the limit when @xmath33 is far from the left edge , i.e. , for @xmath281 ( corresponding to the left large deviation tail of @xmath10 @xcite ) , and then match this result with the typical regime , for @xmath282 .
this program was carried out in detail in a similar albeit different context , involving discrete ops in ref . @xcite .
since we are interested in the first correction , this actually requires a very precise ( and tedious ) analysis of this regime @xmath281 which goes beyond the scope of the present paper . hence our result for the first correction @xmath283 in eq .
( [ eq_h1 ] ) does determine this function only up to a constant .
we have not found any simple solution to this equation ( [ eq_h1 ] ) , which could indicate that the corrections to scaling in this case are actually more complicated , as found recently in the case of real wishart matrices @xcite .
to conclude , we have provided a direct computation of the cumulative distribution @xmath78 of the smallest eigenvalue of complex wishart random matrices ( [ pjoint ] ) , for arbitrary parameter @xmath29 .
this was done by studying a set of semi - classical orthogonal polynomials as defined in eq .
( [ eq : def_op ] ) for which we derived ( i ) the schlesinger ( [ schlesinger_intro ] ) and ( ii ) the laguerre - freud ( [ lf_intro ] ) equations . by combining these equations
, we showed that @xmath78 can be expressed in terms of a solution of a painlev v equation ( [ painleve5 ] ) , thus recovering the result of tracy and widom @xcite using a quite different method . in the large @xmath4 limit , @xmath78 ,
properly shifted and scaled , converges to a limiting distribution @xmath67 which can be expressed in terms of a solution of a painlev iii equation ( [ painleveiii ] ) in the hard edge limit ( corresponding to @xmath284 ) and of a painlev ii equation ( [ pii_bis ] ) in the soft edge limit ( corresponding to @xmath285 ) .
furthermore , we have computed explicitly the first correction to the limiting distribution when @xmath52 . in the hard edge case ( [ conjecture_finite_n ] ) ,
we confirmed a conjecture by edelman , guionnet et pch in ref .
@xcite . in this case
, the first correction can be simply understood as a correction to the scale of the fluctuations of @xmath10 , see eq .
( [ eq : correction ] ) . on the other hand , in the soft edge limit
, we found that this correction is a solution of a second order linear differential equation with varying coefficients ( [ eq_h1 ] ) .
solving this equation remains a challenging open problem , which could suggest that the first correction does not correspond to a simple shift or rescaling of the scaling variable , as found for the hard edge ( [ eq : correction ] ) .
this could be reminiscent of the result found for real wishart matrices @xcite and certainly deserves further investigations .
we would like to thank y. chen , p. j. forrester and n. s. witte for useful correspondence and discussions .
to compute the terms which enter the laguerre - freud equations , we need to compute several integrals containing the ops @xmath286 ( [ eq : def_op ] ) . to perform these computations
, we will follow the method developed by belmehdi and ronveaux @xcite , using turn determinants .
we introduce four types of integrals @xmath287 for @xmath288 and @xmath164 , while @xmath90 is an arbitrary integer . using the recurrence relation ( [ 3term_rec ] ) , we can calculate the integrals @xmath289 s and the @xmath290 s : @xmath291 the terms @xmath292 , @xmath293 and @xmath294 can easily be calculated using that @xmath101 s are monic ops : @xmath295 the three last quantities , @xmath296 and @xmath297 , require more work . for this purpose , we introduce a new object : the turn determinant @xmath298 defined as : @xmath299 using the three term recurrence ( [ 3term_rec ] ) , we find a recurrence for the turn determinant @xmath300 where we have introduced the auxiliary quantity @xmath301 .
finally , we can write , using @xmath302 and the initial condition @xmath303 @xmath304 this equation allows us to rewrite @xmath305 where the prime denotes a derivative w.r.t . @xmath200 . on the other hand , by using the definition of the turn determinant we find @xmath306 we obtain finally @xmath307 we now have all the ingredients needed to calculate the last three integrals . using the three term recurrence and the orthogonality condition
, we can express @xmath308 as @xmath309 from eq .
( [ scal_pipi2 ] ) for @xmath310 , we can express the last scalar product in terms of integrals that we have computed before .
using eq .
( [ j0j1l0 ] ) , we find @xmath311 using the three term recurrence and the orthogonality condition , we can express @xmath312 as @xmath313 finally , we have from eqs .
( [ r_en_h ] ) , ( [ j0j1l0 ] ) and ( [ l1 ] ) , @xmath314 similarly , using again the three term recurrence and the orthogonality of the family , we can express @xmath297 as @xmath315 using eq .
( [ scal_pipi2 ] ) for @xmath316 , we can express the last scalar product in terms of integrals that we have computed before .
@xmath317.\end{aligned}\ ] ] we can simplify the telescoping sums and find @xmath318 equipped with the computations of these integrals , we can now derive the laguerre - freud equations .
the derivation presented here follows the one of ref .
@xcite where slightly different ops were considered . to derive the laguerre - freud equations , we start from equation ( [ lf01 ] ) : @xmath319 it is useful to introduce the following variables @xmath320 in terms of these variables , we can rewrite the first equation of ( [ lf01_a2 ] ) as @xmath321 and the second equation of ( [ lf01_a2 ] ) as @xmath322 similarly , eq .
( [ lf_omega2 ] ) can be written as @xmath323 we also rewrite eq .
( [ lf_omega1 ] ) as @xmath324 finally , by multiplying together eqs .
( [ lfo1_omega ] ) and ( [ lfo2_omega ] ) , we have , with the notation @xmath325 , @xmath326 for @xmath327 , we find @xmath328 using the initial condition @xmath329 , we can solve this equation and find @xmath330 for @xmath331 , we find @xmath332 again , using the initial condition @xmath329 , we can solve this equation and find @xmath333 finally , using the explicit expression of @xmath334 , this equation can also be written as @xmath335 thanks to eq .
( [ theta_km ] ) , we can substitute @xmath336 in ( [ theta_km2 ] ) . using the definition of @xmath337 in ( [ def_omega ] ) , eq .
( [ theta_km2 ] ) yields a quadratic equation for @xmath337 in terms of @xmath338 and @xmath88 @xmath339 which yields the equation given in the text in eq .
( [ lf02_bonne_limite ] ) .
in this section , we present a numerical check of our formula for the first finite @xmath4 corrections to the limiting density of @xmath10 in the hard edge case , i.e. , keeping @xmath27 finite while @xmath52 .
our simulations have been carried out for integer values of the parameter @xmath27 for which the function @xmath64 in eq .
( [ finfty(x ) ] ) can be computed explicitly in terms of bessel @xmath340 functions @xcite : @xmath341_{1\le j , k \le a}}}{\det{[i_{j - k}(2\sqrt{x})]_{1\le j , k \le a}}}\ , , \ ; a \ ; { \in } \ ; { \mathbb n}^ * \;,\end{aligned}\ ] ] and @xmath342 for @xmath69 . in fig .
( [ fig : fn_ordre_2 ] ) we show a plot of @xmath343 obtained by sampling @xmath344 independent wishart matrices of size @xmath247 ( dots ) for different values of @xmath345 and @xmath346 . the comparison with @xmath347 , evaluated exactly from eqs .
( [ finfty(x ) ] ) and ( [ fdetbessel ] ) shows a good agreement between theory and numerical simulations .
the main objective of these simulations is to test the formula for the first order correction , given in eq .
( [ finfty2 ] ) which can also be written as @xmath348 in the left panel of fig .
[ fig : fn_ordre_2 ] , we show a plot of @xmath349}$ ] as a function of @xmath200 and compare it to @xmath350 , where @xmath64 is given in eq .
( [ fdetbessel ] ) .
this comparison illustrates the accuracy of the asymptotic formula characterizing the finite @xmath4 correction to @xmath67 . for a review see s. n. majumdar , les houches lecture notes on complex systems ( 2006 ) , ed . by j .-
bouchaud , m. mzard and j. dalibard [ arxiv : cond - mat/0701193 ] .
_ random matrices , the ulam problem , directed polymers & growth models , and sequence matching _ | we study the probability distribution function ( pdf ) of the smallest eigenvalue of laguerre - wishart matrices @xmath0 where @xmath1 is a random @xmath2 ( @xmath3 ) matrix , with complex gaussian independent entries .
we compute this pdf in terms of semi - classical orthogonal polynomials , which are deformations of laguerre polynomials . by analyzing these polynomials , and their associated recurrence relations , in the limit of large @xmath4 , large @xmath5 with @xmath6
i.e. for quasi - square large matrices @xmath1 we show that this pdf , in the hard edge limit , can be expressed in terms of the solution of a painlev iii equation , as found by tracy and widom , using fredholm operators techniques .
furthermore , our method allows us to compute explicitly the first @xmath7 corrections to this limiting distribution at the hard edge .
our computations confirm a recent conjecture by edelman , guionnet and pch .
we also study the soft edge limit , when @xmath8 , for which we conjecture the form of the first correction to the limiting distribution of the smallest eigenvalue . |
this work is supported by the polish national science center grant harmonia - dec-2012/04/m / st7/00799 .
s.v.d . acknowledges financial support from the european research council ( erc-2012-stg 307502-e - control ) .
s.z acknowledges the dean s grant 15.11.230.198 .
t. nan , z. zhou , m. liu , x. yang , y. gao , b. a. assaf , h. lin , s. velu , x. wang , h. luo , j. chen , s. akhtar , e. hu , r. rajiv , k. krishnan , s. sreedhar , d. heiman , b. m. howe , g. j. brown , and n. x. sun , sci .
rep . * 4 * , 3688 ( 2014 ) . | dynamic properties of nife thin films on pmn - pt piezoelectric substrate are investigated using the spin - diode method .
ferromagnetic resonance ( fmr ) spectra of microstrips with varying width are measured as a function of magnetic field and frequency .
the fmr frequency is shown to depend on the electric field applied across the substrate , which induces strain in the nife layer .
electric field tunability of up to 100 mhz per 1 kv / cm is achieved . an analytical model based on total energy minimization and the llg equation , with magnetostriction effect taken into account , is developed to explain the measured dynamics .
based on this model , conditions for strong electric - field tunable spin diode fmr in patterned nife / pmn - pt structures are derived .
electric field control of magnetism at room temperature can lead to the development of efficient and low - power memories,@xcite magnetic field sensors@xcite , voltage - tunable microwave filters,@xcite and oscillators.@xcite application of multiferroic materials can additionally lead to the design of new electronic devices in which both the electron spin and charge are affected by an external electric field .
voltage control of magnetic anisotropy ( vcma ) in multiferroics can be realized by strain transfer from a ferroelectric or piezoelectric layer to a ferromagnetic film , as the deformation of the ferromagnet changes the magnetoelastic anisotropy _ via _ inverse magnetostriction.@xcite recently , the influence of electric - field induced strain on magnetic anisotropy has been demonstrated in a variety of unpatterned and patterned multiferroic heterostructures including batio@xmath0/fm @xcite , pmn - pt / fm @xcite and pzn - pt / fm @xcite with fm = ni @xcite , nife @xcite , nico @xcite , co @xcite , cofe @xcite , cofeb @xcite , fe @xcite , fegab @xcite , and fe@xmath0o@xmath1 @xcite .
electric - field tuning of ferromagnetic resonance ( fmr ) has also been studied . in most reports , strong tuning of fmr in continuous ferromagnetic films on piezoelectric substrates
is inferred from microwave cavity or vector network analyzer fmr measurements @xcite . in this letter , we experimentally study electric - field tuning of fmr in patterned ni@xmath2fe@xmath3 microstrips on pmn - pt substrates using a spin diode ( sd ) measurement technique .
patterning of the ferromagnetic film is anticipated to introduce a magnetostatic shape anisotropy , which competes with the magnetoelastic anisotropy that is induced via transfer of piezoelectric strain . to systematically study this effect
, we consider nife microstrips of different width . an analytical model for electric - field tunable microwave signals in confined ferromagnetic geometries
is also presented . using this model
, we derive a phase diagram of the fmr frequency shift as a function of microstrip width and magnetic field strength . on a polished pmn - pt ( pb(mg@xmath4nb@xmath5)o@xmath0- pbtio@xmath0 ) ( 011)-oriented piezoelectric substrate , a 20 nm thick layer of ni@xmath2fe@xmath3
was deposited using magnetron sputtering .
the bottom side of the crystal was covered by a 5 nm ti/50 nm au layer , in order to apply high voltage perpendicular to the piezoelectric substrate .
afterwards , nife microstrips of 1.5 , 2.6 , 6.7 m width and 90 m length along the [ 01 - 1 ] direction of the pmn - pt crystal were fabricated using electron beam lithography and ion - beam etching .
the sample was vacuum annealed at 330@xmath6c in an in - plane magnetic field to increase the anisotropic magnetoresistance ( amr ) ratio .
the dc resistance of the respective strips was : 842 @xmath7 , 576 @xmath7 , 201.5 @xmath7 .
a radio frequency ( rf ) current of amplitude @xmath8 , flowing through the nife strips deposited on the pmn - pt substrate , generates a time - dependent spin transfer torque ( stt ) and oersted field .
these effects lead to magnetization dynamics and , because of amr , resistance oscillations .
mixing of the oscillating current and resistance generates a dc sd voltage @xmath9 .
the amplitude of this signal is proportional to the real part @xmath10 of the complex amplitude of the resistance change,@xcite @xmath11 .
the amr effect is given by : @xmath12 , where @xmath13 and @xmath14 ( @xmath15 ) denote the resistance of the strip when the magnetization is perpendicular ( parallel ) to the current .
the change in resistance due to a small change in the angle @xmath16 between magnetic moment and current is thus given by @xmath17 . from this , one finds @xmath18 .
we note that @xmath19 represents the real part of the complex angular changes .
as mentioned above , the magnetization dynamics responsible for the @xmath9 signal is driven by the uncompensated oersted field and by the stt effect .
even for a single - layer ferromagnetic strip , an uncompensated oersted field can be induced , since electron scattering processes at both interfaces are generally different .
the stt , on the other hand , may appear due to some inhomogeneities of the magnetization distribution within the permalloy strip.@xcite small changes of the angle @xmath16 can be derived from the landau - lifshitz - gilbert ( llg ) equation for a unit vector along the magnetization @xmath20 : @xmath21 where @xmath22 is a damping constant ( of the order of @xmath23 ) , and @xmath24 is a torque acting on the magnetic moment @xmath25 , with @xmath26 being the gyromagnetic ratio . the second term in eq .
[ eq : gama ] corresponds to the stt effect .
its amplitude , @xmath27 , is proportional to the magnitude of the rf current density @xmath28 and its spin polarization p , i.e. @xmath27 @xmath29.@xcite we have omitted here the non - adiabatic term in stt , as its amplitude is usually much smaller than that of the adiabatic term@xcite .
the averaged magnetic energy @xmath30 includes the shape anisotropy , zeeman - like terms due to static and dynamic ( oersted ) magnetic fields , as well as terms related to the stress due to the deformation of the pmn - pt substrate under external electric field .
the energy @xmath31 can be written as : @xmath32}(e ) \cos^2 \theta \nonumber \\
-[k_{[100]}- \frac{3}{2 } \lambda \sigma_{[100]}(e)]\sin^2 \theta\sin^2\phi -\vec{m}\cdot\vec{h}_{oe[100,011]},\end{aligned}\ ] ] where @xmath33 is the demagnetizing field , @xmath34 is the external magnetic field , and @xmath35}$ ] is the uniaxial magnetocrystalline anisotropy that is induced during magnetic - field annealing .
the coordinate system with defined angles describing orientation of the magnetization ( @xmath16 ) and magnetic field ( @xmath36 ) with respect to the crystallographic axes of the pmn - pt crystal is shown in fig .
[ fig : schemat2](a ) .
the easy axis in our sample is transverse to the strip long axis , i.e. along the [ 100 ] direction .
in turn , @xmath37}$ ] in eq . 3 denotes the time - dependent oersted field components in the [ 100 ] or [ 011 ] directions .
because its amplitudes are constant within the sample , the oersted field may be written as a gradient with respect to @xmath38 of the energy term @xmath39 } $ ] .
finally , @xmath40 in eq .
3 . is the magnetostriction constant .
the electric field - dependent stress @xmath41 acting on permalloy in the [ 100 ] and [ 01 - 1 ] directions can be calculated from the relation between the strain of the pmn - pt substrate and the applied electric field , @xmath42 , where @xmath43 denotes the strain of the pmn - pt along the @xmath44 direction ( @xmath45,[01 - 1]$ ] ) , @xmath46 is a matrix of piezoelectric constants , and @xmath47 is the electric field applied in the @xmath48 direction ( @xmath49 $ ] ) . since the electric field is applied perpendicularly to the substrate ( in [ 011 ] direction ) ,
the only strains that influence the permalloy microstrip are those in [ 100 ] and [ 01 - 1 ] directions .
in particular , the relation between the strain within pmn - pt and stress transmitted to permalloy can be written as:@xcite @xmath50}(e ) = \frac{y(d_{31}+\nu d_{32 } ) } { ( 1-\nu^2 ) } e_{[011 ] } , \\
\sigma_{[100]}(e ) = \frac{y(d_{32}+\nu d_{31 } ) } { ( 1-\nu^2 ) } e_{[011]},\end{aligned}\ ] ] where y=200 gpa is the young s modulus , and @xmath51 is poisson s ratio of nife.@xcite from the above equations one can see that stresses in both in - plane directions ( [ 100 ] and [ 01 - 1 ] ) of permalloy may be different .
thus , depending on their signs , they can induce either easy or hard magnetization axes along the [ 100 ] or [ 01 - 1 ] directions of the nife microstrips . in our case
, @xmath52 equals @xmath53 ( discussed in experimental part ) , which agrees well with data reported by m. shanthi et al . in ref . for a pmn - pt crystal .
based on this , we assume the remaining piezoelectric coefficients : @xmath54[01 - 1]}=723\pm 20 \frac{pc}{n}$ ] and @xmath55[100]}=-1761\pm 13 \frac{pc}{n}$ ] . as a consequence , the strain - induced easy magnetization axis is aligned along the [ 01 - 1 ] direction , which is parallel to the magnetostatic shape anisotropy of the microstrip . from the llg equation one
finds @xmath19 as a function of the driving frequency @xmath56 .
the solution has a general resonance - curve form:@xcite @xmath57 + \sin\psi [ af ( f^2 - f_0 ^ 2 ) - b\sigma f ] } { ( f^2 - f_0 ^ 2)^2 + \sigma^2 f^2}\ ] ] where a and b describe the amplitudes of the symmetric and antisymmetric contributions to the @xmath9 signal , while @xmath58 denotes the resonance frequency given by : @xmath59 ^ 2 \right)(1+\alpha^2)\csc^2\theta + \cot\theta \left ( \frac{\partial u}{\partial \theta } \frac{\partial^2 u}{\partial \phi \theta}\alpha \csc\theta \right . \\ \left .
+ \left ( 2 \frac{\partial u}{\partial \phi } \frac{\partial^2 u}{\partial \theta \partial \phi}\alpha^2 + \frac{\partial u}{\partial \phi}\frac{\partial^2 u}{\partial \phi \partial \theta } - \frac{\partial u}{\partial \theta } \frac{\partial ^2 u}{\partial \phi^2 } \right)\csc^2\theta - \frac{\partial u}{\partial \theta}\frac{\partial^2u}{\partial \phi^2 } \alpha \csc^3 \theta \right ) % \left [ \frac{\partial^2 u}{\partial\phi\partial\theta } \right]^2 \right ) ( 1+\alpha^2)\csc\theta % \frac{\partial^2 u}{\partial \theta^2 } \end{array } } \label{eq : disp_gen}\ ] ] with the partial derivatives calculated at the stationary angles ( @xmath60 ) determined by minimum magnetic energy ( eq .
( [ eq : energy ] ) ) in the absence of rf current . because of lack of dc current , the resonance frequency does not depend on the stt- and oersted - field - related terms . on the other hand ,
the resonance frequency implicitly depends on applied electric field due to the presence of static electric - field related energy terms .
is the angle of the external magnetic field and @xmath16 and @xmath61 denote the polar and azimuthal angles of the magnetization direction .
( b ) experimental setup for sd effect measurements with external voltage applied to the pmn - pt piezoelectric crystal to electrically strain the nife microstrips.,width=309 ] a schematic of the experimental setup for sd measurements of the fmr effect is presented in fig .
[ fig : schemat2](b ) .
a microwave signal of 10 dbm was applied to the nife microstrip using a rf probe and generator , and the dc voltage produced by mixing of the rf current with resistance oscillations was detected by a dc voltmeter .
experiments with unpatterned pmn - pt / nife sample were also conducted as reference .
x - ray diffraction ( xrd ) scans of the ( 022 ) pmn - pt reflection were measured in an electric field ranging form 0 to 12 kv / cm ( fig .
[ fig : xrd](a ) ) .
the electric - field induced piezoelectric strain @xmath62 that is derived from these measurements is summarized in fig .
[ fig : xrd](b ) .
the slope of the curve , which corresponds to the @xmath52 piezoelectric constant , equals @xmath63 cm / kv .
-2@xmath16 scans of the ( 022 ) reflection of the pmn - pt substrate for electric fields ranging from 0 to 12 kv / cm .
( b ) electric field induced relative change of the out - of - plane ( 011 ) lattice parameters of the pmn - pt substrate .
the red line represents a linear fit .
, width=294 ] with @xmath35}=785\frac{j}{m^3}$ ] , @xmath64 , @xmath65 ( determined form vsm measurements ) , and @xmath66 .
( c ) fmr frequency as a function of applied electric and magnetic field for nife microstrips with a width of 6.7 m.,width=317 ] we used the sd effect to characterize voltage - tunable fmr in our patterned pmn - pt / nife structures for dc voltages in the range of 0100 v ( 0 - 2 kv / cm ) .
examples of electric - field tunable fmr spectra are shown in fig .
[ fig : dysp](a ) .
the spectra were measured in a constant magnetic field ( 20 - 500 oe ) applied at @xmath67 , which corresponds to the angle for which maximum sd voltages have been observed@xcite . the evolution of fmr with applied magnetic field strength for three different nife strips in an electric field up to 2 kv / cm are shown in fig .
[ fig : dysp](b ) . the largest voltage - induced shift in fmr frequency ( 202 mhz )
is observed for the widest strip at low magnetic field ( 21.5 oe ) . for more narrow strips and larger magnetic fields , the voltage - induced frequency shifts are smaller .
fig .
[ fig : dysp](c ) shows the fmr frequency as a function of electric and magnetic fields obtained for the widest strip , compared with theoretical model calculations ( solid lines ) . to additionally proof the consistency of the derived model , amr loops on 6.7 m wide strips
were measured ( fig .
[ fig : rh](a ) ) .
the theoretical curves , indicated by solid lines , were calculated for 0 kv / cm and 4 kv / cm using the same parameters as for the calculated fmr shifts .
loop measured at @xmath68 .
the solid lines represent the macrospin calculations .
( b ) calculated shift of the resonance frequency , @xmath69 , as a function of the external magnetic field and strip s width.,width=340 ] the small discrepancy in switching fields between the theoretical and experimental results may be caused by magnetic domain formation or thermally activated switching , which are not taken into account in the macrospin model . apart from the hysteretic region , however , the macrospin model predictions fit the experimental data well .
this allows us to model the frequency shift in an electric field of 2 kv / cm as a function of nife strip width and external magnetic field using system parameters that are derived from dynamic and static measurements . the result is shown in fig .
[ fig : rh](b ) . the calculated phase diagram can be used to identify the parameter space for which strong electric - field tuning of fmr can be attained in nife microstrips on piezoelectric pmn - pt .
the results indicate that the largest effects are obtained in wide microstrips at modest external magnetic field .
the reduction of strain sensitivity in more narrow strips can be attributed to stronger magnetostatic shape anisotropy , which reduces the effect of the piezostrain induced magnetoelestic anisotropy . to summarize ,
we have explored electric - field tuning of fmr in nife microstrips on the piezoelectric pmn - pt substrate .
our results indicate that electrical control of fmr spectra depends sensitively on the shape of the nife microstructures and applied magnetic field .
a newly developed analytical model accounts for these effects and can be used to calculate fmr phase diagrams based on well - known material parameters . |
the promise of a qualitative advantage of quantum computers over classical ones in solving certain classes of problems has led to a massive effort in theoretical and experimental investigation of controlled , quantum - coherent systems .
the standard circuit model ( cm ) of quantum computing is analogous to classical computing in the sense of requiring a sequence of logic gate operations .
however , the requirement of precise time - dependent control of individual qubits in the quantum case is hard to achieve experimentally while still maintaining the quantum coherence of the system .
a number of alternative approaches have been proposed , of which _ adiabatic quantum computing _ ( aqc ) is a promising example .
this involves the evolution of a quantum system from a simple hamiltonian with an easily - prepared ground state to a hamiltonian that encodes the problem to be solved , and whose ground state encodes the solution .
if the system is prepared in the initial ground state and the time evolution occurs slowly enough to satisfy the adiabatic theorem , the final state will have a large overlap with the ground state .
measurement in the computational basis will then yield the desired solution with high probability @xcite .
several authors have demonstrated polynomial equivalence between aqc and the cm , mapping the latter onto an aqc with @xmath3-local interactions between qubits or @xmath4-local interactions between @xmath5-state qudits in two dimensions @xcite , or with @xmath4-local interactions between qubits on a two - dimensional lattice ( but requiring two or more control hamiltonians ) @xcite . despite these proofs of equivalence between aqc and cm ,
it is clear that there are classes of problems more suited to one or the other ; in addition , aqc is believed to be more robust against decoherence @xcite , although the effects of decoherence @xcite and noise @xcite imply an optimal computation time , beyond which errors increase .
the type of problems most suited to aqc include optimization problems , where the requirement is to find the global minimum of a cost function @xmath6 , and the related decision problems , where the requirement is to demonstrate the existence of a good solution @xmath7 obeying @xmath8 for some specified @xmath9 .
thus the existence proof of a polynomial - time aqc implementation of shor s prime factorization algorithm does not help practical implementation : one rather starts afresh and maps factorization onto an optimization problem , as in the recent nmr factorization of @xmath10 @xcite .
simulations of the travelling salesman problem show faster decay of residual energy ( i. e. , tour length ) through aqc than through classical simulated annealing , although other classical algorithms are faster @xcite .
applications have also been found in graph theory , most recently in the evaluation of ramsey numbers @xcite .
the task of aqc is to find the ground state of a hamiltonian @xmath11 ; this hamiltonian encodes the problem under consideration and its ( unknown ) eigenvalues determine the cost function . a hamiltonian @xmath12 interpolates between a simple initial hamiltonian , @xmath13 , at time @xmath14 and the desired final hamiltonian , @xmath11 , at the end of the computation @xmath15 .
many interpolation schemes have been considered , which may optimize final - state fidelity but require some knowledge of the energy - level structure @xcite or phase cancellation @xcite .
we therefore restrict consideration to the simple linear interpolation @xmath16 where @xmath17 $ ] is the reduced time @xmath18 .
the eigenvalues and eigenstates of the hamiltonian @xmath12 of an @xmath19-qubit system are given by @xmath20 the instantaneous state of the system is given by @xmath21 , the solution of schrdinger s equation , which in reduced time ( and @xmath22 ) reads @xmath23 the system is prepared in the ( non - degenerate ) ground state of @xmath13 : @xmath24 .
at the end of the evolution a suitable figure of merit is the closeness of the state vector , @xmath25 , to the desired result , @xmath26 .
this is provided by the _ success probability _ @xmath27 the subscript @xmath19 , denoting the number of qubits , will be omitted except where a distinction needs to be made . in practical optimization problems , a low - cost solution that is
not necessarily the global optimum often suffices . here
the _ energy error _
@xmath28 is a suitable figure of merit .
approximate adiabatic quantum computing ( aaqc ) aims to reduce this error @xcite . for
some purposes other characterizations of the final - state distribution @xmath29 may be more appropriate figures of merit . in the present work we shall concentrate on the success probability .
we require @xmath30 parameters to specify the hamiltonian @xmath11 .
one of the aims of this work is to investigate to what extent the success probability ( [ eq : p ] ) can be approximated as a function @xmath31 , where @xmath32 is a small ( @xmath19-independent ) number of parameters characterizing the initial and final hamiltonians .
the most important dependence is expected to be on the _ minimum gap _ between ground state and first excited state @xmath33 which occurs at the reduced time(s ) @xmath34 : @xmath35 while it has long been known @xcite that this probability tends to unity for slow evolution : @xmath36 the precise statement of this _ adiabatic theorem _ has been the subject of much debate in recent years @xcite .
the original statement in the context of aqc @xcite was that the adiabatic condition @xmath37 where @xmath38 guarantees @xmath0 to be very close to @xmath39 .
while this only considers transitions into the first excited state , such transitions will dominate in most situations .
et al _ @xcite derived such a result , with @xmath40 taken as the maximum over all matrix elements to excited states .
if @xmath40 is considered constant ( of the order of a typical eigenvalue of @xmath13 ) , @xmath1 determines the required @xmath2 .
it is however sufficient ( see , for example , ref .
@xcite ) to require an evolution time @xmath41 for all excited states @xmath42 .
( in the present context we are restricting consideration to evolution of the ground state . )
some authors @xcite have claimed counterexamples to the above criterion .
however , these counterexamples include a resonant term , which is absent from our interpolating hamiltonian ( [ eq : aqch ] ) . for practical purposes
the knowledge that the success probability tends to unity in the infinite - time limit is of less interest than knowledge of parameters governing success for finite evolution times ; it is this question that motivates our study . the minimum gap @xmath1 is usually considered to be the dominant parameter determining the success probability for a given evolution time .
these two variables , @xmath0 and @xmath1 , are both used in the literature to quantify the performance of a given computation , and are assumed to increase monotonically with each other .
the question of how either of these variables varies with system size is an important one that is often addressed .
however , the exact nature of the correlation between these two important figures of merit has not been fully explored .
we explore the relationship between @xmath0 and @xmath1 by looking at the statistical distributions of these two variables over an ensemble of problem hamiltonians ( @xmath11 ) for fixed computation times @xmath2 .
we start by considering a simple two - qubit system and show that a rich structure arises in the scatter plots of success probability against @xmath1 .
we then go on to look at the scatter plots in three- , four- and five - qubit systems and find that , although some of the finer details of the structure are washed out , some remain .
we wish to look at the distribution of the success probability and @xmath1 over a large set of problem hamiltonians .
we use a simple , yet generic , model that is scalable and can be readily solved numerically .
for @xmath13 we use a transverse field of unit magnitude acting on all the qubits : @xmath43 where @xmath44 denote the usual pauli matrices , @xmath19 is the number of qubits in the system ; the @xmath45 matrix @xmath46 acts on the @xmath47th qubit .
the ( non - degenerate ) ground state of @xmath13 is an equal superposition @xmath48 of all @xmath49 computational basis states . for @xmath11
, we use a random - energy hamiltonian , diagonal in the computational basis , where all @xmath50-axis couplings between the @xmath19 qubits are realized : @xmath51 here @xmath52 is the @xmath47th digit in the binary representation of @xmath53 . where there are @xmath54 non - zero bits in the binary representation of @xmath53 , the coupling constant @xmath55 represents a @xmath54-local interaction ( a non - trivial interaction between @xmath54 qubits ) .
the @xmath56 will be selected from a suitable random distribution ; we fix the trivial energy shift @xmath57 .
@xmath11 is diagonal in the computational basis so that the binary - ordered set of states @xmath58 is a permutation of the energy - ordered set of states @xmath59 defined in eq .
[ eq : eigen ] ( in the generic case where the latter are non - degenerate ) .
a hamiltonian of this type can be used to encode any finite computational optimization problem ( minimization of a function @xmath6 ) by choice of the @xmath60 .
it is important to note that only @xmath39- and @xmath4-local interactions are experimentally feasible ; however , higher - order interactions may be reducible to such terms at the cost of auxiliary qubits @xcite . for each sample in the scatter plots , we solve the schrdinger equation numerically over the reduced time range @xmath61 for a given computation time , @xmath2 , using the dormand - prince method @xcite .
this is an adaptive step - size algorithm ; solutions accurate to fourth- and fifth - order in the step size @xmath62 are used to estimate the local error in the former .
if it is less than the desired tolerance , then the fifth - order solution is used for the integration .
otherwise @xmath62 is decreased .
( abscissa ) at ( 6,-0.5 ) minimum gap ( @xmath1 ) ; ( coodinate ) [ rotate=90 ] at ( -0.2,3.5 ) probability ( @xmath0 ) ; for comparison with later scatter plots , figure [ fig : p1 ] plots the probability @xmath63 against minimum gap @xmath64 for a single qubit . since the final hamiltonian is specified by a single parameter
, @xmath0 is a ( not quite monotonic ) function of @xmath1 and @xmath2 .
a test of accuracy of the simulation is that the small component of the final state should be real : @xmath65 ( which can be verified analytically ) .
our numerical calculations reproduce this to high precision .
for two or more qubits , the success probability will no longer collapse to a function @xmath66 of the minimum gap and computation time .
figure [ fig:2qb_uniform ] is a scatter plot of success probability @xmath67 against minimum gap @xmath68 for a large set of two - qubit problem instances , with the coupling coefficients @xmath69 drawn from the uniform distribution @xmath70 , and a short computation time @xmath71 .
observe the sharp upper and lower edges .
the lower bound of the success probability is always @xmath72 for infinitesimally small @xmath1 .
this arises when @xmath73 : with a four - fold degeneracy at @xmath74 the system remains in its original ground state ( [ eq : gs ] ) .
( abscissa ) at ( 4.6,-0.7 ) minimum gap ( @xmath1 ) ; ( coodinate ) [ rotate=90 ] at ( -1,3.075 ) probability ( @xmath0 ) ; ( 0a ) at ( 0.0,-0.15 ) @xmath75 ; ( 1a ) at ( 2.525,-0.15 ) @xmath76 ; ( 2a ) at ( 5.05,-0.15 ) @xmath39 ; ( 3a ) at ( 7.575,-0.15 ) @xmath77 ; ( 4a ) at ( 10.1,-0.15 ) @xmath4 ; ( 0c ) [ left ] at ( 0,0.1 ) @xmath75 ; ( 0c ) [ left ] at ( 0,1.31 ) @xmath78 ; ( 0c ) [ left ] at ( 0,2.52 ) @xmath79 ; ( 0c ) [ left ] at ( 0,3.73 ) @xmath80 ; ( 0c ) [ left ] at ( 0,4.94 ) @xmath81 ; ( 0c ) [ left ] at ( 0,6.15 ) @xmath39 ; ( 0b ) [ right ] at ( 10.8,0.1 ) @xmath75 ; ( 0b ) [ right ] at ( 10.8,3.45 ) @xmath77 ; ( 0b ) [ right ] at ( 10.8,6.75 ) @xmath3 ; it is important to verify that this structure is independent of our choice of random distribution of coupling constants and that it is also not an artefact of the pseudo - random number generators used .
figure [ fig:2qb_gaussian ] also shows scatter plots of success probability and minimum gap , but in this case the coupling coefficients are drawn from a gaussian distribution , @xmath82 ( mean @xmath75 , standard deviation @xmath39 ) .
the trends and structure in the distributions are similar to those shown in figure [ fig:2qb_uniform ] .
however , there are some subtle differences in sharpness between the gaussian and uniform cases . for a large minimum gap
, the lowest probability occurs for large @xmath83 , so we see a sharp cutoff in the uniform case but not in the gaussian case . in general
though , this shows that the results are independent of our choice of coupling constants and , as a different pseudo - random number generator routine was used , we can say that the results are not a numerical artefact .
four computation times are shown : @xmath71 , @xmath84 , @xmath85 and @xmath86 .
as @xmath2 increases , the distribution shifts and tends towards a success probability of @xmath87 for any @xmath88 , in agreement with the adiabatic theorem .
+ + + the two interesting features of these scatter plots are the well - defined sharp edges and the densely - populated bands .
we colour the data points according to the strength of the two - qubit interactions , as this is a special direction in the two - qubit parameter space , which will determine the amount of entanglement during the evolution .
it is clear that the bands correspond to groups of hamiltonians with similar @xmath83 .
the bands where @xmath89 can be seen as two separable one - qubit evolutions for @xmath90 and @xmath91 , so the total success probability @xmath92 is simply the product @xmath93 of the one - qubit success probabilities shown in figure [ fig : p1 ] : @xmath94 where @xmath95 another interesting point to note is that the bands of similar @xmath11 gradually reverse in order in the distribution as the computation time @xmath2 is changed . we have supplemented the uniform random data with sets of @xmath96 chosen on a rectangular grid with the same cut - offs .
these have the advantage that all problem hamiltonians with a given value of @xmath97 can be plotted in the @xmath98-plane and coloured by their minimum gap or success probability ; see figure [ fig : stability ] . the energy structure in the case of two qubits can be simply characterized .
the final - state energies are given by [ levels ] @xmath99 the ground - state phase diagram has tetrahedral symmetry @xmath100 , with the regions of parameter space with ground states @xmath101 separated by six planes meeting at the the four lines [ boundary ] @xmath102 the eigenvalue dynamics has lower symmetry @xmath103 , since the degeneracy planes @xmath104 and @xmath105 admit entangled ground states and are inequivalent to the other four planes ; this is borne out by the observation that neither the eigenvalue dynamics @xmath106 nor the success probability is invariant under all permutations of the diagonal elements of @xmath11 .
identification of the symmetry structure of larger systems may cast further light on the @xmath19-qubit case .
figure [ fig : stability ] shows a constant-@xmath107 slice through this phase diagram .
the degeneracy planes are clearly indicated in the minimum - gap plot ; here the gap vanishes at @xmath108 .
the lower two plots demonstrate the non - adiabaticity of the time evolution , with the success probability increasing ( but not completely monotonically ) with distance from the degeneracy planes .
the energy error is non - monotonic : it is small at the degeneracy planes ( since the final state will have only a small admixture orthogonal to the degenerate ground states ) and small where a large gap reduces the probability of transitions . through ground - state phase diagram for an ensemble of two - qubit hamiltonians .
plots show minimum gap @xmath1 ( top left ) , position of minimum gap @xmath109 ( top right ) , success probability @xmath0 for @xmath71 ( bottom left ) and energy error @xmath110 for @xmath71 ( bottom right).,scaledwidth=50.0% ] these plots suggest a projection of a surface onto the @xmath111 plane ; we seek to find a suitable parameterization of the set of hamiltonians to collapse it onto a low - dimensional surface .
we find that a plot of @xmath67 against the minimum gap @xmath68 and the position @xmath112 of the gap indeed shows that all points lie close to a curved surface @xmath113 ( which rises with increasing @xmath2 ) .
this is understandable , since those two parameters largely determine the shape of the lowest two energy levels .
figure [ fig:3 ] shows a projection of this surface onto the @xmath114 plane .
visual inspection shows that the colour is to a good approximation a function only of position in the plane . note
that the position of the points depends only on the hamiltonian parameters , while the success probability depends also on the computation time . against minimum gap @xmath1 .
the @xmath96 have been chosen from the uniform distribution @xmath70 for @xmath115 random problem instances .
points are coloured by the success probability @xmath0 at @xmath71.,scaledwidth=75.5% ]
we have shown that the relationship between the success probability and @xmath1 is not a pure functional relationship for simple two - qubit systems . however , it is important to determine whether the interesting structure in this relationship remains in larger systems . to determine whether these densely - populated bands represent groups of problem instances that have followed similar evolution paths for the state vector ( e.g. the system remaining mostly in the ground state , then being excited at a single avoided crossing ) , we calculated the average overlap with the ground state : @xmath116 the points in figure [ fig : large_systems ] are coloured with respect to this average overlap value , @xmath117 , and we can see a smooth graduation across the figures , with the average overlap with the ground state increasing with the success probability .
the exception to the smooth graduation of @xmath117 is the densely populated band where @xmath118 .
this band must consist of cases with a degenerate or near - degenerate ground state at @xmath119 , as it includes cases which remain close to the instantaneous ground state throughout the majority of the evolution but have a low success probability .
these results also lend credence to the idea that the structure is closely linked to the choice of hamiltonian parameters .
we note that these distributions are reminiscent of the 2d projections of the higher - dimensional equilibrium surfaces seen in catastrophe theory @xcite . in this case success probability ,
@xmath1 and @xmath117 are all internal variables of the system and not independent control parameters , so we are looking at a different situation to those usually studied in catastrophe theory . identifying the nature of this surface and the dimensions of the phase space that it exists in
is an important task , as it could have a major impact on adiabatic algorithm design . at this point
we can conjecture that the constraint originates from an adiabatic invariant of the hamiltonian .
@xmath120 real parameters are required to specify the density matrix of @xmath19 qubits , reducing to @xmath121 for a pure state as discussed here .
the pechukas - yukawa approach to eigenvalue dynamics ( see e. g. ref .
@xcite ) , which can be extended to density - matrix dynamics @xcite , has at least @xmath30 adiabatic invariants , thus reducing the number of parameters required .
we find it strange that , to the best of our knowledge , there has been no research on adiabatic invariants of adiabatic quantum computers .
we speculate that a systematic investigation of adiabatic invariants of quantum computers especially adiabatic and approximately adiabatic computers could yield important information about their behaviour and have a major impact on adiabatic algorithm design .
we have shown that the relationship between the success probability and the minimum ground state gap may not be as straightforward as is often assumed .
there is a rich structure of distinct sharp edges and densely - populated bands in the distribution , particularly in smaller systems .
a partial explanation has been proposed , whereby this is the projection of a higher - dimensional surface ; identification of the parameters governing this surface will guide understanding of the set of problems amenable to adiabatic quantum computing .
we do not propose a definitive explanation of the origin of this rich structure : this remains an open question . | we explore the relationship between two figures of merit for an adiabatic quantum computation process : the success probability @xmath0 and the minimum gap @xmath1 between the ground and first excited states , investigating to what extent the success probability for an ensemble of problem hamiltonians can be fitted by a function of @xmath1 and the computation time @xmath2 .
we study a generic adiabatic algorithm and show that a rich structure exists in the distribution of @xmath0 and @xmath1 . in the case of two qubits
, @xmath0 is to a good approximation a function of @xmath1 , of the stage in the evolution at which the minimum occurs and of @xmath2 .
this structure persists in examples of larger systems . |
this article describes the macaulay 2 package pierimaps , which defines maps of representations of the general linear group @xmath0 of the form @xmath4 and presents them as degree 0 maps @xmath5 of free modules over the polynomial ring @xmath6 . here
dominant weights of @xmath0 are identified with weakly decreasing sequences @xmath7 of length @xmath8 , and @xmath9 denotes the irreducible representation of highest weight @xmath7 .
such maps are of general importance , and appeared recently in the work of eisenbud , flystad and weyman @xcite .
the package also describes certain related maps in characteristic @xmath10 .
we give some context for this work and then describe the contents of this article .
let @xmath11 be a field , and @xmath12 $ ] the polynomial ring in @xmath8 variables .
in a recent paper of eisenbud and schreyer @xcite , a theorem regarding the `` shape '' of minimal free resolutions of cohen
macaulay @xmath13-modules was established . given a cohen
macaulay @xmath13-module @xmath1 , the * betti diagram * @xmath14 is the number of generators of degree @xmath15 in the @xmath3th syzygy module of a minimal free resolution of @xmath1 .
a betti diagram is * pure * if for each @xmath3 , @xmath16 for at most one @xmath15 . in this case , we let @xmath17 be the degree sequence of @xmath18 : that is , @xmath19 for all @xmath3 .
the theorem mentioned above states that any betti diagram of a cohen macaulay module is a rational linear combination of pure betti diagrams .
the herzog khl equations ( * ? ? ?
* theorem 1 ) show that each strictly increasing degree sequence determines the corresponding betti diagram up to a rational multiple . for @xmath20 , eisenbud , flystad , and weyman
@xcite constructed pure free resolutions for each degree sequence which live in the category of @xmath21 representations .
eisenbud and schreyer give a characteristic free construction which gives different rational multiples of the betti diagrams in general .
it is still not completely known which multiples can and can not come from the betti diagram of a graded module .
for example , it is an interesting open problem to determine for a given degree sequence the smallest integer multiple given by the herzog
khl equations which actually comes from a module .
it is the goal of this article to describe how these resolutions can be represented concretely in macaulay 2 @xcite .
one can find @xmath22-forms for these maps and work in positive characteristic , and one such @xmath22-form is implemented , but in general it will not produce pure resolutions . we should mention that this choice of @xmath22-form is not unique .
it would be interesting to investigate how often the characteristic @xmath10 resolutions will be pure , and to construct @xmath22-forms which give equivariant pure resolutions in positive characteristic .
the rest of the article is organized as follows . in section [ representations ] , we review a construction for representations of @xmath0 . in section [ purefreesection ] ,
we give the construction of eisenbud , flystad , and weyman , and its extension to characteristic @xmath10 . in section [ combinatorialsection ]
, we describe the differentials in terms of bases , in the way that it is implemented in pierimaps , and illustrate an example .
finally , in section [ m2code ] we give some examples of macaulay 2 code which show how one can use this package .
in this section , we present a construction for irreducible polynomial representations of the rational algebraic group @xmath0 which is convenient for our purposes .
let @xmath23 ( @xmath24 ) be a partition and @xmath25 .
the * young diagram * of @xmath7 is a pictorial representation of @xmath7 : we draw @xmath8 rows ( some may be empty ) of boxes with @xmath26 boxes in the @xmath3th row , making sure that each row is left - justified .
the notation @xmath27 refers to the box in the @xmath3th row and @xmath15th column .
a * filling * of shape @xmath7 is an assignment of the numbers @xmath28 ( repetitions allowed ) to the boxes of the young diagram of @xmath7 . by picking some order on the boxes of @xmath7 , we get an action of the symmetric group @xmath29 on the fillings of @xmath7 .
we ll say that @xmath30 is * row - preserving * if it permutes the rows of @xmath7 amongst themselves . the * schur module
* , but in characteristic 0 , it is isomorphic to what is usually called the schur module . ]
@xmath9 is the rational vector space with basis given by the fillings @xmath31 of @xmath7 together with the following relations : 1 .
( symmetric relation ) @xmath32 for any row - preserving permutation @xmath33 .
2 . ( shuffle relation ) for @xmath3 and @xmath15 such that @xmath27 and @xmath34 are boxes of @xmath7 , let @xmath35
. then @xmath36 , where the sum is over all permutations which fix all boxes not in @xmath37 . for more details ,
the reader is referred to ( * ? ? ?
* proposition 2.1.15 ) where our notion of schur module is called a weyl functor , and is denoted by @xmath38 .
the above presentation implicitly replaces the use of divided powers with symmetric powers , but this distinction is irrelevant in characteristic 0 .
a filling is a * semistandard tableau * if the numbers are weakly increasing from left to right along rows , and strictly increasing from top to bottom along columns .
a basis for @xmath9 ( over @xmath39 ) is given by the semistandard tableaux of shape @xmath7 .
we define an action of @xmath0 on @xmath9 as follows . given a filling @xmath31 , let @xmath40 be its entries ( in some order )
. for @xmath41 , set @xmath42 where the sum is over all index sets @xmath43 , and @xmath44 is the filling obtained by replacing each @xmath45 by @xmath46 .
we will need * pieri s formula * : if @xmath47 is a partition with one row , then @xmath48 where @xmath49 ranges over all partitions with at most @xmath8 parts obtained from @xmath7 by adding @xmath50 boxes , no two of which are in the same column .
thus there are inclusions ( unique up to scalar multiple ) @xmath51 which we will call * pieri inclusions*. we remark that this direct sum decomposition is only valid in characteristic 0 , and is the main barrier to extending the setup of this article to positive characteristic .
fix a degree sequence @xmath52 . define a partition @xmath53 by @xmath54 , and for @xmath55 , define partitions @xmath56
let @xmath57 $ ] , and define @xmath13-modules @xmath58 for @xmath59 by @xmath60 ( here @xmath61 denotes a grading shift by @xmath62 . )
the natural action of @xmath0 on @xmath63 and on @xmath64 gives an action of @xmath0 on @xmath65 .
note that @xmath66 , and that @xmath67 is obtained from @xmath68 by adding boxes only in the @xmath3th row , so there exists a pieri inclusion @xmath69 identifying @xmath70 gives a degree 0 map @xmath71 given by @xmath72 .
[ equivariantres ] with the notation above , @xmath73 is a @xmath0-equivariant minimal graded free resolution of @xmath74 , which is pure of degree @xmath50 .
furthermore , @xmath75 is isomorphic , as a @xmath0 representation , to the direct sum of all irreducible summands of @xmath76 corresponding to the partitions that do not contain @xmath77 , and in particular is a module of finite length .
see ( * ? ? ?
* theorem 3.2 ) . we should emphasize that our notation for partitions differs from that of ( loc .
cit . ) in that the notions of rows and columns are interchanged .
an implementation of the map @xmath78 is given in the method purefree in pierimaps .
one can then compute the remaining maps and compose them , or compute a minimal free resolution using macaulay 2 .
the extension of this construction to characteristic @xmath10 in general does not produce pure resolutions .
the idea is to clear the denominators in the matrix giving a pieri inclusion , remove the torsion from the cokernel , and then reduce coefficients modulo @xmath10 , for a given prime .
this is also implemented in the method purefree : the user need only specify a characteristic .
we wish to describe how the pieri inclusions work in terms of the bases of semistandard tableaux of the schur modules @xmath9 .
the pieri inclusion which induces the map @xmath79 is given by @xmath80 and @xmath68 is obtained from @xmath67 by removing @xmath81 boxes from the @xmath3th row .
so to describe this map , we first describe the case @xmath82 . in this case , the map was described by olver in @xcite . for the general case
, one can iterate this process of removing one box at a time and compose the maps .
it needs to be proved ( though it is not hard ) , that such a composition is the desired map , and in fact , if one removes boxes from multiple rows , the order in which the boxes are removed is irrelevant ( up to nonzero scalar multiple ) .
suppose we have a partition @xmath7 with @xmath83 .
let @xmath49 be the partition resulting from adding a box to the @xmath84th row of @xmath7 .
set @xmath85 , and for @xmath86 , define @xmath87 . given @xmath88 where @xmath89 $ ] is a monomial and @xmath31 is a filling of @xmath49 , we first interpret the @xmath90 as a `` zeroth '' row of @xmath31 consisting of @xmath91 boxes filled with @xmath92 @xmath3 s .
we ll call this a * shape*. given numbers @xmath93 , define @xmath94 to be the sum of all shapes obtained from @xmath88 by removing a box along with its entry ( and then the boxes to the right of it get shifted one to the left ) from row @xmath15 and moving it to the end of row @xmath3 . then set @xmath95 the desired map @xmath96 is now the alternating sum @xmath97 .
we give an example to illustrate all of the above .
[ combinpieri ] let @xmath98 , @xmath99 , @xmath100 , and consider the element @xmath101 where @xmath31 is the semistandard tableau @xmath102 we think of @xmath31 as having a `` row 0 '' which is an empty row on top of @xmath31 .
if @xmath103 , then @xmath104 which we really think of as @xmath105 in the module @xmath106 \otimes_{\mathbf{q}}{\mathbf{s}}_{(2,1 ) } v$ ] .
the equality follows from the relations described in section [ representations ] . in this case ,
we illustrate some of the main uses of pierimaps .
first , we load the package and define a polynomial ring in 3 variables @xmath108 $ ] : o3 = | 3a 0 b 0 c 0 0 0 0 0 0 0 0 0 0 | | 0 3a 0 b 0 c 0 0 0 0 0 0 0 0 0 | | 0 0 2a 0 0 0 2b 0 c 0 0 0 0 0 0 | | 0 0 0 2a 0 0 0 2b 0 c 0 0 0 0 0 | | 0 0 0 0 2a 0 0 0 b 0 2c 0 0 0 0 | | 0 0 0 0 0 2a 0 0 0 b 0 2c 0 0 0 | | 0 0 0 0 0 0 0 a -1/2a 0 0 0 3b c 0 | | 0 0 0 0 0 0 0 0 0 a -2a 0 0 2b 2c | this is the matrix of free @xmath13-modules induced by the pieri inclusion @xmath110 .
the bases of @xmath111 and @xmath112 can be listed with the commands standardtableaux(3 , \{3,1 } ) and standardtableaux(3 , \{2,1 } ) , respectively .
for example , the first command has \{\{0,0,0 } , \{1 } } as its first basis element , which is meant to represent the semistandard tableau @xmath113 .
alternatively , this map can be produced with the command pieri(\{3,1,0 } , \{1 } , a ) because the partition @xmath114 is obtained by subtracting 1 from the first entry of @xmath115 .
the module we are after is the cokernel of this map .
o6 = | x 0 y 0 0 z 0 0 0 0 0 0 0 0 0 | | 0 x 0 0 y 0 0 z 0 0 0 0 0 0 0 | | 0 0 0 y 0 0 z 0 0 0 0 0 0 0 0 | | 0 0 0 0 0 0 y 0 z 0 0 0 z 0 0 | | 0 0 0 0 0 0 y 0 z 0 0 0 0 0 0 | | 0 0 0 0 0 0 0 0 y z 0 0 y 0 0 |
| 0 0 0 0 0 0 0 0 0 0 x y 0 z 0 | | 0 0 0 0 0 0 0 0 0 0 0 0 x 0 z | now let s look at an example of changing the order of composition of pieri inclusions .
we know that there is a nonzero inclusion of the form @xmath116 .
there are two different ways to get this map with the function pieri .
we could remove a box from the second row of @xmath114 and then remove a box from the first row of @xmath117 to get the composition @xmath118 where @xmath119 is a pieri inclusion and @xmath10 is the quotient map : here , we see that the matrices differ by a scalar multiple of 2 . in general , different orders of box removals will yield the same matrix up to nonzero scalar multiple .
the differences arise from the denominators @xmath122 ( see ) .
the author thanks david eisenbud and jerzy weyman for helpful comments and encouragement while the package pierimaps was written , and for reading a draft of this article .
the author also thanks an anonymous referee for suggesting some improvements .
david eisenbud and frank - olaf schreyer , betti numbers of graded modules and cohomology of vector bundles , _
* 22 * ( 2009 ) , 859888 , available at http://arxiv.org/abs/0712.1843[arxiv:0712.1843 ] . | we describe a software package for constructing minimal free resolutions of @xmath0-equivariant graded modules @xmath1 over @xmath2 $ ] such that for all @xmath3 , the @xmath3th syzygy module of @xmath1 is generated in a single degree .
we do so by describing some algorithms for manipulating polynomial representations of the general linear group @xmath0 following ideas of olver and eisenbud flystad weyman . |
the _ level persistence _ for a real valued map was first considered in @xcite and @xcite and thought as a refinement of the standard persistence ( referred below as _ sub - level persistence _ ) .
it turned out to be a particular case of a more general persistence theory , the _ zigzag persistence _ proposed by carlsson and silva cf.@xcite .
the numerical invariants we have proposed for level persistence are the _ relevant persistence numbers _ and are equivalent with the four types of bar codes which came out from _ zigzag persistence_. their merits consist in the fact that they can be calculated using standard persistence algorithms with minor adjustments . in @xcite we have indicated how to calculate these numbers for a simplicial map via persistence algorithms slightly modified .
the purpose of this note is to make this relationship precise . in order to explain this we review the meaning of level persistence versus sub - level persistence and explain , from our perspective , the significance of bar codes and of relevant persistence numbers .
we inform the reader that the bar codes proposed by carlsson and silva are based on graph representations and derived decomposing the representations associated to the map in indecomposable components .
our approach is different .
we propose here two concepts _ death _
( left and right ) and _ observability or detectability _ ( left and right ) . the class of maps for which level persistence is naturally defined based on these two concepts is the class of _ tame maps_. so far all maps which appear in practice are tame . in particular any simplicial map @xmath0 , where @xmath1 is a finite simplicial complex and @xmath2 is linear on each simplex , and any morse function are tame .
tameness of a map actually signifies that the topology of the level changes at a discrete collection of values ( referred to as critical values ) .
precisely , [ tamemap ] a continuous map @xmath3 is called * tame map * ( cf .
definition 3.5 , @xcite ) if @xmath1 is a compact anr and there exists finitely many values @xmath4 ( so called critical values ) so that \(i ) for any @xmath5 there exists @xmath6 so that @xmath7 and the second factor projection @xmath8 are fiberwise homotopy equivalent .
\(ii ) for any @xmath9 there exists @xmath6 so that canonical inclusions @xmath10 and @xmath11 are deformation retractions .
the definition can be extended to incorporate @xmath1 locally compact anr s and @xmath2 proper maps . instead of finite collection of critical values
one requires that the set of critical values is a discrete sequence of numbers @xmath12 .
one can show that a simplicial map is tame cf.@xcite , with the set of critical values being among the values of @xmath2 on vertices . in practice ,
for a simplicial map , one can treat all values of @xmath2 on vertices as potential critical values .
the sub - level persistence needs a weaker concept , referred here as weakly tame map , which requires the change in the topology of sub - levels appearing only at finitely many @xmath13s .
precisely , [ wtamemap ] a continuous map @xmath3 is called * weakly tame map * if @xmath1 is a compact anr and there exist finitely many values @xmath4 ( so called critical values ) so that for any @xmath14 the inclusion @xmath15}\subseteq x_{(-\infty , t]}$ ] is a homotopy equivalence ( for the purpose of sub - level persistence , homology equivalence suffices ) . as above the definition can be extended to locally compct anr s .
clearly _ tameness _ implies _ weakly tameness_. the main results stated here are theorem [ theorem326 ] and theorem [ theorem327 ] in section [ ral ] and they were formulated in the author s ph.d thesis .
as suggested , we begin this note with recollection of sub - level persistence ( section [ slp ] ) , then general considerations about level persistence ( section [ lp ] ) , and ultimately the relation between the relevant persistence numbers and bar codes ( section [ ral ] ) .
note that when we refer to _ homology _ we mean homology with coefficients in a field @xmath16 fixed once for all .
the case @xmath17 and @xmath18 are the most familiar . in this case
the @xmath19-dimensional homology is a @xmath16-vector space and its dimension is referred to as betti number .
the author thanks d. burghelea for advise and help .
given a continuous map @xmath20 , the sub - level persistent homology introduced in @xcite and further developed in @xcite is concerned with the following questions : q1 .
does the class @xmath21 originates in @xmath22 for @xmath23 ? does the class @xmath21 vanishes in @xmath24 for @xmath25 ?
what are the smallest @xmath13 and @xmath26 such that this happens ?
the information that is contained in the linear maps @xmath27 for any @xmath28 is known as * sub - level persistence * and permits to answer the above questions .
recall that sub - level * persistent homology * is the collection of vector spaces and linear maps @xmath29 .
let @xmath30 .
one says that \(i ) the element @xmath31 is born at @xmath26 , @xmath23 , if @xmath32 is contained in img@xmath33 but is not contained in img@xmath34 for any @xmath6 .
\(ii ) the element @xmath31 dies at @xmath13 , @xmath35 , if its image is zero in img@xmath36 but is nonzero in img@xmath37 for any @xmath38 .
\(iii ) the element @xmath31 survives for ever , if its image is always nonzero in img@xmath36 for any @xmath35 .
note that most papers treat persistence for filtered spaces rather than for a map .
clearly a map provide a filtration by finitely many sub - levels if the map is weakly tame .
conversely , the standard construction _ telescope _ in homotopy theory permits to replace any finite filtered space @xmath39 by a weakly tame map @xmath20 simply by taking @xmath40 \cup_{\phi_1 } k_1\times [ t_1,t_2 ] \cup_{\phi_2}\cdots\cup_{\phi_{n-1}}k_{n-1}\times [ t_{n-1},t_n]\cup_{\phi_n}k_n$ ] where @xmath41 is the inclusion and @xmath42}$ ] the projection of @xmath43 $ ] on @xmath44 $ ] .
the sub - level persistence for a filtered space is the sub - level persistence of the associated weakly tame map .
when @xmath2 is weakly tame , the sub - level persistence for each @xmath45 is determined by a finite collection of invariants referred to as * bar codes for sub - level persistence * @xcite .
the @xmath19-bar codes for sub - level persistence of @xmath2 are intervals of the form @xmath46 or @xmath47 with @xmath48 .
the number @xmath49 of @xmath19-bar codes which identify to the interval @xmath46 is the maximal number of linearly independent homology classes in @xmath50 , which are born at @xmath51 , die at @xmath13 and remain independent in img(@xmath52 ) for any @xmath53 , @xmath54 .
the number @xmath55 of @xmath19-bar codes which identify to the interval @xmath47 is the maximal number of linearly independent homology classes in @xmath50 which are born at @xmath51 , and remain independent in img(@xmath52 ) for any @xmath56 .
it follows from the above definitions that for a weakly tame map the set of @xmath19-bar codes for sub - level persistence is finite and any @xmath19-bar code is an interval of the form @xmath57 or @xmath58 with @xmath59 critical values of @xmath2 and @xmath60 . from these bar codes
one can derive the * betti numbers * @xmath61 , the dimension of @xmath62 , for any @xmath28 and get the answers to questions q1 and q2 . for example , @xmath63.\ ] ] from the betti numbers @xmath64 one can also derive these @xmath19-bar codes . denote @xmath65=number of @xmath19-bar codes which equal to @xmath66 for @xmath67 , where @xmath68 is the smallest critical value .
we have ( see @xcite ) @xmath69 = & \hskip -1.5 mm \left\ { \hskip -1.5 mm \begin{array}{ll } \beta_r(t_i , t_{j-1})-\beta_r(t_{i-1},t_{j-1})-\beta_r(t_i , t_j)+\beta_r(t_{i-1},t_j ) , & t_0<t_i < t_j<\infty\\ \beta_r(t_0,t_{j-1})-\beta_r(t_0,t_j ) , & t_i = t_0 , t_0
< t_j<\infty\\ \beta_r(t_i,\infty)-\beta_r(t_{i-1},\infty ) , & t_0<t_i<\infty , t_j=\infty\\ \beta_r(t_0,\infty ) , & t_i = t_0,t_j=\infty \end{array } \right .
\end{array}\ ] ] the computation of the bar codes for a filtration of simplicial or polytopal complex or equivalently for a simplicial map is discussed in subsection 3.4 of @xcite when the coefficients field for homology groups is @xmath70 or @xmath71 .
the case of the field @xmath72 is taken from @xcite and is , by now , the well known elz - algorithm .
level persistence for a map @xmath73 was first considered in @xcite and was better understood when the zigzag persistence was introduced and formulated in @xcite . given a continuous map @xmath20 , level persistence is concerned with the homology of the fibers @xmath74 and addresses questions of the following type .
q1 . does the image of @xmath75 vanish in @xmath76 , where @xmath35 or in @xmath77 , where @xmath78 ?
q2 . can @xmath75 be detected in @xmath79 where @xmath35 or in @xmath80 where @xmath78 ? the precise meaning of detection is explained below .
what are the smallest @xmath13 and @xmath26 for the answers to q1 and q2 to be affirmative ? to answer such questions one has to record information about the following linear maps @xmath81
the * level persistence * is the information provided by this collection of vector spaces and linear maps considered for all @xmath51 , @xmath13 .
let @xmath82 .
one says that \(i ) @xmath83 dies downward at @xmath84 , if its image is zero in img@xmath85 but is nonzero in img@xmath86 for any @xmath87 .
\(ii ) @xmath83 dies upward at @xmath88 , if its image is zero in img@xmath89 but is nonzero in img@xmath90 for any @xmath91 .
we say that @xmath75 can be detected at @xmath92 , if its image in @xmath93 is nonzero and is contained in the image of @xmath94 .
similarly , the detection of @xmath32 can be defined for @xmath78 also . in case of sub - level persistence for tame maps the collection of the vector spaces and linear maps
is determined up to coherent isomorphisms by a collection of invariants called * bar codes for level persistence * which are intervals of the form @xmath95 $ ] with @xmath28 and @xmath96 , @xmath97 $ ] , @xmath46 with @xmath48 .
these bar codes are called _ invariants _ because two tame maps @xmath98 and @xmath99 which are fiber - wise homotopy equivalent have the same associated bar codes .
the above result can be derived from zigzag persistence but , in view of definitions above can be proven directly .
the details of the derivation are not contained in this paper .
an open end of an interval signifies the death of a homology class at that end ( left or right ) whereas a closed end signifies that a homology class can not be detected beyond this level ( left or right ) .
there exists an @xmath19-bar code @xmath100 if there exists a class @xmath101 for some @xmath102 which is detectable for @xmath103 and dies at @xmath26 and @xmath13 .
the multiplicity of @xmath100 is the maximal number of linearly independent classes in @xmath104 such that \(i ) all remain linearly independent in img(@xmath105 ) for @xmath106 and img(@xmath107 ) for @xmath108 ; \(ii ) all die at @xmath26 and @xmath13 .
notice that the change of @xmath51 above will not affect the multiplicity of @xmath109 .
there exists an @xmath19-bar code @xmath110 $ ] if there exists an element @xmath111 which is not detectable for @xmath112 and detectable for @xmath113 and dies at @xmath26 .
the multiplicity of @xmath110 $ ] is the maximal number of linearly independent elements in @xmath79 such that
\(i ) neither one is detectable for @xmath112 ; \(ii ) all remain linearly independent in img(@xmath114 ) for @xmath113 ; \(iii ) all dies at @xmath26 .
there exists an @xmath19-bar code @xmath115 if there exists an element @xmath116 which is not detectable for @xmath117 and detectable for @xmath118 and dies at @xmath13 .
the multiplicity of @xmath115 is the maximal number of linearly independent elements in @xmath80 such that \(i ) neither one is detectable for @xmath117 ; \(ii ) all remain linearly independent in img(@xmath119 ) for @xmath118 ; \(iii ) all dies at @xmath13 .
there exists an @xmath19-bar code @xmath120 $ ] if there exists an element @xmath116 which is not detectable for @xmath117 or @xmath112 and detectable for @xmath121 .
the multiplicity of @xmath120 $ ] is the maximal number of linearly independent elements in @xmath80 such that \(i ) neither one is detectable for @xmath117 or @xmath112 ; \(ii ) all remain linearly independent in img(@xmath119 ) for @xmath121 .
note , that a priory , the set of linearly independent elements in @xmath74 for each @xmath51 between @xmath13 and @xmath26 might be very different for different @xmath13s .
the tameness hypothesis insures however their consistency . in view of the description above for a tame map ,
the set of @xmath19-bar codes for level persistence is finite .
any @xmath19-bar code is an interval of the form @xmath122 $ ] with @xmath123 critical values or @xmath124 , @xmath125 $ ] , @xmath57 with @xmath126 @xmath59 critical values . given a tame map @xmath20 with critical values @xmath127 , denote by @xmath128 : = \text{the number of intervals } ( t_i , t_j ] \text { in } bl_r(f ) .
\\ n_r[t_i , t_j ) : = \text{the number of intervals } [ t_i , t_j ) \text { in } bl_r(f ) .
\\ n_r[t_i , t_j ] : = \text{the number of intervals } [ t_i , t_j ] \text { in } bl_r(f ) .
\\ \end{array}\ ] ] hence @xmath129 + n_r[t_i , t_j ) + n_r[t_i , t_j]$ ] . in figure [ figure421 ]
, we indicate the bar codes both for sub - level and level persistence for some simple map in order to illustrate their differences and what they have in common .
for example looking at figure [ figure421 ] the class consisting of the sum of two circles at level @xmath51 is not detected on the right , but is detected at all levels on the left up to ( but not including ) the level @xmath13 .
level persistence provides considerably more information than the sub - level persistence @xcite and the bar codes for the sub - level persistence can be recovered from the bar codes for the level persistence .
an @xmath19-bar code @xmath130 for level persistence contributes an @xmath19-bar code @xmath131 for sub - level persistence .
an @xmath19-bar code @xmath132 $ ] for level persistence contributes an @xmath19-bar code @xmath133 for sub - level persistence .
@xmath19-bar codes @xmath134 $ ] and @xmath135 for level persistence contribute nothing to @xmath19-bar codes for sub - level persistence .
an @xmath19-bar codes @xmath135 for level persistence contributes an @xmath136-bar code @xmath137 for sub - level persistence .
see figure [ figure421 ] and lemma [ levsublev ] below .
[ levsublev ] given a tame map @xmath98 with critical values @xmath138 .
we have @xmath139 @xmath140 + \sum_{l=0}^{i-1 } n_{r-1}(t_l , t_i)\ ] ] for any critical values @xmath141 .
item 1 follows from formulas ( [ e1 ] ) and ( [ e2 ] ) .
item 2 is more elaborate .
one uses formula ( [ e2 ] ) which calculates @xmath142 as @xmath143 . a calculation of @xmath144
can be recovered from corollary 3.4 in @xcite which implies that this number is exactly the number of @xmath145-bar codes of the form @xmath146 plus the number of @xmath19-bar codes of the form @xmath147 $ ] with @xmath148 .
clearly @xmath149 should be critical values .
a different derivation can be achieve independently of @xcite .
the bar codes for the level persistence can be also recovered from the bar codes for the sub - level persistence but from the bar codes of a collections of tame maps canonically associated to @xmath2 .
this will be described in the next subsection .
for this purpose one uses an alternative but equivalent way to describe the level persistence based on a different collection of numbers , referred below as relevant persistence numbers , @xmath150 .
@xmath161\txt{$i_r(t , t')$ } \ar@{=>}[rr]^{\hskip -1 cm thm\ ; 4.2 } & & * + [ f]\txt{$ l_r(t ) , l_r^+(t;t')$\\$l_r^-(t;t '' ) , e_r(t;t',t'')$ } \ar@{=>}[rr]^{\hskip 0.5 cm thm\ ; 4.3 } & & * + [ f]\txt{$n_r([t , t'])$\\ $ n_r((t , t'))$\\ $ n_r((t , t'])$\\ $ n_r([t , t'))$ } \ar@/_-4pc/[ll]^{\hskip -.8 cm observation\ ; 4.1 } \ar@/_-8pc/[llll]^{observation\ ; 4.1 } } \ ] ] 1 .
@xmath170 = number of intervals in @xmath171 which contain @xmath95 $ ] ; 2 .
@xmath172 = number of intervals in @xmath171 which contain @xmath51 ; 3 .
@xmath173 4 .
@xmath174 + \sum_{t''\leq t_i
< t < t_j } n_r(t_i , t_j);\ ] ] 5 .
@xmath175 to determine the numbers @xmath194 $ ] , @xmath195 and @xmath196 $ ] , we introduce the following auxiliary numbers @xmath197 : = \text { the number of intervals in } bl_r(f ) \text { which intersect the level } x_{t_i } \\ \hskip 2.1 cm \text { with closed end at } t_j;\\ n_r(t_i , t_j\ } : = \text { the number of intervals in } bl_r(f ) \text { which intersect the level } x_{t_j } \\ \hskip 2.1 cm \text { with open end at } t_i;\\ n_r[t_i , t_j\ } : = \text { the number of intervals in } bl_r(f ) \text { which intersect the level } x_{t_j } \\ \hskip 2.1 cm \text { with closed end at } t_i .
\end{array}\ ] ] the explicit calculation of the relevant persistence numbers @xmath172 , @xmath176 , @xmath177 and @xmath178 is discussed in subsection 4.4 of @xcite and is based on positive and negative bar codes which are defined and calculated in terms of sub - level persistence via minor adjustments of the elz algorithm .
alternatively , we can get the bar codes for the level persistence providing an alternative to the carson - silva algorithm cf .
@xcite which calculates the level persistence bar codes as bar codes for zigzag persistence . | the purpose of this note is to describe a new set of numerical invariants , _ the relevant level persistence numbers _ , and make explicit their relationship with the four types of bar codes , a more familiar set of complete invariants for level persistence .
the paper provides the opportunity to compare level persistence with the _ persistence
_ introduced by edelsbrunner - letscher - zomorodian called in this paper as _ sub - level persistence_. |
by now it has been demonstrated that neutrinos can change flavour . experiments with solar , atmospheric
, reactor and accelerator neutrinos have established that there is a nonzero probability that a neutrino created with a certain flavour is detected with a different one after having propagated for some distance , and that this probability is dependent on the propagated distance , @xmath3 , and the neutrino energy @xmath4 . the standard mechanism that explains neutrino flavour transitions makes use of two different bases : the basis of neutrino mass eigenstates , which have well - defined masses , and the basis of neutrino interaction states -the flavour basis- which are the ones that take part in weak processes such as @xmath5 decay .
the two bases are connected by means of a unitary transformation @xmath6 , so that we can write each of the flavour states @xmath7 as a linear combination of the mass eigenstates @xmath8 , i.e. @xmath9 where the coefficients @xmath10 are components of the unitary mixing matrix that represents the transformation . assuming the existence of three active neutrino families ( @xmath11 ) , as indicated by experiment @xcite , the summation index runs @xmath12 and @xmath6 is a @xmath13 matrix . using this definition for the flavour states it is straightforward to conclude that flavour transitions occur because neutrinos are massive , and because different mass eigenstates have different masses .
this can be seen , within the three - neutrino formalism , by writing the transition probabilities from @xmath14 to @xmath15 as @xmath16 where @xmath17 , with @xmath18 the mass of the @xmath19-th eigenstate , and @xmath20 . clearly , if @xmath21 , no transitions occur .
note the @xmath22 dependence on the energy associated with this standard , mass - driven , oscillation mechanism .
the experiments that have studied neutrino flavour transitions @xcite have been designed to detect neutrinos with energies that range from a few mev ( solar neutrinos ) to the tev scale ( atmospheric neutrinos ) .
the mass - driven mechanism mentioned in the preceding paragraph has been experimentally confirmed within this energy range .
notably , data from the super - kamiokande atmospheric neutrino experiment @xcite was used to find an energy dependence of the oscillation probability of @xmath23 , with @xmath24 at @xmath25 confidence level , thus confirming the dominance of the mass - driven mechanism in this energy range , and relegating any other possible mechanism due to new physics to subdominance .
it is possible , however , that one or more of such subdominant mechanisms become important at higher energies . in the present paper , we have explored a possible scenario where there is an extra oscillation mechanism present which results in an energy - independent contribution to neutrino oscillations .
this mechanism , though subdominant in the mev - tev range , might become dominant at higher energies .
we will deferr the detailed treatment of how the contribution is introduced to section [ section_theory ] and focus now on the possible mechanisms behind it .
an energy - independent contribution to neutrino oscillations is introduced by a vector coupling of the form @xmath26 such a term results in an energy - independent oscillation phase @xmath27 , with the @xmath28 eigenvalues of the @xmath29 matrix , so that the phase in eq .
( [ eqtransprob ] ) becomes @xmath30 and the mixing matrix @xmath6 is modified as well .
specially relevant to the detection prospects of new physics is the fact that not only the oscillation phase -which , as we will see , is lost when considering neutrinos from distant astrophysical sources- , but also its amplitude , is modified by such an energy - independent contribution . a vector coupling like that of eq . ( [ eqlagrangian ] ) could be induced by violation of lorentz invariance which in turn results in a violation of cpt symmetry .
a realisation of this possibility is the standard model extension @xcite , which includes cpt - odd terms in the lagrangian , such as that of eq .
( [ eqlagrangian ] ) .
a different possible mechanism is a nonuniversal coupling to a torsion field @xcite . in the next section
, we will argue that a possible energy - independent contribution to flavour oscillations could become important , with respect to the standard oscillation terms , when the neutrino energy is high ( pev or more ) .
the highest - energy available flux of neutrinos is the expected ultra - high - energy ( uhe ) flux from astrophysical sources -notably , active galaxies- which are located at distances in the order of tens or hundres of mpc . for these conditions , the flavour - transition probability in eq .
( [ eqtransprob ] ) oscillates very rapidly , and so we are moved to use its average value instead , which is obtained by averaging the oscillating terms in the expression , i.e. @xmath31 thus the information in the oscillation phase is lost , but the modifications due to the energy - independent new physics is preserved in the elements of the mixing matrix .
as we saw in the previous section , the standard scenario of neutrino oscillations considers each neutrino of a given flavour to be a linear combination of neutrino mass eigenstantes , namely @xmath32_{\alpha i } \lvert \nu_i \rangle ~,\ ] ] where we have renamed the mixing matrix @xmath33 .
the mass eigenstates @xmath34 satisfy schrdinger s equation and so propagate as @xmath35 where we have assumed that @xmath36 , so that @xmath37 , and that , because neutrinos are highly relativistic particles , @xmath38 ( in natural units )
. thus , after having propagated for a distance @xmath3 , the neutrino created with flavour @xmath39 has become @xmath40_{\alpha i } e^{-i\frac{m_i^2}{2e}l } \lvert \nu_i \rangle ~.\ ] ] we see that each mass eigenstate acquires a phase that depends on the value of its mass : the interference of these phases is the source of the flavour transitions . at detection ,
the probability amplitude of seeing a @xmath15 is @xmath41_{\alpha i}^\ast \left[u_0\right]_{\beta j } e^{-i\frac{m_i^2}{2e}l } \langle \nu_j \lvert \nu_i \rangle = \sum_i \left[u_0\right]_{\alpha i}^\ast \left[u_0\right]_{\beta i } e^{-i\frac{m_i^2}{2e}l}\ ] ] and so the probability @xmath42 for the transition @xmath43 turns out to be eq .
( [ eqtransprob ] ) .
( the reader is referred to @xcite for a more extended treatment of the basics of neutrino oscillations . ) the oscillation hamiltonian in the previous description of standard neutrino oscillations was written in the basis of mass eigenstates .
an alternative , but equivalent , description of the same phenomenon can be achieved by writing the hamiltonian in the flavour basis instead . using a hamiltonian in the flavour basis will allow us to introduce contributions from new physics in a more straightforward manner . if a neutrino is produced with flavour @xmath39 , then , after having propagated for a distance @xmath3 , its evolved state will be @xmath44 where the oscillation hamiltonian @xmath45 is the one corresponding to the standard , mass - driven , mechanism , and is written in the flavour basis .
@xmath45 is related to the hamiltonian in the mass basis -the `` mass matrix''- through a similarity transformation that makes use of the unitary mixing matrix @xmath33 : @xmath46 @xmath33 is the pontecorvo - maki - nakagawa - sakata mixing matrix , which can be written in terms of three mixing angles , @xmath47 , @xmath48 and @xmath49 , and one cp - violation phase , @xmath2 , as @xmath50 with @xmath51 , @xmath52 . using the latest data from solar , atmospheric , reactor ( kamland and chooz ) and accelerator ( k2k and minos ) experiments , the authors of @xcite found the best - fit values ( @xmath53 ) of the standard oscillation parameters to be @xmath54 the are no experimental values for @xmath2 presently .
motivated by the vector coupling considered in eq .
( [ eqlagrangian ] ) , and in analogy to the standard oscillation scenario , we can introduce an energy - independent contribution in the form of the hamiltonian ( also in the flavour basis ) @xmath55 where @xmath27 .
following @xcite , we write the mixing matrix in this case as @xmath56 the mixing angles associated with this hamiltonian are @xmath57 , @xmath58 , @xmath59 , and @xmath60 fills the role of @xmath2 in the standard hamiltonian .
the two extra phases , @xmath61 and @xmath62 , appear because of how the mixing matrix between flavour and mass eigenstates was defined in section [ section_theory_sub_stdosc ] .
@xmath63 is dependent on eight parameters -two eigenvalues ( @xmath64 , @xmath65 ) , three mixing angles ( @xmath57 , @xmath58 , @xmath59 ) and three phases ( @xmath60 , @xmath61 , @xmath62)- whose values are currently unknown . experimental upper limits
@xcite , however , have been set for @xmath64 , using solar and super - kamiokande data , and @xmath66 , using atmospheric and k2k data : @xmath67 the full hamiltonian , including standard oscillations and the energy - independent contribution , is then @xmath68 in section [ section_intro ] , we saw that @xmath45 has been experimentally demonstrated to be the dominant contribution to the oscillations in the low energy ( mev - tev ) regime : according to eq .
( [ eqbijlimits ] ) , the values of the @xmath69 are too low for the new physics , if there is any , to manifest at these energies . the @xmath22 dependence of @xmath45 , however , allows us to explore the possibility that , at higher energies , when its contribution is reduced , the effect of an energy - independent hamiltonian @xmath63 becomes comparable to it or even dominant .
such energy requirement is expected to be fullfilled by the uhe astrophysical neutrino flux ( see section [ section_intro ] . )
we would like to write the flavour transition probability corresponding to this hamiltonian in a form analogous to eq .
( [ eqtransprobavg ] ) . in order to do this , we need to know what is the mixing matrix @xmath70 between the flavour basis and the basis in which @xmath71 is diagonal . according to basic linear algebra , this
is achieved simply by diagonalising @xmath71 , finding its normalised eigenvectors , and building @xmath70 by arranging them in column form .
the components of the resulting matrix are in general complicated functions of the standard mixing parameters ( @xmath72 , @xmath73 , @xmath2 ) and of the parameters of @xmath63 ( @xmath74 , @xmath75 , @xmath76 , @xmath61 , @xmath62 ) . in analogy to eq .
( [ eqtransprobavg ] ) , the average flavour transition probability associated to the full hamiltonian @xmath71 is then @xmath77_{\alpha i}\rvert^2 \lvert\left[u_f\right]_{\beta i}\rvert^2 ~.\ ] ] by comparing the mixing matrix obtained by diagonalisation of @xmath71 with a general pmns matrix , eq . ( [ eqpmnsmatrix ] ) with mixing angles @xmath78 and
phase @xmath79 , we are then able to calculate how the effective mixing angles @xmath78 vary with the parameters of @xmath63 and @xmath2 .
succintly put , we have @xmath80 up to this point , @xmath70 is dependent on @xmath81 free parameters .
however , the standard mixing parameters @xmath82 , @xmath83 , @xmath47 , @xmath48 and @xmath49 have been fixed by neutrino oscillation experiments .
additionally , in order to simplify the analysis , in the present work we have set all of the phases to zero , i.e. @xmath84 .
furthermore , also to simplify our analysis , we have made the eigenvalues of @xmath45 proportional to those of @xmath63 , at an energy of @xmath85 pev , that is , @xmath86 with @xmath87 the proportionality constant .
the upper bounds on the @xmath69 , eq .
( [ eqbijlimits ] ) , are satisfied for @xmath88 .
thus we are left with only four free parameters to vary : @xmath87 , @xmath57 , @xmath58 and @xmath59 .
in the preceding section , we saw that , in order for the energy - independent contribution to flavour transitions to be visible , we would need to use the expected high - energy astrophysical neutrino flux .
as mentioned in section [ section_intro ] , the sources of this flux , e.g. active galaxies , are located at distances of tens to hundreds of mpc , so that the appropriate flavour transition probability to be used is the average one , eq .
( [ eqtransprobavg ] ) .
if , at the sources , neutrinos of different flavours are produced in the ratios @xmath89 , then , because of flavour transitions during propagation , the ratios at detection will be @xmath90 for @xmath91 . evidently , the initial flavour ratios depend on the astrophysics at source , which is currently not known with high certainty .
the standard initial flux @xcite considers that the charged pions created in proton - proton and proton - photon collisions decay into neutrinos and muons , which decay into neutrinos , too : @xmath92 such process yields @xmath93 . in the standard oscillation scenario , i.e. in the absence of the energy - independent contribution , plugging this initial flux into eq .
( [ eqfluxdetected ] ) , and using the best - fit values of the mixing angles , eq .
( [ eqstdmixpar ] ) , yields equal detected fluxes of each flavour , i.e. @xmath94 in a different production process @xcite , the muons produced by pion decay lose all their energy before decaying , so that a pure-@xmath95 flux is generated at the source , i.e. @xmath96 .
in the standard scenario , this initial flux yields a detected flux of @xmath97 . alternatively , a scenario of pure-@xmath98 production ( @xmath99 ) through beta decay has been considered , e.g. in @xcite . in this scenario ,
high - energy nuclei emmited by the source have sufficient energy for photodisintegration to occur , but not enough to reach the threshold for pion photoproduction .
the neutrons created in the process generate @xmath100 through beta decay .
for this initial flux , the resulting detected fluxes , in the standard oscillation scenario , are @xmath101 . in what follows
, we will consider the possibility of observing the energy - independent contribution of @xmath63 assuming these three initial flavour fluxes .
lllll production mechanism & initial flux & std . detected flux & @xmath102 & @xmath103 + & @xmath89 & @xmath104 & & + pion decay & @xmath105 & @xmath106 & @xmath107 & @xmath107 + muon cooling & @xmath108 & @xmath109 & @xmath110 & @xmath107 + beta decay & @xmath111 & @xmath112 & @xmath113 & @xmath107 +
we have defined the ratios @xmath114 and studied the effects of the new physics on them .
the standard values of @xmath115 and @xmath116 , along with the standard values of the detected fluxes @xmath117 , for each production scenario , are shown in table [ tblstdrs ] .
note that @xmath118 for every production mechanism because the value of @xmath49 used in the table was its best - fit value @xmath119 , which ensures yields equal detected fluxes of @xmath95 and @xmath120 due to maximal mixing .
figure [ figscattersr_only120_lambda100 ] shows , in light blue , the allowed region of values of @xmath115 and @xmath116 when @xmath121 and the mixing angles @xmath57 , @xmath58 and @xmath59 are allowed to vary , independently , within the range @xmath122 $ ] , with the standard mixing angles set to their best - fit values , eq . ( [ eqstdmixpar ] ) .
this value of @xmath87 already corresponds to the scenario where @xmath63 dominates over @xmath45 in the full hamiltonian ; higher values of @xmath87 will not modify the shape of the allowed @xmath123 region . for comparison , the darker - coloured region corresponds to the scenario where there is no energy - independent contribution ( i.e. @xmath124 ) and the standard mixing angles are allowed to vary within their @xmath53 bounds .
we have marked four notable points in the plot , a - d .
table [ tblnotablepts ] shows for each of them the values of the effective mixing angles @xmath78 and the detected fluxes if we now consider the two other possible initial fluxes , @xmath96 and @xmath111 , set @xmath121 to make @xmath63 the dominant contribution and allow each of the @xmath125 to vary within @xmath122 $ ] , we obtain the @xmath123 regions in figure [ figscattersr_1sigma_lambda100 ] .
the light purple and light brown regions correspond , respectively , to the initial fluxes @xmath108 and @xmath111 when @xmath121 .
the light blue region is the same region that was shown in figure [ figscattersr_only120_lambda100 ] , assuming an initial flux of @xmath105 and @xmath121 .
the dark blue , dark purple and dark brown regions correpond , respectively , to standard oscillations ( @xmath124 ) , allowing the standard mixing angles @xmath1 to vary within their @xmath53 bounds . cccl case & @xmath126 & @xmath127 & + a & @xmath128 & @xmath106 & standard mixing + b & @xmath129 & @xmath130 & maximal mixing @xmath131 ; @xmath120 s do nt mix + c & @xmath132 & @xmath105 & no effective mixing + d & @xmath133 & @xmath134 & only @xmath135 mix ; @xmath136 + because we are considering neutrinos that travel distances of tens or hundreds of mpc , neutrino decay is a possibility .
assuming that the neutrinos decay into products that are not detectable ( i.e. `` invisible daughters '' such as sterile neutrinos ) , then , following @xcite , the flux of flavour @xmath39 at earth will be @xmath137_{\beta i}\rvert^2 \lvert \left[u_0\right]_{\alpha i}\rvert^2 e^{-l/\tau_i } ~~\underrightarrow{l\gg\tau_i}~~ \sum_{\beta = e,\mu,\tau } \sum_{i\text{(stable ) } } \phi_\beta^0 \lvert \left[u_0\right]_{\beta i}\rvert^2 \lvert \left[u_0\right]_{\alpha i}\rvert^2 ~,\ ] ] where @xmath138 is the lifetime of the @xmath19-th mass eigenstate in the laboratory frame .
as explained in @xcite , this expression corresponds to the case where the decay has been completed when the neutrinos arrive at earth . in a normal hierarchy , @xmath139 is the only stable state and so @xmath140_{\alpha1 } \rvert^2 \sum_{\beta = e,\mu,\tau } \phi_\beta^0 \lvert \left[u_0\right]_{\beta 1 } \rvert^2~,\ ] ] while in an inverted hierarchy @xmath141 is the stable state and @xmath140_{\alpha3 } \rvert^2 \sum_{\beta = e,\mu,\tau } \phi_\beta^0 \lvert \left[u_0\right]_{\beta 3 } \rvert^2~,\ ] ] in figure [ figscattersr_1sigma_lambda100 ] , decay to @xmath139 has been coloured red and decay to @xmath141 , green .
we see that the @xmath123 region for decay into @xmath139 does not intersect the region accessible with @xmath121 , assuming an initial flux of @xmath105 or @xmath108 .
it is , however , totally contained within the allowed region for @xmath111 , and thus there is a possibility of not being able to disentangle the effects of neutrino decay ( in a normal hierarchy ) from an energy - independent contribution , under this production model .
the region allowed for decay into @xmath141 , on the other hand , is superposed to that of @xmath108 , for values of @xmath142 and @xmath143 .
a notable feature of this plot is that it reveals that , while there are many opportunities of establishing the presence of an energy - independent contribution by measuring @xmath115 and @xmath116 , it is a more difficult task to deduce from this measurement what the initial flux was .
an extreme example occurs in the region around @xmath144 , @xmath145 , where the three assumptions of initial flux result in the same prediction for @xmath115 and @xmath116 .
in other words , if we measured values for @xmath115 and @xmath116 around this region , we could conclude ( provided there are no other new physics effects ) that there is an energy - independent contribution present , but we could not conclude what the initial flux was . on the other hand ,
if we measured a value of @xmath146 , then ( again , assuming there are no unaccounted effects ) , we could conclude both that there is an energy - independent contribution _ and _ that the initial flux was @xmath108 .
the same would occur if we measured @xmath147 , @xmath148 : in this case , in addition to concluding that an energy - independent contribution is present , we could also conclude that the initial flux was @xmath111 .
as evidenced in figure [ figscattersr_1sigma_lambda1 ] , if we now make the contributions from standard , mass - driven , oscillations and energy - independent new physics comparable , effectively setting @xmath149 , the regions of allowed values of @xmath115 and @xmath116 do not change significantly with respect to the case with @xmath121 , corresponding to dominance of the new physics .
hence , it is not necessary for the new physics effects to be dominant in order for them to affect the astrophysical flavour neutrino flux in a measurable way . in the @xmath149 case , however , we recover some capability to disentangle neutrino decays to @xmath139 from the allowed region corresponding to the production mechanism @xmath111 .
additionally , for values of @xmath150 and @xmath151 , it seems to be possible to identify , in principle , the production mechanism to be @xmath105 . because @xmath116 depends on the number of @xmath120 s detected , and given that this is expected to be a low number ( e.g. one every two years at icecube @xcite ) , the uncertainty on @xmath116 , using the present generation of neutrino telescopes ,
will be high .
figures [ figscattersr_1sigma_lambda1 ] and [ figscattersr_1sigma_lambda100 ] show estimated error bars at the points corresponding to standard oscillations using the best - fit values of the mixing angles , considering a @xmath152 uncertainty on @xmath116 , due to the low statistics of @xmath120 , and a lower value of @xmath153 for @xmath115 , dominated by the systematic error introduced by the atmospheric @xmath95 flux . nevertheless ,
knowledge of only @xmath115 could be enough to establish whether energy - independent new physics exists at the uhe scale or even to deduce what the initial was .
for instance , we have seen that , if a value of @xmath115 between @xmath154 and @xmath155 is measured then we could establish that there is energy - independent new physics and that the initial flux is @xmath108 .
we have shown that introducing an energy - independent contribution to neutrino flavour oscillations in the form of an additional term in the oscillation hamiltonian affects the mixing angles , which , in turn , modify the flavour - transition probabilities .
this contribution is motivated by scenarios of new physics such as flavour - dependent coupling to the gravitational field and violation of cpt symmetry .
given that the standard oscillation hamiltonian has a @xmath22 dependence on energy , we have considered in our analysis the expected ultra - high - energy astrophysical neutrino flux , so that it becomes possible for the dominant term in the full oscillation hamiltonian to be the energy - independent term .
this flux , however , propagates over distances of tens or hundres of mpc , so that we need to use the average flavour - transition probability and , more importantly , it makes sense to consider the possibility that neutrino decays occur .
the modifications to the mixing angles result in a change in the expected flux of each flavour that arrives at earth , which depends both on the initial flux of each flavour and on the oscillation probabilities . for the initial flavour fluxes ,
we have considered three possibilities : either neutrinos are produced by the decay of pions generated in proton - proton and proton - photon collisions , and of the muons produced by pion decay , which yields an initial flux of @xmath93 ; or , in a situation where the muons lose most of their energy before decaying , they are produced predominantly as direct products of the decay of pions , which yields a pure-@xmath95 flux of @xmath108 ; or else they are produced by the beta decay of neutrons liberated in the photodisintegration of nuclei , yielding a pure-@xmath98 flux of @xmath111 .
we have defined the ratios @xmath156 and @xmath157 , and shown that large deviations from their standard values are possible , in all three neutrino production scenarios , given our current lack of knowledge about the intervening new physics parameters .
both when the effects of the new physics are dominant as well as when they are comparable to the standard oscillation terms , our ability to deduce , from a single measurement of @xmath115 and @xmath116 , the existence of an energy - independent contribution _ and _ the value of the initial flux is dependent on the particular values of @xmath115 , @xmath116 measured . given that the number of @xmath120 s to be detected at current neutrino telescopes is very low ( about one every two years at icecube ) , however , the uncertainty on @xmath116 will be quite high . in spite of this , knowledge of only @xmath115
could be enough to establish the presence of the extra contribution or to deduce what the initial flux was , with particular sensitivity to a @xmath108 initial flux .
this work was supported by grants from the direccion academica de investigacion of the pontificia universidad catolica del peru ( projects dai-4075 and dai - l009 ) and by a high energy latinamerican - european network ( helen ) stt grant .
mb acknowledges the hospitality of ific during the development of this work . | we add an energy - independent hamiltonian to the standard flavour oscillation one .
this kind of physics might appear in theories where neutrinos couple differently to a plausible non - zero torsion of the gravitational field or more dramatically in the presence of cpt - violating physics in the flavour oscillations .
if this contribution exists , experiments at higher energies are more sensitive to their free parameters , and flavour conversion could be severely modified .
we show that this new physics modifies the neutrino mixing angles and find expressions that relate the new , effective , angles to the standard oscillation parameters @xmath0 , @xmath1 and @xmath2 and to the parameters in the new - physics hamiltonian , within a three - neutrino formalism .
we consider scenarios where the new parameters allow for extreme deviations of the expected neutrino flavour ratios at earth from their standard values .
we show that large departures of the standard flavour scenario are plausible , which would be a strong hint of the violation of a conserved symmetry . |
if the galaxy formed primarily through continuous merging of small dwarf systems as demanded by cold dark matter scenarios ( e.g. , navarro , frenk , & white 1997 ; klypin et al.1999 ; moore et al.1999 ; bullock , kravstov , & weinberg 2001 ; boily et al.2004 ) , possibly even until @xmath3 , then we ought to be able to trace this assembly through the ratio of the elemental abundances in stars .
this is because the color - magnitude diagrams of dwarf galaxies in the local group have shown a wide variety of star formation histories and physical properties ( e.g. , grebel 1997 ; mateo 1998 ; tolstoy et al.1998 ; dolphin et al.2003 , skillman et al.2003 ) , which are expected to lead to differences in chemical evolution , as predicted by many evolution scenarios ( e.g. , lanfranchi & matteucci 2004 ; pagel & tautvaisien 1998 ) . a variety of element ratios are expected in the stars of different ages and dwarf galaxy origin .
therefore , `` chemical - tagging '' ( a term from freeman & bland - hawthorn , 2002 ) of stellar populations in the galaxy should be possible if significant formation occurred through hierarchical merging .
the detailed and recent stellar abundances in the smallest , dwarf spheroidal galaxies ( dsphs ; shetrone et al.2001 , 2003 ; tolstoy et al.2003 ; geisler et al.2005 ) , have shown their stars tend to have lower [ @xmath0/fe ] abundance ratios than similar metallicity galactic field stars .
the [ @xmath0/fe ] abundance ratio is an important tracer of the relative contributions of sne ii / ia since only sne ii contribute to the production of the @xmath0 elements , whereas both contribute to iron .
the [ @xmath0/fe ] ratio is particularly sensitive to this difference at higher metallicities , when sne ia have begun to make a significant contribution to the chemical evolution of the system ; the sgr dwarf galaxy remnant is an excellent example of this , where the [ @xmath0/fe ] trend is similar to that of galactic halo stars , however offset such that it appears that sne ia have contributed significant amounts of iron at lower metallicities and [ @xmath0/fe ] is significantly subsolar by [ fe / h]=0 ( bonifacio et al.2004 ) .
however some galaxies , e.g. , draco , show low [ @xmath0/fe ] ratios even at the lowest metallicities ( [ fe / h ] @xmath4 ; shetrone , ct , & sargent 2001 ) , which may suggest that metal - dependent outflows of sne ii ejecta are also important .
many metal - poor dsph stars also show low [ y / eu ] abundance ratios , which depends on the r- and s - process contributions from sne ii and agb stars . because of the slower chemical evolution in the small dwarf galaxies , s - process enrichments from _ metal - poor _ agb stars are important and these stars have lower yields of light s - process elements , such as y ( travaglio et al.2004 ) . in dsph stars , the low y
is not mimiced by low ba ( as predicted ) , which results in high [ ba / y ] ratios compared to the galactic field stars ( venn et al.2004 ) .
this is most evident around metallicities of [ fe / h ] @xmath5 , before the more metal - rich agb stars can also contribute to the s - process yields , even dominate the total contribution which can bring the [ ba / y ] ratio , and others , into agreement with the galactic comparison stars ( e.g. , one metal - rich rgb / agb star in fornax has lower [ ba / y ] than the metal - poor rgb stars in fornax , in excellent agreement with galactic comparison field stars ) .
the chemical comparisons of a large sample of galactic field stars were compared to those in the small dsph galaxies by venn et al.(2004 ) .
there the stars were divided into galactic components based on their kinematics . while there is some marginal evidence that galactic halo stars with extreme retrograde orbits overlap in [ @xmath0/fe ] with the stars in dsphs , this was not evident in the [ ba / y ] ratios .
thus , it appears that no significant component of the galactic halo , nor the galactic thick disk which was also examined , could have formed from the mergers of these small dwarf galaxies .
however , these comparisons were based on the available detailed abundances in seven dsph galaxies , and primarily from stars in their central fields .
early results from the much broader cat survey in three dsphs ( sculptor , fornax , and sextans from the dart survey ; e.g. , see tolstoy et al.2004 for sculptor ) shows that the stars in their central fields tend to have higher metallicities than most of the stars out near their tidal radii .
it is possible then that these outer dsph field stars have different chemical abundance ratios , which may be more similar to the metal - poor stars in the galactic halo ; detailed abundance ratios for these outer metal - poor dsph stars are pending . aside from the selection of stars in the dsphs , the comparisons examined by venn et al.(2004 ) also did not rule out early merging , before the dwarf galaxies had a chance to have a significant and unique chemical evolution history ; merging with larger dwarf galaxies was also not ruled out .
it is also interesting to note that another important physical difference between the large and small dsphs galaxies is that the larger dwarf galaxies have higher masses and contain globular clusters ( gcs ; van den bergh 2000 ) . in this paper , we ask how do the gcs fit into this scheme ? can the gcs be used as a test of the merging history of large dwarf galaxies in the formation of our galaxy ?
if some of the galactic gcs have extragalactic origins , then we may assume that they trace the merging history of the _ large _ dwarf galaxies , given that gcs are only found in the larger milky way dwarf galaxies of the magellanic clouds , sagittarius ( sgr ) , fornax , and canis major ( cma ) .
we caution that the opposite is not true ; not all massive dwarf galaxies have gcs .
therefore we can only say that if we find an extragalactic gc that it must have come from a more massive dwarf galaxy . presumably , differences in star formation histories , chemical evolution , and/or the initial conditions ( mass , pre - enrichment , or environmental factors ) between the various dwarf galaxies could also leave discernable chemical imprints on their gcs that might be chemically tagged when merged into our galaxy .
it has been shown that the sgr gcs terzan 7 ( ter 7 ) and palomar 12 ( pal 12 ) follow the abundance trends of the sgr field stars ( e.g. , cohen 2004 ; tautvaisien et al.2004 ) .
in addition , the gcs and field stars analyzed in the large magellanic cloud ( lmc ; hill et al.2000 ; hill 2004 ) follow the same abundance trends that have lower [ @xmath0/fe ] ratios ( o , ca , and ti ; with the possible exception of mg ) than similar metallicity field stars in the galaxy , which is similar to the field stars in the small dwarf galaxies .
thus , the question is whether there is a population of gcs in the galaxy that have low [ @xmath0/fe ] ratios ( and possibly other chemical signatures ) that could be interpreted as these clusters having formed in dwarf galaxies and later were captured through merging .
there have been many papers that present and/or review the abundances in gcs , or compare a subset of gcs to one another and/or to galactic field stars ( most recently sneden , ivans , & fulbright 2004 ; gratton , sneden , & carretta 2004 ) . in general , the galactic gcs show amazing uniformity in their [ @xmath0/fe ] ratios ( plateau levels @xmath6dex ) , and they follow the abundance trends seen in the galactic field stars when plotted as a function of the [ fe / h ] values .
this seems incredible in terms of understanding hierarchical galaxy formation .
carney ( 1996 ) showed that there is little or no relationship between [ @xmath0/fe ] and age for the gcs .
given sne ia contributions lead to a decrease in the [ @xmath0/fe ] ratios with increasing metallicity as seen in the galactic field stars , carney argued that the lack of a similar turn - down for the gcs implies a lack of sne ia contributions to the gcs . therefore ,
if the timescale of when the type ia supernovae significantly contribute to the interstellar medium is short , the `` old '' halo and disk gcs could not share a common chemical history and one of the populations must have been later accreted .
the exceptions include gcs associated with the sgr dwarf galaxy , e.g. , ter 7 and pal 12 , which have lower [ @xmath0/fe ] ratios than comparison galactic stars ( cohen 2004 ; sbordone et al.2005 ; tautvaisien et al.2004 ) , and both have younger ages than typical galatic halo gcs . also , @xmath1 centauri ( pancino et al.2002 ) and ruprecht 106 ( rup 106 ; lin & richer 1992 ; brown , wallerstein , & zucker 1997 ) show peculiar chemical abundances and are thought to be captured clusters .
but when it comes to gcs , the more common questions have been related to the internal variations in cno and namgal observed in their red giant stars .
these variations are attributed to a combination of initial composition differences coupled with internal mixing mechanisms ( gratton et al.2004 ) , though others have been exploring the possibilities and predictions of enrichments from early agb stars during the cluster formation process ( cottrell & da costa 1981 ) . because these abundance anomalies are seen only in gcs and never in field stars , including in dwarf galaxy field stars , shetrone et al.(1998 , 2001 , 2003 ) concluded that the gcs can not have formed from the small dwarf galaxies , e.g. , during the merging event .
however , gcs formed in large dwarf galaxies that later merged into the galactic halo and survived has not been ruled out . in this paper , we re - examine the element ratios in as many galactic gcs as available in the literature to compare with the galactic field stars and stars in the dwarf galaxies .
the goal is to indentify candidate gcs that show signatures of extragalactic origins either chemically or ideally both chemically and kinematically .
while it has been found that smaller dwarf galaxies ( such as the present - day milky way dsph satellites ) have contributed little or nothing to the galactic halo ( at least from those stars as summarized in venn et al.2004 ) , the gcs should examine the contributions of larger dwarf galaxies .
the field stars with detailed abundance ratios include only stars that are currently in the solar neighborhood , while the gcs sample all parts of the galaxy , including the bulge .
of particular interest are the [ @xmath0/fe ] ratios , as well as the light r - process ratios which have been successful in separating the galactic field stars from stars in the dsphs .
we study the gc abundance ratios with regards to their kinematic populations having adopted the galactocentric velocities ( @xmath7 ) from dinescu et al.(1999a , b , 2000 , 2001 , 2003 , and private communications for updates ) to separate the gcs into standard galactic components ( as discussed in 3 ) . by noting which component of the galaxy the clusters belong to
, we also try to identify clusters that may have unusual abundance signatures for their kinematically assigned galactic component .
detailed elemental abundances are available for 45 gcs ; 32 clusters with @xmath83 stars analyzed .
these clusters are listed in table 1 , along with the number of stars analysed in each cluster and references .
the literature was searched for high - resolution abundances determined within the past 20 years .
the oldest analysis included here is by gratton , quarta , & ortolani ( 1986 ) for three gcs .
a majority of the analyses are from the past five years .
we have found that studies from older than 1990 used lower resolution spectra ( @xmath9 ) .
more modern studies that use lower resolution ( 18000 or less ) include mcwilliam , geisler , & rich ( 1992 ) , mishenina , panchuck , & samus ( 2003 ) , cavallo , suntzeff , & pilachowski ( 2004 ) .
we consider the abundance ratios coming from these studies as being less reliable . in most cases when abundance ratios based on higher resolution spectra are available , those from lower resolution spectra are not included in the final mean ratios .
table 2 lists the elemental abundances and their standard error of the mean for select elements from mg to eu for each of the gcs in table 1 .
we adopted the solar abundances and @xmath10 values used by the lick - texas group as the standard values ( solar abundances , grevesse & sauval 1998 ; @xmath10 , see table 3 in fulbright 2000 for example ) .
all abundance ratios were corrected to these values ( observed - adopted ) in an effort to standardize the different datasets values were given in the published article . ] .
details on the adjustments are given in 4 . in table 2
, the final individual abundance ratios relative to fe ( shown in bold ) were averaged together by weighted averages according to the number of stars that were used in each study .
the [ @xmath0/fe ] and comparison ratios , e.g. [ ba / y ] , were calculated using the final weighted - mean ratios .
the italicized ratios were not used in calculating the final weighted mean abundance ratios .
the gcs in table 2 span a range of @xmath11 ( m15 ) @xmath12 [ fe / h ] @xmath13 ( ngc 6528 ) . in a majority of the cases , the [ fei /
h ] and [ feii / h ] ratios were averaged when both were available .
similarly , the [ ti / fe ] ratio is the average of the [ tii / fe ] and [ tiii / fe ] ratios when both were given for the cluster , otherwise it is the [ tii / fe ] ratio .
ideally , it would be best to match ionization states when determining the ti ratios , e.g. , [ tii / fei ] and [ tiii / feii ] .
however , in most cases not enough information was presented to determine these values .
many papers give the final abundances [ tii / fe ] and [ tiii / fe ] with no clear explanation how [ fei / h ] and [ feii / h ] were averaged together to determine [ fe / h ] . as a result
, we used the mean [ fe / h ] ratio when determining the ti ratios .
our selection of elements was based on those determined regularly in gc stars , and those that have been useful in the analysis of stars in dsph galaxies .
variations in cno , na , mg , and al are regularly found in clusters from star - to - star , and are usually attributed to mixing with cno - cycled gas ( e.g. , kraft 1994 ; sneden 2000 ; freeman & bland - hawthorn 2002 ; sneden et al.2004 ; gratton et al.2004 ) .
we have avoided these elements in this paper ( with the exception of magnesiummg@xmath14 dex ( see discussion in 5.1 ) ] given its relevance as a true @xmath0 element indicator ) because we are primarily concerned with global signatures of galaxy formation and not stellar or chemical evolution within gcs . for abundances in galactic field stars
, we have adopted those in table 2 in venn et al.(2004 ; see references therein ) .
the abundances in seven dsph galaxies are from shetrone et al.(2001 , 2003 ) and geisler et al.(2005 ) .
stellar abundances for 26 stars in the sgr dwarf galaxy remnant have also been added from bonifacio et al .
( 2000 ; 2004 ) and smecker - hane & mcwilliam ( 2005 ) .
we have determined the galactic stellar population component for 29 gcs based on a calculation of their phase space .
phase space distribution functions have been determined using the galactocentric velocity vector components ( v(@xmath7 ) in kms@xmath15 ) from dinescu et al.(1999ab , 2000 , 2001 , 2003 , and private communications for updates ) and the galactocentric positional vector components ( r(x , y , z ) in kpc ; harris 1996 ) .
firstly , we computed the probability for a gc to be associated with the galactic thin disk , thick disk , or halo from its velocity vector using a standard bayesian classification scheme and galactic gaussian velocity ellipsoid components from dehnen & binney ( 1998 ; thin disk ) , soubiran , bienaym , & siebert ( 2003 ; thick disk ) , and chiba & beers ( 2000 ; halo ) .
this method is the same as was used for the galactic field stars by venn et al.(2004 ) .
secondly , the probability for a gcs to be associated with these galactic components was determined from its positional vector . for this , a standard galactic model ( robin , reyle , derrire , & picaud 2003 ) was chosen with a maximum halo extent of r = 150 kpc , a softening parameter @xmath16=1.0 ( which is used to stop the density of the halo from going to infinite at the galactic center ) , and adopting an asympotitic r@xmath17 halo profile .
this was not necessary for the galactic field stars studied by venn et al.(2004 ) since it is a good assumption that those stars are in the solar neighbourhood quadrant .
the kinematic and positional probabilities were combined for a phase space determination of the final probabilities for each gc to be associated with each galactic component .
finally , we allocated any gcs within r = 2.7 kpc of the galactic center to a bulge component unless the kinematics placed it in another component , and did not probe more deeply into variations between bulge gcs and/or those associated with a bar ( e.g. , dinescu et al.2003 ) . table 3 lists the velocity and positional vector components for each gc , as well as which galactic component they belong to .
we have also added kinematic information for m54 and ter 7 which are embedded in the sgr dwarf galaxy remnant ; thus , on the assumption that these clusters are associated with the sgr remnant , then we have adopted the known kinematics for sgr ( ibata et al.1997 ) as a reasonable approximation to their @xmath7 velocities .
this allows us to highlight these clusters in our abundance ratios and kinematic analyses .
rup 106 is another cluster that does not have kinematic information , however it has a large galactocentric radial velocity ( @xmath18 kms@xmath15 ; harris 1996 ) . adopting a plausible range of proper motions ( 0``,@xmath191 '' in ra and dec )
, we investigate its potential space velocities ( @xmath20 and @xmath7 ) . in all cases , the galactocentric radial velocity , @xmath21 , is quite large ( @xmath22 to @xmath23 kms@xmath15 ; see table 4 ) implying rup 106 is a member of the galactic halo , possibly on a plunging orbit that would be consistent with a captured cluster . the final column in table 3 lists the gc classifications from ( mackey & gilmore 2004 ) which are based on the physical properties of gcs . while overall there is good agreement between the two classifications ,
there are some key differences especially for the clusters which have thick disk kinematics , while according to mackey & gilmore they have halo - like properties .
a toomre diagram for the gcs is shown in figure 1 , where the galactocentric rotational velocity @xmath24 is plotted against @xmath25 ( @xmath26 ) . to be consistent with the field stars examined by venn et al.(2004 ) , we plot the gcs as colored points according to the assigned galactic component ( e.g. , cyan = halo , green = thick disk , red = thin disk ) .
gcs assigned to the bulge are not shown in fig . 1 .
only three clusters stand out in this diagram ; pal 12 has thin disk kinematics however its positional vector places it in the halo ( thus cyan colored gc amongst the thin disk field stars ) , m68 which has an unusually high galactocentric rotational velocity ( @xmath27 kms@xmath15 ) , and m22 is a thick disk cluster with an unusually high @xmath25 component ( + 212 kms@xmath15 ) .
figure 2 shows the distribution in [ fe / h ] and [ @xmath0/fe ] vs. @xmath24 for the gcs compared with the field stars in the solar neighborhood .
venn et al.(2004 ) commented on the large and overlapping distribution in these abundances for the field stars in each of the galactic components .
for the [ @xmath0/fe ] plot , we have averaged the mean abundances of ca and ti per gc ( note that this differs slightly from the field star analysis by venn et al.2004 where [ @xmath0/fe ] represented an average of mg , ca , and ti ) .
the most outstanding gcs are pal 12 and rup 106 , with very low [ @xmath0/fe ] ratios .
of course , pal 12 is associated with the sgr dwarf galaxy remnant and rup 106 is thought to have been captured possibly from the magellanic clouds ( lin & richer 1992 ) . for the thick disk clusters ,
it is interesting to see that they are clustered both toward lower [ fe / h ] and higher [ @xmath0/fe ] .
although they are within the range of the thick disk field stars , the thick disk gcs do not show as wide of a spread in metallicity . in any case , we conclude that the [ @xmath0/fe ] ratios from the gcs in each kinematic component are in good to excellent agreement with those of the field stars in the same galactic component .
most of the data on gc stellar abundances comes from the lick - texas group .
therefore we have adopted the solar abundances they used ( grevesse & sauval 1998 ) , along with their @xmath10 values ( see table 3 of fulbright 2000 for references ) , as the fiducial standards on which all abundances are adjusted .
differences in solar abundances and @xmath10 are in the sense of the value from the reference paper minus our adopted values .
a number of clusters were not corrected because either their solar abundances and @xmath10 values matched the adopted ones ( no adjustments were made for differences of less than about 0.04 dex ) or no values were given .
these gcs were noted in table 1 . in the following
we note each cluster where adjustments were made to the abundance ratios .
no adjustments for differences in hyperfine splitting corrections have been made because such abundance ratio adjustments are not as simple as those for solar abundances and @xmath10 .
* ngc 104 ( 47 tucanae ) : * although the values from gratton , quarta , & ortolani ( 1986 ) are not used in the final mean abundances , we adjust them to match our adopted @xmath10 values
. there are no solar abundances given in their paper , so no adjustments were made .
[ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , and [ ba / fe ] are corrected by @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 , respectively , assuming the @xmath10 values are the same as in gratton ( 1987 ) .
the values from brown & wallerstein ( 1992 ) are not used in the final mean abundances , but we adjusted them to match our adopted solar abundances and @xmath10 values . since the @xmath33 values are given , the ratios are determined directly from those values using our adopted solar abundances . from the @xmath10 values , [ mg / fe ] , [ si / fe ] , and [ tiii / fe ]
have been corrected by @xmath34 , @xmath35 , and @xmath36 , respectively .
to avaoid confusion , in table 3 we have combined the results from gratton et al.(2001 ) and james et al.(2004a ) since they are both part of a collaboration that studies the same stars , but different element ratios .
the abundances from james et al.are
weighted means of the turnoff and subgiant star ratios . * ngc 288
, ngc 362 , ngc 5897 , ngc 6352 , and ngc 6362 : * no solar abundance values were given in gratton ( 1987 ) , so the only adjustments that were made were due to the @xmath10 values .
[ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , and [ ba / fe ] are corrected by @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 . for ngc 288 and
ngc 362 , the results from gratton ( 1987 ) are superceeded by those in shetrone & keane ( 2000 ) . *
ngc 2298 : * no changes are needed due to the solar abundance values in mcwilliam , geisler , & rich ( 1992 ) . only [ la / fe ] is adjusted by @xmath34 due to the @xmath10 values . *
ngc 3201 : * we do not use the gratton & ortolani ( 1989 ) ratios because two of the three stars are redone by gonzalez & wallerstein ( 1998 ) . examining the solar abundances , adjustments in [ mg / fe ] , [ tii / fe ] , and [ ba / fe ] needed to be made for the gonzalez & wallerstein abundance ratios .
further corrections need to be made due to differences in the @xmath10 values for [ mg / fe ] , [ tii / fe ] , and [ eu / fe ] .
it happens that the adjustments for the differences in the solar abundances and @xmath10 cancel each other out for [ mg / fe ] and [ tii / fe ] so that no adjustments are necessary .
the total changes for [ ba / fe ] and [ eu / fe ] are @xmath37 and @xmath38 , respectively . in determing
the gonzalez & wallerstein ratios , we have chosen to use the high resolution abundances from 1991 and 1994 for those stars observed multiple times .
* ngc 4590 ( m68 ) : * the results from shetrone et al.(2003 ) and lee , carney , & habgood ( 2004 ) superceed those from gratton & ortolani ( 1989 ) .
the abundance ratios derived from photometric values in lee , carney , & habgood were used . * ngc 5273 ( m3 ) : * the results of kraft et al.(1993 , 1995 ) were redone by sneden et al.(2004 ) .
in addition , the three stars in shetrone , ct , & sargent ( 2001 ) are also done in either sneden et al.or cohen & melendez ( 2005a ) . therefore , we did not use their abundance ratios in the final weighted - mean ratios . for the cohen & melendez study ,
the [ fe / h ] values were adjusted by @xmath39 due to a difference in the solar abundance value .
the abundance ratios [ mg / fe ] , [ ca / fe ] , and [ eu / fe ] were adjusted due to differences in the @xmath10 values by @xmath40 , @xmath41 , and @xmath42 . therefore the final abundance ratios for m3 are a weighted mean of those from sneden et al.(2004 ) and cohen & melendez ( 2005a ) . *
ngc 5466 : * there are no changes to the ngc 5466 abundance ratios .
we note that the only available chemical abundances come from the anomalous cepheid in this cluster .
this type of variable star can derive from either binary mass transfer or younger stars ( demarque & hirshfeld 1975 ; norris & zinn 1975 ; renzini , mengel , & sweigart 1977 ) .
because of the star s variability , we caution about the reliability of the abundance ratios from this single star even though they are consistent with other stars and clusters of similar metallicity . * ngc 5904 ( m5 ) : * although the values from gratton , quarta , & ortolani ( 1986 ) are not used in the final mean abundances , we adjust them to match our adopted @xmath10 values
. there are no solar abundances given in the paper , so no adjustments can be made .
[ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , and [ ba / fe ] are corrected by @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 , respectively , assuming the @xmath10 values are the same as in gratton ( 1987 ) . the ratios from sneden et al.(1992 ) are not used because the stars were reanalyzed by ivans et al.(2001 ) .
the asymptotic giant branch stars in the ivans et al.study were not included in the mean abundance ratios .
for the ramrez & cohen ( 2003 ) abundance ratios , adjustments are necessary for both the solar abundances and the @xmath10 values . for the solar abundances , [ mg / fe ] , [ ca / fe ] , [ tii / fe ] ,
[ fei / h ] , [ feii / h ] , and [ la / fe ] need to be adjusted by @xmath43 , @xmath44 , @xmath39 , @xmath42 , @xmath45 , and @xmath42 , respectively . for the @xmath10 values , [ mg / fe ] , [ ca / fe ] , and [ eu / fe ] need corrections of @xmath42 , @xmath41 , and @xmath42 , respectively .
this leads to a total adjustments of @xmath46 , @xmath47 , @xmath39 , @xmath42 , @xmath45 , @xmath42 , and @xmath42 for [ mg / fe ] , [ ca / fe ] , [ tii / fe ] , [ fei / h ] , [ feii / h ] , [ la / fe ] , and [ eu / fe ] .
after all of the correction were made , the ivans et al.and ramrez & cohen ratios were averaged together by a weighted mean .
* ngc 6093 ( m80 ) : * the solar abundances in cavallo , suntzeff , & pilachowski ( 2004 ) matched our adopted values .
for the @xmath10 values , correction were made to [ ca / fe ] , [ tiii / fe ] , and [ eu / fe ] by @xmath48 , @xmath47 , and @xmath28 , respectively .
* ngc 6121 ( m4 ) : * although the values from gratton , quarta , & ortolani ( 1986 ) are not used in the final mean abundance ratios , we adjust them to match our adopted @xmath10 values .
there are no solar abundances given in the paper , so no adjustments can be made .
[ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , and [ ba / fe ] are corrected by @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 , respectively , assuming the @xmath10 values are the same as in gratton ( 1987 ) .
the values from brown & wallerstein ( 1992 ) are not used in the final mean abundance ratios , but we adjusted them to match our adopted solar abundances and @xmath10 values . since the @xmath33 values are given , the ratios were determined directly from those values using our adopted solar abundances . from the @xmath10 values , [ mg / fe ] and [ si / fe ]
have been corrected by @xmath34 and @xmath35 , respectively .
ivans et al.(1999 ) re - observed one star in gratton , quarta , & ortolani ( 1986 ) and all three stars in brown & wallerstein ( 1992 ) and therefore supersedes these two earlier studies . *
ngc 6205 ( m13 ) : * the eighteen stars in kraft et al.(1997 ) were reanalyzed by sneden et al .
( 2004 ) ( the latter adopted here ) . as such , the [ mg / fe ] values from sneden et al.supercede those from kraft et al .
, while the remaining abundances were used to calculated the final weighted - mean ratios . for cohen & melendez ( 2005a ) , the [ fe / h ] ratios needed to be corrected by @xmath39 .
the ratios [ mg / fe ] , [ ca / fe ] , and [ eu / fe ] need to be corrected due to differences in @xmath10 values by @xmath40 , @xmath41 , and @xmath42 , respectively .
there are four stars that match between the sneden et al.and cohen & melendez . given
this is a small number compared to the total number of stars analyzed , we calculated the final abundance ratios from a weighted average of the two studies . * ngc 6287 , ngc 6293 , and ngc 6541 : *
no adjustments are needed for the lee & carney ( 2002 ) abundance ratios due to the solar abundances . however , [ mg / fe ] has been corrected by @xmath40 due to @xmath10 differences .
the ratios derived from photometric values were used . *
ngc 6341 ( m92 ) : * four of the six stars in shetrone ( 1996 ) have been reobserved by either sneden , pilachoski , & kraft ( 2000 ) or shetrone , ct , & sargent ( 2001 ) .
therefore , we will not use the shetrone ( 1996 ) results in calculating the final abundance ratios .
although , three of the four stars in shetrone , ct , & sargent were also observed by sneden , pilachoski , & kraft , we use the ratios for [ mg / fe ] , [ y / fe ] , and [ eu / fe ] from the former study in the final weighted - mean ratios because they were not calculated by the later study . * ngc 6342 : * the spectra in origlia , valenti , & rich ( 2005 ) were taking in the infrared , which does not allow for a direct comparison with the adopted lick - texas @xmath10 values .
no changes are necessary due to the solar abundances . * ngc 6528 : * no adjustments were made to the carretta et al.(2001 ) results due to the solar abundances .
[ mg / fe ] was corrected by @xmath49 because of the @xmath10 values .
no adjustments are needed for the zoccali et al.(2004 ) ratios due to the solar abundances . due to differences in the @xmath10 values , [ mg / fe ] and
[ ca / fe ] need to be adjusted by @xmath50 and @xmath51 .
the spectra in origlia , valenti , & rich ( 2005 ) were taking in the infrared , which does not allow for a direct comparison with the adopted lick - texas @xmath10 values .
no changes are necessary due to their solar abundances .
* ngc 6553 : * no solar abundances were given in barbuy et al.(1999 ) , so only adjustments were made due to the @xmath10 values .
their abundance ratios for [ mg / fe ] and [ la / fe ] have been corrected by @xmath37 and @xmath52 , respectively .
the only adjustment in the cohen et al.(1999 ) ratios is for [ ca / fe ] by @xmath46 due to difference between the adopted solar abundance .
as a side note , origlia , rich , & castro ( 2002 ) using infrared spectra give a general [ @xmath0/fe ] ratio for ngc 6553 of @xmath53 .
we do not include their results in table 2 because they do not give the individual star values .
* ngc 6656 ( m22 ) : * the values from brown & wallerstein ( 1992 ) have been adjusted to match our adopted solar abundances and @xmath10 values .
[ fei / h ] and [ feii / h ] are corrected by @xmath54 and @xmath48 , respectively .
there is no change to [ y / fe ] and [ la / fe ] . for the other ratios , [ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , [ tiii / fe ] , [ ba / fe ] , and [ eu / fe ]
are adjusted by @xmath54 , @xmath29 , @xmath45 , @xmath34 , @xmath55 , @xmath56 , and @xmath51 , respectively .
there is only one star common between both studies .
the final abundance ratios are a weighted mean of the two studies .
* ngc 6715 ( m54 ) : * due to differences in both the solar abundances and the @xmath10 values , we corrected [ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , [ ba / fe ] , [ la / fe ] , and [ eu / fe ] by @xmath57 , @xmath41 , @xmath58 , @xmath34 , @xmath59 , @xmath47 , and @xmath51 , respectively , for the brown , wallerstein , & gonzalez ( 1999 ) abundance ratios .
the [ fei / h ] and [ feii / h ] ratios need to be corrected by @xmath60 and @xmath61 . * ngc 6752 : * although the values from gratton , quarta , & ortolani ( 1986 ) are not used in the final mean abundances , we adjust them to match our adopted @xmath10 values
. there are no solar abundances given in the paper , so no adjustments can be made .
[ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , and [ ba / fe ] are corrected by @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 , respectively , assuming the @xmath10 values are the same as in gratton ( 1987 ) .
the only adjustments to the yong et al.(2003 ) ratios is for [ mg / fe ] by @xmath62 due to @xmath10 differences .
no changes need to be made to the cavallo , suntzeff , & pilashowski ( 2004 ) ratios for the solar abundances .
however , corrections need to be made due to the @xmath10 values
. [ ca / fe ] , [ tiii / fe ] , and [ eu / fe ] are corrected by @xmath48 , @xmath47 , and @xmath28 . to avoid confusion , in table 3
we have combined the results from gratton et al.(2001 ) and james et al.(2004a ) since they are both part of a collaboration that studies the same stars , but different element ratios .
the abundances from james et al.are
weighted means of the turnoff and subgiant star ratios . *
ngc 6838 ( m71 ) : * although the values from gratton , quarta , & ortolani ( 1986 ) are not used in the final mean abundances , we adjust them to match our adopted @xmath10 values
. there are no solar abundances given in the paper , so no adjustments can be made .
[ mg / fe ] , [ si / fe ] , [ ca / fe ] , [ tii / fe ] , and [ ba / fe ] are corrected by @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 , respectively , assuming the @xmath10 values are the same as in gratton ( 1987 ) .
there are difference in both the solar abundances and the @xmath10 values between our adopted ones and those from ramrez & cohen ( 2002 ) .
[ fe / h ] has been corrected by -0.06 .
the final abundance ratios [ mg / fe ] , [ ca / fe ] , [ y / fe ] , and [ la / fe ] have been corrected by @xmath39 , @xmath63 , @xmath37 , and @xmath64 . in table 2
, we include the results from lee , carney , & balachandran ( 2004 ) , which is an infrared study .
both stars in their study were already observed by sneden et al.(1994 ) with similar results . also , one of the two stars was observed in ramrez & cohen . to allow for a better comparison between the different datasets
, we do not include the ratios from the infrared study of lee , carney , & balachandran in the final weighted - mean ratios .
* ngc 7078 ( m15 ) : * the three stars in sneden et al.(2000 ) were also observed by sneden et al.(1997 ) and sneden , pilachowski , & kraft ( 2000 ) .
still , we include the results from sneden et al.(2000 ) in the final weighted - mean ratios because they are updates from the other two studies .
in addition , ten stars are common between the sneden et al.(1997 ) and sneden , pilachowski , & kraft ( 2000 ) studies .
therefore , in the cases where these two studies have common element ratios , those from sneden , pilachowski , & kraft supercede those from sneden et al.(1997 ) . *
ngc 7492 : * the [ fe / h ] value from cohen & melendez ( 2005b ) needs to be adjusted by @xmath45 due to differences in the solar abundances .
differences in solar abundances also require a change for the [ la / fe ] ratio by @xmath42 . from the @xmath10 difference , [ mg / fe ] , [ ca / fe ] , [ eu / fe ] ,
have been corrected by @xmath40 , @xmath41 , and @xmath42 , respectively . * liller 1 : * although individual stellar abundances were not given in origlia , rich , & castro ( 2002 ) for liller 1 , a mean [ @xmath0/fe ] ratio of @xmath65 was given , with [ fe / h]@xmath66 .
this study was done with infrared spectra and therefore it is not possible to compare their @xmath10 values to those from the lick - texas group .
the solar abundances match our adopted values .
* palomar 6 : * this cluster was observed by lee , carney , & balachandran ( 2004 ) in the infrared and therefore it is not possible to compare their @xmath10 values to those from the lick - texas group .
the solar abundances match our adopted values .
although three stars were observed , only one star had a [ ti / fe ] value , which is @xmath67 and [ fei / h]@xmath68 .
no other elements we used in this paper were given .
* palomar 12 : * the brown , wallerstein , & zucker ( 1997 ) [ mg / fe ] ratio has been corrected by @xmath58 due to @xmath10 differences . for the cohen ( 2004 ) ratios , [ mg / fe ] , [ ca / fe ] , [ la / fe ] , and [ eu / fe ] needed to be corrected by @xmath40 , @xmath41 , @xmath52 , & @xmath42 , respectively , due to differences in @xmath10 and the solar abundances .
the two stars in brown , wallerstein , & zucker were reobserved by cohen so we use only the cohen results for the final abundance ratios . *
ruprecht 106 : * the brown , wallerstein , & zucker ( 1997 ) [ mg / fe ] ratio has been corrected by @xmath58 due to @xmath10 differences .
* terzan 4 and terzan 5 : * the origlia & rich ( 2004 ) ratios for ter 4 and ter 5 are in the infrared and can not be directly compared to the @xmath10 values from the lick - texas group .
the solar abundances match our adopted values .
* terzan 7 : * for the sbordone et al.(2005 ) ratios , the only correction that needs to be made is to the [ mg / fe ] ratio by @xmath42 due to differences in the @xmath10 values .
in figure 3 , we plot the [ mg / fe ] , [ si / fe ] , [ ca / fe ] , and [ ti / fe ] ratios for the gcs in our sample , as well as the field stars from venn et al.(2004 ) . sneden et al.(2004 ) showed that [ ca / fe ] vs. [ fe / h ] in the gcs closely follow the galactic field stars . we confirm this , and note that this trend holds for mg , si , and ti as well , though with more scatter .
the range in the [ mg / fe ] ratios is significantly larger than the other ratios for both the gc and field stars .
the cause of this dispersion may differ between the gc and field stars though . in the gcs
, the [ mg / fe ] variation may be related to stellar evolution effects and internal mixing effects , e.g. , the typical star - to - star variation in a gc is 0.3 dex ( as determined from the maximum range in [ mg / fe ] per gc in our sample ) . the dispersion in the field stars has been shown not to be related to atmospheric parameters or oscillator strengths , but may be due to departures from lte ( carretta , gratton , & sneden 2000 ) .
the agreement between these @xmath0-element ratios in the gcs and field stars is best for metal - poor clusters , [ fe / h ] @xmath69 .
this implies a uniformity in the sne ii yields of @xmath0-elements and iron , and no contributions from sne ia ( nor agb stars ) as expected .
ngc 2419 ( a possible sgr dwarf galaxy gc ; newberg et al.2003 ) has a marginally lower [ ca / fe ] ratio , though [ mg / fe ] , [ si / fe ] , and [ ti / fe ] resemble the halo field stars well .
the bulge cluster ngc 6293 has signficantly lower [ mg / fe ] and marginally lower [ ti / fe ] .
the halo cluster m68 has lower [ ti / fe ] and higher [ si / fe ] ratios as noted by lee , carney , & habgood ( 2004 ) , although its [ mg / fe ] and [ ca / fe ] ratios are consistent with field stars of similar metallicity . at intermediate metallicities ( @xmath70 [ fe / h ] @xmath71 ) , there is a larger dispersion in the [ @xmath0/fe ] ratios .
the [ mg / fe ] ratios vary by 0.5 dex , however we do _ not _ suggest that this dispersion is due to mixing of gcs from a variety of different dwarf galaxy mergers .
the [ ca / fe ] and [ ti / fe ] ratios in the gcs show smaller dispersions ( considering rup 106 as an outlier ) , while the [ si / fe ] ratios have dispersions somewhere in - between .
this suggests that most of the gcs are in good agreement with the field star distribution .
the most outstanding outlier is rup 106 at @xmath72}=-1.4 $ ] .
this gc has been noted for its young age ( buonanno et al.1990 , 1993 ) and lin & richer ( 1992 ) speculated that it may be a recent capture from the magellanic clouds . here
we see it is also outstanding in its low ratios for three of the four @xmath0-elements , typically sub - solar .
both ngc 4833 and ter 4 also stand out with high [ ca / fe ] ratios ; however , the data for ngc 4833 is from only two stars analysed by gratton & ortolani ( 1989 , which is one of the oldest publications included in our database ) , and the data for the highly reddened cluster ter 4 are from infrared echelle spectroscopy which requires a slightly different analysis method and spectral lines from the more commonly used optical spectroscopy and optical lines .
ngc 6362 has high [ si / fe ] and [ ti / fe ] ratios , but they derive from specta of two stars in gartton ( 1987 ) which is an older study and should be reobserved before firm conclusions can be drawn . in figure 4
we compare the [ @xmath0/fe ] ratios of the gcs , the galactic field stars , and the stars in dsph galaxies .
the @xmath0-index is an average of ca and ti , although only ca is available for stars in the sgr dwarf galaxy from bonifacio et al.(2004 ) . in venn
et al.(2004 ) , mg , ca , and ti were averaged for the @xmath0-index for the galactic field stars , however here ( bottom panel ) we have not included mg for field star comparisons since this element can vary significantly from star to star in gcs ( see 2 ) .
the mean [ @xmath0/fe ] in most of the gcs mimics that in the galactic field stars ( top panel ) , as was seen for the individual [ @xmath0/fe ] element abundances in fig .
most of the exceptions to this good agreement have low [ @xmath0/fe ] , in better agreement with the dsph galaxies ( middle panel ) .
we also note that amongst the metal - rich sgr dwarf remnant stars , the half with disk - like [ @xmath0/fe ] ratios are from smecker - hane & mcwilliam ( 2005 ) whereas the half with sub - solar [ @xmath0/fe ] are from bonifacio et al.(2000 ; 2004 ) .
about half ( 4 out of 10 ) of the metal - rich gcs ( [ fe / h ] @xmath73 ) have [ @xmath0/fe ] ratios that are not in good agreement with the galactic field stars .
pal 12 has significantly lower [ @xmath0/fe ] ratios , while ter 5 and ngc 6352 are significantly higher .
ter 5 has higher [ mg / fe ] , [ si / fe ] , and [ ca / fe ] ratios than most of the metal - rich gcs and field stars , however these abundances are from infrared echelle spectroscopy ( like ter 4 , see above ) and may suffer from unrecognized systematic offsets when compared with optical spectroscopic results .
the low [ @xmath0/fe ] ratio for pal 12 have been discussed ( cohen 2004 ) ; combined with its kinematics ( dinescu et al.2000 ) , this provides compelling evidence that this cluster was formed in the sgr dwarf galaxy before merging with the galaxy .
ngc 6352 has a high [ ti / fe ] ratio , but the data come from an older study ( gratton 1987 ) and need updating . just as the [ @xmath0/fe ]
ratios are related to variations in the sne ii / sne ia contributions , the neutron capture element ratios are sensitive to the variations in the massive star imf , variations in sne ii / agb yields , and possibly metallicity effects and star formation efficiencies . in the galaxy ,
interpretation of the metal - poor stellar abundances suggest that s - process contributions do not occur until [ fe / h]@xmath74 and are not significant until near [ fe / h ] @xmath75 ( e.g. , travaglio et al.2004 ) .
the lower star formation efficiencies in dwarf galaxies ( tolstoy et al.2003 ; lanfranchi & matteucci 2004 ; pagel & tautvaisiene 1998 ) mean that metals build up more slowly _ with time _ than in the galaxy ,
and thus we might expect more contributions from metal - poor stars ( e.g. , metal - poor agb stars ) at a given time or metallicity .
coupled with the fact that stellar yields are thought to be metallicity dependent , then the s- and r - process ratios in stars in dwarf galaxies may be different from comparable metallicity stars in the galaxy .
venn et al .
( 2004 ) showed that metal - poor stars in dsph galaxies often have lower [ y / eu ] and higher [ ba / y ] ratios than similar stars in the galaxy , consistent with metallicity dependent yields by travaglio et al.(2004 ) .
thus , the neutron capture element ratios offer another tool to distinguish variations in the dsph abundance ratios compared with the field star population and possibly the gcs . in figure 5 ,
the ratios of [ eu / fe ] , [ y / eu ] , and [ laba / eu ] ( an average of la and ba , which have similar nucleosynthetic histories and s - process contributions in the solar system ) are shown .
the eu abundance primarily samples the pure r - process contributions at all metallicities ( its r - process in the sun is 97% ; burris et al.2000 ) . comparing the [ eu / fe ] ratios ( fig
. 5 , upper panel ) to the [ @xmath0/fe ] ratios ( fig . 4 , upper panel )
, they follow similar trends as the field stars as a function of [ fe / h ] .
this agrees with the idea that @xmath0-elements and eu derive from the same environment ( sne ii s ) . to examine the s - process and r - process contributions separately , we examine y ( its s - process in the sun is 74% ; travaglio et al.2004 ) , la ( its s - process in the sun is 75% ; burris et al.2000 ) and ba ( its s - process in the sun is 81% ; arlandini et al.1999 ) relative to eu .
y versus la and ba are important because they sample different peaks in the neutron magic numbers .
y ( z=39 ) belongs to the first peak that builds through rapid captures around neutron magic number n=50 .
la ( z=57 ) and ba ( z=56 ) belong to the second peak that builds around n=82 . in the most metal - poor stars ( [ fe / h ] @xmath76 ) , the observed variations in [ eu / fe ] in both the galactic field stars and gcs has been interpreted as variations in the r - process contributions , probably due to inhomogeneous mixing due to the location of the forming gc relative to a sne ii event ( possibly even a sne ii event from the formation of the gc itself ) in the early chemical evolution of the galaxy ( mcwilliam 1997 ) .
the [ y / eu ] and [ laba / eu ] ratios are in excellent agreement with the galactic field stars , suggesting similar r - process contributions of these elements to gcs and field stars at low metallicities plus no s - process contributions . around
[ fe / h]@xmath77 , the [ laba / eu ] ratio are noticably above the pure @xmath78-process level ( [ ba / eu]@xmath79 ; burris et al.2000 ) .
johnson & bolte ( 2001 ) argued that this rise may be due to problems with the analysis of the strong ba ii 4554 line and not due to @xmath80-process contributions .
at intermediate metallicities , @xmath81 [ fe / h ] @xmath82 , m80 has a slightly high value of [ eu / fe ] , and rup 106 has a slightly low ratio , though rup 106 is within the range of the halo field stars .
given that the m80 data comes from a large sample of stars from a recent study , this result likely means m80 was contaminated by higher mass sne ii events .
m4 has a slightly elevated [ laba / eu ] ratio .
this is likely to be related to a slightly larger agb s - process contribution than the other gcs , however it is not outside the range of the galactic field stars .
ivans et al.(1999 ) also noted that m4 has slightly higher [ ba / fe ] and [ la / fe ] ratios compared to other galactic gcs , and commented `` that the high ba and la properties of m4 stars is surely a primordial , not an evolutionary , effect . '' among the metal - rich stars and clusters , [ fe / h ] @xmath83 , the two gcs associated with the sgr remnant , pal 12 and ter 7 , stand out with slightly high values of [ eu / fe ] relative to thick and thin disk stars .
the [ y / eu ] ratio in pal 12 is very low though ; this may suggest a lack of s - process contributions from agb stars , however [ laba / eu ] is not lacking , and therefore the [ ba / y ] ratio in pal 12 is remarkably high in figure 6 , where we compare the gcs , field stars , and dsph field stars .
this is consistent with the [ y / eu ] and [ ba / y ] ratios seen in the dsph field stars ( shetrone et al.2003 ; venn et al.2004 ) , and interpreted in terms of more significant contributions from _ metal - poor _ agb stars in the dwarf galaxies .
metallicity dependent yields from agb stars ( travaglio et al . 1999 , 2004 ) suggest that the first - peak r - process elements ( including y , sr , and zr ) are by - passed in favor of second - peak elements ( including ba and la ; and third peak elements as well , but these are not observable in red giant spectra ) in metal - poor stars , due to an excess of neutrons available for capture per seed nucleus .
finally , we noted that in the dsph galaxies , a third of the dsph field stars with @xmath72 } < -1.8 $ ] show a clear offset in the [ ba / y ] ratios relative to the galactic field stars ( venn et al.2004 ) .
the metal - poor galactic gcs do not clearly show the same neutron capture element ratios as in the low mass dsph field stars ( see middle panel of fig .
it is interesting that comparing the halo and thick disk clusters , the halo gcs are offset more toward higher [ ba / y ] similar to dsph field stars .
given that gcs are thought to be the first objects to form in a galaxy , it is very important to determine their true ages .
however , this has been a very challenging task ( see review by sarajedini , chaboyer , & demarque 1997 ) .
while the relative ages between gcs is an easier property to measure , different methods have yielded different absolute ages for the gcs ( e.g. , salaris & weiss 2002 ) . for our purposes , we do not need the absolute ages of the gcs in our sample to be the true ages of the gcs .
given that the wmap results ( spergel et al . 2003 ) find the age of the universe as @xmath84 gyr , most of the absolute gc ages in the literature are too old .
here , we only investigate whether the gcs show any interesting properties or trends in relative age . in figure 7 , we show age as a function of [ fe / h ] , where the ages have been determined by vandenberg ( 2000 ) and salaris & weiss ( 2002 ) .
although rosenberg et al.(1999 ) have a sample of relative gc ages , we chose to use the more recent results of salaris & weiss ( 2002 ) which also includes a larger number of gcs .
vandenberg used stellar evolution models which take into account the @xmath0 element abundance to measure the @xmath85 magnitude difference between the horizontal branch and the main - sequence turn - off ( @xmath86 ) for the gcs .
the [ @xmath0/fe ] ratio was assumed to either be 0.3 or 0.0 .
we contrast these `` @xmath0 enhanced '' isochrone ages with those from salaris & weiss , who divide the gcs into four groups according to their metallicities then assign ages for a reference cluster within each group using @xmath87 .
the ages for the remaining gcs in each group are then found using isochrones ( assuming [ @xmath0/fe ] @xmath88 ) to determine the color difference between the main sequence turn - off and the red giant branch .
of the 45 gcs in our database , 17 and 32 have ages from the vandenberg and salaris & weiss studies , respectively .
we have not searched the literature to find absolute ages for the remaining gcs in our sample because of the difficulties in normalizing those ages to these two studies . comparing the age vs. metallicity diagrams in fig . 7
, both studies show a trend of decreasing age with increasing metallicity .
the most metal - poor thick disk gcs have ages similar to those found in the halo .
the thin disk clusters show a younger age consistent with the idea that the thin disk formed after the halo .
both methods yield the same ( old ) age for the more metal - poor gcs .
the metal - rich gcs associated with the sgr remnant , pal 12 and ter 7 , are significantly younger than the galactic gcs , regardless of the differences in [ @xmath0/fe ] adopted between the two studies ; i.e. , vandenberg ( 2000 ) adopted [ @xmath0/fe ] @xmath89 and salaris & weiss ( 2002 ) used [ @xmath0/fe ] @xmath88 .
we note that m68 , which is amongst the most metal - poor gcs , appears to be slightly younger ( 0.5 to 1 gyr ) than the others .
whether or not this difference is significant will be discussed in 7.6 .
thus , both methods yield broadly similar results for the relative ages of the gcs , however rup 106 is a good example of the difference between the two methods ; salaris & weiss find that it is younger than the majority of galactic gcs , whereas vandenberg finds that it is not much different in age than the other intermediate metallicity gcs .
an updated study of relative gc ages by de angeli et al.(2005 ) has also found there there is a trend of decreasing age with increasing metallicity .
the metal - poor gcs ( [ fe / h]@xmath90 ) are shown to be old and coeval with the exceptions of m68 and m15 .
de angeli et al.found that intermediate metallicity gcs are typically younger than the metal - poor clusters , but there is a significant spread in the ages for these clusters . while there does appear to be a metallicity - age trend , there does not appear to be a @xmath0-age trend as first noted by carney ( 1996 ) . if we exclude the extragalactic gcs rup 106 , pal 12 , and ter 7 from fig .
4 , there is very little scatter among the gcs in [ @xmath0/fe ] .
therefore , we can conclude that there is no trend of [ @xmath0/fe ] with age from the existing data .
given the relatively small numbers of metal - rich gcs with abundance ratios , more observations of these types of clusters should be done before firm conclusions can be drawn about any possible @xmath0-age relation .
ngc 362 is shown to have a younger age than typical halo gc by salaris & weiss ( 2002 ) .
this reaffirms the previous finding that ngc 362 is younger than its second - parameter pair cluster ngc 288 ( catelan et al.2001 ) .
when the horizontal branch of a gc is redder or bluer than expected for its metallicity , a `` second - parameter '' is thought to be affecting its morphology . while it is now thought that there are many different parameters that may affect the horizontal branch morphology , age is considered the most common parameter ( e.g. , fusi pecci & bellazzini 1997 ) , as appears to be the case for ngc 362 .
we make note of this because even though it has a younger age , we see no peculiarities in the abundance ratios of this gc .
we want to clarify that although we discuss the difference in ages between the clusters , we do not attempt to break the gcs into `` old '' or `` young '' halo groups . zinn ( 1993 , 1996 ) looked at the properties of globular clusters and separated them into `` young '' and `` old '' halo and thick disk components based on the metallicity of the cluster and its horizontal branch morphology .
mackey & gilmore ( 2004 ) have updated zinn s study to include the most recent measurements for galactic gcs and their own measurements of extragalactic gcs , separating them into bulge / disk , `` young '' halo , and `` old '' halo components . in column ( 9 ) of table 3 we list the mackey & gilmore classifications for the gcs in our sample .
mackey & gilmore also used the salaris & weiss ( 2002 ) ages to discuss which gcs are truly young or old .
although some of their `` young '' gcs have similar ages to the oldest clusters and some `` old '' gcs have relatively young ages , mackey & gilmore classified the gcs solely on their [ fe / h ] values and horizontal branch morphologies .
we chose not to use this classification scheme because there may be other parameters beyond metallicity and age affecting a cluster s horizontal branch morphology and as a result its classification in zinn s scheme .
several gcs have been associated kinematically with the sgr dwarf galaxy merger remnant , including ter 7 , ter 8 , m54 , and arp 2 ( ibata , gilmore , & irwin 1994 ) , and pal 12 ( irwin 1999 ; dinescu et al.2000 ) .
ngc 2419 was included as a possible member due to its location near an overdensity of stars associated with the sgr tidal tails ( newberg et al.2003 ) .
more recently , ngc 4147 has also been associated with the sgr remnant kinematically ( bellazzini et al.2003 ) , however this has been questioned by martinez - delgado et al.(2004 ) based on an update of the proper motion studies by wang et al.(2000 ) .
m54 and ngc 2419 show no distinctive abundance ratios from the galactic field stars , with the possible exception that ngc 2419 has a slightly lower [ ca / fe ] than seen in the field . however , that is based on only one star ( shetrone , ct , & sargent 2001 ) .
smecker - hane & mcwilliam ( 2005 ) found that their three metal - poor sgr field stars have similar [ @xmath0/fe ] abundance ratios as stars in the galactic halo , although bonifacio et al.(2004 ) suggest that their metal - poor stars may belong to m54 .
unfortunately , the m54 analysis is based on only five stars studied by brown , wallerstein , & gonzalez ( 1999 ) and at least one of the smecker - hane & mcwilliam ( 2005 ) sgr stars does have the same metallicity and [ @xmath0/fe ] ratio as m54 .
furthermore , layden & sarajedini ( 1995 ) suggest that m54 may have a significant metallicity spread due to the dispersion in the red giant branch .
the other sgr clusters , ter 7 and pal 12 , do show distinctive kinematic and chemical properties .
their [ @xmath0/fe ] ratios are lower than the typical metal - rich galactic gc , and more similar to those seen in the dsph field stars .
pal 12 also has low [ y / eu ] and high [ ba / y ] ratios like stars in the dsph galaxies .
both clusters have noticably younger ages than other galactic gc . if these clusters were not already associated with the sgr remnant , their properties ( abundance ratios , kinematics , and ages ) would set them apart from the other galactic gcs as other studies have found ( e.g. , tautvaisien et al.2004 ; cohen 2004 ; sbordone et al.2005 ) .
it has also been shown by these studies that the sgr gcs follow the abundance trends seen in the sgr field stars .
martin et al.(2004 ) discovery of an overdensity of stars suggested to be the remnant of a dwarf galaxy in cma is also associated with four gcs .
these are ngc 1851 , ngc 1904 ( m79 ) , ngc 2298 , and ngc 2808 .
however , we note that pearrubia et al.(2005 ) used the proper motions of ngc 1851 , ngc 1904 , and ngc 2298 to show that they are not associated with the cma dwarf . of these gcs , only m79 and ngc 2298 have abundance ratios in the literature .
neither of these gcs stands out in a significant way in fig . 3 and fig . 4 , although , as seen in table 1 , the data for these two clusters is sparce and from older studies .
m79 does have a slightly lower [ si / fe ] ratio , but the remaining @xmath0-element ratios are consistent with field stars of similar metallicity .
however , we note that pearrubia et al.(2005 ) used the proper motions of ngc 1851 , ngc 1904 , and ngc 2298 to show that they are not associated with the cma dwarf .
martin et al.also calculated the orbit of the cma galaxy for a prograde or a retrograde orbit . with the knowledge that it is in a prograde orbit
, we can also examine the clusters near the cma dwarf galaxy in phase space ( see table 1 , martin et al.2004 ) .
of the gcs associated with a prograde orbit , only ngc 6205 ( m13 ) , ngc 7078 ( m15 ) , ngc 6341 ( m92 ) , ngc 4590 ( m68 ) , and rup 106 are included in our survey . of these gcs , only m68 and rup 106 stand out .
rup 106 is discussed in 7.3 , since it is thought to actually have been captured from the magellanic clouds .
the properties of m68 and its possible membership to the cma dwarf are discussed further in 7.6 .
rup 106 has long been noted as an interesting cluster given that the metallicity from its red giant branch does not match the metallicity from the stellar abundances .
for example , brown , wallerstein , & zucker ( 1997 ) found [ fe / h][email protected] ( for our study we use [ fe / h]@xmath92 , which is an average of [ fei / h ] and [ feii / h ] ) from spectra of two stars , while buonanno et al.(1990 ) found @xmath72}=-1.9\pm0.2 $ ] from the red giant branch .
buonanno et al . also found that rup 106 has a younger age than the typical halo gc ( see fig .
lin & richer ( 1992 ) speculated that rup 106 was acquired by our galaxy from the magellanic clouds based on its position and assorted other odd properties for an outer halo cluster ( young age , high number of blue stragglers , and horizontal branch morphology ) rup 106 has a radial velocity of @xmath93 kms@xmath15(harris 1996 ) , but no other kinematics exist .
as discussed in 3 , we examine potential orbits for rup 106 from reasonable limits on its ( undetected ) proper motion to find that it has either a typical halo orbit or a plunging orbit .
the @xmath0-element ratios for rup 106 are significantly lower than the galactic field stars of similar metallicity , which is similar to the stars in the dsph galaxies ( see figs . 3 and 4 ) .
we also notice that [ eu / fe ] is lower than in most of the gcs and galactic field stars ( see fig . 5 ) , also suggesting that rup 106 has had less sne ii r - process contributions .
if rup 106 had its origin in the magellanic clouds , then we can compare its abundance ratios with other lmc gcs .
hill et al.(2000 ) and hill ( 2004 ) found that the lmc gcs have [ ca / fe ] and [ ti / fe ] ratios below what is seen in the galactic halo stars at all metallicities .
rup 106 falls near these lmc gcs abundances , but typically with abundance ratios about 0.1 dex lower .
johnson et al.(2004 ) also determined the abundance ratios for three old lmc gcs .
they find the [ ca / fe ] ratio for hodge 11 to be similar to the galactic field stars , which is much higher than the other two lmc gcs in their sample and those in the hill study .
the [ ti / fe ] ratios are similar in both studies . while the abundance ratios in rup 106 do not match precisely with the analysed gcs , the tendancy towards lower [ @xmath0/fe ] is more consistent with the lmc clusters and dsph field stars .
as noted in 3 , we classified anything within @xmath94 kpc to be within the bulge , unless the kinematics associate the gc with a different galactic component .
for example , dinescu et al.(2003 ) noted that ngc 6528 and ngc 6553 are confined to the bulge region , and suggest that ngc 6528 is actually a `` genuine milky way bar cluster . ''
although ngc 6553 is within our bulge limit , its kinematics show it has a circular orbit and therefore it is classified as a thin disk cluster . from our model ,
we classify nine clusters as belonging to the bulge .
however , only one of these clusters ( ngc 6528 ) has known phase space velocities .
the bulge stars have been shown to be mostly metal - rich ( [ fe / h]@xmath95 ; mcwilliam & rich 2004 ) . using this limit , six of the nine gcs fall in this range with ter 4 being right at this limit .
the three clusters ngc 6287 , ngc 6293 , and ngc 6541 are more metal - poor than this ( [ fe / h]@xmath96 ) and may not be truly associated with the bulge .
the bulge stars are known to show elevated [ @xmath0/fe ] ratios compared to the metal - rich thin disk stars ( mcwilliam & rich 2004 ; fulbright , rich , & mcwilliam 2004 ) . in general , the bulge gcs show a similar trend , but not as clear as seen in the bulge field stars .
of the @xmath0 elements , [ ca / fe ] shows the clearest difference between the bulge gcs and the thin disk stars .
this is somewhat contrary to the bulge field stars .
they have been found to be more enhanced compared to metal - rich thin disk stars in [ mg / fe ] and [ si / fe ] than [ ca / fe ] and [ ti / fe ] ( mcwilliam & rich 2004 ) .
mcwilliam & rich suggested that this implied sne ii played a greater role in the bulge than in the thin disk .
as noted in 3 , it is interesting to see the majority of thick disk gcs clustered toward the metal - poor range of the thick disk field stars .
of the ten thick disk clusters , seven have @xmath7 velocities , which makes their classifications more certain since it is based on phase space sampling ( not just physical location ) .
as discussed in 6 , the metal - poor thick disk gcs have similar ages to halo clusters , while the two metal - rich ( [ fe / h]@xmath97 ) gcs show a younger age ( see fig .
the implication is that when discussing the formation of the thick disk , it must be taken into account that the metal - poor thick disk gcs formed early on and with similar abundance ratios as the metal - poor thick disk field stars .
thus , the metal - poor thick disk represents the transitional phase between the halo and thick disk formation as the galaxy collapsed according to the eggen , lynden - bell & sandage ( 1962 ) model of galaxy formation . examining the m31 gcs , morrison et al.(2004 )
found that 40% have disk kinematics and are metal - poor ( with no age estimates ) , which is similar to our findings .
contrary to this , a recent study by mould ( 2005 ) found that the thick disks in four edge - on galaxies are old and relatively _ metal - rich_. this suggests that thick disks form relatively early on and from gas - rich mergers ( dalcanton & bernstein 2002 ; brook et al.2005 ) .
further studies of the galactic thick disk field stars and gcs need to be done in order to resolve the formation of the thick disk .
_ ngc 4590 ( m68 ) : _
m68 stands out from the majority of the galactic gcs , primarily due to its high galactocentric rotational velocity ( @xmath98 kms@xmath15 ) . as noted by lee , carney , & habgood ( 2004 )
, m68 also has some interesting abundance ratios . as seen in fig . 3 , the [ ti / fe ] ratio is lower than other gcs and field stars of similar metallicites . on the other hand , the [ si / fe ]
ratio in m68 is unusually high .
m68 is also interesting because its horizontal branch is notably redder than expected for a gc of its metallicity ( harris 1996 ) .
this is an example of the second - parameter effect as discussed in 6 . as shown in fig .
7 , m68 is thought to have a slightly younger age ( @xmath99 gyr , vandenberg 2000 ; see also de angeli et al.2005 ) compared to the typical halo gc ( @xmath100 gyr ) .
this age difference is not large enough to seriously affect the horizontal branch morpology though according to theoretical models ( e.g. , catelan & de freitas pacheco 1993 ; lee , lee , & gibson 2002 ) , and argues that age alone can not be the sole parameter affecting the horizontal branch in m68 .
given the slightly younger age of m68 , its high rotational velocity , and its unique [ si / ti ] ratio , we argue that it is a candidate extragalactic gc .
martin et al.(2004 ) included m68 as a possible cma dwarf cluster .
however , the membership of certain gcs with the cma dwarf has been called into question by pearrubia et al.(2005 ) .
using fig .
11 of pearrubia et al.and the kinematics for m68 ( dinescu 2004 , private communications ) , it seems unlikely that m68 belongs to the cma dwarf . _
@xmath1 centauri : _ we briefly comment on @xmath1 centauri ( @xmath1 cen ) .
this cluster has long been known to have a spread in its metallicity ( dickens & woolley 1967 ) .
because of this spread , @xmath1 cen has been suggested to have once been the nucleus of a now accreted dwarf galaxy ( e.g. , rey et al.2004 ) .
using the kinematic data from dinescu et al.(1999b , and private communications for updates ) , @xmath1 cen is a typical halo gc ( @xmath101 ) .
pancino et al.(2002 ) and origlia et al .
( 2003 ) obtained spectra for the different populations in @xmath1 cen including the recently discovered high metallicity population ( pancino et al.2000 ) .
both studies have found abundance ratios for [ o / fe ] , [ si / fe ] , [ mg / fe ] , and [ ca / fe ] from @xmath72}\sim-1.6 $ ] to @xmath102 that are consistent with the galactic field stars , i.e. , halo field star ratios , with a bend towards lower abundance ratios at higher metallicities ( @xmath72}>-1.0 $ ] ) .
in addition , pancino ( 2004 ) presented the preliminary [ ca / fe ] results of about 700 red giants in @xmath1 cen which illustrated the turn " toward lower [ ca / fe ] for the metal - rich stars .
this shows that the star formation in @xmath1 cen took place over an extended period of time and that it was enriched in a similar manner as seen in galaxies .
out of the 45 gcs with high - resolution abundances , four gcs are excellent candidates for having an extragalactic origin ( rup 106 , ter 7 , pal 12 , and m68 ) .
this is approximately one - eleventh of our sample .
two of the four candidate extragalactic gcs belong to the sgr remnant , ter 7 and pal 12 .
rup 106 has been suggested as being captured from the magellanic clouds or as part of the cma remnant ( prograde orbit ) .
we have included m68 as a possible extragalactic cluster due to its unusual kinematics , its younger age , and its unique [ si / ti ] ratio .
the detection of extragalactic metal - rich gcs seems relatively easy ( e.g. , pal 12 and ter 7 ) because the [ @xmath0/fe ] ratios are clearly distinctive .
the influence of type ia sn occuring ( presumably ) at a lower metallicity in the chemical evolution of a dwarf galaxy causes the turn " towards lower [ @xmath0/fe ] at lower metallicities than in the galactic field stars leading to clear differences between their stellar populations .
mackey & gilmore ( 2004 ) analyzed the physical properties of gcs in the milky way and its dwarf companions to allow a comparison between their gc systems .
they divided up the galactic gcs according to their metallicities and horizontal branch morphologies into bulge / disk , `` young '' halo , and `` old '' halo components , which we list in column ( 9 ) of table 3 .
overall , our classifications match up well with those from mackey & gilmore with the exception of a few cases where the kinematics tend to place a cluster in the thick disk while they classify it as belonging to the halo .
mackey & gilmore concluded that all of the `` young '' halo clusters and a small fraction ( 15 - 17% ) of `` old '' halo clusters have been captured from merged dwarf galaxies , which does not include the sgr gcs .
this is much higher than what we would estimate from the abundance ratios , even including the sgr gcs . in figure 8
, we compare the [ @xmath0/fe ] and [ ba / y ] ratios for the `` old '' and `` young '' halo gcs as classified by mackey & gilmore .
although the [ @xmath0/fe ] ratios in `` young '' halo gcs are marginally lower than the `` old '' halo gcs , the offset is not significant , with the exception of rup 106 which is thought to be a captured gc . as noted above , the lack of noticable variations between the galactic gcs and field stars in our sample argues against a large number of accretions by dwarf galaxies with gcs . because of the noticable offsets between the dsph and galactic field stars at metallicities of [ fe / h]@xmath103 and higher , for example in the [ @xmath0/fe ] ratios ( see fig . 4 ) , it should be relatively easy to pick out metal - rich extragalactic gcs .
however , what about metal - poor extragalactic gcs ?
can we rule out early accretions by dwarf galaxies before major chemical evolution has occured ? are there any abundance indicators , such as [ ba / y ] ( venn et al.2004 ) , that show differences for the metal - poor stars in dwarf galaxies and the galactic field stars ?
the more metal - poor sgr gc m54 does not stand out in any way from the galactic field stars , although it may not be the ideal example given that it may be the nucleus of the sgr dwarf galaxy .
none of the candidate clusters associated with the cma dwarf galaxy stand out in a significant way from the rest of the galactic field according to their abundance ratios either .
none of the metal - poor gcs in our compilation stand out in [ ba / y ] in the same way that the dsph stars do .
however , a comparison between the thick disk and halo gcs at intermediate metallicities ( [ fe / h]@xmath104 ) shows that the halo clusters are more like the dsphs stars than the thick disk clusters in that the halo gcs are shifted toward higher [ ba / y ] values than the thick disk ones . one way to resolve
this problem is to examine the abundances in known extragalactic clusters such as those in the lmc , fornax , and sgr .
many of these clusters have been shown to have ages similar to the galactic gcs ( e.g. , buonanno et al.1998 , 1999 ) .
therefore , it will be key to determine what similarities and , more importantly , what differences there are between metal - poor clusters and galactic field stars of similar metallicities .
for example , as noted in 7.3 , hill ( 2004 ) found that the lmc gcs tend to have [ ca / fe ] and [ ti / fe ] ratios that are below those of the galactic halo stars even for the metal - poor ones .
this is similar to what is seen in rup 106 . for more metal - poor gcs ,
the picture is less clear .
before firm conclusions can be drawn , more work will need to be done on the metal - poor dwarf galaxy stars and their gcs to help determine what kind of impact early mergers may have had in the formation of our galaxy .
lee & carney ( 2002 ) discussed a trend of decreasing [ si / ti ] ratios with increasing galactocentric radius ( @xmath105 ) for the `` old '' halo gcs as classified by zinn ( 1993 ) .
they argued that this gradient could be explained by differences in sne ii contributions by stellar progenitors with different masses according to models by woosley & weaver ( 1995 ) .
the models suggest that more massive stars ( @xmath106 ) yield higher amounts of silicon relative to calcium and titanium .
therefore , lee & carney reasoned that the central regions of the galaxy , with a deeper gravitational potential , retained a higher amount of the more massive sne ii ejecta than the outer regions of the galaxy .
given our larger sample of gcs , we plot [ si / ti ] against @xmath107 in figure 9 , where the larger symbols are for gcs plotted at their apogalactic distance ( dinescu 2004 , private communications ) and the smaller ones are those at their galactocentric distance .
to be consistent with the lee & carney analysis , we have plotted the clusters according to their classification in mackey & gilmore ( 2004 ) .
mean errors are adopted for [ si / ti ] and @xmath107 ( the true erros for some of the individual gcs may be smaller than these ) .
excluding m79 whose ratios come from an older study ( gratton & ortolani 1989 ) , there is no apparent trend of decreasing [ si / ti ] with increasing @xmath108 . in investigating [ si / ti ]
, we noticed that all the gcs with [ si / ti]@xmath109 have [ fe / h]@xmath110 . to examine this further ,
[ si / ti ] , [ si / ca ] , and [ ti / ca ] are plotted as a function of [ fe / h ] in figure 10 . for the stars with [ fe / h]@xmath111 ,
[ si / ti ] and [ si / ca ] are scattered over a wide range of values , but [ ti / ca ] remains relatively flat across all metallicities .
this suggests spatial inhomogeneities where the oldest stars were likely polluted by individual sne ii events from progenitors of varying progenitor stellar masses ( a majority with @xmath106 ; woosley & weaver 1995 ) .
those forming from the ejecta of more massive sne ii events are predicted to have higher amounts of silicon than calcium or titanium , while the calcium and titanium yields are similar over all stellar masses .
shetrone ( 2004 ) found a similar trend in the dsph stars , where the [ o / fe ] and [ mg / fe ] ratios tended to be higher than [ ca / fe ] and [ ti / fe ] for the more metal - poor dsph stars ( [ fe / h]@xmath90 ) .
between [ fe / h]@xmath112 to @xmath104 , there appears to be a trend of decreasing [ si / ti ] and [ si / ca ] .
this is consistent with a decrease in the ejecta of the most massive stars .
all of the different populations , galactic field stars , gcs , and dsph stars , show this same effect . above [ fe / h]@xmath104 ,
the relations of [ si / ti ] and [ si / ca ] are relatively flat for galactic field stars and gcs , which suggests that the galactic interstellar medium has become fully mixed in ejecta from all sne ii masses .
the two gcs with clearly lower [ si / ti ] and [ si / ca ] ratios come from an older study ( ngc 4833 and m79 ; gratton & ortolani 1989 ) and should be reinvestigated .
it is interesting to see that as many as one third of the stars in the dsph galaxies fall below the majority of the galactic field stars in both [ si / ti ] and [ si / ca ] .
given that these same stars show no differences to the galactic field stars in [ ti / ca ] , something must be reducing the silicon abundances .
that the dsph stars in the range @xmath113}<-1.6 $ ] are similar to the galactic field star and the gc abundance ratios argues against the possibility of a truncated initial mass function in the dsph galaxies ( discussed by tolstoy et al.2003 ) because the initially higher silicon yields presumably came from more massive stars as discussed above .
however , a preferential outflow of ejecta from the most massive stars could lead to a reduced amount of silicon in the dsph interstellar medium .
we have produced a standardized dataset of select abundance ratios for the galactic gcs .
we also divided up the gcs into their respective galactic components to better compare them to the galactic field stars .
in general , the galactic gcs follow the trends seen in the field stars for both the @xmath0 and neutron capture elements .
therefore , the similarity in the abundance ratios over a wide range of metallicities seems to indicate that the chemical evolution of the galaxy has been similar throughout .
we also find that gcs assigned to the thick disk tend to be found at the metal - poor end of the thick disk field star distribution .
given the similar ages of these gcs with the metal - poor halo gcs , this seems to indicate that the metal - poor population of the thick disk formed in a manner similar to the halo , although some differences were seen in [ ba / y ] . on the other hand ,
the bulge is a distinct component ; mcwilliam & rich ( 2004 ) and fulbright , rich , & mcwilliam ( 2004 ) found that the bulge stars tend to have enhanced @xmath0 elements compared to field stars and we confirm this tendancy for the bulge gcs as well . from our sample ,
we have found four extragalactic gc candidates : ter 7 , pal 12 , rup 106 , and m68 .
ter 7 and pal 12 are both known to be associated with the sgr dwarf galaxy , while rup 106 is thought to be associated with either the magellanic clouds or the cma dwarf galaxy .
m68 is interesting because of its high galactocentric rotational velocity , its slightly younger age , and its low [ ti / fe ] and high [ si / fe ] ratios .
there is an interesting trend of decreasing [ si / ti ] with increasing [ fe / h ] for galactic field stars and gcs in the approximate range of @xmath113}<-1.6 $ ] .
this implies that the most massive stars ( @xmath114 ; woosley & weaver 1995 ) played a larger role during the initial chemical evolution of the galaxy that lead to higher yields of silicon .
the dsph stars also follow this trend , yet the [ si / ti ] ratios in about one - third of the sample appear lower than in the galactic field stars .
this implies that silicon has been lost in dsph galaxies , presumably by a preferential outflow of the ejecta from massive sne ii events
. thank you to the referee b. carney who gave very valuable comments which aided in clarifying our discussions .
b.j.p . and k.a.v .
would like to thank the national science foundation ( nsf ) for support through a career award , ast 99 - 84073 .
thanks to d. dinescu for sharing her updated kinematic information for the gcs and for discussions concerning those kinematics . thank you to j. cohen , j. melendez , j .- w .
lee , m. shetrone , b. castilho , g. wallerstein , & g. tautvaisien for sharing their solar abundances and @xmath10 values .
martinez - delgado , d. , dinescu , d. i. , zinn , r. , tuntsoff , a. , ct , p. , & boyarchuck , a. 2004 , in satellites & tidal streams , eds .
f. prada , d. martinez - delgado , & t. j. mahohey ( san francisco : asp ) , in press , ( astro - ph/0409118 ) mcwilliam , a. , & rich , r. m. 2004 , in carnegie observatories astrophysics series , vol . 4 : origin and evolution of the elements , ed .
a. mcwilliam and m. rauch ( pasadena : carnegie observatories , http://www.ociw.edu/ociw/symposia/series/symposium4/proceedings.html ) sneden , c. 2000 , in proc .
liege int .
colloq . , the galactic halo : from globular clusters to field stars , ed .
a. noels , p. magain , d. caro , e. jehin , g. parmentier , & a. thoul ( lige : inst .
astrophys . gophys . ) , 159 lrl 104 ( 47 tuc ) & 3 & gratton , quarta , ortolani ( 1986 ) + & 4 & brown & wallerstein ( 1992 ) + & 1 & norris & da costa ( 1995 ) + & 12 & carretta et al.(2004 ) + & 12 & james et al.(2004b ) + 288 & 2 & gratton ( 1987 ) + & 13 & shetrone & keane ( 2000 ) + 362 & 1 & gratton ( 1987 ) + & 12 & shetrone & keane ( 2000 ) + 1904 ( m79 ) & 2 & gratton & ortolani ( 1989 ) + 2298 & 3 & mcwilliam , geisler , & rich ( 1992 ) + 2419 & 1 & shetrone , ct , & sargent ( 2001 ) + 3201 & 3 & gratton & ortolani ( 1989 ) + & 18 & gonzalez & wallerstein ( 1998 ) + 4590 ( m68 ) & 2 & gratton & ortolani ( 1989 ) + & 1 & shetrone et al.(2003 ) + & 7 & lee , carney , & habgood ( 2004 ) + 4833 & 2 & gratton & ortolani ( 1989 ) + 5272 ( m3 ) & 7 & kraft et al.(1993 ) + & 3 & kraft et al.(1995 ) + & 3 & shetrone , ct , & sargent ( 2001 ) + & 28 & sneden et al.(2004b ) + & 13 & cohen & melendez ( 2004a ) + 5466 & 1 & mccarthy & nemec ( 1997 ) + 5897 & 2 & gratton ( 1987 ) + 5904 ( m5 ) & 3 & gratton , quarta , & ortolani ( 1986 ) + & 13 & sneden et al.(1992 ) + & 28 & ivans et al.(2001 ) + & 23 & ramrez & cohen ( 2003 ) + 6093 ( m80 ) & 10 & cavallo , suntzeff , & pilachowski ( 2004 ) + 6121 ( m4 ) & 3 & gratton , quarta , & ortolani ( 1986 ) + & 3 & brown & wallerstein ( 1992 ) + & 24 & ivans et al.(1999 ) + 6205 ( m13 ) & 35 & kraft et al.(1997 ) ; sneden et al.(2004b ) + & 13 & cohen & melendez ( 2004a ) + 6218 ( m12 ) & 1 & mishenina , panchuk , & samus ( 2003 ) + 6254 ( m10 ) & 2 & gratton & ortolani ( 1989 ) + & 14 & kraft et al.(1995 ) + & 2 & mishenina , panchuk , & samus ( 2003 ) + 6287 & 3 & lee & carney ( 2002 ) + 6293 & 2 & lee & carney ( 2002 ) + 6341 ( m92 ) & 6 & shetrone ( 1996 ) + & 30 & sneden , pilachowski , & kraft ( 2000a ) + & 4 & shetrone , ct , & sargent(2001 ) + 6342 & 4 & origlia , valenti , & rich ( 2005 ) + 6352 & 3 & gratton ( 1987 ) + 6362 & 2 & gratton ( 1987 ) + 6397 & 3 & gratton & ortolani ( 1989 ) + & 2 & norris & da costa ( 1995 ) + & 16 & castilho et al.(2000 ) + & 8 & james et al.(2004b ) + 6528 & 4 & carretta et al.(2001 ) + & 3 & zocalli et al.(2004 ) + & 4 & origlia , valenti , & rich ( 2005 ) + 6541 & 2 & lee & carney ( 2002 ) + 6553 & 2 & barbuy et al.(1999 ) + & 5 & cohen et al.(1999 ) + 6656 ( m22 ) & 3 & gratton & ortolani ( 1989 ) + & 7 & brown & wallerstein ( 1992 ) + 6715 ( m54 ) & 5 & brown , wallerstein , & gonzalez ( 1999 ) + 6752 & 3 & gratton , quarta , & ortolani ( 1987 ) + & 6 & norris & da costa ( 1995 ) + & 18 & gratton et al.(2001 ) + & 20 & yong et al.(2003 ) + & 11 & cavallo , suntzeff , & pilachowski ( 2004 ) + & 18 & james et al.(2004a ) + 6809 ( m55 ) & 2 & shetrone et al.(2003 ) + 6838 ( m71 ) & 3 & gratton , quarta , & ortolani ( 1986 ) + & 8 & sneden et al.(1994 ) + & 8 & ramrez & cohen ( 2002 ) + & 2 & mishenina , panchuk , & samus ( 2003 ) + & 2 & lee , carney , & balachandran ( 2004 ) + 7006 & 6 & kraft et al.(1998 ) + 7078 ( m15 ) & 18 & sneden et al.(1997 ) + & 3 & sneden et al.(2000 ) + & 31 & sneden , pilachowski , & kraft ( 2000 ) + 7099 ( m30 ) & 1 & shetrone et al.(2003 ) + 7492 & 4 & cohen & melendez ( 2004b ) + ( lil 1 ) & 2 & origlia , rich , & castro ( 2002 ) + ( pal 5 ) & 4 & smith , sneden , & kraft ( 2002 ) + ( pal 6 ) & 3 & lee , carney , & balachandran ( 2004 ) + ( pal 12 ) & 2 & brown , wallerstein , & zucker ( 1997 ) + & 4 & cohen ( 2004 ) + ( rup 106 ) & 2 & brown , wallerstein , & zucker ( 1997 ) + ( ter 4 ) & 4 & origlia & rich ( 2004 ) + ( ter 5 ) & 6 & origlia & rich ( 2004 ) + ( ter 7 ) & 3 & tautvaisien et al.(2004 ) + & 5 & sbordone et al.(2005 ) + lcccccccccccc 47 tucanae & 6.63 & -2.58 & -3.18 & 19.4 & 10.2 & 177.1 & 7.8 & 40.3 & 3.6 & tk & bd + ngc 288 & 8.58 & 0.04 & -8.80 & 18.7 & 9.5 & -50.5 & 20.1 & 52.8 & 0.4 & h & oh + ngc 362 & 5.43 & -5.01 & -6.14 & 81.4 & 28.4 & -20.2 & 28.5 & -79.0 & 22.1 & h & yh + m79 & 16.14 & -8.25 & -6.32 & 93.8 & 31.2 & 81.2 & 30.4 & 12.4 & 35.8 & h & oh + ngc 2298 & 12.74 & -9.37 & -2.95 & -61.7 & 40.6 & -49.5 & 34.8 & 116.5 & 51.0 & h & oh + ngc 2419 & 84.66 & -0.49 & 35.90 & & & & & & & h &
oh + ngc 3201 & 7.88 & -4.90 & 0.75 & & & & & & & tk & yh + m68 & 4.42 & -7.17 & 6.00 & -115.3 & 27.7 & 303.5 & 34.8 & 13.3 & 25.0 & h & yh + ngc 4833 & 4.94 & -5.36 & -0.91 & & & & & & & tk & oh + m3 & 6.99 & 1.37 & 10.20 & 50.9 & 14.2 & 18.6 & 14.6 & -108.0 & 3.5 & h & yh + ngc 5466 & 5.17 & 3.01 & 15.25 & 256.0 & 83.5 & -51.9 & 83.4 & 199.4 & 24.4 & h & yh + ngc 5897 & -7.66 & 0.24 & 16.65 & 49.7 & 50.8 & 82.7 & 44.7 & 119.5 & 44.8 & h & oh + m5 & 3.38 & 0.35 & 5.47 & -324.1 & 34.8 & 113.4 & 30.5 & -214.1 & 31.7 & h & oh + m80 & -0.85 & -1.20 & 3.33 & 92.5 & 31.5 & 86.6 & 35.0 & -99.5 & 29.9 & h & oh + m4 & 6.41 & -0.33 & 0.61 & -51.2 & 3.0 & -19.7 & 21.9 & -16.0 & 5.8 & h & oh + m13 & 5.50 & 4.99 & 5.04 & 293.6 & 35.5 & -89.9 & 35.4 & -111.8 & 20.3 & h & oh + m12 & 4.27 & 1.19 & 2.17 & -28.2 & 12.6 & 132.7 & 19.4 & -132.9 & 17.1 & tk & oh + m10 & 4.59 & 1.06 & 1.72 & -65.7 & 8.8 & 126.5 & 19.3 & 88.6 & 15.5 & tk &
oh + ngc 6287 & -0.63 & 0.02 & 1.78 & & & & & & & b & oh + ngc 6293 & -0.21 & -0.36 & 1.20 & & & & & & & b & oh + m92 & 6.02 & 6.25 & 4.69 & 82.4 & 14.4 & -16.6 & 14.4 & 79.4 & 20.9 & h &
oh + ngc 6342 & 0.05 & 0.72 & 1.45 & & & & & & & b & bd + ngc 6352 & 3.14 & -1.80 & -0.71 & & & & & & & tn & bd
+ ngc 6362 & 2.53 & -4.10 & -2.29 & -43.0 & 20.2 & 119.4 & 15.5 & 41.4 & 17.1 & tk & oh + ngc 6397 & 6.41 & -0.84 & -0.48 & 23.1 & 6.7 & 124.6 & 12.2 & -108.8 & 11.3 & tk & oh + ngc 6528 & 0.62 & 0.16 & -0.57 & 189.7 & 4.2 & 57.9 & 10.8 & 4.3 & 10.4 & b & bd + ngc 6541 & 1.75 & -1.25 & -1.35 & & & & & & & b & oh + ngc 6553 & 2.53 & 0.55 & -0.32 & 55.4 & 4.6 & 218.6 & 16.4 & 13.7 & 2.2 & tn & oh + m22 & 5.37 & 0.54 & -0.42 & 172.9 & 5.5 & 176.7 & 21.7 & -122.9 & 24.6 & tk & oh + m54 & -17.37 & 2.54 & -6.52 & & & & & & & h & sg + ngc 6752 & 5.19 & -1.44 & -1.73 & -24.7 & 5.6 & 193.9 & 9.1 & 21.9 & 7.2 & tk & oh + m55 & 3.69 & 0.74 & -2.09 & -184.4 & 9.6 & 43.7 & 27.7 & -108.5 & 15.3 & h & oh + m71 & 6.31 & 3.33 & -0.32 & 3.1 & 14.1 & 178.7 & 11.4 & -2.6 & 14.8 & tn & bd
+ ngc 7006 & -8.80 & 35.11 & -13.79 & -179.6 & 44.4 & 167.1 & 71.3 & 147.8 & 68.5 & h & yh + m15 & 4.63 & 8.33 & -4.73 & -142.5 & 23.8 & 167.8 & 34.2 & -96.0 & 31.1 & h & yh + m30 & 3.63 & 2.50 & -5.83 & -15.6 & 25.0 & -120.5 & 27.2 & 50.4 & 2.0 & h & oh + ngc 7492 & 1.63 & 9.25 & -23.09 & & & & & & & h & oh + liller 1 & -1.06 & -0.86 & -0.03 & & & & & & & b & bd + palomar 5 & -7.66 & 0.24 & 16.65 & -78.4 & 14.2 & 110.3 & 38.9 & 14.3 & 13.9 & h & yh + palomar 6 & 2.61 & 0.22 & 0.18 & & & & & & & b & yh + palomar 12 & -2.58 & 6.53 & -14.12 & 1.8 & 39.7 & 243.3 & 31.8 & -20.9 & 18.3 & h & sg + ruprecht 106 & -2.16 & -17.82 & 4.29 & & & & & & & h & yh + terzan 4 & -0.58 & -0.63 & 0.21 & & & & & & & b &
oh + terzan 5 & -1.77 & 0.69 & 0.30 & & & & & & & b & bd + terzan 7 & -13.25 & 1.29 & -7.96 & & & & & & & h & sg + lcc 0,0 & @xmath115 , @xmath116 , @xmath117 & @xmath118 , @xmath119 + 1,1 & @xmath120 , @xmath121 , @xmath122 & @xmath123 , @xmath124 + @xmath1251,@xmath1251 & @xmath126 , @xmath127 , @xmath128 & @xmath129 , @xmath130 + 1,@xmath1251 & @xmath131 , @xmath132 , @xmath133 & @xmath134 , @xmath135 + @xmath1251,1 & @xmath136 , @xmath137 , @xmath138 & @xmath139 , @xmath140 + | we have compiled a sample of globular clusters with high quality stellar abundances from the literature to compare to the chemistries of stars in the galaxy and those in dwarf spheroidal galaxies .
of the 45 globular clusters examined , 29 also have kinematic information .
most of the globular clusters belong to the galactic halo , however a signficant number have disk kinematics or belong to the bulge .
focusing on the [ @xmath0/fe ] and light r - process element ratios , we find that most globular cluster stars mimic those of the field stars of similar metallicities , and neither clearly resembles the presently available stellar abundances in the dwarf galaxies ( including the globular clusters in the large magellanic cloud ) .
the exceptions to these general elemental ratio comparisons are already known in the literature , e.g. , @xmath1 centauri , palomar 12 , and terzan 7 associated with the sagittarius remnant , and ruprecht 106 which has a high radial velocity and low [ @xmath0/fe ] ratio .
a few other globular clusters show more marginal peculiarities .
the most notable one being the halo cluster m68 which has a high galactocentric rotational velocity , a slightly younger age , and a unique [ si / ti ] ratio .
the [ si / ti ] ratios decrease with increasing [ fe / h ] at intermediate metallicities , which is consistent with very massive stars playing a larger role in the early chemical evolution of the galaxy .
the chemical similarities between globular clusters and field stars with [ fe / h]@xmath2 suggests a shared chemical history in a well mixed early galaxy . the differences to the published chemistries of stars in the dwarf
spheroidal galaxies suggests that neither the globular clusters , halo stars , nor thick disk stars had their origins in small isolated systems like the present - day milky way dwarf satellites . |
identifying the universal principles and patterns that shape cities and urban human activities is crucial for both fundamental and practical reasons , as it not only sheds light on social organizations , but is also crucial for planning and designing safer and more sustainable cities .
complex systems approaches @xcite provide us with tools and new ways of thinking about urban areas and the ways they may correspond to living organisms @xcite .
the activities of people in urban areas are responsible for the emergence of patterns on large scales that define the dynamics of cities . until recently , one of the main obstacles to quantify such ideas was the lack of large scale data on flows of people and their activities . however , in the last decade there has been a surge of new technologies that make it possible to obtain real - time data about populations , and these new `` social microscopes '' have changed in fundamental ways the study of social systems @xcite .
recent examples include characterizing @xcite and predicting @xcite individual human mobility patterns using mobile phone and social media data . toward that end
, twitter has been a valuable tool to track and to identify patterns of mobility and activity , especially using geolocated tweets .
geolocated tweets typically use the global positioning system ( gps ) tracking capability installed on mobile devices when enabled by the user to give his or her precise location ( latitude and longitude ) .
geolocated tweets have recently been used to study global mobility patterns @xcite and spatial patterns and dynamics of sentiment @xcite . here
we describe the collective dynamics of the greater metropolitan area of new york city ( nyc ) as reflected in the geographic dynamics of twitter usage .
we observe and quantify the patterns that emerge naturally from the hourly activities at different subareas , and discuss how they can be used to understand the social dynamics of urban areas .
twitter data can be understood not just by considering where people are but also by the extent to which they are preoccupied or have time and attention to devote to twitter posting .
we collected more than 6 million geolocated messages from twitter s streaming application programming interface ( api ) @xcite from which more than 90@xmath0 of geolocated tweets can be downloaded as they occur @xcite . from this data we observe wake / sleep cycles and the daily social `` heartbeat '' of the nyc area , reflecting the commuting dynamics from home to work in the diurnal cycle .
we identify differences in weekday and weekend dynamics , and find specific locations where activity occurs at certain hours , including the early morning at air transit hubs .
we also find anomalous events associated with specific individuals whose high engagement with twitter at specific times can dominate their local region .
we discuss how this dataset reflects the collective patterns of human activity and attention in both space and time .
we collected tweets between the latitudes @xmath1 $ ] and longitudes @xmath2 $ ] from august 19 , 2013 to december 31 , 2013 .
[ fig : map ] shows the geographical twitter coverage .
we aggregated the data of the corresponding days of the week , and in hourly units of time , resulting in @xmath3 time slices describing each hour of a `` typical week . ''
we divided the geographic area into @xmath4 cells ( @xmath5 ) . for each hour of the week we obtained the difference in that cell from the average number of tweets over the week as @xmath6\ ] ] where @xmath7 is the number of tweets in the cell @xmath8 at a given hour , @xmath9 the average number of tweets in a given cell averaged over all days and hours , and @xmath10 a constant that controls the slope of the hyperbolic tangent ( @xmath11 in all figures ) .
note that the @xmath12 function bounds @xmath13 to the range @xmath14 $ ] .
the values of @xmath13 were then used to generate a heat map of twitter activity in the nyc area for a particular hour ( see fig . [ fig : day ] ) .
we also constructed two and three - dimensional movies of those patterns that can be accessed at http://www.necsi.edu .
nyc geographical region with the locations of 6 million tweets shown .
the sharp land - sea boundary is apparent as is the boundary of land area with high population density.,scaledwidth=100.0% ] + the dynamics of twitter can be understood first by recognition of the dominant diurnal wake - sleep cycle and geographic commuting to work in manhattan from surrounding bedroom communities for the conventional workday hours approximately 9:00am - 5:00pm .
earlier in the morning people tweet from their homes , and therefore manhattan has much fewer tweets than average .
tweets are concentrated in manhattan in the morning work hours and peak there at mid - day ( around 13h ) , and become much more widely dispersed after work hours .
the bedroom community activity is high in the evening , clearly visible at 10:00pm , and decreases as people go to sleep .
twitter activity of weekdays compared to weekends , where the difference in the urban life can be clearly seen during the late night ( top panels ) and late afternoon ( bottom panels ) .
colors as in fig .
[ fig : day],title="fig:",scaledwidth=100.0% ] + twitter activity of weekdays compared to weekends , where the difference in the urban life can be clearly seen during the late night ( top panels ) and late afternoon ( bottom panels ) .
colors as in fig .
[ fig : day],title="fig:",scaledwidth=100.0% ] twitter activity of weekdays compared to weekends , where the difference in the urban life can be clearly seen during the late night ( top panels ) and late afternoon ( bottom panels ) .
colors as in fig .
[ fig : day],title="fig:",scaledwidth=100.0% ] + twitter activity of weekdays compared to weekends , where the difference in the urban life can be clearly seen during the late night ( top panels ) and late afternoon ( bottom panels ) .
colors as in fig .
[ fig : day],title="fig:",scaledwidth=100.0% ] in addition to these largest scale daily patterns , patterns emerge when comparing weekdays ( i.e. , monday to friday ) to weekends . in fig .
[ fig : weekdays ] we compare the late - night and late - afternoon twitter activity between sunday and a workday at 2:00am and 5:00pm . while the workday activity is suppressed almost everywhere at 2:00am , the nightlife activity at 2:00am sunday morning ( late saturday night ) has a unique pattern , with high activity in wide swaths of the city and suburbs .
high levels of activity span a band extending from lower manhattan across to brooklyn and hoboken , new jersey .
other high spots of night activity include the bronx , and union city and more specific spots in surrounding communities .
sunday afternoons also present an unusual pattern of widely dispersed but localized spots of activity likely corresponding to tweets in residential community areas of activity .
moreover , a peak of activity is observed in central park most of the day sunday that is not observed on other days of the week .
examples of times of high activity at locations listed in table [ tab : spots ] .
the left panel shows annotated activity plots using the colors as in fig .
[ fig : day ] , and the right panel shows the corresponding height as a three dimensional surface.,title="fig:",scaledwidth=100.0% ] + examples of times of high activity at locations listed in table [ tab : spots ] .
the left panel shows annotated activity plots using the colors as in fig .
[ fig : day ] , and the right panel shows the corresponding height as a three dimensional surface.,title="fig:",scaledwidth=100.0% ] examples of times of high activity at locations listed in table [ tab : spots ] .
the left panel shows annotated activity plots using the colors as in fig .
[ fig : day ] , and the right panel shows the corresponding height as a three dimensional surface.,title="fig:",scaledwidth=100.0% ] + examples of times of high activity at locations listed in table [ tab : spots ] .
the left panel shows annotated activity plots using the colors as in fig .
[ fig : day ] , and the right panel shows the corresponding height as a three dimensional surface.,title="fig:",scaledwidth=100.0% ] we also find other interesting activities at particular locations and times , as can be seen in fig .
[ fig : spots ] .
the top row shows the high activity at 6:00pm on friday at the three main airports of the area : john f. kennedy ( jfk ) airport ( * a * ) @xcite , newark airport ( * b * ) @xcite and la guardia airport ( * c * ) @xcite .
weekend activities are seen in the bottom row , we point out the meadowlands sports complex ( * d * ) @xcite and the statue of liberty ( * e * ) @xcite . .
jfk airport & mon - sat 6 am ; everyday 4pm-6pm ; sun 8 pm + * b*. newark airport & everyday 4pm-7pm + * c*. la guardia airport & sun - fri 4pm-7pm + * d*. meadowlands sports complex & sat 2pm-9pm ; sun 9am-7pm + * e*. statue of liberty & sat 11am-4pm ; sun 2 pm + [ tab : spots ] single user anomaly : the annotated peak corresponds to a single user tweeting more than 10 times the average of other users.,title="fig:",scaledwidth=100.0% ] + single user anomaly : the annotated peak corresponds to a single user tweeting more than 10 times the average of other users.,title="fig:",scaledwidth=100.0% ] + finally , fig .
[ fig : anom ] shows examples where an individual s behavior dominates the collective twitter activity .
we used statistical measures to discover that a single twitter user was responsible for most tweets in a specific region around 5:00am on weekdays .
detecting such anomalies enables distinguishing collective social dynamics from individual behavior that can at times dominate aggregate measures . to identify a single measure that can capture key aspects of the collective behavior of the city we considered the average distance of the tweets from a location in manhattan ( central park , @xmath15 , @xmath16 ) over time .
the distance was calculated using the haversine formula @xcite , @xmath17 \ , \ ] ] where @xmath18 km is the average radius of the earth , @xmath19 and @xmath20 , ( @xmath21 ) are the latitude and longitude of the point @xmath8 .
the results are shown in fig .
[ fig : avgd ] . average distance from central park at the different hours of the day for each day of the week .
the working days of the week , monday to friday , are shown in different shades of blue ( light blue for monday and the darkest blue for friday ) , while the weekends are shown in shades of red ( red for saturday and dark red for sunday).,scaledwidth=100.0% ] we find that while many people are expected to be located in bedroom communities , the average distance to manhattan falls overnight , reflecting the fact that downtown activity continues through the night so that most twitter activity at late hours / early morning ( 12:00am to 5:00am ) happens in areas in or close to manhattan . during weekdays a sharp peak at 6:00am corresponds to people tweeting in suburban areas before commuting .
after 8:00am the distance grows gradually during the day and through the evening , until it falls again overnight .
a change in this pattern is observed during weekends .
there is a horizontal shift in the curves for saturday and sunday of about 3 hours , reflecting people s tendency to tweet later into the night and wake up later the next morning .
the pre - commuting peak is entirely missing .
in this paper we characterized , for the first time , the patterns of weekly activity of the nyc area using more than 6 million geolocated tweets posted between aug 19 , 2013 and dec 31 , 2013 .
we related the collective geographical and temporal patterns of twitter usage to activities of the urban life and the daily `` heartbeat '' of a city .
the largest scale daily dynamics are the waking and sleeping cycle and commuting from the suburbs to office areas in manhattan , while the hourly dynamics reflects the interplay between commuting , work and leisure .
we showed not only that twitter can provide insight into understanding human social activity patterns , but also that our analysis is capable of identifying interesting locations of activity by focusing on departures from global behaviors .
we observed a peak of twitter usage in the suburbs of nyc before people begin their workday , and during evening hours , while the main activity during the workday and late at night is in manhattan .
in addition to the daily differences , we also characterized the weekly patterns , especially differences between the weekend and weekdays .
daytime recreational activities concentrate at identifiable locations spread widely across residential areas of the city .
we determined more specific times and locations of high activity at air transportation hubs , tourist attractions and sports arenas .
we analyze the role of particular individuals where they have large impacts on overall twitter activity , and find effects which may be considered outliers in discussion of social activity , but are an important aspect of human activity in the city .
we explored the use of the average distance from downtown to understand the dynamics of twitter usage , and while weekdays are similar in the geography of twitter usage , the sleep waking cycle is shifted later by about 3 hours during the weekends , and a pre - commuting twitter activity peak is absent . taken together
, those results demonstrate the potential of using social media analysis to develop insight into both geographic social dynamics and activities , and opens the possibility to understand and compare the life of cities on various scales .
a. pentland , reality mining of mobile communications : toward a new deal on data . in the global information technology report 20082009 ,
s. dutta , i. mia i ( world economic forum , geneva ) p. 75
http://hd.media.mit.edu/wef_globalit.pdf a. sadilek , j. krumm , far out : predicting long - term human mobility , in : twenty - sixth aaai conference on artificial intelligence ( 2012 ) , http://research.microsoft.com/en-us/um/people/jckrumm/publications%202012/sadilek-krumm_far-out_aaai-2012.pdf f. morstatter , j. pfeffer ,
h. liu , k. m. carley , is the sample good enough ? comparing data from twitter s streaming api with twitter s firehose , 7th international aaai conference on weblogs and social media ( icwsm ) 2013 , arxiv:1306.5204 [ cs.si ] ( 2013 ) . | describing the dynamics of a city is a crucial step to both understanding the human activity in urban environments and to planning and designing cities accordingly . here
we describe the collective dynamics of new york city and surrounding areas as seen through the lens of twitter usage .
in particular , we observe and quantify the patterns that emerge naturally from the hourly activities in different areas of new york city , and discuss how they can be used to understand the urban areas . using a dataset that includes more than 6 million geolocated twitter messages we construct a movie of the geographic density of tweets .
we observe the diurnal `` heartbeat '' of the nyc area .
the largest scale dynamics are the waking and sleeping cycle and commuting from residential communities to office areas in manhattan .
hourly dynamics reflect the interplay of commuting , work and leisure , including whether people are preoccupied with other activities or actively using twitter .
differences between weekday and weekend dynamics point to changes in when people wake and sleep , and engage in social activities .
we show that by measuring the average distances to the heart of the city one can quantify the weekly differences and the shift in behavior during weekends .
we also identify locations and times of high twitter activity that occur because of specific activities .
these include early morning high levels of traffic as people arrive and wait at air transportation hubs , and on sunday at the meadowlands sports complex and statue of liberty .
we analyze the role of particular individuals where they have large impacts on overall twitter activity .
our analysis points to the opportunity to develop insight into both geographic social dynamics and attention through social media analysis . |
veritas @xcite is sensitive to astrophysical @xmath0-rays in the energy range between @xmath1100 gev and @xmath130 tev .
the array is used to take @xmath11000 h of good - weather data each year , and can detect a 1% crab nebula flux source at zenith in @xmath325 h. the typical systematic errors on veritas measurements are 0.1 for the photon index and 20% in the integral flux . an upgrade of veritas was completed in 2012 .
the new energy threshold is expected to be @xmath160 gev with an overall sensitivity increase of @xmath120% .
the veritas collaboration hopes to improve the understanding of vhe blazars and their related science by expanding the known population and by making precision measurements of their spectra and variability patterns .
contemporaneous multi - wavelength ( mwl ) observations are a key component of this program since blazars are highly - variable at all wavelengths .
modeling of the resulting seds and searches for mwl correlations are used to provide insight into origin of the observed emission . since september 2007
, veritas has observed 128 blazars , for a total of 1981 h ( annual average of @xmath1400 h ) during good weather conditions . until 2010 ,
@xmath180% of the veritas blazar data were used for vhe discovery observations and follow - up studies of new sources . in 2010 , the focus became deep studies of known vhe sources , and @xmath160% of the data are now devoted to known sources .
target - of - opportunity ( too ) observations are a key part of the blazar program and typically average 50 h per year .
the 21 blazars detected by veritas are shown in table [ blazar_table ] ; see @xcite for more details .
.the 21 blazars detected at vhe with veritas .
the 10 vhe discoveries are marked with @xmath4 .
the long - term monitoring targets are shown along with their reason for inclusion in the program and the current annual exposure goals .
both the targets and the exposures may evolve .
the lbl s50716 + 714 is also monitored ( 10 h goal ) , but is not yet detected by veritas . [ cols="^,^,^,^,^",options="header " , ] veritas has regularly observed bllacertae ( an lbl ) since the start of the 2010 - 11 season .
it is not usually bright enough to be detected at vhe .
however , in june 2011 , the tail of a bright vhe flare was observed .
the peak vhe flux observed was 125% of the crab nebula flux , and the flare exponentially decayed in @xmath5 minutes . despite being the brightest vhe flux
ever observed from the object , the observed photon index ( @xmath6 ) is consistent with that measured by magic during a lower - flux ( 3% crab nebula ) flare in 2006 .
more details on the veritas and other mwl observations of this exceptional flare can be found in these proceedings @xcite .
the comparatively distant ( @xmath7 ) hbl 1es0414 + 009 is a known vhe emitter and was observed with veritas for 56 h of quality - selected live time between 2008 and 2011 .
these data resulted in the significant detection of a source with flux @xmath12% of the crab nebula flux @xcite . the observed spectrum is soft @xmath8 , largely due to the effects of ebl absorption .
although an hbl , its sed ( see figure [ blazar_plots ] ) is not well fit by one - zone leptonic models ; a lepto - hadronic model provides a better fit . since september 2007 , @xmath11260 h of veritas data were devoted to the discovery and follow - up observations of more than 100 blazars at vhe .
the candidates observed are largely hbls , but also include ibls , lbls , fsrqs , and fermi - lat sources likely associated with blazars . from 2007 - 09 , the discovery program @xcite focused on nearby ( redshift , @xmath9 ) x - ray - bright hbls , ibls , and fsrqs recommended in the literature and nearby egret - detected blazars .
some of the x - ray - bright targets were objects that met the selection criteria applied in the literature , but were first reported in later catalogs . following the release of the first fermi - lat catalog of mev - gev - bright blazars in 2009 @xcite , the discovery program has focused on objects detected by fermi - lat @xcite .
these observations resulted in the veritas discovery of vhe @xmath0-ray emission from 6 hbls and the first 4 ibls known to emit vhe photons @xcite .
the initial observations of 60% of these new vhe sources were motivated by the x - ray / egret - based selection criteria , and all these objects are detected by fermi - lat . in general , these new sources have soft vhe spectra ( @xmath10 ) , and only three exhibit strong vhe flux variability .
the vhe flux upper limits derived from the unsuccesful discovery observations are typically less than 2% of the crab nebula flux and often the strictest yet .
the successful vhe discovery observations @xcite of rxj0648.7 + 1516 were motivated by the identification of a cluster of @xmath11 gev photons in the fermi - lat data .
a total of 19 h of quality - selected live time was taken in march - april 2010 .
significant vhe emission ( @xmath12 , @xmath13% crab nebula flux ) was detected at the position of rxj0648.7 + 1516 , an unidentified radio , optical and x - ray emitter .
follow - up optical spectroscopy of this object yielded a continuum - dominated spectrum typical of bl lac objects and weak absorption lines compatible with z = 0.179 .
the sed of this blazar indicates it is an hbl , but it is not well fit by an ssc model .
rbs0413 is a bright hbl , at @xmath13 , that was initially targeted based on its inclusion in the sedentary survey @xcite .
it was observed for @xmath126 h of quality - selected live time between 2008 and 2010 @xcite .
the data yielded the detection of a soft - spectrum ( @xmath14 ) source with @xmath11% crab nebula flux .
its sed , shown in figure [ blazar_plots ] , is again not well described by a ssc model .
the modeling favors the inclusion of an additional external - compton ( ec ) component , and can also be fit by a lepto - hadronic scenario .
in 2010 , veritas began a systematic campaign to acquire deep exposures on 14 vhe blazars .
much of these data are acquired via regular monitoring exposures , and data taking is intensified during any flaring behavior .
regular simultaneous mwl observations are also organized .
the 14 selected blazars are indicated in table 1 , and include : * 6 comparatively distant , hard - vhe - spectrum hbls for ebl and igmf studies . * 4 bright hbls to easily enable mwl studies with high vhe statistics .
these may also have the highest likelihood for exhibiting greater than crab nebula flux flares . * 4 non - hbls to provide insights into the blazar sequence , to discover their low vhe states , and to study future vhe flares .
the total targeted exposure on these 14 vhe blazars is @xmath1220 h per year .
a discovery program will also continue , with a goal of @xmath1100 h per year .
it will focus on fermi - lat selected objects but will have a larger emphasis on higher - risk / higher reward targets that may require deep exposures . an active too program on both known vhe sources and vhe discovery candidates will also be maintained .
the triggers will be based on optical , x - ray , mev - gev gamma - ray and tev gamma - ray flares .
the veritas collaboration hopes to increase the amount of blazar too data to @xmath1100 h per year .
this research is supported by grants from the u.s .
department of energy office of science , the u.s . national science foundation and the smithsonian institution , by nserc in canada , by science foundation ireland ( sfi 10/rfp / ast2748 ) and by stfc in the u.k .
we acknowledge the excellent work of the technical support staff at the fred lawrence whipple observatory and at the collaborating institutions in the construction and operation of the instrument . | the veritas array of 12-m atmospheric - cherenkov telescopes in southern arizona began full - scale operations in 2007 , and is one of the world s most - sensitive detectors of astrophysical vhe ( e>100 gev ) @xmath0-rays .
approximately 50 blazars are known to emit vhe photons , and observations of blazars are a major focus of the veritas collaboration .
nearly 2000 hours have been devoted to this program and @xmath1130 blazars have already been observed with the array , in most cases with the deepest - ever vhe exposure .
these observations have resulted in 21 detections , including 10 vhe discoveries .
recent highlights of the veritas blazar observation program , and the collaboration s long - term blazar observation strategy , are presented .
address = harvard - smithsonian center for astrophysics , 60 garden st ms-20 , cambridge , ma 02138 , usa address = http://veritas.sao.arizona.edu/ blazars are a class of agn with relativistic jets pointed along the line of sight to the observer , and are the most numerous class of vhe source
. both sub - classes of blazars , bllac objects and flat spectrum radio quasars ( fsrqs ) , are detected at vhe .
nearly 80% of the vhe blazar population exhibits vhe flux variability .
these variations occur on time scales from minutes to years , but are typically only a factor of 2 to 3 .
the photon spectra of the observed vhe emission is often soft ( @xmath2 ) , largely due to the attenuation of vhe photons on the extragalactic background light ( ebl ) .
fermi - lat has detected almost every vhe blazar . by combining vhe and lat spectra , the highest - energy peak of the double - humped blazar spectral energy distributions ( seds )
is completely resolved . by combining this with x - ray and optical data , sampling the lower - energy peak ,
it is possible to constrain source models .
vhe blazar studies can also lead to strong new constraints on the ebl and the intergalactic magnetic field ( igmf ) , and provide insight into the origin of ultra - high - energy cosmic rays .
approximately 80% of the known vhe blazars are high - frequency - peaked bllac objects ( hbls ) .
the seds of these hbls can usually be fit by a synchrotron self - compton ( ssc ) model .
there are indications that the vhe emission of intermediate and low - frequency peaked bl lac objects ( ibls and lbls , respectively ) has a different leptonic origin ( ssc plus external compton ) .
however , the current ibl / lbl detections are plagued by poor statistics at vhe , and these non - hbl objects are typically only detected during flaring episodes which might have a different physical origin . |
as a fundamental many - body physical model , one - dimensional(@xmath0d ) interacting electron systems are an everlasting research topic .
unlike its high - dimensional counterparts , which are well understood from the quasiparticle picture in the fermi liquid theory,@xcite @xmath0d interacting electron systems can be described by bosonic gapless collective excitations of the fermion density fluctuations within the framework of the tomonaga - luttinger liquid theory.@xcite tomonaga - luttinger liquid behaviors in @xmath0d systems have been revealed in measuring the transport properties of cleaved - edge overgrowth quantum wires,@xcite quantum hall systems,@xcite and single - wall carbon nanotubes.@xcite it opens up the possibilities to test some theoretical predictions even in simplified pure physical models , and triggers intensive attention to the physics of the tomonaga - luttinger liquids.@xcite in mesoscopic fermi - liquid systems , electronic transport is modeled by the transmission of an incident electron through a potential barrier , and the conductance of the system is directly related to the transmission probability.@xcite however , potential barriers play a counterintuitive role in the tomonaga - luttinger liquids . in a seminal paper , kane and fisher@xcite have shown that the barrier is irrelevant for attractive electron - electron interactions , and cuts the system into two pieces for repulsive interactions . in transport measurements , tomonaga - luttinger liquid quantum wires of finite length need to be connected to leads acting as electron reservoirs .
if the reservoirs are modeled as @xmath0d noninteracting systems , then the model of inhomogeneous tomonaga - luttinger liquids with different interaction strengths in different parts is appropriate to investigate the transport properties of the tomonaga - luttinger liquids.@xcite it was argued that the dc conductance is not renormalized by the interactions.@xcite an interesting phenomenon of andreev - like reflections has been shown at the interfaces between the interacting wire and the noninteracting leads due to the mismatch of the interaction strengths.@xcite the presence of an impurity results in interference of the bosonic excitations which can be modulated by an applied bias voltage , and thus leads to characteristic oscillations of the backscattering current as a function of the dc voltage.@xcite feldman et al.@xcite have showed that backscattering off a weak dynamic impurity would enhance the current .
pea et al.@xcite and recher et al.@xcite have demonstrated interesting fabry - perot interference patterns of the non - linear conductance as a function of the bias voltage in carbon nanotubes attached to metallic reservoirs , where backscattering processes mainly occur at the two metal - contact - nanotube interfaces . in this paper , we investigate interference effects of electron tunneling through double barriers in an interacting quantum wire .
we assume that the @xmath0d interacting electron system is adiabatically attached to two @xmath0d noninteracting electron systems , which act as electron reservoirs .
such an assumption excludes the possibility of backscattering at the interfaces between the interacting and noninteracting segments . backscattering events
take place at the positions of the double barriers .
due to the existence of the andreev - like reflections at the interfaces , electron tunneling through the double barriers in a tomonaga - luttinger - liquid wire is more complicated and thus expected to be more interesting .
as we show later , there exists a contribution from the quantum mechanical interference of bosonic excitations to the backscattering current , besides the contributions from backscattering off the separate barriers .
it is the quantum mechanical interference that dominates the oscillatory characteristics of the backscattering current as a function of the applied bias voltage .
the aim of this work is to make clear how the oscillation pattern of the backscattering current in an interacting quantum wire with double barriers is modulated by the electron interaction strength and the arrangement of the double barriers .
since what we are interested in is the low - temperature transport properties , it is convenient to reformulate the problem in the framework of standard bosonization.@xcite first linearize the energy spectra about the fermi points @xmath11 , and introduce two species of electron operators @xmath12 ) to describe the right - moving and left - moving fermions , then the excitations of the fermion system can be described by a bosonic field @xmath13 . in this way the hamiltonian for the fermion system can be recast in the following bosonic form where @xmath15^{-1/2}$ ] is the interaction parameter , @xmath16 , and @xmath17 is the charge density wave velocity . in terms of the bosonic field @xmath18
, the current operator can be written as @xmath19 the current - bias relations of the nonequilibrium system is of interest , it is appropriate to adopt the keldysh formalism . following the path integral technical procedures developed in ref .
@xmath20 for a static impurity , the average current is obtained @xmath21 in eq .
( [ total current ] ) , @xmath22 is the background current in the absence of barriers , and @xmath23 is the backscattering current .
the average backscattering current is the sum of three contributions to leading order in @xmath24 ( see appendix ) @xmath25,\end{aligned}\ ] ] where @xmath26 , @xmath27/2\big \rangle_0 $ ] are the correlation functions of the bosonic field @xmath28 in the clean wire limit .
these correlation functions can be obtained by expanding the bosonic field on the basis of the eigenfunctions , which satisfy a specific inhomogeneous equation.@xcite it is evident that , besides the independent contributions from backscattering by the separate barriers , there exists a quantum mechanical interference contribution to the backscattering current due to the coexistence of double barriers .
it is just such an interference term that gives rise to some interesting transport characteristics in interacting quantum wires with double barriers . at zero temperature ,
the correlation functions yield a simplified expression @xmath29 ^ 2}{[\alpha^2+(2\xi_i+m)^2][\alpha^2+(2\xi_j+m)^2 ] } \big]\big\},\nonumber\end{aligned}\ ] ] where @xmath30 , @xmath31 is the dimensionless cutoff , with @xmath32 being the inverse of the traversal time of the charge density wave and @xmath33 being the high - energy cutoff frequency , @xmath34 is the andreev - like reflection coefficient at the interfaces . the above expressions for the backscattering current @xmath41 allow us to evaluate it for arbitrary values of the barrier positions ( @xmath42 ) , the interaction strength @xmath43 , and the applied voltage @xmath44 .
it is noticed that the backscattering current contributed from an individual barrier is dependent on the relative distance of the barrier to the wire center(@xmath45 ) , while that from the quantum mechanical interference term is dependent on the location details of the double barriers(@xmath46 and @xmath47 ) .
it seems impossible to obtain an explicit expression for the backscattering current , we resort to numerical calculations to investigate the backscattering current in some typical cases .
first , we investigate the dependence of the backscattering current on the interaction strength .
it is well known that the interaction strength is characterized by the interaction parameter @xmath48^{-1/2}$ ] .
we have @xmath49 for repulsive interactions and @xmath50 for noninteracting limit .
it should be noted that the smaller the value of @xmath43 , the stronger the interaction . in fig .
@xmath0 , we present numerical results of @xmath41 as a function of the applied voltage @xmath51 for different typical interaction parameters @xmath52 , as the barriers locate symmetrically near the ends of the wire with @xmath53 .
since the barriers are symmetrically located , the backscattering currents arising from different barriers are the same .
it can be observed that , as the interaction decreases , the oscillation of the incoherent addition of the backscattering currents contributed from the two barriers disappears gradually , while that contributed from the quantum mechanical interference persists and is more prominent .
the oscillatory behavior of the backscattering current is mainly determined by the quantum mechanical interference of the bosonic waves backscattering off the double barriers .
we attribute the less pronounced oscillation of the backscattering current in the strong interaction case to the suppression of the quantum mechanical interference by the electron - electron interaction . as shown by dolcini et al .
@xcite , the oscillation of the backscattering current in the single barrier case arises from a combined effect of the barrier , the finite length and the interaction in the wire .
the phase shift of bosonic excitations traveling between the barrier and the interfaces is responsible for such an oscillation and can be modulated by an applied bias voltage . in our case with double barriers , competition between the single - barrier interference and the double - barrier quantum mechanical interference results in an interesting oscillatory behavior of the backscattering current . in the cases of a given barrier location , the oscillation period of the backscattering current contributed from the quantum mechanical interference
is irrespective of the interaction strength , while the oscillation period from the incoherent addition of the backscattering currents off different barriers is increased as the interaction parameter is decreased , and eventually becomes infinite in the noninteracting limit , which can be found in fig . @xmath0 .
then we consider two special cases for strong interaction ( @xmath54 ) where the double barriers locate symmetrically near the midpoint of the interacting wire ( @xmath55 ) and right at the ends of the wire ( @xmath56 ) .
it can be found from fig . @xmath57
that in both cases the backscattering current oscillates in a more pronounced way , compared to the case where the double barriers are located symmetrically near the interfaces ( @xmath53 , fig .
@xmath0 ( a ) ) . in fig .
@xmath57 , we also observe that , the oscillation frequency of the incoherent addition of the backscattering currents off the separate barriers is twice that due to the quantum mechanical interference when the double barriers are located at the ends of the wire , while the two frequencies are about the same when the double barriers are located near the midpoint of the wire .
the reason is that the round - trip ballistic time for bosonic excitations propagating between a barrier and an interface in the former case is twice that in the later case .
up to the present we have analyzed the current - voltage characteristics in cases where the two barriers are symmetrically located .
as the barriers are located asymmetrically , it is expected that such an asymmetry has trivial effects on the current - voltage characteristics of weakly interacting wires , but influences significantly the transport properties of strongly interacting wires . in fig .
@xmath58 , we provide the results of the backscattering current of the interacting wire with symmetric and asymmetric locations of the barriers in both weak interaction(@xmath59 ) and strong interaction(@xmath54 ) cases .
we find that in the weak interaction case , the backscattering currents are about the same for symmetric and asymmetric barrier locations(figs .
@xmath58(a ) and @xmath58(b ) ) , as long as the spacing between the double barriers is fixed .
however , such a scenario is changed in the strong interaction case , where the current oscillation strongly depend on the symmetry of the barrier locations ( figs .
@xmath58(c ) and @xmath58(d ) ) .
it is interesting to note that , the period of current oscillation depends strongly on the spacing between the double barriers in the weak interaction limit , i.e. , the bigger the barrier spacing , the smaller the oscillation period .
this phenomenon can not be observed in the strong interaction limit .
we attribute the dependence of the current oscillation on the barrier spacing in the weak interaction limit to the resonant tunneling through a double - barrier structure in one - dimensional noninteracting electronic systems.@xcite in the case of strong interactions , different locations of the barriers lead to different interference patterns for bosonic plamonic excitations traveling between a barrier and the interfaces , and then results in different oscillation behaviors of the backscattering currents contributed from the separate barriers .
finally , we would like to point out that , depending on the interaction strength , the backscattering current exhibits different scaling rules with the applied bias voltage in a finite - length tomonaga - luttinger - liquid wire with double barriers , as can be found from figs .
this observation deserves further investigation .
in summary , we have investigated interference effects in electron transport through a double - barrier structure in an interacting quantum wire , which is ideally attached to two noninteracting leads .
we have found that the oscillation of the backscattering current with the applied voltage is mainly determined by the quantum mechanical interference due to the coexistence of the double barriers .
it is contrast to the single barrier case , where the current oscillation is strongly dependent on the interference of the andreev - like reflected plasmonic excitations propagating between the single barrier and the interfaces .
depending on the interaction strength and the locations of the barriers , the competition between these two kinds of interferences results in different oscillation and scaling behaviors of the backscattering current with the bias voltage .
this work is supported by the nsfc under grant no .
10404010 , the project - sponsored by srf for rocs , sem and the excellent talent fund of jiangxi normal university .
in this appendix , we outline the main formulae to calculate the backscattering current based on the keldysh functional approach following ref .
@xmath20 .
the generating functional takes the form @xmath61=\frac{1}{\mathcal { n}_\mathcal { z}}\int d\vartheta \exp\big[\frac{i}{\hbar}\mathcal { s } + \frac{i}{\sqrt{2}}\int d\textbf{x } j{(\textbf{x})}\vartheta{(\textbf{x})}\big],\end{aligned}\ ] ] where the action functional of the system can be written in terms of the boson field @xmath28 as @xmath62\nonumber\\ & & + \frac{e}{\sqrt{2}}\int dx \int dt e(x , t)\vartheta(x , t)-\int dt h_b.\end{aligned}\ ] ] after introduce the standard keldysh time contour and denote by @xmath63 and @xmath64 the complex fields on the upper and lower time branches of the keldysh contour , and define four green s functions averaged with respect to the free hamiltonian @xmath65 , the generating functional is rewritten as @xmath66=\frac{1}{\mathcal { n}_\mathcal { z}}\int d\vartheta^\pm \exp\big\{-\frac{1}{2}\int d\textbf{r } ' d\textbf{r}''\sum_{\eta~\eta'=\pm}\vartheta^\eta(\textbf{r}')\nonumber\\ & & ( c^{-1})^{\eta\eta'}(\textbf{r}';\textbf{r}'')\vartheta^{\eta ' } ( \textbf{r}'')\big\}\exp\big\{\sum_{\eta=\pm}\big(\frac{i\eta e}{\hbar\sqrt{\pi}}\int d\textbf{r}'e(\textbf{r } ' ) \nonumber\\ & & \vartheta^\eta(\textbf{r}')-\frac{i\eta}{\hbar}\int_{-\infty}^{+\infty } dt ' h_b[\vartheta^\eta ] + \frac{i}{\sqrt{2}}\int d\textbf{x}j{(\textbf{x})}\vartheta^\eta(\textbf{x}))\big)\big\},\nonumber\end{aligned}\ ] ] where @xmath67 is the inverse of of a @xmath68 matrix formed from the above four green s functions .
define the following matrices @xmath69 and @xmath70 the generating functional is reduced to the following simplified form @xmath61&=&\frac{1}{\mathcal { n}_\mathcal { z}}\int d\mbox{\boldmath$\vartheta$}e^{(-\mbox{\boldmath$\vartheta$}^t\textbf{c}^{-1}_0\mbox{\boldmath$\vartheta$ } + 2i\textbf{j}^t\textbf{q}\mbox{\boldmath$\vartheta$ } ) /2}\nonumber\\ & & ~~~~~~\times\exp\big\{-\frac{i}{\hbar}\sum_{\eta=\pm}\eta\int dt ' h_b[\vartheta^\eta]\big\}.\end{aligned}\ ] ] after shift the fields @xmath71 , one obtains a factorized form of the generating functional @xmath61&&=e^{-\textbf{j}^t\textbf{q}\textbf{c}_0\textbf{q}^t\textbf{j}/2}\nonumber\\ & & \times\big\langle\exp(-\frac{i}{\hbar}\sum_\eta\eta\int h_b[\vartheta^\eta+a^\eta_j]dt')\big\rangle_0.\end{aligned}\ ] ] from the expression of current operator eq .
( [ current operator ] ) , we have @xmath72}{\delta j(\textbf{x})}\big |_{j=0}\big\rangle\nonumber\\ & = & \frac{e}{\sqrt{\pi}}\partial_t \big [ \frac{ie}{\hbar \sqrt{\pi}}\int d\textbf{r } ' c^r_0(\textbf{x};\textbf{r}')e(\textbf{r}')-\frac{1}{\sqrt 2\hbar}\nonumber\\ & & \big \langle \sum_{\eta=\pm}\int dt ' \frac{\delta h_b[\vartheta^\eta+a^\eta_j]]}{\delta \vartheta^\eta}\frac{\delta a^\eta_j(\textbf{r}')}{\delta j(\textbf{x})}\big \rangle_0 \big]\nonumber\\ & = & e^2v / h - i_{bs},\end{aligned}\ ] ] where @xmath73\rangle_0 $ ] is the retarded green s function , and the backscattering current takes the following form here @xmath75 is the local conductivity of clean wire , and the backscattering current operator @xmath76}{\delta\vartheta(\textbf{x})}$ ] with @xmath77 , @xmath78 denotes an average along the keldysh contour with respect to the shifted hamiltonian @xmath79+h_b[\vartheta+a_0]$ ] . some algebras after substitution of the expression of @xmath9 into eq .
( [ bcc ] ) finally yield @xmath80\big)\nonumber \\ & & ~~~~\sin[\sqrt{4\pi}\vartheta^+(x_i , t')+2k_fx_i+\omega_0t']\big\rangle_0.\end{aligned}\ ] ] 10 _ theory of quantum liquids _ , p. nozieres and d. pines ( westview press 1999 ) .
s. tomonaga : prog .
. phys . * 5 * , 544 ( 1950 ) .
j. m. luttinger : j. math . phys . * 4 * , 1154 ( 1963 ) . f. d. m. haldane , j. phys .
c * 14 * , 2585 ( 1981 ) ; f. d. m. haldane , phys .
lett . * 47 * , 1840 ( 1981 ) .
s. tarucha , t. honda , and t. saku , solid state comm .
* 94 * , 413 ( 1995 ) ; a. yacoby , h. l. stormer , n. s. wingreen , l. n. pfeiffer , k. w. baldwin , and k. w. west , phys .
* 77 * , 4612 ( 1996 ) .
a. m. chang , l. n. pfeiffer , and k. w. west , phys .
lett . * 77 * , 2538 ( 1996 ) .
m. bockrath , d. h. cobden , j. lu , a. g. rinzler , r. e. smalley , l. balents , and p. l. mceuen , nature ( london ) * 397 * , 598 ( 1999 ) ; z. yao , h. w. j. postma , l. balents , and c. dekker , ibid . * 402 * , 273 ( 1999 ) .
_ quantum physics in one dimension _ , t. giamarchi ( clarendon press 2005 ) .
_ electronic transport in mesoscopic systems _ , s. datta ( cambridge university press 1997 ) .
c. l. kane and m. p. a.fisher , phys . rev .
b * 46 * , 15233 ( 1992 ) .
d. l. maslov and m. stone , phys .
b * 52 * , r5539 ( 1995 ) ; v. v. ponomarenko , ibid . * 52 * , r8666 ( 1995 ) .
i. safi and h. j. schulz , phys . rev .
b * 52 * , r17040 ( 1995 ) .
f. dolcini , h. grabert , i. safi , and b. trauzettel , phys .
lett . * 91 * , 266402 ( 2003 ) ; f. dolcini , b. trauzettel , i. safi , and h. grabert , phys . rev .
b * 71 * , 165309 ( 2005 ) .
d. e. feldman and y. gefen , phys .
b * 67 * , 115337 ( 2003 ) . c. s. pea , l. balents , and k. j. wiese , phys . rev .
b * 68 * , 205423 ( 2003 ) .
p. recher , n. y. kim , and y. yamamoto , phys . rev .
b * 74 * , 235438 ( 2006 ) . _ bosonization and strongly correlated systems _ , a. gogolin , a. nersesyan , and a. tsvelik ( cambridge university press 2004 ) .
_ the physics and applications of resonant tunnelling diodes _ , h. mizutz and t. tanoue ( cambridge university press 1995 ) . | we investigate interference effects of the backscattering current through a double - barrier structure in an interacting quantum wire attached to noninteracting leads . depending on the interaction strength and the location of the barriers , the backscattering current exhibits different oscillation and scaling characteristics with the applied voltage in the strong and weak interaction cases .
however , in both cases , the oscillation behaviors of the backscattering current are mainly determined by the quantum mechanical interference due to the existence of the double barriers . |
the most important decay modes of superheavy nuclei are mainly @xmath6 decay and spontaneous fission @xcite . among the many theoretical papers in this field
one should mention @xcite and @xcite . for atomic numbers larger than 121 cluster decay @xcite may compete as well @xcite . in 1928
g. gamow @xcite as well as r.w .
gurney and e.u .
condon @xcite gave the first explanation of @xmath2 decay based on quantum mechanical tunneling of a preformed particle at the nuclear surface .
the microscopic theory had been developed by many scientists , e.g. @xcite .
it was also extended to explain cluster decays @xcite .
simple relationships are also very useful @xcite . in our paper mentioned in the abstract @xcite we reported results obtained within macroscopic - microscopic method @xcite using cranking inertia @xcite and the best two - center shell model @xcite in the plane of two independent variables @xmath7 , where @xmath8 is the separation distance of the fragments and @xmath9 is the mass asymmetry with @xmath10 the mass numbers of the parent and nuclear fragments .
phenomenological deformation energy , @xmath11 , was given by yukawa - plus - exponential model @xcite , and the shell plus pairing corrections , @xmath12 are based on the asymmetric two center shell model ( atcsm ) @xcite .
this time we give more detailed arguments for the neighboring nucleus @xmath3cn . also similar figures are presented for heavy nuclei @xmath4pu and @xmath5cf .
the deep minimum of total deformation energy near the surface is shown for the first time as a strong argument for cluster preformation .
an outline of the model was presented previously @xcite .
we repeat few lines in this section . the parent @xmath13 is split in two fragments : the light , @xmath14 , and the heavy one , @xmath15 with conservation of hadron numbers @xmath16 and @xmath17 .
the corresponding radii are given by @xmath18 , @xmath19 , and @xmath20 .
the separation distance of the fragments is initially @xmath21 and at the touching point @xmath22 with @xmath23 fm .
the geometry for linearly increasing @xmath1 from 0 to @xmath24 is defined by : @xmath25 according to the macroscopic - microscopic method the total deformation energy contains the yukawa - plus - exponential ( y+em ) and the shell plus pairing corrections @xmath26 in units of @xmath27 the shell corrections are calculated with the strutinsky procedure as a sum of protons and neutrons contributions @xmath28 by solving the bcs @xcite system of two equations with two unknowns , we find the fermi energy , @xmath29 , and the pairing gap @xmath30 .
the total pairing corrections are given by @xmath31 and finally the total shell plus pairing corrections in mev @xmath32 the inertia tensor @xcite is given by @xmath33 where @xmath34 is the single - particle hamiltonian allowing to determine the energy levels and the wave functions @xmath35 , @xmath36 , @xmath37 are the bcs occupation probabilities , @xmath38 is the quasiparticle energy , and @xmath39 are the independent shape coordinates . for spherical fragments with @xmath40 deformation parameters the cranking inertia symmetrical tensor will have three components , hence the scalar @xmath41
when we find the least action trajectory in the plane @xmath42 we need to calculate the three components @xmath43 in every point of a grid of 66@xmath4424 ( for graphics ) or 412@xmath4424 ( for the real calculation ) for 66 or 412 values of @xmath45 and 24 values of @xmath46 or @xmath47 .
potential energy surfaces ( pes ) for spontaneous fission of @xmath3cn are shown in figures [ pescn ] and [ pescnl ] for constant radius , @xmath1 , of the light fragment and linearly increasing one , respectively .
the corresponding contour plots are given in figures [ cnc ] and [ cncl ] , where the first and second minima of deformation energy at every value of mass asymmetry are plotted with dashed and dotted white lines .
more details are given in the two tables [ tab ] and [ tabl ] . also , the position and value of maximum y+em model deformation energy versus mass asymmetry , @xmath48 , for fission of @xmath3cn with linearly increasing @xmath1 ( top ) and constant @xmath1 ( bottom ) are shown in figure [ eym ] . in figure [ cen ]
we compare the deformation energies with respect to spherical shapes for symmetrical fission of @xmath3cn with @xmath1 constant and linearly increasing @xmath1 ( lin ) .
one can see a relatively law macroscopic energy @xmath49 on which the shell and pairing corrections , @xmath50 deegs a rather deep minimum not far from the nuclear surface .
a completely different `` classical '' two - humped barrier , @xmath51 , may be seen for @xmath52 constant . while the first minima in figure [ cncl ] and table [ tabl ] ,
are all lying at @xmath53 , this is true only for 10 mass asymetries out of 23 in figure [ cnc ] and the table [ tab ] .
we compare in figure [ shpe2 ] the absolute values of shell and pairing correction energies for symmetrical fission of @xmath3cn with @xmath1 constant ( dashed line ) and linearly increasing @xmath1 ( solid line ) . as expected , the gap for protons , @xmath54 , and neutrons , @xmath55 , solutions of the bcs system of two equations , in figure [ bcs ] are also following similar variations , while the fermi energies , @xmath56 and @xmath57 have only a shallow minima in the surface region .
deep minima around @xmath58 are clearly seen in both figures .
similar results are also obtained for heavy nuclei like @xmath4pu , figure [ pu00lc ] and @xmath5cf ( see figure [ cf00lc ] ) .
see also figures 1 - 3 of the e - print @xcite .
at the touching point , @xmath59 , both kinds of variations of @xmath60 are ariving at the same state , hence the shell effects are identical there , as may be seen in figures [ shpe2 ] , [ pu00lc ] and [ cf00lc ] . the decimal logarithm of the dimensionles @xmath61 component of nuclear inertia tensor for symmetrical fission of @xmath3cn with linearly increasing @xmath1 isplotted in figure [ lgbr ] . at the touching point and beyond , @xmath62
, one should get the reduced mass : @xmath63 .
the proton contribution are more important than the neutron one .
this figure is completely different from the figure 5 of the ref.@xcite where the components @xmath61 for almost symmetrical fission ( with the light fragment @xmath64pd , @xmath65cd and @xmath66sn ) of @xmath3cn are shown for @xmath67 constant .
values larger than @xmath68 are seen in figure [ lgbr ] , compared to smaller than @xmath69 for @xmath67 constant .
if we use in graphics @xmath70 instead of @xmath8 then for @xmath0fl the interval of variation will be @xmath71 .
for the initial parent nucleus one may have either @xmath53 or / and @xmath72 .
this is the reason why the dashed line ends up at the value of @xmath73 . in present calculations
we have used 66 values of @xmath74 from @xmath75 to @xmath76 and 24 values of @xmath77 from @xmath75 to @xmath78 .
for minimization of action we need not only @xmath79 but also the values of @xmath80 in every point of a grid of 66@xmath4424 for 66 values of @xmath45 and 24 values of @xmath46 or @xmath47 .
we expect a dynamical path very different from the statical one shown in fig .
[ cncl ] with a white dashed line .
the optimum value of the parameter zero - point vibration energy , @xmath81 , used to reproduce the experimental value of @xmath3cn spontaneous fission half - life , @xmath82 . in conclusion , with our method of calculating the spontaneous fission half - life including macroscopic - microscopic method for deformation energy based on asymmetric two - center shell model , and the cranking inertia for the dynamical part
, we may find a sequence of several trajectories one of which gives the least action . assuming spherical shapes
, we found that the shape parametrization with linearly increasing @xmath1 is more suitable to describe the fission process of shs in comparison with that of exponentially or linearly decreasing law .
it is in agreement with the microscopic finding concerning the preformation of a cluster at the surface , which then penetrates by quantum tunneling the potential barrier .
cn shown as a pes in the upper panel of fig .
[ pescn ] .
the first and second minima of deformation energy at every value of mass asymmetry are plotted with dashed and dotted white lines .
[ cnc],width=377 ] cn shown as a pes in the upper panel of fig .
[ pescnl ] .
the first and second minima of deformation energy at every value of mass asymmetry are plotted with dashed and dotted white lines . @xmath86 linearly increasing with @xmath8 .
[ cncl],width=377 ] | in a previous article published in phys .
rev .
c 94 ( 2016 ) 014309 we have shown for the first time that the best dynamical trajectory during the deformation toward fission of the superheavy nucleus @xmath0fl is a linearly increasing radius of the light fragment , @xmath1 .
this macroscopic - microscopic result reminds us about the @xmath2 or cluster preformation at the nuclear surface , assumed already in 1928 , and proved microscopically many times .
this time we give more detailed arguments for the neighboring nucleus @xmath3cn . also similar figures are presented for heavy nuclei @xmath4pu and @xmath5cf .
the deep minimum of total deformation energy near the surface is shown for the first time as a strong argument for cluster preformation . |
in 1935 japanese physicist hideki yukawa predicted [ 1 ] the existence of a particle - a meson - with a mass about 200 times that of the electron , which mediated the nuclear forces . at first , the muon , discovered in 1937 , looked like a good candidate for the meson , but soon it became clear that the muon s properties did not match those predicted by yukawa theory .
then , in 1947 , lates , occhialini and powell discovered [ 2 ] the @xmath2 -meson , or pion , which did have the predicted meson properties .
today , we know that the muon is an elementary particle [ 3 ] , cousin to the electron , -leptons and neutrinos , and that mesons ( @xmath2 , @xmath3 , @xmath4 , etc . ) , including the pion , are combinations of quark and an antiquark .
everybody know that the discovery of -meson was an very important finding since this was in fact a fundamental step in the understanding the subatomic world .
it heralded the beginning of a deep revision of the physical concepts on the structure of matter .
however , who can tell us what exactly are those mesons ? also , it is well known that the pion as lightest of mesons has a finite size , with a mean charge radius of 0.66 fm , and that the long range part of the nucleon - nucleon forces necessarily arises from one - pion - exchange . the traditional picture of the nucleus as a collection of neutrons and protons bound together via the strong force has proven remarkable successful in understanding a rich variety of nuclear properties .
however , it is a well grounded achievement of modern nuclear physics that not only nucleons are relevant in the study of nuclear dynamics but that pions and the baryonic resonances like @xmath5 s and n * play an important role too [ 4 ] .
so , when the nucleus is explored at short distance scales the presence of short lived subatomic particles , such as the pion and delta , are revealed as nuclear constituents . at
even shorter distance scales the basic building blocks of matter , the quarks and gluons , are also revealed as nuclear constituents .
the role of pions , deltas , quarks and gluons in the structure of nuclei is one frontier of modern nuclear physics .
this modern picture of the nucleus bring us to the idea to search for new kind of natural radioactivities such as : ( @xmath6 , @xmath7 , @xmath5 , @xmath8 etc.)-emission during the nuclear fission in the region of heavy and superheavy nuclei .
moreover , new mode of nuclear fission such as : deltonic fission an hyperfission , was also suggested and investigated .
so , in 1985 , d. b. ion , m. ivascu and r. ion - mihai initiated [ 5 ] the investigation of the nuclear spontaneous pion emission as a new possible nuclear radioactivity called nuclear pionic radioactivity ( npir ) with possible essential contributions to the instability of heavy and superheavy nuclei .
in this paper we present a short review not only of the main theoretical problems of the npir but also of the essential experimental results obtained in these 18 years of existence of the nuclear pionic radioactivity .
moreover , in this paper by using the recent results on the spontaneous fission half lives @xmath0 of the heavy nuclei with @xmath9 we present new prediction on the pionic yields in the region of superheavy elements .
the nuclear pionic radioactivity of a parent nucleus ( a , z ) can be considered as an inclusive reaction of form : @xmath10 where x denotes any configuration of final particles ( fragments , light neutral and charged particles , etc . ) which accompany emission process .
the inclusive npir is in fact a sum of all exclusive nuclear reactions allowed by the conservation laws in which a pion can be emitted by a nucleus from its ground state .
the most important exclusive reactions which give the essential contribution to the inclusive npir ( 1 ) are the spontaneous pion emission accompanied by two body fission : : @xmath11 where @xmath12 hence , the npir is an extremely complex coherent reaction in which we are dealing with a spontaneous pion emission accompanied by a rearrangement of the parent nucleus in two or many final nuclei . charged pions as well as neutral pions can be emitted during two body or many body fission of parent nucleus .
the energy released in an exclusive nuclear reaction ( 1 ) is given by @xmath13 where @xmath14 is the energy released during spontaneous fission @xmath15 then the energy condition for the pionic radioactivity channel is @xmath16 or @xmath17 other interesting new natural radioactivities predicted at ifin - bucharest are as follows * * muonic radioactivity * [ 25 ] is a nuclear process in which the lepton pair @xmath18 is emitted during two or many body fission of the parent nucleus . for the the nuclear spontaneous emission of @xmath19 in binary fission we can write + @xmath20 + @xmath21 *
* lambdonic radioactivity * [ 26 ] is just a nuclear reaction of form + @xmath22 + @xmath23 * * hyperonic radioactivities * @xmath24 are also the possible nuclear decays of form + @xmath25 + @xmath26 * * hyperfission or weak fission * [ 27 ] is a fission process in which in one of fragment is hypernucleus : + @xmath27 + @xmath28 * * deltonic fission * [ 28 ] is also a fission process in which in one of fragment contain a delta resonance : + @xmath29 + @xmath30 a fission - like model ( see ref.[5 ] ) for the pionic radioactivity was regarded as a first stage in the development of an approximate theory of this new phenomenon that takes into account the essential degree of freedom of the system : @xmath6-fissility , @xmath6-fission barrier height , etc .
therefore , let us consider @xmath31 @xmath32 where @xmath33 and @xmath34 are the usual coulomb energy and surface energy of the parent nucleus given by @xmath35 @xmath36 with @xmath37mev and @xmath380.71 mev.@xmath39 is a parameter defined so that @xmath40 and @xmath41 are the coulombian and nuclear contributions to the pion mass . for @xmath42 , the entire pion mass is obtained only from coulomb energy of the parent nucleus .
so , by analogy with binary fission was introduced the pionic fissility @xmath43 which is given by @xmath44 or @xmath45 in the particular case @xmath46 we have @xmath47 we note that the definitions ( 22)-(24 ) are valid only with the constraints @xmath48 while the total variation @xmath49of parent nucleus at small deformation of type @xmath50is given by @xmath51 = \frac{2\varepsilon ^{2}e_{s}^{\pi f}\left ( 0\right ) } { 5}\left [ 1-x_{\pi f}^{\left ( \alpha \right ) } \right ] \label{26}\ ] ] in fig .
2 we presented the regions from the plane ( a , z ) in which some parent nuclei are able to emit spontaneously pions during the nuclear fission therefore , according to fig .
2 , we have the following important regions : * she ( super heavy elements)-region , indicated by white circles , where @xmath52 and @xmath53 , all the nuclei are able to emit spontaneously pions during the sf since no pion fission barrier exists . *
he ( heavy elements)-zone indicated by signs plus ( + + + ) , corresponding to @xmath52 and @xmath53 , where all the nuclei can emit spontaneously pion only by quantum tunneling of the pionic fission barrier .
* e - region , indicated by signs minus ( ) where the spontaneous pion emission is energetically interdicted since @xmath54 .
the dynamical thresholds for the pionic fission are obtained just as in fission case by using the substitution : @xmath55 so , by analogy with binary fission the barrier height for the pionic fission in a liquid drop model can be written as : @xmath56 \label{28}\ ] ] while the nuclear configuration at the sadle point is given by @xmath57 \label{29}\ ] ] where @xmath58 @xmath59 is the spherical radius and @xmath60 is a scale factor just as in binary spontaneous fission . in fig . 4 and 5 we present the values of e@xmath61 as well as the nuclear configuration at the saddle point compared with those from fission ( f ) or hyperfission ( hf ) .
true barrier height for the pionic fission is @xmath62 where and @xmath63and @xmath64 are correction due to shell effect at saddle point and ground state , respectively .
next , the essential ingredients used for the estimation of the relative pionic yields @xmath65 are as follows : * the variables @xmath66 and @xmath67 defined by + @xmath68 + @xmath69 * the real function @xmath70 defined for spontaneous fission ( sf ) by + @xmath71 * the scaling property : + @xmath72 + where @xmath73 is defined in ref . [ 5 ] . by fit
were obtained the following important results : @xmath74 for even - even parent nuclei and @xmath75 for parent nuclei with a - odd .
for example in fig .
6 we presented the experimental values of the @xmath76 compared with the fitted curve ( 37 ) .
finally , by using ( 36)-(37 ) we obtain the following important prediction @xmath77 ^{\delta _ { \alpha } ( a , z ) } \label{38}\ ] ] where @xmath78 and @xmath79 @xmath80 @xmath81 for even - even parent nuclei , and @xmath82 for a - odd parent nuclei . here , as in ref .
[ 31 ] we also adopt the notions of _ sf - nuclide _ , _ @xmath83-nuclide _ , and _
t - transition nuclide _ , as being those parent nuclei characterized by : @xmath84 , @xmath85 and yield @xmath86 , respectively , with their characteristic features presented in table 1 .
the predicted values of @xmath87 calculated with ( 38)-(40 ) for @xmath42 and @xmath88 , for even - even and a - odd parent nuclei , are presented in table 2a , b .
the pionic radioactivity was experimentally investigated by many authors [ 13]-[24 ] .
a short review of the results obtained on the spontaneous npir yields is presented in table 3 .
recently khryachkov et al .
[ 24 ] presented a new high performance spectrometer for investigation of ternary nuclear fission .
the measured characteristics of this spectrometer allow for its successful use in studies of ternary fission with the emission of @xmath39-particles , tritons , and protons as well as in the search for exotic nuclear fission accompanied by the emission of charged mesons ( @xmath89 . this new spectrometer was tested with a reaction of spontaneous @xmath90 ternary fission .
this choice was determined by the fact that this reaction is well studied and the available data can be employed for the performance check of this facility .
so , a @xmath91 layer 5 mm in diameter with an activity of 15 fissions / s was placed inside the spectrometer .
the measurements were carried out continuously 1.5 months .
a 2d spectrum of the scintillator signals obtained in coincidence with fragments is shown here in fig.8 ( see fig . 9 in ref . [ 24 ] ) . analyzing spectrum in fig .
8 has shown that , _ in the space between the electron and proton regions there are events that can not be explained by an external background or as effect of neutrons_. a detailed analysis of the parameter sf does not allow us to assign these events to protons or electrons .
we estimate these events as an indication of the possibility of existence of a ternary nuclear decay with the participation of charged mesons .
subsequently , it is suggested to study the meson emission during the spontaneous fission on this new facility which has a number of substantial advantages .
the experimental and theoretical results obtained on the npir can be summarized as follows : \2 .
the unusual background , observed by wild et al.[33 ] in ( @xmath93)-energy region below that characteristic for long range alpha emission from @xmath94 was interpreted by ion , bucurescu and ion - mihai [ 9 ] as being produced by negative pions emitted spontaneously by @xmath94 .
then , they inferred value of the pionic yield is : @xmath95 for @xmath96/fission . in a similar way ,
janko and povinec [ 22 ] , obtained the yield : @xmath97 for @xmath98/fission .
the supergiant radiohalos ( sgh ) , discovered by grady and walker [ 34 ] and laemmlein [ 35 ] are interpreted [ 12 ] as being the @xmath96-radiohalos and @xmath98-radiohalos , respectively .
hence these radiohalos are experimental signatures of the integral record in time of the natural pionic radioactivity from radioactive inclusions in ancient minerals .
a new interpretation of the experimental bimodal fission is obtained in terms of the unitarity diagrams .
is obtained in ref .
[ 29 ] so , the presence of the symmetric mode in the fragment mass - distribution at transfermium nuclei can be interpreted as experimental signature of the pionic radioactivity .
hence , it is expected that this new degree of nuclear instability can becomes dominant at she - nuclides .
this idea is illustrated not only by the symmetrization of fragment mass - yields from sf of the heavy nuclei with @xmath99 experimentally observed in fig .
7 but also by high values of @xmath100-yields obtained by using the theoretical prediction ( 38)-(39 ) .
the nuclides @xmath101 with @xmath102 are expected to be all @xmath103-nuclides while the @xmath101-isotopes with @xmath104 are all sf - nuclides ( see table 2a ) .
high pionic yields are obtained for @xmath105 ( 6.67 ) as well as for @xmath106 ( @xmath107 ) and @xmath108 ( @xmath109 ) .
\7 . the nuclides @xmath112 ( @xmath113 ) , @xmath114 ( @xmath115 ) , @xmath116 ( @xmath117 ) , @xmath118 ( @xmath119 ) , @xmath120 ( @xmath121 ) as well as @xmath122 ( @xmath123 ) all
are predicted to be @xmath103-nuclides .
finally , we note that dedicated experiments , using nuclei with theoretically predicted high pionic yields ( e.g. @xmath124 , @xmath125 , @xmath106 , @xmath126 , @xmath127 , @xmath128 and @xmath122 ( @xmath123 ) ) , are desired , since the discovery of the nuclear pionic radioactivity is of fundamental importance not only in nuclear science ( e. g. , for clarification of high instability of she- nuclei ) but also for geochemistry and cosmology . | in this paper a short review of the theoretical problems of the pionic radioactivity is presented .
the essential experimental results obtained in the 18 years of existence of the nuclear pionic radioactivity are reviewed . moreover , using the recent results on the spontaneous fission half lives @xmath0 of the heavy nuclei with @xmath1 new predictions on the pionic yields in the region of superheavy elements are established . |
chiral symmetric strongly coupled lattice gauge theories with fermion and scalar matter fields , @xmath2 and @xmath3 , contain unconfined fermions @xmath4 .
their mass @xmath5 is generated dynamically in the phase with chiral symmetry breaking @xcite and such models can thus be considered as an alternative to the standard higgs - yukawa mechanism of fermion mass generation , provided they are nonperturbatively renormalizable @xcite . here
we consider a 2d model of this type with continuous chiral symmetry , the @xmath6 model , which in the strong coupling limit is equivalent to the chiral gn@xmath0 model .
the observed scaling properties at large but finite gauge coupling indicate that the @xmath1 model belongs also here to the universality class of the gn@xmath0 model .
this would mean that the model is renormalizable and that the shielded gauge mechanism of fermion mass generation , proposed in @xcite , works in 2d .
a similar problem in 4d is addressed in ref .
@xcite and in these proceedings @xcite .
the action of the @xmath1 model consists of three parts : @xmath7 where @xmath8 here @xmath2 is a staggered fermion field with charge one , @xmath9 u(1 ) is a compact abelian gauge field with coupling @xmath10 and @xmath3 is a complex scalar field with charge one and constraint @xmath11 .
the mass term is introduced for technical reasons and the model is meant in the limit @xmath12 , where the action has a global u(1 ) chiral symmetry . in the limiting case @xmath13
one can perform the lee - shrock - transformation @xcite in which the scalar and gauge fields are integrated out and a four fermion term appears .
this leads to the following action ( @xmath14 ) : @xmath15 \biggr ) \nonumber\\ & & + \underbrace{\frac{am_0}{r}}_{\displaystyle am'_0 } \sum_x \bar{\chi}_x\chi_x \;. \label{wirk}\end{aligned}\ ] ] the four fermion coupling @xmath16 is a function of @xmath17 which is shown in fig .
[ fig:1 ] . the transformed action ( [ wirk ] ) for @xmath18 is that of the chiral gn@xmath19 model .
this model has a critical point at @xmath20 , where the fermion mass vanishes as @xmath21
we first verified the predicted scaling law ( [ scal ] ) at @xmath23 .
for @xmath18 the simulation is possible only for small lattices ( @xmath24 ) because of the slow convergence of the fermion matrix inversion .
hence we ` somehow ' had to extrapolate the masses measured at @xmath25 to @xmath18 .
it turned out that all examined schemes which tried to extrapolate for fixed @xmath17 to @xmath18 failed . but to perform the continuum limit @xmath26 it is not necessary to set @xmath18 first .
also a combined limit @xmath27 and @xmath28 can lead to a continuum theory without bare mass .
therefore we looked for a scaling law , in which @xmath25 is allowed for finite @xmath17 but @xmath28 when @xmath26 .
such a generalized scaling law is provided by the sd - equations for the model ( [ wirk ] ) truncated to order @xmath29 @xcite : @xmath30 @xmath31 these two coupled equations for functions @xmath32 and @xmath33 , with @xmath34 , were solved numerically @xcite . for infinite volume and small @xmath35
their approximate analytic solution is @xmath36 for @xmath18 one obtains the scaling behavior ( [ scal ] ) as @xmath37 .
the idea of the combined limit @xmath26 and @xmath38 is to make @xmath39 a function of @xmath17 in such a way , that eq .
( [ eq : n ] ) is solvable and that @xmath40 .
we choose ( see fig . [ fig:2 ] ) @xmath41 with a free parameter @xmath42 obeying @xmath43 .
the generalized scaling law is then @xmath44 and thus @xmath45 obviously , the ratio @xmath46 does not vanish except if @xmath47 .
[ fig:3 ] shows , up to the factor @xmath48 which approaches 1 when @xmath37 , the ratio of the measured fermion mass and the bare mass plotted against @xmath49 . in this ratio the bare part @xmath50 of the fermion mass
is ` divided out ' .
calculations have been performed for @xmath51 .
it turned out to be very difficult to obtain sufficient data for larger values of @xmath42 , as much larger lattices would be needed .
the full lines show the corresponding ratio calculated using the sd - equations for finite lattice and those @xmath25 given by ( [ gl5.1 ] ) .
equation ( [ gl5.3 ] ) is represented by horizontal dashed lines .
good agreement can be observed , especially for not too small @xmath50 and not too large @xmath35 .
_ thus at @xmath13 we have a suitable analytic description of the scaling behavior of the measured fermion mass , at least in the interval @xmath52 . _
_ but how can the data be described for @xmath54 ? _ fig .
[ fig:4 ] shows the same ratio , but for masses measured at @xmath55 . as expected , at the same @xmath35 value the fermion masses are in general smaller than those for @xmath13 , as the gauge coupling becomes smaller
this is seen by comparison of the data with the dotted lines which indicate the ratios calculated by the sd - equations at @xmath13 .
the data suggest a simple modification of the scaling behavior at @xmath53 with respect to @xmath22 . indeed , we have found that a good agreement between the measured data and the sd - equations can be obtained when in these equations the parameter @xmath42 is replaced by a fitted value @xmath56 .
the full lines in fig .
[ fig:4 ] show the calculated ratio using the sd - equations ( [ sdn],[sdf ] ) with a bare mass which is given by ( [ gl5.1 ] ) but using @xmath56 instead of @xmath42 .
the scaling law for @xmath54 corresponding to equations ( [ gl5.3 ] ) and ( [ gl5.1 ] ) when using @xmath56 is : @xmath58 with @xmath59 for the examined interval @xmath52 we thus observe the same scaling law for @xmath54 as that for @xmath22 apart from the @xmath35-independent factor @xmath60 with @xmath61 .
_ thus we found some evidence that the @xmath1 model for @xmath62 and @xmath50 finite but approaching zero belongs to the same universality class as the gn@xmath19 model with similar bare mass . _
a method of extrapolating to the @xmath18 @xmath63 limit has not yet been found , however .
a calculation of @xmath65 for @xmath66 is very difficult because of its low values , and we could not study its scaling behavior any more . but looking at some bulk observables and bosonic masses , we found no indication that the properties of the @xmath1 model substantially change when @xmath67 is increased .
we also found that the kosterlitz - thouless ( kt ) phase transition occuring at @xmath68 , when the @xmath1 model reduces to the xy@xmath0 model and decoupled fermions , gets weaker and dissolves at some finite @xmath67 when @xmath67 is increased .
this remnant of the kt phase transition does not influence the fermion mass in an appreciable way .
these observations suggest that the scaling behaviour of @xmath65 is the same at small and large @xmath67 , and that only the magnitude of @xmath65 decreases with increasing @xmath67 .
the continuum limit is approached at any @xmath69 by turning @xmath26 .
if both above hypothetical extensions of our results are correct , then the @xmath1 model belongs to the universality class of th gn@xmath0 model at all @xmath69 and is thus a renormalizable field theory in which the mass of an unconfined fermion is generated dynamically .
it is thus an example of the shielded gauge mechanisme proposed in @xcite . | by means of the lee - shrock transformation we generalize the 2d gross - neveu ( gn@xmath0 ) model to a u(1 ) gauge theory with charged fermion and scalar fields in 2d ( @xmath1 model ) .
the @xmath1 model is equivalent to the gn@xmath0 model at infinite gauge coupling .
we show that the dynamical fermion mass generation and asymptotic freedom in the effective four - fermion coupling persist also when the gauge coupling decreases .
these phenomena are not influenced by the xy@xmath0 model phase transition at weak coupling .
this suggests that the @xmath1 model is in the same universality class as the gn@xmath0 model and thus renormalizable . |
despite the tremendous effort that has led to the discovery of neutrino mass @xcite the mechanism of neutrino mass generation will remain open for years to come ( a detailed analysis of the three
neutrino oscillation parameters can be found in @xcite ) . the most popular mechanism to generate neutrino masses is the seesaw mechanism @xcite .
although the seesaw fits naturally in so(10 ) unification models , we currently have no clear hints that uniquely point towards any unification scheme .
therefore it may well be that neutrino masses arise from physics having nothing to do with unification , such as certain seesaw variants @xcite , and models with radiative generation @xcite .
here we focus on the specific case of low - energy supersymmetry with violation of r parity , as the origin of neutrino mass .
r parity is defined as @xmath0 with @xmath1 , @xmath2 , @xmath3 denoting spin , baryon and lepton numbers , respectively @xcite . in these models
r parity can be broken either explicitly or spontaneously . in the first case we consider the bilinear r
parity violation model , the simplest effective description of r - parity violation @xcite .
the model not only accounts for the observed pattern of neutrino masses and mixing @xcite , but also makes predictions for the decay branching ratios of the lightest supersymmetric particle @xcite from the current measurements of neutrino mixing angles @xcite . in the second case r
parity violation takes place `` a la higgs '' , i.e. , spontaneously , due to non - zero sneutrino vacuum expectation values ( vevs ) @xcite . in this case one of the neutral cp - odd scalars is identified with the majoron , @xmath4 .
in contrast with the seesaw majoron , ours is characterized by a small scale ( tev - like ) and carries only one unit of lepton number . in previous studies
@xcite it was noted that the spontaneously broken r parity ( ` sbrp ` ) model leads to the possibility of invisibly decaying higgs bosons , provided there is an 1 singlet superfield @xmath5 coupling to the electroweak doublet higgses , the same that appears in the ` nmssm ` .
we have reanalyzed @xcite this issue taking into account the small masses indicated by current neutrino oscillation data @xcite .
we have shown explicitly that the invisible higgs boson decay eq .
( [ eq : hjj ] ) , @xmath6 can be the most important mode of higgs boson decay .
this is remarkable , given the smallness of neutrino masses required to fit current neutrino oscillation data .
since ` brpv ` has been discussed in the literature several times @xcite we will repeat only the main features of the model here . we will follow the notation of @xcite . the simplest bilinear model
( we call it the ` mssm ` ) is characterized by the superpotential @xmath7 in this equation @xmath8 is the ordinary superpotential of the ` mssm ` , @xmath9\ ] ] where @xmath10 are generation indices , @xmath11 are @xmath12 indices .
we have three additional terms that break r - parity , @xmath13 these bilinear terms , together with the corresponding terms in the soft supersymmetric ( ` susy ` ) breaking part of the lagrangian , @xmath14 define the minimal model , which we will adopt throughout this paper .
the appearance of the lepton number violating terms in eq .
( [ eq : wrpv ] ) leads , in general , to non - zero vacuum expectation values for the scalar neutrinos @xmath15 , called @xmath16 in the rest of this paper , in addition to the vevs @xmath17 and @xmath18 of the ` mssm ` higgs fields @xmath19 and @xmath20 .
together with the bilinear parameters @xmath21 , the @xmath16 induce mixing between various particles which in the ` mssm ` are distinguished ( only ) by lepton number ( or r parity ) .
mixing between the neutrinos and the neutralinos of the ` mssm ` generates a non - zero mass for one specific linear superposition of the three neutrino flavor states of the model at tree - level while 1-loop corrections provide mass for the remaining two neutrino states @xcite . in the basis
@xmath22 the neutral fermions mass terms in the lagrangian are given by @xmath23 where the neutralino / neutrino mass matrix is @xmath24\ ] ] with @xmath25 \quad ; \quad m=\left [ \begin{array}{c } a_1 \cr
a_2 \cr a_3 \end{array } \right]\ ] ] where @xmath26 .
this neutralino / neutrino mass matrix is diagonalized by @xmath27 if the parameters are small ( we will show below that this is indeed the case ) , then we can block - diagonalize @xmath28 approximately to the form diag(@xmath29 ) @xmath30 with @xmath31 the matrices @xmath32 and @xmath33 diagonalize @xmath34 and @xmath35 @xmath36 where @xmath37 so we get at tree level only one massive neutrino , the two other eigenstates remaining massless
. the tree level value will give the atmospheric mass scale , while the other states will get mass at one loop level .
therefore the ` brpv ` model produces a hierarchical mass spectrum .
we will show below how the one loop masses are generated .
the self energy for the neutralino / neutrino is @xcite , @xmath38 -\left [ p_l \pi^l_{ij } + p_r \pi^r_{ij } \right]\right\}\ ] ] ( 0,0 ) ( -0.5,10 ) then the pole mass is , @xmath39 with @xmath40_{\delta=0}\ ] ] where @xmath41 an the parameter that appears in dimensional reduction is , @xmath42 explicit expressions can be found in @xcite .
[ cols="^,^ " , ] in ref.@xcite a complete analysis of the parameter space where this happens is presented .
the ` brpv ` model is a simple extension of the ` mssm ` that leads to a very rich phenomenology .
we have calculated the one loop corrected masses and mixings for the neutrinos in a completely consistent way , including the rg equations and correctly minimizing the potential .
we have shown that it is possible to get easily maximal mixing for the atmospheric neutrinos and large angle msw , as it is preferred by the present neutrino data .
we have also obtained approximate formulas for the solar mass and solar mixing angle , that we found to be very good , precisely in the region of parameters favored by this data .
we emphasize that the ` lsp ` decays inside the detectors , thus leading to a very different phenomenology than the ` mssm ` .
no matter what supersymmetric particle is the ` lsp ` measurements of branching ratios at future accelerators will provide a definite test of the ` brpv ` model as a viable model for explaining the neutrino properties .
we have also discussed the possibility of invisibly decaying higgs boson in the context of the ` sbrp ` model .
one of the neutral cp - odd scalars in this model , is the nambu - goldstone boson associated to the breaking of lepton number , the majoron @xmath4 .
neutrino masses and mixings can easily be explained .
in contrast to the ` mssm ` , where the higgs boson can decay invisibly only to supersymmetric states , in our case the higgs boson can decay mainly invisibly through the decay @xmath43 . as a result , our analysis indicates that invisibly decaying higgs bosons should be an important topic in the agenda of future accelerators , such as the large hadron collider and the next linear collider .
super - kamiokande , y. fukuda _ et al .
lett . * 81 * , 1562 ( 1998 ) , [ hep - ex/9807003 ] .
sno , q. r. ahmad _ et al .
* 89 * , 011301 ( 2002 ) , [ nucl - ex/0204008 ] .
kamland , k. eguchi _
et al . _ ,
lett . * 90 * , 021802 ( 2003 ) , [ hep - ex/0212021 ] . m. maltoni , t. schwetz , m. a. tortola and j. w. f. valle , new j. phys . * 6 * , 122 ( 2004 ) , [ hep - ph/0405172 ] . m. gell - mann , p. ramond and r. slansky , ( 1979 ) , print-80 - 0576 ( cern ) .
r. n. mohapatra and g. senjanovic , phys . rev .
* d23 * , 165 ( 1981 ) . y. chikashige , r. n. mohapatra and r. d. peccei , phys . lett . *
b98 * , 265 ( 1981 ) . j. schechter and j. w. f. valle , phys .
rev . * d22 * , 2227 ( 1980 ) .
r. n. mohapatra and j. w. f. valle , phys . rev . *
d34 * , 1642 ( 1986 ) .
a. zee , phys .
* b93 * , 389 ( 1980 ) .
k. s. babu , phys
b203 * , 132 ( 1988 ) . c. s. aulakh and r. n. mohapatra , phys .
lett . * b119 * , 136 ( 1982 ) .
m. a. diaz , j. c. romao and j. w. f. valle , nucl .
b524 * , 23 ( 1998 ) , [ hep - ph/9706315 ] .
m. a. diaz , m. hirsch , w. porod , j. c. romao and j. w. f. valle , phys .
rev . * d68 * , 013009 ( 2003 ) , [ hep - ph/0302021 ] . m. hirsch , m. a. diaz , w. porod , j. c. romao and j. w. f. valle , phys .
rev . * d62 * , 113008 ( 2000 ) , [ hep - ph/0004115 ] , err - ibid.d65:119901,2002 .
j. c. romao , m. a. diaz , m. hirsch , w. porod and j. w. f. valle , phys . rev . *
d61 * , 071703 ( 2000 ) , [ hep - ph/9907499 ] .
m. hirsch and j. w. f. valle , new j. phys .
* 6 * , 76 ( 2004 ) , [ hep - ph/0405015 ] . m. hirsch , w. porod , j. c. romao and j. w. f. valle , phys . rev .
* d66 * , 095006 ( 2002 ) , [ hep - ph/0207334 ] . w. porod , m. hirsch , j. romao and j. w. f. valle , phys . rev . *
d63 * , 115004 ( 2001 ) , [ hep - ph/0011248 ] .
d. restrepo , w. porod and j. w. f. valle , phys . rev . *
d64 * , 055011 ( 2001 ) , [ hep - ph/0104040 ]
. m. hirsch and w. porod , phys . rev . *
d68 * , 115007 ( 2003 ) , [ hep - ph/0307364 ] .
a. masiero and j. w. f. valle , phys
b251 * , 273 ( 1990 ) .
j. c. romao , c. a. santos and j. w. f. valle , phys .
b288 * , 311 ( 1992 ) .
m. shiraishi , i. umemura and k. yamamoto , phys . lett . *
b313 * , 89 ( 1993 ) .
j. c. romao and j. w. f. valle , nucl .
b381 * , 87 ( 1992 ) .
j. c. romao , f. de campos and j. w. f. valle , phys .
b292 * , 329 ( 1992 ) , [ hep - ph/9207269 ] .
f. de campos , j. w. f. valle , a. lopez - fernandez and j. c. romao , hep - ph/9405382 .
f. de campos , o. j. p. eboli , j. rosiek and j. w. f. valle , phys .
rev . * d55
* , 1316 ( 1997 ) , [ hep - ph/9601269 ] . m. hirsch , j. c. romao , j. w. f. valle and a. villanova del moral , phys . rev . * d70
* , 073012 ( 2004 ) , [ hep - ph/0407269 ] . f. de campos , m. a. garcia - jareno , a. s. joshipura , j. rosiek and j. w. f. valle , nucl . phys . * b451 * , 3 ( 1995 ) , [ hep - ph/9502237 ] . a. g. akeroyd , m. a. diaz , j. ferrandis , m. a. garcia - jareno and j. w. f. valle , nucl . phys . *
b529 * , 3 ( 1998 ) , [ hep - ph/9707395 ] . t. banks , y. grossman , e. nardi and y. nir , phys . rev . *
d52 * , 5319 ( 1995 ) , [ hep - ph/9505248 ] . g. passarino and m. j. g. veltman , nucl . phys . *
b160 * , 151 ( 1979 ) .
b. mukhopadhyaya , s. roy and f. vissani , phys . lett . *
b443 * , 191 ( 1998 ) , [ hep - ph/9808265 ] . | we review supersymmetric models where r parity is broken either explicitly or spontaneously .
the simplest unified extension of the ` mssm ` with explicit bilinear r
parity violation provides a predictive scheme for neutrino masses and mixings which can account for the observed atmospheric and solar neutrino anomalies . despite the smallness of neutrino masses r - parity violation
is observable at present and future high - energy colliders , providing an unambiguous cross - check of the model .
this model can be shown to be an effective model for the , more theoretically satisfying , spontaneous broken theory .
the main difference in this last case is the appearance of a massless particle , the majoron , that can modify the decay modes of the higgs boson , making it decay invisibly most of the time . |
the @xmath0 spectrum of the laplacian operator , denoted by @xmath1 , has been analyzed and computed for large classes of complete noncompact manifolds .
for example , when that manifold has a soul whose exponential map is a diffeomorphism , supposing nonnegative sectional curvature and some other additional conditions under dimension of manifold , escobar @xcite and escobar and freire @xcite proved that @xmath2 where @xmath3 denotes the essential spectrum of laplacian operator .
the additional conditions under dimension of @xmath4 assumed by them were proved to be unnecessary by zhou @xcite .
now , by assuming that the complete noncompact manifold has nonnegative ricci curvature and euclidean volume growth , the essential spectrum of the laplacian operator was proved to be @xmath5 by donnelly @xcite .
it is classical that the essential spectrum of laplacian operator on compact manifold is empty , and thus , its @xmath6 spectrum is discrete . given that the @xmath6 spectrum of laplacian operator for a large class of noncompact manifolds is essential , it becomes important to estimate the bottom ( that is , the greatest lower bound ) of essential spectrum of @xmath7 denoted by @xmath8 .
there exists many interesting results on the estimates for @xmath8 , for instance , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite . the problem of estimating the @xmath8 of complete noncompact manifold @xmath4 under simple geometric assumptions has been intensively studied in the past decades . for example , donnelly proved in @xcite that @xmath9 when the ricci curvature is bounded from below by the constant @xmath10 where @xmath11 and @xmath12 in @xcite and @xcite , brooks generalized this donnelly s estimate .
he showed that if @xmath4 has infinite volume , then @xmath13 where @xmath14 ; and if @xmath4 has finite volume , @xmath15 where @xmath16 .
later , higuchi @xcite improved these brooks estimates . in many occasions
, it is natural to consider a @xmath17dimensional riemannian manifold @xmath18 endowed with a weighted measure of the form @xmath19 , where @xmath20 is a smooth function on @xmath4 , called _
weight function _ , and @xmath21 is the volume element induced by the metric @xmath22 . a _ weighted manifold _ is a triple @xmath23 the _ weighted laplacian operator _
@xmath24 , defined by @xmath25 is associated to @xmath19 as well as @xmath26 is associated to @xmath27 moreover , @xmath24 is a self - adjoint operator on the @xmath28 space of square integrable functions on @xmath4 with respect to the measure @xmath29 and therefore , the @xmath28 spectrum of @xmath24 on @xmath4 , denoted by @xmath30 , is a subset of @xmath31 also , @xmath30 can be decomposed into the disjoint union @xmath32 , where @xmath33 is the set of isolated eigenvalues of finite multiplicity , called _ discrete spectrum _ , and its complement @xmath34 called _ essential spectrum _ , is the set of eigenvalues of infinite multiplicity and accumulation points of the spectrum .
a natural extension of the ricci curvature tensor to this new context is the bakry - mery ricci curvature tensor , see @xcite , given by @xmath35 where @xmath36 is the hessian of @xmath20 on @xmath4 .
it is known that a complete weighted manifold satisfying @xmath37 for some constant @xmath38 is not necessarily compact .
one of the examples is the gaussian shrinking soliton @xmath39 with the canonical metric @xmath40 and @xmath41 if @xmath42 is complete noncompact manifold with @xmath43 and @xmath44 , the @xmath28 essential spectrum of @xmath45 was proved to be @xmath5 by silvares @xcite . moreover , for a compact weighted manifold with @xmath46 for constant @xmath38 , the @xmath28 spectrum of @xmath45 is discrete , see @xcite .
let @xmath47 be the geodesic ball of @xmath4 with center in a fixed point @xmath48 and radius @xmath49 .
the _ weighted volume _ of @xmath47 is given by @xmath50 and the _ weighted volume _ of @xmath4 is defined by @xmath51 manifolds with the bakry - mery ricci curvature bounded below have been studied by many authors in recent years , particularly some interesting weighted volume estimates and splitting theorems , which can be found in @xcite and @xcite , for example .
some results about estimates of weighted volume can be also seen in @xcite and @xcite .
the aim of this paper is to give estimates for @xmath52 for complete noncompact manifolds in function of the weighted volume growth of their geodesic balls and spheres .
precisely , we have the following theorem . [ it4 ]
let @xmath42 be a complete noncompact weighted manifold . 1 . if @xmath53 , then @xmath54 where @xmath55 .
2 . if @xmath56 , then @xmath57 where @xmath58 furthermore , if @xmath59 with @xmath60 such that @xmath61 is a nonnegative smooth function satisfying @xmath62 then occurs equality in . here
@xmath63 denotes the standard metric on @xmath64-dimensional unit sphere @xmath65 and @xmath66 is the euclidian distance to origen .
in theorem 6.2 of @xcite , bessa , pigola , and setti gave an upper bounds for @xmath52 in terms of the volume growth .
indeed , they show that : 1 . if @xmath53 , then @xmath67 2 .
if @xmath56 , then @xmath68 therefore , theorem [ it4 ] gives an improved estimate for @xmath52 .
now , by supposing @xmath69 been a constant function , we have @xmath70 moreover , @xmath71 is the bottom of essential spectrum of the laplacian acting on @xmath72 .
therefore , as a consequence of theorem [ it4 ] , it follows : [ itc11 ] let @xmath4 be a noncompact complete manifold . 1 . if @xmath73 , then @xmath74 where @xmath75 .
2 . if @xmath76 , then @xmath77 where @xmath78 corollary [ itc11 ] was proved by higuchi , see corollary 2 of @xcite . let @xmath79 be a compact domain .
the bottom of the spectrum of @xmath24 on @xmath80 with the dirichlet boundary condition on @xmath81 admits the usual variational characterization @xmath82 where @xmath83 denotes the set of the smooth functions @xmath84 with compact support on @xmath80 .
moreover , it can be seen in theorem 6.1 of @xcite that @xmath85 where @xmath86 runs over the set of compact domains of @xmath4 .
let @xmath87 be the geodesic sphere of @xmath4 with center in a fixed point @xmath48 and radius @xmath49 .
the _ weighted volume _ of @xmath87 is given by @xmath88 where @xmath89 is the volume form on @xmath87 .
now , we are ready to enunciate the second result of this paper , namely : [ it1 ] let @xmath42 be a complete noncompact weighted manifold and let @xmath86 be a compact subset of @xmath4 . if there exists a positive constant real @xmath90 such that @xmath91 then @xmath92 consequently , @xmath93 furthermore , if @xmath94 ) with @xmath95 such that @xmath61 is a nonnegative smooth function satisfying @xmath96 then occurs equality in ( [ e10 ] ) and ( [ at11 ] ) for any compact @xmath97 here @xmath63 denotes the standard metric on @xmath64-dimensional unit sphere @xmath65 and @xmath66 is the euclidian distance to origen .
[ ri1 ] the condition @xmath98 implies that @xmath99 where @xmath100 .
but @xmath4 does not necessarily have @xmath101 look to @xmath94 ) with @xmath102 and @xmath103 such that @xmath61 is a nonnegative smooth function satisfying ; it has finite weighted volume , i.e. , @xmath104 and hence , @xmath105 this can be viewed in expression ( [ er1 ] ) of example [ ri3 ] . there is a class of complete riemannian manifolds that satisfies the conditions of theorem [ it1 ] ( see example [ ri3 ] ) .
now , by supposing @xmath20 been a constant function , it follows directly from theorem [ it1 ] that let @xmath4 be a noncompact complete manifold and let @xmath86 be a compact subset of @xmath4 . if @xmath106 then @xmath107 consequently , @xmath108 furthermore , if @xmath109 with @xmath103 such that @xmath61 is a nonnegative smooth function satisfying @xmath110 then occurs equality in ( [ aaa5 ] ) and ( [ aaa6 ] ) for any compact @xmath97 here @xmath63 denotes the standard metric on @xmath64-dimensional unit sphere @xmath65 and @xmath66 is the euclidian distance to origen . before beginning to enunciate some applications , we need to introduce some definitions that will be necessary to understand the results .
let @xmath111 be a weighted manifold , i.e. , @xmath112 let @xmath113 be an isometric immersion of a riemannian orientable manifold @xmath114 into weighted manifold @xmath115 the function @xmath116 restricted to @xmath4 induces a weighted measure @xmath19 on @xmath4 .
thus , we have an induced weighted manifold @xmath117 .
the _ second fundamental form _
@xmath118 of @xmath119 is defined by @xmath120 where @xmath121 symbolizes the projection above the normal bundle of @xmath4 .
the _ weighted mean curvature vector _ of @xmath4 is defined by @xmath122 with @xmath123 the hypersurface @xmath4 is called _
_ when its weighted mean curvature vector @xmath124 vanishes identically ; and when there exists real constant @xmath125 such that @xmath126 with @xmath127 being unit normal vector field , we say the hypersurface @xmath4 has _ constant weighted mean curvature_. the operator @xmath128 is called the _ @xmath20-stability operator _ of the immersion @xmath119 and it is associated with the quadratic form @xmath129 for each compact domain @xmath79 , define the index , @xmath130 , of @xmath131 in @xmath86 as the maximal dimension of a subspace of @xmath132 where @xmath133 is a negative definite .
the _ index _ , @xmath134 , of @xmath131 in @xmath4 ( or simply , the index of @xmath4 )
is then defined by @xmath135 where the supreme is taken over all compact domains @xmath136 for more details , see @xcite .
the following results are applications of the estimates of bottom essential spectrum of weighted laplacian operator given in theorem [ it4 ] and theorem [ it1 ] .
[ it5 ] let @xmath137 be an isometric immersion of a complete noncompact manifold @xmath114 into an oriented complete weighted manifold @xmath138 with unit normal vector field @xmath127 .
let @xmath139 if @xmath4 has constant weighted mean curvature @xmath140 and @xmath141 , then @xmath142 and @xmath143 for any constant @xmath144 , where @xmath145 if @xmath53 and @xmath146 if @xmath56 .
particularly , if @xmath147 for some constant positive @xmath148 , then @xmath4 is a @xmath20-minimal hypersurface . here @xmath149 denotes a generalization of the bakry- ' emery ricci curvature . as a consequence of the inequality ( [ aa12 ] ) of theorem [ it5 ]
, it follows [ aic1 ] let @xmath138 be an oriented complete weighted manifold with @xmath150 , where @xmath151 is a fixed constant .
then , there is no complete noncompact @xmath20-minimal hypersurface @xmath152 immersed into @xmath138 with @xmath153 and satisfying either @xmath154 if @xmath53 or @xmath155 if @xmath56 .
it is said that the weighted volume of @xmath4 has _ polynomial growth _ if there exists positive numbers @xmath90 , @xmath125 , and @xmath156 such that @xmath157 for any @xmath158 thus , @xmath159 i.e. , @xmath160 .
therefore , it follows from theorem [ it5 ] the following consequence : [ ic6 ] let @xmath137 be an isometric immersion of a complete noncompact manifold @xmath114 into an oriented complete weighted manifold @xmath138 with unit normal vector field @xmath127 .
assume that the weighted volume of @xmath4 is infinite and it has polynomial growth .
if @xmath4 has constant weighted mean curvature @xmath140 and @xmath141 , then there exists a constant @xmath161 such that for all @xmath162 , @xmath163 and @xmath164 where @xmath148 is a positive constant .
in particular , if @xmath165 for some real @xmath166 then @xmath4 is a @xmath20-minimal hypersurface .
observe also that if we assume @xmath167 , then @xmath168 therefore , as a consequence of theorem [ it5 ] , we obtain the following : [ ic7 ] let @xmath137 be an isometric immersion of a complete noncompact manifold @xmath114 into an oriented complete weighted manifold @xmath138 with unit normal vector field @xmath127 .
assume that the weighted volume of @xmath4 is infinite and it satisfies @xmath167 for some constant @xmath125 , @xmath90 , and @xmath169 .
if @xmath4 has constant weighted mean curvature @xmath140 and @xmath141 , then there exists a constant @xmath161 such that for all @xmath162 , @xmath170 and @xmath171 where @xmath148 is a positive constant . particularly , if @xmath172 for some real @xmath166 then @xmath4 is a @xmath20-minimal hypersurface . when @xmath20 is a constant function , corollary [ ic6 ] was obtained by alencar and do carmo , see theorem 1.1 of @xcite , and improved by do carmo and zhou , see theorem 4.1 of @xcite .
already , corollary [ ic7 ] was proved by do carmo and zhou , see theorem 4.4 of @xcite . by using the inequality ( [ e13 ] ) of the corollary [ ic6 ] and the inequality ( [ e122 ] ) of the corollary [ ic7 ]
, we can acquire the next two results : [ ic1 ] let @xmath138 be an oriented complete weighted manifold with @xmath150 , where @xmath151 is a fixed constant .
then , there is no complete noncompact @xmath20-minimal hypersurface @xmath152 immersed into @xmath138 with @xmath173 , @xmath153 , and polynomial growth weighted volume .
[ ic8 ] let @xmath138 be an oriented complete weighted manifold with @xmath150 , where @xmath151 is a fixed constant .
then there is no complete noncompact @xmath20-minimal hypersurface @xmath152 immersed into @xmath138 with @xmath153 , @xmath53 , and @xmath167 for any @xmath174 and @xmath175 .
now , using the estimate of @xmath176 viewed in theorem [ it1 ] , we obtain the application : [ it3 ] let @xmath137 be an isometric immersion of a complete noncompact manifold @xmath114 into an oriented complete weighted manifold @xmath138 with unit normal vector field @xmath127 .
assume that @xmath177 if @xmath4 has constant weighted mean curvature @xmath140 and @xmath141 , then there exists a constant @xmath161 such that for all @xmath162 , @xmath178 and @xmath179 particularly , if @xmath172 for some real @xmath166 then @xmath4 is a @xmath20-minimal hypersurface . as a consequence of the inequality ( [ e12 ] ) of theorem [ it3 ] , it follows [ ic2 ] let @xmath138 be an oriented complete weighted manifold with @xmath150 , where @xmath151 is a fixed constant . then , there is no complete noncompact @xmath20-minimal hypersurface @xmath152 immersed into @xmath138 with @xmath153 and @xmath180 for all @xmath181
let @xmath182 be a weighted manifold and let @xmath183 be a compact set .
for a positive number @xmath184 define the set @xmath185 with @xmath186 denoting the distance between @xmath119 and @xmath187 moreover , consider the amount @xmath188 where @xmath87 is the geodesic sphere of radius @xmath49 and center in a fixed point @xmath48 . established the above notations , we get the following : [ al1 ] let @xmath42 be a complete noncompact weighted manifold . for any fixed @xmath189 , we set @xmath190 with @xmath191 defined in . 1 . if @xmath53 , then @xmath192 besides that , + @xmath193 if @xmath194 2 .
if @xmath56 , then @xmath195 for arbitrary compact domain @xmath79 , let @xmath196 be the bottom of the spectrum of @xmath24 on @xmath80 .
it is well known that @xmath197 and , see @xcite , @xmath198 therefore it suffices to prove the following : for any fixed @xmath189 and compact domain @xmath79 , and for arbitrary sufficiently small @xmath199 there exists a function @xmath200 with compact support in @xmath201 such that @xmath202 where @xmath203 as @xmath204 .
let s establish a test function @xmath205 with compact support in @xmath201 . to this , let @xmath48 be a fixed point and let @xmath206 denote the distance from @xmath207 to @xmath208 .
we define @xmath209 , see figure [ fig1 ] , and @xmath210 , see figure [ fig2 ] , as follows : for @xmath66 sufficiently large such that @xmath211 , @xmath212 and for a fixed number @xmath213 , and for a positive integer @xmath214 , @xmath215 for @xmath216 , we have that @xmath217 and @xmath218 note that @xmath219 is supported in @xmath220 and @xmath221 . in addition , @xmath222 .
therefore , @xmath223 for @xmath214 and @xmath66 sufficiently large such that @xmath224 and @xmath225 for all @xmath226 , there exists a real constant @xmath125 independent of @xmath66 and @xmath214 such that @xmath227 \(i ) assume that @xmath53 .
then @xmath228 if @xmath229 it follows from the definition of @xmath230 that , for any @xmath231 there exists a sequence @xmath232 with @xmath233 for every @xmath234 such that @xmath235 by choosing @xmath236 , we obtain @xmath237 for all @xmath238 and @xmath239 therefore , by ( [ ee9 ] ) , ( [ ee10 ] ) , ( [ ee11 ] ) and ( [ ee13 ] ) , we can select @xmath234 and @xmath214 such that @xmath240 and @xmath241 for any @xmath242 where @xmath243 .
now , if @xmath244 , then , for any @xmath245 satisfying @xmath246 there exists a sequence @xmath232 such that @xmath247 setting @xmath248 and @xmath249 we have @xmath250 hence , for any @xmath251 we can select @xmath234 such that @xmath252 where @xmath253 \(ii ) now , assume @xmath56 . in this case ,
@xmath254 we can assume @xmath255 it follows from the definition of @xmath256 that , for any sufficiently small @xmath231 there exists a sequence @xmath232 such that @xmath257 and @xmath258 here , we can assume that this sequence @xmath232 satisfies @xmath259 .
choosing @xmath260 and @xmath261 , we obtain @xmath262 where @xmath263 resulting @xmath264 and @xmath265 in the same way as in the case of infinite weighted volume , selecting sufficiently large @xmath234 and @xmath214 , we obtain the desired estimate .
observe that @xmath266 and @xmath267 if @xmath268 .
thus , there exists the limits @xmath269 with @xmath230 defined in ( [ a13 ] ) .
[ al2 ] let @xmath42 be a complete noncompact weighted manifold .
then @xmath270 where @xmath271 if @xmath53 and @xmath272 if @xmath56 .
it follows direct from lemma [ al1 ] that @xmath273 if @xmath42 has infinite weighted volume and @xmath274 if @xmath42 has finite weighted volume .
now , if @xmath53 and @xmath275 then there exists @xmath189 such that @xmath194 using lemma [ al1 ] ( i ) , we obtain @xmath276 this concludes the proof of this lemma .
we are going to use lemma [ al1 ] and lemma [ al2 ] to prove theorem [ it4 ] that was enunciate in introduction .
_ of theorem [ it4 ] _
( i ) assume @xmath53 .
recall that @xmath277 since @xmath266 whenever @xmath278 then @xmath279 satisfies @xmath280 . for arbitrary fixed @xmath189 , we can select @xmath66 sufficiently large such that @xmath281 thus , @xmath282 and @xmath283 for all @xmath284 if @xmath285 then it follows from lemma [ al2 ] that @xmath286 moreover , @xmath193 if @xmath287 .
\(ii ) assume that @xmath56 .
recall that @xmath288 and @xmath289 since @xmath267 if @xmath278 then @xmath290 satisfies @xmath291 the second inequality of the above expression follows from the inequality @xmath292 let @xmath293 for any @xmath245 satisfying @xmath294 there exists @xmath295 such that @xmath296 for any @xmath297 then we get , for any @xmath298 @xmath299 thus , we have @xmath300 resulting in @xmath301 since we can select arbitrary small @xmath231 then @xmath302 therefore , by ( [ eee3 ] ) , @xmath303 besides on , @xmath304 if @xmath305 .
therefore , @xmath306 this concludes the part ( ii ) .
let @xmath59 with @xmath60 such that @xmath61 is a nonnegative smooth function satisfying @xmath307 note that @xmath308 where @xmath309 for all @xmath310 .
therefore , @xmath200 is a positive solution of @xmath311 for all @xmath312 since , see corollary 6.4 of @xcite , @xmath313 for any positive function , then @xmath314 for all @xmath312 using remark [ ri1 ] , we have that @xmath315 thus , it follows from ( [ e111 ] ) and ( [ 1111 ] ) , and the first part of theorem [ it4 ] that @xmath316 this concludes the proof of theorem [ it4 ] .
we are interested in estimating @xmath317 , where @xmath86 is an arbitrary compact subset of @xmath4 . for this
, we will look at the oscillatory behaviour of solutions of one second order ordinary differential equation .
we will need the following result obtained by fite ( see @xcite , theorem i ) : [ t5 ] assume that @xmath318 and @xmath319 are continuous functions of @xmath320 such that @xmath321 and @xmath322 for every @xmath323 , where @xmath90 and @xmath324 are constants such that @xmath325 , then every solution of differential equation @xmath326 changes sign an infinite number of times , i.e. , every solution of ( [ e5 ] ) is oscillatory .
the oscillation criteria in the integral form was firstly used by do carmo and zhou @xcite to obtain estimates of @xmath176 of the laplacian operator supposing polynomial or exponential growth of volume without weight .
other authors have also used this criterion to estimating the first eigenvalue of other elliptical operators assuming volume conditions , for example , see @xcite , @xcite , and @xcite . in this last reference , impera and rimoldi
have given an upper estimates of @xmath327 of the weighted laplacian operator supposing polynomial or exponential growth of weighted volume and infinite weighted volume .
now , we can use the previously oscillation result to prove theorem [ it1 ] .
_ of theorem [ it1 ] _ let @xmath328 denote the weighted area of geodesic sphere @xmath329 with center in @xmath48 and radius @xmath330 .
then , by coarea formula , we have @xmath331 choose @xmath332 such that @xmath333 .
put @xmath334 . since by hypothesis @xmath335 for all @xmath336
, we obtain that the differential equation @xmath337 satisfies the condition of theorem [ t5 ] .
therefore , for any @xmath38 such that @xmath338 the differential equation ( [ ee7 ] ) is oscillatory , i.e. , every solution @xmath339 , @xmath340 , of ( [ ee7 ] ) changes sign an infinite number of times .
fixed the initial values @xmath341 and @xmath342 , we have that any solution of ( [ ee7 ] ) can be extended to @xmath343 .
let @xmath344 a non - trivial oscillatory solution of ( [ ee7 ] ) with @xmath345 .
thus , there exists two numbers @xmath346 and @xmath347 in @xmath343 such that @xmath348 and @xmath349 , and @xmath350 for any @xmath351 .
write @xmath352 , @xmath353 and @xmath354 .
it follows by coarea formula that @xmath355 however , by equation ( [ ee7 ] ) , @xmath356 resulting @xmath357 since @xmath319 is an arbitrary positive constant larger than @xmath358 , then @xmath359 by implying , see ( [ 1111 ] ) , that @xmath360 now , let @xmath94 ) with @xmath361 such that @xmath61 is a nonnegative smooth function satisfying @xmath362 now , we consider the function @xmath363 .
note that @xmath364 where @xmath365 for all @xmath310 .
therefore , @xmath200 is a positive solution of @xmath366 for all @xmath312 since , see corollary 6.4 of @xcite , @xmath313 for any positive function , then @xmath367 therefore , @xmath368 for any compact @xmath97 using remark [ ri1 ] , we have that @xmath369 thus , it follows from theorem [ it1 ] that @xmath370 using inequalities ( [ e11 ] ) along with the above inequality and equality ( [ 1111 ] ) , we conclude that @xmath371 for any compact @xmath97 [ ri3 ] there is a class of complete riemannian manifolds that satisfies the conditions of theorem [ it1 ] .
in fact , let @xmath372 with weight @xmath373 and a smooth metric @xmath103 , where @xmath63 denotes the standard metric on @xmath64-dimensional unit sphere @xmath374 and @xmath61 is a nonnegative smooth function satisfying @xmath375 , @xmath376 , and @xmath377 for all @xmath312 thus , @xmath378 and @xmath379 where @xmath380 is @xmath381-dimensional volume of @xmath374 .
therefore , @xmath382 for all @xmath297
let @xmath111 be a weighted manifold , i.e. , @xmath112 let @xmath113 be an isometric immersion of a riemannian orientable manifold @xmath114 into weighted manifold @xmath115 the function @xmath116 restricted to @xmath4 induces a weighted measure @xmath19 on @xmath4 .
thus , we have an induced weighted manifold @xmath117 . the _ weighted mean curvature vector _ of @xmath4 is defined by @xmath383 with @xmath123 the hypersurface @xmath4 is called _
_ when its weighted mean curvature vector @xmath124 vanishes identically ; when there exists real constant @xmath125 such that @xmath126 with @xmath127 being unit normal vector field , we say the hypersurface @xmath4 has _ constant weighted mean curvature_. let @xmath384 , @xmath385 for all @xmath386 and @xmath387 be a variation of the immersion @xmath119 associated with the normal vector field @xmath388 , where @xmath389 the corresponding variation of the _ functional weighted area _
@xmath390 satisfies @xmath391 where @xmath140 is such that @xmath392 .
the expression ( [ ie1 ] ) is known as _ first variation formula_. in @xcite , mcgonagle and ross obtained the first variation formula to hypersurface @xmath393 i.e , in the case where @xmath394 ( see @xcite , ( 2.1 ) , pp .
to the general case , we can proceed in the same way as @xcite .
the @xmath20-minimal hypersurfaces are critical points of the functional weighted area . yet
, the hypersurfaces with constant weighted mean curvature can be viewed as critical points of the functional weighted area restricted to variations which preserve the _ enclose weighted volume , _
i.e. , to functions @xmath395 which satisfy the additional condition @xmath396 for such critical points , the _ second variation _ of the functional weighted area is given by @xmath397 where @xmath398 is the bakry- ' emery ricci curvature and @xmath118 is the second fundamental form . the second variation formula of functional weighted area , @xmath399 , can be viewed in @xcite , lemma 2.3 , to isometric immersion of hypersurface @xmath114 into @xmath400 with constant weighted mean curvature .
the case which @xmath401 is any weighted manifold , it follows from analogue arguments shown in the proof of lemma 2.3 , @xcite . when @xmath20 is a constant function , the first and second variation formula were given by barbosa and do carmo @xcite and barbosa , do carmo and eschenburg @xcite .
the operator @xmath128 is called the _ @xmath20-stability operator _ of the immersion @xmath119 . in the @xmath20-minimal case
, the @xmath20-stability operator is viewed as acting on @xmath402 ; in the case of the hypersurfaces with constant weighted mean curvature , the @xmath20-stability operator is viewed as acting on @xmath403 associated with @xmath131 is the quadratic form @xmath129 for each compact domain @xmath79 , define the index , @xmath404 , of @xmath131 in @xmath86 as the maximal dimension of a subspace of @xmath405 where @xmath133 is a negative definite .
the _ index _
, @xmath406 , of @xmath131 in @xmath4 ( or simply , the index of @xmath4 ) is then defined by @xmath407 where the supreme is taken over all compact domains @xmath136 for more details , see @xcite . in the case of constant weighted mean curvature immersion ,
if we analyze the index of @xmath131 on @xmath4 , with @xmath131 acting on @xmath408 instant of @xmath409 , we denote the _ stronger index _ of @xmath131 on @xmath4 by @xmath134 . however , it is easy to verify that @xmath410 is equivalent to @xmath411 so in the statement of theorem [ it5 ] and theorem [ it3 ] along with their respective corollaries , they are immaterial whether one takes @xmath412 or @xmath413 _ of theorem [ it5 ] _ since @xmath141 by hypothesis , then using proposition 5 of @xcite , there exists a compact set @xmath86 and a positive function @xmath200 on @xmath4 such that @xmath420 on @xmath201
. let @xmath48 and let @xmath161 be such that @xmath421 therefore , it follows from equality ( [ 1111 ] ) and theorem [ it4 ] , @xmath422 where @xmath145 if @xmath53 and @xmath146 if @xmath56 .
now , by using corollary 6.4 of @xcite and the previously inequality , we obtain @xmath423 because @xmath424 .
let @xmath144 and since @xmath425 then it follows from lemma [ l1 ] that @xmath426 and @xmath427 resulting in @xmath428 and @xmath429 therefore , @xmath430 and @xmath431 the second part of this theorem is immediate , just use the previous inequality by doing @xmath432 _ of theorem [ it3 ] _ since @xmath141 by hypothesis , then using proposition 5 of @xcite , there exists a compact set @xmath86 and a positive function @xmath200 on @xmath4 such that @xmath420 on @xmath201 .
let @xmath48 and let @xmath161 be such that @xmath421 therefore , it follows from theorem [ it1 ] that @xmath433 now , by using corollary 6.4 of @xcite and previously inequality , we obtain @xmath434 because @xmath424 .
let @xmath144 and since @xmath425 then it follows from lemma [ l1 ] that @xmath426 and @xmath427 resulting in @xmath435 and @xmath436 therefore , @xmath437 and @xmath179 the second part of this theorem is immediate , just use the previous inequality by doing @xmath438 | we obtain upper estimates for the bottom ( that is , greatest lower bound ) of the essential spectrum of weighted laplacian operator of a noncompact weighted manifold under assumptions of the volume growth of their geodesic balls and spheres .
furthermore , we find examples where the equality occurs in the estimates obtained . as a consequence ,
we give estimates for the weighted mean curvature of complete noncompact hypersurfaces into weighted manifolds .
+ * keywords : * essential spectrum ; weighted laplacian ; volume growth ; estimate ; mean curvature ; index . |
the study of open charm production in high - energy lepton - hadron and hadron - hadron collisions is considered as a test of the general applicability of perturbative quantum chromodynamics ( qcd ) and also provides information on the parton distribution functions ( pdfs ) of protons and photons .
the present analysis is to explore our potential to access a new dynamical regime , namely the high - energy regge limit , which is characterized by the condition @xmath2 , where @xmath3 is the total collision energy in the center - of - mass ( cm ) reference frame , @xmath4 is the asymptotic scale parameter of qcd , and @xmath5 is the typical energy scale of the hard interaction . in the processes of heavy - quark production , one has @xmath6 , where @xmath7 is the heavy - quark mass . in this high - energy regime ,
the contribution from partonic subprocesses involving @xmath8-channel parton ( quark or gluon ) exchanges to the production cross section can become dominant .
thus , the transverse momenta of the incoming partons and their off - shell properties can no longer be neglected , and we deal with _
reggeized _ @xmath8-channel partons .
the quasi - multi - regge - kinematics ( qmrk ) approach @xcite is particularly appropriate for this kind of high - energy phenomenology .
it is based on an effective quantum field theory implemented with the non - abelian gauge - invariant action , as suggested a few years ago @xcite .
our previous analyses of charmonium and bottomonium production at the fermilab tevatron @xcite demonstrated the advantages of the high - energy factorization scheme over the collinear parton model as far as the description of experimental data is concerned .
these observations were substantiated for @xmath9-meson production at the tevatron in ref .
@xcite , where the experimental data were again well described using the fadin - lipatov effective reggeon - reggeon - gluon vertex @xcite . in ref .
@xcite , where the effective photon - reggeon - quark vertex was obtained and for the first time , the hypothesis of quark reggeization was successfully used to describe experimental data on single prompt - photon production and on the proton structure functions @xmath10 and @xmath11 .
the cdf collaboration measured the differential cross sections @xmath12 for the inclusive production of @xmath13 , @xmath14 , @xmath15 , and @xmath16 mesons in @xmath17 collisions in run ii at the tevatron as functions of transverse momentum ( @xmath18 ) in the central rapidity ( @xmath19 ) region @xcite .
these measurements were compared with theoretical predictions obtained at next - to - leading order ( nlo ) in the collinear parton model of qcd @xcite taking into account quark and hadron mass effects , and it was found that the latter improve the description of the data . the differential cross sections @xmath12 and @xmath20 for inclusive @xmath21 and @xmath22 photoproduction measured by the h1 @xcite and zeus @xcite collaborations at the desy hera collider were compared with nlo predictions in the collinear parton model . for @xmath21 mesons , this was done in three approaches : the zero - mass variable - flavor - number scheme ( zm - vfns ) @xcite , the fixed - flavor - number scheme ( ffns ) @xcite , and the general - mass variable - flavor - number scheme ( gm - vfns ) @xcite .
the experimental results were found to generally lie above the nlo expectations . for @xmath22 mesons ,
the calculations were performed in the ffns @xcite and in the model suggested by berezhnoy , kiselev , and likhoded @xcite . in this paper , we study @xmath1-meson production under hera and tevatron experimental conditions as well as the charm structure function @xmath0 of the proton for the first time in the framework of the qmrk approach @xcite complemented with the quark - reggeization hypothesis . this paper is organized as follows . in sec .
[ sec : two ] , we present the basic formalism of our calculations and briefly recall the qmrk approach in connection with the quark - reggeization hypothesis . in sec .
[ sec : three ] , we consider the charm structure function @xmath0 of the proton and compare our results with experimental data . in secs .
[ sec : four ] and [ sec : five ] , we describe @xmath1-meson production via @xmath23-quark fragmentation at hera and the tevatron , respectively . in sec .
[ sec : six ] , we summarize our conclusions .
in the phenomenology of the strong interactions at high energies , it is necessary to describe the qcd evolution of the pdfs of the colliding particles ( hadrons or photons ) starting from some scale @xmath24 which controls a non - perturbative regime up to the typical scale @xmath5 of the hard - scattering processes , which is typically of the order of the transverse mass @xmath25 of the produced particle ( or hadron jet ) with ( invariant ) mass @xmath26 and transverse momentum @xmath27 . in the region of very high energies , which corresponds to the so - called regge limit , the typical ratio @xmath28 becomes very small , @xmath29 .
this leads to large logarithmic contributions of the type @xmath30^n$ ] , where @xmath31 is the strong - coupling constant , which are conveniently resummed in the balitsky - fadin - kuraev - lipatov @xcite formalism by the evolution of unintegrated gluon and quark distribution functions @xmath32 , where @xmath33 and @xmath34 are the longitudinal - momentum fraction and transverse momentum of the reggeized parton w.r.t . the parent particle , respectively .
correspondingly , in the qmrk approach @xcite , the initial - state @xmath8-channel gluons and quarks are considered as reggeons , or reggeized gluons ( @xmath35 ) and quarks ( @xmath36 ) .
they carry finite transverse momenta @xmath34 with respect to the hadron or photon beam from which they stem and are off mass shell .
the advantages of the qmrk approach in comparison with the conventional @xmath37-factorization scheme @xcite include : firstly , it uses gauge - invariant amplitudes and is based on a factorization hypothesis that is proven in the leading logarithmic approximation ; secondly , it carries over to non - leading orders in the strong - coupling constant , as recently proven @xcite .
the reggeization of amplitudes provides the opportunity to efficiently take into account large radiative corrections to processes in the regge limit beyond what is included in the collinear approximation , which is of great practical importance .
recently , the feynman rules for the induced and some important effective vertices of the effective theory based on the non - abelian gauge - invariant action @xcite have been derived in ref .
however , these rules only refer to processes with reggeized gluons in the initial state . as for @xmath8-channel quark - exchange processes ,
such rules are still unknown , so that it is necessary to construct effective vertices involving reggeized quarks using qmrk approach prescriptions in each application from first principles .
of course , a certain set of reggeon - reggeon - particle effective vertices are known , for example for the transitions @xmath38 @xcite , @xmath39 @xcite , and @xmath40 @xcite .
the effective @xmath41 vertex , which describes the production of a quark in the collision of a virtual photon with a reggeized quark , has been recently obtained in ref .
@xcite . in our numerical calculations below
, we adopt the prescription proposed by kimber , martin , ryskin , and watt @xcite to obtain unintegrated gluon and quark distribution functions for the proton from the conventional integrated ones , as implemented in watt s code @xcite . to obtain the analogous unintegrated functions for the photon , we modify watt s code @xcite . as input for this procedure ,
we use the martin - roberts - stirlin - thorne @xcite proton and the glck - reya - vogt @xcite photon pdfs .
on the experimental side , the charm structure function @xmath0 of the proton was measured by h1 @xcite and zeus @xcite in deep inelastic scattering ( dis ) of electrons and positrons on protons at hera . in this section
, we consider this quantity in the framework of the qmrk approach endowed with the quark - reggeization hypothesis .
we thus need the partonic cross section for the production of a @xmath23 quark in the collision of a virtual photon and a reggeized charm quark .
the relevant vertex was found in ref .
@xcite and reads : @xmath42,\ ] ] where the four - momenta of the virtual photon , the proton , the reggeized charm quark , and the outgoing charm quark are denoted as @xmath43 , @xmath44 , @xmath45 , and @xmath46 , respectively .
we concentrate on photons with large virtuality @xmath47 , so that the massless approximation for describing dis structure functions is appropriate @xcite .
we then obtain the following master formula for @xmath0 : @xmath48 where @xmath49 is the fractional electric charge of the @xmath23 quark and @xmath50 , with @xmath51 being the bjorken variable .
for definiteness , we choose the factorization scale to be @xmath52 . in fig .
[ fig1 ] , we compare the @xmath51 distributions of @xmath0 for various values of @xmath53 with the h1 @xcite and zeus @xcite data .
we find good agreement for all values of @xmath53 , except for the highest one , @xmath54
gev@xmath55 , where our prediction is about @xmath56 below the data .
this disagreement shows the importance of higher - order corrections at large values of @xmath53 , which are beyond the scope of our present study .
on the experimental side , zeus measured the @xmath57 distributions of @xmath21 @xcite and @xmath16 @xcite mesons with rapidity @xmath58 inclusively produced in photoproduction at hera i , with proton energy @xmath59 gev and lepton energy @xmath60 gev in the laboratory frame , in the ranges @xmath61 gev and @xmath62 gev , respectively . in this section , we compare this data with our qmrk predictions . at leading order ( lo )
, we need to consider only three @xmath63 partonic subprocesses , namely @xmath64 for direct photoproduction and @xmath65 and @xmath66 for resolved photoproduction , where the subscript indicates the mother particle . exploiting the factorization theorem
, the @xmath57 distribution of direct photoproduction takes the form @xmath67 where @xmath68 with @xmath69 being the transverse momentum of the produced @xmath23 quark .
we evaluate the quasi - real - photon flux @xmath70 in weizscker - williams approximation using @xmath71,\ ] ] where @xmath72 is sommerfeld s fine - structure constant , @xmath73 is the electron mass , @xmath74 , and @xmath75 is determined by the experimental setup , with @xmath76 gev@xmath55 in our case @xcite . as for the @xmath77 fragmentation function
( ff ) @xmath78 , we adopt the non - perturbative @xmath21 and @xmath16 sets determined in the zm - vfns with initial evolution scale @xmath79 @xcite from fits to opal data from cern lep1 .
we choose the renormalization and initial- and final - state factorization scales to be @xmath80 , where @xmath81 is the @xmath1-meson mass . using the reggeized - quark photon effective vertex from ref .
@xcite , the square of the hard - scattering amplitude is found to be @xmath82 it is understood that also the contribution from the charge - conjugate partonic subprocess is to be included in eq .
( [ eq : dir ] ) . in the case of resolved photoproduction via the partonic subprocess @xmath65 , the factorization formula reads : @xmath83 where @xmath84 with @xmath85 being the angle enclosed between @xmath27 and @xmath86 . using the reggeized - quark reggeized - gluon effective vertex from ref .
@xcite , we have @xmath87 again , the charge - conjugate partonic subprocess is to be included in eq .
( [ eq : res ] ) . resolved photoproduction via the partonic subprocess @xmath66
is treated very similarly . in figs .
[ figdhera](a ) and ( b ) , our results for @xmath21 and @xmath16 mesons , respectively , are broken down to the @xmath64 , @xmath65 , and @xmath66 contributions and are compared with the zeus data @xcite .
we find that the theoretical predictions are dominated by direct photoproduction and agree rather well with the experimental data over the whole @xmath57 range considered .
cdf @xcite measured the @xmath57 distributions of @xmath13 , @xmath14 , @xmath21 , and @xmath16 mesons with rapidity @xmath88 inclusively produced in hadroproduction in run ii at the tevatron , with @xmath89 tev . to lo in the qmrk approach ,
the factorization formula for the @xmath90 channel reads : @xmath91 where @xmath92 is given by eq .
( [ eq : ms ] ) , @xmath93 the result for the @xmath94 channel is similar and has to be included in eq .
( [ eq : had ] ) together with those from the charge - conjugate partonic subprocesses . in figs .
[ figdtevn](a)(d ) , our results for @xmath13 , @xmath14 , @xmath21 , and @xmath16 mesons , respectively , are compared with the cdf data @xcite .
we find that the theoretical predictions generally agree rather well with the experimental data , except perhaps for the slope .
in fact , the predictions exhibit a slight tendency to undershoot the data at small values of @xmath57 and to overshoot them at large values of @xmath57 . however , we have to bear in mind that these are just lo predictions , so that there is room for improvement by including higher orders . in the framework of the collinear parton model , comparisons with the experimental data of ref .
@xcite were performed beyond lo , namely in the fixed - order - next - to - leading - logarithm ( fonll ) scheme @xcite and at nlo in the gm - vfns @xcite .
the fonll predictions systematically undershoot the cdf data @xcite .
the gm - vfns predictions of ref .
@xcite , which are evaluated with ffs determined in the zm - vfns @xcite , describe these data within their errors , but are still somewhat on the low side .
the degree of agreement is further improved @xcite by evaluating the gm - vfns predictions of ref .
@xcite using ffs extracted @xcite from a global fit to @xmath9- and @xmath95-factory data of @xmath96 annihilation in the very same scheme .
in this paper , we explored the usefulness of the quark - reggeization hypothesis in the framework of the qmrk approach by studying several observables of inclusive charm production at lo , namely the charm structure function @xmath0 of the proton measured at hera @xcite as well as the one - particle - inclusive cross sections of @xmath21 and @xmath16 photoproduction in @xmath97 collisions at hera @xcite and of @xmath13 , @xmath14 , @xmath21 , and @xmath16 hadroproduction in @xmath17 collisions at the tevatron @xcite . in all three cases , we found satisfactory agreement between our default predictions and the experimental data , which is quite encouraging in view of the simplicity of our lo expressions for the partonic cross sections .
by contrast , in the collinear parton model of qcd , the inclusion of nlo corrections is necessary to achieve such a degree of agreement .
we thus recover the notion that the qmrk approach is a powerful tool for the theoretical description of qcd processes in the high - energy limit and automatically accommodates an important class of corrections that lie beyond the reach of the collinear parton model at lo @xcite .
we thank l. n. lipatov and o. v. teryaev for useful discussions .
a. v. s. is grateful to the international center of fundamental physics in moscow and to the dynastiya foundation for financial support .
this work was supported in part by the german federal ministry for education and research bmbf through grant no .
05 ht6gua and by the german research foundation dfg through grant no .
kn 365/71 v. a. saleev and d. v. vasin , phys .
d * 68 * , 114013 ( 2003 ) [ arxiv : hep - ph/0304114 ] ; phys .
nucl . * 68 * , 94 ( 2005 ) [ yad .
fiz . * 68 * , 95 ( 2005 ) ] ; phys .
nucl . * 38 * , 635 ( 2007 ) ; b. a. kniehl , d. v. vasin , and v. a. saleev , phys . rev .
d * 73 * , 074022 ( 2006 ) [ arxiv : hep - ph/0602179 ] ; b. a. kniehl , v. a. saleev , and d. v. vasin , _ ibid . _ * 74 * , 014024 ( 2006 ) [ arxiv : hep - ph/0607254 ] .
b. a. kniehl , g. kramer , i. schienbein , and h. spiesberger , phys .
d * 71 * , 014018 ( 2005 ) [ arxiv : hep - ph/0410289 ] ; eur .
j. c * 41 * , 199 ( 2005 ) [ arxiv : hep - ph/0502194 ] ; phys .
lett . * 96 * , 012001 ( 2006 ) [ arxiv : hep - ph/0508129 ] ; phys .
d * 77 * , 014011 ( 2008 ) [ arxiv:0705.4392 [ hep - ph ] ] .
j. binnewies , b. a. kniehl , and g. kramer , z. phys .
c * 76 * , 677 ( 1997 ) [ arxiv : hep - ph/9702408 ] ; b. a. kniehl , g. kramer , and m. spira , _ ibid . _ * 76 * , 689 ( 1997 ) [ arxiv : hep - ph/9610267 ] ; j. binnewies , b. a. kniehl , and g. kramer , phys . rev
. d * 58 * , 014014 ( 1998 ) [ arxiv : hep - ph/9712482 ] .
s. frixione , m. l. mangano , p. nason , and g. ridolfi , phys .
b * 348 * , 633 ( 1995 ) [ arxiv : hep - ph/9412348 ] ; s. frixione , p. nason , and g. ridolfi , nucl .
* b454 * , 3 ( 1995 ) [ arxiv : hep - ph/9506226 ] .
e. a. kuraev , l. n. lipatov , and v. s. fadin , sov .
jetp * 44 * , 443 ( 1976 ) [ zh . eksp .
* 71 * , 840 ( 1976 ) ] ; i. i. balitsky and l. n. lipatov , sov .
j. nucl .
* 28 * , 822 ( 1978 ) [ yad .
* 28 * , 1597 ( 1978 ) ] .
l. v. gribov , e. m. levin , and m. g. ryskin , phys .
* 100 * , 1 ( 1983 ) ; j. c. collins and r. k. ellis , nucl .
* b360 * , 3 ( 1991 ) ; s. catani , m. ciafaloni , and f. hautmann , _ ibid .
_ * b366 * , 135 ( 1991 ) .
v. s. fadin , m. i. kotsky , and l. n. lipatov , phys .
b * 415 * , 97 ( 1997 ) ; d. ostrovsky , phys . rev .
d * 62 * , 054028 ( 2000 ) [ arxiv : hep - ph/9912258 ] ; v. s. fadin , m. g. kozlov , and a. v. reznichenko , phys .
nucl . * 67 * , 359 ( 2004 ) [ yad .
fiz . * 67 * , 377 ( 2004 ) ] [ arxiv : hep - ph/0302224 ] ; j. bartels , a. sabio vera , and f. schwennsen , jhep * 0611 * , 051 ( 2006 ) [ arxiv : hep - ph/0608154 ] .
m. a. kimber , a. d. martin , and m. g. ryskin , phys .
d * 63 * , 114027 ( 2001 ) [ arxiv : hep - ph/0101348 ] ; g. watt , a. d. martin , and m. g. ryskin , eur .
j. c * 31 * , 73 ( 2003 ) [ arxiv : hep - ph/0306169 ] ; phys .
d * 70 * , 014012 ( 2004 ) ; * 70 * , 079902(e ) ( 2004 ) [ arxiv : hep - ph/0309096 ] .
j. binnewies , b. a. kniehl , and g. kramer , phys .
d * 58 * , 014014 ( 1998 ) [ arxiv : hep - ph/9712482 ] ; b. a. kniehl and g. kramer , _ ibid . _ * 71 * , 094013 ( 2005 ) [ arxiv : hep - ph/0504058 ] ; * 74 * , 037502 ( 2006 ) [ arxiv : hep - ph/0607306 ] .
b. a. kniehl , in _ proceedings of the xvi international workshop on deep - inelastic scattering and related subjects ( dis 2008 ) _ , london , england , 2008 , edited by r. devenish and j. ferrando ( science wise publishing , 2008 ) , ` http://dx.doi.org/10.3360/dis.2008.195 ` [ arxiv:0807.2215 [ hep - ph ] ] . | we study open charm production at high energies in the framework of the quasi - multi - regge - kinematics approach applying the quark - reggeization hypothesis implemented with reggeon - reggeon - particle and reggeon - particle - particle effective vertices . adopting the kimber - martin - ryskin unintegrated quark and gluon distribution functions of the proton and photon
, we thus nicely describe the proton structure function @xmath0 measured at desy hera as well as the transverse - momentum distributions of @xmath1 mesons created by photoproduction at hera and by hadroproduction at the fermilab tevatron . |
an rna sequence is a linear , oriented sequence of the nucleotides ( bases ) * a , u , g , c*. these sequences `` fold '' by establishing bonds between pairs of nucleotides .
bonds can not form arbitrarily : a nucleotide can at most establish one watson - crick base pair * a - u * or * g - c * or a wobble base pair * u - g * , and the global conformation of an rna molecule is determined by topological constraints encoded at the level of secondary structure , i.e. , by the mutual arrangements of the base pairs @xcite .
secondary structures can be interpreted as ( partial ) matchings in a graph of permissible base pairs @xcite .
they can be represented as diagrams , i.e. graphs over the vertices @xmath8 , drawn on a horizontal line with bonds ( arcs ) in the upper halfplane . in this representation one
refers to a secondary structure without crossing arcs as a _ simple _ secondary structure and pseudoknot structure , otherwise , see figure [ f : rnap ] .
folded configurations are energetically somewhat optimal . here
energy means free energy , which is dominated by the loops forming between adjacent base pairs and not by the hydrogen bonds of the individual base pairs @xcite .
in addition sterical constraints imply certain minimum arc - length conditions for minimum free energy configurations @xcite .
in particular , only configurations without isolated bonds and without bonds of length one ( formed by immediately subsequent nucleotides ) are observed in rna structures . in this paper , optimize a problem we meas maximize the score but not to minimize the free energy . for a given rna sequence polynomial - time dynamic programming ( dp ) algorithms
can be devised , finding such minimal energy configurations .
the most commonly used tools predicting simple rna secondary structure ` mfold ` @xcite and the ` vienna rna package ` @xcite , are running at @xmath9 space and @xmath10 time solution . in the following we omit `` simple '' and refer to secondary structures containing crossing arcs as pseudoknot structures . generalizing the matrices of the dp - routines of secondary structure folding @xcite to gap - matrices @xcite ,
leads to a dp - folding of pseudoknotted structures @xcite ( ` pknot - r&e ` ) with @xmath11 space an @xmath12 time complexity .
the following references provide a certainly incomplete list of dp - approaches to rna pseudoknot structure prediction using various structure classes characterized in terms of recursion equations and/or stochastic grammars : @xcite .
the most efficient algorithm for pseudoknot structures is @xcite ( ` pknotsrg ` ) having @xmath13 space and @xmath11 time complexity .
this algorithm however considers only a few types of pseudoknots .
rna secondary structures are exactly structures of topological genus zero @xcite .
the topological classification of rna structures @xcite has recently been translated into an efficient dp algorithm @xcite . fixing
the topological genus of rna structures implies that there are only finitely many types , the so called irreducible shadows @xcite .
sparsification is a method tailored to speed up dp - algorithms predicting mfe - secondary structures @xcite .
the idea is to prune certain computation paths encountered in the dp - recursions , see figure [ f : spar_sec ] . to make the key point ,
let us consider the case of rna secondary structure folding . here
sparsification reduces the dp - recursion paths to be based on so called candidates .
a candidate is in this case an interval , for which the optimal solution can not be written as a sum of optimal solutions of sub - intervals , see figure [ f : cand ] .
tracing back these candidates gives rise to `` irreducible '' structures and the crucial observation is here that these irreducibles appear only at a low rate .
this means that there are only relatively few candidates , which in turn implies a significant reduction in time and space complexity .
sparsification has been applied in the context of rna - rna interaction structures @xcite as well as rna pseudoknot structures @xcite . in difference to rna secondary structures , however , not every decomposition rule in the dp - folding of rna pseudoknot structures is amendable to sparsification . by construction ,
sparsification can only be applied for calculating mfe - energy structures . since the computation of the partition function
@xcite needs to take into account _ all _ sub - structures , sparsification does not work .
is obtained from the optimal solutions @xmath14 , @xmath15 and @xmath16 . based on the recursions of the secondary structures ,
@xmath14 and @xmath15 produce an optimal solution of @xmath17 .
similarly , @xmath15 and @xmath16 produce an optimal solution of @xmath18 .
now , in order to obtain an optimal solution of @xmath19 it is sufficient to consider either the grouping @xmath17 and @xmath16 or @xmath14 and @xmath18 . ]
is irreducible .
( c ) is another computation path with distinct leftmost irreducible over a different interval , hence representing a new candidate that can not be reduced to ( a ) by the sparsification . ] for the mfe - folding of rna secondary structures considerable attention has been paid in order to validate that the set of candidates is small .
the idea here is that irreducibles are contained in short , `` rainbow''-like arcs . to be precise ,
the gain is @xmath20 , if secondary structure satisfy the so called _ polymer - zeta property _
@xcite : the latter quantifies the probability of an arc of length @xmath4 to be @xmath21 , where @xmath22 , @xmath23 .
note that these arcs confine in case of secondary structures irreducible structures , that is arcs and irreducibility are tightly connected . in pseudoknotted rna structures
however , we have crossing arcs and the associated notion of irreducible structures differs significantly from that of rna secondary structures .
the polymer - zeta property is theoretically justified by means of modeling the 2d folding of a polymer chain as a self - avoiding walk ( saw ) in a 2d lattice @xcite .
more evidence of the polymer - zeta property for rna secondary structures has been collected via the ncbi database @xcite of mfe - rna structures . in this paper
we study the sparsification of the decomposition rule @xmath0 that splices an interval @xcite in the context of the dp - folding of rna pseudoknot structures of fixed topological genus .
our paper provides a combinatorial framework to quantify the effects of sparsifying the @xmath0-decomposition rule .
we shall prove that the candidate set @xcite is indeed small .
our argument is based on assuming a specific distribution of irreducible structures within mfe - structures .
namely we assume these irreducibles to appear with probability @xmath24 , where we assume @xmath25 to be a fixed parameter and @xmath26 to be a bivariate ( energy - filtered ) generating function whose associated generation function of irreducibles is @xmath27 .
while this energy - filtration seems to be reparameterization of the notion of `` stickiness '' @xcite , it is really fundamentally different .
this becomes clear when considering loop - based energies which distinguishes energy and arcs .
clearly when folding random sequences one weights the latter around @xmath28 , reminiscent of the probability of two given positions to be compatible .
the energy however is fairly independent as it really depends on the particular loop - type .
we obtain these energy - filtered gfs also for rna pseudoknot structures of fixed topological genus .
this provides new insights into the improvements of the sparsification of the concatenation - rule @xmath0 in the presence of cross serial interactions .
our observations complement the detailed analysis of backofen @xcite .
we show that although for pseudoknot structures of fixed topological genus @xcite the effect of sparsification on the global time complexity is still unclear , the decomposition rule that splits an interval can be sped up significantly .
the general idea of sparsification @xcite is following : let @xmath29 be a set whose elements @xmath30 are unions of pairwise disjoint intervals . let furthermore @xmath31 denote an optimal solution ( a positive number or score ) of the dp - routine over @xmath32
. by assumption @xmath31 is recursively obtained .
suppose the optimal solution @xmath31 is given by @xmath33 , where @xmath34 .
then , under certain circumstances , the dp - routine may interpret @xmath31 either as @xmath35 or as @xmath36 , see figure [ f : spar_idea ] . to be precise , this situation is encountered iff * there exists an optimal solution @xmath37 for a sub - structure over @xmath38 where @xmath39 via @xmath40 and @xmath31 is obtained from @xmath37 and @xmath41 via @xmath42 , * there exists an optimal solution @xmath43 for a sub - structure over @xmath44 where @xmath45 via @xmath46 and @xmath31 is obtained by @xmath47 and @xmath43 via @xmath42
. given a decomposition @xmath48 we call @xmath40 @xmath49-compatible to @xmath42 if there exists a decomposition rule @xmath46 such that @xmath50 note that if @xmath40 is @xmath49-compatible to @xmath42 then @xmath46 is @xmath49-compatible to @xmath42 . to summarize * ( @xmath49-compatible ) * suppose @xmath31 is the optimal solution for @xmath51 over @xmath32 , @xmath52 under decomposition rule @xmath42 .
@xmath47 is obtained from two optimal solutions @xmath47 and @xmath53 under rule @xmath40 .
then @xmath40 is called _
@xmath49-compatible _ to @xmath42 if there exist some rule @xmath46 such that @xmath54 and @xmath55 .
is alternatively realized via @xmath47 and @xmath43 , or @xmath37 and @xmath41 .
thus it is sufficient to only consider one of the computation paths . ]
figure [ f : spar_idea ] depicts two such ways that realize the same optimal solution @xmath31 .
sparsification prunes any such multiple computations of the same optimal value .
we next come to the important concept of candidates .
the latter mark the essential computation paths for the dp - routine . * ( candidates ) * suppose @xmath31 is an optimal solution .
we call @xmath32 is a @xmath56-_candidate _ if for any @xmath57 obtained by @xmath56 and @xmath58 , we have @xmath59 and we shall denote the set of @xmath56-candidates set by @xmath60 .
[ l:11]@xcite suppose @xmath40 is @xmath49-compatible to @xmath42 then any optimal solution @xmath31 can be obtained via @xmath40-candidates . by construction
a @xmath40-candidate @xmath32 is a union of disjoint intervals such that its optimal solution @xmath31 can not be obtained via a @xmath40-splitting .
this optimal solution allows to construct a non - unique arc - configuration ( sub - structure ) over @xmath32 @xcite and the above @xmath40-splitting consequently translates into a splitting of this sub - structure .
this connects the notion of @xmath40-candidates with that of sub - structures and shows that a @xmath40-candidate implies an sub - structure that is @xmath40-irreducible . in the case of sparsification of rna
secondary structures we have one basic decomposition rule @xmath0 acting on intervals , namely @xmath0 splices an interval into two disjoint , subsequent intervals . the implied notion of a @xmath0-irreducible sub - structure is that of a sub - structure nested in an maximal arc , where maximal refers to the partial order @xmath61 iff @xmath62 .
this observation relates irreducibility to that or arcs and following this line of thought @xcite identifies a specific property of polymer - chains introduced in @xcite to be of relevance for the size of candidate sets : * ( polymer - zeta property)*[d : polymer - zeta ] let @xmath63 denotes the probability of a structure over an interval @xmath64 $ ] under some decomposition rule @xmath56 .
then we say @xmath56 follows the polymer - zeta property if @xmath65 for some constant @xmath66 .
this property is theoretically justified by means of modeling the 2d folding of a polymer chain as a self - avoiding walk ( saw ) in a 2d lattice @xcite . in this section
we recall some results of @xcite on the sparsification of rna secondary structures .
secondary structure satisfies a simple recursion which gives the optimal solution over @xmath64 $ ] by @xmath67 , where @xmath68 denotes the optimal solution in which @xmath69 is a base pair , and @xmath70 denotes the optimal solution obtained by adding the optimal solutions of two subsequent intervals , respectively .
note that the optimal solution over a single vertex is denoted by @xmath71 .
we have the recursion equation for @xmath68 and @xmath70 : @xmath72 where @xmath73 is the score when @xmath69 form a base pair , see figure .
[ f : sec ] . in case
two positions , @xmath74,@xmath75 in the sequence are incompatible then we have @xmath76 . an interval @xmath64 $ ] is a @xmath0-candidate if the optimal solution over @xmath64 $ ] is given by @xmath77 .
indeed , @xmath64 $ ] is a candidate iff @xmath64 $ ] is in the candidate set of @xmath0 , and we denote the set @xmath78 by @xmath79 .
suppose the optimal solution @xmath70 is given by @xmath80 and suppose we have @xmath81 .
then since @xmath82 $ ] is not a candidate , lemma [ l:11 ] shows that we can compute @xmath83 , where @xmath84 $ ] is a candidate .
accordingly , the recursion for @xmath70 can be based on candidates , i.e. @xmath85\in q } \{l_{i , k}+l_{k+1,j}\}$ ] .
clearly , the bottleneck for computing the recursion is the calculation of @xmath70 , which requires @xmath86 time . applying sparsification
, this recursion is based on candidates @xmath84 $ ] .
suppose we have @xmath87 such candidates , then the time complexity reduces to @xmath88 , since the optimal solution is necessarily based on a candidate .
once the latter is identified the expression @xmath18 requires only @xmath20 time complexity . in the worst case
, @xmath79 contains @xmath13 elements .
the polymer - zeta property however implies that the expectation of @xmath87 is given by @xmath89 where @xmath90 and @xmath91 are constants and @xmath23 .
we can conclude from the polymer - zeta property that @xmath92 and accordingly the runtime reduces to @xmath93 .
sparsification can also be applied to the dp - algorithm folding rna structures with pseudoknots @xcite .
in contrast to the decomposition rule @xmath0 that spliced an interval into two subsequent intervals , we encounter in the grammar for pseudoknotted structures additional more complex decomposition rules @xcite . as shown in @xcite there exist some decomposition rules which are not @xmath49-compatible and which can accordingly not be sparsified at all , see figure [ f : spar_g ] .
for instance , given a decomposition rule @xmath56 in ` pknot - r&e ` subsequent decomposition rules which are @xmath49-compatible to @xmath56 are referred to as split type of @xmath56 @xcite . in the following
we will study rna pseudoknot structures of fixed topological genus , see section * diagrams , surfaces and some generating functions * for details .
an algorithm folding such pseudoknot structures , ` gfold ` , has been presented in @xcite .
the decomposition rules that appear in ` gfold ` are reminiscent to those of ` pknot - r&e ` but as they restrict the genus of sub - structures the iteration of gap - matrices is severely restricted and the effect of sparsification of these decompositions is significantly smaller . , which is @xmath49-compatible to itself .
we show that for @xmath0 we obtain a linear reduction in time complexity .
( b ) three decomposition rules @xmath94 where @xmath95 are @xmath49-compatible to @xmath42 .
a quantification of the candidate set is not implied by the polymer - zeta property .
( c ) three decomposition rules @xmath94 where @xmath95 are not @xmath49-compatible to @xmath42 . ] in the following , we restrict our analysis to the decomposition rule @xmath0 which splices an interval into two subsequent intervals . expressed in combinatorial language , @xmath0 cuts the backbone of an rna pseudoknot structure of fixed genus @xmath96 over one interval without cutting a bond .
in this section we recall some facts about diagrams and pass from diagrams to surfaces in order to be able to formulate what we mean by an rna pseudoknot structure of fixed genus @xmath96 .
most of this section is derived from @xcite with the exception of lemma [ l : irr_rec ] and theorem [ t : irr ] , which are new and key for the subsequent analysis of @xmath0-candidates .
a diagram is a labeled graph over the vertex set @xmath97=\{1 , \dots , n\}$ ] in which each vertex has degree @xmath98 , represented by drawing its vertices in a horizontal line .
the backbone of a diagram is the sequence of consecutive integers @xmath99 together with the edges @xmath100 .
the arcs of a diagram , @xmath69 , where @xmath101 , are drawn in the upper half - plane . we shall distinguish the backbone edge @xmath102 from the arc @xmath103 , which we refer to as a @xmath7-arc .
a stack of length @xmath104 is a maximal sequence of `` parallel '' arcs , @xmath105 and is also referred to as a @xmath104-stack , see figure [ f : diagram ] . .
the arcs @xmath106 and @xmath107 are crossing and the dashed arc @xmath108 is a @xmath7-arc which is not allowed .
this structure contains @xmath109 stacks with length @xmath110 , @xmath111 and @xmath112 , from left to right respectively . ]
we shall consider diagrams as fatgraphs , @xmath113 , that is graphs @xmath114 together with a collection of cyclic orderings , called fattenings , one such ordering on the half - edges incident on each vertex .
each fatgraph @xmath113 determines an oriented surface @xmath115 @xcite which is connected if @xmath114 is and has some associated genus @xmath116 and number @xmath117 of boundary components .
clearly , @xmath115 contains @xmath114 as a deformation retract @xcite .
fatgraphs were first applied to rna secondary structures in @xcite and @xcite .
a diagram @xmath113 hence determines a unique surface @xmath115 ( with boundary ) . filling the boundary components with discs we can pass from @xmath115 to a surface without boundary .
euler characteristic , @xmath118 , and genus , @xmath96 , of this surface is given by @xmath119 and @xmath120 , respectively , where @xmath121 is the number of discs , ribbons and boundary components in @xmath113 , @xcite .
the genus of a diagram is that of its associated surface without boundary and a diagram of genus @xmath96 is referred to as @xmath96-diagram . a @xmath96-diagram without arcs of the form @xmath103 ( @xmath7-arcs )
is called a @xmath96-structure . a @xmath96-diagram that contains only vertices of degree three
, i.e. does not contain any vertices not incident to arcs in the upper halfplane , is called a @xmath96-matching .
a stack of length @xmath122 is a maximal sequence of `` parallel '' arcs , @xmath123 a diagram is called irreducible , if and only if it can not be split into two by cutting the backbone without cutting an arc .
let @xmath124 and @xmath125 denote the number of @xmath96-matchings and @xmath96-structures having @xmath126-arcs and @xmath126 vertices , respectively , with gf @xmath127 the gf @xmath128 has been computed the context of the virtual euler characteristic of the moduli - space of curves in @xcite and @xmath129 can be derived from @xmath128 by means of symbolic enumeration @xcite .
the gf of genus zero diagrams @xmath130 is wellknown to be the gf of the catalan numbers , i.e. , the numbers of triangulations of a polygon with @xmath131 sides , @xmath132 as for @xmath133 we have the following situation @xcite [ t : genus ] suppose @xmath134 .
then the following assertions hold + * ( a ) * @xmath135 is algebraic and @xmath136 in particular , we have for some constant @xmath137 depending only on @xmath96 and @xmath138 : @xmath139{\bf d}_{g}(z)\sim a_g\,n^{3(g-\frac{1}{2 } ) } \gamma^n.\ ] ] * ( b ) * the bivariate gf of @xmath96-structures over @xmath126 vertices , containing exactly @xmath4 arcs , @xmath140 , is given by @xmath141 in the context of @xmath0-candidates we observed that irreducible substructures are of key importance .
it is accordingly of relevance to understand the combinatorics of these structures . to this end
let @xmath142 denote the gf of irreducible @xmath96-structures .
[ l : irr_rec ] for @xmath143 , the gf @xmath144 satisfies the recursion @xmath145 for a proof of lemma [ l : irr_rec ] , see section * proofs*. [ t : irr ] for @xmath133 we have + * ( a ) * the gf of irreducible @xmath96-structures over @xmath126 vertices is given by @xmath146 where @xmath147 , @xmath148 and @xmath149 are both polynomials with lowest degree at least @xmath150 , and @xmath151 , @xmath152 . in particular , for some constant @xmath153 and @xmath154 : @xmath155 * ( b ) * the bivariate gf of irreducible @xmath96-structures over @xmath126 vertices , containing exactly @xmath4 arcs , @xmath156 , is given by @xmath157 where @xmath158 .
we shall postpone the proof of theorem [ t : irr ] to section * proofs*. in section * sparsification * we observed that sparsification applies to the decomposition rule @xmath0 , which effectively splices off an irreducible sub - structure ( diagram ) .
this notion of @xmath0-irreducibility is indeed compatible by the notion of combinatorial irreducibility introduced in section * diagrams , surfaces and some generating functions * , see figure [ f : diagram10 ] .
an optimal solution for the original structure is obtained from an optimal solution of the spliced , @xmath0-irreducible , sub - structure and an optimal solution for the remaining sub - structure .
splitting @xmath159 to @xmath160 and @xmath161 , @xmath162 is not @xmath0-irreducible , while @xmath163 and @xmath164 are .
however , for the decomposition rule @xmath40 , which removes the outmost arc , @xmath164 is not @xmath40-irreducible while @xmath163 is . ]
folded configurations are energetically optimal and dominated by the stacking of adjacent base pairs @xcite , as well as minimum arc - length conditions @xcite discussed before . in the following we mimic some form of minimum free energy @xmath96-structures : inspired by the nussinov energy model
@xcite we consider the weight of a @xmath96-structure over @xmath126 vertices to be given by @xmath165 , where @xmath104 is the number of arcs for some @xmath166 @xcite .
note that the case @xmath167 corresponds to the uniform distribution , i.e. all @xmath96-structure have identical weight .
this approach requires to keep track of the number of arcs , i.e. we need to employ bivariate gf . in theorem [ t : genus ] * ( b ) * we computed this bivariate gf and in theorem [ t : irr ] * ( b ) * we derived from this bivariate gf @xmath168 , the gf of irreducible @xmath96-structures over @xmath126 vertices containing @xmath104 arcs .
the idea now is to substitute for the second indeterminant , @xmath169 , some fixed @xmath170 .
this substitution induces the formal power series @xmath171 which we regard as being parameterized by @xmath172 .
obviously , setting @xmath167 we recover @xmath129 , i.e. we have @xmath173 . note that for @xmath174 , the polynomial @xmath175 has no real root .
thus we have for @xmath174 the asymptotics @xmath176 with identical exponential growth rates as long as the supercritical paradigm @xcite applies , i.e. as long as @xmath177 , the real root of minimal modulus of @xmath178 is smaller than any singularity of @xmath179 .
in this situation @xmath172 affects the constant @xmath180 and the exponential growth rate @xmath181 but _ not _ the sub - exponential factor @xmath182 .
the latter stems from the singular expansion of @xmath128 .
analogously , we derive the @xmath172-parameterized family of gf @xmath183 . assuming a random sequence has on average a probability at most @xmath28 to form a base pair we fix in the following @xmath184 , where @xmath25 is the euler number . by abuse of notation
we will omit the subscript @xmath172 assuming @xmath185 .
the main result of this section is that the set of @xmath0-candidates is small . to put this size into context we note that the total number of entries considered for the @xmath0-decomposition rule is given by @xmath186 [ t : eumel1 ]
suppose an mfe @xmath96-structure over an interval of length @xmath4 is irreducible with probability @xmath187 , then the expected number of candidates of @xmath96-structures for sequences of lengths @xmath126 satisfies @xmath188 and furthermore , setting @xmath189 we have @xmath190 where @xmath191 is a constant .
we provide an illustration of theorem [ t : eumel1 ] in figure [ f : ecand ] .
-structures , @xmath192 and @xmath193 : we compute the expected number of candidates obtained by folding @xmath194 random sequences for secondary structures ( a)(solid ) and @xmath7-structures ( b)(solid ) .
we also display the theoretical expectations implied by theorem [ t : eumel1 ] ( a)(dashed ) and ( b)(dashed ) . ]
we proof the theorem by quantifying the probability of @xmath64 $ ] being a @xmath0-candidate . in this case
any ( not necessarily unique ) sub - structure , realizing the optimal solution @xmath19 , is @xmath0-irreducible , and therefore an irreducible structure over @xmath64 $ ] .
let @xmath195 , by assumption , the probability that @xmath64 $ ] is a candidate conditional to the existence of a substructure over @xmath64 $ ] is given by @xmath196 \mid [ i , j ] \
\text{\rm is a candidate}\ ) = \frac{{\bf d}^*_{g}(m)}{{\bf d}_{g}(m)},\ ] ] note that @xmath197 \mid [ i , j ] \
\text{\rm is a candidate}\ ) $ ] does not depend on the relative location of the interval but only on the interval - length .
let @xmath198 , then according to theorem [ t : genus ] , @xmath199 for @xmath200 where @xmath201 and @xmath202 are constants . on the one hand @xmath203 where @xmath204 is a constant . on the other hand , we have @xmath205 setting @xmath206
, we can conclude that @xmath207 , see fig .
[ f : prob ] .
( a ) and @xmath208 ( b ) . ]
we next study the expected number of candidates over an interval of length @xmath4 . to this end
let @xmath209\mid [ i , j ] \
\text{\rm is a $ \lambda^*$-candidate of length $ m$ } \ , \}\vert.\ ] ] the expected cardinality of the set of @xmath0-candidates of length @xmath195 encountered in the dp - algorithm is given by @xmath210 since there are @xmath211 starting points for such an interval @xmath64 $ ] .
therefore , by linearity of expectation , for sufficiently large @xmath212 , @xmath213 with @xmath214 being a small constant
. thus we have @xmath215 consequently , the expected size of the @xmath0-candidate set is @xmath5 .
we proceed by comparing the expected number of candidates of a sequence with length @xmath126 with @xmath216 , @xmath217 for sufficient large @xmath218 , @xmath219 .
furthermore @xmath220 from which we can conclude @xmath221 and the theorem is proved . in this section
we discuss the more realistic loop - based energy model of rna secondary structure folding .
to be precise we evoke here instead of two trivariate gfs @xmath222 and @xmath223 counting secondary structures over @xmath126 vertices that filter energy and arcs .
this becomes necessary since the loop - based model distinguishes between arcs and energy .
the `` cancelation '' effect or reparameterization of stickiness @xcite to which we referred to before does not appear in this context .
thus we need both an arc- as well as an energy - filtration .
a further complication emerges . in difference to the gfs @xmath224 and @xmath168 the new gfs are not simply obtained by formally substituting @xmath225 into the power series @xmath129 and @xmath144 as bivariate terms .
the more complicated energy model requires a specific recursion for irreducible secondary structures .
the energy model used in prediction secondary structure is more complicated than the simple arc - based energy model .
loops which are formed by arcs as well as isolated vertexes between the arcs are considered to give energy contribution .
loops are categorized as hairpin loops ( no nested arcs ) , interior loops ( including bulge loops and stacks ) and multi - loops ( more than two arc nested ) , see figure [ f : loops ] .
an arbitrary secondary structure can be uniquely decomposed into a collection of mutually disjoint loops . a result of the particular energy parameters
@xcite is that the energy model prefers interior loops , in particular stacks ( no isolated vertex between two parallel arc ) , and disfavors multi - loops .
base on this observation , we give a simplified energy model for a loop @xmath226 contained in secondary structure by * @xmath227 if @xmath104 is a hairpin loop , * @xmath228 if @xmath104 is an interior loop , * @xmath229 if @xmath104 is a multi - loop , where @xmath226 is a loop .
the weight for a secondary structure @xmath230 accordingly is given by @xmath231 let @xmath232 and @xmath233 be the gfs obtained by setting @xmath234 and @xmath235 in @xmath223 and @xmath222 , where @xmath25 is the euler number .
this means we find a suitable parameterization which brings us back to a simple univariate gf . [ l : eumel3 ] the weight function of rna secondary structures , @xmath232 , satisfies @xmath236 and @xmath237 is uniquely determined by the above equation .
furthermore @xmath238 we first consider the gf @xmath232 whose coefficient of @xmath239 denotes the total weight of irreducible secondary structures over @xmath126 vertexes , where @xmath240 is an arc .
thus it gives a term @xmath241 .
isolated vertex lead to the term @xmath242 where @xmath243 denotes the minimum number of isolated vertexes to be inserted . depending on the types of loops formed by @xmath244
, we have * hairpin loops : @xmath245 , * interior loops : @xmath246 , * multi - loops : there are at least two irreducible substructures , as well as isolated vertices , thus @xmath247 we compute @xmath248 which establishes the recursion . the uniqueness of the solution as a power series follows from the fact that each coefficient can evidently be recursively computed . an arbitrary secondary structure can be considered as a sequence of irreducible substructure with certain intervals of isolated vertexes .
thus @xmath249 [ l : eumel33 ] @xmath232 and @xmath233 have the same singular expansion .
@xmath250 where @xmath251 and @xmath252 are constants and @xmath253 solving eq .
[ e : sec_rec_loop ] we obtain a unique solution for @xmath232 whose coefficient are all positive . observing the dominant singularity of @xmath232
it is @xmath254 .
@xmath233 is a function of @xmath232 and we examine the real root of minimal modulus of @xmath255 is bigger than @xmath256 . then by the supercritical paradigm @xcite applying , @xmath233 and @xmath232 have identical exponential growth rates
furthermore , @xmath232 and @xmath233 have the same sub - exponential factor @xmath257 , hence the lemma .
[ t : eumel22 ] suppose an mfe secondary structure over an interval of length @xmath4 is irreducible with probability @xmath258 , then the expected number of candidates for sequences of lengths @xmath126 is @xmath259 and furthermore , setting @xmath260 , we have @xmath261 where @xmath262 . by lemma
[ l : eumel33 ] we have @xmath263 where @xmath90 is a constant .
the proof is completely analogous to that of theorem [ t : eumel1 ] .
we show the distribution of @xmath264 and @xmath192 in figure [ f : prob1l ] .
( a ) and @xmath192 obtained by folding @xmath194 random sequences on the loop - based model ( b)(solid ) , as well as the theoretical expectation implied by theorem [ t : eumel22 ] ( b)(dashed ) . ]
in this paper we quantify the effect of sparsification of the particular decomposition rule @xmath0 .
this rule splits and interval and thereby separates concatenated substructures .
the sparsification of @xmath0 alone is claimed to provide a speed up of up to a linear factor of the dp - folding of rna secondary structures @xcite .
a similar conclusion is drawn in @xcite where the sparsification of rna - rna interaction structures is shown to experience also a linear reduction in time complexity .
both papers @xcite base their conclusion on the validity of the polymer - zeta property discussed in section * sparsification*. for the folding of pseudoknot structures there may however exist non - sparsifiable rules in which case the overall time complexity is not reduced .
the key object here is the _ set of candidates _ and we provide an analysis of @xmath0-candidates by combinatorial means . in general , the connection between candidates , i.e. unions of disjoint intervals and the combinatorics of structures is actually established by the algorithm itself via backtracking : at the end of the dp - algorithm a structure is being generated that realizes the previously computed energy as mfe - structure .
this connects intervals and sub - structures .
so , does polymer - zeta apply in the context of rna structures ?
in fact polymer - zeta would follow _ if _ the intervals in question are distributed as in uniformly sampled structures .
this however , is far from reasonable , due to the fact that the mfe - algorithm deliberately designs some mfe structure over the given interval . what the algorithm produces is in fact antagonistic to uniform sampling .
we here wish to acknowledge the help of one anonymous referee in clarifying this point .
our results clearly show that the polymer - zeta property , i.e. the probability of an irreducible structure over an interval of length @xmath4 satisfies a formula of the form @xmath265 $ } ) = b \
, m^{1+c } , \quad \text{\rm where } \ b , c>0.\ ] ] does not apply for rna structures .
the theoretical findings from self - avoiding walks @xcite unfortunately do not allow to quantify the expected number of candidates of the @xmath0-rule in rna folding . that the polymer - zeta property does not hold for rna
has also been observed in the context of the limit distribution of the 5-3 distances of rna secondary structures @xcite .
here it is observed that long arcs , to be precise arcs of lengths @xmath20 _ always _ exist .
this is of course a contradiction to eq .
( [ e : mmn ] ) .
the key to quantification of the expected number of candidates is the singularity analysis of a pair of energy - filtered gf , namely that of a class of structures and that of the subclass of all such structures that are irreducible .
we show that for various energy models the singular expansions of both these functions are essentially _ equal_modulo some constant .
this implies that the expected number of candidates is @xmath5 and all constants can explicitly be computed from a detailed singularity analysis .
the good news is that depending on the energy model , a significant constant reduction , around @xmath6 can be obtained .
this is in accordance with data produced in @xcite for the mfe - folding of random sequences .
there a reduction by @xmath3 is reported for sequences of length @xmath266 .
our findings are of relevance for numerous results , that are formulated in terms of sizes of candidate sets @xcite
. these can now be quantified .
it is certainly of interest to devise a full fledged analysis of the loop - based energy model .
while these computations are far from easy our framework shows how to perform such an analysis . using the paradigm of gap - matrices backofen
has shown @xcite that the sparsification of the dp - folding of rna pseudoknot structures exhibits additional instances , where sparsification can be applied , see fig .
[ f : spar_g ] ( b ) .
our results show that the expected number of candidates is @xmath5 , where the constant reduction is around @xmath267 .
this is in fact very good new since the sequence length in the context of rna pseudoknot structure folding is in the order of hundreds of nucleotides .
so sparsification of further instances does have an significant impact on the time complexity of the folding .
in this section , we prove lemma [ l : irr_rec ] and theorem [ t : irr ] . * proof for lemma [ l : irr_rec ] : * let @xmath268 and @xmath269 be the bivariate gf @xmath270 , and @xmath271 .
suppose a structure contains exactly @xmath75 irreducible structures , then @xmath272 and @xmath273{\bf d}^*(z , u)=-[u^g]\frac{1}{{\bf d}(z , u)},\quad g\geq 1,\ ] ] as well as @xmath274\frac{1}{{\bf d}(z , u)}$ ] .
. then @xmath276 , whence for @xmath134 , @xmath277{\bf{f}}(z , u){\bf{d}}(z , u)=0,\ ] ] and @xmath278 , where @xmath279{\bf{f}}(z , u)=[u^g]\frac{1}{{\bf{d}}(z , u)}$ ] .
furthermore , we have @xmath280 and @xmath281 which implies @xmath282 and @xmath283 * proof for theorem [ t : irr ] * let @xmath97_k$ ] denotes the set of compositions of @xmath126 having @xmath284 parts , i.e. for@xmath285_k$ ] we have @xmath286 and @xmath287 .
+ _ claim . _ @xmath288_{g+1-j}}\prod_{i=1}^{g+1-j } { \bf{d}}_{\sigma_i}(z)\right).\ ] ] we shall prove the claim by induction on @xmath96 . for @xmath289
we have @xmath290 whence eq .
( [ e : irr_exp ] ) holds for @xmath289 . by induction hypothesis
, we may now assume that for @xmath291 , eq .
( [ e : irr_exp ] ) holds .
according to lemma [ l : irr_rec ] , we have @xmath292_{g_1-j}}\prod_{i=1}^{g_1-j } { \bf{d}}_{\sigma_i}(z)\right)\right){\bf{d}}_{g+1-g_1}(z).\end{aligned}\ ] ] we next observe @xmath293_{g+1-(g-1)}}\prod_{i=1}^{g+1-(g-1 ) } { \bf{d}}_{\sigma_i'}(z)\right),\ ] ] and setting @xmath294 we obtain , @xmath295_{g_1-j}}\prod_{i=1}^{g_1-j } { \bf{d}}_{\sigma_i}(z)\right){\bf{d}}_{g+1-g_1}(z)\\ & = & \sum_{g_1=1}^{g}\sum_{h=2}^{g_1}\frac{(-1)^{h+2 } } { { \bf{d}}_{0}(z)^{h+2 } } \left(\sum_{\sigma\in[g_1]_{h}}\prod_{i=1}^{h } { \bf{d}}_{\sigma_i}(z)\right){\bf{d}}_{g+1-g_1}(z)\\ & = & \sum_{h=2}^{g}\frac{(-1)^{h+2 } } { { \bf{d}}_{0}(z)^{h+2}}\left(\sum_{g_1=h}^{g } \left(\sum_{\sigma\in[g_1]_{h}}\prod_{i=1}^{h } { \bf{d}}_{\sigma_i}(z)\right){\bf{d}}_{g+1-g_1}(z)\right)\\ & = & \sum_{h=2}^{g}\frac{(-1)^{h+2 } } { { \bf{d}}_{0}(z)^{h+2 } } \left(\sum_{\sigma'\in[g+1]_{h+1}}\prod_{i=1}^{h+1 } { \bf{d}}_{\sigma_i'}(z)\right)\end{aligned}\ ] ] and setting @xmath296 @xmath297_{g+1-j}}\prod_{i=1}^{g+1-j } { \bf{d}}_{\sigma_i'}(z)\right).\ ] ] consequently , the claim holds for any @xmath133 . for any @xmath134 , we have @xcite @xmath298 where @xmath299 is a polynomial with integral coefficients of degree at most @xmath300 , @xmath301 , @xmath302{\bf p}_g(u ) \neq 0 $ ] and @xmath303{\bf p}_g(u ) = 0 $ ] for @xmath304 .
let @xmath147 , the claim provides in this context the following interpretation of @xmath305 @xmath306_{g - j}}\prod_{i=1}^{g - j } { \bf p}_{\sigma_i}(u)}{(1 - 4u)^{3g-\frac{g - j}{2}}},\ ] ] and @xmath307_{g - j}}\prod_{i=1}^{g - j } { \bf p}_{\sigma_i}(u)}{(1 - 4u)^{3g-\frac{g - j}{2}}}\\ & = & \sum_{j=0}^{g-2}\sum_{k=0}^{g+1-j } \left(-\frac{1}{2}\right)^{g+1-j}\binom{g+1-j}{k } \frac{\sum_{\sigma\in[g]_{g - j}}\prod_{i=1}^{g - j } {
\bf p}_{\sigma_i}(u)}{(1 - 4u)^{3g-\frac{g - j+k}{2}}}\\ & = & \sum_{j=0}^{g-2}\sum_{s = g - j}^{2g+1 - 2j } \left(-\frac{1}{2}\right)^{g+1-j}\binom{g+1-j}{s - g+j } \frac{\sum_{\sigma\in[g]_{g - j}}\prod_{i=1}^{g - j } {
\bf p}_{\sigma_i}(u)}{(1 - 4u)^{3g-\frac{s}{2}}}.\end{aligned}\ ] ] as @xmath308 and @xmath309 , we have @xmath310
. consequently we arrive at @xmath311 where @xmath312_{g - j}}\prod_{i=1}^{g - j } { \bf p}_{\sigma_i}(u)\right)(1 - 4u)^{\frac{s-1}{2}},\end{aligned}\ ] ] and @xmath313_{2}}\prod_{i=1}^{2 } { \bf p}_{\sigma_i}(u)\right)+3\left(-\frac{1}{2}\right)^{3 } \left(\sum_{\sigma\in[g]_{2}}\prod_{i=1}^{2 } { \bf p}_{\sigma_i}(u)\right)(1 - 4u)\\ & + & \sum_{j=0}^{g-3}\sum_{g - j\leq s\leq 2g+1 - 2j\atop{s\text { is even } } } \left(-\frac{1}{2}\right)^{g+1-j}\binom{g+1-j}{s - g+j } \left(\sum_{\sigma\in[g]_{g - j}}\prod_{i=1}^{g - j } {
\bf p}_{\sigma_i}(u)\right)(1 - 4u)^{\frac{s-2}{2}}.\end{aligned}\ ] ] we have for @xmath314_{k}$ ] , @xmath315 @xmath316\left(\sum_{\sigma\in[g]_{k}}\prod_{i=1}^{k } { \bf p}_{\sigma_i}(u)\right)=\sum_{\sigma\in[g]_{k } } \prod_{i=1}^{k}[u^{h_i}]{\bf p}_{\sigma_i}(u),\ ] ] where @xmath317 , @xmath318
. then we obtain that @xmath316\left(\sum_{\sigma\in[g]_{k}}\prod_{i=1}^{k } { \bf p}_{\sigma_i}(u)\right)=0 , \quad 0\leq h\leq 2g-1.\ ] ] since @xmath319{\bf p}_{\sigma_i}(u)=0 $ ] , @xmath320 , @xmath321{\bf p}_{\sigma_i}(u)\neq 0 $ ] and @xmath322 .
thus for @xmath323 , @xmath316{\bf u}_g(u)=0\quad\text{and}\quad [ u^h]{\bf v}_g(u)=0.\ ] ] as shown in @xcite we have @xmath324 and we obtain @xmath325 .
furthermore , @xmath326_{2}}\prod_{i=1}^{2 } { \bf p}_{\sigma_i}(1/4)\right ) = \frac{1}{8}\left(4{\bf p}_g(1/4)-\sum_{j=1}^{g-1}{\bf p}_j(1/4 ) { \bf p}_{g - j}(1/4)\right)\neq 0.\ ] ] we can recruit the computation of @xcite in order to observe @xmath327 . in order to compute the bivariate gf , @xmath168 , we only need to replace in eq .
( [ e : irr_exp ] ) @xmath129 by @xmath224 and the proof is completely analogous .
fenix w.d . huang and christian m. reidys contributed equally to research and manuscript .
we want to thank thomas j.x .
li for discussions and comments .
we want to thank an anonymous referee for pointing out an incorrect assumption of first version of this paper .
his comments have led to a much improved version of the paper .
\(a ) an rna secondary structure and ( b ) an rna pseudoknot structure .
suppose the optimal solution @xmath19 is obtained from the optimal solutions @xmath14 , @xmath15 and @xmath16 . based on the recursions of the secondary structures ,
@xmath14 and @xmath15 produce an optimal solution of @xmath17 . similarly , @xmath15 and @xmath16 produce an optimal solution of @xmath18 .
now , in order to obtain an optimal solution of @xmath19 it is sufficient to consider either the grouping @xmath17 and @xmath16 or @xmath14 and @xmath18 .
what sparsification can and can not prune : ( a ) and ( b ) are two computation paths yielding the same optimal solution .
sparsification reduces the computation to path ( a ) where @xmath328 is irreducible .
( c ) is another computation path with distinct leftmost irreducible over a different interval , hence representing a new candidate that can not be reduced to ( a ) by the sparsification .
sparsification : @xmath31 is alternatively realized via @xmath47 and @xmath43 , or @xmath37 and @xmath41 . thus it is sufficient to only consider one of the computation paths .
the recursion solving the optimal solution for secondary structures .
\(a ) three decompositions via the rule @xmath0 , which is @xmath49-compatible to itself . we show that for @xmath0 we obtain a linear reduction in time complexity .
( b ) three decomposition rules @xmath94 where @xmath95 are @xmath49-compatible to @xmath42 .
a quantification of the candidate set is not implied by the polymer - zeta property .
( c ) three decomposition rules @xmath94 where @xmath95 are not @xmath49-compatible to @xmath42 .
a diagram over @xmath329 .
the arcs @xmath106 and @xmath107 are crossing and the dashed arc @xmath108 is a @xmath7-arc which is not allowed .
this structure contains @xmath109 stacks with length @xmath110 , @xmath111 and @xmath112 , from left to right respectively .
the rule @xmath0 splitting @xmath159 to @xmath160 and @xmath161 , @xmath162 is not @xmath0-irreducible , while @xmath163 and @xmath164 are .
however , for the decomposition rule @xmath40 , which removes the outmost arc , @xmath164 is not @xmath40-irreducible while @xmath163 is .
we compute the expected number of candidates obtained by folding @xmath194 random sequences for secondary structures ( a)(solid ) and @xmath7-structures ( b)(solid ) .
we also display the theoretical expectations implied by theorem [ t : eumel1 ] ( a)(dashed ) and ( b)(dashed ) .
the probability distribution of @xmath264 ( a ) and @xmath208 ( b )
\(a ) hairpin loop , ( b ) interior loop , ( c ) multi - loop .
the distribution of @xmath264 ( a ) and @xmath192 obtained by folding @xmath194 random sequences on the loop - based model ( b)(solid ) , as well as the theoretical expectation implied by theorem [ t : eumel22 ] ( b)(dashed ) . | * background : * we study the sparsification of dynamic programming folding algorithms of rna structures .
sparsification applies to the mfe - folding of rna structures and can lead to a significant reduction of time complexity .
* results : * we analyze the sparsification of a particular decomposition rule , @xmath0 , that splits an interval for rna secondary and pseudoknot structures of fixed topological genus .
essential for quantifying the sparsification is the size of its so called candidate set .
we present a combinatorial framework which allows by means of probabilities of irreducible substructures to obtain the expected size of the set of @xmath0-candidates .
we compute these expectations for arc - based energy models via energy - filtered generating functions ( gf ) for rna secondary structures as well as rna pseudoknot structures . for rna
secondary structures we also consider a simplified loop - energy model .
this combinatorial analysis is then compared to the expected number of @xmath0-candidates obtained from folding mfe - structures . in case of the mfe - folding of rna
secondary structures with a simplified loop energy model our results imply that sparsification provides a reduction of time complexity by a constant factor of @xmath1 ( theory ) versus a @xmath2 reduction ( experiment ) . for the `` full '' loop - energy model
there is a reduction of @xmath3 ( experiment ) .
* conclusions : * our result show that the polymer - zeta property , describing the probability of an irreducible structure over an interval of length @xmath4 does not hold for rna structures . as a result sparsification of the @xmath0-decomposition rule does not lead to a linear reduction of the set of candidates .
we show that under general assumptions the expected number of @xmath0-candidates is @xmath5 , the constant reduction being in the range of @xmath6 .
the sparsification of the @xmath0-decomposition rule for rna pseudoknotted structures of genus @xmath7 leads to an expected number of candidates of @xmath5 .
the effect of sparsification is sensitive to the employed energy model .
[ 1995/12/01 ] |
it is often useful in the visual arts to depict a scene composed within a very wide angle of view . for scenes that are wide but not tall ( lanscapes ) one can project on a cylinder and then unfold this developable surface isometrically onto a plane to get the so - called cylindrical or panoramic perspective .
when the scene is wide all around an axis , it can be projected onto a half - sphere and then deformed onto a disc on the plane .
this is the so - called spherical perspective , and has been described thoroughly in @xcite in the 1960s .
it is in fact a hemi - spherical perspective , that allows a depiction of up to 180 degrees around an axis , wherein images of lines can be easily constructed with ruler and compass within a reasonable approximation .
it is often useful , however , to depict an even wider scene .
the author , having taught a course in art and mathematics for a few years to a varied audience ( urban sketchers , architects , school teachers , programmers ) has often received requests from students regarding two questions : how to plot ( either freehand or with minimal instruments ) a view wider than the @xmath0 allowed by ( hemi)spherical perspective , and how to draw a sphere reflection .
this paper solves the former question and relates it to the latter .
it also hopes to help clarify some concepts of anamorphosis and perspective .
in their 1968 work , barre and flocon @xcite described a ruler and compass method to plot a 180 degrees spherical perspective .
this was a work with a focus on the artistic practice of actual freehand drawing . since then several works of a computational nature have proposed various types of wide angle perspectives , by expanding the angle of view up to 360 degrees or by generalizing the shapes of the projection surfaces ( @xcite and @xcite ) .
these works are of a computational nature and not concerned with the artistic practice of freehand or ruler and compass drawing . as far as this author
can tell , there has been no publication wherein a system is proposed to allow a depiction of a 360 degree spherical perspective in a way adequate for drawing from observation or from orthogonal plans with the use of minimal equipment such as ruler and compass .
regarding contributions by artists themselves , dick thermes is well - known for his paintings on spherical surfaces .
he has published a book @xcite on the subject .
his approach is based on gridding in the manner that follows from @xcite . his solution to go beyong the 180 degrees
is simply to draw the two complementary 180 degree views separately and place them adjacently to each other .
when drawing on a sphere he can put each view on on its hemisphere , but that is a work of anamorphosis and not of perspective proper . the most common artistic device for representating views beyond 180 degrees
is that of drawing sphere reflections from observation , and we will consider the relation of these with spherical perspective . in @xcite sphere reflections have been proposed as way to obtain a wide field of view ( with difficulties that we will discuss ahead ) .
again this is a work of a computational nature .
perspectives are ways of representing spatial scenes on a plane , with relation to an observer . we take a _ scene _ to mean any subset of the euclidean 3-dimensional space .
we represent an _ observer _ by a point in 3-space , usually denoted by @xmath1 . _ _ _ _ our purpose is to map each point of a scene in 3-space onto a point on a plane , the latter point being called the perspective image of the former .
mostly we will be concerned with sets of points , lines , planes , or circles , and will use these to approximate more complicated objects .
a map from @xmath2 to @xmath3 that fails to project these elementary objects with simplicity is not adequate for the purposes of a perspective to be used `` by hand '' .
all the maps that are usually called perspectives ( plane , cylindrical , spherical ) can be constructed as particular cases of a same scheme that we will now describe .
let @xmath1 be an observer and @xmath4 a point in a scene .
we call the ray @xmath5 the _ ray of sight _ from @xmath1 to @xmath4 .
let @xmath6 be the unit sphere centered at @xmath1 .
we call @xmath7 the _ direction of sight _ from @xmath1 to @xmath4 .
points will be equivalent if they are radial from @xmath1 .
these equivalence classes are naturally formalized into points of the projective space in classical perspective . for more general perspectives
, however we have to take into account the need to represent both a direction and its diametrically opposite , hence the sphere is the natural manifold of directions . ] _ _ _ _ a point @xmath4 in a scene will define a vector @xmath5 and a ray with the same notation
. we will let the ambiguity stand and let context distinguish them .
the unit vector @xmath8 on the sphere can be seen as the equivalence class of all the points that have it s direction of sight from @xmath1 , i.e. , all the points in the ray of sight @xmath5 .
now recall a few generalities about circles on spheres : a _ great circle _ is a circle on a sphere , defined by the intersection of the sphere with a plane through the origin . _ _ given a point p on the sphere we call _ antipode point _ of p to the its diametrically opposite point on the sphere .
_ _ each point @xmath4 on the sphere defines a family of great circles that covers the sphere , all of those circles crossing both @xmath4 and its antipode .
each two non - antipodal points @xmath4 and @xmath9 on the sphere define a unique great circle , the intersection of the sphere with the plane @xmath10 .
a _ meridian _ is one contiguous half of a great circle .
given a point @xmath4 and its antipode @xmath11 we will call @xmath4-meridian or @xmath12-meridian to a meridian that is an arc of a great circle whose endpoints are at @xmath4 and at its antipode . _ _
we are now ready to consider the problem of vanishing points . [ missing ]
let @xmath13 be a line not crossing @xmath1 .
there is a single plane @xmath14 through @xmath1 containing @xmath13 .
this plane projects radially into a great circle @xmath15 on the sphere .
the set of rays of sight of @xmath13 will form a cone , the set @xmath16 which is a half - plane contained in @xmath14 .
the boundary of this half - plane is the line @xmath17 , the translation of @xmath13 to the origin .
this line corresponds to two rays from @xmath1 none of which is a ray of sight of an actual point of @xmath13 but correspond to the limit of the directions of sight of an observer that follows the line to infinity in both directions .
the intersection of @xmath18 with the sphere is the set of directions of sight of @xmath13 .
it is half of the great circle @xmath15 , and it does not contain the two antipodal endpoints , which are the intersection of @xmath17 with the sphere . the two missing directions in example [ missing ] correspond to the intuitive notion of vanishing point .
as the eye follows a line @xmath19 to infinity , the ray of sight will in the limit become parallel to the line it follows .
this will happen at both ends of the line @xmath19 .
but since this happens for no actual point on the line , the projection of the line @xmath19 on the sphere of directions will be missing two points at its end .
such missing points , for lines or more complex sets , can be added to the cone of directions by taking the topological closure .
let @xmath20 be a scene .
let @xmath1 be an observer .
let @xmath21 be the topological closure of @xmath20 .
the rays of sight of the points of @xmath21 define a cone @xmath22 with vertex at @xmath1 .
let @xmath23 be the closure of the intersection of this cone with the unit sphere @xmath6 .
we call @xmath24 the _ cone of sight _ of @xmath20 relative to @xmath1
. we will abuse the term to also mean the corresponding cone of rays in 3-space stemming from @xmath1 that project radially to @xmath24 , and we will denote that cone of rays by @xmath25 .
_ _ taking the topological closure of the cone of sight allows us to obtain its vanishing points in the following definition .
note that we took the closure of @xmath20 in order to avoid the appearance of `` false '' vanishing points . in what follows
we will always assume the scenes to be closed sets , so this step may be ignored .
we call _ vanishing points _ of a scene @xmath20 to the set @xmath26 . _
_ that is , vanishing points are the frontier points of the cone of directions that are not directions of actual points in ( the closure of ) the scene .
going back to example [ missing ] we take the closure of the half circle and obtain its two endpoints .
we see that the set of vanishing points of a line @xmath13 is the intersection of its translation to the origin , @xmath17 , with the sphere of directions .
the cone of sight of @xmath13 is one half of the great circle of @xmath14 , including its end points .
the corresponding cone of rays is the half - plane of h plus the set of two diametrically opposite rays passing through the vanishing points .
let @xmath14 be a plane not passing through @xmath1 .
the cone of rays of the individual points of @xmath14 forms a half - space whose frontier is a plane parallel to @xmath14 passing through @xmath1 .
this frontier is not contained in the set of rays of sight of individual points of @xmath14 .
it is however contained in the closure of that set . on the sphere ,
the cone of sight will be a hemisphere and the set of vanishing points will be the great circle that forms its boundary .
( in classical perspective the visible part of this circle will form a line , called a vanishing line . )
we say that a surface @xmath27 is radial with respect to a point @xmath1 if any ray stemming from @xmath1 will hit the surface on at most on point .
we will say that a surface is radial if the point @xmath1 is obvious from context . a _ perspective
_ is a map from euclidean 3-space to a region of the plane , that is achieved by the composition of two maps .
the first we call an _ anamorphism _ and the second a _ flattening _ . _ _ _ _ _ _ given an observer @xmath1 , a surface @xmath27 that is radial for @xmath1 , and a scene @xmath28 , we say that @xmath29 is the _ anamorphosis _ ( or _ anamorphic image _ ) of @xmath28 on @xmath27 relative to @xmath1 . to the map that takes each point to its anamorphic image we call the anamorphism onto @xmath27 relative to @xmath1 . is an image painted on @xmath27 such that the observer at @xmath1 would mistake that image for the spatial scene itself .
we should notice here that we are using the word in two senses : there is `` anamorphosis '' as the problem of tromp loeil , and there is anamorphosis as the mathematical construction we just defined .
that the latter is the solution of the former can not be demonstrated mathematically ; it is an empirical fact that depends on the approximate validity of linear optics and on the physiology vision . ] . _ _ _ _ a _ flattening _ is a map from @xmath27 to a region of the plane @xmath3 .
it takes the anamorphosis of a scene to its perspective image . _ _
a perspective is a map @xmath30 where @xmath31 is an anamorphosis and @xmath32 is a flattening .
we note that the anamorphism is fully defined by the choice of surface @xmath27 and observer @xmath1 .
the map itself is just the conical projection from @xmath1 onto that surface . in classical perspective
the anamorphic surface is a plane and the flattening can be seen as the identity map up to scaling .
lines will project onto lines , but for each line only at most one of the vanishing points will be present in the perspective image .
lines that are parallel to the anamorphic plane will have no vanishing points at all in their perspective image .
in cilyndrical perspective , @xmath27 is a cylinder with @xmath1 on the axis , and the flattening consists in unrolling the cylinder over a plane .
the anamorphic images of straight lines will be arcs of ellipses and the flattening will unroll them onto sinusoids .
the pairs of vanishing points of lines on the sphere of directions will project onto the cylinder except if their directions are along the axis of the cylinder
. in the the so - called spherical perspective of @xcite @xmath27 is a hemisphere of a sphere around @xmath1 and the flattening is a restriction to a hemisphere of the azimuthal equidistant projection .
lines will project onto arcs of great circle on the hemisphere , and will have one vanishing point on both the anamorphic image and the perspective image ; or two points if these are on the equator . in the the full spherical perspective considered ahead
@xmath27 is a sphere around @xmath1 and the flattening is the azimuthal equidistant projection .
lines will project onto halves of great circle on the sphere , and will have always two vanishing points on the anamorphic image . because the anamorphic surface is a sphere , just like the sphere of directions , the anamorphism is a homothety , and can be identified with the identity map .
we note an interesting symmetry between spherical and plane perspective . in classical perspective
the flattening is trivial but the anamorphosis is not .
in spherical perspective the reverse is true .
this is because in classical perspective the plane of the anamorphosis can be identified with the plane of the perspective , while in the spherical perspective the anamorphic sphere can be identified with the sphere of directions , so the flattening in the former case and the anamorphosis in the latter can be identified with the identity map .
we have not placed any constraints on the nature of @xmath27 or on the maps .
the fact is that in artistic practice there are examples where the surface @xmath27 in anything from a set of disconnected planes to a myriad of dust - like particles in suspension .
@xmath27 is certainly not a necessarily a smooth , nor a connected surface , although it is so in all the usual perspectives such as the one we treat here .
we will now define our spherical perspective , within the general scheme outlined above .
first we must define the anamorphic surface and the place of our observer .
we will take @xmath27 to be a spherical surface of arbitrary positive radius @xmath33 , with the observer @xmath1 at its center .
choosing these elements fully defines our anamorphism .
what can we say about it ? as we already discussed , the anamorphic sphere being a surface , the anamorphic image coincides with the cone of sight in the sphere of directions .
hence the anamorphosis of a line @xmath13 onto a sphere relative to its center is an arc of a great circle : the rays of sight from @xmath1 to the points of the line @xmath13 define ( half of ) a plane that crosses @xmath1 .
the anamorphic image of that plane is a great circle , and the image of the line is a half of the great circle , delimited by two vanishing points , determined thus : if @xmath19 is a line , and @xmath34 is the translation of line @xmath19 that passes through @xmath1 , then the vanishing points of @xmath19 are the intersection of @xmath34 with @xmath27 .
we will now define our flattening map .
we must first define a system of coordinates .
we consider a ray stemming from @xmath1 , representing a privileged direction of sight .
we call it the _ central ray of sight _ and to its axis we call the _ central axis of sight _ .
we place an orthonormal reference frame @xmath35 in @xmath1 , such that the positive side of the @xmath36 axis coincides with the central ray of sight . for quick reference
we name the points where the three axes cut the sphere : we call * * n**orth to the intersection of the central ray of sight with the sphere and * * s**outh to its antipode point ; * * e**ast to the point where the @xmath37 axis touches the sphere and * * w**est to it s antipode ; * * z**enith to the point where the positive @xmath38 axis touches the sphere and na**d**ir to its antipode , and we represent these points by the letters that we wrote in bold . _ _ _ _ we call the plane orthogonal to the central axis of sight ( the xz - plane ) the _ observer s plane _ ; we call the @xmath39-plane the _ sagital plane _ .
the observer s plane intersects the sphere in a great circle we call the _ equator _ . _
_ _ _ _ _ we call the half - space @xmath40 the anterior half - space ( representing everything in front of the observer ) and to the half - space @xmath41 we call the posterior half - space ( representing all that is behind the observer ) .
we construct a flattening map that will project all the points of the sphere onto a disc on a plane , with the exception of the south pole .
this map composed with the anamorphosis will result in a perspective that images all points of 3-space except those points located on the @xmath42 ray .
we define the flattening thus : we map the sphere minus its south pole one - to - one to a disc on the plane in such a way that each north - south meridian goes to a line segment and \i ) distances are preserved along each meridian .
\ii ) angles between meridians are preserved at point @xmath43 .
condition i ) means that the map is an isometry for each meridian separately .
since distances measured along great circles of the sphere are proportional to angles from the center , this means that if @xmath44 are points on the same meridian and if @xmath45 are their images , then @xmath46 up to multiplication by a scale factor to mean that these equalities are valid modulo product by the adequate scale factors . ] .
condition i ) also implies that @xmath43 will be mapped to the center of the disc with the segments corresponding to the meridians radiating from it .
the lines stemming from the image of @xmath43 we will call measuring lines , because angular distances are preserved along them
. they will be the essential tool for our plotting of points .
condition ii ) means that the angles between measuring lines at the image of @xmath43 will be equal to the angles of their meridians at @xmath43 ( which are the angles between the planes that define them ) .
it ensures the meridians will be distributed radially preserving their tangents on @xmath43 , that is , they will look as if orthogonally dropped onto the tangent plane to the sphere at @xmath43 .
we call _ longitude _ of an n - meridian to the angle between its tangent and that of the east meridian at point @xmath43
. the longitude of a meridian will equal the angle of its measuring line with the @xmath47 measuring line .
_ _ condition i ) and ii ) together imply that the two meridians of each great circle through @xmath43 form a diameter of the perspective disc and that distances are preserved along interior points of these diameters . in intuitive terms , we look at the north - south meridians as inextensible threads , connected to the sphere on the north and south poles .
we cut them free at the south pole only , and , pulling them straight along their tangents at @xmath43 , make from them a disc on the plane tangent to the sphere at @xmath43 .
at the end , the threads are radiating from @xmath43 , each under the orthogonal shadow of its former position , so that the angles their tangents make at @xmath43 are the same as before .
distances have not changed between points on the same thread .
the threads together form a disc of radius @xmath48 , where @xmath33 is the radius of the sphere , with @xmath43 at the center and each point of the sphere corresponds to a single point on the disc , except for the south pole , that was blown - up and is missing from the outer edge of each thread , hence making the disc open at its boundary .
we take the closure of the disc and get a frontier circle which we call the south circle , or the blow - up of the south pole , each of whose points corresponds to one of the directions that the original pole could be apporached from ( each of them can be identified with one of the meridians , or with a ray of the tangent plane at @xmath27 ) .
when context is clear we will use the same letters for the @xmath49 points and their images , and will call equator to the perspective image of the equator . see fig .
[ disc360 ] for a picture of the perspective image of these points and lines .
points in the anterior half - space will be projected by the anamorphosis into the hemisphere in front of the equator , and these will be flattened in turn onto the inner part of the perspective disc that is contained within the image of the equator .
this inner disc will have half of the radius of the perspective disc .
the posterior half space will be projected onto the outer ring between the image of the equator and the circle of the blow - up of the south pole @xmath27
. in terms of the @xmath50 coordinates , the flattening composed with anamorphosis @xmath51 gives the perspective map @xmath52 where @xmath27 is the sphere of radius @xmath33 and @xmath53 is a disc of radius @xmath48 , @xmath54 that is , we project orthogonally against the @xmath55plane , take the unit vector , and then scale to a length equal to the value of the angle @xmath56 .
the natural set of spherical coordinates for this map is @xmath57 where @xmath58 that is , @xmath59 is the latitude and @xmath60 is the longitude with respect to the north pole . in these coordinates @xmath61 we see that the perspective map does not in fact depend on the norm @xmath62 of the point , which is to be expected since the anamorphosis is a central projection . , height=340 ]
_ solving a scene _ in perspective means finding the perspective images of all points on the scene - classically , and in the interest of solving a scene with simple instruments , we are concerned with the images of points grouped into lines , and especially with their vanishing points .
we will now show how to solve a scene in our 360 degree spherical perspective using ruler and compass . _ _ a common technique to solve scenes in classical perspective is to make the plane of the perspective image do double or triple duty by superposing on various orthogonal projections .
this technique also works in spherical perspective
. we will illustrate it in the following construction that we will make use of in the next section .
[ on_equator ] construction of the perspective image of points on the observer s plane : let @xmath63 be a point on the observer s plane . then the ray @xmath5 crosses the equator at a point @xmath9
. then the perspective image @xmath28 of @xmath4 will be at the equator of the perspective disc . also , if @xmath60 is the longitude of the meridian crossing @xmath9 , then @xmath64 .
hence the following construction plots the perspective image of @xmath4 : make the perspective plane do double duty , using it to represent also the orthogonal back view of the observer s plane , with the image of @xmath43 in the perspective view coinciding with @xmath1 in the orthogonal view and the perspective scaled in such a way that the equator s perspective image coincides with its orthonal image .
now plot @xmath4 in the orthogonal image from its coordinates and trace the ray @xmath5 .
the point where @xmath5 touches the equator will be the perspective image of @xmath4 .
the problem of solving a scene can be divided into two parts : plotting points and lines in the anterior half - space and in the posterior half - space .
plotting inside the anterior half - space is solved in @xcite .
we will very quickly give a review of the method .
it is shown in @xcite that the perspective image of lines in the anterior half - space is well approximated by arcs of circle .
this is important for two reasons : the first is that in drawing practice , arcs of circle are easy to trace with ruler and compass or even freehand ; the second is that three points determine a unique arc of circle on a plane there passes a single circle . to find that circle
find the perpendicular bisectors of @xmath65 and @xmath66 and intersect them at x , then open the compass from x to p. ] .
on what follows we always assume that the lines being plotted do not cross @xmath1 unless stated .
( the case where it does is easy , the perspective image consisting of two antipodal points that we will also learn how to plot ) .
we have to consider two cases .
we say that a plane is _ frontal _ if it is perpendicular to the central axis .
let @xmath13 be a line on a frontal plane @xmath14 .
first suppose that @xmath14 is not the observer s plane .
translating @xmath13 to @xmath1 we find it has two vanishing points which are diametrically opposite points on the equator .
they are found by drawing the translated line directly on the perspective disc , to obtain its intersection with the equator ( the perspective plane doing double duty as in construction [ on_equator ] ) .
next , we find a third point .
if the line is not vertical , it intersects the sagital plane at some point p. we measure the angle of @xmath56 and plot the measure of this angle on the vertical measuring line .
if the line is vertical then it crosses the @xmath67 plane and we measure instead the angle with the central ray of sight at this point , and plot it on the horizontal measuring line .
the image of the line is well approximated by the arc of circle that crosses the two vanishing points and the third point ( see fig .
[ anterior_lines ] ) .
_ now suppose that @xmath14 is the observer s plane .
we get two vanishing points in the same way as above .
the third point is obtained in the same way as before but will now be found on top of the equator of the perspective disc .
the arc of circle will be one half of this disc .
notice that measuring angles along the vertical plane and along the horizontal plane is very natural in practice , and the kind of thing you can do using an improvised theodolite when drawing from nature .
[ arbitrary_anterior ] we note that we now know how to plot an arbitrary point located on the anterior half - space . given a point @xmath4 , consider the frontal plane going through @xmath4 and on it a vertical line @xmath68 and a horizontal line @xmath69 going through @xmath4 .
we have just learned how to obtain the perspective images of these lines .
the perspective image of @xmath4 will be found at the intersection of the images of @xmath68 and @xmath69 .
we say that a line is a _ receding line _ if it intersects the observer s plane at a single point .
let @xmath4 be the point of intersection of a receding line @xmath13 with the observer s plane .
we construct the image of @xmath4 as in construction [ on_equator ] .
denote the image also by @xmath4 .
the plane @xmath14 defined by @xmath1 and @xmath13 it must also intersect the equator at the antipodal point @xmath11 . to find a third point , we translate @xmath13 to @xmath1 and intersect it with the sphere to find the two vanishing points . one of these will be on the anterior hemisphere . we find it by construction [ arbitrary_anterior ] for plotting an arbitrary anterior point .
let its image be @xmath70 .
we trace the auxiliary arc of circle @xmath71 that is the image of the plane @xmath14 .
the image of @xmath13 will be the part of the arc that lies between @xmath70 and @xmath4 .
a case of particular interest is that of the central lines .
we say that a line is central if it is perpendicular to the plane of the observer .
in this case point @xmath70 will project onto @xmath43 , hence will be between @xmath4 and @xmath11.the image of @xmath14 will be the straight line segment @xmath12 and the image of @xmath13 will be the segment @xmath72 ( see fig . [ anterior_lines ] ) . _
_ is the image of a frontal line .
arc @xmath73 is the image of the plane of a receding line , and the arc from @xmath74 to @xmath70 is the image of the line itself .
the line segment @xmath75 is the image of a central line .
the image of its plane proceeds until the image of the antipode of @xmath76 .
see text for details.,height=453 ] this ends the review of ( hemi)spherical perspective as presented in @xcite . outside of the anterior space disc
the images of lines are no longer well approximated by circles , and that is perhaps why spherical perspective was kept limited to the anterior 180 degrees in its original formulation . however , as it turns out , the generalization is easily constructed both with ruler and compass and even in freehand drawing , by approximating the images of great circles on the posterior hemisphere by `` fat lines '' consisting of segments of circles , in a reasonably easy construction .
let @xmath13 be a spatial line .
together with @xmath1 , @xmath13 defines a plane @xmath14 through the origin , and this plane defines a great circle . we wish to plot , to a good approximation , the full perspective image of this great circle .
the image of @xmath13 will be contained in it and lie between its vanishing points .
what we want is to use the arc of circle approximation obtained in last section for the anterior region and use it to obtain a full plot of the great circle .
the key lies in plotting antipodal points .
let @xmath4 such that @xmath78 be a point in space with antipode point @xmath11 .
let @xmath28 and @xmath79 be the perspective images of @xmath4 and @xmath11 respectively . then @xmath79 is the point on @xmath80 such that @xmath81 equals the radius of the perspective disc .
[ ruler ] the plane @xmath82 defines the single great circle @xmath15 through @xmath43 and @xmath4 .
it is the union of two diametrically opposite n - meridians , @xmath83 .
. then @xmath85 .
the images of @xmath86 and @xmath87 are diametrically opposite measuring lines @xmath88 , respectively , that together form a diameter of the perspective disc .
let @xmath89 be the images of @xmath90 respectively .
then @xmath28 is on @xmath91 , and therefore @xmath79 is on @xmath92 . since @xmath93 is the length of a meridian and @xmath88 are measuring lines , then @xmath81 is the length of a measuring line , or a radius of the perspective disc .
this proposition allows us to easily plot the antipode of an already plotted point @xmath28 .
just draw a line to @xmath43 from @xmath28 , open the compass with radius equal to the radius of the perspective disc , and intersect with the line to find @xmath79 . or ,
if using a marked ruler , pass the ruler through @xmath28 and @xmath43 with the zero marker at @xmath28 , and plot the point at the mark of the lenght of the radius .
for the purposes of freehand drawing of a perspective it is often useful , when plotting points nearer to the equator than to @xmath43 , to use instead the following result : [ freehand ] let @xmath4 such that @xmath78 be a point in space with antipode point @xmath11 .
let @xmath28 and @xmath79 be the perspective images of @xmath4 and @xmath11 respectively .
let @xmath94 be the intersection of @xmath80 with the blow - up of @xmath27 .
then @xmath79 is the point on @xmath80 such that @xmath95 .
also , @xmath96 , where @xmath97 is the point on the perspective disc diametrically opposite to @xmath94 .
the plane @xmath82 defines a great circle @xmath15 that contains @xmath98 , and @xmath27 . on that plane ,
the lines @xmath99 and @xmath100 intersect at @xmath1 , and therefore we have the equalities between opposing angles @xmath101 and @xmath102 . since @xmath15 is a great circle through @xmath43 , it projects to a diameter of the perspective disc .
on @xmath15 we have a cyclic order of points @xmath103 .
the perspective map is continuous outside of @xmath27 and preserves the order on the image .
we will have @xmath104 where @xmath94 and @xmath105 are the points of the blow - up corresponding to the directions of the two meridians of @xmath15 at @xmath27 .
because distances are preserved along measuring lines , the two angle equalities above imply @xmath106 and @xmath107 respectively .
the practical interest of this proposition lies in the fact that for freehand plotting of lines in the full spherical perspective it is often easier to transport the measurement @xmath108 by eye than to transport the radius of the disc without an actual compass or ruler .
but , having a compass at hand , or a marked ruler , the use of proposition [ ruler ] makes for very efficient plotting of antipodes .
this allows us to plot the image of a great circle s posterior meridian from the image of its anterior meridian .
[ makefatline ] construction of _ fat lines _ :
let @xmath15 be the perspective image of the anterior meridian of a great circle on the sphere . to obtain an approximation of the posterior image of @xmath14 , trace an arbitrary number of measuring lines @xmath109 through @xmath43 .
intersect each of these lines with @xmath15 to get points @xmath110 , and use proposition [ ruler ] to to obtain the antipodes @xmath111 . through each successive three of these points
we trace an arc of circle , thus getting overlapping arcs @xmath112 , @xmath113 , etc .
these overlapping arcs from a _ fat line _ that approximates @xmath14 .
the degree to which successive arcs fail to exactly overlap ( how `` fat '' the envelope of these arcs is ) indicates the amount of error in the approximation and the need to increase the number of measuring lines @xmath114 ( see fig .
[ fatlinesexample ] ) .
_ _ _ _ the practical draughtsman , being given @xmath15 , and armed with a marked ruler , will follow the following procedure : stick a nail on the center of the perspective disc , and sliding the ruler against the nail to ensure that it always touches @xmath43 , make its zero mark slide against along the curve @xmath15 , plotting points at the mark of the length of the radius of the disc at desired intervals . with this procedure a great number of points can be marked very quickly , to the point where the antipodal curve can be interpolated by hand with good precision .
we are now ready to plot arbitrary lines in the full perspective .
we have the following cases : let @xmath13 be a line in a frontal posterior plane .
let @xmath14 be the plane defined by @xmath13 and @xmath1 and @xmath15 its great circle .
suppose @xmath13 is not vertical .
then it crosses the sagital plane at a point @xmath4
. the image of @xmath4 will be @xmath28 such that @xmath115 , on @xmath116 or on @xmath117 according to whether @xmath4 is above or below the observer .
the antipode of @xmath4 will map to the point @xmath79 on the same axis and at a distance equal to the radius of the perspective disc .
this point will be in the inner disc corresponding to the anterior half space .
taking point @xmath79 and the two vanishing points at the equator we plot the approximation of the line through and arc of circle by the method of @xcite described above .
let this arc be @xmath118 . to obtain an approximation to the posterior image of @xmath15
we can now trace an arbitrary number of measuring lines through @xmath43 and follow the procedure of construction [ makefatline ] to trace a fat line approximation of the posterior half of @xmath15 from antipode points of @xmath118 .
thus we obtain the full image of @xmath15 . to obtain the image of the line @xmath13 simply discard the anterior half of @xmath15 ( see fig .
[ posteriorlinesfig ] ) .
we can now locate an arbitrary point @xmath4 on the posterior half - space : just pass vertical and horizontal lines through it , plot them according to the procedure just described , and intersect them to find the image of @xmath4 .
let @xmath13 be a line that crosses the observer s plane at a single point @xmath4 .
let @xmath14 be the plane defined by @xmath13 and @xmath1 and @xmath15 its great circle .
displacing @xmath13 to the origin we obtain two vanishing points ; one on the anterior hemisphere , let it be @xmath119 , the other on the posterior hemisphere , its antipode @xmath120 .
plot @xmath119 according to the methods of @xcite , then use proposition @xmath121 to plot @xmath120 . let their images be called @xmath122 and @xmath123 . through construction [ on_equator ]
we obtain the points @xmath28 and @xmath79 , the perspective images of @xmath4 and its antipode respectively , which are on the equator .
trace the arc of circle @xmath124 . from that arc use construction [ makefatline ] to trace the fat line of its antipodal arc .
this plots the full image of the great circle @xmath15 . to get the image of @xmath13 ,
discard the arc of the curve defined by @xmath125 ( see fig .
[ posteriorlinesfig ] ) . in the particular case
in which @xmath13 is a central line , its great circle @xmath15 is an n - meridian , hence its image will be a diameter of the disc .
one of the vanishing points will be at @xmath43 and the other will be split into two diametrically opposite points @xmath122 and @xmath126 at the blow - up circle , each at the longitude of one of the two meridians from which @xmath27 can be approached along @xmath15 .
let @xmath28 be the image at the equator of the point @xmath4 where @xmath13 crosses the observer s plane , which we can obtain by construction [ on_equator ] .
@xmath28 will either be on @xmath127 or on @xmath128 .
suppose it is the former case .
then the image of @xmath13 will be the measuring line @xmath129 . put in another way
, the image of @xmath13 will be the measuring line that contains the image of @xmath4 ( see fig .
[ posteriorlinesfig ] ) . what we have just learned is enough to solve a scene when we have the cartesian coordinates of its points - for instance when drawing from an arquitectural plan .
when drawing from observation , however , the artist is not in the position of the architect , but in that of the astronomer .
what he measures are the angles subtended by objects .
now we have already seen what the natural spherical coordinates are for this perspective ( the angles @xmath59 and @xmath60 defined above ) , and it is possible to construct a simple device to measure such angles directly , but the more habitual set of angles are the horizontal angle @xmath130 together with the angular elevation @xmath131 , defined thus : @xmath130 is the angle between the central ray @xmath132 and the orthogonal projection of @xmath5 against the horizontal plane .
@xmath131 is the angle between @xmath5 and its orthogonal projection on the horizontal plane .
these are the angles one would measure with a standard theodolite .
lines of constant horizontal angle @xmath130 are the images of vertical lines and we already know how to plot them
. lines of constant angular elevation are circles on the anamorphic sphere obtained by intersection with horizontal planes . for short we will call these circles and their images _
parallels__. _ _ _ in the anterior hemisphere , following @xcite we approximate parallels by arcs of circles and plot them in the following way : let @xmath69 be parallel on the anamorphic sphere , of constant angular altitude @xmath131 .
@xmath15 intersects the equator in two points @xmath133 and @xmath134 on the west and east side of the sagital plane respectively and intersects the sagital plane in a point @xmath9 . @xmath133 and
@xmath134 will be mapped to the equator at @xmath135 and @xmath136 respectively in such a way that @xmath137 .
@xmath9 will be mapped to @xmath122 on the vertical line @xmath138 in such a way that @xmath139 .
we draw the arc of circle @xmath140 and take this as an approximation of the image of the parallel @xmath15 . to plot the posterior part of the parallel we make use of the following proposition : [ plotparallels ] let @xmath69 be a parallel on the anamorphic sphere .
let @xmath4 be a point of @xmath69 , with perspective image @xmath28 .
let @xmath141 where @xmath142 is the image of the equator .
let @xmath122 be the point such that @xmath143 is the middle point of @xmath144 .
then @xmath122 is the perspective image of a point of @xmath69 .
parallels and n - meridians lines are both invariant with regard to reflection on the observer plane ( because so are their defining planes and the sphere itself and hence their intersection )
. then the intersection set of parallels and measuring lines is itself invariant for reflection on the observer plane .
this intersection is made up of no points at all or of two symmetric points .
let @xmath69 be a parallel and let @xmath4 be a point on @xmath69 .
there is a single n - meridian @xmath145 through @xmath4 , and this meridian must cross the parallel at another point @xmath33 . by mirror symmetry ,
if @xmath9 is the point where the meridian line crosses the equator , we must have @xmath146 , and , since @xmath145 is an n - meridian its image @xmath19 is a measuring line through the origin and the points @xmath147 ( image points of @xmath148 respectively ) will verify @xmath149 . to plot the posterior half of a parallel @xmath69 , plot first the anterior half @xmath150 as an arc of circle , then plot a set of measuring lines @xmath151 , intersect them with @xmath150 at points @xmath152 , find the points @xmath111 from proposition [ plotparallels ] , and trace a fat line through the @xmath111 .
figure [ elevationfig].a ) shows a computer plot of a uniform grid of parallels and verticals calculated directly from map [ exactxyz ] .
figure [ elevationfig].b ) shows the approximation of the parallels of elevation 10 , 45 , 80 , and 85 degrees plotted by the method above .
it is quite evident that near the equator the approximation is not very good ; the curves are not smooth at the transition from the anterior to the posterior hemisphere .
this is an artefact of the approximations , as it is easy to see from equation [ exactxyz ] that the perspective images of constant elevation curves are differentiable .
the error comers not from proposition [ plotparallels ] , which is exact , but from the initial approximation of the parallel by an arc of circle inside the anterior disc . near the equator
the draughtsman might do well to avoid plotting vanishing points from parallels and verticals and use horizontals and verticals instead .
as is well known in classical perspective drawing , as long as we can plot a grid of squares we can plot any object to any required precision , by caging it inside a fine enough grid and interpolating through the intermediate points .
we will therefore concern ourselves with the basic examples of grid construction . in fig .
[ onepoint360 ] we build the image of a central perspective grid , i.e , we consider the perspective image of a horizontal grid of squares with one axis perpendicular to the observer s plane . we assume for simplicity that a vertex is directly under the observer ,
hence one axis is directly under the @xmath153 axis of the sphere .
the plane of perspective does triple duty here , serving also to represent a top orthogonal view of the scene and a back view of the observer s plane .
we take the sphere s arbitrary radius to be equal to the height of the observer relative to the ground plane , so that a horizontal line through the nadir point @xmath53 coincides with an axis of the grid and will be made to represent both the ground plane on the back view and the observer s plane on top view . on the top view the receding lines of the grid intersect the ground line of the observer s plane at uniformly spaced points @xmath154 .
the perspective of each such point is obtained by intersecting the line that joins them to @xmath43 with the equator .
the same straight line is the perspective image of the central receding line of the grid that crosses that point .
thus the image of the receding lines of the grid is a set of straight lines @xmath155 going from @xmath43 to the blow up and passing through the @xmath154 .
note that this is exactly the same construction as in classical perspective , though with a different interpretation . to plot the frontal lines we first trace a line @xmath19 on the plane of the grid , such that it makes a 45 degree angle to the right of the observer and crosses the nadir point .
this line will cross each row of squares , touching a vertex of each row .
we plot the great circle @xmath15 of the plane defined by @xmath1 and @xmath19 .
first we plot the anterior part by drawing the arc @xmath156 where @xmath119 is the 45 degree vanishing point on the horizontal axis .
then we note that each of the receding lines @xmath155 crosses a vertex of the grid where it touches the anterior part of the 45-degree line . to find these points we intersect each @xmath155 with the upper half of @xmath15 on the anterior disc and take the antipodal point of that intersection .
since this point will be on the posterior space and will belong to @xmath19 and to @xmath155 , then it will be the required vertex .
so we can plot a posterior horizontal line through it to get a posterior planar line of the grid .
note that we have used the receding lines of the grid as the natural measuring lines to construct the posterior fat line of @xmath19 . in this fashion we can plot the full 360 degree grid to any required precision and extension . in fig .
[ room45 ] we represent a tiled cubic room drawn from the point of view of an observer at its center , looking straight into the center of one of the walls .
the whole setup is drawn very simply from a judicious use of vertical and horizontal frontal lines at 45 degrees to the observer ; these lines do double duty , as , for instance , the frontal vertical at 45 degrees to the right of the observer has the same projection as the line on the plane of the horizon that goes from the nadir to the 45 degree mark on the @xmath153 measuring line .
often we will want our grids to be oriented at some arbitrary angle to the ray of sight . in fig .
[ arbitraryback ] we represent a square @xmath157 on a horizontal plane , below , behind , and to the left of the observer , such that one side of the square makes a 60 degree angle with the ray of sight .
once again the plane of perspective does triple duty , serving also to represent a top orthogonal view of the scene and a back view of the observer s plane and we take the sphere s arbitrary radius to be equal to the height of the observer relative to the ground plane , so that a horizontal line through the nadir point represents both the ground plane on the back view and the observer s plane on top view . on this top view
we draw the square @xmath157 and project its sides until they intersect the line of the observer s plane . from the back view
we can now trace lines to these intersection points and find their projections on the equator . from these projections and
the vanishing points we can find the arcs of circle representing the lines that extend the sides of the square ( note that the vanishing points are all on the horizontal measuring line , one set of lines converging to the points at @xmath158 and @xmath159 and the other to @xmath160 and @xmath161 . from the arc of the anterior perspective
we can obtain the corresponding fat lines of the posterior perspective . by intersecting these lines
we find the perspective images of the points @xmath162 . from this square
we can plot a grid by an adaptation of the previous methods .
it is apparent from the plot of the 45 degree room in fig [ room45 ] that our perspective bears striking resemblances to a reflection on a sphere .
it is natural to ask if it is indeed a reflection .
recall how a reflection works ( see fig [ obstruction ] ) : an observer at point @xmath163 will see a point @xmath4 reflected at a point @xmath33 on the sphere according to these rules : @xmath33 will be on the intersection of the sphere with the plane @xmath164 and @xmath165 ( angle of incidence equals angle of reflection ) .
we notice several difficulties . given @xmath33 , it is easy to find the incident and reflected rays , but the inverse problem of obtaining @xmath33 from @xmath4 is non - trivial . in general
it requires solving an algebraic equation of order four ( see @xcite ) .
also , obstructions are non - trivial . in fig .
[ obstruction ] we can see that points @xmath4 and @xmath9 will have the same reflection @xmath33 even though they are not in the same ray from either the center @xmath1 or the observer @xmath163 . by contrast , obstructions are always radial for perspectives , being fully determined before the flattening , at the step of the anamorphosis . finally , our spherical perspective has an angle of view of @xmath77 .
the angle of view captured by a reflection depends on the distance of the observer to the sphere .
the points of the sphere define a cone with the observer @xmath163 at the vertex , the cone of shadow , and every point outside of this cone of shadow will be viewable on the sphere . in angular terms , the field of view will be @xmath166 with @xmath167 , where @xmath19 is the radius of the sphere and @xmath168 the distance of the observer from the center of the sphere .
there is however a limiting case wherein these problems disappear .
there are two ways of making @xmath169 go to zero : you can move away from the sphere ( and look at it from a telescope to compensate ) or stay put and shrink it ( and look at it with a microscope ) , or some combination of the two that will make the @xmath19 negligible small compared to @xmath168 .
either way will , in the limit , make for a @xmath77 angle of view ( the first option will leave a tubular shadow of finite section , however , while the second will reduce the shadow to a ray ) . in any case , in the limit @xmath170 , the rays coming from the observer @xmath163 to the sphere with center @xmath1 become parallel to @xmath171 , and the angle of reflexion @xmath172 becomes equal to the angle @xmath173 .
hence @xmath174 ( see fig . [ parallelreflection ] ) .
still , the position of @xmath4 is not determined by @xmath175 but by the angle @xmath59 of fig .
[ obstruction ] .
we can however make these angles equal by either restricting the reflection to objects at infinte distance ( say , plotting a reflection of the celestial sphere ) or by making @xmath19 go to zero .
when @xmath19 goes to zero , @xmath33 goes to @xmath1 and @xmath59 becomes @xmath176 .
in that limit , the projection becomes radial ( therefore making obstructions trivial ) , and the whole space of directions is mapped onto the hemisphere visible from @xmath163 .
this can be seen as a sphere anamorphosis followed by a linear contraction onto a hemisphere by halving the angle @xmath59 of each point on the sphere .
seen from point @xmath163 , since all rays are parallel to the axis @xmath177 , the reflection will look like the orthogonal projection along @xmath177 of the image on the sphere .
hence the reflection , seen from @xmath163 , is anamorphically equivalent to a perspective obtained by the trivial anamorphism onto the sphere composed with a flattening which is the composition of a linear compression onto a hemisphere followed by an orthogonal projection . in the spherical coordinates of [ natcoords ] ( with the @xmath36 axis on @xmath178 and @xmath179 in the perpendicular plane through @xmath1 ) and rescaling the sphere to @xmath180 ,
this perspective is the map @xmath181 where the first map is the trivial anamorphosis , the second is the crunching into the hemisphere and the last step is the orthogonal projection .
this is a 360 degree perspective , but different from our spherical perspective .
it is not linear along @xmath59 , squashing the outer angles more , and can not be easily used for drawing by hand without the help of pre - computed grids ( since we lose the isometry along measuring lines ) .
but we can see why there is a qualitative similarity between the two
. becomes equal to @xmath172 .
if p goes to infinity then @xmath59 goes to @xmath172.,width=264 ] it has been noted in @xcite that reflections on a sphere could be used as a form of wide angle `` perspective '' .
this is well inspired in art history , in fact , as reflections drawn from observation have been the time honoured tool of the artist to represent a wide angle of view , escher s self portrait being a well known example .
we have already seen the difficulties in this approach .
first , reflections are hard to calculate .
second , they are not perspectives in our sense of the word ( i.e. , `` radial '' perspectives ) , and they have non - trivial obstructions .
as noted in @xcite this causes difficulties for hidden - face removal algorithms .
if the purpose is to represent a wide angle view , spherical perspective is a much more natural proposal .
it allows for ( up to ) a 360 degree view ; it is a perspective in the sense we defined above and therefore , like all such perspectives , has trivial obstructions , with hidden - face algorithms working exactly as in the classical case , being calculated at the anamorphosis step .
furthermore , spherical perspective is easy to calculate by map [ exactxyz ] and , unlike the limiting case of reflections mentioned above , it can actually be used by an artist armed only with ruler and compass or even , after some practice , in freehand drawing from nature .
note : further notes , computer code and illustrations will be made available at the author s page : http://www.univ-ab.pt/~aaraujo/full360.html
this work was supported by fundao para a cincia e a tecnologia ( fct ) project uid / mat/04561/2013 .
trapp , m. and dllner , j. ( 2008 ) , a generalization approach for 3d viewing deformations of single - center projections , in proceedings of international conference on computer graphics theory and applications ( grapp ) , 162 - 170 .
link : url[http://www.hpi.unipotsdam.de / fileadmin / hpi / fgdoellner / publications/ 2008/td08/paper107nonplanarprojectiontrappdoellner.pdf ] | we describe a general setup for anamorphosis and perspective , then obtain a ruler and compass construction for the 360 degrees spherical perspective .
we consider its uses in freehand drawing and computer visualization , and its relation with reflections on a sphere . |
the two - nucleon spectral function , yielding the probability to remove two nucleons with momenta @xmath1 and @xmath2 from nuclear matter leaving the residual system with excitation energy @xmath3 , is defined as ( see , e.g. , ref.@xcite ) : @xmath4 where @xmath5 is the @xmath6-particle nuclear matter ground state with energy @xmath7 , @xmath8 denotes a @xmath9-nucleon intermediate state with energy @xmath10 and @xmath11 is the usual creation operator .
the two - nucleon spectral function carries relevant information on the short range structure of the nuclear medium and nucleon - nucleon ( nn ) correlations .
therefore , a considerable effort aimed at extracting experimental information on @xmath12 is currently being undertaken . in this context , a very important role
is played by electron - nucleus scattering experiments ( a number of theoretical and experimental topics in the field of electron - nucleus scattering are reviewed in ref.@xcite ) .
unambiguous evidence of strong correlation effects has been provided by single nucleon knock out @xmath13 reactions , showing that in a nucleus only about 70 % of the nucleons are in states of low momentum and low removal energy , that can be described by a mean field theory .
the remaining 30 % of the nucleons belong to strongly correlated pairs , whose occurrence is mainly to be ascribed to the one - pion - exchange tensor force and to the repulsive core of the nn interaction .
the available @xmath13 data at low missing momentum and low missing energy clearly show the depletion of the occupation probabilities of single particle states predicted by the nuclear shell model ( for a recent review see , e.g. , ref.@xcite ) , thus providing a somewhat indirect measurement of correlation effects .
complementary information will soon be available from single nucleon knock out experiments specifically designed to investigate the kinematical region corresponding to large missing momentum and large missing energy , where correlation effects are believed to be dominant@xcite . over the past decade
, the availability of a new generation of high energy 100 % duty - cycle electron beams has made it possible to carry out double coincidence @xmath14 experiments , in which two knocked out nucleons are detected . in principle , these experiments may allow for direct measurements of _ correlation observables _ , such as the momentum space two - nucleon distribution or the two - nucleon spectral function .
however , as pointed out in ref.@xcite , extracting the relevant dynamical information from the measured @xmath14 cross section may turn out to be a challenging task , requiring both a careful choice of the kinematical setup and a quantitative theoretical understanding of the final state interactions ( fsi ) of the knocked out nucleons . regardless of the difficulties involved in the interpretation of the experimental data , reliable theoretical calculations of the two - nucleon emission cross section carried out within the plane wave impulse approximation ( pwia ) , in which all
the nuclear structure information is contained in the two - nucleon spectral function , have to be regarded as a minimal starting point , which can provide guidance for the optimization of future experiments .
the basic assumptions underlying the pwia scheme are that i ) the exchanged virtual photon couples to either of the two outgoing nucleons and ii ) fsi effects are negligible .
the pwia cross section of the process in which an electron of initial energy @xmath15 is scattered into the solid angle @xmath16 with energy @xmath17 , while two nucleons of kinetic energies @xmath18 and @xmath19 are ejected into the solid angles @xmath20 and @xmath21 , respectively , takes the simple factorized form @xmath22 where @xmath23 describes the structure of the electromagnetic vertex , @xmath24 is the nucleon mass and @xmath25 and @xmath26 denote the momenta of the detected nucleons .
the missing momentum @xmath27 and missing energy @xmath28 are defined as @xmath29 @xmath30 being the momentum transfer , and @xmath31 where @xmath32 is the kinetic energy of the recoiling @xmath9-particle system .
the function @xmath33 appearing in eq.([pwia : xsec ] ) can be written in terms of the two - nucleon spectral functions @xmath34 and @xmath35 .
pioneering studies of the pwia two - nucleon emission cross section of the @xmath36 reaction are discussed in ref.@xcite , whereas a g - matrix perturbation theory calculation of the @xmath37 two - proton spectral function is described in ref.@xcite . in this paper
we discuss a calculation of the nuclear matter @xmath12 performed using correlated basis function ( cbf ) perturbation theory .
our theoretical approach , in which the effects of the nonperturbative components of the nn interaction are incorporated in the basis function using a variational approach , has proved to be particularly suited to describe quantities which are strongly affected by nn correlations , such as the single nucleon spectral function @xmath38@xcite , the nucleon momentum distribution @xmath39@xcite and the off - diagonal density matrix @xmath40@xcite .
following ref.@xcite , we evaluate the dominant contributions to @xmath12 at the zero - th order of cbf , as well as the perturbative corrections that have been shown to be relevant in the calculation of @xmath38 .
cbf theory of infinite nuclear matter is built on the set of _ correlated _ states : @xmath41 |n\rangle_{fg}\ , \label{basis}\ ] ] obtained by applying a symmetrized product of two - body correlation operators , @xmath42 , to the fermi gas states @xmath43 .
the structure of @xmath42 is similar to that of the nn interaction : @xmath44 the operators @xmath45 include four central components ( @xmath46 ) ) for @xmath47 and the isoscalar and isovector tensor ones @xmath48 and @xmath49 for @xmath50 .
additional spin - orbit components are sometimes introduced , but we will neglect them in this work . with transparent and widely employed notation , the components are also denoted as @xmath51 ( @xmath52 ) , @xmath53 , @xmath54 and @xmath55 ( tensor ) .
realistic correlated basis states are expected to be close to the eigenstates of the hamiltonian .
an efficient recipe to choose the correlation functions consists in their variational determination by minimizing the expectation value of the hamiltonian in the correlated ground state : @xmath56 the scalar ( or jastrow ) component @xmath57 heals to unity for @xmath58 , whereas @xmath59 .
the tensor components have the longest range , since they are related to the one - pion exchange potential and @xmath60 ( @xmath61 is the hamiltonian associated with a two - nucleon pair ) is a _
well - behaved _ operator healing to @xmath61 at large interparticle distances , like the g - matrix .
the results presented in this work have been obtained using the urbana @xmath62 potential supplemented by the tni three - nucleon interaction @xcite . at the zero - th ( or _
variational _ ) order of cbf , @xmath63 is obtained using the correlated states ( [ basis ] ) in eq.([tbsf ] ) . the variational estimate is then corrected by inserting perturbative corrections in the correlated basis .
this procedure has been already adopted for many nuclear matter properties and it has been particularly successful in describing inclusive electromagnetic responses at both intermediate and high momentum transfers ( see , e.g. , ref.@xcite ) . in section 2 we will discuss the relationships between the two - nucleon spectral function , the two - nucleon momentum distribution , and the two - body density matrix , @xmath64 .
section 3 will present theory and results for the variational two - nucleon spectral function , as well as the main cbf perturbative corrections , evaluated with realistic interactions .
finally , section 4 is devoted to conclusions and perspectives .
the nuclear matter two - body density matrix @xmath64 is defined as @xmath65 @xmath66 being the ground state wave function . for @xmath67 and @xmath68
, the density matrix reduces to the two - nucleon density distribution @xmath69 , yelding the joint probability of finding two nucleons at positions @xmath70 and @xmath71 in the nuclear matter ground state .
the two - nucleon momentum distribution , i.e. the probability that two nucleons carry momenta @xmath1 and @xmath2 , is related to the two - nucleon spectral function @xmath12 , defined in eq.([tbsf ] ) , through @xmath72 from eqs.([def : dm ] ) and ( [ def : nkk ] ) it follows that @xmath73 can be written in terms of the density matrix according to @xmath74 where @xmath75 denotes the degeneracy of the system ( in symmetric nuclear matter @xmath75 = 4 ) , @xmath76 is the normalization volume , @xmath77 and @xmath78 .
it is well known@xcite that in a fermi liquid _ with no long range order _ the two - nucleon momentum distribution factorizes according to @xmath79 where @xmath39 is the single nucleon momentum distribution , defined as @xmath80 the @xmath39 resulting from the nuclear matter calculation of ref.@xcite is shown in fig .
it exhibits a discontinuity at @xmath81 ( the fermi momentum @xmath0 is related to the density @xmath82 through @xmath83 ) , and a tail , extending to very large momenta , reflecting the structure of the nuclear wave function at small interparticle distance .
eq.([factor : nkk ] ) implies that in the a @xmath84 limit @xmath73 contains the same dynamical information carried by @xmath39 . some insight on the role of short range correlations
can be obtained rewriting @xmath73 in terms of the relative and center of mass momenta of the pair , defined as @xmath85 and @xmath86 , respectively , and studying the quantity @xmath87 fig .
2 illustrates the behaviour of @xmath88 at saturation density ( corresponding to @xmath89 @xmath90 ) , evaluated using the momentum distribution of ref.@xcite ( solid line ) , compared to the prediction of the fermi gas model ( dashed line ) .
the inclusion of nn interactions leads to a quenching of the peak , located at @xmath91 0.7 - 0.8 @xmath90 , and to the appearance of a sizeable tail at @xmath92 .
it is interesting to single out the contributions to @xmath88 corresponding to @xmath93 , @xmath94 and @xmath95 , and @xmath96 .
they are represented in fig . 2 by diamonds , squares and crosses , respectively ( note that the results shown by the crosses have been enhanced by a factor 10 ) .
the @xmath93 component provides about 72 @xmath97 of the normalization of @xmath88 and becomes vanishingly small at @xmath98 . on the other hand , the contributions coming from pairs in which at least one of the nucleon momenta is larger that @xmath0 are much smaller in size ( the curves marked with squares and crosses yield 26 @xmath97 and 2 @xmath97 of the normalization , respectively ) but extend up to @xmath91 4 @xmath90 .
similar results have been recently obtained @xcite using the momentum distribution resulting from a g - matrix perturbation theory calculation @xcite .
the propagation of two hole states in nuclear matter is described by the hole - hole part of the two - body green s function@xcite , @xmath99 where @xmath100 is the nuclear matter hamiltonian .
@xmath101 is defined for @xmath102 , @xmath103 being the fermi energy .
the two - nucleon spectral function , given in eq.([tbsf ] ) , is straightforwardly related to the imaginary part of @xmath101 by @xmath104 cbf perturbation theory employes the set of correlated states given in eq.([basis ] ) .
note that the basis states are non orthogonal to each other .
orthogonalization can been implemented using the procedure proposed in ref.@xcite , which preserves the diagonal matrix elements of the hamiltonian between correlated states .
this technique has been succesfully applied to the calculation of the one - body spectral function in ref.@xcite ( hereafter denoted i ) .
the perturbative expansion is obtained separating the nuclear hamiltonian @xmath100 into two parts : @xmath105 , and using standard rayleigh - schrdinger type expansions .
the unperturbed and interaction terms , denoted by @xmath106 and @xmath107 , respectively , are defined through @xmath108 @xmath109 the expectation value in eq.([2body : green ] ) is calculated by expanding @xmath110^n\vert 0 \rangle } \over { \vert\langle { \bar 0}\vert{\bar 0}\rangle \vert^{1/2 } } } \ , \label{cbf2}\ ] ] where @xmath111 .
perturbative corrections to the intermediate states @xmath112 are taken into account carrying out an analogous expansion in powers of @xmath113 for the propagator appearing in eq.([2body : green ] ) , as discussed in i : @xmath114 we will now describe the structure of the @xmath115 ( i.e. variational ) and higher order terms of the cbf perturbative expansion of @xmath12 . at the lowest order in the perturbative expansion the spectral function
is given by @xmath116 where the correlated states ( [ basis ] ) , and their energies @xmath117 and @xmath118 are used .
the main contributions to the sum in eq.([tbsf_v ] ) come from correlated two - hole ( @xmath119 ) and three - hole one - particle ( @xmath120 ) intermediate states . in the uncorrelated fermi gas
only @xmath119 states contribute to @xmath121 , leading to the well known result @xmath122 where @xmath123 denotes the usual step function .
the presence of correlations , besides changing the structure of the @xmath119 contribution in the way we will discuss below , allows for nonvanishing contributions from other intermediate states , with a consequent quenching of the @xmath119 peak since part of the strength is moved to higher excitation energies .
the variational @xmath119 spectral function is given by : @xmath124 where @xmath125 is the variational single particle energy and the @xmath119 overlap matrix element , @xmath126 , is @xmath127 the correlated @xmath128 overlap matrix element , @xmath129 , was computed in i using cluster expansion techniques .
the same method is used for the @xmath119 overlap , with the result that only _ unlinked _
cluster diagrams ( _ i.e. _ diagrams where the points reached by the @xmath130-lines are not connected to each other by any dynamical , @xmath131 and @xmath132 , or statistical correlations ) contribute .
the _ linked _ diagrams contribution turn out to be of order 1/a , and therefore vanish in nfinite nuclear matter .
note that the factorization property of the two - nucleon momentum distribution , illustrated by eq.([factor : nkk ] ) , is also a consequence of the fact that _ linked _ diagrams do not contribute in the a@xmath84 limit .
the definitions , as well as the technicalities entering the cluster expansion method can be found in i and in the related references and will not be repeated here . because of the factorization property , in nuclear matter we obtain @xmath133 the cbf expression of @xmath134 being given in i ( eq.(a.19 ) ) .
the correlated @xmath119 spectral function retains the fermi gas delta shaped peak .
however , correlations quench the peak itself ( via the @xmath119 overlap matrix element ) and move it to @xmath135 .
the @xmath120 contribution reads @xmath136 the overlap matrix element being given by @xmath137 in the limit a@xmath84 the contributions of fully linked diagrams vanish in the @xmath120 overlap as well . as a consequence , @xmath138 in the other cases we obtain @xmath139 and @xmath140\ , \label{phi3_3h1p}\ ] ] where the sums include all permutations of the @xmath141 indices , while the structure of the @xmath142 overlap @xmath143 is discussed in i ( appendix b ) . unlike the @xmath119 contribution ,
@xmath144 is no longer delta shaped , but it is rather spread out in energy .
it may be interpreted as a background contribution to be added to the two - quasiparticle part of the spectral function .
the next intermediate state to be considered is the @xmath145 one , whose contribution may be written in terms of products of two squared @xmath142 overlap matrix elements .
however , it is numerically negligible with respect to the @xmath120 term ( several order of magnitudes smaller ) and can be safely neglected .
the above intermediate states completely exhaust the momentum distribution sum rule in nuclear matter at the variational level : @xmath146 the admixture of @xmath24-hole @xmath147-particle correlated states in @xmath148 originate cbf perturbative corrections .
as far as the ground state is concerned , they correspond to the @xmath149 terms in eq.([cbf2 ] ) .
following the strategy developped in i , the perturbative corrections are classified as @xmath150 and @xmath151 , coming from ground and intermediate state corrections , respectively .
the total spectral function is then given by @xmath152 here we consider @xmath153 admixtures to the ground state and @xmath142 admixtures to the @xmath128 intermediate states .
the first ones contribute for all @xmath154 values ( both below and above @xmath0 ) , and their contribution can be expressed in terms of the correlation self energy @xmath155 , whose imaginary part is given by @xmath156 at @xmath157 and @xmath158 and @xmath159 at @xmath160 and @xmath161 .
the self energy is computed using correlated states and retaining two- and three - body separable contributions in the cluster expansion of the relevant matrix element @xcite .
the @xmath142 intermediate state contribution involves the polarization self energy @xmath162 .
the corresponding imaginary part reads @xmath163 at @xmath160 and @xmath158 and @xmath164 at @xmath157 and @xmath161 .
the effects of these corrections on the spectral functions at @xmath165 are : _
i ) _ a quenching of the variational two - quasiparticle strength and a shift of its position by a quantity @xmath166 , with @xmath167 from ground state corrections , and _
ii ) _ a broadening of the delta - like peak itself , roughly proportional to @xmath168 , from intermediate state admixtures .
the expression for the total @xmath169 at @xmath165 reads @xmath170\right .
\\ \nonumber & \times & \left [ \alpha({\bf k}_1,{\bf k}_2)(-e^v_{k_1}-e^v_{k_2}-e ) -\delta e_{k_1}-\delta e_{k_2 } \right .
\\ \nonumber & - & \re \left ( \sigma^{po}({\bf k}_1,e)- \sigma^{po}({\bf k}_1,e =- e^v_{k_1 } ) + 1\rightarrow 2 \right ) \\ & - & i \left . \left .
\left ( \sigma^{po}({\bf k}_1,e ) + 1\rightarrow 2 \right ) \right]^{-1 } \right\}\ , \label{deltap2 } \end{aligned}\ ] ] where @xmath171 .
the quantities @xmath172 , @xmath173 and @xmath174 are defined in i. fig .
[ spec:1 ] gives two examples of the two - nucleon spectral function below @xmath0 .
the figure shows the variational one - hole ( @xmath175 ) and two - hole ( @xmath176 ) peaks and the total spectral function ( solid line ) .
the one - nucleon spectral function below @xmath0 can be separated into a _ single particle _ and
a _ correlated _ part @xcite .
the former originates from @xmath128 intermediate states and their admixtures , the latter from @xmath147-hole @xmath177-particle states , and it would be strictly zero in absence of correlations .
the corresponding contributions for the two - nucleon spectral functions are shown in the figure as dashed and dot - dash lines , respectively .
the correlated part has been enhanced by a factor 10@xmath178 , while the single particle and total ones by a factor 10 .
the figure also gives the barely visible variational correlated part ( dotted lines ) .
ground state @xmath153 corrections give the leading contributions to @xmath169 when one of the momenta is above @xmath0 .
so , for @xmath179 and @xmath180 , we obtain @xmath181 where @xmath182 is again given in i. the spectral functions for two such cases are shown in fig .
[ spec:2 ] , together with the one - hole variational peaks corresponding to the momenta below @xmath0 .
the whole spectral function now comes from the correlated part .
the figure gives the total ( solid lines ) and variational ( dotted lines ) @xmath183 , as well as the perturbative corrections ( dashed lines ) .
corrections at @xmath184 are quadratic in the self energies and have not been considered .
we have carried out a calculation of the nuclear matter two - nucleon spectral function @xmath12 within the framework of cbf perturbation theory . the zero - th order approximation ,
corresponding to a variational estimate , has been supplemented with higher order corrections associated with both @xmath153 admixture to the ground state and @xmath142 admixtures to the @xmath128 states .
the results show that the inclusion of nn interactions produces drastic changes in the behavior of @xmath12 , with respect to the predictions of the fermi gas model .
the peaks correponding to the single particle states get quenched already at zero - th order , and aquire a width when higher order corrections are included .
due to strong short range nn correlations , the strength removed from the quasi particle peak is pushed to large values of momenta and removal energy , giving rise to a broad background . the factorization property exhibited by the two - nucleon momentum distribution of a normal fermi liquid
has been found to hold at the level of the amplitudes entering the calculation of @xmath12 as well . as a result ,
all the linked cluster contributions to the two - nucleon spectral function vanish in the infinite nuclear matter limit . in principle
, the results discussed in this paper may be used to estimate _
correlation observables _ , related to the small components of @xmath12 , whose measurement represents the ultimate goal of the @xmath14 experimental programs . however , it has to be emphasized that , in order to make fully quantitative predictions , our approach should be extended to finite systems , working along the line proposed in refs.@xcite .
the analysis of the a dependence of @xmath12 would also provide very interesting information on the relevance of the 1/a linked contributions and the applicability of the factorized approximations for the two - nucleon spectral function and momentum distribution . | the two - nucleon spectral function in nuclear matter is studied using correlated basis function perturbation theory , including central and tensor correlations produceded by a realistic hamiltonian .
the factorization property of the two - nucleon momentum distribution into the product of the two single nucleon distributions shows up in an analogous property of the spectral function .
the correlated model yields a two - hole contribution quenched whith respect to fermi gas model , while the peaks acquire a quasiparticle width that vanishes as the two momenta approach @xmath0 .
in addition , three - hole one - particle and more complicated intermediate states give rise to a background , spread out in energy and absent in the uncorrelated models .
the possible connections with one- and two - nucleon emission processes are briefly discussed .
# 1#2#3#4#1 * # 2 * , # 3 ( # 4 ) nuclear structure . nuclear matter . two - nucleon emission . |
diffusion strategies @xcite-@xcite were first invented to solve distributed estimation problems in real - time environments where data are continuously streamed . here , all nodes employ adaptive filter algorithms to process the streaming data , and simultaneously share their instantaneous estimates with their neighbors .
these approaches are also very useful to model many self - organizing systems @xcite .
recently , in @xcite-@xcite , diffusion lms schemes have been used to estimate sparse vectors , or equivalently , to identify fir systems that have most of the impulse response coefficients either zero or negligibly small . in these papers , certain sparsity promoting norms of the filter coefficient vectors
have been used to regularize the standard lms cost function , prominent amongst them being the @xmath0 norm of the coefficient vector that leads to the sparsity aware , zero attracting lms ( za - lms ) @xcite-@xcite form of weight adaptation .
these diffusion sparse lms algorithms manifest superior performance in terms of lesser steady state network mean square deviation ( nmsd ) compared with the simple diffusion lms . in this paper , we show that the minimum level of the steady state nmsd achieved using za - lms based update at _ all _ the nodes of the network can also be obtained by a _
heterogeneous _ network with only a fraction of the nodes using the za - lms update rule ( referred as sparsity aware nodes in this paper ) while the rest employing the standard lms update ( referred as sparsity agnostic nodes in this paper ) , provided the nodes using the za - lms are distributed over the network maintaining some uniformity " .
note that reduction in the number of sparsity aware nodes reduces the overall computational burden of the network , especially when more complicated sparsity aware algorithms involving significant amount of computation are deployed to exploit sparsity . as shown in this paper , the only adjustment to be made to achieve the above reduction in the number of sparsity aware nodes is a proportional increase in the value of the optimum zero attracting coefficient .
analytical expressions explaining the above behavior are provided and the claims made are validated via detailed simulation studies . finally , the proposed analysis ,
though restricted to the @xmath0-norm regularized algorithm ( i.e. , za - lms ) only , can be trivially extended to the case of more general norms and thus similar behavior can also be expected from the corresponding heterogeneous networks .
we consider a connected network consisting of @xmath1 nodes that are spatially distributed . at every time index @xmath2
, each @xmath3 node collects some scalar measurement @xmath4 and some @xmath5 vector @xmath6 which are related by the following model : @xmath7 where @xmath8 is the measurement noise at the @xmath3 node and @xmath9 is the unknown @xmath10 vector , known a priori to be sparse , which is required to be estimated .
both @xmath6 and @xmath8 are variates generated from some gaussian distributions , with @xmath6 and @xmath11 being mutually independent for all @xmath12 . in the diffusion scheme , every @xmath13-th node , @xmath14 deploys a @xmath10 adaptive filter @xmath15 to estimate @xmath9 , which takes @xmath4 and @xmath16 respectively as the local desired response and input vectors .
the estimates of @xmath9 , i.e. , @xmath17 for each @xmath13 are exchanged with the neighbors of the @xmath13-th node , i.e. , nodes directly connected to it , and are used to refine the estimates in one of the two following manners : ( a ) adapt - then - combine ( atc ) where @xmath17 is first updated to an intermediate estimate @xmath18 , which is then linearly combined with similar estimates received from the neighbors , and ( b ) combine - then - adapt ( cta ) where @xmath17 is first linearly combined with similar estimates received from the neighbors and then updated .
originally , the diffusion schemes were proposed assuming lms form of weight adaptation at each node @xcite-@xcite . in the context of sparse estimation , certain sparsity exploiting norms of @xmath17 were added to the corresponding lms cost function @xcite-@xcite , the most popular of them being the @xmath0 norm penalty @xmath19 which results in the introduction of the zero attracting terms @xmath20 $ ] in the lms update equations @xcite-@xcite .
the resulting diffusion za - lms algorithm for the atc scheme , popularly termed as za - atc diffusion algorithm @xcite , is shown in table i and is considered by us in this paper .
the parameter @xmath21 in table i is the zero - attracting coefficient which is a very very small , positive constant taken same for all the nodes and @xmath22 denotes the set of nodes in the neighborhood of the node @xmath13 ( including itself ) .
[ table : za_atc ] the combining coefficients @xmath23 are non - negative constants which are usually chosen satisfying the following @xcite : @xmath24 there exist several standard schemes in the literature to choose the coefficients @xmath25 , e.g. , the uniform combination rule , the metropolis rule , the laplacian rule and the nearest neighbor rule to name a few . using these coefficients ,
a combination matrix @xmath26 is defined for the network , where @xmath27_{i , j}=c_{i , j}^{'}.$ ]
before presenting the proposed heterogeneous network and its at par behavior with the za - atc based diffusion network of @xcite , it will be useful to consider some of the major results of @xcite here .
for this , we first define the average network mean - square deviation at the @xmath28 time index as , @xmath29 where @xmath30 is the individual mean - square deviation of the @xmath3 node at the @xmath28 time index , i.e. , @xmath31 = e[\|{\bf { \tilde w}}_k(n)\|^2]\ ] ] where @xmath32 is the weight deviation vector for the @xmath13-th node at @xmath2-th index .
the expression for steady - state @xmath33 ( i.e. , @xmath34 ) of the za - atc algorithm was derived analytically in @xcite . however , @xcite considered a more general form of diffusion , in which both @xmath35 and @xmath36 are also exchanged with the neighbors along with the local estimates @xmath37 .
in contrast , in this paper , we consider exchange of only @xmath38 which is also the most common form of diffusion .
additionally , we introduce a few more simplifications in @xcite .
firstly , we assume same step - size @xmath39 for all nodes . next
, both the input signal and noise at each node are assumed to be spatially and temporally i.i.d . under these , it is easy to check that the @xmath34 expression for the za - atc algorithm @xcite simplifies to the following : @xmath40^t({\bf i } - { \bf f})^{-1}{\bf q } \nonumber\\ & + & \frac{1}{n}(\beta(\infty ) - \alpha(\infty ) ) , \label{eq : msd}\end{aligned}\ ] ] with @xmath41^t{\boldsymbol \varomega}{\bf c}{\bf c}^t({\bf i } - \mu{\bf d}){\bf { \tilde w}}(\infty)]\ ] ] and @xmath42\|_{{\boldsymbol \varomega}{\bf c}{\bf c}^t{\boldsymbol \varomega}}^{2}],\ ] ] where @xmath43 is an operator that stacks the columns of its argument matrix on top of each other , @xmath44 , @xmath45 and @xmath46 , with @xmath47 denoting an operator that carries out stacking of its argument column vectors on top of each other , and @xmath48 and @xmath49 are the variances of the noise and input signal respectively .
the matrices @xmath50 , @xmath51 , @xmath52 and @xmath53 are defined as follows : + @xmath54 [ @xmath55 defines the right kronecker product . ]
+ @xmath56 + @xmath57 , + @xmath58 .
[ also note that for a vector @xmath59 and a matrix @xmath60 , @xmath61 indicates @xmath62 . ]
it is noticed that the first term in the r.h.s . of ( 6 )
is actually the steady - state network msd of simple atc diffusion lms @xcite and is independent of @xmath21 .
let us denote the second term as @xmath63 , i.e. , @xmath64 .
it is easy to see that one can express @xmath63 as @xmath65 , where , @xmath66^t{\bf c}{\bf c}^t({\bf i } - \mu{\bf d}){\bf { \tilde w}}(\infty)]\ ] ] and @xmath67\|_{{\bf c}{\bf c}^t}^{2}]~(>\;0).\ ] ] the function @xmath63 has two zero - crossing points , one at @xmath68 and the other at @xmath69 , and between them , @xmath63 takes only negative values with the minima occurring at @xmath70 , which , from ( 6 ) , also minimizes @xmath34 . for systems that are highly sparse
, it follows from @xcite that @xmath71 and conversely , for non - sparse systems , @xmath72 . since for proper zero attraction ,
@xmath21 must be positive , the optimum value of @xmath21 is then given by @xmath73.\ ] ] the corresponding minimum value of @xmath74 ( when @xmath75 ) is then given as @xmath76 * the proposed heterogeneous diffusion network :* + + in this section , we show that the same level of @xmath77 as given by ( 10 ) and therefore , the same @xmath78 $ ] can be reached by a heterogeneous network as well , where only a fraction of the nodes are sparsity aware and rest are sparsity agnostic , provided the network is designed satisfying the assumptions i.a and i.b as given in the box below where @xmath79 denotes the set of indices of the sparsity aware nodes and : + + in order to show the above , we replace the matrix @xmath80 by a new one defined as @xmath81,$ ] where @xmath82 . using this and the fact that @xmath83 , @xmath84 and @xmath85 modify to @xmath86 and @xmath87 , given as follows : @xmath88^t{\boldsymbol \varomega}_s{\bf c}{\bf c}^{t}{\bf { \tilde w}}(\infty)]\end{aligned}\ ] ] and @xmath89^t{\boldsymbol \varomega}_s{\bf c}{\bf c}^{t}{\boldsymbol \varomega}_s sgn[{\bf w}(\infty ) ] ] \end{aligned}\ ] ] note that unlike @xmath84 and @xmath85
, it is lot more difficult to express @xmath86 and @xmath87 as a function of @xmath21 , since unlike @xmath53 , @xmath90 can not be written simply as @xmath91 .
instead , one needs to analyze the rhs of ( 11 ) and ( 12 ) to express @xmath86 and @xmath87 in terms of @xmath21 . towards this
, we make the following assumptions : + it is then possible to prove the following : + for a network satisfying the @xmath92 and @xmath93 as given above , we have , @xmath94 n_s.\end{aligned}\ ] ] proof : skipped due to page limitation .
+ for a network satisfying the @xmath95 , @xmath96 and @xmath97 as given above , we have , @xmath98n_{s}^{2}}{n}.\end{aligned}\ ] ] proof : skipped due to page limitation . + substituting @xmath99 and @xmath100 in @xmath101 , then differentiating w.r.t . @xmath21 and equating the derivative to zero , we obtain , @xmath102n}{\mu tr[{\boldsymbol \psi}]n_s}].\ ] ] the corresponding minimum value of @xmath63 [ when @xmath75 , i.e. , the system is sparse ] , say , @xmath103 is given as @xmath104 ^ 2}{tr[{\boldsymbol \psi}]}.\ ] ] note that @xmath103 as given in ( 16 ) is independent of @xmath105 . therefore , _ its value remains same when @xmath106 , i.e. , when the network becomes _ _ homogeneous with all nodes being sparsity aware_. this also implies that if @xmath77 as given by ( 10 ) is analyzed using the assumptions i and ii , it would give rise to the same expression as that of @xmath103 ( i.e. , ( 16 ) ) . from this and ( 15 ) , we then make the following two conclusions : + + @xmath107 the @xmath78 $ ] does not change when the network changes from being homogeneous to heterogeneous , with only @xmath105 of the total @xmath1 ( @xmath108)nodes employing sparsity aware adaptation .
+ @xmath107 for sparse systems , the @xmath109 minimizing @xmath63 and thus @xmath110 $ ] ( i.e. , @xmath111n}{\mu tr[{\boldsymbol \psi}]n_s}$ ] as given in ( 15 ) ) is inversely proportional to @xmath105 , meaning that while maintaining the same @xmath78 $ ] , one can reduce the number of sparsity aware nodes by introducing proportional increase in the value of @xmath21 .
to test the performance of the heterogeneous networks , we use a strongly connected network of @xmath112 nodes placed randomly in a geographic region .
the weights of the edges are determined by the uniform combination rule @xcite .
the goal of the network is to estimate a @xmath113 vetor @xmath9 which is highly sparse ( only one coefficient being non - zero ) .
we choose the same step - size @xmath114 for all the nodes . among these @xmath115 nodes ,
@xmath105 number of nodes use the za - lms and rest of the nodes use simple lms update , with the former spaced uniformly ( i.e. , satisfying assumptions i.a and i.b ) over the network .
the input signals and noise variables are drawn from gaussian distributions , and they are temporally and spatially independent .
also , the input and noise statistics are same for all the nodes , with @xmath116 , and @xmath117 . to start with
, the value of @xmath21 is kept fixed at @xmath118 for all the @xmath105 sparsity aware nodes .
the simulation is then carried out for @xmath119 iterations and the network steady state msd is evaluated by taking ensemble average over @xmath120 independent runs .
this is done for different values of @xmath105 ( ranging from @xmath121 to @xmath115 ) and based on this , the network steady state msd is plotted as a function of @xmath105 . the value of @xmath21 is then increased progressively to take the following five values : @xmath122 , one at a time for all the za - lms based nodes .
1 displays the network steady state msd vs. @xmath105 plots with @xmath21 as a parameter .
it is easily seen from fig .
1 that ( i ) the minima reached by each msd - vs-@xmath105 plot is same for all the plots , and ( ii ) as @xmath21 increases , the value of @xmath105 where the minima occurs reduces and vice versa . in other words ,
1 validates the theoretical conjectures made in the previous section .
1 a. h. sayed , `` diffusion adaptation over networks , '' in _ e - reference signal processing _
, r. chellapa and s. theodoridis , eds .
amsterdam , the netherlands : elsevier , available online at http://arxiv.org/abs/1205.4220 , to be published .
y. gu , y. chen and a. o. hero ,
sparse lms for system identification " , _ proc .
ieee intl .
taipei , taiwan , _ apr .
k. shi and p. shi , `` convergence analysis of sparse lms algorithms with @xmath0-norm penalty based on white input signal , '' _ signal process .
3289 - 3293 , dec . 2010 .
bijit kumar das and m. chakraborty `` sparse adaptive filtering by an adaptive convex combination of the lms and the za - lms algorithms , '' _ ieee trans .
circuits syst.i , reg.papers_ , vol .
5 , pp . 1499 - 1507 , may 2014 . | in - network distributed estimation of sparse parameter vectors via diffusion lms strategies has been studied and investigated in recent years . in all the existing works ,
some convex regularization approach has been used at each node of the network in order to achieve an overall network performance superior to that of the simple diffusion lms , albeit at the cost of increased computational overhead . in this paper , we provide analytical as well as experimental results which show that the convex regularization can be selectively applied only to some chosen nodes keeping rest of the nodes sparsity agnostic , while still enjoying the same optimum behavior as can be realized by deploying the convex regularization at all the nodes . due to the incorporation of unregularized learning at a subset of nodes ,
less computational cost is needed in the proposed approach .
we also provide a guideline for selection of the sparsity aware nodes and a closed form expression for the optimum regularization parameter . * index terms*adaptive network , diffusion lms , sparse systems , excess mean square error , adaptive filter , @xmath0 norm . |
the discovery of the @xmath2 meson by the collider detector at fermilab(cdf ) collaboration @xcite in @xmath4 collisions at @xmath5 tev and the subsequent measurement of its lifetime have provided a new window for the analysis of the heavy - quark dynamics and thus for an important test of quantum chromodynamics .
recently the cdf and d0 collaborations announced some new measurements of the @xmath2 meson lifetime and mass @xcite , @xmath6 ps @xcite , @xmath7 mev @xcite , and @xmath8 mev @xcite .
the lhc is expected to produce around @xmath9 @xmath2 events per year @xcite .
this will provide more detailed information on the decay properties of the @xmath2 meson .
since the @xmath2 mesons carry flavor explicitly(@xmath10 and @xmath11 ) and can not annihilate into gluons , they are stable against strong and electromagnetic annihilation processes .
the decays of the @xmath2 meson are therefore only via weak interactions , which can be divided into three classes at the quark level : ( 1 ) the @xmath12 ( @xmath13 ) transition with the @xmath11 quark being a spectator , ( 2 ) the @xmath14 ( @xmath15 ) transition with the @xmath10 quark being a spectator , and ( 3 ) the weak annihilation channels .
although the phase space of the @xmath16 transitions is much smaller than the phase space of the @xmath17 transitions , the cabibbo - kobayashi - maskawa(ckm ) matrix elements are greatly in favor of the @xmath11 quark decay , i.e.@xmath18 .
in fact , the @xmath11-quark decays provide about @xmath19 of the @xmath2 decay width while the @xmath10-quark decays and the weak annihilation yield about 20@xmath20 and 10@xmath20 , respectively @xcite .
this indicates that both @xmath10-and @xmath11-quark decay processes contribute to the @xmath2 decay width on a comparable footing .
because the @xmath10 and @xmath11 quarks can decay individually and the @xmath2 meson has a sufficiently large mass , one can study a great variety of decay channels .
there have been many theoretical efforts to calculate the semileptonic @xcite and non - leptonic @xcite decays of the @xmath2 meson .
the semileptonic @xmath2 decays provide a good opportunity to measure not only the ckm elements such as @xmath21 , @xmath22,@xmath23 and @xmath24 but also the weak form factors for the transitions of @xmath2 to bottom and charmed mesons .
the nonleptonic @xmath2 decays , in which only hadrons appear in the final state , are strongly influenced by the confining color forces among the quarks .
while in the semileptonic transitions the long - distance qcd effects are described by a few hadronic form factors parametrizing the hadronic matrix elements of quark currents , the nonleptonic processes are complicated by the phenomenon of the quark rearrangement due to the exchange of soft and hard gluons .
the theoretical description of the nonleptonic decays involves the matrix elements of the local four - quark operators .
although the four - quark operators are more complicate than the current operators involved in the semileptonic decays , the nonleptonic decays of the heavy mesons are useful for exploring the most interesting aspect of the qcd , i.e. its nonperturbative long - range character . in our recent paper @xcite , we analyzed the semileptonic @xmath2 decays such as @xmath25 and @xmath26 using our light - front quark model(lfqm ) based on the qcd - motivated effective lf hamiltonian @xcite .
the weak form factors @xmath27 for the semileptonic decays between two pseudoscalar mesons are obtained in the @xmath28 frame ( @xmath29 ) and then analytically continued to the timelike region by changing @xmath30 to @xmath31 in the form factor .
the covariance ( i.e. , frame independence ) of our model has been checked by performing the lf calculation in the @xmath28 frame in parallel with the manifestly covariant calculation using the exactly solvable covariant fermion field theory model in @xmath32 dimensions .
we also found the zero - mode contribution to the form factor @xmath33 and identified the zero - mode operator that is convoluted with the initial and final state lf wave functions . in this paper , we extend our previous lfqm analysis of the semileptonic @xmath2 decays @xcite to the nonleptonic two - body decays of @xmath2 mesons such as @xmath34 and @xmath35(here @xmath36 and @xmath37 denote pseudoscalar and vector mesons , respectively ) . the qcd factorization approach is widely used since it works reasonably well in heavy - quark physics @xcite .
the factorization approximates the complicated non - leptonic decay amplitude into the product of the meson decay constant and the form factor . a justification of this assumption is usually based on the idea of color transparency @xcite .
we shall use the form factors for semileptonic @xmath38 decays as well as the meson decay constants obtained in our lfqm @xcite as input parameters for the nonleptonic @xmath2 decays . as done by many others @xcite
, we consider only the contribution of current - current operators at the tree level and calculate the decay widths for various nonleptonic @xmath2 decays . as far as the decay width is concerned , the contribution from the tree diagram is much larger than that from the penguin diagram .
the penguin contribution may be important in evaluating the @xmath39 violation and looking for new physics beyond the standard model , which we do not consider in this work .
the paper is organized as follows . in sec .
ii , we discuss the weak hamiltonian responsible for the nonleptonic two - body decays of the @xmath2 meson . in sec .
iii , we present the input parameters such as the weak decay constants and the form factors obtained in our lfqm @xcite based on the qcd - motivated effective hamiltonian @xcite .
the mixing angles between @xmath40 and @xmath41 mesons are also analyzed , both in octet - singlet and quark - flavor bases , to extract the decay constants relevant to @xmath40 and @xmath41 mesons .
section iv is devoted to the numerical results .
a summary and conclusions follow in sec.v .
the nonleptonic weak decays are described in the standard model by a single @xmath42 boson exchange diagram at tree level . in the standard model ,
the nonleptonic @xmath2 decays are described by the effective hamiltonian , which was obtained by integrating out the heavy @xmath42 boson and top quark . for the case of @xmath17 and @xmath16 transitions at the quark level , neglecting qcd penguin operators , one gets the following effective weak hamiltonian : [ hbcu ] h^bc(u)_eff&=&\ { v_cb[c_1()o^cb_1 + c_2()o^cb_2 ] + & & + v_ub[c_1()o^ub_1 + c_2()o^ub_2 ] + h.c . } , and [ hcsd ] h^cs(d)_eff&=&\ { v_cd[c_1()o^cd_1 + c_2()o^cd_2 ] + & & + v_cs[c_1()o^cs_1 + c_2()o^cs_2 ] + h.c . } , where @xmath43 is the fermi coupling constant and @xmath44 are the corresponding ckm matrix elements .
we use the central values of the ckm matrix elements quoted by the particle data group(pdg ) @xcite that we summarize in table [ t1n ] .
.values for ckm matrix elements used in this work . [ cols="^,^,^,^,^,^",options="header " , ]
in our numerical calculations of exclusive @xmath2 decays , we use two sets of model parameters ( @xmath45 ) for the linear and ho confining potentials given in table [ t2n ] to compute the weak form factors for semileptonic @xmath38 decays and the unmeasured decay constants as given in tables [ t3n ] and [ t4n ] , respectively . using them together with the ckm matrix elements given by table [ t1n ] we finally predict the branching ratios which are given in tables [ t5 ] and [ t6 ] . although we show the form factors in table [ t3n ] only at a maximum recoil point @xmath46 , we use the form factor at @xmath47 obtained from @xcite for the corresponding nonleptonic @xmath48 decays . in table [ t5 ] we show the nonleptonic decay widths of the @xmath2 meson for a general value of the wilson coefficients @xmath49 and @xmath50 , whereas in table [ t6 ] we give the corresponding branching ratios ( in @xmath20 ) at the fixed choice of wilson coefficients @xcite : @xmath51 and @xmath52 relevant for the nonleptonic decays of the @xmath53 quark . for the lifetime of the @xmath2 , we take the central value @xmath54 ps ( i.e. @xmath55 gev ) presented in pdg @xcite . our branching ratios for both @xmath10 and @xmath11 induced decays listed in table [ t6 ] are generally close to the other quark model results @xcite but differ substantially from the ones obtained by refs .
@xcite .
the relative size of the branching ratios for various decay modes may be estimated from power counting of the wilson coefficients @xmath56 and the ckm factors with respect to the small parameter of the cabibbo angle @xmath57 in the wolfenstein parametrization @xcite , e.g. the ckm matrix elements can be expanded in terms of @xmath58 as @xmath59 , @xmath60 , @xmath61 , @xmath62 , @xmath63 , and @xmath64 . from tables
[ t5 ] and [ t6 ] , we make the following observations : \(1 ) the class i decay modes determined by the wilson coefficient @xmath49 have comparatively large branching ratios .
the ckm favored @xmath11 decays such as @xmath65 decays with the ckm factor @xmath66 have branching ratios of the order of @xmath67 ( e.g. @xmath68 in our model predictions ) , which are the most promising class i decay modes shown in tables [ t5 ] and [ t6 ] .
the ckm - suppressed @xmath11 decays such as @xmath69 with @xmath70 and @xmath71 with @xmath72 , as well as the ckm - suppressed @xmath10 decays such as @xmath73 with @xmath74 , have branching ratios of the order of @xmath75 and should still be accessible at high luminosity hadron colliders . however , the branching ratios of the @xmath76 induced decay modes are too small @xmath77 to be measured experimentally .
\(2 ) the branching ratios for the class ii decay modes determined by @xmath50 are relatively smaller than those for the class i decay modes
. however , the @xmath78 decays with @xmath79 have branching ratios of the order of @xmath75 and these modes should be accessible experimentally . of interest is the abnormally small branching ratio of @xmath80 compared to that of @xmath81 . as stated in @xcite ,
the reason for such a small branching ratio is not only because the available physical phase space is too small but also because there are large destructive interferences between @xmath82 and @xmath83 due to the serious cancellation between the ckm factors @xmath84 and @xmath85 .
\(3 ) the class iii decay modes involve the pauli interference . taking into account the negative value of @xmath50 with respect to @xmath49
, one can see that the class iii decay modes shown in table [ t5 ] should be suppressed in comparison with the cases in which the interference is switched off . in order to test the effects of the interference
, one may put the widths in the form of @xmath86 , where @xmath87 and @xmath88 , and then compute the @xmath89 ( in @xmath20 ) as done in @xcite .
our absolute values of @xmath89 obtained from the linear ( ho ) model are 34.0 ( 34.0)@xmath20 for @xmath90 , 37.3 ( 42.4)@xmath20 for @xmath91 , 18.4 ( 15.1)@xmath20 for @xmath92 , and 20.2 ( 18.9)@xmath20 for @xmath93 , respectively .
this indicates that the interference is the most significantly involved in the @xmath91 decay compared to others .
in particular , the @xmath94 and @xmath95 decay modes have been proposed in @xcite for the extraction of the ckm angle @xmath96 through amplitude relations .
in this work , we have studied the exclusive nonleptonic @xmath97 decays where the final state @xmath98 mesons are factored out in the qcd factorization approach .
the inputs used to obtain their branching ratios were the weak form factors for the semileptonic @xmath38 decays in the whole kinematical region and the unmeasured weak decay constants obtained from our previous lfqm analysis @xcite .
for the measured values of decay constants , we use the central values extracted from the experimental measurements @xcite .
our predictions for the branching ratios are summarized in tables [ t5 ] and [ t6 ] and compared with other theoretical results .
overall , the class ii decay modes have more discrepancies among the theoretical models than the class i and iii decay modes do .
the upcoming experimental measurements of the corresponding decay rates can examine various theoretical approaches .
the most promising measurable decay modes appear to be the ckm favored @xmath11 decays such as @xmath65 decays .
it is thus expected that the dominant contribution to the @xmath2 total rate comes from the @xmath11 induced decays .
the more @xmath11 induced @xmath99 and @xmath100 decay modes seem to deserve further consideration . the work of h .-
m.choi was supported by the korea research foundation grant funded by the korean government(krf-2008 - 521-c00077 ) and that of c .-
r.ji by the u.s .
department of energy(no .
de - fg02 - 96er40947 ) .
99 f. abe _
et al._(cdf collaboration ) , ; .
a. abulencia _
et al._(cdf collaboration ) , ; t. aaltonen _ et al._(cdf collaboration ) , .
v.m . abazov _ et al._(d0 collaboration ) , .
_ et al._(quarkoninum working group ) , report no .
cern-2005 - 005 ; m.p .
altarelli and f. teubert , arxiv:0802.1901 .
i. p. gouz , v. v. kiselev , a. k. likhoded , v. i. romanovsky , and o. p. yushchenko , phys .
nucl . * 67 * , 1559 ( 2004 ) ; yad . fiz .
* 67 * , 1581 ( 2004 ) . v. v. kiselev , a. e. kovalsky , and a. k. likhoded , ; v. v. kiselev , a. k. likhoded , and a. i. onishchenko , . m. a. ivanov , j. g. krner and p. santorelli , .
d. ebert , r. n. faustov and v. o. galkin , .
d. ebert , r. n. faustov and v. o. galkin , .
chang and y .- q . chen , .
f . liu and k .-
a. abd el - hady , j. h. munoz , and j. p. vary , . p. colangelo and f. de fazio , .
e. hernndez , j. nieves and j. m. verde - velasco , .
d. du and z. wang , .
r. dhir , n. sharma , and r.c .
p. colangelo , g. nardulli , and n. paver , .
t. huang and f. zuo , .
m. a. ivanov , j. g. krner and p. santorelli , .
m. a. ivanov , j. g. krner and p. santorelli , .
d. ebert , r. n. faustov and v. o. galkin , .
m. a. nobes and r. m. woloshyn , .
w. wang , y .- l . shen , and c .- d .
m. lusignoli and m.masetti , .
s. godfrey , .
anisimov , p. yu . kulikov , i.m .
narodetskii , and k.a .
ter - martirosyan , phys .
atom nucl .
* 62 * , 1739 ( 1999)[yad
* 62 * , 1868 ( 1999 ) ] .
verma and a. sharma , ; * 64 * , 114018(2001 ) .
a.k . giri , b. mawlong , and r. mohanta , ; * 76 * , 099902(e ) ( 2007 ) ; a.k . giri , r. mohanta , and m.p .
kiselev , .
x. liu and x.q .
cheng , d.s .
du , and c.d .
s. fajfer , j.f .
kamenik , and p. singer , .
r. fleischer and d. wyler , .
kiselev , o.n .
pakhomova , and v.a .
castro , h.b .
mayorga , and j.h .
saleev , phys . at
. nucl . * 64 * , 2027 ( 2001 ) ; o.n .
pakhomova and v.a .
salev , phys . at .
nucl . * 63 * , 1999 ( 2000 ) .
y.s . dai and d.s .
du , ; d.s .
du and z.t .
m. masetti , .
q.p . xu and a.n .
j. sun , y. yang , w. du , and h. ma , .
ivanov , j.g .
krner , o.n .
pakhomova , .
f. hussain and m. d. scadron , .
choi and c .- r .
choi and c .- r .
choi and c .- r .
ji and h .-
choi , c .- r .
ji , and l.s .
kisslinger , .
choi , ; j. korean phys .
53 * , 1205 ( 2008 ) .
m. bauer , b. stech , and m. wirbel , .
m. neubert and b. stech , adv .
high energy phys .
* 15 * , 294 ( 1998)[hep - ph/9705292 ] .
cheng , ; l.l .
chau and h.y .
a. szczepaniak , e. henley , and s.j .
brodsky , .
n. deshpande , m. gronau , and d. sutherland , ; .
j. d. bjorken , nucl .
suppl . ) * 11 * , 325 ( 1989 ) . c. amsler _ et al._(particle data group ) , .
m. beneke , g. buchalla , m. neubert , c. t. sachrajda , ; .
a. j. buras and l. silvestrini , . h .-
choi and c .- r .
d. scora and n. isgur , .
choi and c .- r .
ji , ; ; s. j. brodsky and d. s. hwang , .
w. jaus , .
t. feldmann , p. kroll , and b. stech , ; .
t. feldmann , .
h. leutwyler , nucl .
b(proc . suppl . )
* 64 * , 223 ( 1998 ) .
adler , phys . rev .
* 177 * , 2426 ( 1969 ) ; j.s .
bell and r. jackiv , nuovo cimento a * 60 * , 47 ( 1969 ) ; j. gasser and h. leutwyler , ; j.f .
donoghue , b.r .
holstein , and y.c.r .
m. artuso _ et al._(cleo collaboration ) , .
et al._(cleo collaboration ) , .
k.w . edwards _ et al._(cleo collaboration ) , .
l. wolfenstein , . | we study exclusive nonleptonic two - body @xmath0 decays with @xmath1(pseudoscalar or vector meson ) factored out in the qcd factorization approach .
the nonleptonic decay amplitudes are related to the product of meson decay constants and the form factors for semileptonic @xmath2 decays . as inputs in obtaining the branching ratios for a large set of nonleptonic @xmath2 decays , we use the weak form factors for the semileptonic @xmath3 decays in the whole kinematical region and the unmeasured meson decay constants obtained from our previous light - front quark model .
we compare our results for the branching ratios with those of other theoretical studies . |
the dawn and dusk terminators of hd 189733b represent some of the best - studied atmospheric regions among the known transiting exoplanets , with the spectral coverage ranging from the ultraviolet ( uv ) to the mid - infrared ( mir )
. however , there remain discrepancies between spectra measured by the same instruments ( specifically , the planet - star radius ratio as denoted by @xmath2 ) , including : ( i ) disagreement in @xmath3/irac measurements @xcite , thought to be due to different data reduction processes applied to the same transit measurements @xcite ; and ( ii ) disagreement in @xmath4/nicmos observations due to different approaches to transit curve decorrelation @xcite .
the discrepancies between different groups stem partly from difficulties in unequivocally correcting for instrumental effects . at wavelengths shorter than 1 @xmath0 m
, the transmission spectrum decreases monotonically from the uv to the near - infrared ( nir ) , as measured by @xmath4/acs ( 0.551.05 @xmath0 m ) @xcite and @xmath4/stis ( 0.290.57 @xmath0 m ) @xcite .
this sloped spectrum is consistent with the presence of particles in the rayleigh scattering regime and suggests the presence of aerosols in the terminator regions .
@xcite suggested that a thick and vertically extended aerosol layer consisting of silicate grains such as enstatite ( mgsio@xmath1 ) or forsterite ( mg@xmath5sio@xmath6 ) is able to reproduce the rayleigh slope . by applying stellar spot corrections for all available measurements between 0.324 @xmath0 m
, @xcite suggested that the opacity from an optically thick aerosol layer governs the entire transmission spectrum and that the poor quality of the present data can not be used to identify molecular signatures in the terminators .
given this uncertainty in the data , and the degeneracies inherent in their reduction and analysis , the atmospheric properties for the terminators differ from study to study . in this study
we revisit the atmosphere of the day / night terminators of hd 189733b , considering the transmission spectrum reassessed by @xcite , where the maximum information content from the previous transit data were extracted .
it clearly reveals a precipitous slope in the @xmath71 @xmath0 m spectrum and a flatter spectrum in the ir . in particular
, the spectrum near 1 @xmath0 m is particularly important in determining which opacity source is more plausible in the terminators in the ir , i.e. , optically thick aerosols washing out all other molecular signatures @xcite ; or a combination of aerosols and molecules @xcite .
the main objective of this study is therefore to assess the family of plausible aerosol models that are consistent with the measurements currently available by exploring degeneracies between aerosol and other atmospheric parameters , i.e. , temperature , planetary radius and molecular composition .
we also aim to investigate the influence of different types of aerosols on the transmission spectrum between 0.324 @xmath0 m , demonstrating the importance of measuring a wide spectral range to break the degeneracies between competing atmospheric parameters .
atmospheric properties were derived using the nemesis spectral retrieval tool @xcite , seeking the best - fits across a pre - defined parameterized space of aerosol properties : nadir ( or radial ) optical depth ( @xmath8 ) and particle radius ( @xmath9 ) .
the vertical locations and extent of the aerosol layers are poorly characterised with existing measurements , so we choose to define a simple and phenomenological aerosol model following @xcite , where aerosols are uniformly distributed throughout the pressures of the model atmosphere ( 10@xmath1010 bar ) ( uc , the number of particles per cubic centimeter is vertically uniform ) .
this allows us to reduce the number of free parameters during the retrieval process .
more sophisticated aerosol models are commonly used for the study of sub - stellar objects ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but we find that the existing data quality and low signal - to - noise does not warrant the use of more complex aerosol models at this time . in this study , we explore broad ranges for two aerosol parameters , i.e. , @xmath8 and @xmath9 for various atmospheric conditions so that we are able to characterise the degeneracies between the aerosol properties and other parameters , e.g. , thermal structure , planetary radius , aerosol composition , gaseous compositions .
nemesis finds the best - fitting atmospheric state ( composition and aerosol properties ) that reproduces the observed transit spectrum iteratively , where the next step of retrieval is analytically calculated using the technique of optimal estimation @xcite .
this spectral inversion software was modified to permit analyses of exoplanet spectroscopy and to seek the range of plausible solutions for exoplanetary atmospheres with realistic uncertainties @xcite . to achieve the aerosol information that can be derived from the given data , we grid @xmath8 and @xmath9 over a wide range of values ( see section [ s22 ] ) .
we consider a range of molecules and atoms ( h@xmath5o , co , co@xmath5 , ch@xmath6 , na and k ) that have signatures in the visible and ir .
for the sources of opacity , we used the line lists of hitemp for h@xmath5o , co , and co@xmath5@xcite , stds for ch@xmath6@xcite as explained in @xcite , and vald for na and k@xcite .
these opacity sources are the only fitting variables that are being retrieved at each grid point of the aerosol models . assuming that molecular and atomic abundances are well - mixed with altitude
, we define a single scale factor for these profiles as fitting variables in the retrieval . h@xmath5 and
he are considered to have a fixed abundance ratio , @xmath11 0.097 , i.e. , assuming that they have solar abundance , during retrieval . for every iteration of the retrieval
the summation of the volume mixing ratio of all species must be unity , i.e. , @xmath12=1 , to obtain physically meaningful solutions , where the condition of @xmath13 is always satisfied . and @xmath3 space telescopes . a data point at 589.5 nm
is not considered here due to its resolution , @xmath14=2.2 nm , which is narrower than the 5 nm spectral resolution of our spectral models .
the bottom level of the model atmosphere is shown in the dashed line and equivalent to the radius of the planet at 10 bar , with which we can find the best - fitting solution for the ir measurements .
see main text for more details .
the upper and lower limits of @xmath15 are displayed as dotted lines . at @xmath16 , the continuum due to collisional induced absorption and rayleigh scattering by h@xmath5 and he are shown with @xmath17 ( black ) , @xmath18k ( @xmath19 ) ( red ) and @xmath20k ( @xmath21 ) ( blue ) ( see the context for more detail ) . a rayleigh slope by h@xmath5 and
he may be able to fit the @xmath4/stis and acs data points if the planetary radius has a large uncertainty .
_ however , we do not consider this scenario in the present study .
_ ] since transmission spectra have a low sensitivity to the vertical temperature profile @xcite , we have insufficient constraints for an independent retrieval of the vertical structure of temperature . however , the atmospheric scale height , which mainly determines the depth of the transmission spectrum , is a function of temperature , implying the necessity of the exploration of the temperature degeneracy with aerosol parameters and chemical composition .
hence , we assume that the day / night terminators have the same p - t profile as the dayside atmosphere @xcite .
specifically , we test the robustness of our retrieved composition and aerosol properties by repeating the analysis with the temperatures shifted by @xmath22200 k either side of the best - fitting p - t profile of the dayside hemisphere ( @xmath23 ) ( figure [ f1 ] ) . the details are discussed in section [ s321 ] . in addition to temperature degeneracy , we also consider the planetary radius ( @xmath24 ) at the bottom level of the model atmosphere : @xmath25 .
this is an additional free parameter adding considerable uncertainty to atmospheric parameters during retrieval .
this is because a change in the radius of planet , which defines the transit depth of the bottom of the model atmosphere , results in a change in compositional amount of molecules and aerosols ( e.g. * ? ? ?
* ; * ? ? ?
in other words , retrievals with a smaller @xmath24 favor opacity increased due to enhanced molecular abundances or aerosol amount , whereas a larger @xmath24 requires decreased opacity due to reduced amount of composition to fit the transit measurements . for the dayside emission spectrum of hd 189733b ,
@xcite used @xmath26 ( = @xmath27 ) , based on that measured in the optical @xcite , as a model input , where @[email protected] . in the present analysis
, we perform an additional retrieval to determine @xmath15 using the fixed best - fit temperature profile and molecules from the dayside atmosphere .
we fitted the ir measurements including nicmos , irac and mips channels and obtained the best - fit planet - to - star ratios ( @xmath29 ) of 0.1506 for @xmath23 , which is smaller than the value in @xcite .
furthermore , we assigned the upper and lower limits to @xmath15 so that @xmath30 and @xmath31 , an uncertainty envelope that is seven times larger than suggested by@xcite .
we investigate the sensitivity of the retrieved composition and aerosol properties by varying the p - t profile and planetary radius within these conservative uncertainty ranges . for more details
, we will show our results on the degeneracy due to the planetary radius in section [ s321 ] .
n , contours for the uniform aerosol models with single particle sizes are shown , where @xmath23 and @xmath26 .
the best fit is marked with white cross corresponding spectrum is shown in black in the bottom panel .
the contour plot demonstrates a constrained upper boundary of particle size ( @xmath32 m ) despite of the degeneracy with the aerosol optical depth ( @xmath8 ) .
the bottom panel shows differences between the spectra , where each color corresponds to the points in the contour above .
discrepancies appear mostly longward of 8 @xmath0 m , where the strong extinction by silicate is shown . at @xmath338 @xmath0 m , silicate is responsible for the spectral morphology and degrades the fitting quality for a region where aerosol has small @xmath9 and thick @xmath8 in the top panel.,title="fig : " ] n , contours for the uniform aerosol models with single particle sizes are shown , where @xmath23 and @xmath26 .
the best fit is marked with white cross corresponding spectrum is shown in black in the bottom panel .
the contour plot demonstrates a constrained upper boundary of particle size ( @xmath32 m ) despite of the degeneracy with the aerosol optical depth ( @xmath8 ) .
the bottom panel shows differences between the spectra , where each color corresponds to the points in the contour above .
discrepancies appear mostly longward of 8 @xmath0 m , where the strong extinction by silicate is shown . at @xmath338 @xmath0 m
, silicate is responsible for the spectral morphology and degrades the fitting quality for a region where aerosol has small @xmath9 and thick @xmath8 in the top panel.,title="fig : " ] the nadir optical thickness due to the extinction by an aerosol layer is given by , @xmath34 where @xmath35 is the number density in @xmath36 , @xmath37 is the column height and @xmath9 is the radius of the spherical particle .
@xmath38 is the extinction efficiency of the grain , a sum of the absorption efficiency ( @xmath39 ) and the scattering efficiency ( @xmath40 ) .
since we consider only the first approximation of a single - scattering calculation , we assume that a photon that is either absorbed or scattered by a particle is not returned to the line of sight .
we consider that @xmath35 is vertically uniform in between the pre - defined pressures of 10@xmath10 bar and 10 bar , which corresponds to about 15 atmospheric scale heights for @xmath411500 k and a mean molecular weight of 2.3 . as in @xcite ,
these pressure levels are defined to be the top and bottom of the model atmosphere for the vertically uniform aerosol model ( uc ) and the pressure levels covers the range of aerosol heights ( e.g. @xmath4210@xmath43 bar , * ? ? ?
we choose not to vary these upper and lower boundary pressures in this study because the main objective of this study is to explore the general influence of aerosols on the retrieval of the temperature , planetary radius , and molecular degeneracies consistent with the spectrum , irrespective of aerosol layer geometry .
formally , the aerosol properties should be part of the retrieval .
however , as a first approach , we fix @xmath8 and @xmath9 and fit the spectrum by varying the gaseous abundances for each retrieval and perform a suite of retrievals over a broad range of @xmath4410 @xmath0 m and @xmath4510 .
this approach allows us to explore the parameter space associated with the aerosol properties and to understand the inherent degeneracies . in the next section
, we will derive the best - fit scenario with an aerosol layer consisting of mgsio@xmath1 and explore how the aerosol properties are correlated with temperature , planetary radius , other candidate materials for the aerosols , and gaseous composition .
we perform the retrieval with the full transmission spectrum of hd 189733b between 0.324 @xmath0 m , using the ranges of @xmath8 and @xmath9 described in section [ s22 ] . in this first study ,
the aerosol is modelled as enstatite ( mgsio@xmath1 ) grains @xcite , which is a potential candidate for aerosols in sub - stellar atmospheres ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the temperature and radius ratio are fixed to @xmath17 and @xmath15 for this case .
the fitted spectra are assessed using the goodness - of - fit , i.e. weighted least mean square error , @xmath46n , where n=27 is the number of measurements .
the fitting parameters include h@xmath5o , carbon chemicals ( co , co@xmath5 and ch@xmath6 ) , and alkalis ( na and k )
. since opacities for carbon - based molecules at high temperatures are either unavailable or weak for wavelengths shorter than 1 @xmath0 m , h@xmath5o , na , k and extinction by mgsio@xmath1 are predominant in the @xmath4/stis and acs bandpasses .
figure 2 illustrates the contours of @xmath47/n for a range of @xmath8 and @xmath9 . for clarity , we only illustrate reduced ranges of aerosol parameter space throughout the study due to poor fits for all @xmath8 and @xmath9 larger than 0.1 and 1 @xmath0 m , respectively .
the aerosol models having a particle radius smaller than @xmath480.1 @xmath0 m are able to fit the data for @xmath49n ( @xmath50n@xmath51n ) @xmath72 , provided that @xmath8 lies between 0.0020.02 .
the retrievals effectively rule out any scenario involving particles larger than @xmath480.1 @xmath0 m in radius , with @xmath8 outside of the 0.0020.02 range . in the bottom panel in figure [ f2 ]
, it is found that different combinations of @xmath8 and @xmath9 are indistinguishable up to wavelengths of 8 @xmath0 m for all fitted spectra , the overall shape of which shows a decreasing absorption slope due to an extinction by aerosols combined with a feature of na , k and h@xmath5o at wavelengths shorter than 1 @xmath0 m and molecular absorptions by h@xmath5o , co , co@xmath5 and ch@xmath6 at wavelengths between 18 @xmath0 m . differences between the radius and optical depth combinations can only be seen in spectra at longer wavelengths , where a strong absorption by mgsio@xmath1 is expected .
this is the reason for an anti - correlation between @xmath8 and @xmath9 as shown in the contour plot , i.e. , the larger the size of the particles , the smaller the opacity required to fit the spectrum beyond 8 @xmath0 m . in figure [ f3 ] , we find that , for a fixed @xmath38 ( or @xmath8 ) at 1 @xmath0 m , an aerosol layer with @xmath52 0.1 @xmath0 m shows a smaller @xmath8 at @xmath338 @xmath0 m than that with small @xmath530.1 @xmath0 m due to the intrinsic characteristic of @xmath38 for mgsio@xmath1 . as a result , the measurements of @xmath3/mips and irac at 8 @xmath0 m place strong constraints that move @xmath8 to be decreased and @xmath9 to be increased by an order of magnitude in the contour plot .
this shows that the spectrum at wavelengths @xmath338 @xmath0 m contains a large amount of information about the aerosols as seen at wavelengths @xmath71 @xmath0 m .
this is , again , due to the wavelength - dependent nature of the @xmath38 of mgsio@xmath1 .
the dependence of @xmath38 on the fitting quality and aerosol properties will be discussed in section [ s322 ] . as a function of wavelength .
an increase in @xmath38 at 1 @xmath0 m due to an increase in @xmath9 is more sensitive than an increase at 10 @xmath0 m . given the same @xmath8 at 1 @xmath0 m , the smaller @xmath9 is , the greater @xmath8 is expected at wavelengths longer than 8 @xmath0 m , where an upturn in @xmath38 is shown . ] an aerosol layer composed of small particles ( @xmath54 1 ) produces a negligible opacity between 18 @xmath0 m , where molecules , primarily h@xmath5o , are able to account for the @xmath4/nicmos and @xmath3/irac spectrum .
this implies that any conclusions for aerosols made in this study are unaffected by the diverse interpretations of the @xmath4/nicmos data by different studies @xcite .
it is also possible that the large errors on measurements in the 18 @xmath0 m region permit solutions where the flat spectrum could be produced by a thick aerosol layer with large particles ( @xmath551 ) , for which @xmath38 is nearly flat across wavelengths . in this sense
, @xcite suggested that the whole transmission spectrum could be reproduced if small grains at high altitude are responsible for the spectrum at wavelengths @xmath71 @xmath0 m , and deeper , larger grains ( @xmath52 1@xmath0 m ) underlying the lower aerosol height forming the flat spectrum across the ir .
although more complex solutions with multiple particle sizes may be considered , such complexity is not required to fit the data . for the best - fit scenario for the full spectral range ,
we infer @xmath56 0.1 @xmath0 m and @xmath57 0.004 .
molecular abundances in the terminator regions are comparable with theoretical predictions of chemistry
the h@xmath5o and co abundances ( @xmath5810@xmath59 ) are consistent across the globe ( @xmath60 for both hemispheres ) , and a high co@xmath5 abundance and a low ch@xmath6 abundance ( @xmath6110@xmath62 ) are found compared to a disequilibrium chemistry model ( @xmath63 and @xmath64 ) ( e.g. * ? ? ?
* ) , resulting in a low c / o ratio compared to the dayside atmosphere , which is close to the solar value of about 0.50.6 @xcite . however , the best - fitting solutions for molecules listed here have a broad uncertainty due to degeneracies with temperature , planetary radius , and aerosol composition .
therefore , uncertainties in these best - fitting values will be assessed in subsequent sections . during retrieval , temperature and planetary radius change opacity , which defines the absorption feature at each wavelength .
given the fundamental degeneracy between how these parameters affect the spectrum , however , the retrieval of these parameters is not independently constrainable @xcite . as a result , a change in opacity due to temperature and planetary radius alters the aerosol properties significantly . in this section , we will therefore investigate how the effects of these degeneracy sources change the aerosol solution .
firstly , to test the sensitivity of the aerosol properties to temperature , we repeat the aerosol retrieval analysis for colder and warmer thermal structures , i.e. , @xmath65 and @xmath66 , for which the temperature of the entire profile is shifted by 200k from the dayside temperature profile ( @xmath17 ) as introduced in section [ s22 ] . the planetary radius is fixed to @xmath26 .
setting up the same retrieval parameters listed in section [ s22 ] , suites of @xmath8 and @xmath9 with @xmath65 and @xmath66 are explored . in figure
[ f4 ] , it is shown that the colder temperature profile increases @xmath8 compared to the @xmath17 case whereas the warmer temperature shows a thinner @xmath8 than the @xmath17 case .
this is because warmer temperature ( 1 ) increases the atmospheric scale height and ( 2 ) enhances the continuum absorption from the h@xmath5he cia and rayleigh scattering across wavelengths ( see figure [ f1 ] ) .
an enhanced continuum absorption by @xmath66 , in turn , results in reduced molecular abundances ( particularly h@xmath5o ) compared to the @xmath23 case that is necessary to fit the @xmath4/nicmos and @xmath3/irac fluxes .
however , this yields a poor fit to the @xmath4/acs channels in the 0.81.1 @xmath0 m ( spectrum in red color in figure [ f6 ] ) , resulting in the best fitting quality for @xmath67 worsened over @xmath23 . for @xmath68 ,
the continuum absorption by h@xmath5 and he is reduced and the molecular abundances are enhanced to fit the measurements in the ir . since @xmath68 causes the scale height of atmosphere to be reduced , which reduces the scale of the spectral features , the best - fit spectrum for @xmath68 ( spectrum in orange color in figure [ f6 ] ) at @xmath71 @xmath0 m reproduces the @xmath4/acs measurements ( 0.61.0 @xmath0 m ) with a spectrum that starts to rise at 0.9 @xmath0 m towards shorter wavelengths , which gives an improved fitting quality .
regardless of temperature degeneracy , we find that the upper limit of @xmath9 is consistent , irrespective of the chosen temperature profile .
this is because the particle size determines the shape of @xmath38 across all wavelengths , which is responsible for the slope at @xmath71 @xmath0 m and an absorption at @xmath338 @xmath0 m ( see figure [ f3 ] ) .
independence of the maximum particle size with respect to temperature tells us that the degeneracy between the two parameters can be broken from the current measurements .
secondly , we fix the temperature profile to be @xmath23 and repeat the grid of retrievals for different @xmath8 and @xmath9 , this time shifting the planetary radius by 0.5 per cent from @xmath15 .
as we discussed in the analysis of temperature sensitivity , it is found that the upper limit of @xmath9 is still well - constrained even though the radius of the planet varies ( figure [ f5 ] ) .
again , the degeneracy between the largest particle size allowed and planetary radius is broken . for a smaller @xmath69
, the optical thicknesses can be larger than the @xmath26 case that is required to fit the slope in the uv and visible .
it is found that the well - fitted solutions also appear at large @xmath8 ( @xmath480.04 ) and small @xmath9 ( @xmath70.01 @xmath0 m ) , the region where a high extinction by aerosols is shown beyond 8 @xmath0 m ( c.f .
spectrum in cyan color in figure [ f2 ] ) .
we see that an increase in the radius by 0.5 per cent from @xmath15 mimics the effect of an increase in the temperature by 200 k because the main effect from the temperature is an increase in the scale height .
a large @xmath24 favors a reduced aerosol @xmath8 at @xmath71 @xmath0 m to find a solution for the @xmath4/stis and acs data .
in the ir , the continuum absorption by cia reaches as high as the radius ratio at the bottom of the measured spectra and therefore a small amount of h@xmath5o is required to fit the @xmath4/nicmos spectrum .
this leads to the shallower best - fit spectrum at wavelengths @xmath331 @xmath0 m than the best - fit spectrum for the @xmath67 , @xmath26 case ( see spectra in red and cyan color in figure [ f6 ] ) , even though the @xmath49n contours are nearly identical .
we conclude that constraining the molecular abundance in the terminators with a narrow range of uncertainty is not achievable from the current measurements because the extinction by aerosols is highly degenerate with temperature and planetary radius .
we note here that if the scale height of the model atmosphere is reduced due to a high mean molecular weight found from the retrieval , the model top - of - atmosphere pressure required to reproduce the @xmath4/stis spectrum would be lower than 10@xmath10 bar , as the physical extent of the atmosphere is reduced by increasing the mean molecular weight , leading to a poor fit to the entire spectrum .
we tested the effect on our results of reducing the model top - of - atmosphere pressure to 10@xmath70 bar ; we found that , in general , the required total cloud optical depth was reduced by a factor of two in this case , whereas the upper limit on the particle size is unchanged .
this implies that although the aerosol solutions are degenerate with the top pressure level , reproducing the slope at wavelengths @xmath71 @xmath0 m is much more weighted by the particle size as it primarily governs the shape of the spectrum in this wavelength region . , mg@xmath5sio@xmath6 @xcite , astronomical silicate @xcite , na@xmath5s @xcite and tholin @xcite .
the bottom panel shows the extinction efficiency ( @xmath38 ) of the five materials with @xmath560.01 @xmath0 m , all normalized to the @xmath38 at 1 @xmath0 m .
tholin has a steep slope at wavelength @xmath71 @xmath0 m , resulting in a family of solutions with a thin @xmath8 in the contour plot in figure [ f8 ] that is different from the other aerosol candidates .
, title="fig : " ] , mg@xmath5sio@xmath6 @xcite , astronomical silicate @xcite , na@xmath5s @xcite and
tholin @xcite .
the bottom panel shows the extinction efficiency ( @xmath38 ) of the five materials with @xmath560.01 @xmath0 m , all normalized to the @xmath38 at 1 @xmath0 m .
tholin has a steep slope at wavelength @xmath71 @xmath0 m , resulting in a family of solutions with a thin @xmath8 in the contour plot in figure [ f8 ] that is different from the other aerosol candidates .
, title="fig : " ] , mg@xmath5sio@xmath6 @xcite , astronomical silicate @xcite , na@xmath5s @xcite and tholin @xcite .
the bottom panel shows the extinction efficiency ( @xmath38 ) of the five materials with @xmath560.01 @xmath0 m , all normalized to the @xmath38 at 1 @xmath0 m .
tholin has a steep slope at wavelength @xmath71 @xmath0 m , resulting in a family of solutions with a thin @xmath8 in the contour plot in figure [ f8 ] that is different from the other aerosol candidates .
, title="fig : " ] in addition to uncertainties associated with the thermal structure and planetary radius , we also consider the differences in aerosol extinction cross section ( calculated assuming mie theory for spherical particles ) for different aerosol composition dominating the transmission spectrum of hd 189733b .
this constitutes one of the largest uncertainties when modelling the aerosol opacity for a hot jupiter atmosphere . we examine how the choice of aerosol materials affect the best - fitting aerosol parameters .
each candidate material has its own complex refractive index spectrum , which determines a wavelength - dependent extinction efficiency ( @xmath38 ) , yielding different @xmath47/n contours in the @xmath8 and @xmath9 space .
previously , given warm atmospheres under diverse conditions , a range of different candidates have been suggested in the literature .
they discussed thermal conditions of condensation of each candidate to constrain the lower deck pressure where a saturation process is triggered ( e.g. * ? ? ?
* ; * ? ? ?
the aerosol formation mechanism in hot jupiter atmospheres , however , may differ from that in other warm atmospheres ( e.g. low - mass brown dwarfs ) due to a strong irradiation from the parent star .
we choose aerosols that may be plausible in a warm atmosphere and investigate the effect of @xmath38 on our aerosol solution . by doing so
, we can find the specific shape of @xmath38 , which is required to fit the spectrum best in each wavelength .
we selected four materials ; forsterite ( mg@xmath5sio@xmath6 ) , astronomical silicate @xcite , sodium sulfide ( na@xmath5s ) , and tholin .
some of these species have a substantially different extinction efficiency compared to the mgsio@xmath1 grains used in our previous retrieval testing .
we perform the retrievals for a range of @xmath8 and @xmath9 with a well - mixed aerosol consisting of each material , for which @xmath38 is computed from the refractive index as shown in figure [ f9 ] .
the refractive indices are taken from @xcite for mg@xmath5sio@xmath6 , @xcite and @xcite for astronomical silicate , @xcite , @xcite and @xcite for na@xmath5s , and @xcite for tholin . to study the aerosol degeneracy due to @xmath38 alone , the vertical temperature profile and the radius of the planet at the bottom of the atmosphere
are fixed to be @xmath23 and @xmath71 .
the @xmath47/n contours for the cases with the four different aerosols are shown in figure [ f8 ] .
for mg@xmath5sio@xmath6 and astronomical silicate , where the overall shape of @xmath38 resembles that of mgsio@xmath1 ( see figure [ f9 ] ) , the best - fit @xmath8 and @xmath9 become thicker and smaller than the mgsio@xmath1 case , respectively , showing an improved fit to the spectrum at wavelengths @xmath71 @xmath0 m ( see spectra in orange and cyan color in figure [ f7 ] ) .
a small change in the spectrum at short wavelengths leads to a large improvement in goodness - of - fit even though a degraded fitting quality at the @xmath3/mips channel is found .
an increased @xmath8 induces a peak in the spectrum at 10 @xmath0 m due to aerosol extinction that produces no significant change in fitting quality because there are no constraints in this wavelength region .
the test implies that spectroscopic measurements at @xmath33 8@xmath0 m contain valuable information to understand the composition of aerosol in a hot jupiter atmosphere . on the other hand , we find that the extinction efficiency of na@xmath5s becomes flat at wavelengths @xmath331 @xmath0 m due to the flat real refractive index and the small imaginary refractive index , whereas a high imaginary index is apparent at wavelengths @xmath70.6 @xmath0 m .
this means that , except for the region at wavelengths @xmath70.6 @xmath0 m , the best - fit spectrum with na@xmath5s aerosols ( spectrum in red in figure [ f7 ] ) resembles that with the mgsio@xmath1 in all ir regions , where an extinction by mgsio@xmath1 is hardly shown . as a result , only the @xmath4/stis and acs measurements are useful to characterise the properties of the na@xmath5s aerosols , which adds more uncertainty to @xmath8 and @xmath9 ( see the bottom right panel in figure [ f8 ] ) .
this implies that constraining aerosols in an atmosphere is highly dependent on the extinction efficiency of materials and that the properties of aerosols can only really be constrained in wavelength regions where strong extinction features are evident .
a similar interpretation can be also made in the case of the tholin aerosol , which yields @xmath8 that is thinner than 0.01 in the best - fit case ( see the bottom - right panel in figure [ f8 ] ) .
the shape of the @xmath47/n contour that is different compared to other candidates also comes from the difference in @xmath38 . for a small particle ( @xmath530.1 @xmath0 m ) ,
the tholin s extinction coefficients at wavelengths shorter than 1 @xmath0 m decrease more rapidly with wavelength than the other materials as the bottom panel of figure [ f9 ] shows , which induces a high absorption at @xmath71@xmath0 m , producing good fits to the spectrum even for small @xmath8 ( @xmath70.001 ) at 1 @xmath0 m . for a large particle ( @xmath520.1 @xmath0 m ) , tholin becomes also optically active at wavelengths longward of 2 @xmath0 m , where the spectra fitted exhibit large absorption features that provide a large deviation from the ir measurements , if @xmath720.01 .
therefore , we find that the tholin @xmath8 must be very small to reproduce the spectrum , which is constrained by the measurements in between 0.324 @xmath0 m except @xmath4/nicmos .
overall , we found that all of the aerosol materials considered in the present study are capable of fitting the transmission spectrum of hd 189733b .
the properties of each aerosol constrained are highly diverse , mainly due to the unique characteristics of @xmath38 and the divergence of the aerosol properties comes from the uncertainty in condensation state in the terminators .
we found that the retrieval of @xmath8 depends on the choice of material , which adds more uncertainty to the retrieval value . despite the large uncertainty of @xmath8 retrieved , the conclusion on the upper limit of @xmath9 ( @xmath480.1 @xmath0 m ) is still robust , regardless of the kind of aerosol material .
this is because of the appearance of a peak ( the first harmonic ) of @xmath38 at short wavelengths if @xmath73 0.1 @xmath0 m , which results in poor fits at the @xmath4/stis bandpasses .
temperature and other atmospheric parameters such as planetary radius , aerosol properties are highly correlated with gaseous molecules , which are the most significant degeneracy sources in the exoplanet atmospheres .
previously , inverse model studies have shown that the atmospheric molecular abundance in the transiting exoplanets can not be constrained uniquely from existing transit spectra and therefore the degeneracy between molecules and aerosols , temperature , planetary radius introduces an uncertainty to the atmospheric retrieval method @xcite .
@xcite and @xcite demonstrated that the parameters defining the aerosol deck pressures and its vertical structure of compositional amount modify the molecular abundances , and the uncertain composition of the aerosols increase the uncertainty on the gaseous composition in the retrievals .
furthermore , the choice of value of the planetary radius adds more uncertainty to the determination of molecular abundances .
the vertical sensitivity of the spectrum to variations in molecular composition , temperature and aerosol opacity is shown in figure [ f10 ] . ] .
it is found that the effect of na and k are predominant in the uv and visible range , but also influences the near ir and they play a crucial role in fitting the data points at @xmath71 @xmath0 m . on the other hand ,
our mgsio@xmath1 aerosol model provides strong sensitivities at short ( @xmath71 @xmath0 m ) and long ( @xmath338 @xmath0 m ) wavelengths due to a low extinction of rayleigh slope in the middle range , which is broadly relevant for all other materials tested in the present study .
this tells us that the aerosol properties are weakly correlated with the molecules that are spectrally active between 18 @xmath0 m . in table
[ t1 ] , we present the ranges of molecular abundances constrained with different aerosol materials .
we find that the range of the abundance of h@xmath5o for @xmath47/n@xmath72 are well constrained with a low uncertainty and this tendency is consistent for all aerosol materials , confirming that there are marginal degeneracies between h@xmath5o and aerosol .
however , this is only true where the aerosols have negligible influence in the 18 @xmath0 m range .
if the extinction spectrum of aerosols is rather different and there are absorption features in this wavelength range , there would be more uncertainty in the h@xmath5o abundance .
a possible interpretation of the narrow range of the h@xmath5o abundance is that there are only 8 data points in the ir available and they provide strong constraints that determine the molecular abundance . for all scenarios that we considered in table [ t1 ] ,
we note that the abundances of co , co@xmath5 , and ch@xmath6 have a small range compared to other gaseous molecules .
this is because , for any initial atmospheric conditions that are different from the best fit scenario , modifying the abundance of h@xmath5o resolves most discrepancies between the measured and fitted spectra in the ir , requiring nearly the same amount of carbon chemicals to fit the data whose opacity contribution to the total spectrum is negligible , i.e. , an insignificant contribution to @xmath74 .
we note that retrievals without carbon chemicals marginally change the @xmath74/n contours in the @xmath8 and @xmath9 space and the best - fit fitting quality , e.g. , @xmath74/n=2.42 to 2.55 for the case of mgsio@xmath1 , @xmath23 , and @xmath26 .
therefore the existence of co , co@xmath5 , ch@xmath6 is highly uncertain . particularly , the spectral sensitivities to the carbon - bearing molecules are correlated with the signatures of h@xmath5o , such that the transmission spectrum is unable to place strong constraints on the presence of co , co@xmath5 and ch@xmath6 due to the degeneracy between them . even when the temperature profile is fixed to
that derived from the dayside emission spectrum , some of these degeneracies still remain .
o , co@xmath75 , co , ch@xmath76 , na and k , plotted against pressure and wavelength .
these calculations use the result of the spectrum with the best - fit case in figure [ f2 ] .
the contours are normalised to the maximum values in each panel . ]
conversely , the range of alkali abundances from retrievals as presented in table [ t1 ] appears rather large ( @xmath77 , @xmath78 for @xmath49n@xmath72 ) .
according to figure [ f10 ] , the sensitivities of the spectrum to na and k in the visible and nir overlap with that of the aerosols .
their abundances are sensitive to both optical depth and particle size of aerosol , resulting in broad uncertainties in their amounts .
this clearly shows a significant degeneracy between alkalis and aerosol the extinction efficiency of aerosols for any particle sizes smaller than @xmath480.1 @xmath0 m produces a rayleigh slope at @xmath71.2 @xmath0 m , where the total optical depth of the atmosphere is the sum of an aerosol @xmath8 ( @xmath79 due to rayleigh scattering ) and optical depth by alkalis .
therefore , a small change in aerosol properties ( @xmath8 and @xmath9 ) leads to a larger change in @xmath80 to fit the slope .
this was suggested theoretically for cloudy exoplanets ( e.g. * ? ? ?
* ) and is shown in the analysis of the reflection spectrum of this planet @xcite . despite this degeneracy , an additional retrieval test where na and k were excluded produces a radius ratio slope at @xmath71.2 @xmath0 m with @xmath49n@xmath335 , compared to the best fit case in section [ s31 ] , for all combinations of aerosol @xmath8 and @xmath9 and h@xmath5o ( not shown ) .
this test adds more credibility to the presence of na and k to reproduce the transmission spectrum @xcite .
figure [ f10 ] informs us that the spectrum is insensitive to temperature shortward of 1 @xmath0 m , where a high sensitivity to the presence of aerosols is shown due to the predominance of rayleigh scattering .
longward of 1 @xmath0 m , the spectrum is sensitive to both temperature and h@xmath5o abundance , confirming that an independent retrieval of temperature is not achievable from the primary transits @xcite .
the molecular degeneracy with temperature therefore increases the uncertainty on the chemistry in the atmosphere because temperature is one of the primary driving forces that determine the opacity of the molecules and therefore the shape of the transmission spectrum .
given @xmath71 , the derived abundance of h@xmath5o with @xmath23 for @xmath49n@xmath72 , @xmath8110@xmath59 is an order of magnitude smaller than @xmath68 , @xmath8210@xmath59 and larger than @xmath67 , @xmath8310@xmath59 .
similarly , given @xmath23 , the range of the h@xmath5o abundance with @xmath69 , @xmath84 are two and four orders of magnitude enhanced compared to the cases with @xmath26 , @xmath8510@xmath59 and @xmath86 , @xmath8710@xmath59 , thus limiting our ability to determine the abundance of h@xmath5o in the terminators within a narrow uncertainty range ( see table [ t1 ] ) .
it is expected that this uncertainty will become broader when we add the ambiguity of the aerosol material to the error budget . in summary
, the spectrum shows a low sensitivity to gaseous constituents ( except h@xmath5o ) and characterising their abundances is challenging because of ( i ) a high correlation with aerosol ( na and k ) or ( ii ) the uncertainty in their presence in the atmosphere ( carbon molecules ) .
therefore we conclude that we are unable to robustly determine abundances for gaseous constituents from the spectroscopic measurements used in this study , due to a considerable degeneracy between multiple molecules contained in the model atmosphere and other parameters that determine the shape of spectrum . [
cols="^,^,^,^",options="header " , ]
we have performed an optimal estimation retrieval to reproduce the transmission measurements for hd 189733b , utilising the broad spectral coverage between 0.324 @xmath0 m .
our retrieval analysis with a simple 2-parameter aerosol model confirms that we can characterise the terminator regions of hd 189733b using vertically uniform aerosol models that have an optical depth at 1 @xmath0 m of between [email protected] and particle radii @xmath420.1 @xmath0 m .
exploring a suite of aerosol models for a wide wavelength region , we demonstrated that there are considerable correlations between aerosol parameters and temperature , planetary radius , aerosol composition and even gaseous alkali abundances that are not easily broken if the temperature or planetary radius increases , the opacity required from the aerosols to fit the transit spectrum decreases . at the same time , na and k compete to fit the data in the visible and nir .
moreover , we showed that different materials for the aerosol affect the best - fit @xmath8 and @xmath9 because they have different extinction efficiencies ( @xmath38 ) . due to the wide variety of processes that could cause aerosols in a hot jupiter atmosphere ,
the choice of candidates increases the degeneracy that leads to a high uncertainty in the retrieval of alkalis and aerosol properties .
however , the conclusion on the upper boundary of particle size ( @xmath530.1 @xmath0 m ) is robust , regardless of any degeneracies inherent in the spectrum . combining the constraints from wavelengths at @xmath71 @xmath0 m and
@xmath338 @xmath0 m does allow us to draw conclusions on the aerosol solutions for required optical depth and particle size by fitting these transmission spectra .
furthermore , our aerosol models are consistent with a companion retrieval analysis of the dayside spectrum of hd 189733b ( see * ? ? ?
we found that there is a weak correlation between the molecular abundances ( h@xmath5o , co , co@xmath5 , and ch@xmath6 ) and the aerosol properties .
we confirmed that a strong correlation between h@xmath5o and temperature and planetary radius exists , causing a broad uncertainty in the h@xmath5o abundance .
the existence of carbon - bearing molecules is rather uncertain , because a lack of information is suggested from the measurements , where there exists a strong degeneracy driven by a small number of constraints , i.e. , the under - constrained problem .
we emphasize that the strongest constraints on aerosol properties resulted from measurements at the shortest ( @xmath71 @xmath0 m from @xmath4/stis and acs ) and longest ( @xmath338 @xmath0 m from @xmath3/irac and mips ) wavelengths , with a broad trough in the middle due to molecular absorptions .
the long - wavelength region is often overlooked in the literature , but from the refractive index of materials , the aerosol extinction at these wavelengths can be used to break the degeneracy at shorter wavelengths .
a highlight of our study is combining the short and long wavelengths to break degeneracies associated with the aerosol properties .
we suggest that measuring data in these complementary spectral windows will render interpretations of ir spectra more robust .
any future space - based characterization mission should consider this .
+ _ jl and kh acknowledge the support from the swiss - based merac foundation , the university of bern and the university of zrich .
pgji acknowledges the support of the united kingdom science and technology facilities council .
lnf is supported by a royal society research fellowship .
jkb is supported by the john fell fund by the university oxford press .
the calculations were performed using the @xmath88 computing cluster at the university of zrich thanks to support from doug potter , simon grimm and joachim stadel .
we are grateful to david sing , frederic pont , suzanne aigrain and neale gibson for very helpful discussions . _ | a number of observations have shown that rayleigh scattering by aerosols dominates the transmission spectrum of hd 189733b at wavelengths shortward of 1 @xmath0 m . in this study , we retrieve a range of aerosol distributions consistent with transmission spectroscopy between 0.324 @xmath0 m that were recently re - analyzed by @xcite . to constrain the particle size and the optical depth of the aerosol layer , we investigate the degeneracies between aerosol composition , temperature , planetary radius , and molecular abundances that prevent unique solutions for transit spectroscopy . assuming that the aerosol is composed of mgsio@xmath1 , we suggest that a vertically uniform aerosol layer over all pressures with a monodisperse particle size smaller than about 0.1 @xmath0 m and an optical depth in the range 0.0020.02 at 1 @xmath0 m provides statistically meaningful solutions for the day / night terminator regions of hd 189733b . generally , we find that a uniform aerosol layer provide adequate fits to the data if the optical depth is less than 0.1 and the particle size is smaller than 0.1 @xmath0 m , irrespective of the atmospheric temperature , planetary radius , aerosol composition , and gaseous molecules .
strong constraints on the aerosol properties are provided by spectra at wavelengths shortward of 1 @xmath0 m as well as longward of 8 @xmath0 m , if the aerosol material has absorption features in this region .
we show that these are the optimal wavelengths for quantifying the effects of aerosols , which may guide the design of future space observations .
the present investigation indicates that the current data offer sufficient information to constrain some of the aerosol properties of hd189733b , but the chemistry in the terminator regions remains uncertain . |
the polynomials @xmath1 , and @xmath2 defined on the unit interval @xmath3 are called monic orthogonal polynomials with respect to a weight @xmath4 if @xmath5 where @xmath6 denotes the kronecker @xmath7 sequence . when renormalized , these monic orthogonal polynomials give rise to an orthonormal system of polynomials @xmath8 , where @xmath9 the monic op are constructed by performing the gram - schmidt process on the monomials @xmath10 , with respect to the inner product @xmath11 .
examples of such polynomials include the legendre polynomials defined on @xmath12 $ ] , and using the weight function @xmath13 , numerical information about the op can be obtained via a fundamental recursive equation called the three - term recursion formula .
for example , the values of the polynomials at different points in @xmath3 , and their zeros can be obtained from the three - term recurrence relation . for comparison with our results ,
we state the general form of this recurrence relation and refer to @xcite for details on op . * theorem a. ( * ? ? ?
* theorem 1.27 ) * _ the orthogonal polynomials @xmath14 satisfy a three - term recurrence relation . more specifically , @xmath15 and for all
@xmath16 we have : @xmath17 where @xmath18 _ our goal in this paper is to construct and investigate the properties of the analogs of the above op on @xmath0 . but first what is a polynomial on a fractal set such as @xmath0 ?
let @xmath19 denote the laplacian on @xmath20.$ ] if we let @xmath21 , for @xmath22 $ ] , it is immediate that @xmath23 is a solution to a differential equation @xmath24 for some @xmath25 . this observation can be used to define polynomial on fractals in general and on the sierpinski gasket in particular . in particular , on @xmath0 , one defines a _ polynomial _
@xmath26 to be any solution of @xmath27 for some @xmath28 , where @xmath29 is the ( fractal ) laplacian to be defined below . using this analogy
, monomials were introduced on the sierpinski gasket @xcite based on the fractal laplacian constructed by kigami @xcite , see also @xcite .
we also note that a theory of polynomials based on a different laplacian , the so - called energy laplacian , is developed in @xcite . consequently , our construction of op on @xmath0 is based on these polynomials .
in particular , we apply the gram - schmidt orthogonalization algorithm to produce the analogs to the legendre polynomials on @xmath0 . in the process
we construct both a family of symmetric and antisymmetric op on @xmath0 , and the concatenation of the two families yields a tight frame for a proper closed subspace of @xmath30 .
the reason that our construction does not yield a dense set of op in @xmath30 is due to the fact that the polynomials are not dense in this space ( * ? ? ?
* theorem 4.3.6 ) , ( * ? ? ?
* section 5.1 ) .
in addition , we investigate thoroughly a corresponding three - term recursion formula , paying a particular attention to the asymptotics of the coefficients involved in this recursion .
our three - term recurrence is fundamentally different from , and this is essentially due to the fact that the product of two polynomials on @xmath0 is not a polynomial @xcite ! in addition , our analog of is very unstable and thus can not be used to generate the values of the op on @xmath0 .
however , we are able to exploit this three - term recurrence to recursively express the corresponding op in terms of a well - known basis of polynomials on @xmath0 .
this leads to the development of some numerical tools that allow us to plot graphs of many of these op . in addition
, we present preliminary results pertaining the zeros of these op .
in particular , we graph the zero sets of certain of these op . these results are mainly experimental but seems to indicate that these zero sets have some structure .
we have not been able to prove anything along these lines but hope that our experimental results lead to more investigations on the zero sets of the op .
nonetheless , we have generated many graphs related to the op on @xmath0 that we did not include in the present paper due to space constraint .
the interested reader is referred to the following web site which contains many figures and algorithms that were generated in the course of this research project @xcite .
our paper is organized as follows : in section [ analysis ] we review the basic facts of analysis on fractals needed to state and prove our results .
we also prove some new results about the green s function that may be of independent interest .
section [ sec2 ] is devoted to the construction of the op on @xmath0 and to the investigation of the three - term recursion and the asymptotic analysis of its coefficients .
we also consider the jacobi matrix associated with these coefficients . in section [ sec3 ]
we carry out the numerical computations that enable us to plot many of the op we constructed .
in addition , we give some numerical evidence concerning the asymptotics of the coefficients involved in the three - term relation . finally , we give numerical description of the zero sets of the op and display some of these zero sets as well as some of their corresponding nodal sets on @xmath0 .
for more information on the theory of calculus on fractals see @xcite ; more specifically for calculus on the sierpinski gasket see @xcite . however , we collect below some key facts needed in the formulation of our results .
the sierpinski gasket ( @xmath0 ) , shown in figure [ fig : sg ] , is the attractor of iterated function system ( ifs ) consisting of three contractions in the plane @xmath31 , @xmath32 , defined by @xmath33 where @xmath34 are the vertices of an equilateral triangle .
@xmath0 satisfies the self - similar identity @xmath35 , and the @xmath36 are called _ cells of level 1_. by iteration we can write @xmath37 for @xmath38 , each @xmath39 .
we call each @xmath40 a word of length @xmath41 , and we have the self - similar identity @xmath42 . each @xmath43 is called a _ cell of level @xmath44_. @xmath0 can also be viewed as the limit of a sequence of graphs @xmath45 ( with vertices @xmath46 and edge relation @xmath47 ) defined inductively as follows : @xmath48 is the complete graph on @xmath49 , and @xmath50 with @xmath47 if and only if x and y belong to the same cell of level @xmath44 .
then @xmath51 , the set of all _ vertices _ is the analog of the dyadic points in @xmath52 $ ] and is dense in @xmath0 .
we consider @xmath53 the set of boundary points of @xmath0 , and @xmath54 is the set of _ junction points_. note that every junction point in @xmath46 has four neighbors in the graph @xmath45 .
the graph laplacian @xmath55 is defined by @xmath56 and the laplacian @xmath29 on @xmath0 is defined as the renormalized limit @xmath57 at the three boundary points of @xmath0 , we have two different derivatives , the normal and tangential derivative , respectively defined by @xmath58 identifying the boundary points @xmath59 and @xmath60 when @xmath61 .
using the above definitions , the analog of the gauss - green formula takes the following form : @xmath62 where @xmath63 is the natural probability measure which assigns weight @xmath64 to each cell of order @xmath44 .
these analytical tools can now be used to solve the boundary value problem @xmath65 whose solution is given by @xmath66 @xmath67 is called the green s function , and is given by @xmath68 where @xmath69 is zero when @xmath70 and @xmath71 are not in the same cell of level @xmath72 , @xmath73 @xmath74 and @xmath75 is the piecewise harmonic function satisfying @xmath76 and vanishing outside the @xmath77th cell to which @xmath70 belongs .
we shall need an exact estimate of @xmath78 , the @xmath79 norm of the green function @xmath67 .
it was conjectured in @xcite and proved in @xcite that @xmath80 this immediately implies that @xmath81 .
however , we obtain exact values for @xmath78 and a related quantity in the next result . though the result seems simple , to our knowledge , it has not appeared anywhere in the literature .
first , we briefly recall a description of the spectrum of @xmath29 on @xmath0 , that was given in @xcite using the method of spectral decimation introduced in @xcite .
in essence , the spectral decimation method completely determines the eigenvalues and the eigenfunctions of @xmath29 on @xmath0 from the eigenvalues and eigenfunctions of the graph laplacians @xmath55 .
more specifically , for every dirichlet @xmath82 of @xmath83 on @xmath0 , that is a solution of @xmath84 there exists an integer @xmath85 , called the _ generation of birth _ , such that if @xmath86 is a @xmath87-eigenfunction and @xmath88 then @xmath89 is an eigenfunction of @xmath90 with eigenvalue @xmath91 .
the only possible initial values @xmath92 are @xmath93 , @xmath94 and @xmath95 , and subsequent values can be obtained from @xmath96 where @xmath97 can take the values @xmath98 .
the sequence @xmath91 is related to @xmath82 by @xmath99 [ nmgreen ] let @xmath100 be the eigenvalues of the laplacian on @xmath0 .
then , @xmath101 and @xmath102 it is easily seen that @xmath67 can be expressed as @xmath103 where @xmath104 are the laplacian eigenvalues corresponding to the eigenfunction @xmath105 @xcite .
it then follows that @xmath106 where @xmath107 is a finite set that depends on the multiplicity of the graph - laplacian eigenvalues @xmath92 .
we shall now find explicit formula for the above sum .
suppose that we know the graph - laplacian eigenvalues @xmath108 whose multiplicity are @xmath109 : if @xmath110 it has multiplicity @xmath111 , if @xmath112 it has multiplicity @xmath113 and if @xmath114 it has multiplicity @xmath115 .
then , at the next level ( @xmath72 ) , the graph - laplacian eigenvalues @xmath116 come from the spectral decimation method .
moreover , for every @xmath117 with multiplicity @xmath118 there will be two eigenvalues @xmath119 with multiplicity @xmath118 and given by @xmath120 note that @xmath121 now @xmath122 where @xmath123 , since the eigenvalue @xmath110 with multiplicity @xmath111 contributes @xmath124 to the sum .
therefore , by the spectral decimation , we have @xmath125 with @xmath126 consequently , @xmath127 similarly , if we let @xmath128 , then @xmath129 one can check that @xmath130 next , we can use the spectral decimation method to prove that @xmath131 where @xmath132 moreover , @xmath133 with @xmath134 and @xmath135 consequently , @xmath136 for any integer @xmath28 , the set of polynomials of degree less than or equal to @xmath137 will be denoted @xmath138 and consists of the solutions of @xmath139 .
@xmath138 is a space of dimension @xmath140 and it has a basis @xmath141 characterized by @xmath142 where @xmath143 and @xmath144 in particular , @xmath145 is the space of _ harmonic functions _ , i.e. , a function @xmath146 belongs to @xmath147 if and only if it satisfies @xmath148 .
we refer to @xcite for details on polynomials on @xmath0 .
our construction of op on @xmath0 will be based on another basis for the space @xmath138 introduced in @xcite .
functions in this basis are called _ monomials _ and are essentially the fractal analogs of @xmath149 .
fix a boundary point @xmath150 for @xmath151 , the monomials @xmath152 for @xmath153 and @xmath154 are defined to be the functions in @xmath138 satisfying @xmath155 with @xmath156 when @xmath157 we will sometimes delete the upper index and just write @xmath158 for @xmath159 .
notice that for a fixed @xmath160 , the set of monomials @xmath161 forms a basis for @xmath162 .
moreover , the monomials in this basis satisfy @xmath163 , and possess some symmetry properties .
when @xmath164 or @xmath93 , @xmath152 is symmetric with respect to the line passing through @xmath150 and the midpoint of the side opposing @xmath150 .
in fact , these symmetric polynomials can be viewed as analogs of the even polynomials @xmath165 on the unit interval .
when @xmath166 , @xmath167 is antisymmetric with respect to the line passing through @xmath150 and the midpoint of the side opposing @xmath150 . in this case , @xmath167 can be viewed as analogs of the odd polynomials @xmath168 . in our construction of op on @xmath0
, we will need the following result which was proved in @xcite and which gives recursively values of the above monomials and their derivatives at boundary points .
* theorem b ( * ? ? ?
* theorem 2.3 ) * _ for @xmath154 , let @xmath169 _ the following recursion relations hold : @xmath170 \beta_j & \frac{2}{15(5^j -1 ) } \sum_{\ell=0}^{j-1 } ( 3\cdot 5^{j-\ell}-5^{\ell+1}+6)\alpha_{j-\ell}\beta_{\ell } & j & 1\\ [ .3ex ] \gamma_j & 3\alpha_{j+1 } & j&1\\ [ .3ex ] \eta_j & \frac{5^j + 1}{2 } \alpha_j + 2\sum_{\ell=0}^{j-1 } \eta_{\ell}\beta_{j-\ell } & j&1 , \end{array}\right.\ ] ] where the initial values are : @xmath171 , and @xmath172 .
moreover , @xmath173 we can use theorem b along with the green - gauss formula to compute the inner product among the monomials @xmath174 .
[ lemma1 ] consider the monomials @xmath175 for @xmath176 then the following formulas hold : @xmath177 { \langlep_{j2},p_{k2}\rangle } & -2 \sum_{\ell = j - m_*}^j \beta_{j-\ell}\alpha_{k+\ell+1 } - \beta_{k+\ell+1}\alpha_{j-\ell}\\ [ .2ex ] { \langlep_{j3},p_{k3}\rangle } & 18 \sum_{\ell = j - m_*}^j \alpha_{j-\ell+1}\eta_{k+\ell+2 } - \alpha_{k+\ell+2}\eta_{j-\ell+1}\\ [ .2ex ] { \langlep_{j1},p_{k2}\rangle } & -2 \sum_{\ell=0}^j \alpha'_{j-\ell}\alpha'_{k+\ell+1 } - \beta_{k+\ell+2}\eta_{j-\ell+1},\end{array}\right.\ ] ] where @xmath178 for @xmath179 and @xmath180 , and where @xmath181 .
moreover , @xmath182 in addition , for @xmath166 the inner products among the different monomials associated with each of the boundary points @xmath183 , @xmath184 are given by @xmath185 { \langlep_{j3}^{(n)},p_{k3}^{(n')}\rangle } & -\frac{1}{2}{\langlep_{j3}^{(0)},p_{k3}^{(0)}\rangle } \end{array}\right.\ ] ] for @xmath186 .
it is enough to prove that for each @xmath187 and any @xmath188 we have : @xmath189 notice that by a symmetry argument we can assume that @xmath190 and proved the result by induction on @xmath137 .
fix @xmath191 and notice that for all @xmath16 green - gauss formula gives : @xmath192 & = & \int_{sg } \delta p_{0i}p_{ki ' } \,d\mu+ \sum_{n=0}^2 p_{0i}(q_n)\partial_n p_{ki'}(q_n ) - p_{ki'}(q_n)\partial_n p_{0i}(q_n ) \\[.2ex ] & = & \sum_{n=0}^2 p_{0i}(q_n)\partial_n p_{(k+1)i'}(q_n ) - p_{(k+1)i'}(q_n)\partial_n p_{0i}(q_n ) , \end{aligned}\]]where we have used the fact that @xmath193 .
this establishes for @xmath194 , and all @xmath16 .
now assume that for some @xmath137 and for @xmath190 we have @xmath195 we just have to show that this last relation holds for @xmath196 and @xmath197 .
but , by the gauss - green formula we have @xmath198 & = & { \langlep_{ji},p_{(k+1)i'}\rangle } + \sum_{n=0}^2 \left ( p_{(j+1)i}(q_n)\partial_n p_{(k+1)i'}(q_n ) - p_{(k+1)i'}(q_n)\partial_n p_{(j+1)i}(q_n ) \right ) \\[.2ex ] & = & \sum_{\ell=0}^j \sum _ { n=0}^2 \left ( p_{(j-\ell)i}(q_n ) \partial_n p_{(k+2+\ell)i ' } ( q_n ) - p_{(k+2+\ell)i'}(q_n ) \partial_n p_{(j-\ell)i } ( q_n)\right)\\[.2ex ] & + & \sum_{n=0}^2 \left ( p_{(j+1)i}(q_n)\partial_n p_{(k+1)i'}(q_n ) - p_{(k+1)i'}(q_n)\partial_n p_{(j+1)i}(q_n ) \right ) \\[.2ex ] & = & \sum_{\ell=1}^{j+1 } \sum _ { n=0}^2 \left ( p_{(j+1-\ell)i}(q_n ) \partial_n p_{(k+1+\ell)i ' } ( q_n ) - p_{(k+1+\ell)i'}(q_n ) \partial_n p_{(j+1-\ell)i } ( q_n ) \right ) \\[.2ex ] & + & \sum_{n=0}^2 \left ( p_{(j+1)i}(q_n)\partial_n p_{(k+1)i'}(q_n ) - p_{(k+1)i'}(q_n)\partial_n p_{(j+1)i}(q_n ) \right ) \\[.2ex ] & = & \sum_{\ell=0}^{j+1 } \sum _ { n=0}^2 \left ( p_{(j+1-\ell)i}(q_n ) \partial_n p_{(k+1+\ell)i ' } ( q_n ) - p_{(k+1+\ell)i'}(q_n ) \partial_n p_{(j+1-\ell)i } ( q_n ) \right)\end{aligned}\ ] ] this shows that holds for all @xmath199 and @xmath190 .
next , follows from by symmetry consideration and by observing that the values and normal derivatives at @xmath200 vanish .
in addition , theorem b , especially can be used to further simplify and establish .
note that is trivial because of the symmetry of @xmath201 and @xmath202 and the anti - symmetry of @xmath203 . to prove ,
it is enough to show that @xmath204 since the other inner products can be computed using similar arguments . using equation , and canceling out zero terms yields @xmath205
we are now ready to construct the families of orthogonal polynomials on @xmath0 .
these op will be obtained by applying the gram - schmidt algorithm on the polynomials @xmath206 , where @xmath207 , and @xmath208 or @xmath93 . in subsections
[ subsec3 ] , and [ subsec4 ] , we plot the graphs of various op obtained from @xmath209 , and @xmath207 .
as mentioned in the introduction , more graphs of these op can be found at @xcite .
but we first derive a three - term recursion formula for our op and give estimates on the size of the coefficients appearing in this recursion formula .
these coefficients are extremely small .
this creates major instability problems when plotting the corresponding op . to circumvent this problem
, we carry out our numerical simulation to arbitrary precision .
a similar phenomenon was already observed in @xcite , and rational arithmetic was used .
we also considered the use of rational arithmetic , but because it involves a prohibitive time cost , we settled for arbitrary precision instead .
[ def2 ] fix @xmath210 or @xmath211 and denote by @xmath212 the orthogonal polynomials obtained from @xmath213 by the gram - schmidt process , i.e. , @xmath214 for each @xmath215 .
consequently , there exists a set of coefficients @xmath216 , with @xmath217 , and such that @xmath218 moreover , @xmath219 by normalizing the orthogonal polynomials @xmath220 we obtain the family of orthonormal polynomials @xmath221 characterized by @xmath222 [ nmpk ] given @xmath210 or @xmath211 , and for each @xmath28 the following holds : @xmath223 moreover , for any @xmath224 , there exist constants @xmath225 such that for all @xmath28 @xmath226 in particular , @xmath227 recall that by the definition of @xmath228 we can write that for @xmath215 , the gram - schmidt process yields @xmath214 for each @xmath215 .
therefore , @xmath229 which establishes the first estimate .
now recall from ( * ? ? ?
* theorem 2.7 ) and ( * ? ? ?
* theorem 2.13 ) , that for each @xmath230 , there are @xmath231 and @xmath232 such that @xmath233 when @xmath234 , we get from lemma [ lemma1 ] that @xmath235 from which follows . when @xmath236 or @xmath237 similar arguments are used to obtain the same estimate .
it then follows that @xmath238 with at least an exponential rate of decay .
our first main result is the following theorem that establishes the three - term recursion formula for op on @xmath0 .
[ main1 ] let @xmath239 be the orthogonal polynomials defined above .
let @xmath240 and for @xmath16 , let @xmath241 be the polynomial defined by @xmath242 set @xmath243 and @xmath244 . then , for each @xmath16 @xmath245 where @xmath246 consequently , @xmath247 using the definition of @xmath241 we can write @xmath248 notice that @xmath249 is a polynomial of degree @xmath196 .
thus , @xmath249 is orthogonal to all @xmath250 with @xmath251 .
therefore , @xmath252 for some coefficients @xmath253 and @xmath254 .
it is easy to see that @xmath255 this follows from the fact @xmath256 is orthogonal to @xmath257 .
thus , @xmath258 .
taking inner product with @xmath250 yields the first equation in .
taking inner product with @xmath259 yields @xmath260 which is exactly the second equation in .
observe that the last equation in is equivalent to @xmath261 , which implies .
the above results deserve some discussions .
while resembles , these two relations are fundamentally different . in fact , is essentially the statement that @xmath262 is a monomial of degree @xmath263 , and thus can be expressed in terms of @xmath264 , @xmath265 . because the product of polynomials is not a polynomial on @xmath0 , @xmath266 is replaced by @xmath267 .
therefore , unlike in the classical case in which all information on @xmath268 can be gathered from the coefficients in , on @xmath0 we must also evaluate the auxiliary polynomial @xmath241 .
this presents an additional difficulty in carrying out any numerical simulations with our op .
we now prove a version of theorem [ main1 ] dealing with the three - term recurrence for the orthonormal polynomial @xmath269 .
[ main2 ] let @xmath270 be the orthogonal polynomials defined above .
let @xmath271 and for @xmath16 @xmath272 then , for each @xmath16 @xmath273 where @xmath274 @xmath275 and @xmath276 and @xmath254 were defined in theorem [ main1 ] .
the proof follows from the fact that @xmath277 and .
we prove below certain properties about the coefficients @xmath276 and @xmath254 appearing in the three - term recursion formula .
[ boundbkck ] for each @xmath16 , we have @xmath278 @xmath279 and @xmath280 in particular , @xmath281 where @xmath282 @xmath283 using the expression of @xmath284 given in the proof of theorem [ nmgreen ] , we can write @xmath285 hence , @xmath286 .
similarly , @xmath287 using the fact that @xmath288 , we have @xmath289 from which the second estimate follows .
the last estimate follows by observing that @xmath290 consequently , @xmath291 the last estimate in theorem [ boundbkck ] along with the estimate on the @xmath292 given in theorem [ nmgreen ] , gives another proof that the sequence @xmath293 decays exponentially fast .
we have now proved via two different arguments that the coefficients @xmath293 are very small .
this , together with the fact that we must also evaluate the _ auxiliary _ polynomial @xmath241 prevents us to use theorem [ main1 ] to directly plot the op we construct .
rather we use together with the representation of @xmath250 in definition [ def2 ] to plot these op .
more specifically , we identify each @xmath228 with a vector @xmath294 where @xmath295 for @xmath296 and @xmath297 .
this can be simply written as @xmath298 we now derive a recursion relation for the coefficients @xmath299 appearing in the above representation of @xmath228 .
we shall later , implement this recurrence using arbitrary precision to give accurate graphs of @xmath300 .
[ pkjet ] the recurrence relation for the orthogonal polynomials @xmath250 is accomplished by evaluating the following equations for each @xmath16 . @xmath301 or more specifically , @xmath302 where @xmath303 where defined in theorem [ main1 ] and @xmath304 with @xmath305 given in . using
we can write @xmath306 we now derive a version of the christoffel - darboux formulas for the op @xmath307 . as an application
, we use these formulas to find the expansion of @xmath308 in terms of @xmath309 .
[ christoffel - darboux ] let @xmath310 , @xmath311 , and define @xmath312 . using the notations of the last theorem we have @xmath313 + \sum_{k=0}^{n}\tilde{f}_{k+1}(x)\delta q_{k}(y).\ ] ] in particular , for each @xmath314 and @xmath310 , @xmath315 + \sum_{k=0}^{n}\tilde{f}_{k+1}(x)\delta q_{k}(x).\ ] ]
let @xmath316 , and @xmath311 . by we
have @xmath317-\\ & \qquad \qquad \sqrt{c_{k}}[q_{k}(x)q_{k-1}(y)-q_{k}(y)q_{k-1}(x)].\end{aligned}\ ] ] therefore , summing both sides we obtain @xmath318=\sqrt{c_{n+1}}[q_{n+1}(x)q_{n}(y)-q_{n+1}(y)q_{n}(x)].\ ] ] now , observe that @xmath319 & = & \sum_{k=0}^{n}\tilde{f}_{k+1}(x)q_{k}(y ) + \int_{sg}\sum_{k=0}^{n}g(y , z)q_{k}(z)q_{k}(x)dz\\ & = & \sum_{k=0}^{n}\tilde{f}_{k+1}(x)q_{k}(y ) + \int_{sg}g(y , z)k_{n}(z , x)dz\end{aligned}\ ] ] using this last equation and taking the laplacian with respect to @xmath320 of both sides of yields , @xmath321\bigg ) & = \sqrt{c_{n+1}}[q_{n+1}(x)\delta q_{n}(y)- q_{n}(x)\delta q_{n+1}(y))\\ & = \sum_{k=0}^{n}\tilde{f}_{k+1}(x)\delta q_{k}(y)- k_{n}(x , y)\end{aligned}\ ] ] consequently , @xmath322 + \sum_{k=0}^{n}\tilde{f}_{k+1}(x)\delta q_{k}(y).\ ] ] we now use this theorem to find the coordinates of @xmath323 in terms of lower order polynomials .
that is since , @xmath324 , we have @xmath325 where @xmath326 .
these coefficients can be computed recursively using theorem [ christoffel - darboux ] .
more specifically , we have [ cd - cor ] let @xmath327 , and assume @xmath328 then , @xmath329 and for @xmath330 we have @xmath331 with the initial condition @xmath332 where @xmath333 furthermore , for each @xmath327 and @xmath334 , @xmath335 use @xmath336 in , gives @xmath337 , and integrating this equality leads to @xmath338 .
hence , @xmath339 .
next , write @xmath340 .
using @xmath341 in gives @xmath342 + \tilde{f}_{2}(x)\delta q_{1}(x)\ ] ] where we use the fact that @xmath343 . integrating this last expression leads to @xmath344+\tfrac{1}{\sqrt{c_{1}}}{\langle\tilde{f}_{2},q_{0}\rangle}=\sqrt{c_{2 } } a^{(2)}_{1 } + \tfrac{1}{\sqrt{c_{1 } } } \sqrt{c_{1}}.\ ] ] therefore , @xmath345 . to compute @xmath346
we utilize with @xmath341 . in particular , we multiply @xmath347 by @xmath348 and integrate the resulting equation with respect to both @xmath349 and @xmath320 . this will give @xmath350 consequently , @xmath351 observe that @xmath352 was used in the above arguments .
assume now that we have determined the coefficients @xmath353 of @xmath308 with respect to the orthonormal set @xmath354 . we can use an induction argument to find the coordinates of @xmath355 in the orthonormal set @xmath356 .
set @xmath357 in , and integrate the resulting equation to obtain @xmath358 . to obtain the remaining coefficients , @xmath359 for @xmath360 , set @xmath357 in ,
multiply the resulting equation by @xmath361 and integrate with respect to both @xmath349 and @xmath320 . to obtain the positivity of the coefficients @xmath362
, we proceed by induction on @xmath363 . clearly , @xmath364 , and @xmath365 since @xmath366 for all @xmath363 .
in addition , @xmath367 .
for assume , that for @xmath368 , and each @xmath334 , we have @xmath369
. then , @xmath370 and for each @xmath334 we have @xmath371 since @xmath286 and @xmath369 by the induction hypothesis .
this concludes the proof .
note that theorem [ christoffel - darboux ] and corollary [ cd - cor ] hold for the symmetric op @xmath372 defined in subsection [ subsec4 ] .
we end this section with an investigation of the tri - diagonal symmetric jacobi matrix corresponding to the op @xmath373 . as in the classical case of op defined on the real line ,
each of the families of orthogonal polynomials we constructed above is related to a tri - diagonal , symmetric jacobi matrix that we denote @xmath374 .
in particular , we can write of theorem [ main2 ] in a matrix form that will involve this jacobi polynomial .
to do this we use the following notations : @xmath375^{t}\qquad q^{(n)}=[q_0 , q_1 , \hdots , q_{n-1}]^{t},\ ] ] and @xmath376 by writing in a matrix - vector form we have for each @xmath377 @xmath378 where @xmath379^{t}$ ] . for the op we consider in this paper , though @xmath374 is invertible for all @xmath380 , its determinant is extremely small .
recall ( * ? ? ?
* theorem 0.9.10 ) that the determinant of a tri - diagonal symmetric jacobi matrix obeys the following recursion formula .
let @xmath381 , then @xmath382 with initial conditions @xmath383 and @xmath384 . to illustrate the fact that the jacobi matrices associated to the op considered here have small determinant , we show below a graph of @xmath385 as a function of @xmath380 .
this graph is generated using the antisymmetric op that will be constructed explicitly in the next section . of the determinants of the jacobi matrix , @xmath386 ]
we shall now generate graphs of the op we constructed above starting from two families of monomials on sg .
we also present graphs of the sequences appearing in the three - terms recursion formula .
we start by considering a family of antisymmetric monomials .
recall the monomials @xmath387 in definition [ def2 ] are centered around the point @xmath200 and antisymmetric with respect to that point .
we now use these monomials to generate the family of antisymmetric orthogonal polynomials @xmath388 and their corresponding normalized version @xmath389 .
we do this inductively using theorems [ main1 ] and [ pkjet ] . in the process
, we compute the sequence @xmath390 and @xmath276 .
more specifically , it is easily seen that @xmath391 .
next we use theorem [ pkjet ] to find the auxiliary polynomial @xmath392 notice that we impose the condition @xmath393 .
thus , @xmath394 can now be computed , which , together with the polynomial @xmath395 , is used to find the polynomial @xmath396 .
next , @xmath397 is computed again using theorem [ pkjet ] , from which one gets @xmath398 .
the next step is to compute @xmath399 . using this , @xmath400 and the polynomial @xmath401
are calculated yielding @xmath402 .
continuing these recursions , we construct @xmath250 , @xmath368 and all the related sequences .
to better display the results of these computations , we use a logarithmic scale to plot the numerical sequences obtained above .
in particular , we plot @xmath403 , @xmath404 , and @xmath405 versus @xmath363 .
these graphs are displayed in figure [ fig : dk ] .
in particular , these graphs illustrate theorems [ nmpk ] and [ boundbkck ] .
that is @xmath406 tends to infinity exponentially fast .
in addition the graphs of @xmath404 and @xmath407 suggest that @xmath408 for some @xmath409 .
this seems to be the case for @xmath410 , but we do not have data beyond this range of @xmath363 to confirm or infirm the observation . , @xmath404 , and @xmath407 in log scale for the antisymmetric op @xmath250.,title="fig : " ] , @xmath404 , and @xmath407 in log scale for the antisymmetric op @xmath250.,title="fig : " ] , @xmath404 , and @xmath407 in log scale for the antisymmetric op @xmath250.,title="fig : " ] in figure [ fig:6aop ] , we plot the orthonormal polynomials @xmath411 for @xmath412 , on @xmath413 the level @xmath414 approximation to @xmath0 . to do this
, we use the recursion formula given by theorem [ pkjet ] along with the second half of the representation of @xmath228 given in , as well as the graphs of the monomials @xmath415 which were given in @xcite .
moreover , most of the computations have been carried out to arbitrary precision since the coefficients @xmath416 are very small .
we refer to table [ tab : pk ] for approximate values of these coefficients for @xmath228 , @xmath417 .
the fact that @xmath418 are antisymmetric is apparent from figure [ fig:6aop ] .
in addition , we observe that for @xmath419 , @xmath420 is larger compare to @xmath421 for any @xmath422 .
this is consistent with the behavior of classical op such as the legendre polynomial on @xmath52 $ ] ; see @xcite .
.@xmath250 and their coefficients @xmath423 [ cols= " > , > , > , > , > , > , > , > " , ] four of the symmetric orthogonal polynomials @xmath424 , corresponding to @xmath425 and @xmath95 are shown in figures [ fig:6sop ] . as observed for the antisymmetric op , @xmath426 , which is constant for @xmath427 ( due to symmetry ) is very large compare to the values of @xmath372 at non - boundary points in @xmath0 . by combining both the antisymmetric and symmetric op , we can form an orthogonal system of polynomials in @xmath30 .
notice , that this system will not span the whole space @xmath79 , since the space of polynomials is not dense in @xmath30 either . using lemma [ lemma1 ]
we can prove that the set @xmath428 forms a tight frame for its span @xmath429 .
that is , there exists a constant @xmath430 such that for each @xmath431 we have @xmath432 for more on frame theory we refer to @xcite .
[ tf ] for each @xmath154 , the set @xmath433 forms a tight frame for its two - dimensional span @xmath429 with frame bound @xmath434 .
the frame bounds of @xmath433 are determined by the eigenvalues of its gram matrix @xmath67 given by @xmath435 g has eigenvalues @xmath436 .
this automatically shows that the frame operator @xmath437 is the @xmath438 diagonal matrix with @xmath439 on the diagonal .
thus , @xmath433 is a tight frame with frame bound @xmath440 recall that @xmath441 are the symmetric orthonormal polynomials , and for each @xmath442 , corresponds a family of antisymmetric orthonormal polynomials @xmath443 . with these notations , the following result holds .
[ opcollection ] the set @xmath444 is an orthonormal basis for a subspace @xmath445 of @xmath446 . for @xmath28 and @xmath447 ,
let an element in the above system be defined by @xmath448 then , the inner product of two orthonormal system elements is : @xmath449 recall that @xmath450 is fully symmetric and @xmath451 is antisymmetric , so their inner product is zero for all @xmath137 and @xmath380 .
moreover , if @xmath452 , by the gauss green formula , we can rewrite @xmath453 we now evaluate the right - hand side of the last equation .
when @xmath454 or @xmath455 , @xmath456 , and this eliminates two terms for each @xmath137 .
let @xmath457 , then @xmath458 ( by symmetry ) .
if we let @xmath459 , then @xmath460 . if we let @xmath461 , then @xmath462 since @xmath463 for all j. then we have @xmath464=2 a_j c_j = 0\ ] ] more generally , the same argument can be used to prove that , @xmath465 . consequently , @xmath466 notice
that using a different normalization we can show that the set @xmath467 is also an orthogonal system for @xmath446 .
four of these orthonormal polynomials are plotted in figure [ fig:6onb ] , and as might be expected , these op do not seem to possess any obvious symmetry .
classical orthogonal polynomial theory suggests that zeros of higher order polynomials should interweave between the zeros of the next higher order polynomial .
specifically , if @xmath468 is a zero of @xmath469 , then there exist @xmath470 and @xmath471 zeros of @xmath472 such that @xmath473 . on @xmath0
, we have not been able to establish a similar result .
in fact , it is not clear what interweaving will mean in the fractal setting .
more generally , on @xmath0 , the zero sets of the op seem very difficult to fully describe .
however , using we see that an exact characterization of the zero set of @xmath474 is the set of all points @xmath475 such that @xmath476 though this looks similar to the characterization of the zeros of classical orthogonal polynomials in terms of the eigenvectors of a jacobi matrix , there is a main difference in that @xmath477 is a function of the auxiliary polynomials @xmath478 , @xmath479 despite this difficulty , we have some numerical data , suggesting some structures in these zero sets .
in particular , we shall display below nodal domains corresponding to some of these op .
we first consider the zeros of the op on the edges of @xmath0 .
notice that there are two types of edges , which we call bottom edge ( the edge across from @xmath200 ) and side edges ( edges which meet @xmath200 ) .
we present some numerical data in figures [ fig : qbottomedges ] , [ fig : sbottomedges ] , and [ fig : edges ] , which seem to indicate that the zeros on these edges behave like zeros of classical orthogonal polynomials . because @xmath372 is fully symmetric its behavior on side edges is the same as their behavior on the bottom edge .
the restriction of @xmath418 to the side edge of @xmath0 as illustrated in figure [ fig : edges ] .
, @xmath480 on the bottom edge of @xmath0,title="fig : " ] , @xmath480 on the bottom edge of @xmath0,title="fig : " ] , @xmath480 on the bottom edge of @xmath0,title="fig : " ] , @xmath480 on the bottom edge of @xmath0,title="fig : " ] , @xmath481 on the bottom edge of @xmath0,title="fig : " ] , @xmath481 on the bottom edge of @xmath0,title="fig : " ] , @xmath481 on the bottom edge of @xmath0,title="fig : " ] , @xmath481 on the bottom edge of @xmath0,title="fig : " ] , @xmath482 on a side edge of @xmath0,title="fig : " ] , @xmath482 on a side edge of @xmath0,title="fig : " ] , @xmath482 on a side edge of @xmath0,title="fig : " ] , @xmath482 on a side edge of @xmath0,title="fig : " ] more generally , we have been able to numerically generate the nodal domains of both the first @xmath483 antisymmetric and first @xmath483 symmetric op on @xmath0 . for example , the nodal domains of a polynomial @xmath484 on @xmath0 are defined as follows : let @xmath485 be the zero set ( or the nodal set ) of @xmath484 .
then , @xmath486 can be partitioned into finitely many connected domains @xmath487 , where @xmath488 depends on the degree of the polynomial @xmath484 .
these domains @xmath489 are _ the nodal domains _ of @xmath484 .
we refer to @xcite for color graphs that support the above bound on @xmath488 .
as mentioned in the introduction , we have generated many data for all the op constructed here .
these data as well as the codes used to generate them are available at @xcite .
certain data that are not included in the present article , deal with the dynamics of the op at certain points of @xmath0 .
that is , a choice of op type and a fixed point @xmath490 yield a set of points @xmath491 in the plane .
the passage from @xmath363 to @xmath263 is thought of as a dynamical system in the plane with some `` attractor '' toward which the points tend as @xmath492 . with our numerics , we had hoped to get a `` snapshot '' of this attractor .
the computational complexity associated with our construction , limited us to to generate data for these dynamics only for @xmath493 .
thus , we could not make some conclusive statements about the long term behavior of these dynamics .
however , we refer the interested reader to @xcite for graphs of some of these dynamics .
our investigation in this context is parallel to recent results for the dynamics of certain op related to self - similar measures on the unit interval @xcite .
k. a. okoudjou was partially supported by the alexander von humboldt foundation , by onr grants : n000140910324 @xmath494 n000140910144 , and by a rasa from the graduate school of umcp .
he would also like to express its gratitude to the institute for mathematics at the university of osnabrck for its hospitality while part of this work was completed .
r. s. strichartz was supported in part by the national science foundation , grant dms-0652440 .
e. k. tuley was supported in part by the national science foundation through the research experiences for undergraduates at cornell , and by the national science foundation graduate research fellowship , grant dge-0707424 . n. ben - gal , a. shaw - krauss , r. s. strichartz , and c. young , _ calculus on the sierpinski gasket .
point singularities , eigenfunctions , and normal derivatives of the heat kernel , _ trans .
* 358 * ( 2006 ) , no . 9 , 38833936 . | the construction of a laplacian on a class of fractals which includes the sierpinski gasket ( * @xmath0 * ) has given rise to an intensive research on analysis on fractals .
for instance , a complete theory of polynomials and power series on @xmath0 has been developed by one of us and his coauthors .
we build on this body of work to construct certain analogs of classical orthogonal polynomials ( op ) on @xmath0 .
in particular , we investigate key properties of these op on @xmath0 , including a three - term recursion formula and the asymptotics of the coefficients appearing in this recursion .
moreover , we develop numerical tools that allow us to graph a number of these op .
finally , we use these numerical tools to investigate the structure of the zero and the nodal sets of these polynomials . |
the problem of pricing american and bermudan style options is of fundamental importance in option pricing theory . in the continuous time settings
mckean ( 1965 ) proposed an algorithm for pricing an american put option with an infinite maturity via ordinary differential equations and partial differential equations .
further developments of this technique and applications to other american style securities are considered in peskir and shiryaev ( 2006 ) .
another approach is pricing via monte - carlo simulations that is described by glasserman ( 2004 ) .
one of the most difficult tasks in the theory of pricing of american and bermudan options is the determination of an optimal stopping rule and the valuing of the option under such a rule .
longstaff and schwartz ( 2001 ) proposed an algorithm for pricing american and bermudan style options via monte carlo simulations , least squares monte carlo or lsm .
this technique is especially useful when we deal with multi - factor processes . in this case
the methods based on binomial , trinomial trees , or partial differential equations become slow and thus inefficient due to the high dimensionality of the problem . as in the majority of the numerical algorithms the starting point of lsm for american options is a substitution of the continuous time interval with a discrete set of exercise dates .
practically , by doing this we substitute the american option with a bermudan one .
then for each exercise time ( except the first and the last one ) we project the value of continuation onto a set of basis functions via linear regression .
clement , lamberton and protter ( 2002 ) investigated the convergence of the algorithm with the growth of the number of the basis functions and the monte carlo simulations .
under fairly general conditions they proved the almost sure convergence of the complete algorithm . also , they obtained the rate of convergence when the number of monte carlo simulations increases and showed that the normalized error of the algorithm is asymptotically gaussian .
however , they considered a fixed partition of the time interval and thus , essentially , they discussed the properties of the bermudan , not american option .
glasserman and yu ( 2004 ) investigated the behavior of lsm with the simultaneous grows of the number of the basis functions and the number of the monte - carlo simulations and estimated the rate of convergence in some more specific settings .
moreno and navas ( 2001 ) considered the lsm for different basis functions , namely , power series , laguerre , legendre , chebyshev polynomials , and deduced that the algorithm converges at least for american put options when the underlying problem has a small number of factors .
stentoft ( 2004 ) obtained the rate of convergence of the algorithm in the two period multidimensional case . in the present work we consider the stability of lsm algorithm , when the number of exercise dates increases in such a way that there are exercise dates close to an initial time , which we assume to be equal to zero without loss of generality .
we prove that the algorithm is unstable when the time parameter is close to zero , because the underlying regression problem is ill - conditioned .
the remainder of this work is organized as follows . in section [ description ] , we describe the algorithm . in section [ ill - conditioning ] , we prove the main result , which is formulated in proposition [ mainprop ] , instability of the algorithm for the small values of the time parameter due to the ill - conditioning of the corresponding matrix in the regression problem .
in addition we present the results of the numerical simulations that illustrate the assertions of proposition [ mainprop ] .
in section [ conclusion ] , we give the concluding remarks .
assume that the stock price process @xmath0}$ ] is given by the strong soluton to the following stochastic differential equation : @xmath1 here @xmath2 is a constant , @xmath3}$ ] is a brownian motion ( possibly multidimensional ) on a filtered probability space @xmath4 } , \mathbb{p}\right),$ ] where @xmath5 is the risk - neutral probability measure , @xmath6 and @xmath7 are the progressively measurable functionals , such that the strong solution to equation ( [ sde ] ) exists and unique on the time interval @xmath8 $ ] , see karatzas and shreve @xcite , chapter 5 , for the discussion of this topic .
the time horizon is @xmath9 which we assume to be a finite constant .
usually equations of such a form are used to describe the evolution of the stock prices in practice .
let the payoff of an american option at the time of exercise @xmath10 is given by @xmath11 where @xmath12 is the corresponding payoff function .
then the value of the option is determined by the formula : @xmath13.\ ] ] here @xmath14}$ ] is an interest rate process , which we assume to be deterministic for simplicity ; @xmath15 is the set of the stopping times with respect to the filtration @xmath16}.$ ] to approximate @xmath17 numerically let us conduct @xmath18 monte - carlo simulations of the process @xmath19 .
first , we need to divide the time interval @xmath20 $ ] into @xmath21 subintervals @xmath22 $ ] of the length @xmath23 , where @xmath24 , @xmath25 .
thus at every moment @xmath26 we obtain @xmath18 realizations of the process @xmath27 , @xmath28 .
second , for each simulation we compute the value of the option at time @xmath29 ( under the assumption that the option was not exercised before @xmath30 ) : @xmath31 discounting these values we get a cash flow vector @xmath32 where @xmath33 is the discount factor . to obtain the value of the option at @xmath34 @xmath35 ( under the assumption the option
was not exercised before @xmath36 ) , we chose a hypothesis of linear regression and project the cash flow vector @xmath37 for example , on a constant , @xmath38 and @xmath39 . according to @xcite
this is one of the simplest yet successful regression models . according to @xcite a good alternative choice of basis functions can be hermite , laguerre , legendre , or chebyshev polynomials .
if we use @xmath40 and @xmath41 as the basis , the estimate of the conditional expectation becomes @xmath42= \alpha + \beta x_{t_{m-1 } } + \gamma x^2_{t_{m-1}},\ ] ] where @xmath43 , @xmath44 , @xmath45 are some constants .
then along every path we compare values of immediate exercise , @xmath46 , with values of continuation that are obtained by substitution of @xmath47 into equation ( [ condexpreg ] ) .
the bigger of two gives @xmath48 , @xmath49 . if value of immediate exercise is bigger we set @xmath50 .
similarly we obtain @xmath51 $ ] for each @xmath52 via solving linear regression problems @xmath53 where components of the matrix @xmath54 depend on the regression hypothesis and the outcomes of the simulations , @xmath55 is an unknown vector of coefficients , and the vector @xmath56 is given by equation @xmath57 where @xmath58 is the discount factor .
finally we discount the cash flow up to the moment of time @xmath59 and compare it with the value of immediate exercise at time @xmath60 , @xmath61 .
the bigger is the value of the option .
let @xmath63 be a partition of the interval @xmath64.$ ] in order to compute the estimates of the value of the option at each @xmath65 ( under the assumption that it was not exercised before ) we solve the linear regression problem @xmath66 where @xmath67 is an unknown vector of coefficients , vector @xmath68 is determined by equation ( [ cashflowvector2 ] ) , and the matrix @xmath69 depends on the regression hypothesis and the outcome of the monte carlo simulations .
assume that we have chosen @xmath70 continuous functions @xmath71 as the hypothesis . examples of such functions are power series , laguerre , legendre , hermite polynomials , etc .
in this case @xmath69 has the following form @xmath72 where @xmath18 is the number of the simulations .
we show that , if the underlying process @xmath19 is almost surely continuous , then for small @xmath62 the problem ( [ regression ] ) is ill - conditioned . for a matrix @xmath73 in the @xmath74 norm
the * condition number * is defined as @xmath75 where @xmath76 and @xmath77 are maximal and minimal singular values of the matrix @xmath73 respectively .
a problem with a low condition number is called * well - conditioned * , while a problem with a high condition number is called * ill - conditioned*. usually , problem ( [ regression ] ) is solved via one of the following methods : householder triangularization , gram - schmidt orthogonalization , singular value decomposition , or normal equations .
let @xmath78 be the condition number of matrix @xmath73 .
if @xmath79 exists then the exact solution to the least - squares problem is given by the vector @xmath80 i.e. it is a product of the left - inverse of the matrix @xmath73 and the vector @xmath81 one can see that the solution to the problem ( [ regression ] ) , obtained via normal equations , is governed by @xmath82 , whereas the solution obtained via svd , householder or gram - schmidt is governed by @xmath78 .
consequently , normal equations are the least stable with respect to the grows of the condition number .
nevertheless , the analytical solution to the least - squares problem is defined in terms of the normal equations .
let @xmath83 denote the condition number of the matrix @xmath69 for @xmath84.$ ] we show below that @xmath85 @xmath5-almost surely .
therefore , no matter what algorithm one uses ( householder , gram - schmidt , svd or normal equations ) , for small values of time the underlying regression problem is ill - conditioned , and thus the algorithm is unstable .
we prove ill - conditioning for an arbitrary number of basis functions in the following proposition .
in addition , we illustrate the phenomenon with the results of the numerical simulations in the case when regression is done on three basis functions @xmath86 , @xmath87 , and @xmath88 as a function of @xmath89 @xmath90,$ ] @xmath19 - lognormal process , @xmath91 @xmath92 @xmath93 paths , @xmath94 time steps , milstein discretization scheme.,scaledwidth=90.0% ] as a function of @xmath89 @xmath95,$ ] @xmath19 - normal process ,
@xmath91 @xmath96 @xmath93 paths , @xmath94 time steps , euler discretization scheme.,scaledwidth=90.0% ] [ mainprop ] assume that the process @xmath19 is given by equation ( [ sde ] ) , @xmath97 , @xmath98 , are continuous functions , such that @xmath99 let for each @xmath84 $ ] the matrix @xmath69 is defined by equation ( [ matrixa ] ) and @xmath83 is the condition number of @xmath100 then @xmath101 = 1.\ ] ] * proof.*consider equation ( [ matrixa ] ) .
it follows from equation ( [ sde ] ) that all rows of @xmath102 are identical .
consequently the rank of @xmath102 equals to @xmath86 .
let us look at the following matrix : @xmath103 the matrix @xmath104 has two eigenvalues : @xmath105 is the first one ( with multiplicity @xmath86 ) , @xmath106 is the second one ( with multiplicity @xmath107 ) .
thus @xmath102 has singular values @xmath108 and @xmath109 consequently @xmath110 . note that the matrix @xmath111 is real and symmetric , thus the eigenvalues of @xmath111 are real for all @xmath112s . since eigenvalues are continuous functions of the components of the matrix and the components in turn are a.s .
continuous processes , we deduce that for small @xmath112s the first eigenvalue is in the neighborhood of @xmath113 which is greater then zero by the assumption of proposition , whereas all other eigenvalues are is in the neighborhood of @xmath106 .
the conclusion , @xmath114 = 1 $ ] follows from continuity of the underlying process @xmath19 and the basis functions @xmath115 s , @xmath116 ( 6,6)[br ] + intuitively proposition [ mainprop ] shows that for small values of @xmath62 the condition number is large .
we proved that for a continuous underlying stock price process , lsm algorithm for pricing american options is unstable when time parameter is small .
an interesting question is to obtain an exact bound on applicability of this algorithm .
a possible criterion of applicability is the condition number of matrix @xmath69 , @xmath117 .
for example , if @xmath118 exceeds a certain value , one can treat ( [ regression ] ) as a rank deficient least squares problem ( see @xcite for details ) , or switch to another method : backward induction or the method introduced by mckean @xcite of option pricing via ordinary differential equations or partial differential equations considered on a smaller domain . for certain problems
it is possible to obtain desired accuracy using relatively small number of time intervals , then one does not have to solve the regression problem for small @xmath112s , and consequently the algorithm can be stable . also , if the underlying process @xmath19 is discontinuous with high probability the algorithm can be stable even for small values of the time parameter . | consider least squares monte carlo ( lsm ) algorithm , which is proposed by longstaff and schwartz ( 2001 ) for pricing american style securities .
this algorithm is based on the projection of the value of continuation onto a certain set of basis functions via the least squares problem .
we analyze the stability of the algorithm when the number of exercise dates increases and prove that , if the underlying process for the stock price is continuous , then the regression problem is ill - conditioned for small values of the time parameter .
_ * keywords : * _ option pricing , optimal stopping , american option , least squares monte carlo , monte carlo methods , stability , ill - conditioning . |
metals that contract upon heating are exceedingly rare , if they exist at all .
this observation can be understood by considering the types of cohesive interactions bonding atoms into solids . as a material is heated
, interatomic distances expand due to the asymmetric shape of the potential energy well and therefore to anharmonic effects.@xcite a chemical bond with a deeper potential well will confine atomic vibrations to remain more harmonic and consequently , covalently bonded materials such as diamond exhibit lower thermal expansion than metals such as copper . in some covalent crystals with open framework structures ,
the low thermal expansion can even become negative . among the negative thermal expansion ( nte ) materials
most studied are framework oxides such as zrw@xmath3o@xmath4@xcite and more recently metal cyanide frameworks such as zn(cn)@xmath3,@xcite all of which are electrical insulators or semiconductors .
in contrast , the oxide reo@xmath0 is a superb electrical conductor with a room temperature resistivity close to that of copper.@xcite recently , chatterji _ et al . _
have demonstrated through neutron powder diffraction and first - principles calculations that reo@xmath0 also exhibits isotropic nte,@xcite making it the only known simple binary oxide to combine metallic conductivity with nte . [ !
reo@xmath5 octahedra connect via corner - sharing in infinite chains in all three directions .
the atomic displacement ellipsoids of the oxygen atoms ( in red ) are represented at 90 % probability ; their flat shape indicates transverse displacement normal to the re o re bonds and into the void in the middle of the cubic cell.,title="fig : " ] a topological feature that can lead to nte behavior in an open framework structure is a linear chain consisting of a ligand coordinated to two metals or _ vice versa_.@xcite for example , simple binary oxides such as cu@xmath3o , ag@xmath3o ( o m
o chains)@xciteand certain phases of sio@xmath3 ( si o si chains)@xcite demonstrate nte behavior . while this is not a requisite for nte behavior ( magnetostriction and phase transitions can also result in nte ) , it is a simple principle by which new nte materials can be pursued .
the structure of rhenium trioxide contains such linear chains in a three dimensional framework consisting of corner - sharing reo@xmath5 octahedra ( see fig.[reo3_structure ] ) .
the simple reo@xmath0 structure is a prototype for closely related materials such as the perovskite ferroelectrics , ferromagnets , and high t@xmath6 superconductors and can also be thought of as the cubic perovskite abo@xmath0 with a void in the place of the @xmath7 cation .
thermal expansion investigations of reo@xmath0 have presented conflicting results .
a laser interferometry study on a single crystal concluded that reo@xmath0 displayed nte between 100k and 340k,@xcite .
in contrast to the single crystal work , a catalogue of powder x - ray diffraction ( xrd ) studies reported the coefficient of thermal expansion of reo@xmath0 to be small but positive for all temperature ranges.@xcite the latter study has been more frequently cited in reviews of nte in oxide frameworks,@xcite leaving reo@xmath0 largely unrecognized as a potential nte material until the recent article by chatterji _ et al _ , which showed through diffraction studies nte up to 200 k.@xcite in the present study , we show that reo@xmath0 can exhibit isotropic nte up to room temperature , that the positive to negative transition is dependent on sample quality , and that specific heat measurements show a lack of any low energy einstein - type modes that lead to the nte behavior in other oxides .
furthermore , we show through careful analysis of the atomic displacement parameters the role of static disorder of the oxygen atoms in impeding nte behavior and therefore explain the discrepancy between past thermal expansion studies of reo@xmath0 .
the thermal expansion properties of various reo@xmath0 powder samples were investigated by performing variable temperature neutron powder diffraction ( npd ) with time - of - flight neutron data .
the first sample investigated , labeled reo@xmath0-a throughout this paper , was a powder prepared by the decomposition of an re@xmath3o@xmath81,4 dioxane adduct at 140@xmath9c as first described by nechamkin _
et al_.@xcite the second sample studied , reo@xmath0-b , was a powder purchased from alfa aesar ( 99.99% chemical purity ) .
finally , the third sample , reo@xmath0-c , was prepared by crushing single crystals of reo@xmath0 grown by chemical vapor transport ( cvt ) . by using hgcl@xmath3 as the transport agent in an evacuated pyrex glass ampoule with reo@xmath0 powder ( reo@xmath0-b as the starting powder ) , the crystals were grown across a temperature gradient of about 100@xmath9c as described by feller _ _ et al.__@xcite .
in addition to growing single crystals , cvt also separates the lower oxide phase impurities from the freshly grown reo@xmath0 crystals .
hence , sample reo@xmath0-c was expected to be of higher chemical and phase purity than the others .
npd patterns were obtained on the high intensity powder diffractometer ( hipd ) at the lujan neutron scattering center at the los alamos neutron science center ( lansce ) .
hipd is suitable for studying the behavior of materials as a function of temperature and pressure due to its relatively high data acquisition rates .
hipd data were collected on all samples between 15k and 300k .
the lujan center employs a pulsed spallation neutron source , so the npd histograms were collected in time - of - flight mode on fixed detector banks .
the backscattering detector banks , located @xmath10 with respect to the incident beam , provide the highest resolution pattern and higher - index reflections while the normal ( @xmath11 ) and forward scattering ( @xmath12 ) banks have a lower resolution but probe higher @xmath13-spacings . in the study by chatterji _
_ , the neutron diffraction measurements were performed on a constant wavelength source ( @xmath14 ),@xcite limiting the diffractogram up to @xmath15 @xmath2 in momentum transfer @xmath16 .
more accurate structural parameters such as atomic displacement parameters can be accessed with time - of - flight neutron data since higher @xmath17 ranges can be probed . in our measurements ,
a @xmath18 of @xmath19 @xmath2 was used for all the structural refinements .
the specific heat data were collected upon warming from 2k to 20k on a 28.5 mg single crystal using a quasiadiabatic method as implemented on a quantum design physical property measurement system ( ppms ) . like sample reo@xmath0-c , the single crystal was grown by cvt .
least squares structural refinements by the rietveld method were performed with all six histograms of the npd data using the gsas software package.@xcite in space group @xmath20 , only four variables were refined for each temperature : the lattice parameter ; the isotropic displacement parameter @xmath21 of re , and the anisotropic displacement parameters @xmath22 ( @xmath23 ) , and @xmath24 of o. to determine the lattice parameters of the sample accurately , x - ray powder diffraction was performed at room temperature with si standard reference material 640c purchased from national institute of standards ( see fig.[reo3_xrd_hipd]a ) . the neutron powder profile for the sample prepared from single crystals , reo@xmath0-c ,
is shown in fig.[reo3_xrd_hipd]b . in this sample ,
the lattice parameter plotted as a function of temperature clearly demonstrates nte behavior from 15k up to 300k ( see fig.[lattice_temp]c ) .
nte in the other powder samples was indeed observed but limited to a narrower temperature range ( see fig.[lattice_temp]a , b ) . for sample reo@xmath0-a ,
the nte behavior was observed only up to 110k , and for reo@xmath0-b , only up to 220k .
clearly sample quality has an effect on the thermal expansion properties and this may , in part , explain the discrepancy between past findings on the thermal expansion behavior of reo@xmath0 and the reason why chatterji _
found nte only up to 200 k from their neutron measurements while their dynamical lattice calculations predicted nte up to 350k.@xcite [ !
b ] from x - rays shown in open circles and the calculated profile as a smooth red line .
the x - ray powder patterns contain both sample reo@xmath0-c and a si standard .
upper tickmarks correspond to reo3-c and lower to the si standard .
( b ) observed profile from neutrons is shown in open circles and the calculated as a smooth red line for sample reo@xmath0-c only with corresponding bragg reflection tickmarks below .
the difference curve between the observed and calculated patterns are represented by the smooth purple line below the powder profiles.,title="fig : " ] the linear coefficient of thermal expansion for a cubic system @xmath25 is defined as @xmath26 where @xmath27 is the cubic cell parameter and @xmath28 the temperature .
a least squares polynomial fit ( second - order ) to the lattice parameters obtained from the neutron data affords @xmath29 , which can be differentiated with respect to @xmath28 and divided by the experimental values of @xmath27 to obtain @xmath25 as a function of @xmath28 . for sample
reo@xmath0-c the average @xmath25 below 300k was found to be @xmath30k@xmath2 . for sample reo@xmath0-a , @xmath25 was found to be @xmath31k@xmath2 , and @xmath32k@xmath2 for sample reo@xmath0-b .
thus , lowering of the transition temperature @xmath33 is correlated with a smaller , negative value for @xmath25 . for sample reo@xmath0-c ,
the values for @xmath34 are close to those determined by matsuno _
_ in their interferometry study on single crystals where the value of @xmath25 is @xmath35k@xmath2 at about 120k and slowly rises to @xmath36 k@xmath2 at about 230k.@xcite while we did not make measurements above room temperature , matsuno _ et al .
_ found @xmath25 to be positive and smaller than @xmath37k@xmath2 between 340k and 500k.@xcite in an xrd study carried out on single crystals as well as powder samples , chang _
_ found a value of @xmath38 k@xmath2 above room temperature.@xcite the volumetric @xmath25 calculated by chatterji _
_ is not linear as in our simplified model from the fitted curves , but nevertheless shows a volumetric @xmath25 ( @xmath39 linear @xmath25 ) varying between @xmath35 to zero below 350 k.@xcite [ !
b ] -a , reo@xmath0-b , and reo@xmath0-c .
the smooth black curves are the least - squares , polynomial fits to the lattice parameters .
error bars are obtained from standard uncertainties in the structural refinements.,title="fig : " ] the atomic displacement parameters obtained from the npd data offer some clues as to the mechanism responsible for nte in reo@xmath0 . due to the high range of momentum transfer @xmath17 available from time - of - flight neutrons as well as the lack of any decay with @xmath17 in the atomic form factors for neutron diffraction
, accurate atomic displacement parameters were obtained from the rietveld analysis ( see fig.[thermal_temp]a ) .
the re atom is located on the special position @xmath40 ( site symmetry @xmath41 ) , which constrains the atomic displacement parameter to remain isotropic .
the oxygen atoms are located on the position @xmath42 ( site symmetry @xmath43 ) , constraining them to have only two values , @xmath22 ( @xmath44 ) and @xmath45 . while the displacement of the o atoms along the re o re direction remains fairly constant , the transverse motion increases upon heating , as also observed by chatterji _
et al_.@xcite the most striking feature is the gap between @xmath22 ( @xmath23 ) and @xmath45 as represented by the flatness of the displacement ellipsoids of the o atoms ( see fig.[reo3_structure ] ) .
the anisotropy only increases with temperature , suggesting that the transverse motion contracts the cubic lattice upon heating and is therefore the most likely mechanism for the observed nte behavior .
[ ! b ] -c , where @xmath21 represents the rhenium isotropic displacement parameters and @xmath46 the oxygen anisotropic displacement parameters .
( b ) the ratio @xmath22/@xmath45 for sample reo@xmath0-a and ( c ) the same ratio for samples reo@xmath0-b and reo@xmath0-c . @xmath22 and
@xmath47 represent atomic displacement of the oxygen atoms normal to the re o re bond , while @xmath45 to displacement along the bond .
error bars , shown within the symbols in ( a ) , are obtained from the standard uncertainties in the structural refinements.,title="fig : " ] further inspection of the oyxgen displacement parameters of all three samples reveals differences that may explain the varying thermal expansion behavior .
though sometimes loosely referred to as thermal parameters or debye - waller factors , the values for @xmath48 have contributions from both dynamic and static displacements of the atoms .
extrapolation of the @xmath48 parameters for reo@xmath0 down to 0k reveals non - zero values , which indicates significant contribution from static displacement , especially in the case of @xmath22 ( see fig.[thermal_temp]a ) .
the large gap between @xmath22 and @xmath45 at low temperatures was observed for all three samples . a useful parameter to describe the extent of anisotropy in the atomic displacements is the ratio of the maximum value to the minimum value of @xmath48;@xcite in our case , it is @xmath22/@xmath45 .
when plotted versus temperature ( see fig.[thermal_temp]b , c ) , this measure of anisotropy reveals that it is constant ( within the error bars ) and large for sample reo@xmath0-a , while it is increases upon heating for reo@xmath0-b and reo@xmath0-c .
most importantly , the higher @xmath22/@xmath45 ratios for reo@xmath0-a and reo@xmath0-b than for reo@xmath0-c at low temperatures suggest more static disorder in the first two , which should affect the lattice dynamical properties .
the @xmath49 mode calculated by chatterji _
et al _ involves rotation of the reo@xmath5 octahedra in the @xmath50 plane and arises from transverse motion of the o atoms in this plane.@xcite since the @xmath49 mode has the largest contribution to the negative value for @xmath25 at low temperatures , static displacement of the o atoms , transversely from the re o re bond , would diminish the nte behavior .
( a ) data fit below 4k to the function @xmath51 ( where @xmath52 ; @xmath53 is the gas constant ) .
the fitting allows the sommerfeld coefficient @xmath54 = 3.6mjmol@xmath2k@xmath55 and a debye temperature of @xmath56 = 344k to be obtained .
( b ) data plotted as @xmath57 _ vs. _ @xmath58 in order to reveal any low energy phonon spectral weight .
the absence of any peaks indicates a lack of contribution from any einstein - type modes ; the debye contribution appears as a constant up to @xmath56 and the rise at low @xmath59 is due to the electronic contribution ( @xmath60 ) of the delocalized electrons.,title="fig : " ] specific heat measurements on a single crystal of reo@xmath0 were carried out to understand the type of lattice vibrations contributing to the specific heat .
the low temperature behavior in the heat capacity of reo@xmath0 is that of a straightforward debye solid , with a debye temperature @xmath56 of 344k and a sommerfeld coefficient @xmath54 of 2.8mjmol@xmath2k@xmath55 ( see fig.[heat_cap]a ) .
these values agree very well with the values of 327k for @xmath56 and 2.85mjmol@xmath2k@xmath55 for @xmath54 found in the early studies by king _ _ et al.__@xcite fig.[heat_cap]b shows the specific heat over @xmath61 vs. @xmath59 , which is an approximate representation of the one - dimensional phonon density of states of a solid.@xcite a signature of an einstein mode in this representation is a gaussian - like peak clearly missing in the low temperature region of the specific heat .
these measurements strengthen the argument by chatterji _ et al . _ from their lattice dynamical calculations that nte in reo@xmath0 is caused by low energy acoustic modes with the @xmath49 mode contributing the most to the nte behavior.@xcite also , the specific heat results rule out the suggestion by matsuno _
_ that a low energy , bending - like optical mode is responsible for nte in reo@xmath0.@xcite although both zrw@xmath3o@xmath4 and reo@xmath0 have cubic symmetry , it is interesting to compare how their structures lead do different nte behavior . the primitive cubic structure of reo@xmath0 ( given the absence of the @xmath7 cation in its structure and thus an extremal value of the perovskite tolerance factor ) is of course the consequence of metallicity : the large volume is maintained by the fermi pressure of the delocalized d - electrons in the @xmath62 conduction band .
nevertheless , the proximity to a ` buckling transition ' is revealed by a pressure - induced collapse of the structure produced by cooperative counter - rotation of the octahedra,@xcite in this case very similar to zrw@xmath3o@xmath4.@xcite the linear @xmath25 obtained for reo@xmath0 is smaller than that observed in zrw@xmath3o@xmath4 , where @xmath63 k@xmath2 .
indeed , the smaller absolute value of the nte and the smaller temperature range at which it occurs in reo@xmath0 compared to zrw@xmath3o@xmath4 is consistent with the much weaker precursor effects in the pressure - induced collapse in the former than in the latter.@xcite in turn , the weakness of the fluctuations in reo@xmath0 is possibly a consequence of it possessing fewer soft degrees of freedom than the more open framework structure of zrw@xmath3o@xmath4 . in addition , the low energy einstein - type modes observed in zrw@xmath3o@xmath4 and implicated in that material s nte behavior,@xcite are clearly missing in reo@xmath0 . the relatively small nte effect observed in reo@xmath0 remains , nevertheless , remarkable because it is isotropic and unprecedented for a material with the electronic properties of a simple metal .
while classifying an oxide as a simple metal is atypical , the free - electron model works well enough in reo@xmath0 that it qualifies under the scheme of electrical transport properties ( other examples include the na@xmath64wo@xmath0 bronzes and ruo@xmath3).@xcite most simple metals have a much larger @xmath25 than that of reo@xmath0 .
for example , for copper , @xmath25 is @xmath65k@xmath2 at room temperature.@xcite another notable exception is the intermetallic ygage , which has such small @xmath25 that its behavior has been termed zero thermal expansion.@xcite like ygage , reo@xmath0 is also remarkable in that its @xmath25 above room temperature is as low as in a highly covalent material like diamond where @xmath25 is @xmath66k@xmath2.@xcite while reo@xmath0 does exhibit highly metallic behavior , the strong covalency between the re d - states and o p - states and its unique framework topology allow for negative thermal expansion , demonstrating the propensity for transition metal oxides to challenge our conventions on bonding and properties in solid materials .
furthermore , reo@xmath0 s relatively simple structure has provided a way to quantify the role of static disorder of the oxygen atoms on interesting behavior such as negative thermal expansion .
this work has benefited from the use of hipd at the lujan center at los alamos neutron science center , funded by doe office of basic energy sciences .
los alamos national laboratory is operated by los alamos national security llc under doe contract de - ac52 - 06na25396 .
we would also like to thank j. c. lashley and a. lawson from los alamos for stimulating discussions and l. l. daeman , also from los alamos , for help with sample preparation .
bcm and rs acknowledge the national science foundation for support through a career award ( nsf - dmr-0449354 ) , and for the use of mrsec facilities ( award nsf - dmr0520415 ) . | time - of - flight neutron powder diffraction and specific heat measurements were used to study the nature of thermal expansion in rhenium trioxide , an electrically conducting oxide with cubic symmetry .
the temperature evolution of the lattice parameters show that reo@xmath0 can exhibit negative thermal expansion at low temperatures and that the transition from negative to positive thermal expansion depends on sample preparation ; the single crystal sample demonstrated the highest transition temperature , 300 k , and largest negative value for the coefficient of thermal expansion , @xmath1 k @xmath2 . for the oxygen atoms , the atomic displacement parameters are strongly anisotropic even at 15 k , indicative of a large contribution of static disorder to the displacement parameters .
further inspection of the temperature evolution of the oxygen displacement parameters for different samples reveals that the static disorder contribution is greater for the samples with diminished nte behavior .
in addition , specific heat measurements show that reo@xmath0 lacks the low energy einstein - type modes seen in other negative thermal expansion oxides such as zrw@xmath3o@xmath4 . |
soit @xmath1 un nombre premier et @xmath6 .
dans la recherche dune correspondance ventuelle entre ( certaines ) reprsentations @xmath1-adiques @xmath7 de @xmath3 de dimension @xmath8 et ( certaines ) reprsentations @xmath1-adiques @xmath9 de @xmath10 , un des premiers cas regarder est certainement celui o @xmath11 et @xmath7 est cristalline , absolument irrductible et @xmath12-semi - simple .
cette dernire condition signifie que le frobenius @xmath12 sur le @xmath12-module filtr @xmath13 associ par fontaine @xmath7 est semi - simple .
la reprsentation @xmath7 a des poids de hodge - tate distincts @xmath14 et , si lon note @xmath15 la reprsentation algbrique de @xmath0 de plus haut poids @xmath16 et @xmath17 la reprsentation lisse irrductible de @xmath0 associe par la correspondance locale de hecke la reprsentation non ramifie de @xmath3 envoyant un frobenius arithmtique sur @xmath18 , la reprsentation @xmath9 est simplement le complt @xmath1-adique de la reprsentation localement algbrique @xmath19 par rapport un rseau stable par @xmath0 et de type fini sous laction de @xmath0 .
ainsi , @xmath9 est un espace de banach @xmath1-adique muni dune action continue unitaire de @xmath0 ( i.e. laissant une norme invariante ) .
notons que , lorsque @xmath17 est de dimension @xmath20 , cette dfinition de @xmath9 doit tre modifie .
le problme est quil
nest pas du tout vident quun tel rseau existe , ou , de manire quivalente , que @xmath9 soit non nul ! dans ( * ? ? ?
* thorme 1.3 ) , la non nullit de @xmath9 est dmontre pour @xmath21 ( essentiellement ) , et dans ( * ? ? ?
* thorme 1.3.3 ) , son admissibilit ( au sens de @xcite ) et son irrductibilit topologique ( avec une condition supplmentaire pour cette dernire ) .
la mthode repose sur le calcul de la rduction dune boule unit de @xmath9 modulo @xmath1 ( plus exactement modulo lidal maximal des coefficients ) .
la restriction ci - dessus sur les poids nest pas thorique mais simplement due au fait que les calculs deviennent de plus en plus difficiles quand les poids grandissent .
lorsque @xmath7 est de dimension @xmath2 , absolument irrductible mais cette fois semi - stable non - cristalline , @xmath9 est dfini dans @xcite et @xcite et des conjectures analogues ( non - nullit , admissibilit , etc . )
formules et trs partiellement dmontres via la thorie globale dans @xcite ou via des calculs de rduction modulo @xmath1 sur le demi - plan @xmath1-adique dans @xcite .
colmez dans @xcite a vu que la thorie des @xmath5-modules de fontaine permettait de dmontrer lgamment ces conjectures dans le cas semi - stable en construisant un modle de la restriction au borel suprieur de la reprsentation duale @xmath22 partir du @xmath5-module de @xmath7 .
il tait donc naturel de regarder si un tel modle existait aussi dans le cas _ a priori _ plus simple des reprsentations @xmath7 cristallines et sil permettait de dmontrer les conjectures ci - dessus .
la rponse est affirmative et fait lobjet du prsent article .
comme dans @xcite , on dmontre donc un isomorphisme borel - quivariant entre @xmath22 et @xmath23 ( thorme [ phigammabanach ] ) o @xmath24 est le @xmath5-module associ la reprsentation cristalline @xmath7 et o la limite projective consiste en les suites @xmath25-compatibles bornes dlments de @xmath24 .
pour montrer cet isomorphisme , il est ncessaire de passer par une autre description intermdiaire de @xmath9 comme espace fonctionnel @xmath1-adique , i.e. comme un espace de fonctions continues sur @xmath4 dun certain type ( thorme [ complete ] )
. les rsultats ct @xmath5-modules permettent alors de dduire le rsultat principal de cet article ( [ resul ] ) : thorme supposons @xmath7 de dimension @xmath2 , cristalline , absolument irrductible et @xmath12-semi - simple , alors @xmath9 est non nul , topologiquement irrductible et admissible . on obtient
aussi deux autres corollaires , lun concernant tous les rseaux possibles stables par @xmath0 dans @xmath19 ( corollaire [ reseaurigolo ] ) , lautre concernant les vecteurs localement analytiques dans @xmath9 ( corollaire [ anal ] ) .
une version prliminaire des rsultats de cet article a fait lobjet dun cours hangzhou en aot 2004 et les notes de ce cours sont disponibles ( en anglais , voir @xcite ) .
terminons cette introduction en insistant sur le fait que des noncs non triviaux sur des reprsentations de @xmath0 ( irrductibilit des @xmath9 etc . )
se dmontrent finalement en passant par des reprsentations de @xmath3 . on fixe @xmath26 une clture algbrique
de @xmath4 , @xmath27 la valuation sur @xmath26 telle que @xmath28 , @xmath29 la norme @xmath1-adique @xmath30 et @xmath31 le complt de @xmath26 pour @xmath32 . on normalise
lisomorphisme de la thorie du corps de classes local en envoyant les uniformisantes sur les frobenius gomtriques . on note
@xmath33 le caractre cyclotomique @xmath1-adique de @xmath3 vu aussi comme caractre de @xmath34 .
en particulier , @xmath35 et @xmath36 est lidentit . on note
@xmath37 le caractre non ramifi de @xmath34 envoyant @xmath1 sur @xmath38 . on dsigne par @xmath39 une extension finie
de @xmath4 , @xmath40 son anneau dentiers , @xmath41 son corps rsiduel et @xmath42 une uniformisante . on note
@xmath43 lextension finie de @xmath4 dans @xmath26 engendre par les racines @xmath44-imes de lunit et @xmath45 . on note @xmath7 une reprsentation @xmath1-adique de @xmath3 , cest - - dire un @xmath39-espace vectoriel
de dimension finie muni dune action linaire et continue de @xmath3 , et @xmath46 un @xmath40-rseau de @xmath7 stable par @xmath3 .
enfin , @xmath47 dsigne les matrices triangulaires suprieures dans @xmath0 et @xmath48 le sous - groupe de ces matrices qui sont dans @xmath49 .
le but de cette partie est de donner des prliminaires sur les @xmath5-modules @xmath24 et leurs treillis @xmath50 qui sont utiliss dans les parties suivantes pour ltude des reprsentations cristallines de @xmath0 .
le but de ce paragraphe est dintroduire certains anneaux de sries formelles ( @xmath51 , @xmath52 et @xmath53 ) , ainsi que certaines des structures do nt ils sont munis et do nt nous avons besoin dans la suite de cet article . * on note
@xmath51 lanneau form des sries @xmath54 telles que @xmath55 et @xmath56 quand @xmath57 .
cest un anneau local de corps rsiduel @xmath58 . * on note @xmath59 $ ] , cest un corps local de dimension @xmath2 . * on note @xmath53
lanneau des sries formelles @xmath60 $ ] qui convergent sur le disque unit , ce qui fait que @xmath61 $ ] sidentifie au sous - anneau de @xmath53 constitu des sries formelles coefficients borns .
le corps @xmath52 est muni de la norme de gauss @xmath62 dfinie par @xmath63 si @xmath64 .
lanneau des entiers de @xmath52 pour cette norme est @xmath51 , et la norme de gauss induit sur @xmath51 la topologie @xmath42-adique .
lapplication naturelle @xmath65 est alors continue , si lon donne @xmath51 la topologie @xmath42-adique et @xmath58 la topologie discrte . on peut dfinir
une topologie moins fine sur @xmath51 , la topologie faible , dfinie par le fait que les @xmath66\}_{i , j \geq 0}$ ] forment une base de voisinages de zro , et la topologie faible sur @xmath67 qui est la topologie de la limite inductive .
cette topologie induit la topologie @xmath68-adique sur @xmath69 $ ] , et lapplication naturelle @xmath65 est alors continue , si lon donne @xmath51 la topologie faible et @xmath58 la topologie @xmath70-adique .
si @xmath71 , alors on peut dfinir une norme @xmath72 sur @xmath53 par la formule : @xmath73 .
lensemble des normes @xmath74 dfinit une topologie sur @xmath53 qui en fait un espace de frchet .
[ dordrer ]
si @xmath75 et @xmath76 , on dit que @xmath77 est dordre @xmath78 ( il serait plus correct de dire dordre @xmath79 ) si pour un @xmath80 tel que @xmath81 , la suite @xmath82 est borne . il est facile de voir que si cest vrai pour un choix de @xmath83 , alors cest vrai pour tout choix de @xmath81 .
un exemple de srie dordre @xmath20 est donn par @xmath84 .
soit @xmath85 .
les anneaux @xmath51 , @xmath52 et @xmath53 dfinis ci - dessus sont munis dune action de @xmath86 , telle que si @xmath87 , alors @xmath88 agit par un morphisme de @xmath39-algbres , et @xmath89 o @xmath90 est le caractre cyclotomique . on vrifie
facilement que @xmath86 agit par des isomtries , pour toutes les normes et topologies dfinies ci - dessus .
ces anneaux sont aussi munis dun morphisme de frobenius @xmath12 , qui est lui - aussi un morphisme de @xmath39-algbres , tel que @xmath91 .
cette application est continue pour toutes les topologies ci - dessus et commute laction de @xmath86 .
remarquons enfin que @xmath12 et laction de @xmath86 stabilisent @xmath92 $ ] .
lanneau @xmath51 est un @xmath93-module libre de rang @xmath1 , do nt une base est donne par @xmath94 .
si @xmath95 , on peut donc crire @xmath96 .
[ defpsianno ] on dfinit un oprateur @xmath97 par la formule @xmath98 si @xmath96 .
on dfinit de manire similaire un oprateur @xmath99 .
cet oprateur vrifie alors @xmath100 et commute laction de @xmath86 .
il ne commute pas @xmath12 et nest pas @xmath51-linaire .
rappelons quune reprsentation @xmath39-linaire est un @xmath39-espace vectoriel @xmath7 de dimension finie muni dune action linaire et continue de @xmath3 .
il est assez difficile de dcrire un tel objet , et nous allons voir dans ce paragraphe quune certaine classe de @xmath5-modules sur @xmath52 permet den donner une description explicite .
un @xmath12-module sur @xmath51 est un @xmath51-module de type fini @xmath101 muni dun morphisme @xmath12-semi - linaire @xmath102 . on crit @xmath103
pour le @xmath51-module engendr par @xmath104 dans @xmath101 et on dit que @xmath101 est tale si @xmath105 .
@xmath12-module sur @xmath52 est un @xmath52-espace vectoriel de dimension finie @xmath101 muni dun morphsime @xmath12-semi - linaire @xmath106 .
on dit que @xmath101 est tale si @xmath101 a un @xmath51-rseau stable par @xmath12 et tale .
un @xmath5-module est un @xmath12-module muni dune action de @xmath86 par des morphismes semi - linaires ( par rapport laction de @xmath86 sur les coefficients ) et commutant @xmath12 .
si @xmath101 est un @xmath5-module tale sur @xmath52 , et si @xmath107 est lanneau construit par fontaine dans @xcite , alors @xmath108 est un @xmath39-espace vectoriel de dimension @xmath109 muni dune action linaire et continue de @xmath3 : cest une reprsentation
@xmath39-linaire de @xmath3 . on a alors
le rsultat suivant ( @xcite ) : [ fontequiv ] le foncteur @xmath110 est une quivalence de catgories * de la catgorie des @xmath5-modules tales sur @xmath52 vers la catgorie des reprsentations @xmath39-linaires de @xmath3 ; * de la catgorie des @xmath5-modules tales sur @xmath51 vers la catgorie des @xmath40-reprsentations de @xmath3 . linverse de ce foncteur est not @xmath111 .
les @xmath5-modules tales nous donnent donc un moyen commode de travailler avec les reprsentations @xmath39-linaires .
en thorie , les @xmath5-modules permettent de retrouver tous les objets que lon peut associer aux reprsentations @xmath39-linaires ( et en pratique , cest souvent le cas ) .
dans ce paragraphe et les suivants , nous allons rappeler quelques unes de ces constructions .
si @xmath101 est un @xmath12-module tale et si @xmath112 , alors on peut crire @xmath113 o les @xmath114 sont bien dtermins . [ defipsi ] on dfinit un oprateur @xmath115 par la formule @xmath98 si @xmath96 .
cet oprateur vrifie alors @xmath100 et @xmath116 si @xmath95 ( ou @xmath52 ) et @xmath117 , et commute laction de @xmath86 si @xmath101 est un @xmath5-module .
cest cet oprateur qui permet de faire le lien entre les reprsentations @xmath39-linaires , les reprsentations de @xmath0 , et la cohomologie diwasawa , do nt nous rappelons la dfinition ci - dessous .
pour allger les notations , on pose @xmath118 si @xmath119 .
[ defiiw ] si @xmath7 est une reprsentation @xmath39-linaire , si @xmath46 est un @xmath40-rseau @xmath3-stable de @xmath7 , et @xmath120 , alors on dfinit : @xmath121 le groupe @xmath122 ne dpend alors pas du choix dun @xmath40-rseau @xmath3-stable @xmath46 de @xmath7 .
les @xmath123$]-modules @xmath124 ont t tudis en dtail par perrin - riou , qui a notamment montr ( cf .
@xcite ) : [ bpriw ] si @xmath7 est une reprsentation @xmath1-adique de @xmath3 , alors on a @xmath125 si @xmath126 . de plus : * le sous - module de @xmath123$]-torsion de @xmath127 est isomorphe @xmath128 et @xmath129 est un @xmath130$]-module libre de rang @xmath131 ; * @xmath132 .
les espaces @xmath133 et @xmath134 sont eux aussi des @xmath135$]-modules et on a le rsultat suivant ( cf .
@xcite ) : [ cciw ] on a des isomorphismes canoniques de @xmath135$]-modules @xmath136 et @xmath137 .
en dautres termes , la cohomologie du complexe @xmath138 calcule la cohomologie diwasawa de @xmath7 .
[ dpsinozero ] si @xmath7 est une reprsentation @xmath39-linaire , alors @xmath139 .
remarquons que le corollaire i.7.6 de @xcite affirme que @xmath133 contient en fait une base de @xmath24 sur @xmath52 .
nous donnerons dans le corollaire [ iwnonzero2 ] une autre dmonstration du corollaire [ dpsinozero ] , qui ne passe pas par la cohomologie diwasawa ( mais qui suppose que @xmath7 est cristalline ) .
nous allons maintenant nous intresser la topologie de @xmath101 .
si @xmath101 est un @xmath51-module libre de rang @xmath140 , alors le choix dune base
de @xmath101 donne un isomorphisme @xmath141 et on peut munir @xmath101 de la topologie faible .
un petit calcul montre quune application @xmath51-linaire @xmath142 est ncessairement continue et donc que la topologie dfinie sur @xmath101 par @xmath141 ne dpend pas du choix dune base de @xmath101 .
[ bornetreillis ] si @xmath143 est une partie dun @xmath51-module libre @xmath101 , et @xmath144 est le @xmath69$]-module engendr par @xmath143 , alors les proprits suivantes sont quivalentes : * @xmath143 est borne pour la topologie faible ; * @xmath144 est born pour la topologie faible ; * pour tout @xmath145 , limage de @xmath144 dans @xmath146 est un @xmath69$]-module de type fini .
choisissons une base de @xmath101 , et notons @xmath147 le @xmath69$]-module engendr par cette base , ce qui fait que la topologie faible sur @xmath101 est dfinie par le fait que les @xmath148 forment une base de voisinages de zro .
en particulier , une partie @xmath143 de @xmath101 est borne si et seulement si pour tout @xmath149 , il existe @xmath150 tel que @xmath151 .
comme le @xmath69$]-module engendr par @xmath152 est @xmath153 lui - mme , on voit que les proprits ( i ) et ( ii ) sont quivalentes .
il reste donc montrer que les @xmath69$]-modules borns sont ceux qui satisfont ( iii ) , cest - - dire quun @xmath69$]-module @xmath154 est born si et seulement si pour tout @xmath145 , limage de @xmath154 dans @xmath146 est un @xmath69$]-module de type fini .
si @xmath154 est born alors par dfinition , pour tout @xmath155 , il existe @xmath156 tel que @xmath157 ce qui fait que limage de @xmath154 dans @xmath146 est contenue dans celle de @xmath158 et est de type fini puisque @xmath69 $ ] est un anneau noetherien .
rciproquement , si @xmath154 satisfait ( iii ) , alors pour tout @xmath145 limage de @xmath154 dans @xmath146 est un @xmath69$]-module de type fini et est donc contenue dans @xmath159 pour un @xmath160 , ce qui fait que @xmath161 et donc que @xmath154 est born pour la topologie faible
. un treillis de @xmath101 est un @xmath69$]-module born @xmath162 tel que limage de @xmath154 dans @xmath163 en est un @xmath164$]-rseau .
un treillis dun @xmath52-module @xmath101 est un treillis dun @xmath51-rseau de @xmath101 .
les treillis font lobjet dune tude dtaille dans @xcite .
rappelons le rsultat de base sur les treillis stables par @xmath25 dun @xmath12-module @xmath101 ( cest la proposition 4.29 de @xcite ) : [ col429 ] si @xmath101 est un @xmath12-module tale sur @xmath51 , il existe un unique treillis @xmath165 de @xmath101 vrifiant les proprits suivantes : * quels que soient @xmath112 et @xmath166 , il existe @xmath167 tel que @xmath168 si @xmath169 ; * loprateur @xmath25 induit une surjection de @xmath165 sur lui - mme .
de plus : * si @xmath170 est un sous-@xmath69 $ ] module born de @xmath101 et @xmath166 , il existe @xmath171 tel que @xmath172 si @xmath173 ; * si @xmath170 est un treillis de @xmath101 stable par @xmath25 tel que @xmath25 induise une surjection de @xmath170 sur lui - mme , alors @xmath174 et @xmath175 est annul par @xmath70 .
la fin de ce paragraphe est consacr au dbut de ltude des limites projectives dfinies ci - dessous .
[ defilimproj ] * on note @xmath176 le @xmath39-espace vectoriel des suites @xmath177 dlments de @xmath24 telles que lensemble @xmath178 est born ( do le b ) pour la topologie faible et telles que @xmath179 . *
si @xmath46 est un @xmath40-rseau de @xmath7 stable par @xmath3 , on note @xmath180 le @xmath40-module des suites @xmath25-compatibles dlments de @xmath181 , muni de la topologie de la limite projective .
[ ddiese ] linjection @xmath182 induit un isomorphisme topologique @xmath183 .
il est clair que lapplication ci - dessus est injective .
si @xmath184 , alors par dfinition lensemble @xmath185 est born pour la topologie faible et par le lemme [ bornetreillis ] , le @xmath69$]-module @xmath144 engendr par @xmath143 est born ( pour la topologie faible ) .
quitte multiplier @xmath38 par une puissance de @xmath42 , on peut dailleurs supposer que @xmath186 pour tout @xmath187 .
si @xmath188 , alors la proposition [ col429 ] applique @xmath189 montre que @xmath190 si @xmath8 est assez grand .
comme cest vrai pour tout @xmath191 , on en dduit que @xmath192 pour tout @xmath193 et donc que lapplication @xmath194 est une bijection .
cest un homomorphisme car la topologie de @xmath181 est la topologie induite par la topologie faible de @xmath195 via linclusion @xmath196 .
rappelons que @xmath197 est surjective , ce qui fait que les applications de transition dans @xmath180 le sont .
le lemme ci - dessous et son corollaire seront utiliss dans le paragraphe [ laouiw ] . [ psietddiese ] lapplication naturelle @xmath198 est un isomorphisme .
cette application est videmment surjective , et nous allons montrer quelle est injective , cest - - dire que si @xmath199 , avec @xmath200 , alors @xmath201 .
soit @xmath202 tel que @xmath203 .
pour tout @xmath187 , il existe @xmath204 tel que @xmath205 et on a alors @xmath206 .
loprateur @xmath207 est bijectif sur @xmath208 ( un inverse tant donn par @xmath209 ) et il existe donc @xmath210 tel que @xmath211 .
pour tout @xmath212 , soit @xmath213 tel que @xmath214 .
@xmath180 est un espace topologique compact , la suite @xmath215 a une valeur dadhrence @xmath216 et comme @xmath217 quand @xmath218 ( par la continuit de @xmath25 , voir la proposition [ actgcont ] ci - dessous ) , on voit que @xmath219 et donc que @xmath201 .
[ h2ddiese ] on a un isomorphisme de @xmath220$]-modules @xmath221 .
la proposition 4.43 de @xcite affirme que @xmath222 , et la remarque ii.3.2 de @xcite nous dit que @xmath223 ce qui fait que lon a une succession disomorphismes : @xmath224
le but de cette partie est de dcrire explicitement @xmath50 en fonction de @xmath13 lorsque @xmath7 est une reprsentation @xmath39-linaire cristalline de dimension @xmath2 de @xmath3 ( et absolument irrductible ) .
un nonc similaire est dj paru dans @xcite , mais sans preuve .
dans le dernier paragaphe de ce chapitre , nous dfinissons une action dun sous - groupe du borel @xmath47 sur @xmath225 et nous montrons quelle est continue , topologiquement irrductible , et que le lemme de schur est vrai pour cette action .
le but de ce paragraphe est de rappeler la classification des reprsentations cristallines @xmath12-semi - simples @xmath7 de dimension @xmath2 de @xmath3 en fonction de leur @xmath12-module filtr @xmath13 .
afin de classifier certaines reprsentations @xmath39-linaires du groupe @xmath3 , fontaine a introduit les anneaux @xmath226 et @xmath227 .
ces anneaux vrifient les proprits suivantes : * lanneau @xmath226 est une @xmath4-algbre munie dune action de @xmath3 , telle que @xmath228 et dun frobenius @xmath12 qui commute laction de @xmath3 ; * le corps @xmath227 est le corps des fractions dun anneau complet de valuation discrte @xmath229 ( do nt le corps rsiduel est @xmath31 ) , et il est donc muni de la filtration dfinie par les puissances de lidal maximal .
il est aussi muni dune action continue de @xmath3 , telle que @xmath230 et la filtration est donc stable sous laction de @xmath3 ; * on a une inclusion naturelle @xmath231 et une suite exacte ( dite fondamentale ) : @xmath232 * il existe @xmath233 tel que @xmath234 et @xmath235 est un gnrateur de lidal maximal de @xmath229 .
le choix dun tel @xmath235 dtermine une application injective @xmath236 telle que @xmath237 , et cette injection commute @xmath3 et @xmath12 ; * si @xmath187 , alors on a une application toujours injective @xmath238 et la filtration de @xmath53 induite par cette application est donne par la filtration ordre dannulation en @xmath239 o @xmath240 est une racine primitive @xmath241-ime de @xmath20 dtermine par le choix de @xmath235 .
tant donne une reprsentation @xmath39-linaire @xmath7 , on pose : @xmath242 en gnral , ces @xmath4-espaces vectoriels sont de dimensions infrieure ou gale @xmath131 et on dit que @xmath7 est cristalline ( resp . de de
rham ) si @xmath243 ( resp .
@xmath244 ) est gale @xmath131 .
comme @xmath231 , une reprsentation cristalline est ncessairement de de rham .
toutes les reprsentations que nous considrons dans cet article sont supposes cristallines .
le frobenius @xmath12 de @xmath226 commute laction de @xmath3 et la filtration de @xmath227 est stable par @xmath3 , ce qui fait que @xmath13 est un @xmath12-module et que @xmath245 est un @xmath4-espace vectoriel filtr .
si @xmath7 est une reprsentation cristalline , alors lapplication naturelle de @xmath13 dans @xmath245 est un isomorphisme et @xmath13 est donc un @xmath12-module filtr .
si @xmath7 est @xmath39-linaire , alors @xmath13 et @xmath245 sont naturellement des @xmath39-espaces vectoriels , et @xmath246 ainsi que la filtration sur @xmath245 sont @xmath39-linaires .
si @xmath7 est une reprsentation de de rham , les poids de hodge - tate de @xmath7 sont par dfinition les entiers @xmath247 tels que @xmath248 , compts avec la multiplicit @xmath249 .
si @xmath101 est un @xmath12-module de dimension @xmath20 , on dfinit @xmath250 o @xmath251 est la matrice de @xmath12 dans une base de @xmath101 et si @xmath101 est un espace vectoriel filtr de dimension @xmath20 , on dfinit @xmath252 comme tant le plus grand @xmath253 tel que @xmath254 .
si @xmath101 est un @xmath12-module de dimension @xmath255 , on dfinit @xmath256 et @xmath257 , ce qui fait que @xmath252 est aussi loppos de la somme des poids de hodge - tate de @xmath7 , compts avec multiplicits .
si @xmath101 est un @xmath12-module filtr , on dit que @xmath101 est admissible si @xmath258 et si @xmath259 pour tout sous-@xmath12-module @xmath260 .
le fait que pour tout @xmath261 , on a @xmath262 permet de montrer que si @xmath7 est une reprsentation cristalline , alors @xmath13 est admissible , et on a alors le thorme de colmez - fontaine ( voir ( * ? ? ?
* thorme a ) et aussi @xcite ) : [ cfthm ] le foncteur @xmath263 est une quivalence de catgories de la catgorie des reprsentations @xmath39-linaires cristallines dans la catgorie des @xmath12-modules filtrs @xmath39-linaires admissibles .
ce thorme nous permet de faire une liste des reprsentations cristallines qui nous intressent dans cet article , qui sont les reprsentations cristallines @xmath39-linaires @xmath12-semi - simples do nt les poids de hodge - tate sont @xmath264 et @xmath265 avec @xmath266 .
soit @xmath7 une telle reprsentation .
quitte remplacer @xmath39 par un corps plus gros , on peut supposer que @xmath39 contient les valeurs propres @xmath267 et @xmath268 de loprateur @xmath269 .
quitte changer @xmath251 et @xmath270 , on peut supposer @xmath271 .
lhypothse de semi - simplicit de cet oprateur implique que lon peut crire @xmath272 o @xmath273 et @xmath274 .
il existe alors une droite @xmath275 de @xmath13 telle que la filtration sur @xmath13 est donne par : @xmath276 isomorphisme prs ( et quitte multiplier @xmath277 ou @xmath278 par un lment de @xmath39 ) cette droite est soit @xmath279 , soit @xmath280 , soit @xmath281 . le fait que @xmath13 est admissible se traduit par le fait que @xmath282 et par ces conditions sur @xmath275 : * si @xmath283 , alors @xmath284 , @xmath285 et @xmath7 est irrductible ; * si @xmath286 , alors @xmath287 et : * * si @xmath288 , @xmath7 est rductible non - scinde ; * * si @xmath289 , @xmath7 est rductible scinde .
[ defdalbet ] si @xmath283 , on note @xmath290 le @xmath12-module filtr admissible construit en ( i ) ci - dessus .
dans ce paragraphe , nous faisons le lien entre la thorie des @xmath5-modules et celle des reprsentations cristallines de @xmath3 .
lide de dpart est que si @xmath7 est une reprsentation @xmath1-adique quelconque , alors son @xmath5-module @xmath24 est un objet assez compliqu , mais dans le cas o @xmath7 est cristalline , il est possible den choisir une base particulirement sympathique .
ce rsultat est lobjet de la thorie des modules de wach ( voir @xcite ) .
rappelons le rsultat principal de la thorie ( cf .
@xcite ) . [ wachmod ]
si @xmath7 est une reprsentation @xmath1-adique de @xmath3 , alors @xmath7 est cristalline poids de hodge - tate dans un intervalle @xmath291 $ ] si et seulement sil existe un sous-@xmath61$]-module libre de rang fini @xmath292 , tel que : * @xmath293 } n(v)$ ] ; * @xmath294 est stable par @xmath12 et @xmath295 est annul par @xmath296 ; * @xmath297 est stable sous laction de @xmath86 et celle - ci est triviale sur @xmath298 . si @xmath46 est un @xmath40-rseau de @xmath7 , alors @xmath299 est un @xmath69$]-module libre de rang fini qui satisfait lanalogue des proprits ( i ) , ( ii ) , ( iii ) ci - dessus .
ltude des modules de wach fait lobjet dune bonne partie de @xcite .
partir de maintenant , on suppose que @xmath7 est une reprsentation cristalline .
lobjet de ce paragraphe est de dmontrer le thorme [ recupwach ] ci - dessous .
avant de lnoncer , nous devons faire un certain nombre de rappels .
il existe ( cf .
@xcite ) une application injective naturelle ( en particulier , @xmath61$]-linaire et commutant @xmath12 , @xmath25 et laction de @xmath86 ) : @xmath300 \otimes_l \dcris(v),\ ] ] application qui donne lieu un isomorphisme ( rappelons que @xmath237 ) : @xmath301 \otimes_{l \otimes \o[[x ] ] } n(v ) \buildrel\sim\over\rightarrow \calr^+[1/t ] \otimes_l \dcris(v).\ ] ] si les poids de hodge - tate de @xmath7 sont positifs , on a en fait : @xmath302 et la question se pose alors de caractriser limage de @xmath297 dans @xmath303 .
cest cette question que rpond le thorme [ recupwach ] .
avant de donner cette caractrisation , rappelons que pour tout @xmath187 , on a une application injective de @xmath39-algbres @xmath304 $ ] , qui est dtermine par le fait quelle envoie @xmath305 sur @xmath306 ( ici @xmath240 est une racine primitive @xmath241-ime de @xmath20 qui dpend du choix de @xmath307 , voir le ( v ) du dbut du paragraphe [ cristallin ] ) .
pour allger les notations , on crit @xmath308 au lieu de @xmath309 .
lanneau @xmath310 $ ] est muni dune filtration par les puissances de lidal engendr par @xmath235 , et cela dfinit une filtration sur @xmath310 \otimes_l \dcris(v)$ ] .
rappelons ( cf .
dfinition [ dordrer ] ) quun lment @xmath311 est dordre @xmath78 si , pour un ( ou de manire quivalente tous les ) @xmath312 , la suite @xmath313 est borne . on pose
alors @xmath314 ( cette norme dpend bien sr du choix de @xmath80 , mais pas la topologie quelle dfinit ) .
on renvoie au lemme [ ordre ] ci - dessous pour une caractrisation quivalente de la proprit tre dordre @xmath78 . on a alors
le rsultat suivant , qui prcise le thorme d de @xcite ( pour simplifier , on suppose que @xmath12 est semi - simple sur @xmath13 ) : [ recupwach ] soit @xmath7 une reprsentation cristalline do nt les poids de hodge - tate sont positifs ou nuls et @xmath315 avec @xmath316 et @xmath317 .
si @xmath318 , alors @xmath319 si et seulement si : * pour tout @xmath320 , @xmath321 est dordre @xmath322 ; * pour tout @xmath323 , @xmath324 \otimes_l \dcris(v))$ ] .
afin de dmontrer ce thorme , nous devons utiliser certains rsultats et notations que nous navons pas rappels dans cet article car ils ne sont pas utiliss ailleurs .
commenons par montrer que si @xmath319 , alors il satisfait les conditions
( i ) et ( ii ) .
on peut supposer que @xmath325 o @xmath46 est un rseau @xmath3-stable de @xmath7 .
nous allons utiliser le fait que @xmath326 est naturellement un sous - ensemble de @xmath327 \otimes_{\z } t$ ] o @xmath328 est un sous - anneau de @xmath329 tel que @xmath330 est bijectif , et qui est born pour la topologie @xmath1-adique .
en particulier , si @xmath325 , alors lensemble @xmath331 est born et si @xmath332 , alors @xmath333 ce qui fait que pour tout @xmath334 , la suite des @xmath335 est borne , et @xmath321 est donc dordre @xmath336 parce que si @xmath337 et @xmath75 , alors @xmath338 .
cela montre le ( i ) .
le ( ii ) rsulte du fait que , comme @xmath339 \otimes_{\z } t$ ] , on a @xmath340 pour tout @xmath323 et donc : @xmath341 \otimes_l \dcris(v ) = { \rm fil}^0 ( l_m [ [ t ] ] \otimes_l \dcris(v)).\ ] ] montrons maintenant que si @xmath342 satisfait ( i ) et ( ii ) , alors @xmath319 . pour cela , remarquons que @xmath38 satisfait les conditions ( 1 ) , ( 2 ) et ( 3 ) du thorme d de @xcite ( la condition ( 1 ) est automatique car @xmath7 est cristalline , la condition ( 2 ) est quivalente ( ii ) et la condition ( 3 ) ( i ) ) . on en dduit que , dans les notations de @xcite , @xmath343 o @xmath344 est un certain sous - ensemble de @xmath24 .
nous allons utiliser le fait que si @xmath7 est cristalline , alors @xmath345 } n(v)$ ] o @xmath346 est le sous - corps de @xmath52 constitu des sries formelles @xmath347 ayant la proprit quil existe @xmath348 tel que @xmath77 converge
sur la couronne @xmath349 .
le fait que @xmath350 et que , comme on la rappel plus haut , @xmath351 \otimes_l \dcris(v ) = \calr^+[1/t ] \otimes_{l \otimes \o[[x ] ] } n(v)$ ] montre que : @xmath352 \cap \cale^{\dagger } ) \otimes_{l \otimes \o[[x ] ] } n(v).\ ] ] les coefficients de @xmath38 dans une base de @xmath297 sont donc des fonctions appartenant @xmath351 \cap \cale^{\dagger}$ ] .
@xmath353 \cap \cale^{\dagger}$ ] , alors il existe @xmath354 tel que @xmath355 ) [ 1/\varphi^{\ell}(x)]$ ] .
en effet , le fait que @xmath356 implique que @xmath357 ne peut avoir quun nombre fini de ples , et donc , puisque lensemble des zros de @xmath358 est la runion de lensemble des zros des @xmath359 , quil existe @xmath354 tel que @xmath360 $ ] .
le fait que @xmath355 ) [ 1/\varphi^{\ell}(x)]$ ] rsulte alors du fait que @xmath361 $ ] .
il existe donc @xmath354 tel que @xmath362 $ ] .
le dernier rsultat que nous allons utiliser sans dmonstration ( mais qui suit des rsultats de @xcite ) est que si @xmath323 , alors @xmath363 contient une base de @xmath364 \otimes_l \dcris(v))$ ] sur @xmath310 $ ] , ce qui fait que la condition ( ii ) du thorme implique que les coefficients de @xmath38 dans une base de @xmath297 ne peuvent pas avoir de ples ailleurs quen zro , et donc que @xmath365 $ ] .
en effet , la filtration induite sur @xmath366 \otimes n(v))$ ] correspond lordre dannulation en @xmath239 , et le fait dtre dans le @xmath367 correspond donc labsence de ple .
on termine en remarquant que les calculs de ( * ? ? ? * et
iii.2 ) montrent que les diviseurs lmentaires de linclusion : @xmath368 } n(v ) \subset \calr^+ \otimes_l \dcris(v)\ ] ] sont premiers @xmath70 et donc que si @xmath342 , et @xmath365 $ ] , alors @xmath319 .
soit @xmath369 lespace de banach des @xmath140-uplets de fonctions holomorphes @xmath370 o @xmath371 est dordre @xmath372 , muni de la norme @xmath373 , et soit @xmath170 le sous - espace de @xmath369 form des fonctions qui satisfont la condition ( ii ) du thorme [ recupwach ] .
remarquons que @xmath297 muni de la topologie @xmath1-adique est un espace de banach ( on peut par exemple dcider que @xmath326 en est la boule unit ) .
nous aurons besoin du lemme suivant : [ istopon ] lespace vectoriel @xmath170 est ferm dans @xmath374 et lapplication naturelle @xmath375 est un isomorphisme topologique despaces de banach .
si @xmath323 , alors @xmath376 \otimes_l \dcris(v))$ ] est de codimension finie dans @xmath310 \otimes_l \dcris(v)$ ] et le noyau de lapplication : @xmath377 \otimes_l \dcris(v ) / { \rm fil}^0 ( l_m [ [ t ] ] \otimes_l \dcris(v))\ ] ] qui @xmath370 associe limage de @xmath378 est donc un sous - espace ferm @xmath379 de @xmath374 ce qui fait que @xmath380 est lui aussi ferm dans @xmath374 . lapplication naturelle @xmath375 est un isomorphisme par le thorme [ recupwach ] .
par le thorme de limage ouverte , pour montrer que @xmath375 est un isomorphisme topologique , il suffit de montrer que lapplication @xmath381 est continue , cest - - dire quelle est borne .
pour cela , il suffit dobserver quune boule unit de @xmath297 ( pour la topologie @xmath1-adique ) est donne par @xmath326 , qui est un @xmath69$]-module de type fini .
le lemme [ istopon ] montre que lapplication @xmath382 est continue pour la topologie forte .
[ continj ] lapplication @xmath383 est continue si lon donne @xmath297 la topologie faible et @xmath53 sa topologie de frchet .
cela rsulte du fait que lapplication @xmath61 \rightarrow \calr^+$ ] est continue , et du fait que @xmath297 est un @xmath61$]-module de type fini .
le but de ce paragraphe est de donner une description explicite du @xmath384$]-module @xmath50 en fonction du @xmath39-espace vectoriel @xmath13 .
tant donns les calculs du paragraphe prcdent , il suffit de faire le lien entre @xmath297 et @xmath50 . on suppose
tout au long de ce paragraphe que @xmath7 est une reprsentation cristalline do nt les poids de hodge - tate sont dans un intervalle @xmath385 $ ] .
la proposition [ noslopezero ] ci - dessous nous dit qualors @xmath386 ( sous une petite condition technique pour la deuxime inclusion ) .
[ wachstabpsi ] si les poids de hodge - tate de @xmath7 sont positifs ou nuls et si @xmath46 est un @xmath40-rseau de @xmath7 stable par @xmath3 , alors les modules de wach @xmath297 et @xmath326 sont stables par @xmath25 .
rappelons que @xmath247 est un entier tel que les poids de hodge - tate de @xmath7 sont dans lintervalle @xmath385 $ ] .
si @xmath387 , alors par le ( ii ) du thorme [ wachmod ] , on a @xmath388 et donc @xmath389 ce qui fait que @xmath390 .
ceci tant vrai pour tout @xmath387 , on a donc @xmath391 et donc @xmath392 .
le fait que @xmath393 en rsulte en inversant @xmath1 . [ noslopezero ] si les poids de hodge - tate de @xmath7 sont positifs ou nuls , alors on a @xmath394 .
si les pentes de @xmath12 sur @xmath13 sont de plus toutes strictement ngatives , alors on a @xmath395 . comme @xmath326 est un @xmath69$]-module qui engendre @xmath195 , cest un treillis de @xmath195 et le lemme [ wachstabpsi ] montre que @xmath326 est en plus stable par @xmath25 .
il suit alors de ( * ? ? ?
* lemme 4.27 , ( iv ) et proposition 4.29 ) que @xmath396 .
par le thorme [ wachmod ] , on a @xmath397 , ce qui fait que @xmath398 .
on sait par ailleurs ( par le lemme [ wachstabpsi ] ) que @xmath392 , et on conclut de ces deux observations que la suite des @xmath399 est une suite croissante de @xmath69$]-modules tous contenus dans @xmath326 , ce qui fait quil existe @xmath400 tel que @xmath401 pour tout @xmath402 , et donc que @xmath403 est un treillis qui contient @xmath404 et sur lequel @xmath25 est surjectif .
le lemme rsulte alors du ( iv ) de la proposition [ col429 ] .
montrons finalement que si les pentes de @xmath12 sur @xmath13 sont toutes @xmath405 , alors @xmath395 .
on vient de voir que @xmath396 .
pour terminer la dmonstration , nous allons montrer que pour tout @xmath166 , il existe @xmath406 tel que si @xmath325 , alors @xmath407 si @xmath408 .
le fait que @xmath409 implique alors que @xmath410 , et comme cest vrai pour tout @xmath149 , cest que @xmath395 .
nous allons donc montrer que @xmath411 sur le @xmath39-espace vectoriel de dimension finie @xmath412 .
pour cela , rappelons que , par ( * ? ? ?
* thorme iii.4.4 ) , on a un isomorphisme de @xmath12-modules @xmath413 .
si @xmath414 , alors on peut crire @xmath415 puisque @xmath416 , ce qui fait que @xmath417 et donc que , via lidentification @xmath418 qui @xmath419 associe @xmath420 , lapplication @xmath421 correspond @xmath422 .
comme on a suppos que toutes les pentes de @xmath12 sont @xmath405 , on a bien @xmath423 sur @xmath424 , ce qui fait que @xmath395 .
[ weakispt ] la topologie faible sur @xmath195 induit une topologie sur @xmath326 qui nest autre que la topologie @xmath425-adique , ce qui donne une description plus agrable de la topologie faible sur @xmath181 .
ceci montre dautre part quune partie de @xmath297 est borne pour la topologie @xmath1-adique si et seulement si elle est borne pour la topologie faible .
[ iwnonzero2 ] les modules @xmath426 et donc @xmath133 sont non nuls ( comparer avec le corollaire [ dpsinozero ] ) .
il suffit de montrer que @xmath427 .
si @xmath38 est non nul dans @xmath326 , alors @xmath428 comme on le vrifie aisment , et la srie @xmath429 converge pour la topologie faible vers un lment @xmath216 non nul dans @xmath326 tel que @xmath430 . on a donc @xmath431 et @xmath427 .
nous en venons maintenant au dernier point de ce paragraphe , la description de @xmath432 en termes de fonctions valeurs dans @xmath13 .
de lapplication @xmath383 , on dduit une application encore injective : @xmath433 on
suppose comme ci - dessus que @xmath12 est semi - simple sur @xmath13 , et de plus que @xmath12 est de pentes strictement ngatives sur @xmath13 .
[ ddiesedcrisproj ] soient @xmath315 avec @xmath316 et @xmath434 , et soient @xmath435 @xmath140 suites dlments de @xmath53 telles que @xmath436 .
alors @xmath437 si et seulement si : * pour tout @xmath438 et tout @xmath320 , la fonction @xmath439 est dordre @xmath440 et @xmath441 est born indpendamment de @xmath8 ; * pour tous @xmath442 , @xmath443 , on a : @xmath444 \otimes_l \dcris(v));\ ] ] * pour tout @xmath445 et tout @xmath320 , @xmath446 .
par le thorme [ recupwach ] , les conditions ( i ) et ( ii ) entranent @xmath447 . de plus , le fait que @xmath441 est born indpendamment de @xmath8 est alors quivalent , par le lemme
[ istopon ] , ce que la suite @xmath448 est borne dans @xmath297 pour la topologie @xmath1-adique , et donc aussi pour la topologie faible ( cf .
remarque [ weakispt ] ) .
enfin , la condition ( iii ) est quivalente ce que @xmath449 ce qui quivaut @xmath450 et donc finalement @xmath451 par la proposition [ ddiese ] .
le but de ce paragraphe est de dfinir une action du groupe : @xmath452 sur @xmath453 et de montrer que cette action est continue et topologiquement irrductible et que le lemme de schur est vrifi . on suppose toujours
que @xmath7 est une reprsentation cristalline absolument irrductible do nt les poids de hodge - tate sont dans un intervalle @xmath385 $ ] .
tout lment de @xmath454 peut scrire : @xmath455 o @xmath456 , @xmath457 et @xmath458 .
si @xmath459 , on tend la fonction @xmath460 @xmath461 en posant @xmath462 pour @xmath463 .
si @xmath457 , on note @xmath464 \in \gamma$ ] lunique lment tel que @xmath465)=a$ ] .
[ actg ] si @xmath459 , on pose pour @xmath6 : @xmath466(v_n ) \\ \left ( \begin{pmatrix } 1 & z \\ 0 & 1 \end{pmatrix } \cdot
v \right)_n & \= \psi^m((1+x)^{p^{n+m } z } v_{n+m}),\ n+m\geq -{\rm val}(z).\end{aligned}\ ] ] on laisse le soin au lecteur de vrifier que les formules ci - dessus dfinissent bien une action du groupe @xmath467 sur @xmath180 .
on dfinit aussi une structure de @xmath69$]-module sur @xmath180 en posant @xmath468 pour @xmath469 .
[ actgcont ] lapplication @xmath470 est continue .
on vrifie quil suffit de montrer que lapplication @xmath471 est continue et que lapplication @xmath472 est continue .
si @xmath473 et @xmath474 sont des espaces topologiques et si pour tout @xmath475 , on se donne une application continue @xmath476 , alors lapplication @xmath477 donne par @xmath478 est continue . en effet , la diagonale @xmath479 de @xmath480 y est ferme ,
et lapplication @xmath478 est la composition des applications : @xmath481 on se ramne donc montrer que si chaque @xmath482 est continue , alors @xmath483 est continue ( pour la topologie produit ) ce qui est laiss en exercice facile au lecteur .
afin de montrer la proposition , il suffit donc de montrer que lapplication @xmath484 est continue , et que lapplication @xmath485 est continue .
commenons par montrer que si @xmath7 est une reprsentation cristalline , alors ( a ) lensemble @xmath486 est une base de voisinages de zro pour la topologie faible et ( b ) lensemble @xmath487 est aussi une base de voisinages de zro pour la topologie faible .
la proposition [ noslopezero ] montre que @xmath488 , ce qui montre le point ( a ) puisque @xmath326 est un @xmath69$]-module libre qui engendre @xmath195 sur @xmath51 .
pour montrer le point ( b ) , et comme @xmath489 , il suffit de montrer ( par exemple ) que pour tout @xmath212 , il existe @xmath490 tel que @xmath491 .
pour cela , remarquons que si @xmath492 est tel que @xmath493 , alors @xmath494 et on peut donc prendre @xmath495 .
le fait que loprateur @xmath496 est continu pour la topologie faible rsulte alors du fait que @xmath497 .
montrons maintenant que lapplication naturelle @xmath498 est continue ( pour la topologie faible ) .
comme @xmath48 agit par multiplication par des lments de @xmath69 $ ] ou par action de @xmath86 , et que la topologie faible de @xmath181 est dfinie par une base de voisinages de @xmath264 qui sont des @xmath69$]-modules stables par @xmath86 , on voit pour chaque @xmath499 laction de @xmath500 sur @xmath181 est continue .
il reste donc montrer que si @xmath501 est un voisinage de zro dans @xmath181 , alors il existe un sous - groupe normal ouvert @xmath502 de @xmath48 tel que @xmath503 pour tous @xmath504 .
cela rsulte des faits suivants : * si @xmath505 et @xmath438 , alors @xmath506 ; * si @xmath507 et @xmath438 , alors @xmath508 , que nous laissons en exercices au lecteur .
la proposition suivante est le lemme de schur pour la reprsentation de @xmath467 que lon vient de dfinir .
[ schur ] toute application continue @xmath467-quivariante : @xmath509 est scalaire , i.e. est la multiplication par un lment de @xmath39 . notons @xmath510 lapplication @xmath511 .
commenons par montrer que si @xmath512 , alors @xmath513 ne dpend que de @xmath514 .
soit @xmath515 lensemble des @xmath516 do nt les @xmath8 premiers termes sont nuls , ce qui fait que pour @xmath517 , @xmath515 est un sous-@xmath69$]-module ferm et stable par @xmath25 et @xmath86 de @xmath180 et que @xmath518 . on en dduit que @xmath519 est
un sous-@xmath69$]-module ferm et stable par @xmath25 et @xmath86 de @xmath181 .
par ( * ? ? ?
* lemme 4.56 ) , et comme @xmath7 est absolument irrductible , il existe une constante @xmath520 telle que @xmath521 .
le fait que @xmath518 et que @xmath522 implique @xmath523 .
enfin , on voit que @xmath524 ( en ce sens que tout voisinage ouvert de @xmath264 dans @xmath180 contient @xmath515 pour @xmath525 ) , ce qui fait que @xmath526 , et donc finalement que @xmath527 pour tout @xmath517 , ce qui veut dire que @xmath528 et donc que @xmath513 ne dpend que de @xmath529 .
pour tout @xmath530 , soit @xmath531 un lment de @xmath180 tel que @xmath532 .
les calculs prcdents montrent que lapplication @xmath533 donne par @xmath534 est bien dfinie , et quelle est @xmath69$]-linaire et commute @xmath25 et laction de @xmath86 . par ( * ? ? ?
* proposition 4.7 ) , elle stend en une application de @xmath5-modules @xmath535 qui est scalaire car @xmath7 est irrductible .
comme @xmath536 car @xmath357 commute @xmath537 , on en dduit que @xmath357 est aussi scalaire . on a galement : [ actgirred ] laction
de @xmath47 sur @xmath180 est topologiquement irrductible .
cest la runion de la dmonstration du corollaire 4.59 et du ( iii ) de la remarque 5.5 de @xcite ( puisque @xmath7 et donc @xmath538 est absolument irrductible ) .
le but de cette partie est de dfinir des espaces de banach @xmath1-adiques @xmath9 munis dune action linaire continue de @xmath0 et den commencer ltude . les reprsentations @xmath9 sont associes aux reprsentations cristallines irrductibles
@xmath12-semi - simples @xmath7 de dimension @xmath2 de @xmath3 .
le but de ce paragraphe est de rappeler sans preuve les dfinitions et noncs ( classiques ) danalyse @xmath1-adique utiliss dans la suite .
soit @xmath39 une extension finie de @xmath4 .
si @xmath540 est une fonction quelconque , on pose pour @xmath8 entier positif ou nul : @xmath541 soit @xmath78 un nombre rel positif ou nul .
[ classecr ] une fonction @xmath540 est de classe @xmath539 si @xmath542 dans @xmath543 quand @xmath544 .
rappelons que @xmath357 est continue si et seulement si @xmath545 tend @xmath1-adiquement vers @xmath264 quand @xmath8 tend vers linfini , de sorte que les fonctions de classe @xmath546 au sens de la dfinition [ classecr ] sont prcisment les fonctions continues sur @xmath547 .
toute fonction de classe @xmath539 est aussi de classe @xmath548 pour @xmath549 et est donc en particulier toujours continue . on note
@xmath550 le @xmath39-espace vectoriel des fonctions @xmath540 de classe @xmath539 .
toute fonction continue @xmath540 scrit ( dveloppement de mahler ) : @xmath551 o @xmath552 et @xmath553 , @xmath554 si @xmath445 .
de plus , @xmath555 concide avec @xmath556 .
on vrifie facilement que lespace @xmath550 est un espace de banach pour la norme @xmath557 .
la terminologie de classe @xmath539 provient du fait que , lorsque @xmath78 est entier strictement positif , les fonctions de @xmath550 sont aussi les fonctions sur @xmath547 qui , moralement , admettent @xmath78 drives avec la @xmath78-ime drive continue , ce qui explique la terminologie .
par exemple , les fonctions localement analytiques sur @xmath547 sont de classe @xmath539 pour tout @xmath558 .
[ amice ] soit @xmath140 un entier tel que @xmath559 .
le sous-@xmath39-espace vectoriel @xmath560 de @xmath550 des fonctions @xmath540 localement polynomiales de degr ( local ) au plus @xmath140 est dense dans @xmath550 . on appelle distribution tempre dordre @xmath78 un lment
de lespace @xmath561 , cest - - dire une forme linaire continue sur lespace de banach des fonctions de classe @xmath539 . on dit parfois
aussi distribution tempre dordre @xmath79 .
nous donnons maintenant deux descriptions des distributions tempres dordre @xmath78 .
par le thorme [ amice ] , linclusion @xmath562 induit lorsque @xmath559 une injection : @xmath563 o @xmath564 est lespace vectoriel dual de @xmath560 .
[ amice2 ] soit @xmath565 et supposons que @xmath559 .
alors @xmath566 si et seulement sil existe une constante @xmath567 telle que , @xmath568 , @xmath569 et @xmath570 : @xmath571 o @xmath572 ( @xmath573 est la fonction caractristique de @xmath574 ) .
notons que le plus petit entier @xmath140 tel que le thorme [ amice2 ] sapplique est la partie entire de @xmath78 .
lorsque @xmath575 est dordre @xmath78 et @xmath559 , on pose : @xmath576 on peut montrer que @xmath577 est une norme sur @xmath561 qui redonne la topologie despace de banach de @xmath561 et qui est quivalente la norme : @xmath578 en particulier , la majoration ( [ borneamice ] ) est quivalente la mme majoration pour tout @xmath579 ( et tout @xmath580 , @xmath6 ) , quitte peut - tre modifier @xmath581 .
besoin du lemme suivant : [ lemmech ] soit @xmath558 et @xmath140 la partie entire de @xmath78 .
soit @xmath6 et : @xmath582 o @xmath583 .
alors : @xmath584 par ( [ normedistrd ] ) , on voit que le rel de gauche est plus petit que celui de droite .
pour @xmath585 , il nest pas difficile de construire une forme linaire @xmath586 support dans @xmath574 satisfaisant ( [ borneamice ] ) telle que @xmath587 si @xmath588 ( @xmath589 ) , @xmath590 et @xmath591 ( les dtails sont laisss en exercice au lecteur ) .
en particulier , @xmath592 est infrieur au rel de gauche .
comme cela est vrai pour tout @xmath585 , on en dduit le rsultat .
lespace vectoriel des fonctions localement analytiques sur @xmath547 , @xmath593 ( muni de sa topologie despace de frchet , cf .
@xcite ) , est _ a fortiori _ dense dans @xmath550 et on dispose donc aussi dune injection continue entre duaux continus : @xmath594 la transforme damice - mahler : @xmath595 induit un isomorphisme topologique entre @xmath596 et @xmath597 .
rappelons que ( voir [ series ] ) : @xmath598 et que cet espace est muni de la topologie despace de frchet induite par la collection des normes @xmath599 pour @xmath600 .
rappelons galement ( dfinition [ dordrer ] ) quun lment @xmath601 est dordre @xmath78 si , pour un ( ou de manire quivalente tous les ) @xmath6020,1[$ ] , la suite @xmath603 est borne dans @xmath604 .
[ ordre ] soit @xmath605 . *
un lment @xmath605 est dordre @xmath78 si et seulement si @xmath606 est born ( dans @xmath604 ) lorsque @xmath8 varie . *
les normes @xmath607 et @xmath608 sont quivalentes pour tout @xmath6020,1[$ ] . voir @xcite .
le rsultat suivant dcoule immdiatement des dfinitions et du lemme [ ordre ] : [ amice3 ] soit @xmath609 .
alors @xmath566 si et seulement llment @xmath610 est dordre @xmath78 dans @xmath597 . on peut montrer que
@xmath611 est une norme sur @xmath561 qui redonne la topologie despace de banach de @xmath561 .
le but de ce paragraphe est de donner une premire dfinition de @xmath9 comme espace fonctionnel .
soit @xmath7 une reprsentation cristalline @xmath12-semi - simple absolument irrductible comme au [ cristallin ] ( rappelons que @xmath7 pour poids de hodge - tate @xmath612 avec ncessairement @xmath613 sinon @xmath7 nest pas absolument irrductible ) .
les valeurs propres @xmath267 et @xmath268 de @xmath12 sur @xmath614 sont alors telles que @xmath615 , @xmath616 , @xmath617 , @xmath618 et @xmath619 ( quitte changer @xmath251 et @xmath270 ) .
soit @xmath620 lespace de banach suivant .
son @xmath39-espace vectoriel sous - jacent est form des fonctions @xmath621 vrifiant les deux conditions : * @xmath622 est une fonction de classe @xmath623 ; * @xmath624 se prolonge sur @xmath547 en une fonction de classe @xmath623 .
comme espace vectoriel , on a donc : @xmath625 o , pour @xmath552 , @xmath626 et @xmath627 .
on munit @xmath620 de la norme : @xmath628 qui en fait un espace de banach en vertu du [ analyse ] . on fait agir @xmath39-linairement @xmath0 gauche sur les fonctions de @xmath620 comme suit : @xmath629 notons que @xmath630 agit par la multiplication par @xmath631 .
[ action2 ] si @xmath632 et @xmath633 , alors @xmath634 et laction de @xmath0 sur @xmath620 se fait par automorphismes continus . on pose @xmath635 , @xmath140 la partie entire de
@xmath78 ( donc @xmath636 ) et on munit lespace de banach @xmath550 de la norme induite par son bidual , cest - - dire : @xmath637 qui redonne la topologie despace de banach de @xmath550 par ( * ? ? ?
* lemme 9.9 ) et les rsultats du [ analyse ] .
lassertion du lemme est triviale si @xmath500 est scalaire .
par la dcomposition de bruhat et le cas scalaire , on est rduit montrer la stabilit et la continuit pour les matrices @xmath500 de la forme @xmath638 , @xmath639 et @xmath640 ( @xmath641 ) . le
premier cas est vident puisquil envoie @xmath642 sur @xmath643 ( cf .
( [ f1f2 ] ) ) .
quitte conjuguer par @xmath638 et multiplier par un scalaire convenable , on peut prendre @xmath644 dans le deuxime et @xmath500 envoie alors @xmath645 sur @xmath646 ( multiplication par des scalaires prs ) .
il suffit donc de montrer que lapplication @xmath647 est bien dfinie
et continue de @xmath550 dans @xmath550 .
par le thorme [ amice ] , il suffit de montrer quil existe @xmath648 tel que , si @xmath649 , alors @xmath650 . en crivant , pour un @xmath6 assez petit et des @xmath651 convenables : @xmath652 un calcul donne : @xmath653 on en dduit grce ( [ normedistrd ] ) : @xmath654 o la dernire ingalit rsulte du lemme [ lemmech ] et du fait que @xmath655 .
passons au dernier cas .
quitte conjuguer @xmath656 @xmath640 par un lment convenable de @xmath657 , on peut supposer @xmath658 et @xmath500 envoie alors @xmath659 sur @xmath660 ( multiplication par des scalaires prs ) .
il suffit donc
de montrer que les applications @xmath661 et @xmath662 sont bien dfinies et continues de @xmath550 dans @xmath550 .
cela se vrifie par un argument analogue au prcdent avec un calcul utilisant ( [ normedistrd ] ) et ( [ normedistr ] ) do nt on laisse les dtails au lecteur .
la reprsentation @xmath620 doit tre pense comme une induite parabolique . sans donner un sens formel ce qui suit , on a un isomorphisme @xmath0-quivariant : @xmath663 o lespace de gauche est celui des fonctions @xmath664 qui sont de classe @xmath623 ( oublions que nous navons pas dfini de telles fonctions dans ce cadre ! ) telles que : @xmath665 avec action de @xmath0 donne par @xmath666 . on passe de @xmath667 une fonction @xmath632 en posant : @xmath668 on dfinit de mme @xmath669 muni dune action de @xmath0 par automorphismes en changeant partout @xmath251 et @xmath270 .
voici des exemples importants de fonctions dans @xmath620 : [ appartient ] pour @xmath670 et @xmath671 , les fonctions @xmath672 et les fonctions : @xmath673 ( prolonges par @xmath264 en @xmath674 ) sont dans @xmath620 . en faisant agir @xmath675 sur @xmath676 , il suffit de traiter les deuximes fonctions .
en faisant agir @xmath677 , il suffit mme par le lemme [ action2 ] de traiter le cas @xmath678 et comme @xmath672 est clairement de classe @xmath623 sur @xmath547 , il suffit de montrer que @xmath679 est de classe @xmath623 sur @xmath547 .
soit @xmath680 la fonction nulle sur @xmath547 et , pour @xmath681 , @xmath682 , posons @xmath683 si @xmath684 , @xmath685 sinon .
la fonction @xmath686 est de classe @xmath623 sur @xmath547 puisquelle est localement polynomiale .
il suffit de montrer que @xmath687 dans @xmath688 quand @xmath689 ( car @xmath690 puisque @xmath688 est complet ) . par ( *
* lemme 9.9 ) , il suffit de vrifier que @xmath687 dans @xmath691 , i.e. que : @xmath692 mais : @xmath693 do : @xmath694 o @xmath695 en utilisant ( [ normedistr ] ) et en se rappelant que @xmath618 .
cela implique @xmath696 do le rsultat puisque @xmath697 . on a un lemme analogue en changeant @xmath251 et @xmath270 . on note
@xmath698 ladhrence dans @xmath620 du sous-@xmath39-espace vectoriel engendr par les fonctions @xmath676 et @xmath699 pour @xmath671 et @xmath700 entier , @xmath670
. de mme , on note @xmath701 ladhrence dans @xmath669 du sous-@xmath39-espace vectoriel engendr par les fonctions @xmath676 et @xmath702 pour @xmath671 et @xmath700 entier , @xmath703 .
notons que , lorsque @xmath704 , @xmath701 est de dimension finie et sidentifie aux polynmes en @xmath216 de degr au plus @xmath705 .
le sous - espace @xmath698 ( resp .
@xmath701 ) est stable par @xmath0 dans @xmath620 ( resp .
@xmath669 . exercice . on pose @xmath706 .
il sagit encore dun @xmath39-espace de banach ( avec la topologie quotient ) muni dune action de @xmath0 par automorphismes continus .
nous allons voir que lapplication @xmath707 est continue , que @xmath9 est unitaire et que lon a un morphisme continu @xmath0-quivariant @xmath708 qui est un isomorphisme lorsque @xmath709 .
le but de ce paragraphe est de donner une description plus intrinsque de @xmath710 pour en dduire certaines proprits de laction de @xmath0 ( continuit , unitarit , entrelacements ) peu videntes sur la dfinition prcdente . on conserve les notations du [ definition1 ] .
soit : @xmath711 la reprsentation de @xmath0 produit tensoriel de la reprsentation algbrique @xmath712 par linduite parabolique lisse @xmath713 ( cest - - dire lespace des fonctions localement constantes @xmath714 vrifiant une galit analogue ( [ equfonc ] ) avec action gauche de @xmath0 par translation droite ) .
on munit @xmath715 de lunique topologie localement convexe ( au sens de @xcite ) telle que les ouverts sont les sous-@xmath40-modules gnrateurs ( sur @xmath39 ) .
la reprsentation @xmath715 est dite localement algbrique ( cf .
lappendice de @xcite ) et nest autre que la reprsentation @xmath716 de lintroduction .
on identifie @xmath712 lespace vectoriel des polynmes @xmath717 de degr au plus @xmath705 coefficients dans @xmath39 munis de laction gauche de @xmath0 : @xmath718 comme en ( [ fonction ] ) , on identifie @xmath719 lespace vectoriel des fonctions @xmath621 localement constantes telles que @xmath720 se prolonge sur @xmath4 en une fonction localement constante avec action gauche de @xmath0 comme en ( [ action ] ) mais sans le facteur @xmath721 . on en dduit une injection @xmath0-quivariante continue : @xmath722 par le thorme [ amice ] , limage de @xmath715 est dense dans @xmath620 .
en particulier , on a une injection continue @xmath723 .
on dfinit de mme @xmath724 et une injection @xmath0-quivariante continue dimage dense @xmath725 .
ces injections induisent des applications @xmath0-quivariantes continues @xmath726 et @xmath727 . si @xmath728 est un sous-@xmath40-module gnrateur dun @xmath39-espace vectoriel @xmath729 , rappelons quon appelle complt de @xmath729 par rapport @xmath728 lespace de banach : @xmath730 on a un morphisme canonique dimage dense @xmath731 qui nest pas injectif en gnral ( si @xmath732 , on a @xmath733 ) .
le dual continu @xmath734 de @xmath735 sidentifie en tant quespace de banach @xmath736 ( avec @xmath737 comme boule unit ) .
si @xmath729 est un espace localement convexe tonnel ( cf .
@xcite ) muni dune action continue dun groupe topologique localement compact @xmath467 telle que @xmath728 est ouvert et stable par @xmath467 , il est facile de vrifier en utilisant le thorme de banach - steinhaus ( cf .
* proposition 6.15 ) ) que @xmath735 et @xmath734 sont des @xmath467-banach unitaires et que la flche canonique @xmath731 est continue .
[ complete ] lapplication @xmath726 induit un isomorphisme topologique @xmath0-quivariant entre @xmath738 et le complt de @xmath715 par rapport un quelconque sous-@xmath40-module gnrateur de @xmath715 stable par @xmath0 et de type fini comme @xmath739$]-module . on a le mme rsultat en remplaant @xmath251 par @xmath270 .
notons que le complt ne dpend pas du choix du sous-@xmath739$]-module de type fini gnrateur de @xmath715 car ces @xmath40-modules sont tous commensurables dans @xmath715 . en utilisant @xmath740 et le fait que @xmath49 est compact , on voit facilement quil suffit de complter par rapport un sous-@xmath741$]-module de type fini gnrateur quelconque , par exemple : @xmath742({\mathbf 1}_{\z}(z)z^j)+\sum_{j=0}^{k-2 } \o[\b]\left(\left(\frac{p\alpha}{\beta}\right)^{\!\!{\rm val}(z)}\!\!\!\!\!\!\!\!{\mathbf 1}_{\q-\z}(z)z^j\right)\subset \pi(\alpha)\ ] ] o @xmath743 est la fonction caractristique de louvert @xmath502 .
le dual du complt cherch est donc isomorphe au banach : @xmath744 en utilisant lintgralit du caractre central , il est quivalent de prendre @xmath745 @xmath746 dans ( [ thedual ] ) .
pour @xmath747 , vue comme fonction sur @xmath4 via ( [ commefonction ] ) , et @xmath502 ouvert de @xmath4 , on crit @xmath748 pour @xmath749 .
donne alors que les conditions sur @xmath575 dans ( [ thedual ] ) sont quivalentes lexistence dune constante @xmath750 indpendante de @xmath575 telle que , pour tout @xmath671 , tout @xmath751 et tout @xmath681 : @xmath752 ( si @xmath656 @xmath753 , poser @xmath754 et @xmath755 ; en fait , on peut prendre @xmath756 ) . de ( [ chaud1 ] ) , on dduit facilement en notant que @xmath757 si @xmath758 et quitte modifier @xmath759 : @xmath760 en dcomposant @xmath761 , puis @xmath762 etc . jusqu arriver @xmath763 avec @xmath463 , on dduit de ( [ chaud3 ] ) que ( [ chaud2 ] ) pour @xmath697 dcoule de ( [ chaud1 ] ) et de ( [ chaud2 ] ) pour @xmath442 ( utiliser ( [ chaud3 ] ) pour les morceaux compacts dans la dcomposition et ( [ chaud2 ] ) avec @xmath764 pour le restant ) .
si @xmath765 , en dcomposant @xmath766 , puis @xmath767 etc .
jusqu arriver @xmath768 avec @xmath769 et @xmath770 , on dduit de ( [ chaud3 ] ) que ( [ chaud2 ] ) pour @xmath765 et @xmath771 dcoule de ( [ chaud1 ] ) et de ( [ chaud2 ] ) pour @xmath678 et @xmath772 ( utiliser ( [ chaud3 ] ) pour les morceaux compacts dans la dcomposition puis dvelopper @xmath773 et utiliser ( [ chaud2 ] ) avec @xmath678 et @xmath774 pour le restant ) .
autrement dit , @xmath775 et @xmath776 sont quivalents : * ( [ chaud1 ] ) ; * ( [ chaud2 ] ) pour @xmath442 ; * ( [ chaud2 ] ) pour @xmath678 et @xmath772 .
par ailleurs , tout @xmath777 scrit @xmath778 o @xmath779 ( si @xmath747 , @xmath780 ) . un calcul facile ( laiss au lecteur ) montre que @xmath781 et @xmath782 vrifient ( [ borneamice ] ) pour @xmath783 et @xmath784 avec @xmath785 ( i.e. @xmath786 par le thorme [ amice2 ] avec leurs normes bornes , i.e. @xmath575 est dans une boule de @xmath787 ) si et seulement si @xmath575 vrifie ( quitte modifier @xmath759 ) : @xmath788 pour tout @xmath789 , tout @xmath790 et tout entier @xmath445 , puis : @xmath791 pour tout @xmath792 , tout @xmath790 et tout entier @xmath793 , et enfin : @xmath794 pour tout @xmath790 et tout entier @xmath772 . en dveloppant @xmath795 dans ( [ ch1 ] ) ,
un calcul montre que , quitte modifier @xmath759 , ( [ ch0 ] ) , ( [ ch1 ] ) et ( [ ch2 ] ) sont quivalents : * ( [ chaud1 ] ) pour @xmath765 ; * ( [ chaud1 ] ) pour @xmath678 et @xmath442 ; * ( [ chaud2 ] ) pour @xmath678 et @xmath772 . si @xmath575 est comme en ( [ thedual ] ) , i.e. si @xmath575 vrifie ( i ) ( iii ) , alors _ a fortiori _
@xmath575 vrifie ( iv ) ( vi ) et donc @xmath796 .
mais on a plus .
en faisant tendre @xmath8 vers @xmath797 dans ( [ chaud1 ] ) lorsque @xmath678 , on voit que ( [ chaud1 ] ) pour @xmath697 et @xmath678 implique que @xmath575 annule les fonctions @xmath798 . en faisant tendre @xmath8 vers @xmath799 dans ( [ chaud2 ] ) , on voit que ( [ chaud2 ] ) pour @xmath697 implique que @xmath575 annule les fonctions @xmath800 .
un examen plus approfondi ( sans difficult mais que nous omettons pour ne pas allonger la preuve ) montre que les conditions ( i ) ( iii ) prcdentes sont en fait _ quivalentes _ aux conditions ( iv ) ( vi ) avec les deux conditions supplmentaires que @xmath575 annule les fonctions @xmath676 pour @xmath697 et les fonctions @xmath801 pour @xmath671 et @xmath697 , cest - - dire les fonctions de @xmath698 .
autrement dit , on obtient que le banach dual du complt cherch est isomorphe dans @xmath802 au sous - espace de banach de @xmath803 form des @xmath575 qui annulent @xmath698 , cest - - dire @xmath804 . en particulier , @xmath804 est un @xmath0-banach unitaire .
comme les banach ne sont pas rflexifs , nous allons devoir faire un passage par les topologies faibles pour dduire lisomorphisme de lnonc .
linjection @xmath805 tant une immersion ferme
@xmath0-quivariante , @xmath738 est aussi un @xmath0-banach unitaire .
cela entrane facilement que lapplication @xmath806 induit un morphisme @xmath0-quivariant continu du complt unitaire ci - dessus de @xmath715
vers @xmath738 , donc un morphisme continu sur les duaux munis de leur topologie faible ( qui sont des modules compacts isognie prs au sens de @xcite ) .
mais on vient de voir que ce morphisme sur les duaux tait bijectif ( et mme un isomorphisme topologique pour les topologies fortes ) . par ( * ? ? ?
* lemme 4.2.2 ) , on en dduit que cest aussi un isomorphisme topologique pour les topologies faibles .
par dualit ( cf .
* thorme 1.2 ) ) , on obtient lisomorphisme topologique @xmath0-quivariant de lnonc .
le cas @xmath270 se traite de mme .
rappelons quil existe , multiplication par un scalaire non nul prs , un unique morphisme non nul @xmath0-quivariant : @xmath807 qui est un isomorphisme lorsque @xmath709 et qui a un noyau et un conoyau de dimension @xmath20 lorsque @xmath808 ( voir @xcite par exemple ) .
en termes de fonctions localement constantes sur @xmath4 , ce morphisme est donn explicitement par : @xmath809 o @xmath810 est la mesure de haar sur @xmath4 ( valeurs dans @xmath811 ) .
comme la thorie des reprsentations lisses est algbrique , il ny a pas de problmes de convergence dans les intgrales ci - dessus car on peut toujours remplacer les sommes infinies aux voisinages de @xmath264 ou de @xmath797 par des expressions algbriques en @xmath812 parfaitement dfinies .
en tensorisant par lapplication identit sur @xmath712 , on en dduit un morphisme non nul @xmath0-quivariant : @xmath813 qui est un isomorphisme lorsque @xmath709 .
[ entrepadique ] les reprsentations @xmath738 et @xmath814 sont des @xmath0-banach unitaires et on a un diagramme commutatif @xmath0-quivariant : @xmath815 o @xmath816 est le morphisme @xmath0-quivariant de ( [ intertw ] ) .
lorsque @xmath709 , les flches @xmath816 et @xmath817 sont des isomorphismes .
cela dcoule du thorme [ complete ] car limage par une flche @xmath0-quivariante dun @xmath739$]-module de type fini est aussi un @xmath739$]-module de type fini .
lorsque @xmath808 ( ce qui implique @xmath818 ) , il est vident que @xmath814 est non nul puisque @xmath701 est dans ce cas une reprsentation de dimension finie ( isomorphe via ( [ algebraic ] ) la reprsentation algbrique @xmath712 une torsion non ramifie prs ) .
dans les autres cas , le thorme [ complete ] ne dmontre en rien que les espaces de banach @xmath738 et @xmath814 sont non nuls .
mais on a : [ existencereseau ] si @xmath709 , le banach @xmath738 ( resp .
@xmath814 ) est non nul si et seulement si @xmath715 ( resp .
@xmath724 ) possde un @xmath40-rseau stable par @xmath0 .
si @xmath808 , le banach @xmath738 est non nul si et seulement si @xmath715 possde un @xmath40-rseau stable par @xmath0 . rappelons quun @xmath40-rseau
est par dfinition un sous-@xmath40-module gnrateur qui ne contient pas de @xmath39-droite .
supposons dabord @xmath709 , de sorte que les reprsentations @xmath715 et @xmath724 sont ( algbriquement ) irrductibles .
si @xmath819 , lapplication canonique @xmath726 est injective car non nulle ( car dimage dense ) et une boule unit de @xmath738 stable par @xmath0 induit un rseau stable par @xmath0 sur @xmath715 .
inversement , supposons que @xmath715 possde un @xmath40-rseau stable par @xmath0 , alors pour tout @xmath357 non nul dans @xmath715 , @xmath739f\subset \pi(\alpha)$ ] est un @xmath40-rseau de @xmath715 de type fini comme @xmath739$]-module . il est gnrateur car @xmath715 est irrductible et il ne contient pas de @xmath40-droite car , multiplication prs par un scalaire , il est contenu dans un @xmath40-rseau stable par @xmath0 de @xmath715 .
lapplication de @xmath715 dans son complt par rapport @xmath739f$ ] , qui est @xmath738 par le thorme [ complete ] , est alors injective et en particulier @xmath819 .
lorsque @xmath808 , ce qui suppose @xmath820 , @xmath715 nest plus irrductible et a un quotient isomorphe @xmath712 .
si @xmath819 , lapplication non nulle @xmath726 reste injective sinon elle induirait une injection non nulle @xmath821 ce qui est impossible car , pour @xmath820 , @xmath712 ne possde pas de @xmath40-rseau stable par @xmath0 .
notons que , lorsque @xmath808 , @xmath724 ne peut possder de @xmath40-rseau stable par @xmath0 car sa sous - reprsentation irrductible @xmath712 nen possde pas .
dans ce cas , lapplication @xmath727 est non injective ( son noyau est prcisment @xmath712 ) . via la proposition
[ existencereseau ] , on peut montrer pour des petites valeurs de @xmath191 ou pour les valeurs de @xmath822 provenant des formes modulaires que les banach @xmath738 et @xmath814 sont non nuls , voir par exemple @xcite , @xcite , @xcite , @xcite . on va voir dans la suite que la non nullit pour tout @xmath191 ( et tout @xmath251 , @xmath270 ) dcoule de la thorie des @xmath5-modules .
une autre approche possible , purement en termes de thorie des reprsentations , est prsente dans @xcite .
le but de cette partie est de dmontrer lexistence dun isomorphisme topologique canonique entre @xmath176 et le dual @xmath22 ( muni de sa topologie faible ) et den dduire que @xmath9 est toujours non nul , topologiquement irrductible et admissible .
ces noncs taient conjecturs ( et des cas particuliers dmontrs ) dans @xcite et @xcite .
le fait remarquable est que ces noncs , entirement du ct @xmath823 , se dmontrent en passant par le ct _ galoisien_. on fixe une fois pour toutes une reprsentation cristalline irrductible @xmath7 comme au [ cristallin ] avec @xmath614 .
[ explicit ] soit @xmath824 , @xmath825 , @xmath826 et @xmath827 , @xmath828 les distributions localement analytiques sur @xmath547 correspondantes par ( [ yvette ] ) .
la condition : @xmath829\otimes_ld(\alpha,\beta))\ ] ] est quivalente aux galits dans @xmath26 : @xmath830 pour tout @xmath751 et toute racine primitive @xmath831-ime @xmath240 de @xmath20 . on a : @xmath832 dans @xmath833 $ ] ( voir [ rcm ] ) .
en posant @xmath834 et @xmath835 ( @xmath836 , @xmath837 ) , la condition sur @xmath367 est quivalente : @xmath838(e_{\alpha}+e_{\beta})\oplus ( { \rm exp}(t / p^m)-1)^{k-1}(l_m[[t]]e_{\alpha } ) \end{gathered}\ ] ] en notant que @xmath839 engendre @xmath840)=\q\overline t$ ] .
on peut supposer @xmath39 aussi grand que lon veut en ( [ filo ] ) , et en particulier contenant @xmath841 . en utilisant : @xmath842 et en dveloppant @xmath843 , un calcul facile montre que la condition
( [ filo ] ) est quivalente aux galits dans @xmath39 : @xmath844 pour tout @xmath751 et _ toute _ racine primitive @xmath241-ime @xmath240 de @xmath20 .
noter que les sries en ( [ combin ] ) convergent bien car @xmath845 .
en se souvenant que @xmath846 et en utilisant le dveloppement de mahler ( [ mahler ] ) : @xmath847 on obtient : @xmath848 ( la srie @xmath849 convergeant vers @xmath850 dans @xmath593 ( cf .
@xcite ) , on peut inverser @xmath851 et @xmath852 ) . on a la mme
galit avec @xmath853 et @xmath828 .
avec ( [ combin ] ) , on dduit le rsultat . via lidentification @xmath854/{\mathbb z}$ ] , on peut dfinir le nombre complexe algbrique @xmath855 pour tout @xmath856 ( par exemple , @xmath857 si @xmath552 ) . en fixant des plongements @xmath858 et @xmath859 , on peut voir @xmath855 comme un lment de @xmath26 . on obtient ainsi
un caractre additif localement constant @xmath860 , @xmath861 trivial sur @xmath547 .
ce que lon fera dans la suite ne dpendra pas du choix de ce caractre , i.e. du choix des plongements .
notons @xmath862 le @xmath39-espace vectoriel des fonctions localement polynomiales support compact @xmath621 de degr ( local ) au plus @xmath140 .
si @xmath575 est une forme linaire sur @xmath862 , @xmath502 un ouvert compact de @xmath4 et @xmath621 une fonction localement polynomiale de degr au plus @xmath140 support quelconque , on note comme dhabitude @xmath863 . * pour tout @xmath751 , tout @xmath865 et tout @xmath866 , on a : @xmath867 * pour tout @xmath747 support compact ( comme fonction sur @xmath4 via ( [ commefonction ] ) ) tel que @xmath868 est aussi support compact ( comme fonction sur @xmath4 via ( [ commefonction ] ) ) , on a : @xmath869 comme @xmath615 et @xmath870 , la constante dans ( [ debutinter ] ) est bien dfinie et non nulle .
soit @xmath871 une fonction localement constante support compact et @xmath872 la transforme de fourier ( usuelle ) de @xmath247 .
rappelons que @xmath872 est aussi une fonction localement constante sur @xmath4 support compact telle que @xmath873 et @xmath874 o @xmath810 , @xmath875 dsignent la mesure de haar sur @xmath4 .
pour @xmath876 , on a @xmath877 et on voit que @xmath878 est support compact dans @xmath4 si et seulement si @xmath879 .
supposons donc @xmath879 et soit @xmath880 tel que @xmath247 et @xmath878 ont leur support dans @xmath881 et tel que @xmath882 .
pour @xmath751 et @xmath883 , on a : @xmath884 en dcomposant @xmath885 , un calcul facile ( laiss au lecteur ) donne pour @xmath866 : @xmath886 do : @xmath887 pour @xmath883 et @xmath888 sinon
. de mme , on a : @xmath889 pour @xmath883 et @xmath890 sinon .
notons que ( [ fourier1 ] ) and ( [ fourier2 ] ) sont en fait des sommes finies sur le mme ensemble ( fini ) de valeurs de @xmath891 .
en remplaant @xmath892 et @xmath893 dans ( [ variant ] ) ci - dessous par les sommes finies ( [ fourier1 ] ) et ( [ fourier2 ] ) , on voit que ( i ) entrane ( ii ) .
rciproquement , supposons que pour tout @xmath751 , tout @xmath894 et toute fonction @xmath247 comme ci - dessus localement constante support dans @xmath881 , on a : @xmath895 soit @xmath865 tel que @xmath866 , @xmath896 et : @xmath897 alors @xmath898 est aussi support compact car @xmath882 et un calcul via ( [ fourier1 ] ) montre que : @xmath899 on peut donc appliquer lgalit ( [ variant ] ) @xmath247 qui est alors exactement lgalit ( [ rajout ] ) multiplie par @xmath900 .
cela montre que ( ii ) entrane ( i ) et achve la preuve .
soit @xmath902 un @xmath40-rseau stable par @xmath3 .
on reprend les notations du [ phigamma ] , en particulier on dispose du @xmath69$]-module de type fini @xmath181 muni de la surjection @xmath903 et de laction semi - linaire de @xmath86 qui commute @xmath25 . on dispose
aussi de lisomorphisme topologique de la proposition [ ddiese ] qui permet de remplacer @xmath904 par @xmath905 et on sait par le corollaire [ ddiesedcrisproj ] que @xmath905 concide avec les suites dlments @xmath906 de @xmath907 telles que : * @xmath908 , @xmath909 ( resp .
@xmath910 ) est dordre @xmath911 ( resp .
@xmath912 ) dans @xmath597 et @xmath913 ( resp .
@xmath914 ) est born indpendamment de @xmath8 ; * @xmath908 et @xmath915 , on a : @xmath916\otimes_l\dcris(v));\ ] ] * @xmath917 , @xmath918 et @xmath919 .
nous allons dabord dfinir une application @xmath39-linaire @xmath920 .
et @xmath922 les distributions sur @xmath547 correspondant @xmath923 et @xmath924 par la transforme damice - mahler ( [ yvette ] ) . on associe
@xmath925 et @xmath926 deux distributions localement analytiques @xmath827 et @xmath828 sur @xmath4 support compact ( i.e. deux formes linaires continues sur lespace vectoriel des fonctions localement analytiques sur @xmath4 support compact ) en posant : @xmath927 ( resp .
avec @xmath270 au lieu de @xmath251 ) o @xmath621 est localement analytique ( support quelconque ) et @xmath502 est un ouvert compact de @xmath4 contenu dans @xmath881 .
si @xmath575 est une distribution localement analytique sur @xmath547 correspondant @xmath601 par ( [ yvette ] ) , il est facile de voir que la distribution localement analytique @xmath929 correspondant @xmath930 vrifie : @xmath931 on a donc : @xmath932 en remarquant que @xmath933 est support dans @xmath934 . [ versfin ] il y a une manire unique de prolonger @xmath828 et @xmath827 comme lments respectivement de @xmath941 et @xmath802 telle que , pour tout @xmath942 ( vue comme fonction sur @xmath4 par ( [ commefonction ] ) ) : @xmath943 un calcul facile partir de ( [ integrale ] ) donne pour @xmath751 : @xmath944 comme @xmath945 et @xmath946 sont bien dfinis , on voit en utilisant ( [ extension ] ) et ( [ calculc ] ) que : @xmath947 est uniquement dtermin , puis en utilisant ( [ calculnc ] ) ( et ( [ extension ] ) ) que : @xmath948 est aussi uniquement dtermin . il est facile de vrifier que cela dfinit bien un prolongement unique de @xmath828 et @xmath827 comme formes linaires sur respectivement @xmath724 et @xmath715 . en particulier , le lemme [ actiongl2 ] induit une action du groupe @xmath47 sur @xmath180 , qui concide bien sr avec celle de la dfinition [ actg ] ( en faisant agir les scalaires par multiplication par le caractre central de @xmath802 ) .
linjectivit dcoule via ( [ prolong ] ) de linjectivit dans lisomorphisme @xmath963 ( cf .
( [ yvette ] ) ) et la @xmath47-quivariance du lemme [ actiongl2 ] .
montrons que lapplication @xmath964 est continue .
notons @xmath965 ( resp .
@xmath966 ) le sous-@xmath39-espace vectoriel des fonctions @xmath747 ( resp .
@xmath942 ) support compact dans @xmath4 . par ( [ calculc ] ) , on voit que la flche @xmath967 est surjective et induit une immersion ferme entre espaces de frchet : @xmath968 il suffit donc de montrer la continuit des deux applications @xmath969 et @xmath970 , ce qui dcoule aprs passage la limite projective via ( [ prolong ] ) de la continuit de @xmath963 et de celle de linjection @xmath971 ( cf .
remarques [ continj ] et [ weakispt ] ) .
notons @xmath972 lespace de banach dual du module compact @xmath180 par lanti - quivalence de catgorie de @xcite .
il est muni dune action continue unitaire de @xmath47 par la proposition [ actgcont ] ( on peut utiliser les arguments de dualit de la preuve de ( * ? ? ?
* proposition 1.6 ) pour la continuit de laction ) et on a par ce qui prcde un morphisme @xmath47-quivariant ( continu ) @xmath973 .
par la proprit universelle du complt de @xmath715 par rapport un sous-@xmath741$]-module gnrateur de type fini et par le thorme [ complete ] , ce morphisme stend par continuit en un morphisme @xmath47-quivariant continu @xmath974 .
en redualisant , ce dernier induit un morphisme continu @xmath975 qui est le morphisme de lnonc .
soit @xmath977 et @xmath978 o @xmath817 est le morphisme @xmath0-quivariant du corollaire [ entrepadique ] . on dfinit une suite @xmath925 de distributions
localement analytiques sur @xmath547 en posant : @xmath979 et on dfinit de mme @xmath926 .
soit @xmath980 les lments correspondant @xmath981 par ( [ yvette ] ) .
la condition ( iii ) est vidente partir de ( [ psidis ] ) et la condition ( ii ) dcoule des dfinitions , du lemme [ crucial ] et du corollaire [ equival ] .
vrifions la condition ( i ) .
revenant la preuve du thorme [ complete ] , on a en particulier que @xmath827 satisfait ( [ chaud1 ] ) ce qui entrane : @xmath982 pour tout @xmath580 , tout @xmath983 et tout @xmath6 .
avec les notations du [ analyse ] , cela entrane pour tout @xmath894 : @xmath984 pour une constante @xmath985 . on a une borne analogue pour les @xmath986 . on en dduit ( i ) .
par le lemme [ routine ] et le corollaire [ ddiesedcrisproj ] , on a une application @xmath39-linaire : @xmath987 et il est immdiat partir des dfinitions et du lemme [ versfin ] de vrifier quelle est inverse de celle du lemme [ unsens ] .
[ phigammabanach ] il y a un unique isomorphisme topologique ( multiplication prs par un scalaire non nul ) entre les @xmath39-espaces vectoriels localement convexes ( pour la topologie faible des deux cts ) : @xmath988 tel que laction de @xmath989 sur @xmath22 correspond @xmath990 , laction de @xmath991 celle de @xmath86 et laction de @xmath992 @xmath993 .
lexistence dun tel isomorphisme dcoule des rsultats prcdents , sachant quune application bijective continue entre deux modules compacts isognie prs est un isomorphisme topologique ( cest la version duale par @xcite du thorme de limage ouverte entre espaces de banach ) .
il reste dmontrer lunicit ( scalaire prs ) mais cela rsulte de la proposition [ schur ] .
[ iwasawa2 ] le @xmath995$]-module @xmath998 sidentifie aussi aux coinvariants de @xmath22 sous laction de @xmath989 et en fait , ces deux espaces sont nuls par le ( ii ) de la proposition [ bpriw ] parce que @xmath7 est irrductible .
en effet , le thorme [ phigammabanach ] et le corollaire [ h2ddiese ] nous disent que les coinvariants dun rseau de @xmath22 sous laction de @xmath989 sidentifient @xmath999 .
on ignore si le @xmath40-module compact @xmath180 est stable par @xmath49 dans @xmath22 ( via le thorme [ phigammabanach ] ) mais on peut le remplacer par le @xmath40-rseau de @xmath22 : @xmath1004 qui est un sous-@xmath69$]-module compact stable par @xmath0 dans @xmath22 ( on vrifie quil est stable par @xmath1005 en utilisant la dcomposition diwasawa de @xmath0 ) .
le @xmath40-module @xmath1006 possde alors _ deux _ structures naturelles de @xmath69$]-modules : lune est celle dj dfinie
et lautre est : @xmath1007\times { \mathcal m}\mapsto \begin{pmatrix}0&1\\1&0\end{pmatrix}\lambda\begin{pmatrix}0&1\\1&0\end{pmatrix}v.\ ] ] la premire structure est telle que la multiplication par @xmath1008 correspond laction de @xmath992 et la deuxime est telle que la multiplication par @xmath1008 correspond laction de @xmath1009 .
soit @xmath1010 la projection sur la premire composante et @xmath1011 : @xmath154 est un sous-@xmath69$]-module ( de type fini ) de @xmath181 .
posons @xmath1012 .
lapplication : @xmath1013 est injective : si @xmath1014 a pour image @xmath264 , sa distribution associe @xmath1015 par ( [ versbanach ] ) et ( [ f1f2 ] ) est nulle sur les deux copies de @xmath688 , donc est nulle dans @xmath22 .
en pensant encore en termes de distributions , on voit que @xmath1016 est un @xmath69$]-module pour la premire structure mais seulement un @xmath1017)$]-module pour la deuxime structure .
de plus , pour cette deuxime structure , linjection ( [ tordu ] ) est @xmath1017)$]-linaire .
comme @xmath154 est de type fini sur @xmath69 $ ] , donc sur @xmath1017)$ ] , on obtient que le @xmath1017)$]-module @xmath1016 pour la deuxime action de @xmath1017)$ ] est de type fini .
fixons maintenant des lments @xmath1018 ( resp .
@xmath1019 ) tels que les @xmath1020 ( resp .
les @xmath482 ) engendrent @xmath154 sur @xmath69 $ ] ( resp .
@xmath1021 sur @xmath1017)$ ] ) .
soit @xmath1022 .
il existe @xmath1023 dans @xmath69 $ ] tels que @xmath1024 et il existe @xmath1025 dans @xmath1017)$ ] tels que @xmath1026 .
comme les @xmath1027 correspondent laction dlments de lalgbre de groupe de @xmath992 et les @xmath1028 laction dlments de lalgbre de groupe de @xmath1029 , on voit que @xmath1030 est _ a fortiori _ de type fini sur lalgbre de groupe de @xmath49 , do ladmissibilit .
les corollaires [ irreductible ] et [ admissible ] taient conjecturs et dmontrs par un argument de rduction modulo @xmath1 pour @xmath1000 ( et @xmath1002 si @xmath1003 ) dans @xcite avec lhypothse supplmentaire @xmath1031 pour le premier . [ reseaurigolo ] supposons @xmath709 , alors @xmath715 ( resp .
@xmath724 ) possde des @xmath40-rseaux stables par @xmath0 et tous les @xmath40-rseaux stables par @xmath0 dans @xmath715 ( resp .
@xmath724 ) sont commensurables entre eux .
supposons @xmath1032 , alors on a le mme rsultat pour @xmath715 .
lexistence de tels @xmath40-rseaux rsulte du corollaire [ cestnonnul ] et de la proposition [ existencereseau ] .
pour montrer quils sont tous commensurables entre eux , il est quivalent de montrer quils sont tous commensurables aux @xmath40-rseaux de type fini sur @xmath739 $ ] .
le @xmath40-dual dun @xmath40-rseau stable par @xmath0 est toujours contenu dans le @xmath40-dual dun @xmath40-rseau de type fini sur @xmath739 $ ] .
par le thorme [ complete ] et le corollaire [ admissible ] , ce dernier dual est de type fini sur lalgbre de groupe complte de @xmath49 .
comme cest une algbre noethrienne , il en est de mme du premier dual .
cela entrane que le complt de @xmath715 ( ou @xmath724 si @xmath709 ) par rapport un @xmath40-rseau stable par @xmath0 quelconque est aussi admissible , et donc topologiquement isomorphe @xmath9 par le corollaire [ irreductible ] et le fait que la catgorie des @xmath49-banach admissibles est ablienne ( @xcite ) .
tous les @xmath40-rseaux stables par @xmath0 induisent donc des normes quivalentes sur @xmath715 ( ou @xmath724 si @xmath709 ) ce qui achve la preuve .
comme dans @xcite , on note @xmath1033 le sous-@xmath39-espace vectoriel de @xmath9 des vecteurs localement analytiques , i.e. des vecteurs @xmath1034 tels que lapplication orbite @xmath1035 , @xmath1036 est localement analytique .
il est muni dune topologie naturelle despace localement convexe de type compact ( cf .
@xcite ) .
soit : @xmath1037 linduite parabolique localement analytique
au sens de @xcite .
on dfinit de mme @xmath1038 en changeant @xmath251 et @xmath270 . on a des injections naturelles continues @xmath0-quivariantes @xmath1039 et @xmath1040 .
par @xcite , @xmath715 ( resp .
@xmath724 ) est le seul sous - objet topologiquement irrductible non nul dans @xmath1042 ( resp .
@xmath1043 ) .
par le thorme [ complete ] , on dduit que les injections ci - dessus induisent encore des injections @xmath1044 et @xmath1045 .
le rsultat dcoule alors du corollaire [ entrepadique ] .
berger l. , breuil c. , _ towards a @xmath1-adic langlands programme _ , notes dun cours donn
lcole dt de hangzhou ( aot 2004 ) , disponibles ladresse : ` www.ihes.fr/{}^{\sim}breuil/publications.html ` | dans @xcite , des reprsentations unitaires de @xmath0 sur des espaces de banach @xmath1-adiques sont associes aux reprsentations cristallines irrductibles de dimension @xmath2 de @xmath3 .
certaines conjectures sur ces banach y sont formules ( non nullit , irrductibilit topologique , admissibilit ) .
nous dmontrons ces conjectures en rinterprtant ces espaces de banach comme espaces de fonctions sur @xmath4 dun certain type , puis en utilisant la thorie des @xmath5-modules associs aux reprsentations cristallines correspondantes . in @xcite ,
some unitary representations of @xmath0 on @xmath1-adic banach spaces are associated to @xmath2-dimensional irreducible crystalline representations of @xmath3 .
some conjectures are formulated concerning those banach spaces ( non triviality , topological irreducibility , admissibility ) .
we prove those conjectures by reinterpreting those banach as spaces of functions of a certain type on @xmath4 , and then by using the theory of @xmath5-modules associated to the corresponding crystalline representations . |
the large - scale structure of the universe is one of the pillars of our modern understanding of cosmology , encoding information about the contents and evolution of the universe , the physics of the growth of density fluctuations with time , and the formation and evolution of galaxies within the underlying network of dark matter haloes . in particular , the large - scale clustering pattern of galaxies is sensitive to the properties of the cosmic dark energy component which is currently poorly understood .
dark energy influences both the rate of growth of structure and the geometrical distance - redshift relations .
one of the cleanest probes of dark energy is to delineate as a function of redshift the apparent tangential and radial size of the baryon acoustic oscillation scale , a known `` standard ruler '' preferred separation imprinted into the galaxy distribution ( cooray et al .
2001 ; eisenstein 2002 ; blake & glazebrook 2003 ; seo & eisenstein 2003 ; hu & haiman 2003 ; linder 2003 ; glazebrook & blake 2005 ) .
this cosmological probe is helping to motivate a new generation of massive spectroscopic galaxy surveys .
cosmic structure has been mapped out by a succession of galaxy redshift surveys of increasing size and depth . the local universe ( redshifts @xmath16 )
has been studied in exquisite detail by the 2-degree galaxy redshift survey ( 2dfgrs ; colless et al .
2001 ) and the sloan digital sky survey ( sdss ; york et al .
the sdss luminous red galaxy component extended this programme to a mean redshift @xmath17 using a specific type of tracer galaxy ( eisenstein et al .
indeed , the cosmological conclusions reached should be independent of the galaxy type used , given that the `` bias '' with which galaxies trace the underlying dark matter fluctuations is expected to be a simple linear function on large scales ( coles 1993 ; scherrer & weinberg 1998 ) . in this sense
, the choice of the `` tracer population '' of galaxies can be determined by observational considerations such as telescope exposure times , the availability of input imaging data for target selection , and secondary science goals .
the wigglez dark energy survey , using the aaomega multi - object spectrograph at the 3.9 m anglo - australian telescope ( aat ) , is designed as the next leap forwards in redshift coverage , targetting the range @xmath18 .
the survey is scheduled to cover a sky area of 1000 deg@xmath0 , mapping a cosmic volume @xmath19 gpc@xmath8 sufficient to measure the imprint of baryon oscillations in the clustering pattern at a significantly higher redshift than has been previously achieved by 2dfgrs ( cole et al .
2005 ; percival et al . 2007 ) and sdss ( eisenstein et al , 2005 ; huetsi 2006 ; percival et al . 2007 ; gaztanaga et al
the survey redshift range is motivated by the optimal redshift location for testing a cosmological constant model in a spatially - flat universe ( parkinson et al .
2007 ) , which is the sensible initial hypothesis to reject in the dark energy parameter space .
the target galaxy population is bright emission - line galaxies selected from uv imaging by the galaxy evolution explorer ( galex ) satellite ( martin et al . 2005 ) .
this choice is motivated by the short ( 1-hr ) exposure times required to obtain redshifts at the aat .
the survey commenced in august 2006 and is scheduled to finish in july 2010 , using the equivalent of 165 clear nights of telescope time ( 220 awarded nights ) .
secondary science goals involve the study of star formation and galaxy evolution as a function of redshift and environment .
in this initial study we focus on the small - scale clustering properties of the first @xmath2 of the wigglez sample .
the clustering strength is an important parameter in the survey design and cosmological parameter forecasts : the signal - to - noise with which we can recover the galaxy power spectrum depends on the bias of the galaxies with respect to the dark matter fluctuations , which affects the balance between sample variance and shot noise in the power spectrum error budget .
these initial clustering measurements allow us to determine the bias parameter and complete the survey forecast . furthermore , the joint uv - optical selection in the redshift interval @xmath20 places the wigglez survey in an interesting location in the parameter space of galaxy evolution . in this context
, the clustering strength of a set of galaxies provides a direct indication of the density of the environment or ( equivalently ) the typical mass of the dark matter haloes hosting the galaxies .
the clustering strength of uv - selected samples has been studied at low redshift @xmath21 ( milliard et al . 2007 ; heinis et al .
2007 ) and the corresponding rest - frame samples have been selected at much higher redshift @xmath22 through studies of the clustering of lyman break galaxies ( lbgs ; e.g. giavalisco & dickinson 2001 ; ouchi et al . 2001 ; arnouts et al . 2002 ; foucaud et al . 2003 ; adelberger et al .
2005 ; allen et al .
2005 ; ouchi et al . 2005 ;
lee et al . 2006 ; yoshida et al . 2008 ) .
the wigglez survey samples a redshift range which is intermediate to these previous studies .
moreover , the clustering strength of optically - selected star - forming galaxies at high redshift has been studied over small areas by the deep extragalactic evolutionary probe ( deep2 ) project ( coil et al .
2008 ) and the vimos vlt deep survey ( vvds ; meneux et al . 2006 ) .
wigglez is mapping an area @xmath23 times larger , and is therefore able to measure accurately the clustering strength of the most luminous star - forming galaxies , for which these smaller surveys are limited by small - number statistics and sample variance .
the backdrop to these studies is the recent concept of `` down - sizing '' ( cowie et al .
1996 ; glazebrook et al . 2004 ; van dokkum et al .
2004 ) whereby the stars in more massive galaxies appear to have formed earlier , and the typical mass of the most actively star - forming galaxies is expected to decrease with time .
a recent study of lbgs ( yoshida et al .
2008 ) has emphasized the importance of studying the clustering segregration with both uv and optical luminosities , which crudely trace ongoing star formation rate and stellar mass , respectively .
wigglez is well - suited for undertaking such studies over the redshift range @xmath18 .
the plan of this paper is as follows : in section [ secdata ] we introduce the wigglez survey strategy and target selection and describe the data sample used in this study . in section [ secanalysis ]
we describe the methodology used to produce the small - scale clustering measurement including the generation of random ( unclustered ) realizations of the dataset with the correct selection functions and redshift completeness map .
we also explain how we derive the statistical error in the clustering measurement .
we present the clustering results in section [ secresults ] ( split by redshift , absolute magnitude and rest - frame colour ) together with initial comparisons to other studies .
section [ secforecast ] contains the cosmological parameter forecasts for the full wigglez survey , and section [ secconc ] summarizes our conclusions . when converting redshifts to co - moving co - ordinates we assume a spatially - flat universe with cosmological parameters @xmath24 and @xmath25 .
the design and implementation of the wigglez dark energy survey will be fully described in a forthcoming `` survey paper '' ( drinkwater et al .
2009 , in preparation ) which will accompany our mid - term data release .
we include a brief outline here both for ease of reference and to emphasize the key points relevant to the small - scale clustering analysis .
the wigglez survey strategy is to harvest low signal - to - noise spectra of a large number of uv - selected emission - line galaxies in relatively short exposure times ( 1-hr integrations at the aat ) .
the survey tolerates a relatively low redshift completeness of @xmath26 but generates a large statistical sample of galaxy redshifts .
the survey goal is to cover @xmath27 deg@xmath0 of the equatorial sky , gathering @xmath28 spectra of which @xmath29 are expected to yield successful redshifts .
the survey was designed such that the average galaxy number density @xmath30 is related to the amplitude of the galaxy clustering power spectrum @xmath31 on the relevant baryon oscillation scales by @xmath32 , implying that the contributions of sample variance and shot noise to the clustering error are equal .
this is the optimal survey strategy for fixed number of galaxies . the wigglez survey area , illustrated in figure [ figsurvey ] , is split into seven equatorial regions to facilitate year - round observing .
we require that each region should possess a minimum angular dimension of @xmath33 deg , corresponding to a spatial co - moving scale that exceeds by at least a factor of two the standard ruler preferred scale [ which projects to ( 8.5 , 4.6 , 3.2 , 2.6 ) deg at @xmath34 .
the survey coverage within individual regions should also be highly ( @xmath35 ) contiguous , otherwise the significance of the detection of the acoustic features is degraded by convolution with the survey window function .
the survey duration is forecast to be @xmath36 clear nights between august 2006 and july 2010 , using the multi - object capability of the 2df positioner system coupled to the aaomega spectrographs ( saunders et al .
2004 ; sharp et al .
2006 ) .
galaxy redshifts are obtained from the bright emission lines associated with star - forming galaxies , in particular redshifted [ oii ] 3727 , h@xmath37 4861 and [ oiii ] 4959 , 5007 .
low - resolution ( 5 fwhm ) spectra are obtained spanning the ( observed - frame ) wavelength range @xmath38 , hence the majority of successful redshifts in the range @xmath39 are confirmed by multiple emission lines .
single - line redshifts are almost invariably [ oii ] , for which we usually resolve the doublet in the range @xmath40 , increasing our confidence in the line identification .
redshifts are obtained by visual inspection of each spectrum using the interactive software tool `` runz '' , and are classified by a quality flag @xmath41 , where the range @xmath42 denotes a `` reliable '' redshift ( see colless et al .
the fraction of stellar contamination is very small ( @xmath43 ) and we find a similarly low fraction of high - redshift quasar interlopers . the galaxy continuum is typically detected with low signal - to - noise ( an average of @xmath44 per resolution element ) .
wigglez targets are chosen by a joint uv - optical selection .
the primary selection dataset is the medium imaging survey undertaken by the galaxy evolution explorer ( galex ) uv satellite , which provides typical exposure times of 1500 seconds in two filter bands , @xmath45 ( @xmath46 ) and @xmath47 ( @xmath48 ) .
the galex point - spread function is too broad to allow for accurate placement of the spectrograph optical fibres , therefore the uv imaging is cross - matched with optical data . for our ngp regions ,
the sdss imaging data are used . for our sgp regions ,
the sdss @xmath49 stripes are too narrow compared to the preferred baryon oscillation scale hence we use imaging data from the second red cluster sequence ( rcs2 ) project instead ( yee et al . 2007 ) .
sources are cross - matched between the galex and optical catalogues with a matching tolerance of 2.5 arcsec ( which produces a negligible fraction of incorrect matches ) . in each imaging dataset , the majority of galaxies possess relatively low signal - to - noise ( @xmath50 ) but their detection in both datasets ensures a robust sample .
we note that acquisition of our eventual requirement of @xmath51 galex orbits of data is still ongoing . about @xmath26 of this total had been obtained at the end of 2008 .
targets are chosen from the uv - optical matched sample using a series of magnitude and colour cuts .
these cuts are tuned to optimize the fraction of targets lying at high redshift @xmath52 .
firstly the galaxy magnitudes are de - reddened using standard dust corrections based on the local value of @xmath53 measured from the schlegel , finkbeiner & davis ( 1998 ) dust maps .
the primary galex selection criterion is a red @xmath54 colour ( @xmath55 or @xmath45 drop - out ) , motivated by the lyman break passing through the @xmath45 filter for @xmath52 , and tuned by looking at galaxy templates . at the depth of the medium imaging survey
this colour is noisy , resulting in a significant amount of contamination by low - redshift ( @xmath56 ) galaxies which are partially removed by the additional cuts described below .
we also impose a faint uv magnitude limit @xmath57 and an additional signal - to - noise requirement @xmath58 for the detected nuv flux ( which becomes relevant for fields with unusually high dust content or low exposure time ) .
the galex field - of - view is circular with radius @xmath59 deg ; we only select sources within the central @xmath60 deg because of concerns over the photometry at the edge of the field .
our primary optical selection cuts are derived from sdss @xmath61-band imaging .
we require a uv - optical colour in the range @xmath62 based on the expected model tracks of star - forming galaxies .
we impose a bright @xmath61-band limit @xmath63 ; the uv - optical colour cut implies that the median optical magnitude of our targets is @xmath64 .
finally we increase the high - redshift efficiency by introducing optical colour cuts .
different cuts are used for the sdss and rcs2 regions , governed by the available imaging bands and depths .
for the sdss regions analyzed in this paper , we apply cuts for those ( brighter ) galaxies with good detections in the sdss @xmath65- and @xmath66-bands .
specifically , for targets with @xmath67 and @xmath68 we reject galaxies in the colour space defined by @xmath69 and @xmath70 which is occupied by low - redshift galaxies both theoretically and empirically ( more details will be given in drinkwater et al .
2009 , in preparation ) .
the final fraction of @xmath71 galaxies obtained is @xmath72 .
the redshift distribution is displayed in figure [ fignz ] .
an average of 3 - 4 pointings of the 2df spectrograph per patch of sky is required in order to achieve the required target density of 350 deg@xmath73 . for any observing run the optimal placement of field centres ( based on the current availability of targets )
is achieved using the `` metropolis '' ( simulated annealing ) algorithm ( campbell , saunders & colless 2004 ) .
galaxies are prioritized for spectroscopic follow - up on the basis of optical @xmath61-band magnitude , in the sense that fainter targets are observed first .
the motivation for this strategy is to combat the potential inefficiency of `` mopping up '' residual galaxies in the final pointing for any patch of sky : the brighter remaining galaxies can be observed in a shorter exposure time by configuring fewer fibres . in this paper
we analyze the subset of the wigglez sample assembled from our first observations in august 2006 up until the end of the 08a semester ( july 2008 ) . at this point
we had utilized 108 of our allocated nights , of which the equivalent of 70 nights were clear .
the available galaxy database included @xmath74 reliable ( @xmath75 ) wigglez unique galaxy redshifts .
in this analysis we only use those galaxies lying in the sdss regions of our optical imaging because work is still on - going on the rcs2 portion of the angular selection function .
specifically , we include the wigglez 9-hr , 11-hr , 15-hr and 0-hr ( sdss ) regions illustrated in figure [ figsurvey ] . the number of existing aaomega pointings in these regions is ( 42 , 98 , 140 , 48 ) . the numbers of galaxies in each region with reliable redshifts satisfying the final survey selection criteria are ( 5782 , 14873 , 21629 , 4383 ) , constituting a total sample of @xmath76 for this initial analysis ( about @xmath2 of the final sample ) .
we quantify the small - scale clustering of the galaxy distribution using a standard set of techniques based on the 2-point correlation function .
this statistic compares the number of observed close galaxy pairs to that expected by random chance , as a function of spatial separation .
the key requirement is an ensemble of random ( unclustered ) realizations of the survey possessing the same selection function as the observed galaxy distribution . with this in place
we convert the data ( d ) and random ( r ) galaxy angle - redshift distributions into a grid of co - moving co - ordinates @xmath77 using an assumed cosmological model ( we use a flat model with @xmath24 ) .
we then bin the number of data - data ( dd ) , data - random ( dr ) and random - random ( rr ) pairs in the two - dimensional space of separation perpendicular to the line - of - sight ( denoted by @xmath78 ) and parallel to the line - of - sight ( denoted by @xmath79 ) .
this decomposition is motivated by the influence of galaxy peculiar velocities ( redshift - space distortions ) which shift galaxies in @xmath79 , but not in @xmath78 .
each of our random realizations contains the same number of targets as the data sample , and is generated by a method described below .
the pair counts @xmath80 and @xmath81 are determined by averaging over 10 random realizations .
the 2d redshift - space correlation function @xmath82 is derived using the estimator proposed by landy & szalay ( 1993 ) : @xmath83 ( where this last equation assumes an equal number of data and random galaxies ) .
we bin galaxy pairs by the absolute value of the line - of - sight separation , i.e. @xmath84 .
the `` real - space '' correlation function ( independent of the redshift - space distortion ) can be obtained by summing equation [ eqxiest ] over @xmath79 .
we first define the projected correlation function @xmath85 ( davis & peebles 1983 ) : @xmath86 where the factor of 2 extrapolates the result to the full range @xmath87 .
if we assume that the real - space correlation function @xmath88 is well - described by a power - law @xmath89 , where @xmath90 is the clustering length , @xmath91 is the slope and @xmath92 , and if we neglect the coherent infall velocities described below , we can then derive @xmath93 where @xmath94 the difficulty with this method is that the measurement of @xmath82 becomes noisy for large @xmath79 and therefore the summation in equation [ eqxiproj ] must be truncated at some @xmath95 , invalidating equation [ eqxireal ] .
we therefore adopted the following approach ( similar to the methodology of coil et al .
2008 ) to convert a model real - space correlation function @xmath89 into a projected correlation function which may be compared with the data . in the linear regime
, the effect of coherent infall velocities on the correlation function can be modelled by @xmath96 where @xmath97 are the legendre polynomials , @xmath98 and @xmath99 is the angle between @xmath61 and @xmath79 .
for a power - law real - space correlation function , @xmath100 \xi_r(r)\end{aligned}\ ] ] where @xmath101 is the redshift - space distortion parameter ( hamilton 1992 ; hawkins et al . 2003 ) and @xmath102 is the linear galaxy bias parameter .
we assumed @xmath103 for this model , consistent with our measurements ( see section [ secreddist ] ) , and for each set of trial values @xmath104 we employed the above set of equations to calculate @xmath82 .
for each value of @xmath78 we then integrated this function in the @xmath79 direction up to @xmath105 in order to compare with the correlation function measurements .
we assumed @xmath106 mpc , and we checked that our results did not depend sensitively on the value of @xmath107 .
we treated each of the four survey regions independently , measuring the correlation function and corresponding error .
we then constructed the `` combined '' correlation function by averaging the measurements in the four regions with inverse - variance weighting . for convenience ,
we plot projected correlation functions in this paper as @xmath108 , where @xmath109 we now discuss the generation of the random survey realizations that are required for calculation of the correlation function .
this determination of the survey `` selection function '' will be described fully in a forthcoming paper ( blake et al .
2009 , in preparation ) and we give a brief summary here .
the calculation begins with the angular selection function of the `` parent '' sample of uv - optical matches .
this function is defined firstly by the boundaries of the galex fields and sdss coverage map . secondly , because the uv magnitudes of our sample lie close to the threshold of the galex mis observations , there is a significant incompleteness in the galex imaging that depends on the local dust extinction and galex exposure time .
we used the galex number counts as a function of dust and exposure time to calibrate the relation between these quantities and the parent wigglez target density .
this angular completeness function is displayed in figure [ figangsel ] for the four survey regions analyzed in this paper .
we used this map to produce a series of random realizations of the parent catalogue for each region .
the next step is to process these random parent catalogues into random realizations of the redshift catalogue .
the spectroscopic follow - up of the parent catalogue comprises a network of overlapping aaomega pointings , with field centres optimized by the simulated annealing algorithm and not lying on a regular grid .
the fraction of successful redshifts in each pointing varies considerably depending on weather conditions .
furthermore , the redshift completeness within each aaomega field exhibits a significant radial variation due to acquisition errors at the plate edges . in figure [ figredcomp ]
we illustrate how the redshift completeness varies across these survey regions by simply taking the ratio of successful redshifts to parent galaxies in each pixel .
this is a useful visualization , but in fact the number of unique sectors defined by the overlapping aaomega fields is so large that this determination of the redshift completeness map is very noisy .
indeed , some unique sectors contain zero parent galaxies .
one possible approach is to smooth this completeness map over larger areas to reduce the poisson noise at the expense of a diminished sensitivity to small - scale completeness variations between aaomega pointings . in this analysis
we use an alternative approach , which is to apply the aaomega pointing sequence to each of the random realizations of the parent catalogue , and thereby create an ensemble of random realizations of the redshift catalogue .
the aaomega pointing sequence is defined by the right ascension and declination of the field centre together with the number of successful and unsuccessful redshifts obtained for that pointing . within each field centre
parent galaxies are chosen randomly to create the synthetic redshift catalogue .
it is also necessary to track the sky coverage of the galex data which was contemporaneous with each aaomega pointing .
because the acquisition of the galex imaging data is ongoing with the spectroscopic follow - up , the boundaries of the angular mask must be modulated in step with the redshift follow - up .
in addition we impose the radial redshift completeness variation across each aaomega field , measured independently for each observing run .
the redshift distribution @xmath110 of observed galaxies varies with position in the sky .
this is due to the magnitude prioritization described in section [ sectarsel ] .
because galaxies with fainter @xmath61-band magnitudes are targeted first , the @xmath110 will be skewed toward higher redshifts for areas of the survey which have been targeted by fewer aaomega observations .
this dependence is accounted for in our random catalogues by measuring the magnitude distribution of targeted galaxies as a function of sky position and drawing a random redshift from the correctly weighted @xmath110 .
the optical fibres of the 2df spectrograph can not be placed closer together than @xmath111 arcmin , and there is a diminishing probability of observing in a single pointing both members of a close pair of parent galaxies separated by an angular distance of less than 2 arcmin [ a projected spatial distance of @xmath112 mpc at @xmath34 .
this restriction will eventually be ameliorated by the requirement of observing each patch of sky with 3 - 4 aaomega pointings to build up the number density of the redshift catalogue . at present , however , there is a deficit of close angular pairs in the redshift catalogue , which artificially suppresses the measured value of the galaxy correlation function on small scales .
the close angular pair deficit is illustrated in figure [ figfibcol ] by plotting the ratio @xmath113 as a function of angular separation @xmath99 , where @xmath114 and @xmath115 are the angular correlation functions of the targeted catalogue and the parent catalogue , respectively .
this ratio provides the fraction of surviving close pairs . in order to correct the galaxy correlation function for the missing close pairs we increased the contribution of each galaxy pair to the data - data pair count as a function of angular separation by a factor @xmath116 ( the inverse of the quantity plotted in figure [ figfibcol ] ) using a 2-parameter model @xmath117\}/2 $ ] fitted to the data in figure [ figfibcol ] .
we note that for a survey with a redshift - dependent galaxy number density @xmath118 , the minimum - variance correlation function measurement for separation @xmath119 is achieved if galaxies are assigned a redshift - dependent weight @xmath120^{-1}$ ] where @xmath121 ( efstathiou 1988 ; loveday et al .
1995 ) . in our case
the galaxy number density is sufficiently low that @xmath122 and this weighting makes a negligible difference to the results and we do not use it . the low signal - to - noise spectra obtained by the wigglez survey imply that a small but significant fraction of galaxies are assigned a `` reliable '' ( @xmath42 ) redshift which proves to be incorrect owing to emission - line mis - identification .
this is monitored in the survey by allocating a small number of fibres ( typically 3 to 5 out of 400 per pointing ) to re - observe galaxies with existing @xmath42 redshifts .
the fraction of repeat observations producing a discrepant redshift may be used to estimate the redshift `` blunder '' rate .
there is a significant difference in the reliability of @xmath123 redshifts and @xmath124 redshifts .
@xmath123 redshifts ( which represent a fraction @xmath125 of reliable redshifts ) are typically based either on noisy spectra or on a single emission line with no confirming spectral features , whereas @xmath124 redshifts are based on multiple detected emission lines . comparing repeat observations consisting of a @xmath123 redshift and a @xmath124 redshift , assuming that the latter provides the correct redshift identification , we conclude that @xmath126 of @xmath123 redshifts are blunders .
this agrees with the internal discrepancy rate amongst repeated pairs of @xmath123 redshifts ( which is @xmath127 , which must be divided by two to obtain the blunder rate per object ) . comparing repeat observations consisting of @xmath124 redshifts
we find that only @xmath128 of these redshifts are blunders .
the blunder rate for @xmath123 spectra varies significantly with the true galaxy redshift , which determines how many emission lines appear in the observed wavelength range .
the dependence is displayed in figure [ figzblund ] based on the comparison of @xmath123 and @xmath129 pairs of repeat observations .
the total blunder rate for all reliable ( @xmath42 ) redshifts is below @xmath130 for the range @xmath131 , increasing to @xmath2 by @xmath132 .
the redshift blunder rate does not depend on galaxy continuum magnitude .
redshift mis - identification reduces the measured value of the galaxy correlation function because a fraction of true close data - data pairs are lost as one or both of the redshifts is randomized .
if @xmath133 is the redshift blunder rate , the correction to the correlation function is a constant factor @xmath134 assuming the blunder redshift is randomly distributed .
we applied this correction to the measured correlation function to deduce the final value : @xmath135 when measuring the galaxy correlation function for a particular redshift or luminosity range , we re - calculated the redshift blunder rate for the corresponding sample in each region as explained below .
we corrected the correlation function for that region using equation [ eqcorrblund ] , before combining together the correlation functions for the different regions .
we determined the redshift blunder rate for each region by weighting the blunder probabilities of the @xmath136 individual objects in that region : @xmath137 for objects with @xmath123 we assigned the probability for each object based on its redshift using figure [ figzblund ] . for objects with @xmath124
we assumed a blunder rate of @xmath138 . in order to determine the error in the measured correlation function we must characterize the statistical fluctuations in the data sample .
it is well - known that these fluctuations are not well - described by poisson statistics , for which the error in the pair count @xmath139 in a separation bin would be equal to @xmath140 .
sample variance , geometrical edge effects and the same galaxy participating in pairs in different separation bins cause the statistical variance of the galaxy pair count to exceed the poisson prediction and induce covariances between the bins . in this analysis
we use jack - knife re - sampling to determine the correlation function error . in this technique
the dataset is divided into @xmath136 equal - area sub - regions on the sky .
the correlation function analysis is repeated @xmath136 times , in each case omitting one of the sub - regions in turn . labelling the different correlation function measurements at separation
@xmath119 as @xmath141 from @xmath142 to @xmath143 , the covariance between separation bins @xmath144 and @xmath145 was deduced as : @xmath146 where @xmath147 .
the factor @xmath148 in equation [ eqcov ] is required because the jack - knife re - samples are not independent , sharing a high fraction of common sources . we defined the jack - knife samples by splitting each survey region into @xmath149 sub - regions using constant boundaries of right ascension and declination .
we tried the alternative technique of using the galex tiles to define the jack - knife regions ; this produced a result that did not differ significantly .
future analyses of the wigglez survey clustering will quantify the statistical fluctuations using mock galaxy catalogues constructed from n - body simulations .
figure [ figdist2d ] illustrates the dependence of the 2d redshift - space correlation function @xmath82 of equation [ eqxiest ] on the separations @xmath79 and @xmath78 perpendicular and parallel to the line - of - sight for the sample of wigglez galaxies spanning the full redshift range @xmath150 .
we measured the correlation function separately for the four independent survey regions and combined the results using inverse - variance weighting .
the non - circularity of the contours of constant @xmath151 trace the imprint of galaxy peculiar velocities ; we use linear scales of @xmath78 and @xmath79 in this plot to focus on the large - scale distortions .
in particular , for scales @xmath152 mpc the increase in the value of @xmath151 with increasing angle to the line - of - sight @xmath153 for fixed total separation @xmath154 is a signature of coherent galaxy infall and can be quantified to measure the redshift - space distortion parameter @xmath37 ( see section [ secreddist ] ) .
we also detect some evidence for `` fingers of god '' , in the form of elongation of the contours of @xmath151 along the @xmath79-axis , due to the virialized motions of galaxies in clusters .
there is some similarity here with the results of coil et al .
( 2008 ) figure 7 for luminous blue galaxies , except that our sample size is significantly larger
. galaxy peculiar velocities change values of @xmath79 but not @xmath78 .
the real - space clustering properties of the galaxies may therefore be deduced by integrating @xmath82 along the @xmath79-axis , as discussed in section [ seccorrest ] .
we summed the 2d correlation function for the @xmath150 sample in 5 logarithmic bins of @xmath79 between @xmath155 mpc and @xmath156 mpc .
the result is plotted in figure [ figprojall ] for the projected separation range @xmath157 mpc , with errors obtained from the jack - knife re - sampling .
the full covariance matrix @xmath158 deduced from the jack - knife re - samples is displayed in figure [ figcov ] by plotting in greyscale the correlation coefficient between two separation bins @xmath66 and @xmath144 : @xmath159 we employed the methodology of section [ seccorrest ] to fit a power - law real - space correlation function @xmath160 to the redshift - space data over the range @xmath161 mpc , by minimizing the @xmath162 statistic using the covariance matrix : @xmath163 where @xmath164 is the vector of offsets between the data and the model , and @xmath165 is the inverse of the covariance matrix .
the fitting range was motivated by our wish to estimate the clustering length @xmath90 for which @xmath166 .
a power - law provides a good fit to the data with a best - fitting @xmath167 ( for 8 degrees of freedom ) .
the marginalized measurements of the power - law parameters are @xmath168 mpc and @xmath5 for the @xmath150 sample .
we compare these measurements to previous studies of uv - selected and optically - selected galaxies in section [ secprevious ] .
in figure [ figprojreg ] we plot the separate projected correlation function measurements for each of the four survey regions analyzed in this paper .
the four regions give consistent results . in order to derive the bias factor of the galaxies with respect to dark matter we generated a model non - linear matter power spectrum at @xmath4 assuming a flat cosmological model with fiducial parameters @xmath169 , @xmath170 , @xmath171 ( where @xmath172 km s@xmath10 mpc@xmath10 ) and @xmath173 , using the `` camb '' software package ( lewis , challinor & lasenby 2000 ) including corrections for non - linear growth of structure using the fitting formula of smith et al .
we used this model power spectrum to determine the non - linear matter correlation function @xmath174 at @xmath4 .
the resulting correlation function satisfied @xmath175 for @xmath176 mpc , which we assumed as our estimate of @xmath177 , the clustering length of dark matter at @xmath4 .
given that the overall amplitude of the power spectrum scales with redshift in the linear regime as @xmath178 , where @xmath179 is the linear growth factor , we can approximate : @xmath180 where @xmath181 .
hence the linear bias factor @xmath102 of a population of galaxies with clustering length @xmath90 can be approximated as : @xmath182 our measured clustering length @xmath183 mpc for a sample at median redshift @xmath184 is hence equivalent to a linear bias factor @xmath185 .
the peculiar velocities generated by large - scale coherent infall can be parameterized by @xmath101 where @xmath102 is the linear bias parameter ( kaiser 1987 ) . for a flat cosmological constant model with @xmath186 , @xmath187 , and
our real - space clustering measurement @xmath188 hence predicts @xmath103 at the median redshift of the sample .
the purpose of this section is to demonstrate that our data contains this self - consistent signal of peculiar velocities ( we leave detailed fits for @xmath37 to a further study ) .
we may quantify the imprint of peculiar velocities by measuring the quadrupole moment , @xmath189 , of the 2d correlation function ( hamilton 1992 ) .
this statistic quantifies the anisotropy evident in figure [ figdist2d ] . if we define the correlation function moment @xmath190 for multipole @xmath191 as : @xmath192 we can then show that @xmath193 - \xi_0(s ) } = \frac { \frac{4}{3 } \beta + \frac{4}{7 } \beta } { 1 + \frac{2}{3 } \beta + \frac{1}{5 } \beta^2 } \label{eqdistq}\ ] ] which is valid for large scales @xmath194 mpc .
figure [ figdistq ] plots the measured quantity @xmath189 as a function of separation @xmath119 , together with the prediction of equation [ eqdistq ] for various values of @xmath37 . in order to construct the quantity @xmath189 we measured the 2d redshift - space correlation function in bins of @xmath119 and @xmath195 , and summed over @xmath195 , weighting in accordance with equation [ eqxil ] .
the result is consistent with our estimate @xmath196 and constitutes a statistically - significant detection of redshift - space distortions in our sample .
our sample of wigglez galaxies is large enough for us to analyze the dependence of the clustering length @xmath90 on redshift , galaxy luminosity and colour .
the situation is complicated by our joint uv - optical selection and strong luminosity - redshift correlation , but we can make some comparisons with previous studies .
we fix the correlation function slope @xmath197 in this section of the analysis . the variation of the clustering length with redshift is plotted in figure [ figr0vsz ] , dividing all wigglez galaxies in the range @xmath198 into redshift bins of width @xmath199 . the clustering length is roughly constant with redshift for the range @xmath200 , with a trend to a reduced clustering strength at low redshifts .
our interpretation of the overall constancy of @xmath201 is that it is a product of two cancelling effects .
galaxy luminosity increases with redshift , which would tend to increase clustering length , but at redshifts @xmath52 optically red galaxies , which possess enhanced clustering strengths , are removed from the sample by the optical colour cuts described in section [ sectarsel ] .
we also analyzed the clustering in absolute magnitude and rest - frame colour bins .
we considered the clustering as a function of rest - frame @xmath45-band and @xmath202-band absolute magnitudes , which are well - matched in wavelength ( for redshift @xmath203 ) to the observed - frame @xmath47-band and @xmath61-band magnitudes which are used to define our target samples .
for this initial analysis we assumed redshift - dependent average k - corrections which we applied to all galaxies regardless of colour .
these k - corrections were derived using the spectral energy distribution of a lyman break galaxy including an intrinsic dust contribution @xmath204 , which produces a very good match to the redshift - dependence of the average observed @xmath205 colour of the wigglez targets .
we note that the @xmath45-band and @xmath202-band absolute magnitudes of our target sample correlate strongly with redshift .
this is depicted by figure [ figselbox ] which plots the target selection box in @xmath206 for 4 different redshifts , also indicating the characteristic absolute magnitudes @xmath207 at each redshift obtained from arnouts et al .
( 2005 ) and willmer et al .
( 2006 ) . between @xmath208 and @xmath132 the average value of @xmath209 brightens by 2 magnitudes ( becoming positive at @xmath210 ) and the average value of @xmath211 brightens by 4 magnitudes ( becoming positive at @xmath212 ) .
the dependence of the clustering length @xmath90 of the @xmath150 wigglez sample on @xmath213 , @xmath214 and @xmath215 is displayed in the panels of figure [ figr0panel ] .
these measurements show that the clustering strength of the sample increases steadily with @xmath202-band absolute magnitude , @xmath45-band absolute magnitude and reddening @xmath215 colour .
sub - samples have values of @xmath90 ranging from @xmath216 mpc to @xmath217 mpc .
figure [ figr0vsmb ] plots the variation of @xmath90 with @xmath213 for the low - redshift and high - redshift halves of the dataset , divided at @xmath218 .
this measurement confirms that at fixed @xmath213 , the clustering length of the sample drops slightly with redshift as the redder galaxies are removed by the colour cuts .
coil et al .
( 2008 ) present clustering measurements as a function of galaxy colour and luminosity for the deep2 galaxy redshift survey , which has measured redshifts for @xmath219 galaxies in the range @xmath220 over an area of 3 deg@xmath0 .
the deep2 subset of luminous blue galaxies ( coil et al .
table 2 , line 5 ) has best - fitting clustering parameters @xmath221 mpc and @xmath222 at @xmath223 ( for a galaxy density @xmath224 mpc@xmath225 and median absolute magnitude @xmath226 assuming @xmath227 ) .
these results lie in good agreement with ours .
milliard et al . ( 2007 ) and heinis et al . (
2007 ) present clustering analyses of galex - selected samples . at low redshift ( @xmath228 )
the clustering strength of the uv - selected sample is @xmath229 mpc , corresponding to low - density environments , and shows no dependence on uv luminosity ( indeed , there is tentative evidence for an anti - correlation between @xmath90 and luminosity ) .
these results may naturally be compared to clustering measurements of @xmath22 lbgs also selected at rest - frame uv wavelengths ( e.g. giavalisco & dickinson 2001 ; ouchi et al .
2001 ; arnouts et al . 2002 ; foucaud et al . 2003 ;
adelberger et al .
2005 ; allen et al . 2005 ;
ouchi et al . 2005 ;
lee et al . 2006 ; yoshida et al .
these results show a qualitatively different conclusion : lbgs are highly clustered and concentrated in overdense regions .
furthermore , the clustering strength for galaxies brighter than @xmath230 increases with @xmath45 luminosity , reaching @xmath231 mpc for the most luminous sub - samples .
yoshida et al . ( 2008 ) demonstrate that the behaviour of the clustering length @xmath90 depends on a combination of uv and optical luminosities : galaxies bright in optical magnitudes are strongly clustered irrespective of uv magnitude , whereas galaxies faint in optical magnitude have correlation lengths increasing with uv luminosity ( see yoshida et al .
fig.15 ) . in figure [ figr0compare ] we overplot the clustering measurements of the @xmath150 wigglez sample as a function of @xmath45 absolute magnitude on the compilation of low - redshift and high - redshift clustering measurements presented by heinis et al .
( 2007 ) . at low @xmath45 absolute magnitudes @xmath232
the clustering strengths of the different uv - selected samples agree well .
this absolute magnitude range corresponds to low redshifts @xmath228 in the wigglez sample ( figure [ figselbox ] ) for which we recover a clustering length @xmath233 mpc , similar to samples of low - redshift quiescent star - forming galaxies . at higher @xmath45 luminosities and redshifts ,
the wigglez clustering strength is more comparable to @xmath6 lbgs rather than @xmath4 uv - selected galaxies .
this is expected as the @xmath234 wigglez selection cut becomes effective for @xmath235 and the nature of the resulting wigglez galaxy population changes to merger - induced starbursts .
the wigglez sample does not recover the very high values of @xmath90 present in very luminous lbgs at @xmath6 ; we suggest that this may be a result of the wigglez colour cuts selecting against redder galaxies .
the clustering amplitude of the wigglez target sample is a required input for forecasting the accuracy with which the full 1000 deg@xmath0 survey will measure the large - scale galaxy power spectrum .
the error in the power spectrum measurement @xmath236 at a given redshift @xmath237 and fourier wavenumber @xmath145 is determined by the quantity @xmath238 , where @xmath118 is the galaxy number density and @xmath239 is the galaxy power spectrum amplitude .
this quantity fixes the balance between sample variance and shot noise in the measurement error such that @xmath240 where @xmath241 is the total number of independent fourier modes contributing towards the measurement ( e.g. feldman , kaiser & peacock 1994 ; tegmark 1997 ) .
the contributions of sample variance and shot noise are equal when @xmath242 .
we model the angle - averaged redshift - space linear galaxy power spectrum as a function of @xmath145 and @xmath237 as : @xmath243^\gamma \left ( 1 + \frac{2\beta}{3 } + \frac{\beta^2}{5 } \right ) \label{eqpkmod}\ ] ] where we assume @xmath244 mpc , @xmath245 mpc , @xmath246 and @xmath103 .
the second term on the right - hand - side of equation [ eqpkmod ] describes the boost from the galaxy linear bias factor @xmath102 ( equation [ eqbias ] ) using the relation @xmath247 .
the third term is the result of redshift - space distortions averaged over angles .
we used the cosmological parameters as listed in section [ secr0 ] to produce the @xmath4 dark matter power spectrum : @xmath169 , @xmath170 , @xmath171 and @xmath248 . in order to incorporate the fraction of redshift blunders @xmath249 we reduced the effective value of the power spectrum by a factor @xmath250 [ i.e. increased the value of @xmath251 by a factor @xmath252 .
figure [ fignpvsz ] plots the dependence of @xmath238 on redshift for a set of different scales @xmath253 mpc@xmath10 relevant for the detection of baryon acoustic oscillations , assuming a source redshift distribution combining the survey regions plotted in figure [ fignz ] .
we further assume a total target density of 350 deg@xmath73 with a @xmath26 redshift completeness .
we note that over a significant range of redshifts and scales our large - scale power spectrum measurement will be limited by sample variance rather than shot noise , i.e. @xmath254 . a useful quantity to describe the survey is the scale - dependent `` effective volume '' @xmath255 which is defined by @xmath256 ^ 2 \frac{dv}{dz } \ , dz\ ] ] where @xmath257 is the co - moving volume element .
the effective volume represents an optimally - weighted stacking of power spectrum measurements at different redshifts ( tegmark 1997 ) . for scales
@xmath258 mpc@xmath10 we find @xmath259 gpc@xmath8 .
thus the survey design will achieve the goal of mapping @xmath260
gpc@xmath261 gpc@xmath8 .
we can use the effective survey volume to forecast the error in the final survey power spectrum @xmath262 in a fourier bin of width @xmath263 ( tegmark 1997 ) : @xmath264 this prediction is plotted for bins of width @xmath265 mpc@xmath10 in figure [ figpksim ] , in which we divide the power spectrum by the `` no - wiggles '' reference spectrum provided by eisenstein & hu ( 1998 ) in order to delineate clearly the baryon acoustic oscillations .
we also generated 100 monte carlo realizations of the final 1000 deg@xmath0 survey using the methods described in blake & glazebrook ( 2003 ) and glazebrook & blake ( 2005 ) .
the scatter in the power spectrum measurements across the realizations was very close to that predicted by equation [ eqpkerr ] .
we used these monte carlo realizations to assess the accuracy with which the full wigglez survey will measure the tangential and radial standard ruler scale imprinted by the baryon acoustic oscillations via the fitting formula described in blake et al .
( 2006 ) . restricting ourselves to the @xmath266 subset , and first considering
an `` angle - averaged '' measured power spectrum @xmath267 , we found that the scatter in the fitted acoustic wavescale was @xmath268 .
measuring instead a 2d power spectrum @xmath269 , where @xmath270 and @xmath271 are wavevectors measured perpendicular and parallel to the line - of - sight , the scatters in the tangential and radial fitted wavescales were @xmath272 and @xmath273 , respectively .
this latter pair of measurements corresponds to the accuracy of determination of the quantities @xmath274 and @xmath275 at an effective redshift @xmath276 , where @xmath277 is the angular diameter distance , @xmath278 is the high - redshift hubble constant , and @xmath119 is the sound horizon at recombination , i.e. the standard ruler scale . dividing the survey into redshift slices we find that the angle - averaged wavescale may be measured with accuracy @xmath279 in redshift slices @xmath280 .
the angle - averaged wavescale measures a quantity proportional to @xmath281 , as discussed by eisenstein et al .
( 2005 ) .
these forecasts should be considered a pessimistic lower limit on expected performance for two reasons .
firstly we have neglected the cosmological information contained in the overall shape of the galaxy power spectrum , which is divided out in the above analysis to focus on the `` standard ruler '' aspect of the acoustic oscillations .
this method produces robustness against systematic errors ( which are expected to affect the shape of the power spectrum but not the oscillatory signature ) .
the power spectrum shape carries information about @xmath282 and @xmath283 which further breaks the degeneracy in cosmological distances between these two parameters and the dark energy .
secondly we have neglected the improvement offered by `` reconstruction '' of the density field , which sharpens the measurement of the acoustic signature by un - doing ( to first order ) the large - scale coherent galaxy motions which smooth out the acoustic peaks ( eisenstein et al .
2007 ) .
[ cols="^,^,^,^,^,^,^,^,^,^ " , ] [ tabmodel ] we investigated improved forecasts using the methodology of seo & eisenstein ( 2007 ) which properly incorporates information from the power spectrum shape , redshift - space distortions and density - field reconstruction .
the predicted tangential and radial measurement accuracies for the @xmath266 sample are @xmath284 and @xmath285 , respectively ( and are correlated with a correlation coefficient @xmath286 , further enhancing the power to constrain the cosmological model ) .
we assume here that reconstruction can improve the parameters ( @xmath287 , @xmath288 ) defined by seo & eisenstein ( 2007 ) by a factor equal to @xmath289 ( eisenstein , priv .
comm . ) . dividing the survey into redshift slices we find that the tangential and radial wavescales may be measured with accuracies @xmath290 for @xmath291 , @xmath292 for @xmath293 and @xmath294 for @xmath295 .
this information is collected in table [ tabmodel ] for ease of reference .
we used this last set of forecasts with reconstruction in 3 redshift bins to determine the expected accuracy of measurement of a constant equation - of - state @xmath13 of dark energy ( assuming the measurements of @xmath277 and @xmath296 are correlated with coefficient @xmath297 ) .
confidence ellipses are displayed in figure [ figomw1 ] in the space of @xmath13 and the matter density @xmath282 for a flat cosmology with fiducial model @xmath298 and @xmath299 .
results are shown for each redshift bin separately and for the combination of all 3 bins . in order to generate this figure
we have used the 5-year - wmap measurement of the cmb acoustic scale @xmath300 ( komatsu et al .
2009 ) in order to cancel the dependence of the baryon oscillation measurement on the sound horizon at recombination . in figure [ figomw1 ] we have not included any further cmb information or other external datasets .
the marginalized errors are @xmath301 and @xmath302 . in figure [ figomw2 ]
we add in information from the 5-year - wmap measurement of the cmb shift parameter @xmath303 ( komatsu et al .
2009 ) , including the correlation between @xmath304 and @xmath305 , together with the latest supernovae data from the essence , snls and hst observations ( see wood - vasey et al .
2007 , astier et al .
2006 , riess et al .
2007 , davis et al . 2007 ) .
the marginalized errors in the cosmological model from the full combination of datasets are @xmath306 and @xmath307 .
the forecast performance of the wigglez survey exceeds that of the current cmb and supernovae data , but the different measurements are also complementary , breaking degeneracies in the @xmath308-plane through independent techniques .
disagreement between any pair of datasets would produce the possibility of discovering non - standard physics ( if it exists ) or systematic measurement errors .
the final accuracy of @xmath13 constitutes a robust and precise test of the dark energy model .
we have measured the small - scale clustering amplitude of high - redshift bright emission - line galaxies using the first @xmath2 of spectra from the aat wigglez dark energy survey ( @xmath309 galaxies in the redshift range @xmath150 ) .
we have successfully developed a methodology to generate random realizations of the survey incorporating the currently sparse selection function and redshift incompleteness .
we find that : * the wigglez galaxy sample in the redshift range @xmath150 possesses a clustering length @xmath168 mpc and slope @xmath5 .
this clustering amplitude significantly exceeds that of uv - selected samples at @xmath310 and agrees well with that of the most luminous blue galaxies observed by the deep2 galaxy redshift survey at @xmath311 .
the clustering amplitude of wigglez targets is comparable to that of lyman break galaxies at a similar uv luminosity . *
the clustering length of the wigglez targets is approximately constant with redshift for the range @xmath235 .
the value of @xmath90 increases with @xmath202-band luminosity , @xmath45-band luminosity , and reddening rest - frame colour . * the redshift - space distortion signature of coherent galaxy motions is detected and its amplitude ( @xmath196 ) is consistent with that predicted from the galaxy bias .
we detect some evidence for `` fingers of god '' due to the virialized motions of galaxies in clusters . using these results
, we forecast the performance of the full 1000 deg@xmath0 wigglez survey in the measurement of the galaxy power spectrum and cosmological model .
we find that : * the survey design is well - tuned to the `` optimal '' mean galaxy number density @xmath312 , where @xmath31 is the amplitude of the galaxy power spectrum on the scales of importance for baryon oscillations .
* the survey will delineate the baryon acoustic oscillations in the large - scale clustering pattern in three independent redshift slices , providing measurements of the cosmic distance and expansion rate in each redshift slice with accuracies of @xmath11 . *
the resulting measurement of a constant dark energy equation - of - state parameter @xmath13 from the wigglez survey , calibrating the standard ruler using the cmb measurement of the acoustic scale , has a higher precision than provided by current supernovae datasets .
these independent dark energy probes lie in a highly complementary direction in the parameter space of @xmath12 and @xmath282 .
the full combination of wigglez , supernovae and cmb datasets provides a measurement of the equation of state with accuracy @xmath313 , constituting a robust and precise test of the dark energy model incorporating cross - checking of systematic errors between different probes .
the final survey will enable a wide range of scientific investigations into the cosmological model and galaxy evolution .
we acknowledge financial support from the australian research council through discovery project grants funding the positions of sb , mp , gp and td .
we also thank the university of queensland for supporting the phd scholarship of rj .
we acknowledge the efforts of nick jones and david barnes in creating the online wigglez database , and emily wisnioski for incorporating principal component analysis sky subtraction into the data reduction pipeline .
we thank karl forster for his assistance in scheduling our galex observations and sebastien heinis , ted wyder and mark seibert for invaluable galex support and discussions .
we acknowledge correlation function modelling performed by carlos contreras and ben jelliffe which revealed a mistake in the submitted version of this paper .
galex ( the galaxy evolution explorer ) is a nasa small explorer , launched in april 2003 .
we gratefully acknowledge nasa s support for construction , operation and science analysis for the galex mission , developed in co - operation with the centre national detudes spatiales of france and the korean ministry of science and technology . finally , the wigglez survey would not be possible without the dedicated work of the staff of the anglo - australian observatory in the development and support of the aaomega spectrograph , and the running of the aat .
adelberger k.l . ,
steidel c.c . , pettini m. , shapley a.e . ,
reddy n.a .
, erb d.k . , 2005 , apj , 619 , 697 allen p.d .
, moustakas l.a . , dalton g. , macdonald e. , blake c.a , clewley l. , heymans c. , wegner g. , 2005 , mnras , 360 , 1244 arnouts s. et al . , 2002 , mnras , 329 , 355 arnouts s. et al . , 2005 , apj , 619 , 43 astier p. et al . , 2006 , a&a , 447 , 31 blake c.a , glazebrook k. , 2003 , apj , 594 , 665 blake c.a .
, parkinson d. , bassett b. , glazebrook k. , kunz m. , nichol r.c . , 2006 ,
mnras , 365 , 255 campbell l. , saunders w. , colless m. , 2004 , mnras , 350 , 1467 coil a. et al . , 2008 , apj , 672 , 153 cole s. et al . , 2005 , mnras , 362 , 505 coles p. , 1993 ,
mnras , 262 , 1065 colless m. et al . , 2001 ,
mnras , 328 , 1039 cowie l.l . , songaila a. , hu e.m .
, cohen j.g . , 1996 ,
aj , 112 , 839 cooray a. , hu w. , huterer d. , joffre m. , 2001 , apj , 557l , 7 davis m. , peebles p.j.e .
, 1983 , apj , 208 , 13 davis t.m .
et al . , 2007 ,
apj , 666 , 716 efstathiou g. , 1988 , in lawrence a. , ed .
3rd iras conf . ,
comets to cosmology , springer , new york , p.312 eisenstein d.j .
, hu w. , 1998 , apj , 518 , 2 eisenstein d.j .
et al . , 2001 ,
aj , 122 , 2267 eisenstein d.j . , 2002 , _ ` large - scale structure and future surveys ' _ in ` next generation wide - field multi - object spectroscopy ' , asp conference series vol . 280 , ed . m.brown & a.dey , p.35 ( astro - ph/0301623 ) eisenstein d.j .
et al . , 2005 ,
apj , 633 , 560 eisenstein d.j . ,
j . , sirko e. , spergel d.n . , 2007 , apj , 664 , 675 feldman h.a . ,
kaiser n. , peacock j.a .
, 1994 , apj , 426 , 23 foucaud s. , mccracken h.j . , le fevre o. , arnouts s. , brodwin m. , lilly s.j . ,
crampton d. , mellier y. , 2003 , a&a , 409 , 835 gaztanaga e. , cabre a. , hui l. , 2008 , submitted ( astro - ph/0807.3551 ) giavalisco m. , dickinson m. , 2001 , apj , 550 , 177 glazebrook k. et al . , 2004 , nature , 430 , 181 glazebrook k. , blake c.a . , 2005 , apj , 631 , 1 hamilton a.j.s .
, 1992 , apj , 385 , 5 hawkins e. et al .
, 2003 , mnras , 346 , 78 heinis s. , treyer m. , arnouts s. , milliard b. , donas j. , gal r. , martin d.c . , viton m. , 2004 , a&a , 424 , l9 heinis s. et al . , 2007 , apjs , 173 , 503 hu w. , haiman z. , 2003 , phys . rev .
d , 68 , 063004 huesti g. , 2006 , a&a , 449 , 891 kaiser n. , 1987 , mnras , 227 , 1 komatsu e. et al . , 2009 , apjs , 180 , 330 landy s.d . , szalay a.s .
, 1993 , apj , 412 , 64 lee k .- s . , giavalisco m. , gnedin o. , somerville r.s .
, ferguson h.c . , dickinson m. , ouchi m. , 2006 , apj , 642 , 63 lewis a. , challinor a. , lasenby a. , 2000 , apj , 538 , 473 linder e.v . , 2003 ,
d , 68 , 083504 loveday j. , maddox s.j .
, efstathiou g. , peterson b.a . , 1995 , apj , 442 , 457 martin d. , et al . , 2005 ,
apj , 619 , l1 meneux b. et al . , 2006 ,
a&a , 452 , 387 milliard b. et al . , 2007 , apjs , 173 , 494 ouchi m. et al . , 2001 ,
apj , 558 , l83 ouchi m. et al . , 2005 ,
apj , 635 , l117 parkinson d. , blake c.a . , kunz m. , bassett b. , nichol r. , glazebrook k. , 2007 , mnras , 377 , 185 percival w. , cole s. , eisenstein d.j .
, nichol r.c . , peacock j.a . , pope a.c . ,
szalay a.s .
, 2007 , mnras , 381 , 1053 riess a. et al .
, 2007 , apj , 659 , 98 saunders w. et al . , 2004 , _
` aaomega : a scientific and optical overview ' _ in ` ground - based instrumentation for astronomy ' , ed .
a.moorwood & i.masanori , proceedings of the spie , vol . 5492 , p.389 scherrer r.j . , weinberg d.h . , 1998 , apj , 504 , 607 schlegel d.j . , finkbeiner d.p . , davis m. , 1998 , apj , 500 , 525 sharp r. et al . , 2006 , _ ` performance of aaomega : the aat multi - purpose fiber - fed spectrograph ' _ in ` ground - based and airborne instrumentation for astronomy ' , ed . i.mclean & i.masanori , proceedings of the spie , vol . 6269 , p.14 ( astro - ph/0606137 ) seo h
j . , eisenstein d.j . , 2003 , apj , 598 , 720 seo h .-
j . , eisenstein d.j . , 2007 , apj , 665 , 14 smith r.e .
et al . , 2003 , mnras , 341 , 1311 tegmark m. , 1997 , prl , 79 , 20 van dokkum p. et al . , 2004 ,
apj , 611 , 703 willmer c. et al .
, 2006 , apj , 647 , 853 wood - vasey w. et al . , 2007 , apj , 666 , 694 yee h. , gladders m. , gilbank d. , majumdar s. , hoekstra h. , ellingson e. , 2007 , _
` the red - sequence cluster surveys ' _ in ` cosmic frontiers asp conference series ' , ed . n.metcalfe
& t.shanks , vol . 379 , p.103 ( astro - ph/0701839 ) york d.g .
et al . , 2000 ,
aj , 120 , 1579 yoshida m. , shimasaku k. , ouchi m. , sekiguchi k. , furuswawa h. , okamura s. , 2008 , apj , 679 , 269 | the wigglez dark energy survey is a large - scale structure survey of intermediate - redshift uv - selected emission - line galaxies scheduled to cover 1000 deg@xmath0 , spanning a broad redshift range @xmath1 .
the main scientific goal of the survey is the measurement of baryon acoustic oscillations ( bao ) in the galaxy clustering pattern at a significantly higher redshift than previous studies . the bao may be applied as a standard cosmological ruler to constrain dark energy models .
based on the first @xmath2 of the dataset , we present initial results concerning the small - scale clustering of the wigglez targets , together with survey forecasts .
the wigglez galaxy population possesses a clustering length @xmath3 mpc , which is significantly larger than @xmath4 uv - selected samples , with a slope @xmath5 .
this clustering length is comparable to @xmath6 lyman break galaxies with similar uv luminosities .
the clustering strength of the sample increases with optical luminosity , uv luminosity and reddening rest - frame colour .
the full survey , scheduled for completion in 2010 , will map an effective volume @xmath7 gpc@xmath8 ( evaluated at a scale @xmath9 mpc@xmath10 ) and will measure the angular - diameter distance and hubble expansion rates in three redshift bins with accuracies @xmath11 .
we will determine the value of a constant dark energy equation - of - state parameter , @xmath12 , with a higher precision than existing supernovae observations using an entirely independent technique .
the wigglez and supernovae measurements lie in highly complementary directions in the plane of @xmath13 and the matter density @xmath14 .
the forecast using the full combination of wigglez , supernovae and cmb datasets is a marginalized error @xmath15 , providing a robust and precise measurement of the properties of dark energy including cross - checking of systematic errors .
surveys , cosmology : observations , large - scale structure of universe , galaxies : starburst |
the fact that the capacitance is a geometrical factor is an important property in courses on electricity and magnetism.@xcite derivations of this property are usually based on the principle of superposition@xcite and the green function formalismuehara , lorenzo .
nevertheless , such derivations are not convenient for calculations .
alternative techniques to calculate the capacitance coefficients based on the green function formalism@xcite and other methods@xcite have been developed . in this paper
we give a simple proof of the geometrical nature of the capacitance coefficients based on laplace s equation .
our approach permits to demonstrate many properties of the capacitance matrix .
the method is illustrated by reproducing some well known results , and applications in complex situations are suggested .
we consider a system of @xmath0 internal conductors and an external conductor that encloses them .
the potential on each internal conductor is denoted by @xmath1 , @xmath2 .
the surface of the external conductor is denoted by @xmath3 , and its potential is denoted by @xmath4 ( see fig . [
fig : ncond ] ) .
one reason to introduce the external conductor is that it provides a closed boundary to ensure the uniqueness of the solutions .
in addition , many capacitors contain an enclosing conductor as for the case of spherical concentric shells .
as we shall see , the case in which there is no external conductor can be obtained in the appropriate limit .
internal conductors with conductor @xmath5 enclosing them .
the normals @xmath6 with @xmath7 point outward with respect to the conductors and inward with respect to the volume @xmath8 ( defined by the region in white ) . the surfaces @xmath9 with @xmath10 are slightly bigger than the ones corresponding to the conductors .
in contrast , the surface @xmath3 is slightly smaller than the surface of the external conductor.,width=245 ] the surface charge density @xmath11 on an electrostatic conductor is given by@xcite @xmath12where @xmath6 is an unit vector normal to the surface @xmath9 pointing outward with respect to the conductor ( see fig .
[ fig : ncond ] ) ; @xmath13 and @xmath14 denote the electrostatic field and potential respectively .
the charge on each conductor is given by @xmath15the surface @xmath9 encloses the conductor @xmath16 and is arbitrarily near and locally parallel to the real surface of the conductor ( see fig .
fig : ncond).@xcite we define the total surface @xmath17 as @xmath18the volume @xmath8 defined by the surface @xmath17 is the one delimited by the external surface @xmath3 and the @xmath0 internal surfaces @xmath9 .
the potential @xmath14 in such a volume must satisfy laplace s equation with the boundary conditions @xmath19because of the linearity of laplace s equation , the solution for @xmath14 can be parameterized as @xmath20where the @xmath21 are functions that satisfy laplace s equation in the volume @xmath8 with the boundary conditions @xmath22the solutions for @xmath21 ensure that @xmath14 is the solution of laplace s equation with the boundary conditions in eq .
( [ cond front fi ] ) .
the uniqueness theorem also ensures that the solution for each @xmath21 is unique ( as is the solution for @xmath14 ) .
the boundary conditions ( [ cond fi ] ) indicate that the @xmath21 functions depend only on the geometry .
if we apply the gradient operator in eq .
( [ fi = fi fi ] ) and substitute the result into eq .
( [ qigen ] ) , we obtain [ q_i_sistema_cargas ] @xmath23 which shows that the @xmath24 factors are exclusively geometric
. the symmetry of the associated @xmath24 matrix can be obtained by purely geometrical arguments .
we start from the definition of @xmath24 in eq .
( q_i_sistema_cargas.b ) and find @xmath25 where we have used the fact that @xmath26 on the surface @xmath9 and zero on the other surfaces . from gauss
s theorem we obtain @xmath27\,dv.\end{aligned}\ ] ] because @xmath28 in @xmath8 , it follows that @xmath29 equation implies that @xmath24 is symmetric,@xcite that is , @xmath30 for certain configuration of conductors , consider two sets of charges and potentials @xmath31 and @xmath32 . from eqs .
( [ q_i_sistema_cargas ] ) and ( [ prop2 ] ) we have that @xmath33 which implies that @xmath34 equation ( [ recip ] ) is known as the reciprocity theorem.@xcite when one or more of the @xmath0 internal conductors has an empty cavity , is well known that there is no charge induced on the surface of the cavity berkeley , jack ( let us call it @xmath35 ) . consequently , although @xmath36 is part of the surface of the conductor , such a surface can be excluded in the integration in eq .
( [ qigen ] ) .
in addition , we can check by uniqueness that @xmath37 in the volume of the cavity @xmath38 so that @xmath39 in such a volume , and hence it can be excluded from the volume integral ( [ cij 2 ] ) . in conclusion neither @xmath40 nor @xmath41 contribute in this case .
the situation is different if there is another conductor in the cavity . in this case
, the surface of the cavity contributes in eq .
( [ qigen ] ) .
similarly the volume between the cavity and the embedded conductor contributes in the volume integral ( [ cij 2 ] ) .
the arguments can be extended for successive embedding of conductors in cavities as shown by fig .
[ fig : ncond2 ] or for conductors with several cavities
. corresponds to the region in white .
the regions corresponding to empty cavities ( and their associated surfaces and volumes ) can be excluded without affecting the calculations . in this picture cavity
a is empty and its surface and volume need not be considered for calculations.,width=245 ]
we define a function @xmath42 @xmath43and see from eq .
( [ cond fi ] ) that @xmath44since @xmath45 throughout the surface @xmath17 , we see by uniqueness that @xmath45 in the volume @xmath8 from which we find that @xmath46 in addition , by summing over @xmath47 in eq .
( [ q_i_sistema_cargas.b ] ) and taking into account eq .
( [ propf ] ) , we find that @xmath48the symmetry of the @xmath24 elements leads also to @xmath49 equations and imply that the sum of the elements over any row or column of the matrix is zero . appendix [ ap : pruebas ] gives some proofs of consistency for these important properties .
taking into account the symmetrical nature of the @xmath24 matrix with dimensions @xmath50 and the @xmath5 constraints in eq .
( [ prop1 ] ) , we see that for a system of @xmath0 conductors surrounded by another conductor @xmath5 , the number of independent capacitance coefficients is @xmath51 -(n+1)=\frac{n(n+1)}{2 } , \label{grad lib}\ ] ] other important properties are that [ prop3 ] @xmath52 equation follows straightforwardly from eq .
( [ cij 2 ] ) . to demonstrate eq .
, we recall that the solutions of laplace s equation can not have local minima nor local maxima in the volume in which the equation is valid.@xcite consequently , the @xmath53 functions must lie in the interval @xmath54 because @xmath55 on any surface @xmath9 for @xmath56 , we see that @xmath21 acquires its minimum value on such surfaces .
therefore the function @xmath57 should point outward with respect to the conductor @xmath16 for @xmath56 . hence @xmath58 we substitute eq .
( [ gradf ] ) into eq .
( [ q_i_sistema_cargas ] ) and obtain @xmath59 for @xmath56 .
an additional derivation of the fact that @xmath60 can be obtained by taking into account that @xmath21 acquires its maximum value on the surface @xmath61 .
equation ( [ prop1 ] ) can be rewritten as @xmath62 from eq .
( [ prop3 ] ) we have that @xmath63 for @xmath64 and @xmath65 . hence [ prop4all ] @xmath66 the following properties follow from eqs .
( [ prop2 ] ) , , , and [ prop5all ] @xmath67 .
a particularly interesting case arises when the external conductor is at zero potential .
in such a case , although the elements of the form @xmath68 do not necessarily vanish , they do not appear in the contributions to the charge on the internal conductors as can be seen from eq .
( q_i_sistema_cargas ) by setting @xmath69 . for this reason ,
the capacitance matrix used to describe @xmath0 free conductors ( that is , not surrounded by another conductor ) has dimensions @xmath70.@xcite
we illustrate our method by deriving the basic properties of a system of two conductors . these examples will show the usefulness of eq .
( q_i_sistema_cargas ) and some of the properties derived from our approach .
we analyze a single internal conductor with an external conductor that is , @xmath71 .
the internal conductor is labeled as conductor 1 . from eqs .
( prop2 ) and ( [ prop1 ] ) we have @xmath72 therefore , there is only one independent coefficient , say @xmath73 ( in agreement with eq . with @xmath74 ) .
the charges on the internal and external conductors can be calculated from eq .
( [ q_i_sistema_cargas ] ) @xmath75 equation is consistent with eq .
( [ ext int ] ) and shows that the charge induced on the surface of the cavity of the conductor 2 is opposite to the charge on the conductor 1 . in table
[ tab : resumen ] we display the results of three well known configurations of two conductors .
the second column shows the @xmath76 functions , which can be found by laplace s equation ( [ cond fi ] ) and used to calculate @xmath73 with eq .
( [ q_i_sistema_cargas ] ) . .@xmath73 and @xmath77 factors for three systems of two conductors with @xmath78 @xmath79 and @xmath80 .
we neglect edge effects for the cylinders and planes . [ cols="<,^,^",options="header " , ]
we use our approach to study a system with embedded of conductors . in addition , the case of two internal conductors is examined , and we show the limit in which the configuration of two conductors without external conductor is obtained .
these examples show how the properties we have derived can be used to calculate the capacitance coefficients . _
example 1_. consider two concentric spherical shells with radii @xmath81 and @xmath82 and a solid spherical conductor ( concentric with the others ) with radius @xmath78 such that @xmath83 .
the potentials are denoted by @xmath84 , @xmath85 , and @xmath86 respectively .
the general solution of laplace s equation for @xmath76 can be written as @xmath87 from eqs .
( [ cond fi ] ) and ( [ f param ] ) we obtain @xmath77 and @xmath88 @xmath89 @xmath90 although @xmath91 can be obtained the same way , it is easier to extract it from eq .
( [ propf ] ) .
the result is @xmath92 the nine capacitance coefficients can be evaluated explicitly from eq .
( q_i_sistema_cargas ) , but it is easier to use eqs .
( [ prop2 ] ) and ( prop1 ) and to take into account that @xmath93 ( @xmath94 for @xmath95 ) .
we have [ prop_3cap_contenidos ] @xmath96 from eq .
( [ q_i_sistema_cargas ] ) the charge on each conductor is [ sol_3cond_contenidos ] @xmath97hence , we only have to calculate @xmath73 and @xmath98 .
@xcite the result gives @xmath99 if @xmath100 , we find that @xmath101 and @xmath102
. it can be shown that eqs .
( [ prop_3cap_contenidos ] ) and ( sol_3cond_contenidos ) are valid even if the conductors are neither spherical nor concentric , because those equations come from eqs . , , and which are general properties independent of specific geometries . _
example 2_. consider two internal conductors and a grounded external conductor .
as customary , we begin with @xmath103 . by transfering charge from one internal conductor to the other we keep @xmath101 . from eq .
( [ q_i_sistema_cargas.a ] ) and defining @xmath104we find@xmath105where we have used eq .
( [ prop1 ] ) .
similarly @xmath106 , and using again eq .
( [ prop1 ] ) we find@xmath107since the system is neutral @xmath108 and hence@xmath109substituting eq .
( [ fi1v ] ) into eq .
( [ q1c ] ) we obtain@xmath110 because @xmath111 only three of the coefficients in the definition of @xmath112 are independent . from eqs .
we see that this effective capacitance is non negative .
the procedure is not valid if @xmath113 , in that case we see by using eqs .
( [ prop1 ] ) and that @xmath114 , and from eq .
( [ q1c ] ) we find @xmath115 which is also non negative . the limit in which there is no external conductor is obtained by taking all the dimensions of the cavity to infinity while keeping the external conductor grounded as discussed in ref . .
we have used an approach based on laplace s equation to demonstrate that the capacitance matrix depends only on purely geometrical factors .
the explicit use of laplace s equation permits us to demonstrate many properties of the capacitance coefficients .
the geometrical relations and properties shown here permits us to simplify many calculations of the capacitance coefficients .
we emphasize that laplace s equations necessary for finding the capacitance coefficients are purely geometrical as can be seen from eqs . and .
laplace s equation is usually easier than green function formalism for either analytical or numerical calculations .
appendix [ ap : pruebas ] shows some proofs of consistency to enhance the physical insight and the reliability of our method .
a proof of consistency for the identity ( [ prop1 ] ) , is achieved by using eq . ( [ q_i_sistema_cargas ] ) to calculate the total charge on the @xmath0 internal conductors@xcite @xmath116 . \label{qint}\]]we use eq .
( [ prop1 ] ) to find @xmath117note that eq .
( [ qint int ] ) requires many fewer elements of the @xmath24 matrix than eq .
( [ qint ] ) .
this difference becomes more significant as @xmath0 increases .
if we again use eq .
( [ q_i_sistema_cargas ] ) , we can find the charge on the cavity of the external conductor @xmath118and therefore @xmath119a property that can also be obtained from gauss s law . @xcite
another proof of consistency for eq .
( [ prop1 ] ) is found by employing eqs .
( [ q_i_sistema_cargas ] ) and ( [ qint int ] ) to calculate @xmath120 ( taking into account that eq
. comes directly from eq . )
@xmath121 we utilize eq .
( [ fi = fi fi ] ) to write @xmath122 as @xmath123 this relation is clearly correct because @xmath124 points inward with respect to the volume @xmath8 . a proof of consistency for eq .
( [ cij 2 ] ) that shows the symmetry of @xmath125 can be obtained by calculating the electrostatic internal energy , which in terms of the electric field is @xmath126 , \end{aligned}\ ] ] where we have used eq .
( [ fi = fi fi ] ) . from eq .
( [ cij 2 ] ) we find @xmath127 consistent with standard results.@xcite
to enhance the understanding of this approach and its advantages , we give some general suggestions for the reader . 1 .
implement a numerical method to solve the laplace s equation ( cond fi ) for the @xmath76 functions associated with a nontrivial geometry ( for example , two non - concentric ellipsoids ) .
use eqs .
( [ propf ] ) and to either simplify your calculations or to check the consistency of your results . then use eq .
( [ q_i_sistema_cargas ] ) to obtain the @xmath24 factors numerically .
( [ prop2 ] ) and eqs . either to simplify your calculations or to check the consistency of your results .
we have emphasized that to calculate the total charge on the internal conductors eq .
( [ qint int ] ) requires many fewer @xmath24 elements than eq .
( [ qint ] ) .
how many fewer elements are required for an arbitrary value of @xmath0 ?
3 . for a successive embedding of concentric spherical shells ,
calculate the capacitance coefficients for an arbitrary number of spheres .
4 . show that for the successive embedding of three conductors with arbitrary shapes , eqs . and
still hold .
generalize your results for an arbitrary number of conductors .
99 w. taussig scott , _ the physics of electricity and magnetism _ ( john wiley & sons , new york , 1966 ) , 2nd ed . ; gaylord p. harnwell , _ principles of electricity and electromagnetism _ ( mcgraw - hill , new york , 1949 ) ; leigh page and norman i. adams jr . , _ principles of electricity _
( d. van nostrand , new jersey , 1958 ) , 3rd ed . ; a. n. matveev , _ electricity and magnetism _
( mir , moscow , 1988 ) .
@xmath128 is not necessarily the total charge on the external conductor , but the charge accumulated on the surface of the cavity that encloses the other conductors .
the value of the charge is calculated with the surface integral ( [ qigen ] ) , which for the case of the internal conductors encompasses the whole surface , but for the external conductor is only the surface of the cavity that encloses the other conductors .
equation ( [ cij 2 ] ) is an integral over the volume for the @xmath24 factors .
we might be tempted to use gauss theorem to obtain an integral of the volume directly from eq .
( [ q_i_sistema_cargas ] ) .
however , @xmath21 is not defined in the region inside the conductors .
the gradient of @xmath21 in eq .
( [ q_i_sistema_cargas ] ) is evaluated in an external neighborhood of the conductor surface . by uniqueness ,
the solution for this problem is equivalent to the solution for a system consisting of the same @xmath0 conductors contained in the cavity of a surrounding conductor , such that all the dimensions of the cavity tend to infinity , and the potential of the external conductor is set to zero . for a derivation of some of these results based on the energy of the electrostatic field
see l. d. landau , e. m. lifshitz , and l. p. pitaevskii , _ electrodynamics of continuous media _
( elsevier butterworth - heinemann , 1984 ) , 2nd ed . | the fact that the capacitance coefficients for a set of conductors are geometrical factors is derived in most electricity and magnetism textbooks .
we present an alternative derivation based on laplace s equation that is accessible for an intermediate course on electricity and magnetism .
the properties of laplace s equation permits to prove many properties of the capacitance matrix .
some examples are given to illustrate the usefulness of such properties . |
the presence of an electron in front of a solid surface redistributes the charge in the solid . as a consequence ,
an attractive potential is induced , which far from the surface approaches the long - range classical image potential @xmath3 ( @xmath4 being the distance from the surface , @xmath5 , and @xmath6 the static bulk dielectric constant ; for a metal , @xmath7 ) .
if the bulk band structure projected onto the surface presents an energy gap near the vacuum level , an electron located in front of the surface can not propagate into the solid .
therefore , the electron may be trapped in the vacuum well , and an infinite series of rydberg - like states appears , which converges , for zero parallel momentum , towards the vacuum energy .
these so - called image states@xcite are localized in the vacuum region of the surface , the penetration of the first ( @xmath1 ) image wave function into the solid varying typically between @xmath8 and @xmath9 @xcite . as a result ,
image states are almost decoupled from bulk electron scattering , and they are much longer lived than bulk excitations : the lifetime of bulk electrons with energies of @xmath10 ev above the fermi level is approximately one order of magnitude smaller than that of the first image state .
furthermore , the lifetime of higher order image states ( @xmath11 ) has been predicted@xcite to scale asymptotically with @xmath12 , which makes the series to be resolvable . during the last decade
the linewidth of image states has been measured by inverse photoemission@xcite , two - photon photoemission @xcite , and time - resolved two - photon photoemission @xcite .
recently , time - resolved two - photon photoemission has been used in combination with the coherent excitation of several quantum states , and the lifetime of the first six image states on the ( 100 ) surface of copper has been accurately determined@xcite .
theoretical calculations of the linewidth of image states were first reported in refs . and , within a many - body free - electron description of the metal surface and with the use of simplified models to approximate both initial and final electronic states and , also , the screened coulomb interaction .
later on , the decay of the first image state on the ( 111 ) surfaces of copper and nickel metals to the @xmath0 crystal - induced surface state was calculated@xcite , in terms of auger transitions , with the use of a three band model to describe the surface band structure . in ref . , hydrogenic - like states with no penetration into the solid were used to describe the image - state wave functions , a simplified parametrised form was used for the surface - state wave functions , and screening effects were neglected .
self - consistent calculations of the linewidths of image states on copper surfaces have been reported recently@xcite , and good agreement with experimentally determined decay times has been found . in ref . , the linewidths of image states were computed , within the gw approximation of many - body theory@xcite , by going beyond a free - electron description of the metal surface .
single - particle wave functions were obtained by solving the schrdinger equation with a realistic one - dimensional model potential@xcite , and the screened interaction was evaluated in the random - phase approximation ( rpa)@xcite . in this paper
we focus our attention on the role that the crystal - induced surface state ( @xmath0 ) plays in the relaxation of the first image state at the ( 111 ) surface of copper , which we find to represent a @xmath2 of the total linewidth .
we present self - consistent calculations along the lines of ref . , and we also consider simplified models for both the electronic wave functions and the screened coulomb interaction , showing that a detailed description of these quantities is of crucial importance in the understanding of the origin and magnitude of linewidths of image states
. we account for potential variation parallel to the surface through the introduction of the effective mass , and we find that the linewidth of the first image state of cu(111 ) is @xmath13 smaller than in the case of free - electron behavior along the surface .
finally , we investigate the dependence of interband@xcite linewidths on the momentum of the image state parallel to the surface , @xmath14 .
our results indicate that for image state total energies lying below the top of the gap , the linewidth of the first image state of cu(111 ) is increased with @xmath15 up to a @xmath13 .
we assume translational invariance in the plane parallel to the surface , which is taken to be normal to the @xmath4-axis , and we evaluate the inelastic linewidth of the image state @xmath16 with energy @xmath17@xcite ( we use atomic units throughout , i.e. , @xmath18 ) , as the projection of the imaginary part of the electron self - energy , @xmath19 , over the state itself : @xmath20 where @xmath21 represents the two - dimensional fourier transform of @xmath19 . in the gw approximation@xcite
, the self - energy is obtained by just keeping the first term of the expansion in the screened interaction ( @xmath22 ) .
then , after replacing the green function by the zero order approximation , one finds @xmath23 where the sum is extended over a complete set of final states @xmath24 with energies @xmath25 .
@xmath26 is the effective mass , which accounts for the departure of the motion along the surface from free - electron behavior , @xmath27 is the fermi energy , and @xmath28 represents the two - dimensional fourier transform of the induced part of the screened interaction .
in particular , if only transitions into the crystal - induced @xmath0 surface state @xmath29 with energy @xmath30 are considered , one finds @xmath31 and introducing this contribution to @xmath32 into eq .
( [ gamma1 ] ) one finds : @xmath33 three different models have been used for the evaluation of the screened interaction , @xmath22 .
first , the specular - reflection model ( srm ) of ritchie and marusak@xcite has been considered . in this model ,
bulk electrons are assumed to be specularly reflected at the surface , the interference between the ingoing and outgoing waves being neglected , and the electronic charge density abruptly terminates at the surface ( @xmath34 ) , which we choose to be located half a lattice spacing beyond the last atomic layer
. within this simplified model@xcite , also called semiclassical infinite barrier model ( scibm ) , the screened interaction is obtained in terms of the wave - vector and frequency dependent bulk dielectric function , which we evaluate in the rpa . secondly , for the vacuum contribution to the linewidth ( @xmath35 , @xmath36 ) the surface response function suggested by persson and
zaremba@xcite ( pz ) has been used .
finally , the screened interaction has been evaluated , as in ref .
, by solving the rpa integral equation for the density response function of inhomogeneous media in terms of the eigenfunctions of the one - electron effective hamiltonian .
these eigenfunctions have been computed by solving the schrdinger equation with the realistic one - dimensional model potential suggested in ref .. this model potential uses as parameters the width and position of the energy gap at the @xmath37 point ( @xmath38 ) and , also , the binding energies of both the @xmath0 crystal - induced surface state at @xmath37 and the first image state . for the evaluation of @xmath0 and @xmath1 surface states , we have first used simplified models for the wave functions inside and outside the solid . in the vacuum side of the surface ( @xmath35 ) , @xmath1 and
@xmath0 states have been approximated by a parametrised @xmath39-like hydrogenic wave function and a mere exponential , respectively , which have been matched to a decaying wave function in the crystal band - gap ( @xmath40 ) obtained within a nearly free electron two - band model@xcite : @xmath41 here , @xmath0 and @xmath1 correspond to the crystal - induced surface state and the first image state , respectively , @xmath42 is the limit of the brillouin zone in the direction normal to the surface , and @xmath43 where @xmath44 and @xmath45 represent the energy gap and the energy of the @xmath46 surface state with respect to the midgap , respectively .
the phase shift @xmath47 is given by @xmath48 , & if $ \;\;\;0 \leq \eta_n \leq 1 $ \cr \tan^{-1}\left[\sqrt{\frac{1}{\eta^2_n}-1}\right ] , & if $ -1 \leq \eta_n \leq 0,$ \cr}\ ] ] with @xmath49 .
the image state on cu(111 ) is located right at the top of the gap ( @xmath50 ) , both the hydrogenic - like wave function in the vacuum ( @xmath35 ) and the decaying @xmath51-like wave function in the bulk ( @xmath40 ) having , therefore , nodes at the surface ( @xmath34 ) .
the @xmath0 crystal - induced surface state on cu(111 ) is located at the bottom of the gap ( @xmath52 ) ; thus , it is described by a @xmath53-like wave function in the bulk .
these approximate wave functions ( awf ) are exhibited in fig .
[ wave ] , together with the corresponding wave functions that we obtain by solving the schrdinger equation with the one - dimensional potential of ref .
( mwf ) . both wave functions , awf and mwf ,
coincide within the bulk , but the hydrogenic - like wave function for the @xmath1 image state appears to be less localized near the surface than our model wave function .
the @xmath0 and @xmath1 surface states on cu(111 ) have binding energies ( measured with respect to the vacuum level ) of @xmath54 and @xmath55 ev , respectively .
the @xmath1 probability - density has a maximum at @xmath56 a.u . outside the crystal edge ( @xmath34 ) . the penetration into the bulk of @xmath0 and @xmath1 surface states is found to be , at the @xmath57 point , of @xmath58 and @xmath59 , respectively .
the results of our calculations for the linewidth , @xmath60 , coming from the decay of the @xmath1 image state into the @xmath0 intrinsic surface state on cu(111 ) are presented in table [ ttb1 ] , with the momentum of the image electron parallel to the surface , @xmath61 , set equal to zero . here ,
the linewidth has been split as follows : @xmath62 where @xmath63 , @xmath64 and @xmath65 represent vacuum , bulk and interference contributions , respectively , as obtained by confining the integrals in eq .
( [ surf ] ) to either vacuum ( @xmath66 ) , bulk ( @xmath67 ) or vacuum - bulk ( @xmath68 ) coordinates .
first we show our full rpa calculations , in which the screened interaction is obtained on the basis of one - electron eigenfunctions computed from the realistic one - dimensional model potential of ref .. within these calculations both @xmath0 and @xmath1 surface - state wave functions are also obtained from the model potential of ref .
( mwf ) , with either @xmath69 or @xmath70@xcite . within the specular reflection model@xcite and the model suggested by persson and zaremba@xcite for the screened interaction , we have used the mwf @xmath0 and @xmath1 surface - state wave functions as well as the simplified models ( awf ) described in the previous section , with @xmath69 .
total linewidths , @xmath71 , as obtained from the decay of the @xmath1 image state on cu(111 ) into any final state with energy @xmath72 , @xmath73 , ( see eqs .
( 1 ) and ( 2 ) ) , are presented in table ii . here , our full rpa calculations are shown , with all wave functions computed from the one - dimensional model potential of ref .
realistic values of the effective mass of final states have been considered , according to the experiment or to _ ab initio _ band structure calculations . as for the @xmath0 surface state ,
we have used @xmath70 , as in table i , and for bulk states we have chosen to increase the effective mass from our computed value@xcite of @xmath74 at the bottom of the gap to @xmath75 at the bottom of the valence band .
our full rpa calculations indicate that the decaying rate of the @xmath1 image state into the @xmath0 crystal - induced surface state results , for @xmath69 , in a linewidth of @xmath76 mev , while use of the more realistic effective mass @xmath70 leads to a linewidth of @xmath77 . with the use of either the free - electron mass or more realistic effective masses for both bulk and crystal - induced surface states ,
@xmath78 approximately represents a @xmath2 of the total linewidth , @xmath79 ( @xmath75 ) or @xmath80 ( @xmath81 ) .
the more realistic result of @xmath80 for the total linewidth is in good agreement with the experimentally measured lifetime@xcite of @xmath82 at @xmath83@xcite . within the vacuum side of the surface
the @xmath1 image state couples dominantly to the @xmath0 surface state ( this coupling approximately represents a @xmath84 of the total @xmath63 linewidth ) ; however , the coupling of image states with all bulk crystal states occurring through the bulk penetration plays an important role and can not , therefore , be neglected if one is to accurately describe the lifetime of image states .
we note that simplified jellium models for the evaluation of the screened interaction lead to unrealistic results for the contribution of the surface state to the linewidth of image states@xcite .
first , we compare our full rpa calculations ( see table i ) with the results we obtain , also with use of our model initial and final wave functions ( mwf ) , when our realistic screened interaction is replaced by that obtained within the specular reflection model ( srm ) and the model of persson and zaremba ( pz ) .
bulk contributions to the linewidth are approximately well described within the specular reflection model , small differences resulting from an approximate description , within this model , of the so - called _ begrenzung _ effects .
as the approximate treatment of ritchie and marusak@xcite ignores the quantum mechanical details of the surface , this model fails to describe both vacuum and interference contributions to the linewidth .
these quantum mechanical details of the surface are approximately taken into account within the jellium model of ref .
, thus resulting in a better approximation for the vacuum contribution to the linewidth .
discrepancies between vacuum contributions obtained within this model ( pz ) and our more realistic full rpa calculations@xcite appear as a result of the jellium model of ref . being accurate provided @xmath85 and @xmath86 ( @xmath87 is the fermi momentum , i.e. , @xmath88 ) .
in order to investigate the dependence of @xmath89 on the details of both @xmath1 and @xmath0 wave functions , we present in table i calculations , within srm and pz models for the screened interaction , in which our realistic wave functions are replaced by the simplified models ( awf ) described in the previous section .
as the hydrogenic - like wave function used to describe the @xmath1 image state on the vacuum side of the surface presents an image - state charge gravity center localized further away from the surface than our more realistic model wave function , both vacuum and interference contributions to the linewidth are largely underestimated within this approximation .
furthermore , we note that the linewidth is highly sensitive to the details of the image - state wave functions .
this is a consequence of the critical behavior of the imaginary part of the non - local self - energy coupling points near the surface , as we will discuss below .
now we focus on our full rpa calculation of the total linewidth of the @xmath1 image state on cu(111 ) ( see table ii ) , with all effective masses set equal to the free - electron mass .
we show in fig .
[ band]b separate contributions to the linewidth , @xmath71 , coming from the decay into the various @xmath90 bulk crystal states , @xmath91 , such that @xmath92 fig .
[ band]a exhibits the bulk band structure projected onto the ( 111 ) surface of copper .
the arrows indicate the available phase space in the decay of the @xmath1 image state at the @xmath93 point ( @xmath61=0 ) into the unoccupied portion of the @xmath0 surface state and a generic @xmath90 bulk state , which are represented by their characteristic @xmath94 and @xmath95 parabolic dispersions , respectively .
@xmath63 , @xmath64 and @xmath96 contributions to @xmath91 are represented , together with the total contribution , @xmath91 , as a function of @xmath97 .
the lower edge , at the @xmath93 point , of the energy gap projected onto the cu(111 ) surface lies below the fermi level ( @xmath98 ) and , consequently , the decay from the ( @xmath99 ) image state occurs through finite parallel momentum transfer . hence , as the coupling of the image state with the crystal occurring through the tails of bulk states outside the crystal is expected to be dominated by vertical transitions ( @xmath100 ) , vacuum contributions to the @xmath91 linewidth are very small , especially for those bulk states located at the bottom of the valence band ( decay into these states is only allowed for large values of the momentum transfer ; also , their vacuum penetration is small ) . actually , the coupling of the @xmath1 image state with all bulk states taking place at the vacuum side results in a linewidth of only @xmath101 mev , which approximately represents a @xmath102 of the total @xmath63 linewidth @xcite .
a realistic description of motion along the surface can be approximated by introduction of the effective mass , as described above .
the effective mass of all final states with energies @xmath72 , @xmath103 , is found to be smaller than the free - electron mass , thus both @xmath78 and @xmath71 being about @xmath13 smaller than in the case of free - electron behavior along the surface ( see tables i and ii ) .
this is the result of two competing effects : first , there is the effect of the decrease of the available phase space , which is easily found to scale as @xmath104 .
secondly , as the effective mass decreases the decay from the image state occurs , for a given energy transfer , through smaller parallel momentum transfer , which may result in both enlarged and diminished screened interactions , depending on momentum and energy transfers .
this is illustrated in fig .
3 , where the impact of the introduction of the effective mass on the evaluation of the various contributions ( vacuum , bulk , and interference ) to both @xmath91 and @xmath78 ( see eq . ( 9 ) ) is exhibited through the percentage ratio @xmath105 as a function of the final state energy @xmath97 .
if the screened interaction were independent of the parallel momentum transfer , all ratios would scale as @xmath104 , which is represented in fig . 3 as @xmath106 .
instead , as the parallel momentum transfer decreases the screened interaction is predominantly larger , which results in the ratio @xmath107 to be larger than @xmath108 , especially in the case of vacuum contributions to the linewidth which are expected to be dominated by vertical transitions .
when the decay from the image state may occur through very small parallel momentum transfer @xmath109 , [ this is the case of final states that are just below the bottom of the gap and also the case of the @xmath0 surface state ] , a decrease in @xmath109 may result in a diminished screened interaction ( see fig .
4 ) , thus @xmath107 being slightly smaller than @xmath108 for these states . now
, we analyze the behavior of the imaginary part of the image - electron self - energy , which results in vacuum and interference contributions to the linewidth to be comparable in magnitude and opposite in sign ( see table [ ttb1 ] ) .
[ self ] shows full rpa ( solid line ) and srm ( dashed - dotted line ) calculations of @xmath110 ( see eq .
( [ sigmas ] ) ) for cu(111 ) with all effective masses set equal to the free - electron mass , together with the @xmath1 image - state wave function , as a function of the @xmath111-coordinate and for @xmath112 .
we note that the probability for electron - hole pair creation ( the dominant channel for the decay of these states is provided by this process ) is underestimated within the srm .
the coupling between electronic states is well known to be maximum , within the bulk , at the position of the electron . nevertheless , as the electron moves into the vacuum the maximum of the imaginary part of the electron self - energy stays ( see fig . [ self ] ) near the surface ( @xmath34 ) , as demonstrated , within a jellium model of the surface , by deisz _
et al_@xcite . hence , for any given value of @xmath35 , main vacuum and interference contributions to the linewidth are determined ( see eq .
( [ gamma1 ] ) ) by the specific shape of the image - state wave function in regions @xmath113 and @xmath114 of fig .
[ self ] . as the image - state on cu(111 ) is located right at the top of the energy gap and the corresponding wave function has , therefore , a node at the surface ( @xmath115 ) , an inspection of fig .
[ self ] leads us to the conclusion that vacuum and interference contributions to the linewidth are comparable in magnitude and opposite in sign . on cu(100 )
the image state is located close to the center of the gap , and the corresponding wave function has a node at @xmath116 a.u . on this surface ,
total vacuum and interference contributions to the linewidth are still opposite in sign , though interference contributions coming from the decay into states at the bottom of the valence band are now positive due to their minor vacuum penetration .
if the image state were located at the bottom of the gap , matching at the surface would occur at maximum amplitude and the total interference contribution might be positive , as the sign of the image - state wave function would be the same on both sides just around the surface ( @xmath34 ) .
finally , we investigate the dependence of interband@xcite linewidths on the momentum of the image electron parallel to the surface , @xmath14 .
first , we use the mwf image state wave function evaluated at the @xmath57 point and introduce , within this model , the dependence on @xmath14 of the quasiparticle self - energy , as indicated in eq .
we find that for image state total energies lying below the top of the gap ( @xmath117 ) , the linewidth of the first image state of cu(111 ) increases less than @xmath8 .
secondly , we also account for the change of the @xmath4-dependent initial wave function along the dispersion curve of the image state , by solving the schrdinger equation for a one - dimensional model potential that we build following ref . with various values of @xmath118 : @xmath119 , @xmath120 , and @xmath121 .
the penetration into the bulk of the @xmath1 image state with these values of @xmath118 is found to vary from @xmath59 at @xmath122 to @xmath123 , @xmath124 and @xmath125 at @xmath126 a.u .
, @xmath120 a.u . and @xmath127 a.u .
, respectively . our results , as obtained with all effective masses set equal to the free - electron mass , are presented in table iii . though the penetration of the image state wave function increases with @xmath118 , the amplitude of this wave function on the bulk side and near the jellium edge decreases , thus the absolute value of both bulk and interference contributions to the linewidth decreasing with @xmath118
nevertheless , the total overlap between image state and final wave functions becomes more efficient as @xmath118 increases , which results in larger values of the total linewidth .
also , it is interesting to notice that especially sensitive to the variation of the momentum of the image state parallel to the surface is the contribution to the linewidth from damping into the @xmath0 surface state .
we have reported calculations of the inelastic broadening of the first image state at the ( 111 ) surface of copper , and we have investigated , in particular , the role that the intrinsic crystal - induced surface state plays in the decay of this image state .
we have presented self - consistent rpa calculations , by going beyond a free - electron description of the metal surface .
we have also considered simplified models for both the electron wave functions and the screened coulomb interaction , showing that a detailed description of these quantities is of crucial importance in the understanding of the origin of linewidths of image states .
we have accounted for potential variation parallel to the surface through the introduction of the effective mass .
we have analyzed the origin and magnitude of the various contributions to the linewidth .
though the dominant contribution to the decay of the first image state into the crystal - induced surface state comes from the coupling between image and surface states within the vacuum part of the surface , it appears to be approximately canceled out by the contribution from the interference between bulk and vacuum coordinates .
for the vacuum contribution to the decaying rate into the intrinsic surface state , we have found that it approximately represents a @xmath84 of the total vacuum contribution .
we also conclude that the coupling of image states with all bulk crystal states occurring through the bulk penetration plays an important role in the determination of lifetimes , and that this penetration can not be neglected if one is to accurately describe the lifetime of image states .
we have found , within our full rpa scheme , that in the case of @xmath128 the decaying rate of the first image state on cu(111 ) into the intrinsic surface state results , with all effective masses set equal to the free - electron mass , in a linewidth of @xmath76 mev , while use of more realistic effective masses leads to a linewidth of @xmath77 . with the use of either the free - electron mass or more realistic effective masses for both bulk and crystal - induced surface states ,
@xmath78 approximately represents a @xmath2 of the total linewidth , @xmath79 ( @xmath75 ) or @xmath129 ( @xmath81 ) .
the more realistic result of @xmath80 for the total linewidth is in good agreement with recent experimental results reported in ref .. we have investigated the dependence of interband linewidths on the momentum of the image electron parallel to the surface , showing that for image state total energies lying below the top of the gap the linewidth of the first image state increases with @xmath118 up to a @xmath13 .
we conclude that this increase appears mainly as a consequence of the change of the @xmath4-dependent initial wave function with @xmath118 , and our results indicate that the contribution to the linewidth from damping into the @xmath0 surface state is responsible for the dependence of the total linewidth with the momentum parallel to the surface .
the authors gratefully acknowledge a. rubio , e. zarate and m. a. cazalilla for fruitful discussions in connection with this research . this project has been supported by eusko jaurlaritza ( basque country ) , the ministerio de educaci ' on y cultura ( spain ) , and iberdrola s. a .. 20 p. m. echenique and j. b. pendry , j. phys .
c * 11 * , 2065 ( 1978 ) ; p. m. echenique and j. b. pendry , prog .
sci . * 32 * , 111 ( 1990 ) . th .
fauster and w. steinmann , in @xmath130 @xmath131 , edited by p. halevi ( elsevier , amsterdam , 1995 ) , pp .
347 - 411 .
r. m. osgood jr . and x. wang , solid state phys .
* 51 * , 1 ( 1997 ) , and references therein .
e. v. chulkov , v. m. silkin , and p. m. echenique ( to be published ) . f. passek and m. donath , phys .
rev . lett . * 69 * , 1101 , ( 1992 ) .
f. passek , m. donath , k. ertl and v. dose , phys . rev .
b * 75 * , 2746 , ( 1995 ) .
r. w. schoenlein , j. g. fujimoto , g. l. eesley and t. w. capehart , phys .
. lett . * 61 * , 2596 , ( 1988 ) .
s. schuppler , n. fischer , th .
fauster and w. steinmann , phys .
b * 46 * , 13539 , ( 1992 ) .
x. y. wang , x. j. shen , r. m. osgood , jr . ,
r. haight and f. j. himpsel , phys .
b * 53 * , 15738 , ( 1996 ) .
t. hertel , e. knoesel , m. wolf and g. ertl , phys .
lett . * 76 * , 535 , ( 1996 ) .
m. wolf , e. knoesel and t. hertel , phys .
b * 54 * , 5295 , ( 1996 )
. j. d. mcneill , r. l. lingle , jr .
ge , c. m. wong , r. e. jordan , and c. b. harris , phys .
b * 79 * , 4645 , ( 1997 ) .
e. knoesel , a. hotzel and m. wolf , j. electron .
relat . phenom .
* 88 * -*91 * , 577 , ( 1998 ) . i. l. shumay , u. hfer , ch .
reu , u. thomann , w. wallauer , and th .
fauster , phys .
b * 58 * , ( 1998 ) .
u. hfer , i. l. shumay , ch .
reu , u. thomann , w. wallauer and th .
fauster , science @xmath132 , 1480 ( 1997 ) .
m. echenique , f. flores and f. sols , phys .
55 * , 2348 ( 1985 ) .
p. de andr ' es , p. m. echenique and f. flores , phys .
b * 35 * , 4529 ( 1987 )
. s. gao and b. i. lundqvist , prog .
suppl . * 106 * , 405 ( 1991 ) ; s. gao and b. i. lundqvist , solid state commun . * 84 * , 147 ( 1992 ) . e. v. chulkov , i. sarra , v. m. silkin , j. m. pitarke , and p. m. echenique , phys .
rev . lett.*80 * , 4947 ( 1998 ) .
l. hedin and s. o. lundqvist , in @xmath133 , edited by f. seitz , h. ehrenreich , and d. turnbull ( academic , new york , 1969 ) , vol.23 , p.2 .
e. v. chulkov , v. m. silkin and p. m. echenique , surf .
lett . * 391 * , 1217 , ( 1997 ) .
see e.g. , d. pines , _ elementary excitations in solids _ , chapter 3 ( addison wesley , new york , 1963 ) . in these processes
the image electron decays into a state having a different @xmath4-dependent ( the plane of the surface is assumed to be perpendicular to the @xmath4-axis ) wave function .
the effective mass of the image state is taken to be equal to the free - electron mass , i.e. , @xmath134 ( see , e.g. , ref . ) .
r. h. ritchie and a. l. marusak , surf .
sci . * 4 * , 234 ( 1966 ) .
p. m. echenique , f. j. garca de abajo , v. h. ponce and m. e. uranga , nucl .
instr . and meth .
b * 96 * , 583 , ( 1995 ) . b. n. j. persson and e. zaremba , phys .
b * 31 * , 1863 ( 1985 )
. a. goldmann , v. dose and g. borstel , phys .
b * 32 * , 1971 ( 1985 ) .
s. d. kevan , phys .
lett . * 50 * , 526 ( 1983 ) .
s. l. hulbert , p. d. johnson , n. g. stoffel , w. a. royer and n. v. smith , phys .
b * 31 * , 6815 ( 1985 ) . the electron density parameter has been taken to be @xmath135 , so as to reproduce the fermi energy of @xmath136 ev obtained within our model calculation of the ground state .
the parameters entering the surface response function of persson and zaremba@xcite have been taken to be @xmath137 , @xmath138 and @xmath139 .
calculations of the bulk band structure have been projected onto the ( 111 ) surface of copper , and the dispersion @xmath140 along the @xmath141 direction has been fitted to @xmath142 with @xmath143 and @xmath144 at the bottom and the top of the gap , respectively . though the effective mass at the bottom of the gap has been reported@xcite for cu(111 ) to be @xmath145 , our calculation of the lifetime is nearly insensitive to the precise choice of either @xmath143 or @xmath145 of ref . .
s. d. kevan and r. h. gaylord , phys .
b * 36 * , 5809 ( 1987 ) .
the lifetime , @xmath146 , is related to the linewidth , @xmath71 , of eqs .
( [ gamma1 ] ) and ( [ surf ] ) by @xmath147 mev @xmath148 fs .
we find that the impact of the band structure in the evaluation of the screened interaction is not large , and differences between our full rpa calculations , on the one hand , and the srm and pz results , on the other hand , appear as a consequence of simplifications introduced , within a jellium model of the surface , in both srm and pz approaches .
notice that both @xmath149 and @xmath150 contributions to the linewidth present a maximum for energies just below the bottom of the energy gap .
this is due to the fact that the vacuum penetration of the bulk states increases as one goes up in energy and then decreases just below the gap due to the stronger bulk character of these states reproducing the bottom of the gap .
this maximum appears here as a cusp because of the relatively thin slab used in the present calculation
. this sharp feature will be significantly smoother for thicker films .
j. j. deisz , a. eguiluz , w. hanke , phys .
lett . * 71 * , 2793 ( 1993 ) .
.calculated linewidth ( in mev ) coming from the decay of the @xmath1 image state on cu(111 ) into the @xmath0 intrinsic surface state , as obtained within three different models for the description of the surface response and two different models for the description of both initial and final wave functions ( see text ) .
rpa accounts for our full rpa realistic calculation of the screened interaction , srm for the simplified specular reflection model of ritchie and marusak @xcite , and pz for the vacuum side surface response suggested by persson and zaremba @xcite .
mwf accounts for the wave functions obtained by solving the schrdinger equation with the realistic one - dimensional model potential of ref . , and awf for the approximate model described in the text .
the effective mass of the @xmath1 image state has been set equal to the free - electron mass , and for the @xmath0 surface state we have used either @xmath69 or @xmath70 .
the momentum of the image electron parallel to the surface is set equal to zero . [ cols="<,^,>,>,>,>,^",options="header " , ] | the role of the intrinsic surface state ( @xmath0 ) in the decay of the first image state ( @xmath1 ) at the ( 111 ) surface of copper is investigated .
inelastic linewidths are evaluated from the knowledge of the imaginary part of the electron self - energy , which we compute , within the gw approximation of many - body theory , by going beyond a free - electron description of the metal surface .
single - particle wave functions are obtained by solving the schrdinger equation with a realistic one - dimensional model potential , and departure of the motion along the surface from free - electron behaviour is considered through the introduction of the effective mass .
the decay of the first image state of cu(111 ) into the intrinsic surface state is found to result in a linewidth that represents a @xmath2 of the total linewidth .
the dependence of linewidths on the momentum of the image state parallel to the surface is also investigated . |
in an undirected , finite , simple and connected graph @xmath1 the _ ( geodetic ) distance _ between two vertices @xmath2 is defined as the length of a shortest path connecting @xmath3 and @xmath4 , and it is denoted by @xmath5 . for many other possibilities of defining a metric space on a graph
see @xcite .
one of the most investigated distance - based graph invariants is the _ wiener index _ , defined as the sum of distances between all pairs of vertices of graph @xmath1 : @xmath6
it has been introduced by wiener @xcite in 1947 , when the correlations between the boiling points of paraffins and their molecular structure has been discovered and noted that in the case of a tree it can be easily calculated from the edge contributions by the following formula : @xmath7 where @xmath8 and @xmath9 denote the number of vertices in connected components @xmath10 and @xmath11 formed by removing an edge @xmath12 from the tree @xmath13 .
bryant and tupper @xcite , have introduced a generalisation of a metric space called _ diversity _ , as a pair @xmath14 where @xmath15 is a set and @xmath16 is a function from the finite subsets of @xmath15 to @xmath17 satisfying two axioms : * @xmath18 , and @xmath19 if and only if @xmath20 and * if @xmath21 then @xmath22 for all finite @xmath23 .
every diversity has an induced metric , given by @xmath24 for all @xmath25 .
see a forthcoming book @xcite for many other generalisations of finite metrics .
the classical steiner tree problem is defined as follows : given a set of points in a metric space , find the shortest network interconnecting all given points .
such a network is called a _ steiner tree _ on the given set . as a geometrical problem it is in a simpler form dating back already to fermat and later to euler and steiner . in the last five decades
it has attracted a significant research interest , due to many applications as diverse as optical and wireless communication networks , vlsi - layout and the study of phylogenetic trees @xcite .
the steiner diversity of a graph introduced by chartrand et al .
@xcite , and called there and in the later publications steiner distance of a graph , generalises in a natural way the concept of geodetic distance in a graph , see also @xcite . _ the steiner diversity _
@xmath26 among the vertices of @xmath27 is defined as the minimum number of edges of a connected subgraph @xmath13 whose vertex set contains @xmath28 .
note that this set is always a tree , called a steiner tree connecting vertices of @xmath28 . if @xmath29 then @xmath13 is also called a @xmath0-steiner tree on @xmath28 .
vertices from @xmath28 are called _ terminal vertices _ of steiner tree @xmath13 , and all other vertices of @xmath13 are called _ non - terminal vertices _ of @xmath13 . for a natural number @xmath0 , @xmath30 , _ the steiner @xmath0-wiener index _
@xmath31 of a graph @xmath1 is defined as @xmath32 _ the average @xmath0-steiner diversity of a graph @xmath33 _ is defined as @xmath34 , where @xmath35 denotes the number of vertices of @xmath1 .
it has been first studied in @xcite .
an application of the @xmath0-steiner wiener index in chemistry is reported in @xcite , where it is shown that the term @xmath36 provides a better approximation for the boiling points of alkanes than @xmath37 itself , and that the best such approximation is obtained for @xmath38 . _ the betweenness centrality _
@xmath39 of a vertex @xmath40 is defined as the sum of the fraction of all pairs of shortest paths that pass through @xmath4 across all pairs of vertices in a graph : @xmath41 where @xmath42 denotes the number of all shortest paths between vertices @xmath43 and @xmath44 in a graph @xmath1 and @xmath45 denotes the number of all shortest paths between vertices @xmath43 and @xmath44 in graph @xmath1 passing through the vertex @xmath4 . in a case
when a graph models a social or communication network , as the name suggests , it measures the centrality of a vertex in a graph , by the influence of a vertex in the dissemination of information over a network .
it has been independently introduced by anthonisse in @xcite and by freeman in @xcite , and among other applications has been applied to detect communities in networks @xcite .
see @xcite for a recent survey on results and problems on betweenness centrality and related notions .
vertex decomposition of the wiener index has been studied by caporossi et al . in @xcite .
krekovski and gutman followed in @xcite with a nice formula for a vertex decomposition of wiener index in the case of a tree and in a case of general graphs they have provided a formula which connects vertex decomposition of a wiener index with the betweenness centrality . in terms of the average betweenness centrality and the average distance of a graph
this has been first observed by gago in @xcite and published also in @xcite .
the aim of this paper is to generalise results from @xcite to steiner diversity by introducing @xmath0-steiner betweenness centrality .
after introducing a convenient notation in the next paragraph we follow in the second and third section of the paper with formulas for edge and vertex decompositions of the steiner @xmath0-wiener index of a tree . in fourth section a @xmath0-steiner betweenness centrality is introduced and a formula for a vertex decomposition of the steiner @xmath0-wiener index for general graphs is proved . in fifth section the average @xmath46-steiner betweenness centrality of a modular graph is expressed in terms of its wiener index . in the last section we conclude by analogous results as in the fourth section for the total steiner betweenness centrality and the total steiner wiener index
. for a graph @xmath1 let @xmath47 denote the number of its vertices . for a forest ( acyclic graph )
@xmath48 with @xmath49 , @xmath50 , ( connected ) components @xmath51 denote by @xmath52 the sum over all partitions of @xmath0 into at least two nonzero parts of products of combinations distributed among the @xmath49 components of @xmath48 : @xmath53 for a tree @xmath13 we define @xmath54 .
note that by the definition @xmath55 , and @xmath56 whenever @xmath57 .
for @xmath58 the notation @xmath52 coincides with the one used in @xcite , but it does not necessarily coincide in the case when @xmath59 .
for a tree @xmath13 and @xmath60 let @xmath61 denote a graph obtained by removing edge @xmath12 from @xmath13 . note that @xmath61 consists of two connected components .
let @xmath13 be a tree on @xmath35 vertices .
then , @xmath62 the expression @xmath63 counts how many times @xmath12 appears as an edge in a steiner tree connecting @xmath0 vertices of @xmath13 .
since the steiner diversity between @xmath0 vertices equals the number of edges of the @xmath0-steiner tree connecting them , the equality in formula [ steinerwiener theorem ] follows . since for any edge @xmath12 , @xmath61 consists of exactly two components @xmath10 and @xmath11 it follows that @xmath64 and therefore in this case formula [ steinerwiener theorem ] coincides with the one from @xcite , if moreover @xmath58 it coincides with the classical wiener result from @xcite .
for a tree @xmath13 and @xmath65 let @xmath66 denote a graph obtained by removing @xmath4 from @xmath13 .
note that @xmath66 may consists of several components and that their number equals the degree of @xmath4 .
let @xmath13 be a tree on @xmath35 vertices .
then , @xmath67 the expression @xmath68 counts how many times is @xmath4 a non - terminal vertex of a steiner tree connecting @xmath0 vertices of @xmath13 .
since the steiner diversity between @xmath0 vertices is by @xmath69 greater than the number of non - terminal vertices in the corresponding @xmath0-steiner tree , adding @xmath69 for each set of @xmath0 vertices , we get the sum of @xmath0-steiner diversities between all @xmath0 sets of vertices , and the equality in formula [ vertexsteinerwiener ] follows many summands in @xmath70 might be equal to zero , since for a pendent vertex @xmath4 of a tree @xmath13 we have that @xmath66 consist of only one component and therefore by the definition @xmath71 and the same holds also when @xmath72 for some @xmath73 , @xmath74 .
_ the @xmath0-steiner betweenness centrality _ @xmath75 of a vertex @xmath40 is defined as the sum of the fraction of all @xmath0-steiner trees that include @xmath4 as its non - terminal vertex across all combinations of @xmath0 vertices of @xmath1 : @xmath76 where @xmath77 denotes the number of all steiner trees between vertices of @xmath78 in a graph @xmath1 and @xmath79 denotes the number of all steiner trees between vertices of @xmath78 in a graph @xmath1 that include also the vertex @xmath4 as a non - terminal vertex .
the term `` steiner '' is included in the above definition to separate our notion from the previously studied @xmath0-betweenness centrality from @xcite , where @xmath0 bounds the length of shortest paths between a pair of vertices .
let @xmath1 be a connected graph on @xmath35 vertices .
then , @xmath80 for a steiner tree @xmath13 on @xmath35 vertices and @xmath81 edges it holds that @xmath82 .
let @xmath83 and @xmath84 denote the number of terminal and non - terminal vertices respectively .
if follows that @xmath85 .
therefore for @xmath86 , and for any steiner tree @xmath13 , with @xmath84 non - terminal vertices , whose set of terminal vertices equals @xmath78 , it follows that @xmath87 .
the sum @xmath88 counts all possible steiner @xmath0-trees between elements of @xmath78 , which equals @xmath89 , with multiplicity of the number of their internal vertices , which equals @xmath90 .
hence it follows that @xmath91 .
moreover @xmath92 for @xmath58 the formula [ maintheorem ] expresses wiener index of graph in terms of the sum of betweenness centralities of its vertices , the main result from @xcite . for a graph @xmath1 on @xmath35 vertices _ the average @xmath0-steiner betweenness @xmath93 _
is defined as @xmath94 let @xmath1 be a connected graph on @xmath35 vertices .
then , @xmath95 for @xmath58 equality [ averagebk ] becomes formula for the average betweenness centrality expressed in terms of the average distance of a graph , which has been observed in @xcite .
the wiener index of any graph @xmath1 can be computed in polynomial time in terms of the number of vertices , and for some graph classes , like trees and benzenoid systems , even in linear time @xcite .
the problem _ steiner tree in graphs _ is described as : given a graph @xmath96 , a weight @xmath97 ( a positive integer ) for each @xmath98 , a subset @xmath99 and a positive integer @xmath100 , is there a subtree of @xmath1 that includes all the vertices of @xmath78 and such that the sum of the weights of the edges in the subtree is no more than @xmath100 ?
this problem belongs to classical np - complete problems from @xcite , and it remains np - complete even if all edge weights are equal , and even for a graph class of hypercubes @xcite .
hence for @xmath101 it is not very likely to find an efficient way to compute steiner @xmath0-wiener index and @xmath0-steiner betweenness centrality for general graphs .
therefore it becomes interesting to either look for lower and upper bounds , approximations in general case and exact results in particular cases . in this section
we do the later .
a graph @xmath1 is _ modular _ if for every three vertices @xmath102 there exists a vertex @xmath103 that lies on a shortest path between every two vertices of @xmath104 , if moreover this vertex is unique then @xmath1 is called a _ median graph _ @xcite .
it is easy to see that all modular graphs are bipartite .
examples of modular graphs are trees , hypercubes , grids , complete bipartite graphs , etc .
the simplest examples of non - modular are cycles on @xmath35 vertices , for @xmath105 .
the following theorem has been observed for trees in @xcite and generalised to median graphs in @xcite and to modular graphs in @xcite .
let @xmath1 be a modular graph on @xmath35 vertices .
then , @xmath106 combining this with the collorary [ averagebk ] we get the following formula for average 3-steiner betweenness centrality of a modular graph .
let @xmath1 be a modular graph on @xmath35 vertices .
then , @xmath107 hence for a modular graph one can compute in polynomial time the average 3-steiner betweenness centrality . for @xmath108-dimensional hypercube @xmath109
it is well known that @xmath110 and @xmath111 .
therefore by using formula [ averageb3 ] we get that the average @xmath46-steiner betweenness centrality of @xmath108-dimensional hypercube equals @xmath112 . since hypercubes are vertex transitive graphs it follows that for each vertex @xmath113 it holds that @xmath114 .
motivated by applications of betweenness centrality in social network theory , one might be interested in all steiner trees that include a given vertex as a non - terminal vertex .
therefore we define _ the total steiner betweenness centrality _ ( or simply steiner betweenness centrality ) @xmath115 of a vertex @xmath40 as the sum of the fraction of all steiner trees that include @xmath4 as its non - terminal vertex across all combinations of vertices of @xmath1 : @xmath116 where @xmath77 denotes the number of all steiner trees between vertices of @xmath78 in a graph @xmath1 and @xmath79 denotes the number of all steiner trees between vertices of @xmath78 in a graph @xmath1 that include also the vertex @xmath4 as a non - terminal vertex . _ the average steiner betweenness @xmath93 _
is defined as @xmath117 .
_ total steiner wiener index _ ( or simply steiner wiener index ) @xmath118 of a graph @xmath1 is defined as @xmath119 set with @xmath35 elements has @xmath120 subsets with cardinality at least two .
therefore we define _ the average steiner diversity of a graph @xmath121 _ as @xmath122 , where @xmath35 denotes the number of vertices of @xmath1 .
let @xmath1 be a connected graph on @xmath35 vertices .
then , @xmath123 using theorem [ maintheorem ] and the equality @xmath124 it follows that , @xmath125 the following corollary then immediately follows .
let @xmath1 be a connected graph on @xmath35 vertices .
then , @xmath126
this work was supported by the deutsche forschungsgemeinschaft within the eurocores programme eurogiga ( project gregas ) of the european science foundation , while the author was working at the department of bioinformatics , university of leipzig , and by max planck institute for mathematics in the sciences , leipzig .
s. gago , j. hurajov ' a , t. madaras , betweenness centrality in graphs , in : m. dehmer , f. emmert - streib ( eds . ) , quantitative graph theory : mathematical foundations and applications , crc press , boca raton ( 2014 ) 233257 .
garey , d.s .
johnson , computers and intractability a guide to the theory of np - completeness , freeman , san francisco , 1979 .
m. girvan , m. e. j. newman , community structure in social and biological networks , proc .
usa 99 ( 2002 ) 78217826 .
k. jiang , d. ediger , d.a .
bader , generalizing @xmath0-betweenness centrality using short paths and a parallel multithreaded implementation , in parallel processing , 2009 , icpp 09 - the 38th international conference on parallel processing vienna , austria , p. 542549 . | a formula based on a vertex contributions of the steiner @xmath0-wiener index is induced by a newly introduced @xmath0-steiner betweenness centrality , which measures the number of @xmath0-steiner trees that include a particular vertex as a non - terminal vertex .
this generalizes krekovski and gutman s vertex version of the wiener theorem and a result of gago on the average betweenness centrality and the average distance in general graphs . |
the historical roots of the subject under investigation go back to the classification of the locally symmetric spaces in the ( proper ) riemannian case by cartan @xcite . _
locally symmetric _ pseudo - riemannian manifolds are characterized by the condition @xmath0 , where @xmath1 denotes the curvature tensor of the manifold .
the pseudo - riemannian manifolds satisfying @xmath2 constitute a logical generalization of the locally symmetric ones .
but in the riemannian case these two conditions are equivalent . moreover , in the riemannian case : @xmath3 a result based on the orthogonal de rham decomposition @xcite . in this riemannian case
the natural generalization are the semi - symmetric spaces ( introduced in @xcite , studied in @xcite ) defined by the relation @xmath4 for all vector fields @xmath5 , where @xmath6 is the curvature operator @xmath7}$ ] . in the lorentzian case ( [ 1 ] ) does not hold , because of the failure of the orthogonal de rham decomposition @xcite .
in fact , @xmath8 , where the metric @xmath9 is extended to the inner product between tensor fields in the standard fashion .
so in the lorentzian case we can introduce the _ second - order symmetric _
( also called _
2-symmetric _ ) spaces @xcite , i.e. , the spaces that satisfy the condition @xmath2 .
the following relations hold : @xmath10 in this contribution we present a specific result from @xcite .
namely , we find the 2-symmetric non - symmetric 4-dimensional spacetimes : the explicit local form of the metric , classification and properties
. the main result is summarized in the following theorem : [ t3 ] a 2-symmetric non - symmetric 4-dimensional spacetime @xmath11 is locally isometric to @xmath12 endowed with the metric @xmath13 where @xmath14 for some constants @xmath15 with @xmath16 .
the starting point for this proof is theorem 4.2 in @xcite . from this theorem
follows that a 2-symmetric non - symmetric spacetime must have a null parallel vector field and therefore its metric is of brinkmann type : @xmath17 where @xmath18 , @xmath19 and @xmath20 are functions independent of @xmath21 , otherwise arbitrary . using the 2-symmetry condition
, one can prove that @xmath22 must be locally symmetric on each slice @xmath23 @xcite .
since the only locally symmetric 2-dimensional riemannian spaces are the constant curvature ones , we have that in appropriate local coordinates @xmath24 in fact , @xmath25 must vanish . to prove this
, we use the petrov classification @xcite .
by @xcite we know that all the semi - symmetric spacetimes ( of dimension @xmath26 ) are of type @xmath27 , @xmath28 or @xmath29 .
moreover , as a consequence of this result , it was also proven that the 2-symmetric spacetimes are of petrov type @xmath28 , or its degenerate case , @xmath29 . and
this only happens when @xmath30 .
hence , the line element of a 2-symmetric non - symmetric 4-dimensional spacetime @xmath11 simplifies to @xmath31 from now on , an overdot means derivative with respect to @xmath32 , @xmath33 and @xmath34 . using the cartan method ( see , for example , @xcite ) to calculate the connection 1-forms and the curvature 2-forms for the null coframe @xmath35 and solving the 2-symmetry equations in this frame ,
we conclude that there exist functions @xmath36 ,
@xmath37 , @xmath38 , @xmath39 and @xmath40 such that @xmath41 @xmath42 @xmath43 where @xmath44 , and the following equations must be satisfied : @xmath45 @xmath46 @xmath47 @xmath48 @xmath49 the proof can then be completed by noticing that eqs.([ww2 ] ) , ( [ ww3 ] ) and ( [ h3 ] ) together with ( [ 2s1])([2s5 ] ) permit the coordinate change needed to arrive at the final form of the metric .
this change of coordinates is given by @xmath50 @xmath51 @xmath52 @xmath53 it is easily verified that the functions @xmath54 , @xmath55 , @xmath56 and @xmath57 can be chosen such that the line - element becomes @xmath58 where @xmath59 for some new functions @xmath60 , because the integrability conditions of the resulting differential equations are always met by virtue of equations ( [ ww2])([2s5 ] ) . to finalize the proof of theorem [ t3 ] , one simply has to solve the equations ( [ 2s1])-([2s5 ] ) appropriately restricted to the new form of the metric ( [ me3 ] ) with ( [ h5 ] )
this readily implies that @xmath61 , @xmath62 and @xmath63 must be linear functions : @xmath64 the condition @xmath16 must be enforced as otherwise @xmath0 and the spacetime would be locally symmetric .
we would like to stress that , even though the family of metrics obtained in therorem [ t3 ] depend on six parameters , a further analysis shows that only four of them are essential parameters , and the other two can be removed by the remaining freedom of coordinates .
similarly , we want to remark that our result is local .
a global classification is also feasible , though some global hypotheses on the spacetime seem unavoidable to that end .
such kind of hypotheses may involve the topology of the manifold , geodesics or causality ( see @xcite for a study of the last two issues in some wave - type spacetimes ) .
in particular , it seems that the local solutions obtained in theorem [ t3 ] are also the unique global 4-dimensional solutions under the requirements of simple connectedness and geodesic completeness .
an open question is whether or not the above pattern is maintained for higher - order symmetric 4-dimensional spacetimes .
specifically , we wonder if the @xmath65-symmetric spacetimes ( those with @xmath66 ) are given , for general @xmath65 , precisely by the line - element ( [ me3 ] ) with ( [ h5 ] ) where the functions @xmath61 , @xmath62 and @xmath63 are polynomials of degree @xmath65 .
a more important open question is the extension of the result to higher dimensions . of course
, many more possibilities appear in this general case , which are being analyzed in @xcite .
jmms is supported by grants fis2004 - 01626 ( micinn ) and giu06/37 ( upv / ehu ) .
ofb and ms are partially supported by grants p06-fqm-01951 ( j. andaluca ) and mtm2007 - 60731 ( spanish mec with feder funds ) .
9 blanco o f , snchez m and senovilla j m m 2010 2nd - order symmetric lorentzian manifolds ii : classification and explicit local form , in preparation cartan 1926 sur une classe remarquable despaces de riemann _ bull .
france _ * 54 * 214 - 64 cartan 1927 sur une classe remarquable despaces de riemann _ bull .
france _ * 55 * 114 - 34 cartan _
leons sur la gomtrie des espaces de riemann _
2nd edn ( paris : gauthier - villars ) chandrasekhar s _ the mathematical theory of black holes _
( oxford university press , oxford , 1983 ) candela a m , flores j l and snchez m 2003 on general plane fronted waves .
geodesics _ gen .
_ * 35 * 631 - 649 .
eriksson i and senovilla j m m 2010 note on ( conformally ) semi - symmetric spacetimes _ class .
quantum grav . _
* 27 * 027001 ( _ arxiv:0908.3246 _ ) flores j l and snchez m 2003 causality and conjugate points in general plane waves _ class .
gravit . _ * 20 * 2275 - 2291 flores j l and snchez m 2008 the causal boundary of wave - type spacetimes _ j. high energy phys .
_ no . 3 * 036 * 43 nomizu k and ozeki h 1962 a theorem on curvature tensor fields _ proc
usa _ * 48 206 - 7 * senovilla j m m 2008 2nd - order symmetric lorentzian manifolds i : characterization and general results , _ class .
quantum grav . _ * 25 * 245011 stephani h , kramer d , maccallum m , hoenselaers c and herlt e , _ exact solutions of einstein s field equations _ , 2nd ed .
( cambrigde university press . ,
cambrigde , 2003 ) szab z i 1982 structure theorems on riemannian spaces satisfying @xmath67 : i. the local version _ j. differ .
* 17 * 531 - 82 szab z i 1985 structure theorems on riemannian spaces satisfying @xmath67 : i. the global versions _ geom . dedicata _
* 19 * 65 - 108 tanno s 1972 curvature tensors and covariant derivatives _ ann . mat .
pura appl . _
* 96 * 233 - 41 wu h 1964 on the de rham decomposition theorem _ illinois j. math . _ * 8 * 291 - 311 wu h 1964 decompositon of riemannian manifolds _ bull . am .
* 70 * 610 - 7 wu h 1967 holonomy groups of indefinite metrics _ pacific j. math . _ * 20 * 351 - 92 | we present the explicit local form of the metric of the second - order symmetric non - symmetric 4-dimensional lorentzian manifolds .
they turn out to be a specific subclass of plane waves . |
type i x - ray bursts are thermonuclear flashes on the surface of weakly - magnetic ( @xmath0 ) accreting neutron stars in low mass x - ray binaries ( lmxbs ) ( lewin , van paradijs , & taam 1995 ; bildsten 1998 ) .
observations with the rossi x - ray timing explorer ( rxte ) have revealed nearly - coherent oscillations during type i x - ray bursts from ten lmxbs , with frequencies in the range 270620 hz ( for reviews , see van der klis 2000 ; strohmayer 2001 ) . these oscillations are interpreted as rotational modulation of surface brightness asymmetries , thus providing a direct measurement of the neutron star spin frequency ( strohmayer et al .
the inferred rotation periods of a few milliseconds support the idea that the neutron stars in lmxbs are the progenitors of the millisecond radio pulsars ( see bhattacharya 1995 for a review ) .
the large modulation amplitudes , high coherence , and frequency stability provide convincing evidence that the burst oscillation frequency is related to the neutron star spin frequency ( strohmayer 2001 and references therein )
. however , many puzzles remain .
open questions include ( i ) the origin of the asymmetry on the surface of the neutron star , particularly during the fading tail of the burst , when the burning is thought to have spread over the whole surface ; ( ii ) why oscillations are seen in some bursts but not others , and in some sources but not others ( e.g. , muno et al .
2001 ) ; and ( iii ) whether the oscillation frequency is the spin frequency or twice the spin frequency , motivated by comparisons of the burst oscillation frequency to the difference between the frequencies of the khz quasi - periodic oscillations ( qpos ) seen in the persistent emission ( van der klis 2000 ) . the oscillation frequency increases by a few hz during the burst .
to explain this , strohmayer et al .
( 1997 ) proposed that the burning shell decouples from the neutron star , and undergoes spin changes due to angular momentum conservation as it expands and contracts .
cumming & bildsten ( 2000 , hereafter cb ) studied the rotational evolution of the neutron star atmosphere during a burst .
rather than addressing the complex problems of how the burning front spreads over the neutron star surface during the burst rise , or what causes the asymmetry at late times , cb asked what could be learned from detailed models of the vertical structure of the atmosphere .
they computed the vertical hydrostatic expansion during the burst , and the resulting spin changes .
they noted that to see a single , coherent frequency requires either the atmosphere be rigidly rotating , or the burning region to be vertically thin , so that the differential rotation across it is small ( see also miller 2000 ) . in order to compare the theoretical results with observations
, cb assumed that the burning shell rotates rigidly .
they found rough agreement between the calculated spin down and observed values .
the composition of the burning shell affects the amount of expansion . for accretion rates relevant to x - ray bursters , the accreted hydrogen ( h ) burns via the hot cno cycle as matter accumulates on the neutron star surface .
the thermally unstable helium ( he ) burning which leads to the type i x - ray burst may occur either before or after the hydrogen has been consumed ( see bildsten 1998 for a review ) , leading to either mixed hydrogen / helium or pure helium ignitions .
the presence of a substantial amount of hydrogen leads to a factor of two greater expansion because of the lower mean molecular weight ( @xmath1 for pure he ; @xmath2 for solar composition ) .
cb noted that the largest frequency shifts observed ( fractional frequency shifts @xmath3% ) were difficult to achieve with pure he ignitions , requiring temperatures at the base of the burning shell extremely close to the limiting value from radiation pressure at the start of the burst .
however , they also noted that no burst oscillations have been seen during type i x - ray bursts with long ( @xmath4 ) cooling tails , characteristic of hydrogen burning ( which is limited by beta decays during the rp process ; bildsten 1998 ) , perhaps indicating a preponderance of he bursts amongst those which show oscillations .
thus they suggested that understanding the largest frequency shifts observed as being due to hydrostatic expansion together with rigid rotation might prove problematic .
recently , two sources have shown extremely large frequency shifts .
wijnands , strohmayer , & franco ( 2001 ) found an increase of @xmath5 in the 567 hz burst oscillation from mxb 1658 - 298 .
in addition to being a large frequency shift ( @xmath6% ) , the spin evolution in this burst was unusual , with a rapid frequency increase occuring several seconds into the burst tail .
this frequency shift is a little larger than the rigidly - rotating mixed h / he models of cb .
galloway et al .
( 2001 ) discovered a 270 hz burst oscillation from the dipping source 4u 1916 - 053 . during 4 seconds in the burst decay ,
the frequency increased by 3.6 hz , a fractional shift of 1.3% , by far the largest fractional frequency shift observed so far . comparing with the results of cb , galloway et al .
( 2001 ) pointed out that the expansion implied by this frequency shift is larger than expected if the atmosphere rotates rigidly , especially given that the measured peak flux of the x - ray burst is sub - eddington ( @xmath7 ) , and that the orbital period of this source ( @xmath8 ; walter et al .
1982 ; white & swank 1982 ; grindlay et al . 1988 ) implies a hydrogen - poor companion ( nelson , rappaport , & joss 1986 ) and therefore a small fraction of hydrogen present at ignition . in this paper , we present new calculations of the hydrostatic expansion and resulting spin - down of the neutron star atmosphere . in 2 , we describe an error in cb s calculation of the change in the moment of inertia of the burning shell , and show that the values of spin - down calculated by cb are a factor of two too large . comparing the new results with observations , we show that , even when hydrogen is present at ignition , _ the theoretical values of spin - down assuming rigid rotation are a factor of three less than the largest observed frequency shifts_. we then go on to consider different neutron star masses and equations of state . in 3 , we derive the angular momentum conservation law in general relativity , presenting analytic results for slow rotation , and numerical calculations of rapidly - rotating stars .
we compare our results with recent work by heyl ( 2000 ) and abramowicz , kluzniak , & lasota ( 2001 ) . in 4 , we show how to include general relativity and rapid rotation in the equations describing the structure of the atmosphere .
we then rescale our fiducial results of 2 to different neutron star masses and equations of state , before presenting some detailed models and rotational profiles .
we again show that the expected spin - down due to angular momentum conservation is smaller than the observed frequency drifts if the atmosphere rotates rigidly .
we discuss the implications of our results in 5 .
in this section , we first show that cb made an error of a factor of two in their calculation of the change in the moment of inertia during a type i x - ray burst , and then compare the new spin down calculations with observations . consider a spherical shell at the surface of the star .
( the arguments of this section are unaffected by including the centrifugal distortion of the isobaric surfaces . )
the mass of the shell is @xmath9 and its moment of inertia is @xmath10 where the factor of @xmath11 comes from integrating over angles .
following cb , we adopt a plane - parallel approximation for the thin atmosphere , and write @xmath12 , keeping only first order terms in @xmath13 . changing integration variables from radius @xmath14 to column depth @xmath15 , where @xmath16 , equation ( [ eq : i ] ) reduces to cb s equation ( 19 ) for the moment of inertia .
if the angular momentum of the atmosphere , @xmath17 , is conserved , the spin - down is given by comparing its moment of inertia before the burst with that during the burst , @xmath18 .
the error in cb s calculation of the change in moment of inertia arises because they assume that the total column depth of the atmosphere is constant during the burst .
we now show that this is not the case when the mass of the layer is conserved .
we write the mass of the shell as @xmath19\ dy=4\pi r^2\delta y\left(1+{2{\langle z\rangle}\over r}\right),\ ] ] where @xmath20 is the total column depth , and we define a mass - weighted thickness @xmath21 .
if the vertical extent of the layer @xmath22 changes , so must its total column depth @xmath20 if the total mass @xmath23 is conserved .
physically , a thin spherical shell has a greater surface area if it moves radially outwards , and so must have a lower column depth in order to conserve its total mass . by taking the column depth at the base during the burst to be the same as that before the burst , cb unwittingly increased the mass of the layer during the burst , and so overestimated the change in its moment of inertia . rather than including the change in the column depth of the layer explicitly
, we write the mass element in equation ( [ eq : i ] ) as @xmath24 , so that there is a unique correspondence between mass and column depth .
the moment of inertia is then @xmath25\ dy,\ ] ] where the integration limits are now held fixed during the burst , thereby keeping the mass of the layer constant , @xmath26 .
we have checked the validity of this expression by integrating the full spherical equations for this problem . in this paper
, we adopt a plane - parallel approximation because it allows us to straightforwardly incorporate the general relativistic angular momentum conservation law .
the spin changes during the burst result from changes in the vertical extent of the atmosphere , which are described by the second term of equation ( [ eq : right ] ) . equation ( 19 ) of cb has a factor @xmath27 rather than @xmath28 , so that the spin - down calculated by cb is too large by a factor of 2 .
specifically , both the spin - down of the convection zone shown in figures 2 and 6 of cb , and the spin - down of a rigidly - rotating atmosphere shown in figure 13 of cb , should be reduced by a factor of 2 .
the spin - down of the radiative layers shown in figures 2 , 4 and 6 of cb is , however , calculated correctly , since in that case cb assumed @xmath29 was constant for each spherical shell .
figure [ fig : newsumm ] shows the spin - down calculated using equation ( [ eq : right ] ) for the moment of inertia , presuming the whole atmosphere rotates rigidly during the burst .
this is an updated version of figure 13 of cb , and detailed discussion of the theoretical models may be found in that paper .
the calculated spin down is exactly a factor of two less than found by cb .
we take @xmath30 , the newtonian gravity for a 1.4 @xmath31 , 10 km star .
we consider two pre - burst models , mixed h / he ignition at @xmath32 ( solid lines and squares ) , and pure he ignition at @xmath33 ( dotted lines and open squares ) .
figure [ fig : newsumm ] shows results for both convective and radiative models . for the radiative models , we show the spin down as a function of flux , since the radiative solution is uniquely determined by the flux .
the convective models depend both on the flux and the extent of the convection zone ; we plot the spin down resulting from the convective models given in table 3 of cb .
the convective models have a composition profile the same as that prior to the burst .
the radiative models have either a pre - burst composition ( upper solid and dotted curves ) or a composition of heavy elements ( lower solid and dotted curves ) , either @xmath34ni for pure he burning , or @xmath35kr for mixed h / he burning ( we choose @xmath35kr as representative of the products of rp process h burning ; see schatz et al .
1998 ; koike et al . 1999 ;
schatz et al . 2001 ) .
as described in cb , the radiative and convective models with a pre - burst composition are appropriate for the burst rise , whereas the radiative atmospheres composed of heavy elements are appropriate for the cooling tail of bursts .
convective models spin down more than radiative models because the steeper temperature gradient in the convection zone leads to a larger base temperature for a given flux . on the left of figure 1 ,
we plot the observed frequency drifts as horizontal bars . a summary of these observations may be found in cb , table 1
; in addition , we include the frequency shifts recently reported for 4u 1916 - 053 ( galloway et al .
2001 ) and mxb 1658 - 298 ( wijnands et al .
for each source , we plot the largest observed frequency shift .
we plot as upper limits those cases for which a frequency was only seen in the burst tail .
the largest observed frequency shift is larger than the theoretical values by at least a factor of 3 .
in this section , we derive the angular momentum conservation law in general relativity . our approach is to calculate the conserved angular momentum of a particle moving on a circular path ( in the appendix , we show that the same result is obtained starting with the conservation equations for a fluid ) .
we show analytic results for slowly rotating stars , and present numerical calculations of rapidly rotating stars .
we compare our results with recent calculations by heyl ( 2000 ) and abramowicz et al .
( 2001 ) . in newtonian physics ,
the specific angular momentum of a particle is given by @xmath29 , where @xmath14 is the distance from the rotation axis .
if the particle s motion conserves angular momentum , the change in its spin frequency with radius is given by @xmath36 .
in this section , we derive the equivalent expression in general relativity .
we start by discussing the angular momentum of a particle moving in the schwarzschild metric .
we then derive the conserved angular momentum in the slow rotation approximation , i.e. neglecting terms of order @xmath37 , where @xmath38 , allowing us to write an analytic expression for the angular momentum conservation law .
however , since @xmath37 may be quite large ( for example , @xmath39 for a star with @xmath40 and @xmath41 spinning at @xmath42 ) , we conclude this section with detailed numerical calculations of rapidly rotating models .
consider first a spherical star , for which the metric is @xmath43 where outside the star the metric functions @xmath44 and @xmath45 are given by @xmath46 where @xmath23 is the total gravitational mass .
note that in this paper we write the ratio @xmath47 as @xmath48 for clarity , but otherwise explicitly write in the constants @xmath49 and @xmath50 .
a particle moving in a circular path has four - velocity ( e.g. , hartle 1970 ) @xmath51 where @xmath52 is the angular velocity measured by an observer at infinity , @xmath53 is the redshift factor between the particle and the observer at infinity , and @xmath54 and @xmath55 are killing vectors of the spacetime .
the redshift factor is @xmath56 where we choose the standard normalization @xmath57 .
associated with the killing vectors @xmath54 and @xmath55 are the conserved quantities @xmath58 and @xmath59 , the specific energy and angular momentum . using equations ( [ eq : fourvel ] ) and ( [ eq : redshift ] ) , we obtain ( see ,
e.g. schutz 1990 , 7.4 ) @xmath60 and @xmath61 differentiating equation ( [ eq : l ] ) and evaluating it at the surface of the star , we find @xmath62 when a particle moves such that angular momentum is conserved . in the limit @xmath63 , this formula reduces to the newtonian result .
it is simple to extend this analysis to the case of slow rotation , for which the metric is ( hartle 1967 ) @xmath64 where @xmath44 and @xmath45 have the same solutions as the spherical case outside the star , and where @xmath65 outside the star ( @xmath66 is the total angular momentum of the star ) . again solving for the conserved angular momentum ,
we find @xmath67 giving @xmath68 \left(1-{2m\over r}\right)^{-1},\ ] ] at the surface of the star , where we presume the particle is initially rotating with the spin of the star , and write the moment of inertia @xmath69 .
for the rapidly - rotating case , we consider the general metric for an axisymmetric stationary star , @xmath70 where the metric potentials @xmath71 and @xmath72 depend only on the coordinates @xmath73 and @xmath74 .
the coordinate @xmath73 is related to the schwarzschild coordinate @xmath14 by @xmath75 , so that @xmath76 is the circumference of a circle with constant @xmath73 and @xmath74 . for further details about the interpretation of these coordinates and potentials ,
see friedman , ipser & parker ( 1986 ) and morsink & stella ( 1999 ) . given an equation of state
, a mass and a rotation rate , the metric ( [ rapid ] ) can be solved numerically .
we use a code written by n. stergioulas , which assumes rigid rotation and is based on methods developed by komatsu , eriguchi & hachisu ( 1989 ) and cook , shapiro & teukolsky ( 1994 ) .
we present the properties of some rapidly - rotating models in table 1 . we consider two different equations of state ( eos ) , eos l which is stiff and eos apr which is softer .
eos l ( pandharipande & smith 1975 ) is one of the stiffest eos in the arnett & bowers ( 1977 ) catalogue .
eos apr is the model a18+@xmath77+uix * computed by akmal , pandharipande & ravenhall ( 1998 ) which uses modern nucleon scattering data and first order special relativistic corrections . for each eos
, we consider two different masses , @xmath78 and @xmath79 , and two spin frequencies , @xmath80 and @xmath81 .
the last column in table 1 gives the break - up spin frequency in each case .
we work in the equatorial plane , and do not consider the variations of the metric potentials with latitude .
the specific angular momentum is @xmath82 ( see also morsink & stella 1999 ) where the three - velocity of a corotating particle as measured by a zero angular momentum observer is given by @xmath83.\ ] ] again holding @xmath84 constant as we vary the schwarzschild coordinate radius @xmath14 , we find @xmath85,\ ] ] where @xmath86 is the equatorial radius of the star , subscripts denote partial differentiation and all quantities are evaluated at the equator . in the slow rotation limit , equations ( [ eq : fastl ] ) and ( [ eq : fast ] ) reduce to equations ( [ eq : slowl ] ) and ( [ eq : slow ] ) .
heyl ( 2000 ) recently calculated @xmath87 , with a rather different result ( compare our eq .
[ [ eq : slow ] ] with his eq .
however , abramowicz et al .
( 2001 ) point out a sign error in heyl s calculation , and , more importantly , that heyl assumes that the quantity @xmath88 is conserved , as , for example , a particle in orbit ( abramowicz & prasanna 1990 ) .
however , the energy @xmath89 of a fluid element in the atmosphere is not conserved during the burst . the correct conserved quantity , which we have considered in this section , is the angular momentum per particle @xmath84 . in the slow rotation approximation , our result ( eq .
[ [ eq : slow ] ] ) agrees with equation ( 9 ) of abramowicz et al .
( 2001 ) .
we write @xmath87 as @xmath90 , where @xmath91 is unity in the newtonian limit .
we give the value of @xmath92 in table [ tab : models ] .
figure [ fig : beta ] shows @xmath92 as a function of neutron star mass for eos apr and eos l , and for @xmath93 and @xmath81 .
we see that including the general relativistic angular momentum conservation law changes @xmath87 by 510% .
in this section , we present calculations of the expansion and spin - down of the atmosphere , including the effects of general relativity and rapid rotation . in the spirit of cb
, we ignore latitudinal variations , and work in the equatorial plane .
first , in 4.1 , we discuss the effects of general relativity on the equations describing the hydrostatic and thermal structure of the atmosphere , and calculate the reduction in gravity at the equator due to rapid rotation . in 4.2 ,
we show how to include the general relativistic angular momentum conservation law that we obtained in 3 . in 4.3 ,
we rescale our fiducial results of 2 for the spin - down of a rigidly - rotating atmosphere to different masses , equations of state , and spin frequencies . in 4.4 , we relax the assumption of rigid - rotation , and present the rotational profiles of the atmosphere for several different cases .
we now write down the differential equations describing the hydrostatic and thermal structure of the atmosphere including general relativity , and compare them to the newtonian equations .
we start by considering a non - rotating star , and then discuss the additional effects of rotation . for a spherical star , thorne ( 1977 )
( see also thorne 1967 ) gives the equation for mass conservation as @xmath94 where @xmath95 is the rest mass ( number of baryons multiplied by baryon rest mass ) within coordinate radius @xmath14 , @xmath96 is the rest mass density , and @xmath97 is the `` volume redshift factor '' . for slow rotation , @xmath98 , where @xmath99 is the gravitational mass interior to @xmath14 .
we adopt a plane parallel approximation in the thin envelope , and write @xmath100 , where @xmath101 is the proper length . here , @xmath97 is the redshift factor at the surface of the star @xmath102 .
we define the rest mass column depth @xmath103 .
the mass conservation equation becomes @xmath104 the equation for hydrostatic balance is ( thorne 1977 ) @xmath105 where @xmath106 is the locally measured pressure , and we neglect contributions to the gravitational mass from the fluid energy density . using the definition of column depth , we rewrite equation ( [ eq : hb ] ) for the thin layer as @xmath107 where @xmath108 rewriting the entropy and flux equations ( thorne 1977 , eqs .
[ 11d ] and [ 11h ] ) in a similar way , we find @xmath109 and @xmath110 where @xmath111 is the heat flux , @xmath112 is the nuclear energy release rate per unit rest mass , @xmath113 is the opacity , and @xmath114 is the locally measured temperature . equations ( [ eq:1 ] ) , ( [ eq:2 ] ) , ( [ eq:3 ] ) , and ( [ eq:4 ] ) , which describe the structure of the atmosphere , are identical to the newtonian equations ( compare cb 2 )
. these results also hold in the slow rotation approximation , for which the star is spherical to first order in @xmath37 .
the inclusion of rapid rotation has three effects .
the first is to change the volume correction factor @xmath115 .
in addition , the mass conservation equation must be covariant so that all observers agree on the rest mass of the star .
this requirement gives an additional lorentz factor in equation ( [ eq : mass ] ) ( e.g. , friedman et al .
1986 ) . however , we find that for the models we consider in this paper , both of these effects are @xmath116% , and so we neglect them .
much more important is the correction to gravity due to the rapid rotation .
this correction factor is the relativistic generalisation of the newtonian reduction in gravitational acceleration due to the centrifugal force . to calculate this , we write the four - velocity as @xmath117 , and use the normalization @xmath118 to solve for the redshift factor @xmath119 where @xmath120 is given by equation ( [ eq : v ] ) .
the equation of hydrostatic balance for a rigidly rotating star is then ( friedman et al 1986 ) @xmath121 where we assume the pressure is much smaller than the rest mass density . to reduce this equation to the form of equation ( [ eq:2 ] ) ,
we write @xmath122 where the correction factor @xmath123 accounts for the reduction in @xmath124 at the equator due to rapid rotation , @xmath125 tending to unity in the limit of zero rotation .
this is the most important effect coming from rotation , giving a correction of up to 25% for a rotation frequency of @xmath42 and a stiff equation of state . in summary ,
the general relativistic equations look the same as the newtonian equations if we use the correct value of surface gravity @xmath124 ( eq . [ [ eq : g2 ] ] ) , and if we identify the column depth @xmath15 with the rest mass column depth , and the thickness @xmath126 with the proper thickness of the layer .
values of @xmath124 , @xmath97 , and @xmath127 are given in table [ tab : models ] for different models .
to incorporate the general relativistic angular momentum conservation law into our calculations of spin down , we write the angular momentum per unit rest mass at the equator in the form @xmath128 where @xmath129 is a function of radius , e.g. @xmath130 in the newtonian limit . if the angular momentum is constant , we find that the change of @xmath131 with radius is determined by the gradient of @xmath132 with radius : @xmath133 . now consider a particle which moves from radius @xmath86 to @xmath134 , where @xmath135 .
conservation of angular momentum gives @xmath136 .
expanding in a taylor series around @xmath137 ,
we find @xmath138 in the newtonian limit , for which @xmath139 , we recover the familiar result @xmath140 . to calculate the moment of inertia
, we must modify equation ( [ eq : right ] ) . again writing the angular momentum in the form of equation ( [ eq :
f ] ) , and expanding @xmath129 around @xmath137 , we find @xmath141\ dy,\ ] ] which reduces to equation ( [ eq : right ] ) in the newtonian limit .
the calculations presented in 2 were for a surface gravity @xmath142 ( the newtonian gravity for a @xmath143 , @xmath144 neutron star ) .
we now rescale these fiducial results to different masses and equations of state , including the effects of rapid rotation and general relativity .
to calculate the thickness of the atmosphere in 2 , we integrated hydrostatic balance in the form @xmath145 , giving the thickness @xmath146 ( see also 2 of cb ) .
in addition , we must include a factor @xmath147 to obtain the correct general relativistic value for @xmath124 , and a factor @xmath127 to include the effects of rotation on gravity .
we thus find the fractional change in thickness of the atmosphere is @xmath148 where @xmath149 is the fiducial value from 2 .
the fractional change in spin frequency is @xmath150 , where @xmath151 is the coordinate thickness .
converting to proper thickness @xmath152 involves another factor of @xmath147 , giving @xmath153 where we define @xmath154 as the scaling factor from the fiducial value , and we took @xmath155 in 2 .
note that the largest contributions to @xmath154 come from rescaling gravity and including the centrifugal force , rather than from general relativistic effects .
figure [ fig : corr ] shows @xmath154 as a function of neutron star mass , for eos apr and eos l , and in each case for @xmath93 and @xmath81 .
we also give @xmath154 in table [ tab : models ] .
figure [ fig : corr ] shows that for massive neutron stars with @xmath156 , the frequency shifts calculated in 2 and shown in figure 1 should be multiplied by @xmath157@xmath158 for eos l , and @xmath159@xmath160 for eos apr . for a @xmath143 star ,
these factors are @xmath161@xmath162 for eos l , and @xmath163@xmath164 for eos apr .
the largest values of @xmath154 are obtained for a low neutron star mass , stiff equation of state , and rapid rotation . however , even @xmath165 for @xmath166 , eos l , and @xmath167 is not large enough to give agreement with the largest observed frequency shifts . in figure
[ fig : omega ] , we present some rotational profiles for some specific models . we take the neutron star mass to be @xmath168 , the global accretion rate to be @xmath169 , and again work in the equatorial plane .
we assume that the atmosphere is rigidly - rotating immediately prior to the burst , and that during the burst , the atmosphere is radiative and carries a flux equal to the solar - composition eddington flux at the photosphere ( @xmath170 ) . for @xmath93 and @xmath167 , and for eos
l and eos apr , we plot the rotational profiles assuming either ( i ) complete rotational coupling , resulting in rigid rotation across the layer , or ( ii ) no angular momentum transport , giving a differentially - rotating layer . if we allow differential rotation to persist , we see that the upper layers of the atmosphere spin down by an amount comparable to or greater than the observed spin changes during bursts
. however , it is not clear why the spin frequency observed should be that of only the outermost layers .
indeed , cb ( 3.4 ) argued that substantial differential rotation would wash out any signal from the deeper cooling layers because of the finite time to transport heat vertically .
we have presented new calculations of the hydrostatic expansion and spin - down of a neutron star atmosphere during a type i x - ray burst .
our main conclusion is that hydrostatic expansion is not enough to explain the observed frequency drifts during type i x - ray bursts if the burning atmosphere rotates rigidly . in 2
, we showed that cumming & bildsten ( 2000 ) ( cb ) overestimated the change in the moment of inertia of the atmosphere , obtaining values of spin - down that were a factor of two too large .
figure 1 compares the new calculations of spin - down with observations , including recent measurements of large frequency drifts during bursts ( galloway et al .
2000 ; wijnands et al .
we find that the largest observed frequency shift is a factor of 3 or more greater than the theoretical values . in 3
, we derived the angular momentum conservation law in general relativity .
we calculated the variation of spin frequency with radial distance , @xmath87 , for a particle moving with constant angular momentum . in the slow rotation approximation ( @xmath171 )
, we obtained the analytic result given by equation ( [ eq : slow ] ) , which agrees with recent work by abramowicz et al .
( 2001 ) . for rapidly - rotating stars
, we calculated @xmath172 by numerically solving for the structure of the neutron star .
the correction to the newtonian angular momentum conservation law , @xmath173 , is shown in figure 2 for different neutron star masses , equations of state and spin frequencies .
contrary to the results of heyl ( 2000 ) , which were also shown to be incorrect by abramowicz et al .
( 2001 ) , we find that the general relativistic correction is small , about 510% . in 4 , we calculated the atmospheric expansion and spin - down , including the effects of rapid rotation and general relativity .
working in the equatorial plane ( in the spirit of cb s calculation , we neglect latitudinal variations in this paper ) , we calculated the scaling factor @xmath154 required to rescale our fiducial results of 2 to different neutron star masses , equations of state , and spin frequencies ( figure 3 ) .
in addition , we presented the rotational profiles for some particular models ( figure 4 ) .
we find that the largest spin - down is for rapid rotation , and low mass stars with a stiff equation of state .
for example , a 1.4 @xmath31 star spinning at 600 hz with the stiff equation of state eos l has @xmath165 .
however , this is not a large enough factor to bring observations and theory into agreement in figure 1 .
the rotational profiles given in figure 4 show that in principle frequency shifts as large as those observed can be obtained by hydrostatic expansion , if we consider only the outermost layers of the atmosphere , and allow differential rotation
. however , it is not at all obvious why the observed frequency would be that of the outermost shells of the atmosphere , which contain a small amount of the mass , particularly since the energy release is in the deeper layers . indeed , cb argued ( see their 3.4 ) that substantial differential rotation in the burning layers would wash out the signal , because of the finite time needed to transport heat vertically in the atmosphere .
another possible objection to the angular momentum conservation picture was given by cb , who pointed out that if the burning layers are threaded by a large scale poloidal magnetic field , this field will be wound up by the differential rotation during the burst .
the wound up toroidal field acts back on the shear , halting and reversing its direction in the time for an alfven wave to cross the atmosphere ( see , for example , spruit 1999 ) . for a magnetic field typical of a millisecond radio pulsar , @xmath174 , this timescale is only @xmath175 .
one possibility is that the surface field is much weaker , @xmath176 , allowing shearing to persist .
observations of burst oscillations in the accreting millisecond x - ray pulsar sax j1808.4 - 3658 , which perhaps has a @xmath177@xmath178 field ( psaltis & chakrabarty 1999 ) , would give an interesting test of this picture .
int zand et al .
( 2000 ) report a marginal detection with bepposax of oscillations at @xmath179 during a type i burst , but are unable to resolve any frequency drift . unfortunately , as yet no burst oscillations have been detected from this source with rxte .
the results we have obtained in this paper suggest that we may have to look elsewhere for an explanation of the frequency drifts .
spitkovsky , levin , & ushomirsky ( 2001 ) , in a detailed investigation of hydrodynamic flows during type i bursts , propose that the frequency drifts may be explained by a combination of radial expansion and variations in the velocity of zonal flows on the neutron star surface . another possibility mentioned by cb is that the burst oscillation is due to a non - radial oscillation in the neutron star surface layers .
such possibilities remain to be investigated in detail .
we thank jeremy heyl and scott hughes for many useful discussions , and yuri levin for a careful reading of the manuscript .
we are grateful to chris matzner for insights into the issues discussed in the appendix .
this work began at the program on spin and magnetism in young neutron stars , itp , santa barbara .
a. c. and s. m. thank the canadian institute for theoretical astrophysics for hospitality during its completion .
this research was supported by the national science foundation under grants phy99 - 07949 and ay97 - 31632 , by nasa via grant nag 5 - 8658 , and by the natural sciences and engineering research council of canada .
l. b. is a cottrell scholar of the research corporation .
abramowicz , m. a. , kluzniak , w. , & lasota , j .-
2001 , , 374 , l16 abramowicz , m. a. , & prasanna , a. r. 1990 , , 245 , 720 akmal , a. , pandharipande , v. r. , & ravenhall , d. g. 1998 , phys .
c , 58 , 1804 arnett , w. d. , & bowers , r. l. 1977 , apjs , 33 , 415 bhattacharya , d. 1995 , in x - ray binaries , ed .
w. h. g. lewin , j. van paradijs , & e. p. j. van den heuvel ( cambridge : cambridge university press ) , 233 bildsten , l. , 1998 , in the many faces of neutron stars , ed .
r. buccheri , j. van paradijs , & m. a. alpar ( dordrecht : kluwer ) , 419 cook , g. b. , shapiro , s. l. , & teukolsky s. a. 1994 , , 424 , 823 cumming , a. , & bildsten , l. 2000 , , 544 , 453 ( cb ) friedman , j.l . ,
ipser , j.r . , & parker , l. 1986 , , 304 , 115 galloway , d. k. , chakrabarty , d. , muno , m. p. , & savov , p. 2001
, , 549 , l85 grindlay , j. e. , bailyn , c. d. , cohn , h. , lugger , p. m. , thorstensen , j. r. , & wegner , g. 1988 , apj , 334 , l25 hansen , c. j. , & kawaler , s. d. 1994 , stellar interiors : physical principles , structure and evolution ( new york : springer - verlag ) hartle , j. b. 1967 , , 150 , 1005 hartle , j. b. 1970 , , 161 , 111 heyl , j. s. 2000 , , 542 , l45 int zand , j. j. m. , et al .
2001 , , in press ( astro - ph/0105324 ) koike , o. , hashimoto , m. , arai , k. , & wanajo , s. 1999 , , 342 , 464 komatsu , h. , eriguchi , y. , & hachisu , i. 1989 , , 237 , 355 lewin , w. h. g. , van paradijs , j. , & taam , r. e. 1995 , in x - ray binaries , ed .
w. h. g. lewin , j. van paradijs , & e. p. j. van den heuvel ( cambridge : cambridge univ .
press ) , 175 mihalas , d. , & mihalas , b. w. 1984 , foundations of radiation hydrodynamics ( oxford : oxford university press ) miller , m. c. 2000 , , 531 , 458 morsink , s. m. , & stella , l. 1999 , , 513 , 827 muno , m. p. , chakrabarty , d. , galloway , d. k. , & savov , p. 2001
, , 553 , l157 nelson , l. a. , rappaport , s. a. , & joss , p. c. 1986 , apj , 304 , 231 pandharipande , v. r. , & smith , r. a. 1975 , phys .
b , 59 , 15 psaltis , d. , & chakrabarty , d. 1999 , , 521 , 332 ravenhall , d. g. , & pethick c. j. 1994 , , 424 , 846 schatz , h. , et al .
1998 , phys . rep .
, 294 , 167 schatz , h. , et al .
2001 , , 86 , 3471 schutz , b. f. 1990 , a first course in general relativity ( cambridge university press : cambridge ) spitkovsky , a. , levin , y. , & ushomirsky , g. 2001 , in preparation spruit , h. c. 1999 , , 349 , 189 strohmayer , t. 2001 , astro - ph/0012516 strohmayer , t. e. , jahoda , k. , giles , b. a. , & lee , u. 1997 , , 486 , 355 strohmayer , t. e. , zhang , w. , swank , j. h. , smale , a. , titarchuk , l. , day , c. , & lee , u. 1996 , , 469 , l9 thorne , k. s. 1967 , in high energy astrophysics , vol . 3 , ed .
c. dewitt , e. schatzman , and p. veron , proc .
1966 les houches summer school ( new york : gordon & breach ) thorne , k. s. 1977 , , 212 , 825 van der klis , m. 2000 , , 38 , 717 walter , f. m. , mason , k. o. , clarke , j. t. , halpern , j. , grindlay , j. e. , bowyer , s. , & henry , j. p. 1982 , apj , 253 , l67 white , n. e. , & swank , j. h. 1982 , apj , 253 , l61 wijnands , r. , strohmayer , t. , & franco , l. m. 2001 , , 549 , l71 llllcllll 1.4 & 11.5 & 0.18 & 1.25 & 1.72 & 0.98 & 0.94 & 0.72 & 1070 + 2.0 & 11.1 & 0.27 & 1.46 & 3.10 & 0.99 & 0.90 & 0.34 & 1330 + 1.4 & 11.9 & 0.17 & 1.24 & 1.48 & 0.91 & 0.94 & 0.81 & 1070 + 2.0 & 11.3 & 0.26 & 1.45 & 2.83 & 0.95 & 0.90 & 0.37 & 1330 + 1.4 & 15.2 & 0.14 & 1.17 & 0.90 & 0.95 & 0.96 & 1.13 & 730 + 2.0 & 15.3 & 0.19 & 1.28 & 1.39 & 0.97 & 0.94 & 0.65 & 850 + 1.4 & 16.7 & 0.12 & 1.15 & 0.55 & 0.72 & 0.95 & 1.70 & 730 + 2.0 & 16.2 & 0.18 & 1.25 & 1.05 & 0.83 & 0.93 & 0.82 & 850 +
in this appendix , we start with the conservation equations for a fluid , which include contributions from the fluid s enthalpy and from heat flow , and show that the appropriate conserved quantity is the angular momentum per baryon @xmath180 as calculated in 3 .
we begin with the stress tensor for a fluid @xmath181 ( see , for example , mihalas & mihalas 1984 ) .
the enthalpy @xmath182 is @xmath183 where the energy density @xmath184 is the sum of rest mass energy , internal energy , and nuclear binding energy .
we write the heat flux as @xmath185 . by setting @xmath186 , and using baryon number conservation @xmath187
, we obtain the angular momentum conservation law @xmath188 assuming the azimuthal heat flux @xmath189 may be neglected when compared to the vertical heat flux @xmath190 , we rewrite equation ( [ eq : a1 ] ) as @xmath191 we now simplify this equation using the first law of thermodynamics , which is @xmath192 where @xmath193 is the entropy per baryon mass , or , using equation ( [ eq : h ] ) , @xmath194 substituting for @xmath195 in equation ( [ eq : a2 ] ) , we find @xmath196 -{u_\phi\over\rho h}{dp\over dt}-{f\over\rho h}{du_\phi\over dz}.\ ] ] the first term in equation ( [ eq : a3 ] ) vanishes , since @xmath197 is the familiar entropy equation .
the nuclear energy production rate is included in the @xmath198 term , since the nuclear binding energy is included in the enthalpy : equations ( [ eq : eps ] ) and ( [ eq : firstlaw ] ) give @xmath199 where @xmath200 is the nuclear energy generation rate . integrating the second term in equation ( [ eq : a3 ] ) in time
, we find it gives a contribution to @xmath201 of order @xmath202 , where we take the pressure scale height @xmath203 .
the third term in equation ( [ eq : a3 ] ) represents radiative viscosity ( see mihalas & mihalas 1984 ) . again integrating in time , we find that this term gives a contribution @xmath204 .
both the second and third terms are of order @xmath205 or smaller , and may be neglected for our purposes . | we present calculations of the spin - down of a neutron star atmosphere due to hydrostatic expansion during a type i x - ray burst .
we show that cumming & bildsten incorrectly calculated the change in the moment of inertia of the atmosphere during a type i burst , resulting in a factor of two overestimation of the magnitude of the spin - down for rigidly rotating atmospheres .
we derive the angular momentum conservation law in general relativity , both analytically for the case of slow rotation , and numerically for rapidly - rotating stars .
we show that contrary to the claims of heyl , but in agreement with abramowicz and coworkers , general relativity has a small effect on the angular momentum conservation law , at the level of 510% .
we show how to rescale our fiducial results to different neutron star masses , rotation rates and equations of state , and present some detailed rotational profiles . comparing our results with recent observations of large frequency shifts in mxb 1658 - 298 and 4u 1916 - 053
, we find that the spin - down expected if the atmosphere rotates rigidly is a factor of two to three less than the observed values .
if differential rotation is allowed to persist , we find that the upper layers of the atmosphere do spin down by an amount comparable to or greater than the observed values . however , there is no compelling reason to expect the observed spin frequency to be that of only the outermost layers of the atmosphere .
we conclude that hydrostatic expansion and angular momentum conservation alone can not account for the largest frequency shifts observed during type i bursts . |
recent observations reveal that outflows are important in the lifecycle of galaxies .
first , @xcite find evidence for chemical enrichment trends throughout star - forming galaxies over three orders of magnitude in stellar mass .
surprisingly , the enrichment trends are observed to continue all the way down to @xmath3 @xcite where @xmath4 is the bolometric luminosity of the galaxy .
indeed , there is evidence that galaxies that exceed the mass of the milky way manage to retain a large fraction of their metals , in contrast to lower mass galaxies where metal loss appears to be anticorrelated with baryonic mass .
recent theoretical work has shown that a low - metal content in dwarf galaxies is attributable to efficient metal - enriched outflows @xcite .
secondly , there is now strong evidence for large - scale outflows in the galaxy across the electromagnetic spectrum @xcite .
finally , the intergalactic medium ( igm ) at all redshifts shows signs of significant metal enrichment consistent with the action of winds @xcite . a high proportion of lyman break galaxies at @xmath5 show kinematic signatures of winds @xcite as do many nearby dwarf starbursts @xcite and ultraluminous infrared galaxies @xcite .
n - body simulations of galaxy formation within cold dark matter ( cdm ) cosmology fail to produce realistic galactic disks @xcite .
over - cooling in cdm+hydrodynamic simulations , due primarily to a lack of resolution , results in too much gas arriving in the centre of the galaxies . but a useful aspect of these incomplete models has been to highlight the possible role of feedback in shaping galaxies . vigorous feedback in the early phase of galaxy formation is a partial solution to the angular momentum problem by preventing baryons from losing too much of their specific angular momentum through the interaction with dark matter @xcite .
indeed , some authors have claimed more realistic disks retaining much more of their angular momentum when one includes the action of agn - driven jets @xcite or starburst driven winds @xcite .
our new survey is preempted by a vast literature across a broad spectrum that can be traced back over two decades @xmath6 comprehensive reviews are given by heckman_etal .
_ ( 1989 ) and veilleux_etal._(2005 ) .
much of the early optical work on galactic outflows was limited to narrowband images and long - slit spectroscopy ( e.g.mccarthy_etal . _ ( 1987 ) ; @xcite ; phillips ( 1993 ) ; lehnert & heckman ( 1996 ) ) although fabry - perot observations have been carried out in a few emission lines ( e.g. bland & tully ( 1988 ) ; cecil ( 1990 ) ) .
the quality of uv , x - ray , infrared and radio continuum observations has greatly improved in recent years ( e.g. hoopes_etal._2005 ; strickland_etal . _ 2004 ; @xcite ) . even without the kinematic separation of compound emission structures that is possible via spectroscopy
, there are clear associations between these bands and the optical maps ( e.g. veilleux_etal . _ 1994 ; cecil_etal . _ 2002 ) .
historically , most winds had been confirmed kinematically from evidence of line - splitting along the galaxy minor axis , but at least one wind has been discovered through its ionisation signature @xcite . in recent times measurement of blueshifted absorption lines
have proved powerful in this regard . in this new study
, we perform wide field integral field spectroscopy of a sample of nearby galactic winds .
the advantages offered by a wide wavelength base line and moderate spectral resolutions allow the determination of physical parameters for wind outflow mechanics based on ionisation and kinematic signatures e.g. rotating disk , high - latitude warm ionizing medium , wind filaments , entrained gas .
one well - studied object in particular convinced us of the potential for wide - field mapping with an integral field spectrograph .
the projected distribution of optical line emission in m82 is exceedingly complex but with the aid of fabry - perot observations , shopbell & bland - hawthorn ( 1998 ) were able to kinematically separate the front side from the back side of the wind flow , and from the diffuse halo of broadened line emission ( see also the recent ifs studies of @xcite ) .
when this is done , both the ionisation and kinematics exhibit well ordered behavior .
the lesson is clear : with complete observations , it is possible to separate distinct dynamical components
. only then does it become possible to determine physical parameters for the outflow mechanism . in this first paper , we present the new observations and describe the reduction procedures specific to integral field spectroscopy ( [ obsred ] ) .
we present emission - line flux and line ratio maps for a wide range of diagnostics ( e.g. ionisation and density ) . for each galaxy
, we use the ionisation diagnostic diagrams ( idds ) to delineate the different sources of ionisation as a function of location and ionisation source ( see [ the galaxy sample ] ) and classify the sample targets according to their individual idds ( [ starbursts ] - starbursts & [ agn ] - agn ) .
these diagnostic diagrams reveal well - ordered behavior across each galaxy . in [ agn sf discussion ]
we consider the source of ionisation from a theoretical perspective and use these considerations to construct an evolutionary sequence for galactic winds in section [ events ] .
our conclusions are presented [ discussion ] . through this work , we assume a flat concordance cosmology such that h@xmath7=71kms@xmath8mpc@xmath8 , @xmath9=0.27 and @xmath10=0.73 .
the aaomega integral - field spectrograph at the aat 3.9 m is ideally suited to our study ( see appendix ) .
the spiral integral field unit ( ifu ) is a 32@xmath1116 element rectangular microlens array coupled via an optical fibre feed to the dual beam aaomega spectrograph @xcite .
mounted at the cassegrain auxilliary focus , the 0.7 ifu pixel scale is well matched to the typical seeing at the aat , and gives excellent sensitivity to extended low surface brightness features .
the single pointing field - of - view is [email protected] . high accuracy ,
guide - probe offset guiding allows simple and reliable mosaicing of large areas .
the default position angle of the ifu places the long axis in an east - west direction .
observations of a number of targets were undertaken at alternate pas to accommodate the demands of other observing programs undertaken alongside the observations reported here .
observations were undertaken over a number of observing runs as detailed in table [ obs runs ] .
the dual beam aaomega spectrograph provides excellent coverage of the spectral features of interest , while maintaining a high spectral resolution .
we observe with the 1500v and 1000r volume phase holographic ( vph ) gratings in the blue and red arms respectively .
we use a 5700 dichroic beam splitter and center the wavelength coverage and vph blazes at 5000 and 6800 , resulting in 2.4 pixel mid - range resolution elements of r@xmath126300 and r@xmath125600 ( velocity resolutions of @xmath13 35 and @xmath1350 km s@xmath8 fwhm ) . on accounting for the sample redshift range , this targets the emission lines h@xmath0/[oiii ] in the blue and [ oi]/h@xmath1/[nii]/[sii ] in the red .
unfortunately the important [ oii ] doublet at 3727 falls outside of the observable spectral range with the chosen settings .
we consider this to be a crucial line for future studies of this kind .
quartz - halogen flatfield exposures and cuar@xmath14fear arc lamp exposures are taken at intervals throughout the observation to allow fibre tracing on the ccds ( both arms of aaomega use e2v 2k@xmath114k ccds ) , flat fielding and wavelength calibration .
stationed in the west coud room , aaomega is kept thermally and gravitationally stable throughout each night s observations in order to minimize the overhead on system calibrations .
the strong night sky oxygen line at 5577 falls outside the observed range and so can not be used to calibrate relative fibre transmission as is the common practice in low resolution multi - object spectroscopy .
therefore twilight flatfield frames are observed in order to account for relative fibre - to - fibre transmission variations .
an assortment of spectrophotometric standard stars were observed throughout each observing block , in order to prepare a flux calibration solution .
the scatter in the calibration function derived from the standards during the observations indicates a 10% uncertainty in the absolute flux calibration level . the relative calibration with wavelength , a quantity more relevant to line ratio determinations , is consistent at the @xmath122 - 3% level , as determined from the comparison of calibration observations over the course of the survey .
each target was observed as part of a multi - pointing mosaic .
an overlap of 1 or 2 ifu elements ( 0.7 or 1.4 ) was allowed to enable relative intensity scaling and accurate mosaicing of the final data sets .
individual science exposures were typically 1200 sec .
mosaic sizes vary between galaxies , but are typically of the order of six to twelve ifu positions , yielding mosaics of the order 20 to 60 on a side .
since our targets completely fill each ifu pointing , it is not possible to generate a sky subtraction solution from fibres within each ifu pointing . to achieve accurate sky subtraction ,
a small number of dedicated offset sky frames was taken each night .
data was processed in real time at the telescope using the ` 2dfdr ` software package .
this allowed monitoring of the observing conditions such that the mosaic pattern could be repeated or adapted if required .
the final data analysis was carried out using a custom suite of idl routines .
there are four isolated `` dead elements '' within the spiral ifu . during the reduction process
, these are replaced with the average spectrum of the four adjacent ifu elements prior to mosaicing the data .
individual frames were then aligned and mosaiced using telescope offset information and a suite of custom written software tools .
frames are scaled , based on an iterative comparison of overlap regions in the mosaic , in order to account for minor variations in transparency , seeing and exposure time .
.[obs runs ] a summary of aat observations with the aaomega@xmath6spiral spectrograph .
[ cols="<,<",options="header " , ] [ bubble ] [ default ] in most simulations to date , the terminal velocity of gas entrained by a galactic wind is @xmath15 km s@xmath8 ( e.g. suchkov_etal._1994 ; cooper_etal . _ 2008 ) .
proper comparisons are difficult because the models deal with clumpiness in the entrained gas , which the scaling relations above do not .
chevalier & clegg ( 1985 ) provide a useful formula for the terminal velocity of a cloud accelerated by a wind ( see also strickland & heckman 2008 ) : @xmath16 where @xmath17 and @xmath18 are the initial distance and column density of the cloud .
( this formula is specific to the constant velocity " phase of the wind before it breaks free of the confining medium . )
the factors @xmath19 and @xmath20 account for the fractions of supernova energy and mass respectively that work to heat the diffuse gas .
these authors explore fractions in the range 0.1 to 1 and ultimately favour the high end of this range . in any event
, the dynamical time of the wind is roughly @xmath21 myr ( see also cooper_etal . _ 2008 ) which is considerably longer than the cooling time of the entrained gas .
the wind timescale is comparable to the dynamical time of the starburst region itself , but is up to an order of magnitude less than the ionisation timescale @xmath22 established by our analysis above .
this is a striking result that suggests a substantial delay ( @xmath23 myr ) after the onset of star formation before the conditions for a large - scale wind are properly established . in the starburst wind objects , the overall ionisation properties are consistent with shock processes ( see 3.1 ) .
a simple calculation can validate this . from dopita & sutherland ( 1996 ) ,
the total shock luminosity is @xmath24 these authors find that about 50% of the mechanical energy is converted into energy capable of ionizing gas .
assuming there is sufficient gas to soak up the ionizing photons , one can derive a useful relation for the h@xmath1 luminosity , such that @xmath25 we now derive a canonical h@xmath1 surface brightness for shock - ionised filaments associated with the galactic wind . for a galaxy at a distance of @xmath26 mpc , the observed h@xmath1 flux from the shock
is @xmath27 which assumes , consistent with dopita & sutherland ( 1996 ) , that only a single shock surface is observed with no correction for possible limb brightening .
the surface brightness of shocked emission at 1 kpc follows : @xmath28 the pre - shock density normalization assumes that the halo gas in starburst galaxies follows an isothermal distribution ( e.g. cooper_etal .
_ 2008 ) normalised to a pre - shock density of @xmath29 @xmath30 at 1 kpc consistent with material swept up in a wind ( q.v .
chevalier & clegg 1985 ; see below ) .
post - shock densities inferred from the [ sii ] doublet ( e.g. mccarthy_etal . _ 1987 ) are typically in the low density limit ( @xmath31 @xmath30 ) beyond @xmath32 pc .
while equation [ surfbri2 ] has a strong dependence on @xmath33 , the shock speed is well constrained by observation .
some authors have associated the terminal velocity of entrained gas with the shock velocity induced by the wind ( e.g. maloney 1999 ) .
but the levels of ionisation tell a different story .
the strong [ nii]/[oiii ] signal favours shock velocities of order 100 km s@xmath8
( e.g.shopbell & bland - hawthorn 1998 ; veilleux & rupke 2002 ) .
this is supported by the soft x - ray emission which appears to be powered by similar shock velocities .
how are we to understand this ? in our picture , the fast hot wind accelerates denser gas to a terminal velocity .
these clouds are ram - pressure confined by the fast wind .
we therefore write down a simple equation for shocks driven into the clouds such that @xmath34 where @xmath35 and @xmath36 are the wind density and velocity .
strickland & heckman ( 2009 ) have recently conducted an excellent re - evaluation of chevalier & clegg s original model for the m82 wind .
the free - flowing or break - out wind speeds can be as high as @xmath37 km s@xmath8 .
for our adopted values of @xmath38 , @xmath39 and @xmath40 above , we can balance equation [ pram ] with @xmath41 @xmath30 , a factor of a few higher than the canonical solution at 1 kpc presented by strickland & heckman ( 2009 ) .
we consider this to be broadly consistent given the uncertainties . in our simple analysis ,
the shock velocity is always several times less than the terminal velocity of clouds caught in the flow .
our simple model leads to an h@xmath1 surface brightness of @xmath42 cgs at 1 kpc .
this approximate value is comparable to what is presented in table [ gas extent ] for the starburst galaxies . in principle , the filling factor of the shocked material can be low ( @xmath43 ) , and there will be moderate levels of extinction internal to the entrained gas ( i.e. @xmath44 mag ; cecil_etal._2001 ) .
but in principle , a moderately fast shock could account for the observed , extended line emission .
we note that the h@xmath1 surface brightness in equation [ surfbri2 ] is comparable to the observed agn surface brightness in table [ gas extent ] , but much fainter than the predicted agn surface brightness in equation [ surfbri1 ] .
this may indicate that the filling factor in the wind filaments is @xmath45 .
fortunately , a discussion of the _ relative _ merits of different ionisation sources does not require knowledge of the gas filling factor @xmath46 .
a lower filling factor affects the theoretical prediction identically in equations [ surfbri1 ] and [ surfbri2 ] , and usefully brings our models into line with the observations , at least in principle .
thus , if agn winds involve similar processes to starburst winds , we would not expect shocks to outshine direct photoionisation from the agn , consistent with our observations .
we have already argued that shocks outshine direct stellar photoionisation from the starburst because the young starburst has dimmed tremendously by the time the wind breaks out .
close to the disk , we may still expect stellar photoionisation to dominate over shocks owing to large - scale star formation in the disk ( bland - hawthorn & maloney 1999 ) .
but much beyond 300 pc or so , wind - driven shocks can dominate the ionisation , as observed in our sample .
if we invoke a lower filling factor consistent with agn winds , we can readily explain ngc 3628 , but now ngc 253 and ngc 1482 appear overluminous .
indeed , a careful study of table [ gas extent ] reveals the need to change @xmath46 over the range ( 0.1,1 ) to match our model to all of the observations .
this may indicate that our model is oversimplified and needs to include a gas filling factor , or even a correction for internal extinction ( although this is found to be typically less than 10% from the spiral observations in the brightest regions ) .
we note two things :
( i ) the inferred filling factors are not unreasonable and can be directly measured in principle ; ( ii ) alternatively , the inferred extinction corrections are plausible , and these can also be directly inferred from the balmer decrement with data of sufficient sensitivity .
we conclude with a `` big picture '' model of the starburst wind phenomenon .
our purpose is to provide a physical explanation for the delay before the galactic wind gets going .
we believe that this simple model explains the broadest features of our new spiral observations .
therefore the stages of evolution for a galactic wind are as follows : \(1 ) stars are born in dense molecular clouds whereupon the hot massive stars produce a strong uv radiation field ( lada & lada 2003 ) ; \(2 ) the mean free path of lyman continuum photons is greatly extended by stellar winds ( and radiation pressure ) that blow bubbles into the surrounding ism .
( good discussions of this process are to be found in castor_etal . _ 1975 ; weaver_etal . _ 1977 ; maclow & mccray 1988 ; tomisaka & ikeuchi 1988 . )
the intense uv radiation ionizes any pre - existing intracloud medium but this may have been cleared out by an earlier wind .
\(3 ) the over - pressured warm gas evaporated from the surfaces of dense clouds helps to fill the wind - blown `` voids '' and intracloud region .
the connection between the evaporating cloud skins and the intracloud wind material may help the wind to entrain denser gas from these same clouds ( cooper_etal .
_ 2008 ) .
\(4 ) the bulk of the supernova ejecta arises after 8 - 10 myr ( fig .
[ starburst99a ] ; leitherer_etal .
_ 1992 ) .
as noted by others ( e.g. chevalier & clegg 1985 ) , the ejecta ultimately thermalizes the diffuse material which escapes as a wind ; any shock heating of the denser medium escapes as infrared radiation ( strickland & stevens 2000 ) .
the uv intensity declines dramatically since many of the high - mass stars have now exploded .
thus we are led to an important conclusion . _
the large - scale wind can not become established until vast quantities of warm diffuse gas are dislodged from the surfaces of dense clouds by uv evaporation . _
the ionised material with a sufficiently long cooling time is then mechanically heated by the supernova mass / energy input .
it is therefore not surprising that , by the time the wind gets going , most of the short - lived massive stars have faded or disappeared .
we would therefore _ not _ expect to see a strong signature of stellar photoionisation in the extended wind filaments if the starburst phenomenon arises from a _ single _ burst of star formation .
_ our new observations may provide the strongest evidence yet that starbursts truly are impulsive events_. the fact that the signature of agn photoionisation _ is _ clearly visible in the wind filaments argues that the accretion - disk uv timescale is considerably longer than the starburst timescale .
furthermore , unlike the starburst , the mechanical energy input _ must _ have occurred early in the life of the agn , presumably at a time when jets can also be launched .
we argue that our observations are a testament to this simple picture .
conversely , if there are starburst sources where stellar photoionisation _
does _ dominate at large radial distances ( cf .
shopbell & bland - hawthorn 1998 ) , we anticipate that the circumnuclear regions in these sources are characterised by an extended episode of star formation ( i.e. continuous star formation ) due to an extended period of fuelling and accretion ( e.g. sustained merger activity ) .
the role of mass / energy feedback and recycling in galactic winds over cosmic time is very poorly understood . in particular , we are far from a detailed understanding of how gas accretion to the circumnuclear regions fuels local star formation and nuclear activity .
so what can we say ? the galaxy wind phenomenon is well established and its influence can be detected in the highest redshift sources ( heckman_etal .
_ 1990 ; strickland_etal .
_ 2004b ; veilleux_etal .
_ 2005 ) .
galactic winds are associated with both star formation ( e.g. starbursts ) , nuclear activity ( e.g. fr i radio sources ) , and a combination of star formation and nuclear activity ( e.g. seyfert galaxies ) .
but this raises many questions .
how do winds get going , how energetic are they , how long do they last , and what is their role in galaxy evolution ?
do we arrive at similar answers for agn - driven winds and starburst - driven winds ?
here we are able to shed light on the first question ( [ events ] ) .
we have carried out the most detailed and extensive survey to date of extended plasma gas in active galaxies .
we constructed ionisation diagnostic diagrams to study the nature of the emission in ten galaxies .
we find that the wholly starburst - powered sources have extended emission that is largely consistent with shock ionisation .
this sets an important lower limit on the time at which the starburst phenomenon began ( @xmath47 myr ) . at this time
, the stellar uv ionisation sources have long since faded from view which essentially rules out a continuous star formation model for the nucleus , at least in our limited subset of starburst galaxies .
an interesting question is whether we can identify those sources that are in an early starburst phase and before the onset of the large - scale wind .
the second question presents difficulties for us , at least in the present analysis . in all of our agn sources
, there is evidence of circumnuclear star formation : the present data are unable to establish which source of activity is responsible for the wind . in future studies
, it may be possible to shed light on this question in one of two ways .
x - ray spectroscopy can , in principle , establish the [ @xmath1/fe ] abundance ratio of the outflowing gas .
enhanced @xmath1 abundances implicate the presence of sn ii , and therefore the central importance of the starburst in driving the outflow .
such an effect may have already been seen in dwarf starburst galaxies ( @xcite , ott , walter & brinks ( 2005 ) ) , but we are unaware of similar studies in seyfert outflow sources .
this is an important avenue for future research .
an alternative approach is to carefully trace the pressure changes over the circumnuclear regions down to small galactic radii .
oconnell & mangano ( 1978 ) noticed that the gas densities , as measured from the [ sii ] doublet , appear to decrease along the minor axis of m82 .
chevalier & clegg ( 1985 ) interpreted this observation as a change of wind pressure in a starburst - driven wind .
this idea has been explored extensively by heckman_etal . _ ( 1990 ) who used the [ sii]doublet to measure the electron densities as a function of radius in a subset of starburst galaxies . in m82
, high pressures have been inferred from the [ siii ] doublet ( houck_etal . _ ( 1984 ) , @xcite ) , from fine structure lines in the far infrared ( lugten_etal . _ 1986 ; duffy_etal . _ 1987 ) , and from the non - thermal radio continuum ( schaaf_etal . _ 1989 ) .
after certain assumptions , pressures can be inferred by combining observations of radio recombination lines and the associated free - free emission ( rodriguez & chaisson 1980 ; seaquist , bell & bignell 1985 ) .
the intrinsically higher resolution of linked radio receivers should allow the pressure profile to be traced to the smallest vlbi spatial scales .
naively , the pressure profile can be relatively steep closer to the black hole compared to an anticipated flatter profile across the energy injection region of a starburst ( chevalier & clegg 1985 ) .
thus , we may be able to distinguish between agn - powered vs. starburst - powered momentum by measuring the gas pressure as a function of galactocentric radius .
we leave this to future work . in the case of the agn sources ,
we find that the extended line emission is always powered by the accretion disk rather than by shock ionisation .
some of these sources are clearly associated with central starbursts that may well contribute momentum to the outflow in which case age - dating the onset of the outflow is more problematic . in the case of the agn - driven outflow ,
these could have commenced as recently as 10 myr or so , although if the starburst dominates , the longer timescale is preferred .
the latter case is interesting because it would indicate that the agn source has a lifetime approaching 100 myr , and that the amount of gas converted to energy is of order 10@xmath48 . in the next paper
, we combine the ionisation diagnostics with the emission - line kinematics to learn more about the energetics of the outflowing gas .
the integral field spectrograph approach is clearly very powerful for studying extended emission line sources .
we stress that our methods for deriving and comparing ionisation and dynamic timescales can be extended to a wider class of active galaxies ( e.g.quasar hosts ) , mergers ( e.g. ulirgs ) and star - forming galaxies ( e.g.dwarf irregulars )
. it will be interesting to compare the inferred timescales with the star formation histories extracted from circumnuclear stellar populations . in high - redshift galaxies where stellar populations are harder to observe , the ionisation diagnostics may provide key information . in the appendix
, we include a brief discussion about how future studies can be improved to advance this work further .
annibali , f. , greggio , l. , tosi , m. , aloisi , a. , leitherer , c. 2003 , , 126 , 2752 avedisova , v.s .
, 1972 , sva , 15 , 708 baldwin j.a , phillips m.m , & terlevich r.1981 pasp 93 5 baldwin j.a , phillips m.m , & terlevich r.1981 pasp 93 817 begelman , m.c . & bland - hawthorn , j. , 1997 , nature , 385 , 22 bergeron , j. , durret , f. , boksenberg , a , 1983 , a&a , 127 , 322 blanco , v.m . , graham , j.a . ,
lasker , b.m .
, osmer , p.s . , 1975 , 198 , l63 bland , j. , taylor , k. , atherton , p.d . , 1987 , 228 , 595 bland , j. & tully , b. , 1988 , nature , 334 , 43 bland - hawthorn , j. , wilson , a.s . ,
tully , r.b . , 1991 , , 371l , 19 bland - hawthorn , j. , & cohen , m. 2003 , apj , 582 , 246 bland - hawthorn , j. , & kedziora - chudczer , l. , 2003 , pasa , 20 , 242 bland - hawthorn , j. , & maloney , p.r . , 1999 , , 510l , 33
bland - hawthorn , j. , veilleux , s. , cecil , g.n . ,
putman , m.e . ,
gibson , b.k . , maloney , p.r . , 1998 , 299 , 611 burbidge , e.m . ,
burbidge , g.r . , 1960 , , 132 , 30 buzzoni , a. , 2005 , , 361,725 , carretta , e. , bragaglia , a. , & gratton , r. 2007 , a&a , 473 , 129 castor , j. , mccray , r. , weaver , r. , 1975 , , 200l , 107 cavaliere , a. , & padovani , p. , 1988
, , 333l 33 cecil , g. , bland , j. , tully , r.b . , 1990 , , 355 , 70 cecil , g. , bland - hawthorn , j. , veilleux , s. , 2002 , , 576 , 745 cecil , g. , bland - hawthorn , j. , veilleux , s. , filippenko , a.v . , 2001 , 555 , 338 cen , r. , & ostriker , j. p. 1999 ,
apj , 514 , 1 chevalier , r.a . , clegg , a.w .
1985 natur 317 44 coccato , l. , corsini , e.m . ,
pizzella , a. , morelli , l. , funes , j.g . ,
bertola , f. , 2004 , a&a , 416 , 507 colbert , e.j.m . ,
wilson , a.s . , &
bland - hawthorn , j. , 1994 , , 436 , 89 colina , l. ; sparks , w. b. ; macchetto , f. 1991apj ... 370 .. 102c collins , j. a. , shull , j. m. , & giroux , m. l. 2007 , apj , 657 , 271 condon , j.j . ,
condon , m.a .
, gisler , g. , puschell , j.j .
, 1982 , 252 102 cooper , j.l . , bicknell , g.v . ,
sutherland , r.s .
, bland - hawthorn , j. , 2008 , , 674 , 157 dalcanton , j. j. 2007 , apj , 658 , 941 dav , r. , oppenheimer , b. d. , & sivanandam , s. 2008 , , 391 110 demoulin , m.h . , burbidge , e.m .
, 1970 , , 159 , 799 de zeeuw , p.t .
, bureau , m. , emsellem , e. , bacon , r. , carollo , c.m . , copin , y. , davies , r.l . ,
kuntschner , h. , _ etal . _ , 2002 , , 329 513 dopita , m.a . ,
sutherland , r.s .
, 1996 , , 102 161 duffy , p.b . ,
erickson , e.f . ,
haas , m.r .
, houck , j.r .
, 1987 , , 315 , 68 edmunds , m.g . , & pagel , b.e.j . , 1982
mnras 198 1089 efstathiou , g. 2000 , mnras , 317 , 697 erb , d. k._etal._. 2006 , apj , 647 , 128 fall , s. m. 2002 , in disks of galaxies : kinematics , dynamics and perturbations , asp conf .
series , vol .
e. athanassoula , a. bosma , & r. mujica ( san francisco : asp ) , p. 389
fabbiano , g. , heckman , t. , keel , w.c . , 1990 , , 355 , 442 flohic , h.m.l.g . ,
eracleous , m. , chartas , g. , shields , j.c . ,
moran , e.c .
, 2006 , , 647 , 140 forbes , d.a . ,
boisson , c. , & ward , m.j . , 1992 , , 259 , 293 forbes , d.a . , & norris , r.p .
1998 , , 300 , 757 freeman , k.c .
, karlsson b. , lynga , g. , burrell , j.f .
, van woerden , h. , goss , w.m . , mebold , u. , 1977 , a&a , 55 , 445 fox , a. j. , wakker , b. p. , savage , b. d. , tripp , t. m. , sembach , k. r. , & bland- hawthorn , j. 2005 , apj , 630 , 332 genzel , r. , burkert , a. , bouch , n. , cresci , g. , frster schreiber , n.m .
, shapley , a. , shapiro , k. , tacconi , l.j .
_ , 2008 , apj , 687 , 59 greenhill , l.j . , booth , r.s . ,
ellingsen , s.p . ,
herrnstein , j.r . , jauncey , d.l . ,
mcculloch , p.m. , moran , j.m . , norris , r.p .
, reynolds , j.e . , tzioumis , a.k .
2003 , , 590 , 162 greenhill , l.j . , kondratko , p.t . ,
lovell , j.e.j . ,
kuiper , t.b.h . ,
moran , j.m .
, jauncey , d.l . ,
baines , g.p . , 2003 , 582l , 11 greve , a. , 2004 , a&a , 416 , 67 grimes , j.p . ,
heckman , t. , aloisi , a. , calzetti , d. , leitherer , c. , martin , c.l .
, meurer , g. , sembach , k. , strickland , d. 2009 , , 181 , 272 hameed , s. , devereux , n. , 1999 , aj , 118 , 730 heckman , t.m , armus , l.m .
, miley , g.k . , 1990 , , 74 , 833 heckman , t.m .
, robert . ,
leitherer , c. , garnett , d.r . , van der rydt , f. , 1998 , , 503 , 646 hjelm , m. , lindblad , p.o . , 1996 ,
a&a , 305 , 727 holtzman , j. a. et al .
, 1993 , aj , 106 , 1826 hoopes , c.g . , heckman , t.m .
, strickland d.k . ,
seibert , m. madore , b.f . , rich , r.m . , bianchi , l. , gil de paz , a. , 2005 , 619l 99 houck , j.r . , shure , m.a . , gull , g.e . ,
herter , t. , 1984 , , 287l , 11 hunter , d.a . , hawley , w.n .
, & gallagher , j.s .
iii , 1993 , , 106 , 1797 irwin , j.a . , sofue , y. , 1996 , , 464 , 738 israel , f.p . , 1998 , a&arv , 8 237 jimnez - bailn , e. , santos - lle , m. , dahlem , m. , ehle , m. , mas - hesse , j.m . ,
guainazzi , m. , heckman , t.m . , weaver , k.a .
2005 , a&a , 442 , 861 junkes , n. , zinnecker , h. , hensler , g. , dahlem , m. , pietsch , w. , 1995 , a&a , 294 , 8 jerseter , s. peterson , c.j . ,
lindblad , p.o . , boksenberg , a. , 1984 , a&as , 58 , 507 keeney , b. a. , danforth , c. w. , stocke , j. t. , penton , s. v. , shull , j. m. , & sembach , k. r. 2006 , apj , 646 , 951 kewley l.j .
, dopita m.a . ,
sutherland r.s .
, heisler c.a . & trevena j. 2001 apj 556 121 kirby , e. n. , simon , j. d. , geha , m. , guhathakurta , p. , & frebel , a. 2008 , apj , 685 , l43 krabbe , a. , sternberg , a. , genzel , r. , 1994 , , 425 , 72 krajnovi , d. , sharp , r. & thatte , n. , 2007 , mnras , 374 , 385 komossa , s. , burwitz , v. , hasinger , g. , predehl , p. , kaastra , j.s .
, , ikebe , y. 2003 apj 582l 15 koo , b .- c . , & mckee , c.f .
, 1992a , , 388 , 93 koo , b .-
c . , & mckee , c.f . , 1992b , , 388 , 103 kraft , r.p . ,
hardcastle , m.j . ,
sivakoff , g.r . ,
jordn , a. , nulsen , p.e.j . ,
birkinshaw , m. , forman , w.r . ,
jones , c. , _ etal . _ 2008 , 677l 97 lada , c.j . , & lada , e.a .
, 2003 , ara&a , 41 , 57 leitherer , c. , robert , c. , drissen , l. , 1992 , , 401 , 596 a&a , 467 , 283 lehnert , m.d . , &
heckman , t.m .
1996 , apj , 462 , 651 lugten , j.b . ,
watson , d.m . ,
crawford , m.k .
, genzel , r. , 1986 , , 311l , 51 laurent , o. , mirabel , i.f . ,
charmandaris , v. , gallais , p. , madden , s.c . , sauvage , m. , vigroux , l. , cesarsky , c. , 2000 , a&a , 359 , 887 lutz , d. , maiolino , r. , moorwood , a.f.m . , netzer , h. , wagner , s.j . ,
sturm , e. , genzel , r. , 2002 , a&a , 396 , 439 mac low , m .-
m . & mccray , r. , 1988 , , 324 , 776 madau , p. , ferrara , a. , & rees , m. j. 2001 , apj , 555 , 92 maloney , p.r . , 1999 , ap&ss , 266 , 207 martin , c.l . , 2006 ,
apj , 647 , 222 martin , c. l. , kobulnicky , h. a. , & heckman , t. m. 2002 , apj , 574 , 663 matt , g. , fabian , a.c . , guainazzi , m. , iwasawa , k. , bassani , l. , malaguti , g. , 2000 , 318 , 173 mccarthy , p.j . , van breugel .
w. , heckman , t. , 1987 , 93 , 264 meurer , g.r .
, freeman , k.c . , dopita , m.a .. cacciari c. , 1992 , , 103 , 60 morganti , r. , holt , j.,saripalli , l. , oosterloo , t.,a . , & tadhunter c.n .
2007 a&a 476 723 morganti , r. , oosterloo , t. , tsvetanov , z , 1998 , , 115 , 915 mueller - snchez , f. , davies , r.i . ,
eisenhauer , f. , tacconi l.j . , genzel , r. , sternberg , a. , 2006 , a&a , 454 , 481 norman , c. , silk , j. , 1983 , , 266 , 502 navarro , j. f. , & white , s. d. m. 1994 , mnras , 267 , 401 nicholson , r.a . ,
bland - hawthorn , j. ; taylor , k. , 1992 , 387 , 503 oconnell , r.w . , & mangano , j.j . , 1978 , , 221 , 62 oliva , e. , salvati , m. , moorwood , a.f.m .
, marconi a. , 1994 , a&a , 288 , 457 osterbrock , d.e .
, `` astrophysics of gaseous nebulae and active galactic nuclei '' , university science books ott , j. , walter , f. , & brinks , e. , 2005 , , 358 , 1453 persic , m. , mariani , s. , cappi , m. , bassani , l. , danese , l. , dean , a.j .
, di cocco , g. , franceschini , a. , _ etal . _ , 1998 , a&a , 339l 33 peterson , b.a . ,
dickens , r.j . , cannon , r.d.m 1975 , pasau , 2 , 366 phillips , a.c . , 1993 ,
aj , 105 , 486 phillips , m.m . ,
edmunds , m.g . , pagel , b.e.j . ,
turtle , a.j .
, 1983 , , 203 , 759 pogge , r.w . & eskridge , p.b .
, 1987 , , 93 , 291 ptak , a. , heckman , t. , levenson , n.a . ,
weaver , k. , strickland , d. , 2003 , , 592 , 782 quillen , a.c . ,
brookes , m.h . ,
keene , j. , stern , d. , lawrence , c.r . , werner , m.w .
, 2006 , , 645 , 1092 quillen , a.c . ,
graham , j.r . ,
frogel , j.a . , 1993 , , 412 , 550 rand , r.j . , 1998 ,
apj , 501 , 137 rice , w. , lonsdale , c.j . , soifer , b.t .
, neugebauer , g. , kopan , e.l . , lloyd , l.a .
, de jong , t. , habing , h.j . , 1988 apjs , 68 , 91 rieke , g.h . , lebofsky , m.j . , thompson , r.i . , low , f.j . , tokunaga , a.t .
, 1980 , , 238 , 24 robertson , b. et al .
2006 , apj , 645 , 986 rodriguez , l.f .
& chaisson , e.j . , 1980 , , 238 , 41 ryan - weber , e. v. , pettini , m. , & madau , p. 2006
, mnras , 371 , l78 saunders , w. , bridges t. , gillingham , p. , _ etal .
_ , 2004 , spie , 5492 , 389 sakamoto , k. , ho , paul t.p . , iono , d. , keto , e.r . ,
mao , r. , matsushita , s. , peck , a.b . ,
wiedner , m.c . ,
wilner , d.j . ,
zhao , j. 2006 , 636 , 685 sandqvist , a. , jerseter , s. , lindblad , p.o . , 1995 , a&a , 295 , 585 schaaf , r. , pietsch , w. , biermann , p.l . ,
kronberg , p.p . ,
schmutzler , t. , 1989 , , 336 , 722 schmelz , j. t. ; baan , w. a. ; haschick , a. d. 1987 , , 320 145 schmelz , j. t. ; baan , w. a. ; haschick , a. d. 1987 , , 315 492 schiano , a.v.r . , 1985 , , 299 , 24 schwartz , c.m . , & martin , c.l . , 2004 ,
apj , 610 , 201 schwartz , c.m . , martin , c.l . ,
chandar , r. , leitherer , c. , heckman , t.m . ,
oey , m.s .
2006 , apj , 646 , 858 seaquist , e.r . , bell , m.b . ,
bignell , r.c . , 1985 , , 294 , 546 sedov , l.i . , 1958 ,
rvmp , 30 , 1077 sharp , r.g . , _ etal .
_ , 2006 , spie , 6269e , 14 shopbell , p.l . & bland - hawthorn , j. , 1998 , , 493 , 129 shull & mckee , 1979 , , 227 , 131 siebenmorgen , r. , haas , m. , pantin , e. , kr gel , e. , leipski , c. , k ufl , h.u . , lagage , p.o . , moorwood , a. , smette , a. , sterzik , m. 2008 , a&a , 488 , 83 smith , l.j . ,
westmoquette , m.s . ,
gallagher , j.s .
, oconnell , r.w . , rosario , d.j . ,
de grijs , r. , 2006 , mnras , 370 , 513 soltan , a. , 1982 , , 200 , 115 steigman , g. , strittmatter , p.a . , williams , r.e . , 1975 , , 198 , 575 steinmetz , m. , & mller , e. 1995 , mnras , 276 , 549 stevens , i.r . ,
forbes , d.a . ,
norris , r.p . , 1999 , , 306 ,
479 strickland , 2007 , mnras , 376 , 523 strickland , d.k
. , heckman , t.m , colbert , e.j.m . ,
hoopes , c.g . , & weaver , k.a . , 2004 , apjs , 151 , 193 strickland , d.k . , heckman , t.m . ,
colbert , e.j.m . , hoopes , c.g . , weaver , k.a .
, 2004 , apj , 606 , 829 strickland , d.k . , heckman , t.m . , weaver , k.a . , & dahlem , m. 2000 , aj , 120 , 2965 strickland , d.k . , & heckman , t.m . , 2009 , , 697 2030 strickland , d.k .
, & stevens , i.r . , 2000 , , 314 , 511 storchi - bergmann , t. , bonatto , c.j .
, 1991 , 250 , 138 suchkov , a.a . ,
balsara , d.s . ; heckman , t.m . , leitherer , c. 1994 , , 430 , 511 sutherland , r.s . , bicknell , g.v . ,
dopita , m.a . , 1993 , , 414 , 510 tadhunter , c. & tsvetanov , z. , 1989 , nature , 341 , 422 taylor , g. , 1950 , rspsa , 201 , 159 terlevich , r. , & melnick , j. , 1985 , 213 841 tomisaka , k. , & ikeuchi , s. , 1988 , , 330 , 695 tremonti , c. a. et al . 2004 , apj , 613 , 898 tully , b.r .
, `` nearby galaxies catalog '' , 1988 , cambridge university press veilleux & bland - hawthorn j. , 1997 , apjl , 479 , 105 veilleux , s. , cecil , g. , bland - hawthorn , j. 2005 , ara&a , 43 , 769 veilleux , s. , cecil , g. , bland - hawthorn , j. , tully , r.b . ,
filippenko , a.v . ,
sargent , w.l.w .
, 1994 , , 433 , 48 veilleux , & osterbrock , 1987 , apjs , 63 , 295 veilleux s. , & rupke , d.s . , 2002 , , 565 , 63l veilleux , s. , shopbell , p.l . ,
rupke , d.s . ,
bland - hawthorn , j. & cecil , g. , 2003 , aj , 126 , 2185 veron - cetty , m .- p . , & veron , p , 1985 , a&a , 145 , 425 veron , p. , lindblad , p.o .
, zuiderwijk , e.j . ,
veron , m.p . ,
adam , g. , 1980 , a&a , 87 , 245 ward , m.j . ,
wilson , a.s . ,
penston , m.v . , elvis , m. , maccacaro , t. , tritton , k.p . , 1978 , apj , 223 , 788 weaver , r. , mccray , r. , castor , j. , shapiro , p. , moore , r. , 1977 , , 218 , 377 westmoquette , m.s . ,
smith , l.j . ,
gallagher , j.s . ,
trancho , g. , bastian , n. , konstantopoulos , i.s . , 2009 , , 696 , 192 westmoquette , m.s . ,
gallagher , j.s . ,
smith , l.j . ,
trancho , g. , bastian , n. , konstantopoulos , i.s . , 2009 , , 706 , 1571 wilson , shopbell , simpson _ etal .
_ 2000 , aj , 120 , 1325 + + + + + + ngc 253 : line ratio maps are constructed following fig .
[ ngc1482 ratios].,title="fig:",width=302 ] ngc 253 : line ratio maps are constructed following fig .
[ ngc1482 ratios].,title="fig:",width=302 ] ngc 253 : line ratio maps are constructed following fig . [ ngc1482 ratios].,title="fig:",width=302 ] ngc 253 : line ratio maps are constructed following fig . [ ngc1482 ratios].,title="fig:",width=302 ] ngc 253 : line ratio maps are constructed following fig . [ ngc1482 ratios].,title="fig:",width=302 ] + + + ngc 1808 : line ratio maps are constructed following fig . [ ngc1482 ratios ] . observations of ngc 1808 were undertaken during a period of poor atmospheric transparency and so emission line maps suffer reduced sensitivity.,title="fig:",width=264 ] ngc 1808 : line ratio maps are constructed following fig . [ ngc1482 ratios ] . observations of ngc 1808 were undertaken during a period of poor atmospheric transparency and so emission line maps suffer reduced sensitivity.,title="fig:",width=264 ] ngc 1808 : line ratio maps are constructed following fig . [ ngc1482 ratios ] . observations of ngc 1808 were undertaken during a period of poor atmospheric transparency and so emission line maps suffer reduced sensitivity.,title="fig:",width=264 ] + + ngc 3628 : line ratio maps are constructed following fig . [ ngc1482 ratios ] . due to the reduce sensitivity of the observations for ngc 3628 , only the [ nii]/h@xmath1 map is recovered.,width=264 ] + + + + + + + + + + + + + + + + + + + ionizing luminosity ( black ) vs. mechanical luminosity ( blue ) evolution in two different star forming scenarios : ( a ) instantaneous burst ; ( b ) continuous . in both cases ,
a salpeter initial mass function is adopted but with two different upper mass cut - offs : 30 m@xmath49 ( dotted line ) , 100 m@xmath49 ( solid line ) . , width=453 ] predicted evolution of h@xmath1 surface brightness of optically thick gas at 1 kpc from a central active source with @xmath50 .
the black and blue curves are described in fig .
[ starburst99a ] .
the red curves correspond to 4 different sets of @xmath51 in our agn model , i.e. ( 0.1 , 20 ) , ( 0.2 , 10 ) , ( 0.4 , 5 ) , ( 0.8 , 2.5 ) where the second value in each pair is the exponential timescale in myr ; larger values therefore produce slower decays .
the predicted values are at least an order of magnitude higher than what we observe as we discuss in 4 .
we do not include the evolution of the agn s mechanical luminosity for comparison because this is highly uncertain .
but since agn photoionization seems to dominate over shocks , we must assume that the mechanical luminosity profile falls below the agn uv curve at all times and therefore does not resemble the starburst profile.,width=453 ]
the use of a double - beam integral field spectrograph has been critical to the success of this project .
it provides both detailed ionisation diagnostics and kinematics over the extent of the outflowing gas .
we have demonstrated how these idds can be used to classify the nature of the wind outflow for each source , and to define spatial zones in which different ionisation sources dominate .
furthermore , we have begun the process of spectral and spatial kinematic and ionisation deconvolution for each source , in an effort to disentangle the complexity of each system .
we conclude that blue sensitivity ( below 500 nm ) is a key factor to many of the diagnostic requirements , in particular sensitivity to the h@xmath0 and [ oiii ] spectral region ( with spiral we can not comment on the utility of [ oii ] emission feature for such work ) . as a clear shock diagnostic ,
[ oi ] is also of great value to classification of the nature of the ionisation source , although in all cases presented in this work , sensitivity to the [ oiii ] emission still presents the primary limitation .
this limitation is imposed in part by the reduced throughput of the blue arm of the aaomega system @xcite but will remain an important consideration for more balanced systems .
spectral resolution is critical to disentangling the multiple emission components of many of the sources discussed , with line splitting in many of the outflows only marginally resolved in our r@xmath125000 spectroscopy ( @xmath1250 km s@xmath8 ) .
one must of course also consider wavelength coverage . while many of the fainter lines common to ism studies may be beyond the reach of a survey such as ours ( for example the temperature sensitive [ nii ] @xmath525756 line is typically rather faint ) we have demonstrated the added value of an extended spectral range in contrast to the limited coverage usually practical with fabry - prot tunable filter systems such as that employed for example by @xcite
. a range of source redshifts , even for a nearby galaxy sample , further complicates the choice of wavelength range selection .
the observations and interpretation presented here have been possible primarily due to the high surface brightness sensitivity of the aaomega - spiral system afforded by the large projected spatial pixel size on the sky .
the 0.7arcsec on a side square elements provide an individual field of view of 0.49arcsec@xmath53 , which represents factors of @xmath122 , 5.4 & 12 increase in contrast to 0.5,0.3 or 0.2arcsec systems , more than accounting for the smaller size of the prime mirror of the aat 3.9 m when compared to systems available on the current generation of large optical telescopes .
this allows spiral to move out of the detector readnoise limit , below which re - binning of adjacent apertures will not recover signal - to - noise quickly . as recently demonstrated by westmoquette _ etal .
_ 09a / b , in truth one must ultimately combine the wide field of view and low surface brightness observations practicable with large pixels , with dedicated narrow field observations at higher spatial resolution of critical regions of each object . | in recent years , we have come to recognize the widespread importance of large - scale winds in the lifecycle of galaxies .
the onset and evolution of a galactic wind is a highly complex process which must be understood if we are to understand how energy and metals are recycled throughout the galaxy and beyond .
here we present 3d spectroscopic observations of a sample of 10 nearby galaxies with the aaomega - spiral integral field spectrograph on the 3.9 m aat , the largest survey of its kind to date .
the double - beam spectrograph provides spatial maps in a range of spectral diagnostics : [ oiii]5007 , h@xmath0 , mg__b _ _ , nad , [ oi]6300 , h@xmath1 , [ nii]6583 , [ sii]6717 , 6731 .
we demonstrate that these flows can often separate into highly ordered structures through the use of ionisation diagnostics and kinematics .
all of the objects in our survey show extensive wind - driven filamentation along the minor axis , in addition to large - scale disk rotation .
our sample can be divided into either starburst galaxies or active galactic nuclei ( agn ) , although some objects appear to be a combination of these .
the total ionizing photon budget available to both classes of galaxies is sufficient to ionise all of the wind - blown filamentation out to large radius .
we find however that while agn photoionisation always dominates in the wind filaments , this is not the case in starburst galaxies where shock ionisation dominates .
this clearly indicates that after the onset of star formation , there is a substantial delay ( @xmath2 10 myr ) before a starburst wind develops .
we show why this behavior is expected by deriving `` ionisation '' and dynamical timescales for both agns and starbursts .
we establish a sequence of events that lead to the onset of a galactic wind . the clear signature provided by
the ionisation timescale is arguably the strongest evidence yet that the starburst phenomenon is an impulsive event .
a well - defined ionisation timescale is not expected in galaxies with a protracted history of circumnuclear star formation .
our 3d data provide important templates for comparisons with high redshift galaxies . |
for heavy charged particle radiotherapy with protons and ions , broad - beam delivery methods ( coutrakon 1991 , kanai 1999 ) are mature technologies with persistent advantages of simplicity and robustness over emerging technologies of pencil - beam scanning methods ( lambert 2005 ) . for a broad - beam system ,
a variety of volumetrically enlarged standard beams are prepared , among which an optimum one is applied to a given target .
target - specific customization is usually made with x - jaw , y - jaw , and multileaf collimators ( xjc , yjc , and mlc ) and custom - made accessories such as a patient collimator ( ptc ) and a range - compensating filter ( rcf ) .
while the downstream collimators form sharp field edges , the upstream collimators , which are mainly for radiation - protection purposes , form gentle field edges
. their combination will be useful for field patching techniques to form an irregular field with gently joining beams for improved robustness ( li 2007 ) . in treatment planning ,
a variety of pencil - beam ( pb ) algorithms are used for dose calculation ( hong 1996 , kanematsu 2006 ) despite intrinsic difficulty with the pencil beams that develop to overreach lateral heterogeneity ( goitein 1978 , petti 1992 , kohno 2004 ) . for electron radiotherapy
, the phase - space theory was rigorously applied to resolve the problem by periodical redefinition of ensemble of minimized pencil beams in the pb - redefinition algorithm ( shiu and hogstrom 1991 ) .
the principle of pb redefinition was applied to heavy charged particles to address the effects of multiple collimators in the monochromatic pb approximation ( kanematsu 2008b ) .
however , its rigorous application to a heterogeneous system requires polychromatic energy spectra , which would be computationally demanding for heavy charged particles with sharp bragg peaks
. it will be thus difficult to cope with range compensation or patient heterogeneity in that approach .
recently , kanematsu ( 2009 ) proposed an alternative approach , the pb - splitting algorithm , where monochromatic pencil beams dynamically split into smaller ones near a lateral density interface .
automatically , fine pencil beams are densely arranged only where they are necessary while otherwise large pencil beams are sparsely arranged for efficient dose calculation . in conjunction with the grid - dose - spreading convolution ( kanematsu 2008a ) ,
the pb - splitting algorithm demonstrated feasibility of accurate patient dose calculation while minimizing the impact of recursive beam multiplication ( kanematsu 2011 ) . in this study
, we further extend the pb - splitting approach to beam - customization devices to deal with their physical structures accurately and efficiently and to complete a consistent algorithmic framework for dose calculation in treatment planning . in the following sections , we define the model elements that were mostly diverted from previous studies ,
construct a novel and original beam - customization model , and examine its validity for a test - beam experiment .
a beam source is defined as the best approximate point from which radiating particles will have the same fluence reduction with distance .
the formulation differs among beam - spreading methods and often between the transverse @xmath0 and @xmath1 axes , _
i.e. _ , at height @xmath2 for @xmath0 and @xmath3 for @xmath1 .
the particles incoming to a point in the field , which is normally the isocenter , are projected back onto the @xmath0 and @xmath1 source planes to define rms source sizes @xmath4 and @xmath5 in the gaussian approximation .
although a range - modulated beam should be ideally subdivided into energy components of different source heights and sizes , it is approximately represented by a single component of average behavior in this study .
a beams - eye - view ( bev ) image is defined as an @xmath6 matrix of @xmath7 square - sized pixels starting at @xmath8 on the isocenter plane . for bev pixel @xmath9 , pixel position @xmath10 and the line connecting to the @xmath0 and @xmath1 sources are defined as @xmath11 following the thick - collimator model ( kanematsu 2006 ) , two identical apertures on the top and bottom faces are associated with every collimator , which are modeled as two - dimensional bitmaps .
matrix @xmath12 describes aperture @xmath13 with @xmath14 elements of transmission @xmath15 1 ( transmit ) or 0 ( block ) for @xmath16 $ ] and @xmath17 $ ] .
the pixel-@xmath18 position is given by @xmath19 where @xmath20 and @xmath21 are the first pixel position and the square pixel size of the bitmap image .
for arbitrary point @xmath22 on the aperture plane , intersecting pixel @xmath18 is determined with the nearest integer function @xmath23 as @xmath24 and the distance to the nearest aperture edge , @xmath25 is quickly referenced from the distance map filled by the distance - transform algorithm ( borgefors 1986 ) .
a rcf made of a tissue - like material of effective density @xmath26 is similarly described by an @xmath27 matrix of range shifts @xmath28 , first pixel position @xmath29 , and pixel size @xmath30 . in this study
, we deal with a single rcf of a flat downstream face at height @xmath31 . the stopping and scattering effects of the rcf
are approximated by a local interaction at the midpoint of the beam path in the structure ( gottschalk 1993 ) , _ i.e. _ , at height @xmath32 for rcf pixel @xmath33 .
following the original pb - splitting algorithm ( kanematsu 2009 ) , the present pb model is based on the fermi - eyges theory ( eyges 1948 ) for stopping and scattering ( kanematsu 2009 , gottschalk 2010 ) excluding hard interactions that are implicitly included in the depth
dose curve .
a gaussian pencil beam is characterized by position @xmath34 , direction @xmath35 , number of particles @xmath36 , residual range @xmath37 , and phase - space variances of the projected angle @xmath38 and transverse displacement @xmath39 , which develop in a tissue - like medium by step @xmath40 as @xmath41 \delta s , \label{eq:7}\end{aligned}\ ] ] where @xmath42 is the stopping - power ratio of the medium to water ( kanematsu 2003 ) and @xmath43 and @xmath44 are the particle mass and charge in units of those of a proton . to limit excessive beam multiplication , pencil beams subject
to splitting should have sufficient particles , _
i.e. _ , @xmath45 , where @xmath46 is the number for the original beam and @xmath47 is a cutoff .
when a pencil beam of rms size @xmath48 spreads beyond the lateral density interface at distance @xmath49 from the beam center , it splits into @xmath50 daughter beams downsized by factor @xmath51 as @xmath52 where @xmath53 is a parameter that limits the fraction of overreaching particles . with respect to the mother beam ,
daughter @xmath54 ( @xmath55 $ ] ) is downscaled , displaced , redirected , and downsized while conserving focal distance @xmath56 and local mean square angle @xmath57 , as @xmath58,\label{eq:10}\\ \fl \delta \vec{r}_{\alpha \beta } = \sigma_t \left ( { \mu_m}_\alpha \vec{e}_t + { \mu_m}_\beta \vec{e}_u \right),\qquad \left(\begin{array}{c}{\bm \mu}_2\\ { \bm \mu}_3\\ { \bm \mu}_4\end{array}\right ) = \left(\begin{array}{cccc } -1/2 & 1/2\\ -1 & 0 & 1 \\ -3/2 & -1/2 & 1/2 & 3/2 \end{array}\right),\label{eq:11}\\ \fl \overline{t^2}_{\alpha \beta } = \sigma_m^2\,\overline{t^2},\qquad \overline{\theta t}_{\alpha\beta } = \sigma_m^2\,\overline{\theta t},\qquad \overline{\theta^2}_{\alpha\beta } = \overline{\theta^2}-(1-\sigma_m^2)\,\overline{\theta t}^2/\overline{t^2 } , \label{eq:12}\end{aligned}\ ] ] where @xmath59 is the share of daughter @xmath60 and @xmath61 and @xmath62 are the displacement factors for transverse directions @xmath63 and @xmath64 . for every bev pixel @xmath9 , pencil beam @xmath65 is placed and the pb parameters are defined at the effective interaction point in the rcf as @xmath66^{-1/2}\left(\frac{x_j}{z_{{\rm s}x}},\frac{y_i}{z_{{\rm s}y } } , 1\right ) , \label{eq:14 } \\
\fl n_b = { n_0}_b = { \phi_0}_{i j}\ , \delta^2 , \qquad r = r_0 , \label{eq:15 } \\
\fl \overline{\theta^2}_b = \frac{1}{2 } \left(\frac{\sigma_{{\rm s}x}}{z_{{\rm s}x}-{z_0}_{i j}}\right)^2 + \frac{1}{2 } \left(\frac{\sigma_{{\rm s}y}}{z_{{\rm s}y}-{z_0}_{i j}}\right)^2 , \label{eq:16 } \\
\fl \overline{\theta t}_b = \frac{\overline{t^2}}{\sqrt{z_{{\rm s}x}-{z_0}_{i j}}\sqrt{z_{{\rm s}y}-{z_0}_{i j}}},\qquad \overline{t^2 } = \frac{z_{{\rm s}x}-{z_0}_{i j}}{z_{{\rm s}x } } \frac{z_{{\rm s}y}-{z_0}_{i j}}{z_{{\rm s}y } } \frac{\delta^2}{12 } , \label{eq:17}\end{aligned}\ ] ] where rcf intersection pixel @xmath33 is determined in analogy with and and open - field fluence @xmath67 and range @xmath68 are usually given by measurement . and axes @xmath0 , @xmath69 , @xmath70 , and @xmath39 ) in the @xmath69@xmath0 view , where symbols _ s _ , _ b _ , and _ p _ indicate the source , a pencil beam , and a particle , respectively.,width=8 ] the upstream collimators restrict angular acceptance of particles incoming to each pb origin .
a pencil beam will be fully blocked when it is far away from any one of the apertures or fully transmitted when it is in the middle of all of them , or @xmath71 \right\}$\\ 1 & for \quad $ \left\ { ^\forall a \mid ( z_a > z_0 ) ; \left({t_a}_{i_a j_a } = 1\right ) \wedge \left [ { d_a}_{i_a j_a } > 3\,{\sigma_\theta}_b ( z_a - z_0 ) \right ] \right\}$ , } \label{eq:18}\end{aligned}\ ] ] where @xmath72 is the transmission factor of beam @xmath65 , @xmath18 is the beam intersection pixel of aperture @xmath13 , and factor @xmath73 to rms projected angle @xmath74 secures three standard deviations for edge distance @xmath75 . for partial transmission , we calculate the geometrical acceptance of particles incoming to the pb origin . as shown in , with small orthogonal angles @xmath38 and @xmath76 about the pb axis , constituent - particle direction @xmath77 is defined as @xmath78 which translates into geometrical line @xmath79,@xmath80 , or @xmath81 only particles passing through all the apertures can get to the pb origin to redefine the number of particles , the direction , and the mean square angle as @xmath82 where @xmath83 is the aperture pixel in which line @xmath84 intersects . in practice , these integrals are made numerically at @xmath85 sampling intervals for @xmath86 regions .
the rcf shortens the residual range of the pencil beam by the thickness of the intersecting pixel as @xmath87 and increases the mean square angle by @xmath88 in before the beam is transported downstream .
every pencil beam is individually transported by through downstream apertures . at an aperture , which is practically either the top or bottom face of an optional ptc ,
pencil beams near the edge will be partially transmitted .
incidentally , edge distance @xmath75 in naturally corresponds to density - interface distance @xmath49 in for pb splitting . at every downstream aperture , multiplicity @xmath89
is appropriately determined while limiting overreaching particles to below 2% by setting @xmath90 . in the case of splitting ,
daughter beams are defined according to and then individually transported downstream starting from the current aperture with possible recursive splitting in the same manner .
the pencil beams that are finally out of the aperture will be blocked by setting @xmath91 , which addresses the partial - blocking effect of the collimator . cross - section view , ( b ) the beam s eye view on the isocenter plane , where the filled areas represent xjc , yjc , mlc , and ptc from upstream to downstream and the hatched area represents rcf.,width=10 ]
as this study shares the objective of beam - customization modeling with the former study in the pb - redefinition approach ( kanematsu 2008b ) , we use the same experimental data , where a broad carbon - ion beam of residual range @xmath92 cm in water was customized with an xjc at height 117137 cm , a yjc at 96116 cm , and a partially effective mlc at 6983 cm , a 3-cm pmma half - plate rcf at 3538 cm , and an 8-cm - square ptc at 2227 cm as shown in .
four lines of the in - air dose profiles on the isocenter plane were measured along the @xmath0 axis at @xmath93 cm and @xmath94 cm and along the @xmath1 axis at @xmath95 cm and @xmath94 cm .
the 20%80% penumbra sizes ( @xmath96 ) were 0.58 cm for the xjc edge and 0.48 cm for the yjc edge , which translate into rms source sizes @xmath97 cm at @xmath98 cm and @xmath99 cm at @xmath100 cm .
the tissue - air ratio for the 3-cm pmma ( @xmath101 ) was measured to be 0.951 . in the calculation , @xmath102 dose grids in a single layer were arranged on the isocenter plane at 1-mm intervals .
the open field of uniform fluence ( @xmath103 ) was subdivided into the bev image pixels of size @xmath104 mm on the isocenter plane , to each of which a pencil beam was defined at the effective scattering point of the rcf .
for every pencil beam , upstream collimation by the xjc , the yjc , and the mlc , range shift and scattering by the rcf , and beam transport including collimation and splitting by the ptc down to the isocenter plane were applied .
the in - air dose distribution on the isocenter plane was calculated with @xmath105 where @xmath106 is the tissue - air ratio or the dose per fluence for beam @xmath65 . to verify the effectiveness of pb splitting for the ptc edge
, we calculated dose distributions at heights 0 cm ( isocenter plane ) and 20 cm ( immediate downstream ) by relocating the dose grids and compared them with corresponding non - splitting calculations , for which we disabled splitting by setting @xmath107 .
in the calculation , 40000 beams were originally defined at the rcf , 23912 of them passed through the upstream collimators , and 20444 of them passed through the ptc to end up with 36704 dose - contributing beams by splitting .
the cpu time of a 2.4 ghz intel core 2 duo processor amounted to 1.30 s and 1.25 s for the calculations with and without pb splitting .
shows the calculated dose distribution . the dip and
bump along the @xmath1 axis are attributed to scattering by the pmma half plate .
sharpness of the field edge was strongly correlated with the distance to the effective collimator .
axis at @xmath93 cm ( a ) , @xmath108 cm ( b ) , along the @xmath1 axis at @xmath109 cm ( c ) , and @xmath110 cm ( d ) , where the solid lines are the calculations and the open circles are the measurements.,width=8 ] in the experiment , the uncertainty of the scanned detector positions was 0.1 mm and that of the collimator positions was @xmath111 mm according to the specifications .
the latter may only shift the edge position and will not influence the penumbra size .
the single - point dose uncertainty was evaluated to be 0.3% in repeated measurements , which is negligible for penumbra analysis .
shows the calculated and measured doses profiles , where the measured doses are in fact the dose ratios of the customized field to the open field to compensate for the fluence non - uniformity .
unexpectedly , the customized - field doses were higher than the open - field doses by a few percent . that may be attributed to the contribution of particles hard - scattered by the collimators , which was not considered in the present model . from these profiles ,
the 20%80% penumbra sizes were obtained by reading 20% and 80% dose positions by linear interpolation of two sampling points , which brings dominant uncertainty amounting to a fraction of the sampling interval of 1 mm .
the measured penumbra sizes were then corrected to quadratically exclude @xmath112 with effective dosimeter size @xmath113 mm for a 2-mm@xmath76 pinpoint chamber .
summarizes the resultant penumbra sizes .
these measurements and calculations agreed to a submillimeter level , which is consistent with the estimated uncertainty .
l l l l l profiling & interested- & effective & + position & edge side & device(s ) & measurement & calculation + @xmath114 cm & @xmath0 left & xjc+rcf & 6.4 & 6.7 + @xmath114 cm & @xmath0 right & xjc & 5.8 & 5.8 + @xmath115 cm & @xmath0 left & mlc+rcf & 4.6 & 4.4 + @xmath115 cm & @xmath0 right & mlc & 3.7 & 3.2 + @xmath116 cm & @xmath1 lower & ptc+rcf & 2.3 & 2.6 + @xmath116 cm & @xmath1 upper & yjc+rcf & 5.6 & 5.7 + @xmath117 cm & @xmath1 lower & ptc & 1.4 & 1.3 + @xmath117 cm & @xmath1 upper & yjc & 4.8 & 4.5 + [ tab:1 ] the effectiveness of beam splitting for the ptc - edge sharpening is shown in , where panels ( c ) and ( d ) show enlarged views of panels ( c ) and ( d ) in with additional lines for non - splitting calculations .
the pb splitting reasonably sharpened the field edges at the immediate downstream and made better agreement with the measurements on the isocenter plane .
ironically , the contamination of collimator - scattered particles happened to compensate substantially for the lack of edge sharpening in the tail regions .
it is one of the algorithmic novelties of this study to originate the pencil beams at the effective scattering points of the rcf regardless of upstream collimation .
then , the lateral heterogeneity of the rcf is naturally irrelevant to the minimized pencil beams .
upstream collimation is reasonably modeled as filtering of particles in the angular distribution to correct the phase - space parameters of the defined pencil beams .
while the pb size and density are generally arbitrary in pb algorithms , small size and high density are required to represent sharp edges of downstream collimation . in the present model , the sharp ptc edge was naturally realized by splitting of the pencil beams . in the former study ( kanematsu 2008b ) , because the pencil beams could not be redefined as monochromatic after range compensation , they were only artificially downsized for edge sharpening .
the downsizing strength was empirically determined to reproduce the 20%80% penumbra size on the isocenter plane while overlooking the other aspects .
in fact , while the resultant penumbra sizes were equivalently good for both models , the dose profiles in the upstream was unphysically bouncy in the former study due to insufficient density , which could be clinically problematic . in the original pb - splitting algorithm ( kanematsu 2009 )
, the overreaching condition was defined as the one - standard - deviation distance ( @xmath118 ) to a 10% density change .
that was because its objective heterogeneity was moderate density variation among body tissues .
this study deals with solid and precisely defined collimator edges , for which the @xmath119 distance to an aperture edge may be more appropriate . in the present example , the ptc was effective only for approximately 1/4 of the field edge .
the pb splitting was limited to the pencil beams around the effective edge and actually increased the number of beams by 80% and the cpu time by 4% .
this discrepancy is mainly attributed to computational overhead for generation , upstream collimation , and range compensation of the pencil beams .
although we only dealt with the planer grids in this case , the pb splitting would not add severe computational load even for volumetric grids when used with the grid - dose - spreading convolution ( kanematsu 2011 ) . in heavy charged particle radiotherapy ,
target doses are predominantly formed by bragg peaks of primary particles .
hard - scattered particles are generally out of the scope of practical pb algorithms due to difficulty in their modeling .
fortunately , the collimator - scattered particles tend to lose large energy in the collimator and thus naturally attenuate with depth ( van luijk 2001 ) .
nevertheless , kimstrand ( 2008 ) included the collimator - scatter contribution in a convolution algorithm using monte - carlo - generated kernels .
that approach may be valid and will further improve the accuracy if combined with the present model .
we have developed a calculation model for customization of a broad beam of heavy charged particles based on the pb - splitting algorithm . in this model , a broad beam is decomposed into pencil beams of various size that is necessarily and sufficiently small to deal with structures of the beam - customization devices accurately and efficiently . also , placement of the pb origins at the effective scattering points in the rcf effectively reduced the relevant heterogeneity and greatly simplified the algorithm using only monochromatic pencil beams .
the performance of the model was tested against existing experimental data , which demonstrated that the penumbra size for various collimator edges in a single field was accurate to a submillimeter level .
this beam - customization part can be naturally combined with the patient - dose - calculation part that is similarly based on the pb - splitting algorithm ( kanematsu 2011 ) to complete an accurate and efficient dose calculation algorithm for treatment planning of heavy - charged - particle radiotherapy .
coutrakon g , bauman m , lesyna d , miller d , nusbaum j , slater j , johanning , j and miranda , j and deluca jr p m and siebers j 1991 a prototype beam delivery system for the proton medical accelerator at loma linda _ med .
phys . _ * 18 * 10939 | a broad - beam - delivery system for heavy - charged - particle radiotherapy often employs multiple collimators and a range - compensating filter , which potentially offer complex beam customization . in treatment planning
, it is however difficult for a conventional pencil - beam algorithm to deal with these structures due to beam - size growth during transport .
this study aims to resolve the problem with a novel computational model .
the pencil beams are initially defined at the range compensating filter with angular - acceptance correction for the upstream collimators followed by the range compensation effects .
they are individually transported with possible splitting near the downstream collimator edges to deal with its fine structure .
the dose distribution for a carbon - ion beam was calculated and compared with existing experimental data .
the penumbra sizes of various collimator edges agreed between them to a submillimeter level .
this beam - customization model will complete an accurate and efficient dose - calculation algorithm for treatment planning with heavy charged particles . |
a very fruitful information on the dynamics can be gained from the study of periodic orbits @xcite .
first , because these particular orbits are generically almost everywhere in phase space , and second because they can be computed easily , i.e. with some short integration time .
these periodic orbits together with their stability organize locally the dynamics .
it is then natural to consider them as a cornerstone of control strategies .
for instance , in order to create invariant tori of hamiltonian systems , cary and hanson @xcite proposed a method based on the computation of an indicator of the linear stability of a set of periodic orbits , namely greene s residue @xcite .
it provides an algorithm to find the appropriate values of some pre - defined parameters in order to reconstruct invariant tori by vanishing some selected residues .
first developed for two - dimensional symplectic maps , it has been extended to four dimensional symplectic maps , and has been applied to stellarators @xcite ( where periodic orbits are closed magnetic field lines ) and particle accelerators @xcite . in this article
, we review and extend this residue method .
the aim is to tune appropriately the parameters of the system such that appropriate bifurcations occur .
it is well - known in the literature that local bifurcations occur when the tangent map associated with the poincar map obtained by a transversal intersection of the flow , has an eigenvalue which is a root of the unity . in particular
, periodic orbits can lose their stability in case of multiple eigenvalues on the unit circle , i.e. when these eigenvalues are equal to 1 or @xmath0 for two - dimensional symplectic maps .
therefore it is natural to consider greene s residues as a way to locate those bifurcations . in this context ,
vanishing residues indicate the specific values of the parameters where significant change occurs in the system and hence will be the basis for the reduction of chaos ( by creation of invariant tori ) as in refs .
@xcite but also for the destruction of regular structures . in sec .
[ sec2 ] , we review some basic notions on periodic orbits of hamiltonian systems and their stability , and we explain the details of the residue method .
we give the condition on the residues of a pair of birkhoff periodic orbits to create an invariant torus in their vicinity , and a similar condition which leads to a destruction of nearby invariant tori . in sec .
[ sec3 ] , we apply this method to the destruction and creation of librational and rotational invariant tori of a particular hamiltonian system , a forced pendulum with two interacting primary resonances , used as a paradigm for the transition to hamiltonian chaos .
we consider an autonomous hamiltonian flow with two degrees of freedom which depends on a set of parameters denoted @xmath1 : @xmath2 where @xmath3 and @xmath4 , and @xmath5 being the two - dimensional identity matrix . in order to determine the periodic orbits of this flow and their linear stability properties , we also consider the tangent flow written as @xmath6 where @xmath7 and @xmath8 is the hessian matrix ( composed by second derivatives of @xmath9 with respect to its canonical variables ) . for a given periodic orbit with period @xmath10 ,
the spectrum of the monodromy matrix @xmath11 gives its linear stability property .
as the flow is volume preserving , the determinant of such a matrix is equal to 1 .
moreover , if @xmath12 is an eigenvalue , so are @xmath13 , @xmath14 and @xmath15 . as the orbit is periodic , @xmath16 is an eigenvalue with an eigenvector in the direction of the flow .
its associated eigenspace is at least of dimension 2 since there is another eigenvector with eigenvalue 1 coming from the conserved quantity @xmath17 .
therefore , according to the remark above , the orbit is elliptic if the spectrum of @xmath11 is @xmath18 ( and stable , except at some particular values ) , or hyperbolic if the spectrum is @xmath19 with @xmath20 ( unstable ) .
the intermediate case is when the spectrum is restricted to @xmath21 or @xmath0 and the orbit is called parabolic . whether or not the parabolic periodic orbit is stable depends on higher order terms . in a more concise form
, the above cases can be summarized using greene s definition of a residue which led to a criterion on the existence of invariant tori @xcite : @xmath22 we notice that the 4 ( instead of 2 for 2d maps ) in the numerator comes from the two additional eigenvalues 1 coming from autonomous hamiltonian flows . if @xmath230,1[$ ] , the periodic orbit is elliptic ; if @xmath24 or @xmath25 it is hyperbolic ; and if @xmath26 and @xmath27 , it is parabolic and higher order expansions give the stability of such periodic orbits .
since the periodic orbit and its stability depend on the set of parameters @xmath28 , the features of the dynamics will change with variations of the parameters .
generically , periodic orbits and their linear stability are robust to small changes of parameters , except at specific values where bifurcations occur .
the proposed residue method to control chaos detects these rare events to yield the appropriate values of the parameters leading to the prescribed behavior on the dynamics .
the residue method which leads to a reduction or an enhancement of the chaotic properties of the system is based on the change of stability of periodic orbits upon a change of the parameters of the system . for @xmath29 ,
let us consider two associated birkhoff periodic orbits ( i.e. periodic orbits having the same action but different angles in the integrable case and having the same rotation number on a selected poincar section ) , one elliptic @xmath30 and one hyperbolic @xmath31 .
let us call @xmath32 and @xmath33 their residues .
we have @xmath34 ( and smaller than one ) and @xmath35 .
we slightly modify the parameters @xmath28 until the elliptic periodic orbits becomes parabolic .
some particular situations arise at some critical value of the parameters @xmath36 : @xmath37 : @xmath38 .
+ @xmath39 : @xmath40 while @xmath41 .
+ @xmath42 : @xmath43 while @xmath41 .
+ the first case is associated with the creation of an invariant torus .
the two latter cases might be associated with the destruction of invariant tori ( the ones around the elliptic periodic orbit ) .
the third one is associated with a period doubling bifurcation . in this latter case , the change of stability of the new elliptic periodic orbit has to be considered .
other interesting cases occur depending on the set of selected periodic orbits .
the situation @xmath37 resembles the integrable situation where all the residues of periodic orbits of constant action are zero .
it is expected that an invariant torus is reconstructed in this case .
it can be associated with a transcritical bifurcation ( an exchange of stability ) , a fold , or another type of bifurcation .
in the situation @xmath39 , a change of stability occurs : the elliptic periodic orbit turns hyperbolic while the hyperbolic one stays hyperbolic .
it is generically characterized by a stationary bifurcation . in this case , the destruction of invariant curves is expected in general whether there are librational ones ( representing the linear stability of an elliptic periodic orbit ) or the neighboring rotational ones .
an extra caution has to be formulated since this method only provides an indicator of the _ linear _ stability of periodic orbits .
the nonlinear stability ( or instability ) has to be checked a posteriori by a poincar section for instance .
this method only states that a bifurcation has occurred in the system , whether it is a stationary , transcritical , period doubling or other types of bifurcations .
a more rigorous and safer control method would require to consider the global bifurcations , like the ones obtained by the intersections of the stable and unstable manifolds of two hyperbolic periodic orbits in the spirit of ref .
however such a control method would be computer - time consuming ( determination of the stable and unstable manifolds ) and hence not practical if some short time delay feedback is involved in the control process .
we consider the following forced pendulum system with 1.5 degrees of freedom @xmath44 a poincar section of hamiltonian ( [ eqn : fp ] ) is depicted on fig .
[ fig1 ] for @xmath45 and on fig .
[ fig2 ] for @xmath46 . in order to modify the dynamics of hamiltonian ( [ eqn : fp ] )
, we add an additional ( control ) parameter @xmath47 : we consider a family of hamiltonians of the form @xmath48 where @xmath47 is not too large in order to consider a small modification of the original system , and minimizing the energy cost needed to modify the dynamics .
other choices of families of control terms are possible ( not restricted to @xmath49 .
in particular , more suitable choices of control terms would include more fourier modes .
we have selected a one - parameter family which originates from another control strategy which has been proved to be effective @xcite .
the goal here is to determine the particular values of the parameter @xmath47 such that suitable modifications of the dynamics ( which will be specified later ) occur .
the algorithm is as follows : first , we determine two periodic orbits of hamiltonian ( [ eqn : fp ] ) , an elliptic and a hyperbolic one with the same rotation number on the poincar section , using a multi - shooting newton - raphson method for flows @xcite
. then we modify continuously the control parameter @xmath47 and follow these two periodic orbits .
we compute their residues as function of @xmath47 . ) with @xmath45 .
the arrows indicate the elliptic periodic orbits for the three cases considered here.,title="fig:",scaledwidth=30.0% ] ) with @xmath45 .
the arrows indicate the elliptic periodic orbits for the three cases considered here.,title="fig:",scaledwidth=30.0% ] ) for @xmath46.,title="fig:",scaledwidth=30.0% ] ) for @xmath46.,title="fig:",scaledwidth=30.0% ] a first analysis is done on librational invariant tori ( around the primary resonance located around @xmath50 ) .
we point out in fig .
[ fig1 ] three particular elliptic periodic orbits ( and their associated hyperbolic ones ) , labeled 1,2 and 3 , with , respectively , @xmath51 , @xmath52 and @xmath53 intersections with the poincar section ( @xmath54 ) .
these orbits will be used for two purposes : first we follow the idea of cary and hanson on the construction of invariant tori .
then we extend the residue method to the destruction of these tori .
a similar analysis is done on rotational invariant tori ( the example of the goldenmean invariant torus is treated ) .
this case allows us to compare the residue method with another approach on the control of hamiltonian systems .
briefly we determine the control parameter @xmath47 such that there is a creation of an invariant torus if the original system does not have one , and the destruction of an invariant torus if the system does have one .
for the case @xmath51 , fig .
[ fig3 ] represents the values of the residues of the elliptic and hyperbolic periodic orbits as functions of the control parameter @xmath47 [ see eq .
( [ eqn : fp2 ] ) ] . at @xmath55 ,
both residues vanish which means that they become parabolic periodic orbits as in the integrable case . by increasing @xmath47
, we notice that both orbits exchange their stability which is the manifestation of a transcritical bifurcation while each of the periodic orbits undergo individually a tangent bifurcation .
this type of bifurcation has been observed in refs .
@xcite . at @xmath56 ,
an invariant torus is reconstructed . in order to check the robustness of the method
, one could argue that since this invariant torus is composed of periodic orbits , it is not expected to be robust .
however , by continuity in phase space , an infinite set of invariant tori is present in the neighborhood of the created invariant torus .
most of them have a frequency which satisfy a diophantine condition and hence which will persist under suitable hypothesis on the type of perturbations .
the locations of the different periodic points on the poincar section as @xmath47 varies are indicated by arrows on fig .
the change of stability of these periodic points is associated with the creation of an invariant torus ( also represented in fig .
[ fig4 ] by the plot of the separatrices ) .
we notice that apart from the exchange of stability , the phase space in the neighborhood of these periodic orbits is still regular ( the chaotic region around the hyperbolic periodic orbits is not well developed ) , and hence the regular nature of phase space has not been changed locally ( or one needs to consider higher values of the parameters ) . of the elliptic and hyperbolic ( bold line ) periodic orbits of case 1 ( @xmath51 ) as functions of the parameter @xmath47 for hamiltonian ( [ eqn : fp2 ] ) with @xmath45.,scaledwidth=30.0% ] ) and hamiltonian ( [ eqn : fp2 ] ) with @xmath45 .
the trajectories in gray are for @xmath57 and the ones in black are for @xmath58 . at @xmath56 , an invariant torus of the system
is represented ( bold line ) .
the arrows indicate the change of locations of the periodic points as @xmath47 increases.,scaledwidth=30.0% ] of the elliptic and hyperbolic periodic orbits of case 2 ( @xmath52 ) as functions of the parameter @xmath47 for hamiltonian ( [ eqn : fp2 ] ) with @xmath45.,scaledwidth=30.0% ] the same analysis can be carried out on a set of periodic orbits which are located in a more chaotic region , like for instance the two cases @xmath52 and @xmath53 of fig .
[ fig1 ] outside the regular resonant island .
the values of the residues as functions of the parameter @xmath47 are respectively represented on figs .
[ fig5 ] and [ fig6 ] for @xmath52 and @xmath53 . for @xmath52
, we notice that the residues do not vanish in the range of @xmath47 we considered although there are small and extremum at the same value of the parameter @xmath59
. however , even if these residues do not vanish ( and therefore no exchange of stability by the creation of an invariant torus ) , there is a significant regularization of the dynamics at this specific value of the parameter ( not shown here ) . for @xmath53 ,
the residues vanish for @xmath60 ( see fig .
[ fig6 ] ) and there is a transcritical bifurcation associated with the creation of a set of invariant tori like in figs . [ fig3 ] and [ fig4 ] .
the associated phase space shows a significant increase of the size of the resonant island in this section , we address the destruction of a resonant island by breaking up librational invariant tori .
we notice that on fig .
[ fig5 ] , a bifurcation occurs at @xmath61 for the case 2 ( @xmath52 ) when the residue of the elliptic periodic orbit becomes equal to 1 .
the neighborhood of this periodic orbit becomes a chaotic layer and the qualitative change in the dynamics is seen since the chaotic layer becomes thicker at this value of the parameter .
however since this periodic orbit was initially ( at @xmath57 ) already in the outer chaotic region ( see fig .
[ fig1 ] ) , the regularization is not drastic . in order to obtain a more significant change in the dynamics and a large chaotic zone , one needs to select a periodic orbit inside a regular region , like for instance the one with @xmath53 .
a bifurcation occurs at @xmath62 where the residue of the elliptic periodic orbit crosses 1 ( see fig . [ fig6 ] ) .
a poincar section for the latter case is depicted on fig .
[ fig7 ] , and shows that a significantly large neighborhood has been destabilized by the control term .
we notice that the ratio between the size of the control term and the one of the perturbation is equal to @xmath63 .
we also notice that this last value of @xmath47 is larger than the one required for @xmath52 .
as expected , one needs a larger amplitude to destabilize a region closer to a regular one .
a more effective destabilization procedure can be obtained with the periodic orbit @xmath51 which is inside the regular region . however , as mentioned , the value necessary for this destabilization ( @xmath64 or for @xmath65 ) is too large ; hence we discard it because of our restriction on energy cost . of the elliptic and hyperbolic periodic orbits of case 3 ( @xmath53 ) as functions of the parameter @xmath47 for hamiltonian ( [ eqn : fp2 ] ) with @xmath45.,scaledwidth=30.0% ] ) for @xmath45 and @xmath66.,title="fig:",scaledwidth=30.0% ] ) for @xmath45 and @xmath66.,title="fig:",scaledwidth=30.0% ] in this section , we apply the same approach on rotational invariant tori .
it allows us to compare the results with the ones obtained by a control method proposed in refs .
first , the idea is to look at the creation of a specific invariant torus .
for instance , we select a torus which has been widely discussed in the literature ( see ref .
@xcite and references therein ) , the goldenmean one , which has a frequency @xmath67 for hamiltonian ( [ eqn : fp2 ] ) .
we choose @xmath46 , and we first notice that when @xmath57 , this hamiltonian does not have such an invariant torus ( since its critical value is @xmath68 @xcite ) .
the purpose here is to find the value of the control parameter @xmath47 needed by the residue method to reconstruct this invariant torus ( such that hamiltonian ( [ eqn : fp2 ] ) has this invariant torus ) .
the idea of doing this follows greene s residue criterion . by performing an appropriate change of stability on higher and higher order periodic orbits
, the amplitude of the control term should be smaller and smaller . for @xmath69 ,
the residues of the elliptic and hyperbolic periodic orbit vanish at @xmath70 , and for @xmath71 , at @xmath72 . for @xmath73 ,
both residues vanish at @xmath74 and also at @xmath75 . at these values of the parameter ,
the phase space is locally filled by invariant tori where it is also expected that the goldenmean invariant torus is present .
we notice that the elliptic periodic orbit with @xmath76 ( the next one in greene s residue approach for the analysis of the golden mean torus ) is destabilized at @xmath77 .
therefore , there is no elliptic periodic orbit with @xmath76 at @xmath46 and the analysis using the coupled elliptic / hyperbolic periodic orbits can not be carried out .
however , by following the two ( initially hyperbolic ) periodic orbits with @xmath76 , we see that both residues vanish at @xmath78 .
we compared these values of stabilization with the one given by a method of local control based on an appropriate modification of the potential to reconstruct a specific invariant torus @xcite .
such method provides explicitly the shape ( and amplitude ) of possible control terms whereas the one used in this article has been guessed from these references . by appropriate truncation ( keeping the main fourier mode ) , this method provides @xmath79 where @xmath67 , as an approximate control term .
therefore the amplitude is @xmath80 which is of the same order as the values obtained by zeroing the residues .
however , we point out that smaller values are obtained by looking at higher periodic orbits .
therefore , an efficient control strategy is to combine the advantages of both methods : first , the specific shape of the terms that have to be added to regularize the system is obtained using the method of ref .
then the amplitudes of these terms are lowered using high order periodic orbits . by considering the control term used in this article
, we expect that zeroing the residues of high period will not be feasible with just this term ( as it is the case for instance in fig .
[ fig5 ] ) . a more suitable form of control terms would be constructed from an exact control term which is @xmath81 however , it should be noticed that a control term given by ref .
@xcite is not always experimentally accessible .
the idea is to use a projection of this control term onto a basis of accessible functions .
this projected control term would give an idea of the type of control terms to be used for the residue method .
we would like to stress that in the absence of elliptic islands an initial guess for the newton - raphson method is not straightforward from the inspection of the poincar section .
in particular , it is not easy to select the appropriate hyperbolic periodic orbits which will lead to a significant change in the dynamics .
however , once it has been located , the method can follow them by continuity in the same way as the elliptic ones since the newton - raphson method does not depend on the linear stability of these orbits .
this makes the method more difficult ( although possible ) to handle for just hyperbolic periodic orbits . in this section , we consider hamiltonian ( [ eqn : fp2 ] ) with @xmath82 .
we notice that for @xmath57 , hamiltonian ( [ eqn : fp2 ] ) does have the rotational goldenmean torus .
the purpose is to find some small values of the parameter @xmath47 where this invariant torus is destroyed .
we notice that this case is easier to find than in the previous section since it is well - known that any additional perturbation will end up by destroying an invariant torus generically . here
it means that there will be large intervals of parameters for which the torus is broken ( contrary to the case of the creation of invariant tori ) .
however we will add an additional assumption that the parameters for which this invariant torus is destroyed has to be small compared with the perturbation .
we also notice that the destruction of the golden mean invariant torus is first obtained for negative values of the control parameter ( see fig .
[ fig8 ] ) .
first we illustrate the method by considering specific elliptic and hyperbolic periodic orbits ( with winding ratio @xmath83 ) near the goldenmean torus which will show the changes of dynamics occurring as the parameter is varied .
we notice that the behaviors described below are generic for all the neighboring periodic orbits .
the residues of these periodic orbits as functions of the parameter @xmath47 are shown in fig .
we notice that the elliptic periodic orbit changes its stability , i.e. becomes hyperbolic , at @xmath84 ( where its residue becomes equal to 1 ) .
a close inspection of the poincar section shows on fig .
[ fig9 ] that it undergoes a period doubling bifurcation into an elliptic periodic orbit with 26 intersections on the poincar section ( and winding ratio @xmath85 ) which has a residue zero at the bifurcation . by following the residue of this elliptic periodic orbit ( depicted by a dashed line in fig .
[ fig8 ] ) we see that it vanishes for @xmath86 . at this value of the parameter and for higher value in amplitude , all the periodic orbits considered here ( the two with @xmath53 and the one with @xmath87 ) are hyperbolic .
therefore it is expected that there is a chaotic zone in this area and it is a value at which the torus is expected to be broken ( confirmed by a close inspection of the poincar section ) . of the elliptic and hyperbolic periodic orbits with @xmath53 and also of the one with @xmath87 ( dashed line ) born out of a period doubling bifurcation for hamiltonian ( [ eqn : fp2 ] ) with @xmath82.,scaledwidth=30.0% ] ( 10,5 ) ( 0,0 ) ( indicated with crosses ) for hamiltonian ( [ eqn : fp2 ] ) for @xmath82 and @xmath88 . the period orbit with period 26 ( indicated by circles ) results from a period doubling bifurcation of the one with period 13 ( represented by crosses on the poincar section).,title="fig:",scaledwidth=30.0% ] ( 1.5,0.7 ) ( indicated with crosses ) for hamiltonian ( [ eqn : fp2 ] ) for @xmath82 and @xmath88 .
the period orbit with period 26 ( indicated by circles ) results from a period doubling bifurcation of the one with period 13 ( represented by crosses on the poincar section).,title="fig:",scaledwidth=17.0% ] it is important to notice that a vanishing residue does not automatically imply that there is a creation of an invariant torus , contrary to the previous cases which were obtained by using jointly the elliptic and hyperbolic periodic orbits ( and vanishing residues in both cases ) . here
the hyperbolic periodic orbit associated with these elliptic periodic orbits ( which is the periodic orbit from which the new elliptic orbit was born out by a period doubling bifurcation ) stays hyperbolic as the residue of the elliptic one vanishes .
this feature is generic : the same analysis has been carried out for higher order elliptic periodic orbits close to the goldenmean invariant torus , i.e. the ones with winding ratio @xmath89 , @xmath90 , @xmath91 , @xmath92 : first , the values of the control parameter for which the residues ( which are around @xmath93 for @xmath57 and increase as @xmath47 decreases ) cross 1 are computed and reported in table i ( denoted @xmath94 ) . at these values of the parameters , a period doubling bifurcation occurs for each of them .
then we follow the residues of the elliptic periodic orbits with double period @xmath95 , @xmath96 , @xmath97 and @xmath98 .
the parameter values at which these residues vanish are also reported in table i. for instance , using the periodic orbit with winding ratio @xmath89 , we obtain @xmath99 as the value at which the residue of the bifurcated elliptic periodic orbit with winding ratio @xmath100 .
if we consider higher order periodic orbits , it happens that the goldenmean invariant torus is destroyed by this additional perturbation but not the ones in the neighborhood .
if one is looking at large scale transport properties , these other invariant tori have to be taken into account .
.values of the parameter @xmath47 at which the residue of the elliptic periodic orbit with period @xmath101 crosses 1 ( denoted @xmath94 ) and at which the residue of the elliptic periodic orbit with period @xmath102 obtained by period doubling bifurcation at @xmath94 vanishes ( denoted @xmath103 ) . [ cols="^,^,^,^,^,^",options="header " , ]
in this article , we reviewed and extended a method of control of hamiltonian systems based on linear stability analysis of periodic orbits .
we have shown that by varying the parameters such that the residues of selected periodic orbits cross 0 or 1 , some important bifurcations happen in the system
. these bifurcations can lead to the creation or the destruction of invariant tori , depending on the situation at hand .
therefore we have proposed a possible extension of the residue method to the case of increasing chaos locally .
moreover , we have compared two methods of chaos reduction , and by taking advantage of both methods , we have devised a more effective control strategy .
it is worth noticing that the extension of cary - hanson s method to four dimensional symplectic maps has been done in refs .
@xcite for the increase of dynamic aperture in accelerator lattices .
the extension to the destruction of invariant surface would be to consider the change of linear stability of selected periodic orbits .
however , it would require to consider new types of bifurcations which occurs in the system , like for instance , krein collisions @xcite . | a method to reduce or enhance chaos in hamiltonian flows with two degrees of freedom is discussed .
this method is based on finding a suitable perturbation of the system such that the stability of a set of periodic orbits changes ( local bifurcations ) .
depending on the values of the residues , reflecting their linear stability properties , a set of invariant tori is destroyed or created in the neighborhood of the chosen periodic orbits .
an application on a paradigmatic system , a forced pendulum , illustrates the method . *
changing the dynamical properties of a system is central to the design and performance of advanced devices based on many interacting particles .
for instance , in particle accelerators , the aim is to find the appropriate magnetic elements to obtain an optimal aperture in order to increase the luminosity of the beam , thus requiring the decrease of the size of chaotic regions . in plasma physics ,
the situation is slightly more complex : inside a fusion device ( like a tokamak or a stellarator ) , one needs magnetic surfaces in order to increase confinement .
these surfaces are invariant tori of some fictitious time dynamics
. a control strategy would be to recreate such magnetic surfaces by an appropriate modification of the apparatus ( magnetic perturbation caused by a set of external coils ) .
on the opposite , in order to collect energy and to protect the wall components , an external modification of the magnetic equilibrium has to be performed such that there is a highly chaotic layer at the border ( like an ergodic divertor ) . therefore these devices require a specific monitoring of the volume of bounded magnetic field lines .
another example is afforded by chaotic advection in hydrodynamics : in the long run to achieve high mixing in microfluidics and microchannel devices in particular , the presence of regular region prevents such mixing , and hence a possible way to enhance mixing is to perturb externally the system according to some theoretical prescriptions , in order to destroy invariant surfaces . * |
since their seminal introduction by @xcite , reaction - diffusion systems ( rds s ) have constituted a standard framework for the mathematical modelling of pattern formation in chemistry and biology .
recent advances in mathematical modelling and developmental biology identify the important role of _ domain evolution _ as central in the formation of patterns , both empirically @xcite and computationally @xcite . in this respect ,
many numerical studies , such as @xcite and @xcite , of rds s on evolving domains are available . yet
, fundamental mathematical questions such as existence and regularity of solutions of rds s on evolving domains remains an important open question @xcite .
we focus on growth functions commonly encountered in the field of developmental biology for which our analysis is valid and show the applicability of our analysis to some of the important reaction kinetics encountered in the theory of biological pattern formation . in [ s6 ] we present numerical results for a rds posed on a periodically evolving domain .
we present a moving finite element scheme and a fixed domain finite element scheme to approximate the solution of a rds posed on the evolving and the lagrangian frame respectively . in [ s7 ]
we summarise our findings and indicate future research directions .
let @xmath0 be a ( @xmath1 ) vector of concentrations of chemical species , with @xmath2 , the time - dependent spatial variable and @xmath3 , \
t>0,$ ] the time variable .
the model problem we wish to consider is a semilinear rds posed on a continuously evolving domain ( see @xcite for details of the derivation ) , given by , @xmath4\left({\ensuremath{\boldsymbol{x}}},t\right)&={\ensuremath{\boldsymbol{f}}}({\ensuremath{\boldsymbol{u}}}\left({\ensuremath{\boldsymbol{x}}},t\right))+{\ensuremath{\boldsymbol{d}}}{\ensuremath{\updelta}}{\ensuremath{\boldsymbol{u}}}\left({\ensuremath{\boldsymbol{x}}},t\right ) \qquad \mbox{for } { \ensuremath{\boldsymbol{x } } } \in \omega_{t } \mbox { and } t \in ( 0,t ] , \end{split}\ ] ] where @xmath5 is a @xmath6 simply connected , bounded and continuously deforming domain with respect to @xmath7 .
the function @xmath8 is a ( @xmath1 ) vector of nonlinear coupling terms that is locally lipschitz , @xmath9 is a diagonal matrix with strictly positive entries on the diagonal , @xmath10 is a flow velocity generated by the evolution of the domain and @xmath11 by definition .
[ a_assum ] we assume that the flow velocity @xmath12 is identical to the domain velocity , i.e. , @xmath13 as is standard in the derivation of rds s on evolving domains on application of reynold s transport theorem @xcite . to simplify the exposition we take boundary conditions to be of homogenous neumann type .
@xcite considers a much wider class of boundary conditions and our analysis may be extended to this more general setting .
we are primarily interested in patterns that arise as a result of self - organisation prompting the consideration of homogenous neumann boundary conditions .
we take the initial condition for each @xmath14 to be bounded and nonnegative .
our model problem thus takes the form ( [ eqn : gen ] ) , equipped with the following boundary and initial conditions : @xmath15({\ensuremath{\boldsymbol{x}}},t)=0 , \quad & { \ensuremath{\boldsymbol{x } } } \in \partial\omega_{t } , t>0,\\ { { \ensuremath{\boldsymbol{u}}}}({\ensuremath{\boldsymbol{x}}},0)={{\ensuremath{\boldsymbol{u}}}}_{0}({\ensuremath{\boldsymbol{x } } } ) , \quad&{\ensuremath{\boldsymbol{x } } } \in \omega_{0}. \end{cases } \end{split}\ ] ] startsection subsection20 mm - 0pt * * lagrangian transformation[lagtran ] for our analysis it is more convenient to work with problems defined on a time - independent domain .
therefore , we introduce a transformation that maps our model problem ( [ eqn : gen ] ) from a time - dependent domain to a fixed domain ( see @xcite for a more detailed discussion of this approach ) . in order to do this , without too many technical complications ,
we will restrict our attention to special evolutions of the domain , described next .
we assume the domain @xmath16 to evolve by obeying a bounded spatially linear isotropic domain deformation , i.e. , @xmath17 \mbox { and all } { \ensuremath{\boldsymbol{x}}}\in\omega_t , \end{split}\ ] ] [ rhodefn]with and where @xmath18 represents the spatial coordinates of the initial domain .
this assumption and assumption [ a_assum ] imply that @xmath19 where @xmath20 .
hence , we obtain the following transformed problem on the initial domain ( see @xcite for details ) , with @xmath21 \mbox { and } { \ensuremath{\boldsymbol{\xi } } } \in \omega_0,\ ] ] we have @xmath22,\\ [ \vec{\nu}\cdot\nabla\hat{{\ensuremath{\boldsymbol{u}}}}]({\ensuremath{\boldsymbol{\xi}}},t)=0 \quad & \mbox{on } \partial\omega_{0 } , t>0,\\ { { \ensuremath{\boldsymbol{\hat{u}}}}}({{\ensuremath{\boldsymbol{\xi}}}},0)={{\ensuremath{\boldsymbol{\hat{u}}}}}_{0}({{\ensuremath{\boldsymbol{\xi } } } } ) \quad&{{\ensuremath{\boldsymbol{\xi } } } } \in \omega_0,\\ 0 \leq { { \ensuremath{\boldsymbol{\hat{u}}}}}_{0}({{\ensuremath{\boldsymbol{\xi } } } } ) < \infty , & \end{cases } \end{split}\ ] ] where @xmath23 is the initial spatial domain , @xmath24 is the spatial dimension and the laplacian is now taken with respect to @xmath18 .
the local and global existence results that we utilise from the existing literature require the coefficients on our diffusion term to be independent of time ; to this end we introduce the following proposition :
we now summarise the existing results for rds s on fixed domains which will form the basis of our analysis on evolving domains . [ localap ]
let conditions ( [ g1 ] ) , ( [ g2 ] ) and ( [ g6 ] ) hold and let @xmath25 be a solution of problem ( [ eqn : fixed ] ) .
the following a priori estimates hold , [ globexis ] if conditions ( [ g1])([g5 ] ) hold , with @xmath26 from condition ( [ g4 ] ) satisfying @xmath27 , @xmath28 , @xmath25 is a solution of problem ( [ eqn : fixed ] ) and if there exists @xmath29 such that @xmath30 then @xmath31 . alternatively ,
if conditions ( [ g1])([g5 ] ) hold , with @xmath26 from condition ( [ g4 ] ) satisfying @xmath32 , @xmath33 and where @xmath24 represents the spatial dimension , @xmath25 solves a problem of the form ( [ eqn : fixed ] ) and if there exists @xmath34 such that @xmath35 then @xmath36 .
specifically if @xmath26 from condition ( [ g4 ] ) satisfies @xmath37 or if @xmath38 with @xmath39 and the remaining conditions ( [ g1])([g6 ] ) are satisfied then @xmath31 .
[ lyap]we assume the lyapunov function introduced in [ gorig ] is of the following form @xmath40 assumption [ lyap ] is somewhat natural .
condition ( [ g5 ] ) is essentially a polynomial type growth restriction on the zero order terms @xcite .
assumption [ lyap ] can be viewed as the explicit analogue of the polynomial growth restriction on the zero order terms implicit in ( [ g5 ] ) .
[ lem : equivalence_lyap ] we denote the zero order term in problem ( [ eqn : scaled ] ) by @xmath41 the positive quadrant remains an invariant region for the solutions of our evolving domain problem since @xmath42 thus assumption [ finv ] implies @xmath43 is an invariant region for the solutions of problem ( [ eqn : scaled ] ) .
let @xmath44 @xmath45 be as defined in
[ gorig ] , for which conditions ( [ g1])([g6 ] ) hold for problem ( [ eqn : fixed ] ) .
denote @xmath46}$ ] and @xmath47}$ ] ; these are well defined real numbers thanks to to assumption [ rhodefn ] .
we now show that conditions ( [ g1])([g6 ] ) hold with the same @xmath48 , @xmath49 replaced by @xmath50 and @xmath26 from ( [ g4 ] ) replaced by @xmath51 , where @xmath52 .
clearly conditions ( [ g1])([g3 ] ) are still satisfied as they depend only on @xmath48 which is unchanged . condition ( [ g4 ] ) holds since @xmath53 by the stability of the fixed domain problem .
assumption [ lyap ] gives , @xmath54 where @xmath55 condition ( [ g5 ] ) holds since @xmath56 where @xmath57 condition ( [ g6 ] ) holds since @xmath58 thus the positive quadrant remains an invariant region for the solutions of problem ( [ eqn : scaled ] ) and the lyapunov stability conditions in [ subsec : lyapunov ] are satisfied , completing the proof .
[ finalthm ]
in this section we illustrate some applications of theorem [ finalthm ] .
we present different forms of admissible domain evolution that fufil assumption [ rhodefn ] .
we show that assumption [ lyap ] is applicable to some commonly encountered models in chemistry and biology .
we concentrate on rds s which admit turing instabilities , as the main focus of our research is biological pattern formation .
we identify and describe lyapunov functions and constants that imply global existence for the fixed domain problem and thus for the evolving domain problem by theorem [ finalthm ] .
startsection subsection20 mm - 0pt * * admissible domain evolution we now provide some commonly encountered examples of domain evolution in developmental biology for which assumption [ rhodefn ] holds : * logistic evolution on any finite positive time interval @xmath59,\ ] ] where @xmath60 is the growth rate and @xmath61 is the carrying capacity ( limiting size of the evolving domain ) . * exponential evolution on any finite positive time interval @xmath62.\ ] ]
* linear evolution on any finite positive time interval @xmath63,\ ] ] where @xmath64
. startsection subsection20 mm - 0pt * * admissible kinetics we now present some of the commonly encountered reaction kinetics of problem ( [ eqn : gen ] ) for which the analysis of @xcite implies global existence of solutions on fixed domains .
we first consider the problem ( [ eqn : gen ] ) with this general reaction term @xmath65 with the following restrictions @xmath66 and nonnegative ( [ eqn : boundary_initial ] ) , we show that assumption [ finv ] is fulfilled , which implies @xmath67 is an invariant region for the solutions .
( [ c3 ] ) we have @xmath68 conditions ( [ c1 ] ) and ( [ c2 ] ) imply @xmath69 thus assumption [ finv ] is fulfilled due to ( [ c5 ] ) .
now we show conditions ( [ g1])([g6 ] ) are fulfilled with @xmath70 . clearly conditions ( [ g1])([g3 ] ) hold .
condition ( [ g4 ] ) holds with @xmath71 @xmath72 , @xmath73 and @xmath70 .
condition ( [ g5 ] ) holds since @xmath74 using ( [ c4 ] ) we have @xmath75 where @xmath76 represent constants that depend on the value of @xmath77 in ( [ c4 ] ) .
condition ( [ g6 ] ) holds since @xmath78 where @xmath79 and @xmath80 .
a straightforward application of theorem [ finalthm ] thus completes the proof .
the generic problem for which we showed global existence of solutions actually encompasses some of the more widely studied models in the theory of pattern formation such as the gray - scott model and the brussellator .
below we present two examples of two species reaction terms for which our analysis implies global existence of solutions , the first of which is a restriction of the reaction term above . * _ activator - depleted _ substrate model : we consider the _ activator - depleted _ substrate model @xcite also known as the brusselator model : @xmath81 where @xmath82 .
the assumption of nonnegative initial data implies that the positive quadrant is invariant for our problem due to the fact that @xmath83
. if we take @xmath84 then conditions ( [ g1])([g6 ] ) are fulfilled with @xmath85 which implies global existence of solutions on evolving domains via theorem [ finalthm ] .
the remaining constants for which conditions ( [ g1])([g6 ] ) hold are given in table [ tab : const ] . *
thomas reaction kinetics : the following model , proposed and studied experimentally by @xcite , is based on a specific reaction involving the substrates oxygen and uric acid which react in the presence of the enzyme uricase : @xmath86 where @xmath87 and @xmath88 .
once again the assumption of nonnegative initial data implies that the positive quadrant is invariant for our problem due to the fact that @xmath83 . if we again take @xmath89 then conditions ( [ g1])([g6 ] ) are fulfilled with @xmath85 which implies global existence of solutions on evolving domains via theorem [ finalthm ] .
the remaining constants for which conditions ( [ g1])([g6 ] ) hold are given in table [ tab : const ] .
it can be shown , utilising the techniques of @xcite , that there exist bounded invariant rectangles for the solutions of the thomas model defined in ( [ eqn : thom ] ) .
this implies global existence of solutions via theorem [ locexis ] .
however the authors can show that it is possible to construct growth functions that fulfil assumption [ rhodefn ] , for which the bounded invariant rectangle can be made arbitrarily large .
this necessitates the lyapunov function approach to show existence and uniqueness of solutions .
.terms from [ s3 ] for which conditions ( [ g1])([g6 ] ) hold for the kinetics defined in ( [ eqn : schnak ] ) and ( [ eqn : thom ] ) respectively . [
cols="^,^,^",options="header " , ] the analysis can be applied to a large number of problems unrelated to the theory of pattern formation .
@xcite provide an example applicable to ecology and the aforementioned paper of @xcite contains further examples as well as the numerous citations of said paper that use the approach on various problems .
in this section we present numerical results on two - dimensional evolving domains to back - up the theoretical results of the previous sections .
the numerical simulation of rds s on growing domains is an extensive research area . @xcite and @xcite study mode doubling and tripling behaviour of the _ activator - depleted _ substrate model on one- and two - dimensional growing domains .
the theoretical results derived above apply for any evolution that fulfils assumption [ rhodefn ] .
we present numerical results on a periodically evolving domain that exhibit spot splitting as well as spot annihilation and merging , which to the authors knowledge is as yet an unstudied area .
startsection subsection20 mm - 0pt * * domain evolution we consider periodic domain evolution defined by @xmath90,\ ] ] with the initial domain defined as @xmath91 ^ 2 $ ] which grows to @xmath92 ^ 2 $ ] before contracting back to the original domain .
startsection subsection20 mm - 0pt * * continuous problems we present results for the aforementioned _ activator - depleted _ model considered on a periodically evolving domain and its equivalent transformed system on a fixed domain as in [ lagtran ] . for example , on a periodically evolving domain the problem is stated as follows : @xmath93({\ensuremath{\boldsymbol{x}}},t)-{{\ensuremath{\updelta}}}{u_1}({\ensuremath{\boldsymbol{x}}},t)=0.1-[u_{2}^2u_{1}]({\ensuremath{\boldsymbol{x}}},t ) , \quad & \mbox{for } { \ensuremath{\boldsymbol{x}}}\in\omega_{t}\\ \partial_{t}{u_2}({\ensuremath{\boldsymbol{x}}},t)+[\nabla\cdot({\ensuremath{\boldsymbol{a}}}{u_2})]({\ensuremath{\boldsymbol{x}}},t)-0.01{{\ensuremath{\updelta}}}{u_2}({\ensuremath{\boldsymbol{x}}},t)=0.9-u_{2}({\ensuremath{\boldsymbol{x}}},t)+[u_{2}^2u_{1}]({\ensuremath{\boldsymbol{x}}},t ) , & \mbox{and } t\in(0,t ] , \\ [ \vec{\nu}\cdot\nabla{\ensuremath{\boldsymbol{u}}}]({\ensuremath{\boldsymbol{x}}},t)=0 , \quad & { \ensuremath{\boldsymbol{x } } } \in \partial\omega_{t } , t>0 , \end{cases}\ ] ] where @xmath94 and @xmath95 @xmath96 equivalently , the following transformed equations are obtained on a fixed domain , @xmath97 , \\ [ \vec{\nu}\cdot\nabla{\ensuremath{\boldsymbol{u}}}]({\ensuremath{\boldsymbol{\xi}}},t)=0 , \quad & { { \ensuremath{\boldsymbol{\xi } } } } \in \partial\omega_{0 } , t>0 . \end{cases}\ ] ] in both cases we take identical initial conditions as small perturbations around the homogenous steady state of ( 1.0 , 0.9 ) obtained in the absence of domain evolution .
startsection subsection20 mm - 0pt * * numerical schemes[subsec : numerical_schemes ] we employ a galerkin finite element method for the spatial approximation and an implicit - explicit modified backward euler scheme for the time integration .
discretising in time we divide the time interval @xmath98 $ ] into a partition of n uniform subintervals , @xmath99 and denote by @xmath100 the time step . for the spatial discretisation
we introduce a regular triangulation @xmath101 of @xmath102 with @xmath103 an open simplex .
we define the following shorthand for a function of time , @xmath104 .
we define the finite element space on the initial domain @xmath105 as , @xmath106 where @xmath107 denotes the space of polynomials no higher than degree 1 . for the numerical simulation of equation ( [ eqn : schnak_moving ] ) we require finite element spaces defined on the evolving domain .
we construct the finite element spaces @xmath108 according to the following relation between the basis functions of @xmath108 and @xmath109 .
@xmath110 thus the family of finite element spaces on the evolving domain @xmath111 may be defined as , @xmath112 where we have used the fact that the domain evolution is linear with respect to space .
we approximate the initial conditions in both schemes by @xmath113 where @xmath114 is the standard lagrange interpolant . the finite element scheme to approximate the solution to equation ( [ eqn : schnak_moving ] ) aims to find @xmath115 such that @xmath116 for all @xmath117 .
similarly the finite element scheme to approximate the solution to equation ( [ eqn : schnak_fixed ] ) aims to find @xmath118 such that @xmath119 for all @xmath120 .
we solved the models in c utilising the fem library by @xcite .
we used the conjugate gradient solver to compute our discrete solutions .
we took an initial triangulation @xmath101 with 8321 nodes , a uniform mesh diameter of @xmath121 and a fixed timestep of @xmath122 .
paraview was used to display our results .
startsection subsection20 mm - 0pt * * results figures [ fig : discrete_schnak_moving ] and [ fig : discrete_schnak_fixed ] show snapshots of the activator profile corresponding to the _ activator - depleted _ system ( [ eqn : schnak ] ) .
the inhibitor profiles have been omitted as they are @xmath123 out of phase to the activator profiles .
we have verified numerically that there is very little difference between the discrete solution corresponding to system ( [ eqn : discrete_schnak_moving ] ) mapped to the fixed domain and the discrete solution corresponding to system ( [ eqn : discrete_schnak_fixed ] ) defined on a fixed domain , as is expected from the results in [ lagtran ] .
the figures illustrate the mode doubling phenomena that occurs as the domain grows as well as the spot annihilation and spot merging phenomena that occurs as the domain contracts .
we note that the mode transition sequence , i.e. , the number of spots , is different when the domain grows to when it contracts .
the difference in the mechanism of mode transitions on growing and contracting domains is an area in which very little work has been done and these initial numerical results indicate the need for further exploration of this area .
many problems in biology and biomedicine involve growth . in developmental biology
recent advances in experimental data collection allow experimentalists to capture the emergence of pattern structure formation during growth development of the organism or species .
such experiments include the formation of spot patterns on the surface of the eel , patterns emerging on the surface of the japanese flounder and butterfly wing patterns forming during the growth development of the imaginal wing disc . in all these examples , patterns form during growth development . since the seminal paper by @xcite which considered linear models that could give rise to spatiotemporal solutions on fixed domains due the process of diffusion - driven instability , a lot of theoretical results on global existence of such solutions have been derived and proved for highly nonlinear mathematical models @xcite .
only recently , mathematical models on growing domains have been derived from first principles in order to incorporate the effects of domain evolution into the models @xcite . in all these studies , very little analysis has been done up to now to extend the theoretical global existence results to models defined on evolving domains . under suitable assumptions ,
we have extended existence results from problems posed on fixed domains to problems posed on an evolving domain .
we have illustrated the applicability of the existence results of @xcite to problems on evolving domains .
we have shown that global existence of solutions to many commonly encountered rds s on fixed domains implies global existence of solutions to the same rds s on a class of evolving domains .
the results are significant in the theory of pattern formation especially in fields such as developmental biology where problems posed on evolving domains are commonly encountered .
our results hold with no assumptions on the sign of the growth rate , which may prove useful in other fields where monotonic domain growth is not valid from a modelling perspective .
the applicability of our results is demonstrated by considering different forms of domain evolution ( linear , logistic and exponential ) . in order to validate our theoretical findings , we presented results on a periodically evolving domain .
our results illustrate the well - known period - doubling phenomenon during domain growth but more interesting and surprising is the development of spot annihilation and spot merging phenomena during contraction .
this raises new questions about bifurcation analysis on growing and contracting domains .
one of our primary goals is the numerical analysis of finite element approximations of rds s on evolving domains
. the classical existence results obtained will be an important tool in future work .
numerical experiments have been carried out and they illustrate the need for further numerical analysis especially in the case of contracting domains .
extension of our work onto domains with more complex evolution is another area for future research .
the research of c.venkataraman is partially supported by an epsrc doctoral training grant and a university of sussex graduate teaching assistantship .
30 [ 1]#1 [ 1]#1 urlstyle [ 1]doi # 1 | we present global existence results for solutions of reaction - diffusion systems on evolving domains .
global existence results for a class of reaction - diffusion systems on fixed domains are extended to the same systems posed on spatially linear isotropically evolving domains .
the results hold without any assumptions on the sign of the growth rate .
the analysis is valid for many systems that commonly arise in the theory of pattern formation .
we present numerical results illustrating our theoretical findings . |
sampling is an efficient methodology that can be used for functional approximations @xcite .
it is based on interpolation through a set of discrete points @xmath3 selected from a function @xmath4 within an interval @xmath5 $ ] .
the sinc function @xmath6 is one of the most popular in sampling @xcite . specifically , due to its remarkable sampling property @xmath4
can be represented as a series approximation @xcite @xmath7 where @xmath2 is a small parameter that at @xmath8 can be regarded as the step .
we have shown recently a new methodology of sampling based on incomplete cosine expansion of the sinc function in numerical integration @xcite .
another very useful sampling tool can be constructed by using the gaussian function @xmath9 .
the sampling property of the gaussian function @xmath9 can be observed by changing its variable as @xmath10 and then considering the limit @xmath11 where the kroneckers delta @xmath12 is a building block of _ the shah function _ , also known as _ the comb function _ : @xmath13 ( see application of _ the shah function _ @xmath14 in sampling in @xcite ) .
consequently , this example shows that the gaussian function can also be utilized for functional approximations by sampling methodology .
it is convenient to modify the gaussian function @xmath9 by multiplying a constant @xmath15 and making change of the variable as @xmath16 .
this results in @xmath17 where the @xmath1 is the fitting parameter .
figure 1 shows three variations of the sampling gaussian function computed at @xmath18 , @xmath19 ( black curve ) , @xmath18 , @xmath20 ( blue curve ) and @xmath18 , @xmath21 ( red curve ) . we may attempt to approximate @xmath4 by replacing the sinc functions in equation with the gaussian functions of kind as given by equation @xmath22 consequently , choosing the sampling points separated equidistantly with step @xmath2 between two adjacent points we obtain @xmath23 the application of the gaussian function of kind can be justified by considering a simplest case when @xmath24 .
figure 2 shows the corresponding approximations at @xmath18 , @xmath19 , @xmath25 ( blue curve ) and @xmath18 , @xmath20 , @xmath25 ( red curve ) .
the multiplication of the number of the sampling points to the step determines the range for approximation @xmath26 . as a consequence , the approximated functions shown in fig .
2 resemble the window function .
in particular , these functions are approximately equal to @xmath27 within the range @xmath28 $ ] and to @xmath29 otherwise . as we can see from this figure , at smaller @xmath1 the approximated function oscillates near the constant @xmath27 at the top of the curve stronger ( blue curve ) . however , as @xmath1 increases the oscillation rapidly decreases ( red curve ) and practically vanishes at @xmath30 .
this signifies that if the change in curve between any two adjacent sampling points is relatively small , then any function can be approximated by sampling with the gaussian function . in this work
we introduce an application of the complex error function to the fourier analysis .
in particular , we show that the use of equation in the fourier integration leads to a weighted sum of the complex error functions . due to remarkable property of the complex error function
this approach provides efficient computational methodology in the fourier transform as a damping harmonic series .
the complex error function , also known as the faddeeva function or the kramp function , is defined as @xcite @xmath31 where @xmath32 is the complex argument .
this function is a solution of the following differential equation @xcite @xmath33 the complex error function finds broad applications in many fields of applied mathematics @xcite , physics @xcite and astronomy @xcite . in applied mathematics
it is closely related to the error function of complex argument @xcite @xmath34 \leftrightarrow { \rm{erf}}\left ( z \right ) = 1 - { e^ { - { z^2}}}w\left ( { iz } \right),\ ] ] the normal distribution function @xcite @xmath35 , \end{aligned}\ ] ] the fresnel integral @xcite @xmath36/2 \end{aligned}\ ] ] and the dawsons integral @xcite @xmath37 in physics and astronomy the complex error function is related to the voigt function @xcite @xmath38 that describes the spectral behavior of the photon emitting or absorbing objects ( photo - luminescent materials @xcite , planetary atmosphere @xcite , celestial bodies @xcite and so on ) .
specifically , the voigt function represents the real part of the complex error function @xmath39 , \quad\quad y \ge 0.\ ] ] other functions that can be expressed in terms of the complex error function are the plasma dispersion function @xcite , the gordeyevs integral @xcite , the rocket flight function @xcite and the probability integral @xcite .
the complex error function can be represented alternatively as ( see equation ( 3 ) in @xcite and @xcite , see also appendix a in @xcite for derivation ) @xmath40 using the change of the variable as @xmath41 in the integral above leads to @xcite @xmath42 further , we will use this equation in derivation of the weighted sum .
a rapid c / c++ implementation ( roofit package from cenrs library ) for computation of the complex error function with average accuracy @xmath43 has been reported in the recent work @xcite .
there are several definitions for the fourier transform . in this work
we will use the following definitions @xcite @xmath44 and @xmath45 thus , the relationships between functions @xmath4 and @xmath46 are performed by two operators @xmath47 and @xmath48 corresponding to the forward and inverse fourier transforms , respectively . in signal processing the arguments @xmath49 and @xmath50 in these reciprocally fourier transformable functions @xmath4 and @xmath46 are interpreted , accordingly , as time vs. frequency @xcite .
we can find an approximation to the fourier transform of the function @xmath4 by substituting approximation into equation .
this leads to @xmath51 taking into account that @xmath52 the equation can be rewritten as @xmath53 since change of the variable @xmath54 leads to @xmath55 the equation can be rearranged as @xmath56 comparing this equation with equation yields @xmath57
lastly , defining the constants @xmath58 we can rewrite approximation in a more compact form as a weighted sum @xmath59}. \end{aligned}\ ] ] the equation can be expressed through any functions that have been considered in the section @xmath60 above .
for example , using the identity after trivial rearrangements we get the fourier transform in terms of error functions of complex argument @xmath61 \right .
\\ & \left .
\qquad \ , + { \alpha _ { - n}}{e^ { - { { \left ( { \pi c\nu - inh / c } \right)}^2}}}\left [ { 1 + { \rm{erf}}\left ( { nh / c + i\pi c\nu } \right ) } \right ] \right\ } . \end{aligned}\ ] ] it can also be shown that the approach based on a weighted sum of the complex error functions can be generalized to the laplace transform .
the derivation of approximation for the inverse fourier transform is straightforward now .
comparing integrals , and using approximation we immediately obtain @xmath62 } , \end{aligned}\ ] ] where the coefficients are calculated as @xmath63 .
since the gaussian function rapidly decreases with increasing @xmath64 , only few terms with negative index @xmath65 actually contribute to shape the curve @xmath4 along the positive @xmath49-axis where the fourier integration takes place according to equation . as a result
, it is sufficient to take into consideration only , say , first three terms @xmath66}^2}}}/\left ( { c\sqrt \pi } \right)$ ] , @xmath67}^2}}}/\left ( { c\sqrt \pi } \right)$ ] and @xmath68}^2}}}/\left ( { c\sqrt \pi } \right)$ ] with negative index @xmath65 .
thus , the number of the sampling points is reduced by almost factor of @xmath60 from @xmath69 to @xmath70 .
consequently , the equations and can be simplified as @xmath71 } \end{aligned}\ ] ] and @xmath72}. \end{aligned}\ ] ] it should be noted that a _ real - time _ computation of the complex error functions in equations and may not be fast enough for the wavelets with extended effective length . however , this problem can be easily resolved by using the following precomputed values @xmath73 and @xmath74 that can be stored as datasets in a computer memory .
in particular , the application of these _
ready - made _ precomputed values enables us to simplify considerably the approximations and for the forward and inverse fourier transforms as given by @xmath75 and @xmath76 accordingly . as an application of the precomputed values @xmath77 , @xmath78 and @xmath79 , @xmath80 in approximations and helps us to avoid a _ real - time _ computation of the complex error functions , this technique significantly accelerates an algorithmic implementation
. however , there is an alternative way for rapid performance in the numerical forward and inverse fourier transforms .
specifically , using a remarkable property of the complex error function @xcite @xmath81 the approximations and can be significantly simplified and expressed in terms of a damping harmonic series .
let us rewrite equation as @xmath82\\ & + \sum\limits_{n = 1}^n { \left [ { { \alpha _ n}w\left ( { - \pi c\nu , - nh / c } \right ) + { \alpha _ { - n}}w\left ( { \pi c\nu , - nh / c } \right ) } \right ] } \\ & + \sum\limits_{n = - n}^ { - 1 } { \left [ { { \alpha _ n}w\left ( { - \pi c\nu , - nh / c } \right ) + { \alpha _ { - n}}w\left ( { \pi c\nu , - nh / c } \right ) } \right ] } \end{aligned}\ ] ] or @xmath83\\ & + \sum\limits_{n = 1}^n { \left [ { { \alpha _ n}w\left ( { - \pi c\nu , - nh / c } \right ) + { \alpha _ { - n}}w\left ( { \pi c\nu , - nh / c } \right ) } \right ] } \\ & + \sum\limits_{n = 1}^n { \left [ { { \alpha _ { - n}}w\left ( { - \pi c\nu , nh / c } \right ) + { \alpha _ n}w\left ( { \pi c\nu , nh / c } \right ) } \right ] } .
\end{aligned}\ ] ] the identity can be expressed through arguments @xmath84 and @xmath85 as @xmath86 this leads to @xmath87 therefore , we have the following identity @xmath88 } \\ & = h\sum\limits_{n = 1}^n { \left [ { f\left ( { nh } \right){e^ { - { { \left ( { \pi c\nu } \right)}^2 } - 2\pi i\nu nh } } + f\left ( { - nh } \right){e^ { - { { \left ( { \pi c\nu } \right)}^2 } + 2\pi i\nu nh } } } \right ] } \\ & \,\,\,\,\ , - \frac{h}{2}\sum\limits_{n = 1}^n { \left [ { { \alpha _ n}w\left ( { \pi c\nu , nh / c } \right ) + { \alpha _ { - n}}w\left ( { - \pi c\nu , nh / c } \right ) } \right ] , } \end{aligned}\ ] ] where @xmath89 . substituting this identity into approximation yields
@xmath90 \\ & + h{e^ { - { { \left ( { \pi c\nu } \right)}^2}}}\sum\limits_{n = 1}^n { \left [ { f\left ( { nh } \right){e^ { - 2\pi i\nu nh } } + f\left ( { - nh } \right){e^{2\pi i\nu nh } } } \right]}. \end{aligned}\ ] ] since ( see @xcite for details ) @xmath91 this approximation can be expressed as @xmath92 } \end{aligned}\ ] ] or @xmath93 the inverse fourier transform can be found from approximation by using the same derivation procedure @xmath94 although the approximations and consist of harmonic functions @xmath95 and @xmath96 , the resultant forward and inverse fourier transformed wavelets remain , nevertheless , non - periodic due to presence of the damping exponential functions @xmath97 and @xmath98 .
consequently , unlike the conventional discrete fourier transform the approximations and have no distortion that may appear as a result of aliasing .
if a function @xmath99 can be separated into the even and odd parts as @xmath100 where @xmath101 and @xmath102 , then according to definition the forward fourier transforms are given by @xmath103 and @xmath104 respectively .
therefore , using approximations and for the forward fourier transforms over the even @xmath105 and odd @xmath106 parts we obtain @xmath107 and @xmath108 respectively . since @xmath109 and @xmath110 from equations and it follows that @xmath111 and @xmath112 consider the following example @xmath113 the equation can be separated into the even @xmath114 and odd @xmath115 parts .
figure 3 shows that the function @xmath99 while fig . 4 depicts its even @xmath116 and odd @xmath117 parts by blue and red curves , respectively . from fig .
4 we can see that the functions @xmath105 and @xmath106 are the wavelets ( or the pulses ) localized within the region @xmath118 $ ] . beyond this region the functions @xmath105 and @xmath106 are equal to zero .
therefore , the effective length of these wavelets is @xmath119 . since according to approximation we applied @xmath69 sampling points , the step @xmath2 between two adjacent sampling points
can be determined from the formula @xmath120 .
thus , by choosing @xmath121 and @xmath122 we can find the corresponding steps to be @xmath123 and @xmath124 , respectively . as the function @xmath117 is odd
, its fourier transform @xmath125 is purely imaginary according to equation .
figure 5 depicts @xmath126 and @xmath127 $ ] corresponding to the functions @xmath116 and @xmath117 by blue and red curves , respectively .
the curves for the functions @xmath126 and @xmath127 $ ] are obtained numerically by using equations and at @xmath121 , @xmath128 .
the fourier transforms for the even @xmath116 and odd @xmath117 parts of the function @xmath99 can be found analytically . in particular , substituting the equations , into approximations , and considering the fact that these wavelets are not zero - valued only at @xmath118 $ ] we get @xmath129 and @xmath130 therefore , it is convenient to define the differences by using these functions @xmath131 and @xmath132 - \left\ { - 2h{e^ { - { { \left ( { \pi c\nu } \right)}^2}}}\sum\limits_{n = 1}^n { f^-\left ( { nh } \right)\sin \left ( { 2\pi \nu nh } \right ) } \right\}\\ & = \frac{{\pi \nu \cos \left ( { \pi \nu } \right ) - \sin \left ( { \pi \nu } \right)}}{{{\pi ^2}{\nu ^2 } } } - \left\ { - 2h{e^ { - { { \left ( { \pi c\nu } \right)}^2}}}\sum\limits_{n = 1}^n { f^-\left ( { nh } \right)\sin \left ( { 2\pi \nu nh } \right ) } \right\}. \end{aligned}\ ] ] figure 6 illustrates the differences @xmath133 ( blue curve ) and @xmath134 ( red curve ) computed at @xmath121 , @xmath128 . as we can see from fig . 6 , the differences @xmath133 and @xmath134 are within the range @xmath135 . further increase of the integer @xmath136 significantly improves the accuracy .
this can be seen from fig .
7 showing that the differences @xmath133 ( blue curve ) and @xmath134 ( red curve ) computed at @xmath122 , @xmath137 remain within the narrow range @xmath138 .
we present a new approach for numerical computation of the fourier integrals based on a sampling with the gaussian function of kind @xmath0 . it is shown that the fourier transform can be expressed as a weighted sum of the complex error functions . applying a remarkable property of the complex error function ,
the weighted sum of the complex error functions can be significantly simplified as a damping harmonic series .
unlike the conventional discrete fourier transform , this methodology results in a non - periodic wavelet approximation .
therefore , the proposed approach may be practically convenient and advantageous in algorithmic implementation .
this work is supported by national research council canada , thoth technology inc . and york university .
the authors wish to thank to prof .
ian mcdade and dr .
brian solheim for discussions and constructive suggestions .
abrarov and b.m .
quine , sampling by incomplete cosine expansion of the sinc function : application to the voigt / complex error function , appl .
comput . , 258 ( 2015 ) 425 - 435 .
http://dx.doi.org/10.1016/j.amc.2015.01.072 b.m . quine and j.r .
drummond , genspect : a line - by - line code with selectable interpolation error tolerance j. quant .
transfer 74 ( 2002 ) 147 - 165 .
http://dx.doi.org/10.1016/s0022-4073(01)00193-5 b.m . quine and s.m .
abrarov , application of the spectrally integrated voigt function to line - by - line radiative transfer modelling .
transfer , 127 ( 2013 ) 37 - 48 .
http://dx.doi.org/10.1016/j.jqsrt.2013.04.020 a. berk , voigt equivalent widths and spectral - bin single - line transmittances : exact expansions and the modtran5 implementation , j. quant .
transfer , 118 ( 2013 ) 102 - 120 .
h. borchert , d. v. talapin , n. gaponik , c. mcginley , s. adam , a. lobo , t. mller and h. weller , relations between the photoluminescence efficiency of cdte nanocrystals and their surface properties revealed by synchrotron xps , j. phys .
b , 107 ( 36 ) ( 2003 ) 9662 - 9668 .
http://dx.doi.org/10.1021/jp0352884 s.j .
mckenna , a method of computing the complex probability function and other related functions over the whole complex plane , astrophys .
space sci .
, 107 ( 1 ) ( 1984 ) 71 - 83 .
http://dx.doi.org/10.1007/bf00649615 s.m .
abrarov and b.m .
quine , master - slave algorithm for highly accurate and rapid computation of the voigt / complex error function , j. math .
research , 6 ( 2 ) ( 2014 ) 104 - 119 . | in this paper we show that a methodology based on a sampling with the gaussian function of kind @xmath0 , where @xmath1 and @xmath2 are some constants , leads to the fourier transform that can be represented as a weighted sum of the complex error functions . due to remarkable property of the complex error function , the fourier transform based on the weighted sum
can be significantly simplified and expressed in terms of a damping harmonic series .
in contrast to the conventional discrete fourier transform , this methodology results in a non - periodic wavelet approximation .
consequently , the proposed approach may be useful and convenient in algorithmic implementation .
+ * keywords : * complex error function , faddeeva function , fourier transform , sampling , gaussian function , numerical integration + |
let @xmath5 denote an @xmath6-dimensional @xmath7-vector space .
a _ reflection _ in @xmath5 is any non - identity element @xmath8 in @xmath9 of finite order that fixes some hyperplane @xmath10 , and a finite subgroup @xmath0 of @xmath9 is called a _ reflection group _ if it is generated by reflections .
[ example : sn ] the action of the symmetric group @xmath11 on @xmath12:=\{1,2,\ldots , n\}$ ] gives rise to a faithful action on @xmath13 via @xmath14 , where @xmath15 denote the standard basis vectors .
since the transpositions act as reflections and generate the group , this representation realizes @xmath11 as a reflection group in @xmath16
. we shall refer to it as the _ defining representation_. for an element @xmath17 and root of unity @xmath18 , let @xmath19 denote the @xmath18-eigenspace of @xmath20 in @xmath5 .
define @xmath21 to be the @xmath0-poset ( partially ordered set ) of all such @xmath18-eigenspaces @xmath22 ordered by reverse inclusion , with @xmath0-action given by @xmath23 .
choosing @xmath24 recovers the lattice @xmath25 of intersections of reflecting hyperplanes for @xmath0 ; see ( * ? ? ?
* lemma 4.4 ) .
the minimal elements of @xmath21 ( i.e. , inclusion - maximal @xmath18-eigenspaces of @xmath0 ) are the focus of springer s theory of regular elements @xcite , and each has dimension equal to the number @xmath26 of _ degrees _
@xmath27 of @xmath0 that are divisible by @xmath4 , the order of @xmath18 ; see proposition [ prop : springer : transitive ] below . when @xmath0 is crystallographic , the poset @xmath21 itself appears in brou , malle , and michel s @xmath28-sylow theory @xcite .
g.i . lehrer and t.a .
springer , _ intersection multiplicities and reflection subquotients of unitary reflection groups , i. _ in : geometric group theory down under ( ed .
j. cossey ) , w. de gruyter , 1999 , 181 - 193 . | the lattice of intersections of reflecting hyperplanes of a complex reflection group @xmath0 may be considered as the poset of @xmath1-eigenspaces of the elements of @xmath0 . in this paper
we replace @xmath1 with an arbitrary eigenvalue and study the topology and homology representation of the resulting poset . after posing the main question of whether this poset is shellable ,
we show that all its upper intervals are geometric lattices , and then answer the question in the affirmative for the infinite family @xmath2 of complex reflection groups , and the first 31 of the 34 exceptional groups , by constructing cl - shellings .
in addition , we completely determine when these eigenspaces of @xmath0 form a @xmath3 ( resp .
free ) arrangement . for the symmetric group
, we also extend the combinatorial model available for its intersection lattice to all other eigenvalues by introducing _ balanced partition posets _
, presented as particular upper order ideals of dowling lattices , study the representation afforded by the top ( co)homology group , and give a simple map to the posets of pointed @xmath4-divisible partitions . [ section ] [ theorem]proposition [ theorem]lemma [ theorem]corollary [ theorem]conjecture [ theorem]definition [ theorem]example [ theorem]observation [ theorem]convention [ theorem]observations [ theorem]question [ theorem]problem [ theorem]remark l |
space - charge - limited current ( sclc ) gives the maximum current that can be transported across a solid of length @xmath1 with a biased voltage @xmath2 , limited by the electrostatic repulsion generated by the _ in - transit _ unscreened charge carriers that are in excess of the thermodynamically allowed population @xcite . in a trap - free bulk crystal ,
sclc exhibits a signature current - voltage ( @xmath8-@xmath2 ) characteristics of @xmath9 known as the mott - gurney ( mg ) law @xcite , which is the solid - state counterpart of the sclc in vacuum as given by the child - langumir ( cl ) law : @xmath10 in classical regime @xcite and @xmath11 in quantum regime @xcite . including defect states or traps in solids ,
sclc becomes _ trap - limited _ as described by the mark - helfrich ( mh ) law @xcite : @xmath12 , where @xmath13 , @xmath14 is temperature and @xmath15 is a parameter characterizing the exponential spread in energy of the traps .
due to the geometrical effect @xcite , the 1d sclc value is enhanced as a result of finite emission area @xcite and weakened coulomb screening in high aspect - ratio nanowire @xcite .
furthermore , sclc is an important tool to probe the trap characteristics in solids , and also for photocurrent measurement since the extraction efficiency of photogenerated carriers is fundamentally limited by sclc @xcite . for organic semiconductors , field - dependent @xcite and density - dependent mobility sclc models @xcite
are commonly employed to characterize the sclc carried by the holes .
similarly , sclc of electrons was found to be universally described by a trap model with gaussian energy distribution in a large class of organic semiconductors @xcite .
recently it is demonstrated that the magnitude of electron sclc can be significantly enhanced via the dilution of traps in conjugated polymer blends of only 10% of active semiconductor @xcite , which opens up an exciting possibility of high - efficiency and low - cost organic light emitting diode .
sclc in the trap - limited regime was re - formulated @xcite with inclusion of the interplay between dopants and traps , poole - frenkel effect @xcite and quantum mechanical tunneling , which has solved the long - standing problem @xcite of the enormously sharp current rise at the trap - limited regime and demonstrated that an exponentially distributed trap is not necessarily required to explain the power - law sharp rises of sclc in the trap - limited regime . remarkably , the model successfully reproduced the anomalous noise - spectrum peak observed in @xcite . in spite of sclc being a classic model first derived in 1940s ,
it remains an active topic for organic materials and nanowires as mentioned above . with the advances in fabricating novel 2d dirac materials @xcite ,
it is of the interests to revisit the sclc model for these 2d dirac materials . to our best knowledge
, there is no theory or model to deal with the sclc transport in dirac materials .
recent experiments reported a typical @xmath8-@xmath2 characteristic in the form of mh law for highly disordered materials like reduced graphene oxide @xcite . on the other hand , sclc in crystalline monolayer mos@xmath7 @xcite and hbn @xcite
was found to exhibit an unusual power law dependence of @xmath16 with @xmath17 , which was claimed to be originated from the different levels of traps in different samples by using the traditional mh law .
this explanation is doubtful as the traditional mh law is only valid for @xmath18 for @xmath19 , which implies that the voltage scaling from the mh law must be @xmath20 theoretically and it can not be used to fit with the mesured scaling of @xmath17 . for @xmath21 ,
the traps are narrowly distributed in energy space and the sclc essentially reduces to single - level shallow trapping with @xmath22 @xcite .
thus , the observation of @xmath23 can not be explained by the mh law or other sclc models such as shallow trap @xcite , gaussian disorder @xcite , field - dependent @xcite and density - dependent @xcite mobility . for a dirac material with finite bandgap ,
the electrons mimics relativistic _ massive dirac fermions _
@xcite whereas the classical sclc models are based on the conduction model of _ non - relativistic _ quasi - free electrons @xcite . in this work ,
we proposed a model of _ relativistic _ sclc of massive dirac fermions , which can explain the peculiar @xmath23 scaling observed in recent experiments using dirac materials and thus circumvents the un - justified assumption of @xmath21 used in the mh law in order to fit the experimental data . according to our model
, the @xmath24-@xmath2 characteristics of sclc in a 1d bulk geometry will vary between the non - relativistic limit of @xmath25 to the ultra - relativistic limit of @xmath26 ( this is different from the cl law - see below for explanation ) .
we present a master equation which is in good agreement with the experimental data and can be used to characterize the transition between the ohmic conduction and sclc regime . by extending the bulk 1d model to a 2d thin film model , the scaling relation becomes @xmath0 with @xmath27 varying between 3/2 and 2 , respectively , at the ultra - relativistic and non - relativistic limits . in doing so , we prove rigorously , using a green s function approach@xcite , that the 1d bulk sclc current - voltage scaling relation can be directly mapped to 2d thin - film sclc . it is shown that for a general transport equation of @xmath28 where @xmath29 is a mobility that depends on carrier density , @xmath30 , and @xmath31 is the electric field , the 1d bulk sclc and 2d thin - film sclc are linked by a universal sclc scaling relation ( see section iii ) .
our analysis provides a convenient tool to deduce the 2d thin - film sclc scaling relation via simple 1d sclc model without the need of explicitly solving the complicated 2d sclc model .
in this section , a relativistic sclc model is developed using semiclassical boltzmann transport equation ( bte ) . for simplicity , we first consider the sclc by assuming a simple 1d poisson equation , which allows semi - analytical scaling relations to be determined . in section iii , we shall show that the simple 1d sclc scaling relation derived in this section can be directly mapped to the case of 2d sclc with thin - film geometry .
the starting point of the trap - free sclc theory , i.e. the mott - gurney s law , is the semiclassical bte which provides a basic equation of current density governing the transport of charge carriers .
the diffusion component is usually not considered except in some cases of polymers due to their highly disordered nature . for a quasi - static system , the bte in the linearized transport regime under relaxation - time approximation is @xmath32 where @xmath33 is the electric field , @xmath34 is the crystal momentum , @xmath35 is the position vector , @xmath36 is the carrier velocity , f is the out - of - equilibrium distribution function , @xmath37 is the equilibrium fermi - dirac distribution function , and @xmath38 is a typical collision time scale . if the system is spatially homogeneous the diffusion component of the transport current , i.e. @xmath39 , can be omitted . by assuming a 3d isotropic parabolic energy dispersion , one arrives at the well - known drift - current density , @xmath40 , for semiconductors : @xmath41 by connecting the drift - equation with the 1d poisson equation via charge density @xmath42
, the mott - gurney current - voltage scaling of @xmath43 can be recovered . for 2d gapped dirac materials , eq .
( 2 ) is no longer valid due to two reasons : ( i ) the dimensionality is reduced to 2d ; and ( ii ) the energy dispersion follows a relativistic dispersion similar to that of the massive dirac fermions . in the following
, we shall formulate the drift - equation for dirac semiconductor based on bte approach and demonstrate that the sclc mediated by relativistic quasiparticles follows a completely different current - voltage scaling relation . for massive dirac fermions ,
the energy dispersion is @xmath44 where @xmath45 is the fermi velocity , @xmath46 is the crystal momentum and @xmath47 is the bandgap .
the group velocity is @xmath48 .
the density of states @xmath49 , is rewritten as @xmath50 where @xmath51 denotes the spin - valley degeneracy and @xmath52 is a heaviside function . the electron density at low temperature
can then be obtained from the two - dimensional ( 2d ) density of states @xmath53 which gives @xmath54 and @xmath55 is the fermi level .
the general expression of the 2d linear current density is @xmath56 where @xmath37 is the fermi - dirac distribution function , @xmath38 is the scattering time and @xmath31 is the externally applied electric field . in the low temperature limit ,
the current density can be analytically solved to give @xmath57 .
eliminating @xmath55 via @xmath30 , we obtain @xmath58 where @xmath59 is a bandgap - dependent _ characteristic charge density _ , @xmath30 and @xmath31 are re - expressed as functions of the transport direction , @xmath60 .
the term @xmath61 in eq .
( [ ce ] ) represents a major difference between the relativistic massive dirac fermions and that of the non - relativistic quasi - free electrons . as @xmath62
, it can be seen from the @xmath63-@xmath46 relation that the electrons are non - relativistic at very large @xmath64 or @xmath65 . for
vanishingly small @xmath64 or @xmath66 , the electrons approach ultra - relativsitic limit and become _ massless _ dirac fermions . by expressing eq .
( 1 ) in the drude form of @xmath67 , a density - dependent _ dirac mobility _ is defined as @xmath68 where @xmath69 .
consequently , the relativistic sclc of massive dirac fermions belongs to the class of _ density - dependent mobility _ sclc .
however , @xmath68 is unique to the massive dirac fermions .
we assume that the sclc is carried solely by electrons injected through an ohmic contact . for simplicity ,
we employ the 1d poisson equation @xmath70 where @xmath71 , @xmath72 is the electrostatic potential , and @xmath73 is the effective dielectric constant . here , we first assume that the 3d carrier density , @xmath74 , is related to @xmath42 via @xmath75 with @xmath76 as an effective thickness ( see modification later ) . by coupling eq .
( [ ce ] ) with the poisson equation via @xmath42 , we obtain the governing equation of the relativistic sclc for dirac solid : @xmath77 we first investigate the solutions of the 1d relativistic sclc , i.e. eq .
( [ sclc ] ) , in two asymptotic limits : ( i ) non - relativistic sclc regime ( @xmath78 ) ; and ( ii ) ultra - relativistic sclc regime ( @xmath79 ) , which allows eq .
( [ ce ] ) to be approximated , respectively , as [ lim ] @xmath80 the mg scaling is readily recovered from the non - relativistic charge dynamics at large @xmath64 . as the electrons reside just slightly above the bandgap where the inequality @xmath81 holds true , we have @xmath82 where @xmath83 .
it can be shown @xcite that the corresponding current density is in the non - relativistic drude form , which recovers the mg scaling of @xmath84 . in the opposite limit of @xmath85
, @xmath86 implies ultra - relativistic dynamics with a scaling of @xmath87 as shown in eq .
coincidentally , this has the same scaling to the cl law @xcite although the underlying physics is fundamentally different
. for dirac materials studied there , the ultra - relativistic sclc is obtained from @xmath88 while sclc in vacuum ( or cl law ) is obtained from @xmath89 a simple dimensional analysis @xcite immediately shows that @xmath90 and @xmath91 are both proportional to @xmath92 albeit their very different origins .
nonetheless , eq . ( [ dense ] ) originates from the @xmath93 dependence owing to the _ ultra - relativistic _ electron dynamics in dirac solids while eq .
( [ cl ] ) originates from the @xmath94 dependence owing to the energy balance of a _ non - relativistic _ free electron accelerating in vacuum .
this demonstrates the fundamentally different mechanism behind the @xmath95 scaling in the two cases .
the two limits : @xmath96 and @xmath22 are , respectively , the extreme limits of the ultra - relativsitic and non - relativistic sclc , an intermediate regime of @xmath97 is expected in the case of massive dirac fermions .
this is in good agreement with the experimental observations of @xmath98 in monolayer mos@xmath7 @xcite and @xmath99 in monolayer hbn @xcite where the charge carriers are essentially massive dirac fermions .
the model proposed here suggests that the @xmath23 scaling is @xmath100 to the relativistic carriers without the unjustified or invalid assumption of introducing traps with @xmath21 , which is also inconsistent with the original formulation of the mh law and other trap - limited sclc models as discussed above . as a function of @xmath101 .
( a ) numerical results of @xmath102 over the full range of @xmath101 .
the dashed line shows the empirical fitting equation ; numerical results with ( b ) @xmath103 ; and at ( c ) @xmath104 .
the fitting constants ( @xmath105 , @xmath106 , @xmath107 , @xmath108 , @xmath109 ) are ( 1.067 , 1.45 , 0.889,0.368,1.092 ) . ] for convenience , we transform eq .
( [ sclc ] ) into a dimensionless form of @xmath110 where @xmath111 and @xmath112 .
the normalized current @xmath102 and the dimensionless parameter @xmath101 are [ para ] @xmath113 for a given value of @xmath101 and boundary conditions : @xmath114 and @xmath115 , eq .
( [ sclc2 ] ) is solved numerically at various @xmath102 . the corresponding space charge limited current ( sclc )
is determined when the value of @xmath102 will cause the onset of @xmath116 , and the calculated sclc @xmath102 is plotted as a function of @xmath101 in fig .
1(a ) , which exhibits contrasting behaviour at @xmath117 and at @xmath104 . for @xmath103 [ fig .
1(b ) ] and @xmath104 [ fig .
1(c ) ] , the numerical results can be fitted by @xmath118 and @xmath119 , respectively , where @xmath120 .
the contrasting @xmath101-dependence can be understood from the dependence of @xmath121 , which corresponds , respectively , to the ultra - relativistic sclc at @xmath103 ( or small @xmath101 ) , and the non - relativistic sclc at @xmath104 ( or large @xmath101 ) . by substituting eq .
( [ para ] ) into the above mentioned fitting equations , @xmath122 and @xmath123 are recovered , thus confirming the analytical solutions given in eq .
( [ lim ] ) .
figure 2 shows the smooth transition of @xmath2 and @xmath1 in between the ultra - relativisitc and non - relativistic regimes .
the dimensionless current density @xmath124-@xmath125 characteristic ( at a fixed @xmath1 ) exhibits a voltage scaling of @xmath3 = 2 at low-@xmath126 and @xmath3 = 3/2 at high-@xmath125 . here
the dimensionless parameters are @xmath127 , @xmath128 , @xmath129 and @xmath130 . at a fixed @xmath1
, we have @xmath131 . and thus low-@xmath125 and high-@xmath125 correspond to the non - relativistic ( @xmath132 ) and the ultra - relativistic ( @xmath133 ) regimes , respectively as shown in fig .
2 ( blue axis , @xmath134-symbols ) . in the intermediate regime
, @xmath124-@xmath125 deviates from the simple power law , and applying a fitting would lead to a sub - quadratic scaling in the range of @xmath6 .
similarly , the dimensionless @xmath135-@xmath136 characteristic ( at a fixed @xmath2 ) shows a length scaling of @xmath137 of @xmath138 at small @xmath136 and @xmath139 at large @xmath136 , as shown in fig .
2 ( green axis , @xmath140-symbols ) . here , the dimensionless parameters are @xmath141 , @xmath142 , @xmath143 and @xmath144 . as @xmath145 , small @xmath136 and large
@xmath136 corresponds to the non - relativistic and the ultra - relativistic regimes , respectively .
the @xmath146 sub - cubic inverse length scaling represents another signature of the relativistic sclc for dirac solids in addition to the @xmath23 voltage scaling .
-symbols ) and length ( green axes , @xmath140-symbols ) characteristics of the current density .
the broken ( red ) and the dotted ( black ) guide - line represents ultra - relativistic and non - relativistic limit , respectively .
the labels of the guide - lines , i.e. 1 , 2 , 3 and 4 , correspond to the scalings @xmath147 , @xmath148 , @xmath149 and @xmath150 , respectively . ] from fig .
1(a ) , the numerical results over a wide range of @xmath101 can be accurately fitted by @xmath151 where @xmath152 . as all parameters are intrinsically contained in @xmath102 and @xmath101 , this empirical relation is universally valid and thus we derive a _
master equation _ that universally describes the sclc transport over a wide range of parameters : @xmath153 where @xmath154 is the total current , @xmath155 is the device width , and @xmath156 it is important to emphasize that eq .
( [ emp3 ] ) is extremely usefully if it is used to fit with the experimental @xmath157-@xmath2 measurement ( in the form of @xmath158 as a function of @xmath159 ) to determine the values of @xmath160 and @xmath161 , which can be subsequently used to determine the collision time scale @xmath38 by using eqs .
( 9 ) if the other parameters are known . for the ultra - relativistic limit at @xmath162 , eq .
( [ emp3 ] ) becomes @xmath163 which confirms the @xmath164 ultra - relativistic scaling of @xmath165 .
for the non - relativistic limit at @xmath166 , eq . ( [ emp3 ] ) reduces to @xmath167 which recovers the classical mg scaling of @xmath168 as expected .
therefore , the intermediate relativistic sclc will produce a positive intercept on the vertical - axis of @xmath169-@xmath159 characteristic whereas the sclc with non - relativistic scaling will have a zero intercept as shown in fig .
interestingly , the @xmath158-@xmath159 characteristic [ suggested in eq . ( 10 ) ] provides a convenient tool to represent the sclc data that can be generally applied to any solids . to illustrate this point
, we consider a trap - free solid in which the conduction transits from ohmic to sclc at increasing @xmath2 as showin in fig . 3a . in the ohmic regime ( large @xmath159 or small @xmath2 )
where @xmath170 , we have @xmath171 , i.e. @xmath158 decreases with increasing @xmath159 ( green dash - dotted line in fig .
in contrast , in the sclc regime ( small @xmath159 or large @xmath2 ) where @xmath172 , we have @xmath173 , i.e. @xmath158 increases linearly with @xmath159 ( blue dashed line in fig 3a ) .
these contrasting behaviors lead to a @xmath174 peak @xcite in the intermediate regime that clearly separates the sclc - dominated and ohmic - dominated conduction as shown in fig .
this finding is confirmed by using various experimental data ( color symbols ) for mos@xmath7 from ref .
@xcite at different temperature as shown in fig .
the above mentioned transitional peak between sclc - dominated regime ( dashed curve ) and ohmic - dominated regime ( dash - dotted curve ) can be clearly observed at all temperatures . a zoom - in view at the small-@xmath159 sclc - dominated regime
is shown in fig .
3(b ) , which indicates that the experimental results ( symbols ) can be explained by linear fitting so to obtain the voltage scaling , which ranges from @xmath3 = 2.11 down to 1.67 according to eq . ( [ emp3 ] ) . the @xmath3 scaling decreases to @xmath23 at elevated temperature because the higher energy levels of the conduction band is increasingly populated by the thermally - liberated electrons from trapping sites .
this leads to a higher - degree of relativistic dynamics of the transport electrons thus reducing @xmath3 .
for the three @xmath23 cases , they can be extrapolated to have positive intercepts on the @xmath175-axis at @xmath176 0 [ see fig .
3(b ) ] . as @xmath3 approaches 2 ,
the intercepts diminishes and becomes approximately zero at @xmath22 .
these observations are in good agreement with the predicted intercepts of eq .
( [ emp3 ] ) , as discussed above .
note that eq .
( [ emp3 ] ) breaks down in the case of @xmath18 where the intercept becomes negative . against @xmath159 in the sclc regime using mos@xmath7 experimental data from ref .
@xcite with ( t,@xmath3 ) of ( 285k , 1.67 ) ( blue square ) , ( 265k , 1.73 ) ( green circle ) , ( 245k , 1.82 ) ( yellow triangle ) , ( 205k , 2.00 ) ( red diamond ) and ( 185k , 2.11 ) ( black star ) . (
a ) the entire @xmath158 range over 0.4 @xmath177 100 .
the green dash - dotted and blue dashed curves denote , respectively , ohmic current ( @xmath170 ) and sclc ( @xmath43 ) fitted to the @xmath178 k data .
the horizontal gray line indicate the transitional regime of @xmath95 ( 1 @xmath179 6 ) separating the green - shaded sclc - dominated and the yellow - shaded ohmic - dominated regime .
( b ) the sclc - dominated regime at small @xmath180 0.7 . ]
the relativistic sclc model derived above is based on solving the 1d poisson equation . as dirac semiconductor
is a 2d thin film , thus a 2d thin - form model is required . in this section ,
we provide a simplified formalism of 2d thin - film relativistic sclc model without the need to explicitly solve for the 2d model @xcite .
the sclc has been previously formulated for thin - film and nanowire using an integral form of 2d electrostatic poisson equation @xcite . here
, we shall formulate a thin - film sclc relativistic model under the assumption of carrier density - dependent mobility to illustrate that the general sclc scaling properties for dirac semiconductor , which has a density - dependent mobility of @xmath182 .
we consider the @xmath181-dimension transport in a solid with density - dependent mobility in a general form of @xmath183 / f_0 $ ] where @xmath184 $ ] is a density - dependent term and @xmath37 is a constant factor .
the dimensionality of @xmath185 and @xmath186 corresponds to bulk and 2d thin film , respectively . in the following analysis
, we consider the case of _ uniform sclc injection _ where the two electrodes are separated by a fixed spacing of @xmath1 as shown in fig .
for uniform sclc injection along the @xmath60-direction , the electric field profiles , i.e. @xmath187 for bulk ( @xmath185 ) and thin film ( @xmath186 ) can be written , respectively , as @xmath188 where @xmath189 and @xmath190 denote the surface and volume carrier density respectively . @xmath191 and @xmath192 is , respectively , the 1d and 2d green s function that are dependent on the geometry of contacts .
figures 4(b ) and ( c ) shows a 2d thin film with two possible contact geometries , i.e. edge and strip contacts , respectively@xcite . by eliminating the @xmath193-integration via @xmath194 and suppressing the argument of @xmath195 in eq .
( 12b ) for simplicity , eq .
( 12 ) can be written compactly as @xmath196 where we have introduced dimensionless variable as @xmath197 , and the subscript of @xmath198 denotes surface and bulk carrier density , respectively .
consider the charge transport mechanism in dirac solid belongs to the class of density - dependent mobility model , we assume a general current density of @xmath199}{f_0 } e_{d}(\xi),\ ] ] where the subscript @xmath200 denotes linear and areal current density respectively .
the density - dependent mobility is given as @xmath201 \equiv \mu_0f[n_\nu(\xi)]/f_0 $ ] where @xmath202 $ ] is a @xmath203-dependent term and @xmath37 is a normalization constant . from eq .
( 14 ) , the bias voltage relation : @xmath204 is written as @xmath205 } = e\mu_0\frac{v}{l}.\ ] ] the solution of eq .
( 15 ) gives the equation of sclc .
its full solution require the knowledge of @xmath203 over the intervals from @xmath206 to @xmath207 , which can be obtained by solving the nonlinear integral equation in eq .
nonetheless , the scaling relations , i.e. @xmath208-@xmath2 and @xmath208-@xmath1 , can be readily deduced via a simple _ dimensional analysis _
@xcite without explicitly solving eqs .
( 13 ) and ( 15 ) .
contact geometry for uniform sclc injection .
current is uniformly injected along the @xmath60-direction . for @xmath185 bulk geometry , the constant-@xmath1 geometry is invariant along both @xmath209-and @xmath175-directions whereas for @xmath186 thin - film geometry , the structure is only invariant along the @xmath209-direction .
( c ) and ( d ) shows the side view of constant-@xmath1 2d thin film with edge and strip contacts , respectively . ] to illustrate this , we first combine eqs . ( 13 ) and ( 14 ) to obtain @xmath210 \int_0 ^ 1 d\xi ' \frac{\partial g_{d}(\xi,\xi')}{\partial \xi}n_{\nu}(\xi ' ) .\ ] ] from the poisson equation , i.e. @xmath211 where @xmath212 is a @xmath181-dimensional dirac delta function , the physical dimension of @xmath213 can be obtained as @xmath214 = \mathcal{l}^{2-d}$ ] , where @xmath215 denotes the fundamental dimension of length and @xmath216 $ ] denotes the unit of physical quantity @xmath217 . correspondingly , the partial derivative , @xmath218 , in eq .
( 16 ) can be non - dimensionalized as @xmath219 where @xmath220 is a dimensionless green s function .
we now rewrite eqs .
( 15 ) and ( 16 ) as @xmath221 where @xmath222 $ ] for simplicity . a direct inspection of eq .
( 18a ) shows that after the integral @xmath223 is fully converted into a dimensionless form , the @xmath208-@xmath2 and @xmath208-@xmath1 scaling relations can be unambiguously determined . in this case
, @xmath223 becomes a _
dimensionless numeric factor _ that does not play any roles in the @xmath208-@xmath2 and @xmath208-@xmath1 scaling relations . the non - dimensionalization of eq .
( 18a ) can be accomplished by appropriately regrouping the constant term in eq .
( 18b ) , i.e. @xmath224 into each of the @xmath225 and @xmath226 terms in the right - hand side of eq .
( 18b ) such that dimensionless terms @xmath227 and @xmath228 can be defined , respectively , for @xmath203 and @xmath226 .
in general , the regrouping of @xmath229 can be expressed in an arbitrary form of @xmath230 where @xmath231 and @xmath232 are terms containing @xmath208 and @xmath233 and @xmath234 are terms containing @xmath235 . the roles of @xmath236 s and @xmath237 s are to pair up with @xmath238 and @xmath226 in eq .
( 18b ) such that the resulting terms are dimensionless . in the following ,
we suppress the argument of @xmath236 s and @xmath237 s for simplicity . as the explicit form of @xmath226 determines the regrouping of @xmath229 , @xmath236 s and @xmath237 s
are both @xmath226-dependent . furthermore ,
@xmath236 s are @xmath181-independent and @xmath237 s are @xmath181-dependent as @xmath235 is deliberately distributed only into @xmath237 s .
we can now recast eq .
( 18b ) as @xmath239 where all terms are dimensionless via the following grouping @xmath240 with @xmath227 and @xmath241 now being dimensionless , eq .
( 18a ) can be rewritten as : @xmath242 or more compactly as @xmath243 where @xmath244 is a dimensionless numeric factor dependent on the @xmath181 and @xmath245 , i.e. @xmath246 which can be explicitly solved from eq .
( 18b ) and it affects only the overall magnitude of sclc without affecting its voltage and length scaling relations .
thus the @xmath208-@xmath2 and @xmath208-@xmath1 scaling relations are determined by @xmath247 equation ( 25 ) represents a _
universal sclc scaling relations _ for uniform sclc injection into either a @xmath185 ( bulk ) or @xmath186 ( thin film ) based solid of length @xmath1 with arbitrary density - dependent mobility @xmath201 $ ] .
several remarkable properties can be extracted from eq .
( 25 ) : ( i ) the @xmath208-@xmath2 scaling , i.e. @xmath248 , is determined solely by the @xmath201 $ ] and is _ completely independent on the device geometry ( @xmath245 ) and dimensionality ( @xmath181 ) _ ; ( ii ) the @xmath208-@xmath1 scaling , on the other hand , is affected by both @xmath226 and @xmath181 ; ( iii ) for a fixed @xmath181 (= 1 for bulk or = 2 for thin film ) , the geometry of contacts affects only @xmath244 and hence both scl @xmath208-@xmath2 and @xmath208-@xmath1 scaling are universal independent on contact . from eq .
( 19 ) , we see the constant term carriers a length scale dependence of @xmath235 , and thus eq .
( 19 ) will be independent of @xmath1 for @xmath186 ( for a thin film setting ) resulting in @xmath249 . in this 2d thin film limit ,
( 23 ) gives the sclc relation for a thin film : @xmath250 note that eq .
( 26 ) includes that both @xmath251-@xmath2 and @xmath251-@xmath1 follow the same scaling relation .
togther with property ( i ) and eq .
( 26 ) , it shows a powerful tool that can be used to directly map the scaling relation of a simple bulk sclc model into 2d thin - film sclc model . by virtual of property ( i ) ,
we conclude that the voltage scaling for thin - film ( @xmath251-@xmath2 ) is identical to the bulk @xmath24-@xmath2 scaling , thus the voltage scaling obtained in sec .
ii are valid for thin film as well . for a thin film , eq .
( 26 ) also dictates that the length ( @xmath251-@xmath1 ) scaling relation is identical to the @xmath251-@xmath2 . to summarize this session
, we provide a rigorous derivation of scaling laws ( both voltage and length ) for uniform scl injection into a 2d thin film seeting .
these properties allow the @xmath251-@xmath2 and @xmath251-@xmath1 scaling relations of a 2d thin - film to be fully determined from a simple 1d bulk sclc model that have been shown in sec .
the full @xmath251-@xmath2 and @xmath251-@xmath1 scaling relations are thus obtained without the need of explicitly solving the complicated coupled equations in eq ( 18 ) . in appendix a , two examples of 2d thin - film sclc are analyzed using our simplified formalism developed here .
the relativistic sclc scaling relation of dirac semiconductor in 2d thin - film geometry can be readily determined by using the simple derivation developed above . since the 1d sclc scaling relations takes the form of @xmath252 and @xmath253 respectively for the ultra - relativistic and non - relativistic regimes , for 2d thin film , our simple analysis yields @xmath254 where @xmath3 = 2 and @xmath3 = 3/2 are for the non - relativistic and ultra - relativistic limit , respectively . in the intermediate regime
, the scaling follows an approximate power - law form with @xmath3 varies continuously from @xmath255 akin to fig .
2d thin film of dirac semiconductor .
the dashed and dotted lines denotes @xmath256 and @xmath257 , respectively . as @xmath258 and @xmath259
, the small-@xmath260 and large-@xmath260 regime corresponds to @xmath261 ( non - relativistic ) and @xmath262 ( ultra - relativistic ) , respectively .
note that the data points exhibits oscillations at small @xmath260 due to numerical error . ] to verify eq .
( 27 ) , the relativistic sclc in 2d thin - film geometry with inclusion of exponential traps is explicitly solved ( the full derivation is presented in appendix b ) . in the presence of exponential traps ,
the relativistic sclc in a 2d thin - film dirac semiconductor is [ 2d ] @xmath263 where @xmath264 , @xmath265 is the 2d carrier density , @xmath266 , @xmath267 with @xmath268 as the effective density of states at the conduction band - edge and @xmath269 as the trap density , @xmath270 and @xmath271 are dimensionless variables . in the ultra - relativistic and non - relativistic sclc regimes ,
we obtain @xmath272 and @xmath273 , respectively . by setting @xmath274 ( which corresponding to a trap - free case )
, we obtain @xmath275 ( non - relativistic ) and @xmath262 ( ultra - relativistic ) , thus confirming the simple derivation in eq .
this agreement demonstrates that the unconventional @xmath208-@xmath2 scaling of @xmath6 is a universal signature of the relativistic charge carrier dynamics in both bulk and 2d thin - film geometries . in comparison to prior works in ref .
@xcite , the variation of @xmath276 to @xmath277 with decreasing temperature was attributed to the transition from @xmath278 ( valid ) to @xmath279 ( invalid ) .
our relativistic sclc model with exponential traps presented here is however ever to take account for such temperature dependence without imposing the invalid @xmath21 condition . to further confirm that the analytical relativistic slcc scaling relation obtained above , the integral equation in eq .
( 28a ) , which belongs to the class of _ nonlinear cauchy singular integral equation _
@xcite is numerically solved for the trap - free case of @xmath274 .
the numerical solved @xmath265 [ from eq .
( 28a ) ] is then integrated in eq .
( 28b ) to obtain the @xmath251 as a function of @xmath2 . for simplicity , eq .
( 28 ) is solved in terms of a dimensionless variables , i.e. @xmath259 and @xmath258 [ see appendix c and eqs .
( c6 ) and ( c7 ) for the definition of @xmath280 and @xmath281 .
the numerical results ( red circles ) of @xmath282 ) as a function of @xmath280 is shown in fig .
5 show good agreement with the derived scaling laws : @xmath256 ( dotted lines ) at small voltage of @xmath283 and @xmath257 at large voltage of of @xmath284 .
thus , the comparison confirms the two corresponding analytical scaling laws [ see eq .
( 27 ) ] for space charge limited conduction in a 2d thin - film dirac solid with finite bangap : @xmath261 and @xmath262 respectively for the non - relativistic and ultra - relativistic limits .
more importantly , the unconventional relativistic sclc scaling of @xmath97 is unambiguously confirmed for the 2d thin - film dirac semiconductor and is in agreement with experiments @xcite . finally , we discuss the screening effect on the relativistic sclc in 2d dirac materials . in 2d dirac materials ,
the charge transport is sensitively influenced by the substrate screening and excess charge screening induced by gate - electrode in field - effect - transistor geometry . despite these screening effects , sclc was unambiguously observed in experiments as reported in refs . [ 31 ] and [ 32 ] .
these experimental observations suggest that the screening effect can not entirely remove sclc in 2d materials . in a previous theoretical work @xcite
, it is demonstrated that the surrounding dielectric screening will affect the transport properties of 2d thin film by modifying the relaxation time @xmath38 and a significant mobility enhancement can be achieved via high- dielectric substrate with vanishingly thin membrane . this theoretical prediction was experimentally confirmed in monolayer mos2 with high- substrate@xcite . in relevance to our sclc model , we point out that as the substrate screening effect alters only the relaxation time @xmath38 which comes into the sclc picture as a proportionality constant , it can be reasonably expected that only the magnitude of the sclc will be altered while the new scaling laws reported here will remain unchanged . a microscopic theory of substrate screening can be formulated via first - principle calculation that takes into account the complex many - body interactions at the interface between 2d materials and the substrate@xcite .
the complete microscopic quantum picture od dielectric screening is beyond the scope of this work .
in summary , we have proposed a theory of relativistic space charge limited conduction ( sclc ) in dirac solids with new scaling laws for both bulk and thin file model .
for the one - dimensional ( 1d ) bulk model , the scaling laws are @xmath285 with @xmath97 and @xmath286 . for 2d thin film model
, we have @xmath287 for uniform sclc injection with @xmath3 remaining the same as the case of 1d bulk model under the assumption of density dependent mobility .
both scaling laws have been verified with numerical calculations and have good agreements with experimental results .
the important finding from this paper is the new voltage scaling of @xmath23 which is a signature of the massive dirac fermions in the 2d dirac materials , and can not be explained by using the traditional sclc models derived decades ago for traditional materials .
the inconsistencies in using such traditional sclc models with unjustified traps condition to fit the experimental measurement is questionable .
our results represent a new class of relativistic space - charge phenomena in dirac solids , which may be used to model dirac - based devices operating in space - charge - limited regime and may also be used as a tool to extract useful parameters by fitting the analytical equations with measurements
. the relativistic sclc model should generate unconventional scl photocurrent response @xcite in relativistic dirac semiconductor such as mos@xmath7 .
the widely - studied photoresponse of mos@xmath7 @xcite can be readily used as an additional platform to verify the proposed relativistic sclc model here .
we thank subhamoy ghatak for providing us the experimental data of mos@xmath7 and insightful discussions .
we thank chun yun kee , kelvin j. a. ooi and shi - jun liang for helpful discussions .
this work is supported by singapore ministry of education t2 grant ( t2moe1401 ) and usa afoar aoard grant ( fa2386 - 14 - 1 - 4020 ) .
in this appendix , we illustrates the simple derivation of the 2d sclc scaling relations developed in sections iii - a and iii - b using two examples . in the first example
, we consider a trivial case with @xmath288 , i.e. the mobility is independent of carrier density . for bulk model of @xmath185 , eq .
( 18b ) becomes @xmath289 which can be fully non - dimensionalized by defining @xmath290 , @xmath291 , @xmath292 and @xmath293 . from eq .
( 23 ) , we obtained @xmath294 , which can be rearranged to give the well - knwon bulk mg law of @xmath295 the numerical factor can be solved as @xmath296 via eq .
( 24 ) by using a 1d green s function @xcite .
we can now map the @xmath24-@xmath2 bulk sclc scaling relation to the 2d thin - film case , which yields @xmath297 .
furthermore , as @xmath251-@xmath1 scales equally with @xmath251-@xmath2 , the 2d thin film sclc scaling relation can now be fully determined as @xmath275 .
one can verify this scaling relation by explicitly solving eq .
( 26 ) with @xmath298 and @xmath299 .
this gives @xmath300 , which can be rearranged to give the well - known 2d thin - film sclc @xcite , i.e. @xmath301 where @xmath302 is a @xmath303-dependent numeric factor . in the second example
, we consider a carrier density dependent mobility in a power - law form , i.e. @xmath304 and @xmath305 where @xmath306 and @xmath307 are some constants .
this particular form of @xmath55 is equivalent to mark - helfrich s exponential - trap model with @xmath308 . for this particular form of @xmath309 ,
( 18b ) can be fully non - dimensionalized by re - grouping the constant factor @xmath310 via the following definitions : @xmath311 , @xmath312 , @xmath313 and @xmath314 .
the bulk sclc can then be obtained from eq .
( 23 ) as @xmath315 which can be simplified as @xmath316 and is in agreement with the mark - helfrich s exponential trap model @xcite . to generalize the bulk sclc to the case of 2d thin film
, we again utilize the facts that : ( i ) @xmath251-@xmath2 follows the same scaling as @xmath24-@xmath2 ; and ( ii ) @xmath251-@xmath1 scales equally with @xmath251-@xmath2 this gives @xmath317 , which is in agreement with the explicitly solution of eq . (
26 ) , i.e. @xmath318
in this appendix , we provide a full derivation of the mark - helfrich s sclc model and dirac semiconductor in 2d thin - film geometry based on grinburg s formalism @xcite . in the presence of traps that follows an exponential energy distribution @xcite , the free and trapped carrier densities are related by @xmath319 where @xmath320 is the free carrier density , @xmath321 is the trapped carrier density , @xmath268 is the effective density of states at the conduction band edge , @xmath269 is the trap density and @xmath322 . here ,
@xmath323 where @xmath15 is a characteristic temperature representing the exponential spread in energy of the traps .
the charge density in a 2d thin - film is given as @xmath324 , \ ] ] where @xmath325 is a dirac delta function and @xmath326 is the charge density induced on the annode by the total @xmath321 residing in the thin film . note that @xmath175 represent the direction that is out - of - plane of the thin film .
for thin - edge contacts , the corresponding green s function is @xmath327^{1/2},\ ] ] and the scalar potential can then be solved as @xmath328^{1/2 } \left(-\frac{4\pi
e}{\epsilon}\right ) \delta(y ' ) \left [ -n_s(x ) + p_s\delta(l - x ) \right].\ ] ] simplify @xmath329 and knowing that @xmath330 , we obtain @xmath331 by defining @xmath332 and @xmath333 , we obtained @xmath334 we now consider a current density equation in drude s form , i.e. @xmath335 by combining eq .
( b5 ) and ( b6 ) , we obtain @xmath336 equation ( b7 ) can be rearranged as followed : @xmath337 by defining @xmath338 eq .
( b8 ) becomes @xmath339 which is an integral equation that can be solved to obtain @xmath340 .
the bias voltage can be obtained from @xmath341 and eq .
( b6 ) as @xmath342 to obtain the exponential trap - limited sclc in 2d thin film geometry with edge - contact , eqs .
( b9 ) and ( b11 ) are combined to give @xmath343 with the definition of @xmath344 , which is a constant that can be solved from the integral equation in eq .
( b10 ) , we obtain @xmath345 equation ( b13 ) gives the exponential trap - limited sclc of a 2d thin film with edge - contact geometry [ see fig .
4(b ) ] . for strip - geometry
[ see fig .
4(c ) ] , the electric field is given as @xmath346 where @xmath347 and @xmath348 .
using similar procedure , we obtain @xmath349 which can be simplified to @xmath350 from eq .
( b12 ) , the sclc current density equation is obtained as @xmath351 where the numeric factor @xmath352 can be obtained by solving @xmath353 from eq .
( b16 ) . in summary ,
the 2d thin - film uniform injection of sclc in the presence of exponential traps follow the following scaling relation of @xmath354 for both edge - and strip - contact geometries .
more importantly , this scaling relation is in agreement with eq .
( a6 ) obtained using the simplified formalism . for 2d dirac semiconductor ,
we obtain @xmath355 and @xmath356 respectively for edge - contact and strip - contact . the applied bias voltage for edge - and strip - contact
geometries become , respectively , @xmath357 and @xmath358 the coupled eqs .
( b19 ) to ( b22 ) can be solved to obtain the relativistic sclc in 2d thin - film geometry .
equations ( b19 ) to ( b22 ) has to be solved numerically . nonetheless , in the non - relativistic and ultra - relativistic limits , semi - analytical scaling relations can be derived .
we first consider the non - relativistic limit of @xmath359 for all @xmath270 with edge - contacts , eqs .
( b19 ) and ( b21 ) can be approximated , respectively , by @xmath360 and @xmath361 by defining @xmath362 we obtain @xmath363 where @xmath364 is a numerical factor which can be solved from the nonlinear integral equation in eq .
( b19 ) . by setting @xmath274
, the current - voltage scaling relation agrees with the 1d bulk model as shown in eq .
( 5a ) of the main text .
the current - voltage scales equally with the current - length which is also in agreement with the simplified derivation of 2d thin - film sclc scaling relation presented in eq .
( 27 ) of the main text . in the ultra - relativistic limit of @xmath365 , eqs .
( b19 ) and ( b21 ) become , respectively , @xmath366 and @xmath367 where @xmath368 which can be rearranged to give @xmath369 the numerical factor , @xmath370 , can again be solved from eq .
( b19 ) . for @xmath274 ,
the current - voltage scaling relation agrees with the ultra - relativistic results in eq .
( 5b ) of the main text . in the intermediate regime
, the scaling relation can be approximated by @xmath371 where @xmath372 and @xmath373 .
equations . ( b28 ) to ( b32 ) can be transformed into a dimensionless form for numerical solution in fig .
5 of main text . for _ edge - contacts _ , we obtain @xmath374 and @xmath375 for _ strip - contacts _ , the dimensionless form give @xmath376 and @xmath377 the dimensionless parameters are defined as @xmath378 @xmath379 and @xmath380 where @xmath381 is current - and material - dependent parameter .
liu , w .- y .
shan , y. yao , w. yao and d. xiao , _ phys .
b _ * 88 * , 085433 ( 2013 ) ; e. cappelluti , r. roldan , j. a. silva - guillen , p. ordejon and f. guinea , _ phys . rev .
b _ * 88 * , 075409 ( 2013 ) ; q. h. wang , k. kalantar - zadeh , a. kis , h. n. coleman and m. s. strano , _ nature nanotech . _ * 7 * , 699 ( 2012 ) . in this limit , the density of states is @xmath382 and the carrier density is @xmath383 .
correspondingly , the current density is @xmath384 and hence the sclc reduces to the mg law , i.e. @xmath385 .
note that this is differ by a factor of 2 in comparison with eq .
( [ lim ] ) .
this is due to the different expansion scheme used in obtaining eq .
( [ lim ] ) .
the expansion of @xmath386 in eq .
( [ sclc ] ) is performed _ after _ the @xmath34-integration of the current density equation . | a theory of relativistic space - charge - limited current ( sclc ) is formulated to determine the sclc scaling , @xmath0 , for a finite bandgap dirac material of length @xmath1 biased under a voltage @xmath2 . in a one - dimensional ( 1d ) bulk geometry ,
our model allows ( @xmath3 , @xmath4 ) to vary from ( 2,3 ) for the non - relativistic model in traditional solids to ( 3/2,2 ) for the ultra - relativistic model of massless dirac fermions . for a two - dimensional ( 2d ) thin - film geometry , we obtain @xmath5 that varies between 2 and 3/2 , respectively , at the non - relativistic and ultra - relativistic limits .
we further provide a rigorous proof based on a green s function approach that for uniform sclc model described by carrier density - dependent mobility , the scaling relations of the 1d bulk model can be directly mapped into the case of 2d thin film for any contact geometries .
our simplified approach provides a convenient tool to obtain the 2d thin - film sclc scaling relations without the need of explicitly solving the complicated 2d problems .
finally , this work clarifies the inconsistency in using the traditional sclc models to explain the experimental measurement of 2d dirac semiconductor .
we conclude that the voltage - scaling @xmath6 is a distinct signature of massive dirac fermions in dirac semiconductor and is in agreement with experimental sclc measurement in mos@xmath7 . |
increasingly accurate measurements of the anisotropy in the temperature of the cosmic microwave background ( cmb ) has ushered in an era of precision cosmology .
a golden decade of cmb anisotropy measurements by numerous experiments was topped by the results from the first year of data obtained by the wilkinson microwave anisotropy probe ( wmap ) @xcite . under simple hypotheses for the spectrum of primordial perturbations , exquisite estimates of the cosmological parameters
have been obtained from the angular power spectrum measurement by wmap combined with other cosmological observations @xcite .
although the assumed , scale free ( with mild deviations ) , initial power spectra may be a generic prediction of the simplest scenarios of generation of perturbations during inflation , initial spectra with radical deviations are known to arise from very reasonable extensions , or , refinements to the simplest scenarios @xcite .
consequently , cosmological parameter estimation from the cmb anisotropy and the matter power spectrum obtained from redshift surveys , weak gravitational lensing and ly-@xmath5 absorption , depends sensitively on the dimensionality , nature and freedom in the parameter space of initial conditions @xcite . the angular power spectrum , @xmath0 , is a convolution of the initial power spectrum @xmath1 generated in the early universe with a radiative transport kernel , @xmath6 , that is determined by the current values of the cosmological parameters . the remarkably precise observations of the angular power spectrum @xmath0 by wmap , and the concordance of cosmological parameters measured from different cosmological observations opens up the avenue to directly recover the initial power spectrum of the density perturbation from the observations .
the richardson - lucy ( rl ) method deconvolution was shown to be a promising and powerful method to measure the power spectrum of initial perturbations from the cmb angular power spectrum @xcite . in this paper , we apply the method to the cmb anisotropy spectrum measured by wmap . we have also devised and implemented an improvement to the rl scheme , whereby the iterative deconvolution algorithm is designed to converge and _
match the measurements only within the given error - bars .
_ the direct numerical deconvolution has clear advantages over the other prevalent approach of obtaining the most likely parameter values for a parametric model primordial spectra space ` bins ' carried out with recent data is a somewhat model independent @xcite .
direct deconvolution with different method @xcite has been attempted @xcite ] .
first , as emerges from our work , the direct method can reveal features that are not anticipated by the parametric model spectra , hence would be completely missed out in the latter approach .
second , in the absence of an accepted early universe scenario ( more narrowly , a favored model of inflation ) , it is difficult to _ a priori _ set up and justify the chosen space of initial conditions .
the complex covariances between the cosmological and the initial parameters are sensitive to the parameterization of the space of initial spectra adopted .
efforts along these lines are further obscured by issues such as the applicability of the occam s razor to dissuade extension of the parameter space of initial conditions .
such deliberations have been recently framed in the more quantitative language of bayesian evidence to evaluate and select between possible parameterizations @xcite . however , this approach can not really point to a preferred parameterization .
whereas , in the direct approach we can evade the issue of appropriate parameterization of the initial power spectrum . for a given set of cosmological parameters , our method obtains the primordial power spectrum that ` maximizes ' the likelihood .
hence , the space of parameters remains that of the ( more widely accepted and agreed up on ) cosmological parameters alone . in principle , it is possible to explore the entire space of cosmological parameters along the lines being done routinely , and instead of simply computing the likelihood for a given model initial power spectrum , one obtains the initial power spectrum that maximizes the likelihood at that point and assigns that likelihood to that point in the space of cosmological parameters .
however , in this paper we limit ourselves to obtaining the primordial spectrum for a few given sets of ` best fit ' cosmological parameters and showing that recovered @xmath1 is robust to small local variations in the cosmological parameters .
we defer the wide exploration of the cosmological parameter space to future work . in section [ meth ]
, we describe the problem and present the deconvolution method , details of its implementation and tests of recovering known primordial spectra from synthetic angular power spectrum data . the recovered spectrum from wmap data using our method
is described in section [ results ] .
the recovered primordial spectrum has statistically significant features that are robust to the variations in the cosmological parameters . in section [ disc ] , the recovered spectrum is shown to be tantalizingly similar to some forms of broken scale invariant spectra known to arise from reasonable variations of the simplest inflationary scenario .
in this subsection we recall the integral equation for the angular power spectrum of cmb anisotropy and set up the inverse problem that we solve using a deconvolution method .
the observed cmb anisotropy @xmath7 is one realization of a random field on the surface of a sphere and can be expressed in terms of the random variates @xmath8 that are the coefficients of a spherical harmonic expansion given by ( * n*)=_l=2^_m =- l^m = la_lmy_lm(*n * ) where @xmath9 are the spherical harmonic functions . for an underlying isotropic , gaussian statistics , the angular power spectrum ,
@xmath0 defined through a_lm a^*_l^m^ = c_l _ ll^ _ mm^ , completely characterizes the cmb anisotropy . in a flat universe the temperature fluctuation in the cmb photons at the location @xmath10 at the present conformal time @xmath11 propagating in a direction
@xmath12 is ( * n * ) ( x_0 , * n * , _ 0)=d^3ke^i*k*(k,*n*,_0 ) . for globally isotropic cosmology
, the temperature fluctuation @xmath13 can be expanded in terms of legendre polynomials leading to _ t ( x,*n * , _ 0)=d^3ke^i*k * _ l=0^(-i)^l ( 2l+1 ) _
l(k,_0 ) p_l(*kn * ) .
the angular power spectrum @xmath0 given by the coefficients of legendre expansion is then expressed as c_l=(4)^2p(k)|_tl(k,_0)|^2 [ clintg ] where @xmath1 , the power spectrum of the _ primordial ( scalar ) metric _ initial perturbation @xmath14 , is given by _ prim(*k*)_prim^*(*k^*)=(*k*-*k^ * ) because the @xmath15 space modes are uncorrelated in a homogeneous space .
the spectrum @xmath1 represents the _ r.m.s .
_ power in the scalar metric perturbations per logarithmic interval @xmath16 at a wavenumber @xmath15 .
it is related to power spectrum of the primordial modes of density perturbations , @xmath17 , as @xmath18 .
for the conventional scale free parameterization , the spectral index @xmath19 is defined through @xmath20 .
_ the power spectrum @xmath1 is a constant for the scale invariant harrison - zeldovich spectrum ( corresponds to @xmath21 ) . _ the harmonics of the temperature fluctuations at the current epoch , @xmath22 ,
is obtained from the solution to the boltzmann equation for the cmb photon distribution . in this paper
we use @xmath22 computed by the cmbfast software @xcite .
numerically , a suitably discretized space of wave - numbers , @xmath23 is used where the following discrete version of integral eq .
( [ clintg ] ) is applicable c_l= _ i g(l , k_i )
p(k_i ) + g(l , k_i ) = |_tl(k_i,_0)|^2 .
[ clsum ] in the above equation , the _ ` target ' _
angular power spectrum , @xmath24 , is the data given by observations , and the radiative transport kernel , @xmath6 is fixed by the cosmological parameters of the _ ` base ' _ model .
( the kernel @xmath6 also includes the effect of geometrical projection from the three dimensional wavenumber , @xmath15 , to the harmonic multipole , @xmath25 on the two dimensional sphere . ) for a given @xmath6 , obtaining the primordial power spectrum , @xmath1 from the measured @xmath0 is clearly a deconvolution problem .
an important feature of our problem is that @xmath26 , @xmath6 and @xmath1 are all positive definite .
however , to get reliable results from the deconvolution , we require high signal to noise measurements of @xmath26 over a large range of multipoles with good resolution in multipole , preferably , from a single experiment to avoid the uncertainties of relative calibration data that combined the heterogeneous cmb band power obtained from different experimental data sets are not always as reliable due to relative calibration uncertainties .
application of our method to binned data of @xcite did not give robust or convincing results @xcite .
however , in principle nothing rules out using binned data from heterogeneous data sets that have good cross - calibration . ] .
ideally , it will be best to measure each @xmath26 independently . in practice , incomplete sky coverage and other effects
limit the resolution in multipole space .
all experiments provide band power estimates , @xmath27 which are averaged linear combinations of the underlying @xmath0 .
since eq .
( [ clsum ] ) is linear , a similar equation holds for all the band powers with a kernel , @xmath28 .
_ for simplicity and brevity of notation , we retain the notation @xmath26 and @xmath6 for band power estimates with @xmath25 denoting the bin center . we also implicitly assume a discrete wavenumber @xmath15 instead of carrying the clumsy notation , @xmath23 . _ as mentioned in more detail in section [ wmapdat ] , the cmb angular power spectrum from wmap ranges from the quadrupole , @xmath29 to @xmath30 .
moreover , the ` full sky ' coverage of wmap implies good resolution in multipole space .
we use wmap tt ( temperature - temperature ) binned power spectrum as @xmath26 in eq .
( [ clsum ] ) .
we use cmbfast software to compute the the @xmath6 matrix for the post - wmap ` best fit ' cosmological parameters .
we emphasize that although the @xmath1 is recovered using binned data , the significance ( performance ) of the recovered spectrum is evaluated using the wmap likelihood of entire unbinned @xmath0 properly accounting for covariances .
the likelihood is computed using the numerical code , data and its error covariance provided by the wmap collaboration with the release of the first year data . in this work
we limit our attention to the angular spectrum of the temperature anisotropy , @xmath31 . including the polarization of cmb photons , equations similar to eq .
( [ clintg ] ) can be also written for the three additional angular power spectra , @xmath32 , @xmath33 and @xmath34 involving their corresponding kernels .
it will certainly be interesting to include these when more complete polarization data is available in the future . at present
, only the @xmath32 spectrum has been published by the wmap team which is not positive definite and hence , not ideally suited for our method . however , once @xmath33 data is made available , our method can readily accommodate both @xmath32 and @xmath33 together since combinations such as @xmath35 are positive definite .
the richardson - lucy ( rl ) algorithm was developed and is widely used in the context of image reconstruction in astronomy @xcite .
however , the method has also been successfully used in cosmology , to deproject the @xmath36-d correlation function and power spectrum from the measured @xmath37-d angular correlation and @xmath37-d power spectrum @xcite .
we employ an improved rl method to solve the inverse problem for @xmath1 in eq .
( [ clsum ] ) .
the advantage of rl method is that positivity of the recovered @xmath1 is automatically ensured , given @xmath6 is positive definite and @xmath0 s are positive .
the rl method is readily derived from elementary probability theory on distributions @xcite . to make this connection
, we consider normalized quantities . ]
_ l c_l=1 ; _ k p(k)=1 ; _
l g(l , k)=1 .
this allows us to view the functions , @xmath1 and @xmath0 as one dimensional distributions , and @xmath6 as a conditional probability distribution .
further , for convenience ( not necessity ) of writing an infinitesimal measure @xmath38 we view @xmath25 to be continuous .
the integrand in the integral eq .
( [ clintg ] ) suggests defining two other probability distributions @xmath39 and @xmath40 , such that g(l , k)p(k)dldk = q(l , k)dldk = l(k , l)c_ldldk [ probeq ] dividing the both side of eq .
( [ probeq ] ) by @xmath41 we obtain l(k , l)=g(l , k ) .
[ rl2 ] the normalization conditions imply p(k)=q(l , k)dl = c_ll(k , l)dl , which in the discrete @xmath25 space reads p(k)=_lq(l , k)=_ll(k , l)c_l .
[ rl1 ] the rl method iteratively solves the eqs .
( [ rl2 ] ) and ( [ rl1 ] ) .
starting from an initial guess @xmath42 , @xmath43 is obtained using eq .
( [ rl2 ] ) as the first step .
the second step is to obtain a revised @xmath44 using eq .
( [ rl1 ] ) .
these two steps are repeated iteratively with the @xmath45 obtained after iteration @xmath46 feeding into the iteration @xmath47 . in principle , the final answer could depend on the initial guess but in practice , for a large variety of problems , rl is known leads to the correct answer even a crude estimation of the initial guess . in particular , for our problem the rl rapidly converges to the same solution @xmath1 independent of the initial guess @xmath42 .
this is demonstrated in appendix [ initguess ] .
the iterative method can be neatly encoded into a simple recurrence relation .
the power spectrum @xmath48 recovered after iteration @xmath49 is given by p^(i+1)(k)- p^(i)(k)=p^(i)(k)_lg(l , k ) [ rlstd ] where @xmath50 is the measured data ( target ) and @xmath51 is the angular power spectrum at @xmath52 iteration obtained from eq .
( [ clsum ] ) using the recovered power spectrum @xmath45 .
( [ rlstd ] ) with eq .
( [ clsum ] ) for obtaining @xmath51 from @xmath45 completely summarizes the standard rl method .
due to noise and sample variance , the data @xmath26 is measured within some non - zero error bars @xmath53 .
the standard rl method does not incorporate the error information at all .
consequently , a well known drawback of the standard rl method is that at large iterations the method starts fitting features from the noise .
modified forms of rl that address this issue have been proposed ( e.g. , see damped rl method in @xcite ) . in our problem , this problem manifests itself as very non - smooth deconvolved spectrum @xmath1 that has poor likelihood with the full wmap spectrum data .
we devise a novel method to make the rl method sensitive to the error @xmath53 by modifying eq .
( [ rlstd ] ) to p^(i+1)(k)- p^(i)(k)=p^(i)(k)_lg(l , k)^2 .
[ rlerr ] the idea is to employ a ` convergence ' function to progressively weigh down the contribution to the correction @xmath54 from a multipole bin where @xmath51 is close @xmath26 within the error bar @xmath53 .
this innovation significantly improves the wmap likelihood of the deconvolved spectrum . for certain @xmath6 ,
the improvement is so dramatic that using irl becomes very crucial to successful recovery of the spectrum ( see [ paramvar ] ) .
the final results are not sensitive to the exact functional form of the convergence function . the choice given in eq .
( [ rlerr ] ) works well but is not unique in any sense . at every iteration of the irl scheme , we compute the @xmath55 of the @xmath51 with respect to the binned data @xmath26 .
we have found that the irl iterations ( as well as the rl ) march almost monotonically toward improved ( smaller ) @xmath55 .
we halt the iterations when the @xmath55 does not change appreciably in subsequent iterations of irl .
the deconvolution algorithm produces a ` raw ' @xmath1 that has to be processed further .
the raw deconvolved spectrum has spurious oscillations and features arising largely out of the @xmath15 space sampling and binning in @xmath25 space .
these numerical artifacts are common to all recovered spectra recovered with same @xmath6 .
[ glk ] shows the plot of the binned kernel @xmath6 for the @xmath25-space bins used by wmap @xcite .
the wavenumbers are scaled by horizon scale , @xmath56 .
the sampling of @xmath15 space used here is also indicated in the figure .
we find that removing these artifacts and smoothing the resultant spectrum improves the wmap likelihood of the corresponding @xmath0 .
@xmath57 \includegraphics[scale=0.22 , angle=-90]{rks1.ps } \includegraphics[scale=0.22 , angle=-90]{diff1.ps } \includegraphics[scale=0.22 , angle=-90]{rnn1.ps } % \mbox{\bf ( a ) } & \mbox{\bf ( b ) } & \mbox{\bf ( c ) } \end{array}$ ] @xmath57 \includegraphics[scale=0.22 , angle=-90]{rks2.ps } \includegraphics[scale=0.22 , angle=-90]{diff2.ps } \includegraphics[scale=0.22 , angle=-90]{rnn2.ps } % \epsfxsize=2.4 in % \epsffile{/dsk4/cmb / people / arman / cmb / project / phase2/continue / wmap / plot4paper / rks1.ps } \\
% \mbox{\bf ( a ' ) } & \mbox{\bf ( b ' ) } & \mbox{\bf ( c ' ) } \end{array}$ ] @xmath57 \includegraphics[scale=0.22 , angle=-90]{sence.ps } \includegraphics[scale=0.22 , angle=-90]{diff3.ps } \includegraphics[scale=0.22 , angle=-90]{rnn3.ps } % \epsfxsize=2.4 in % \epsffile{/dsk4/cmb / people / arman / cmb / project / phase2/continue / wmap / plot4paper / rks1.ps } \\
% \mbox{\bf ( a ' ) } & \mbox{\bf ( b ' ) } & \mbox{\bf ( c ' ) } \end{array}$ ] we illustrate the steps of removing the numerical artefacts and smoothing with test case examples of synthetic @xmath0 from known test primordial spectrum .
the first column of fig .
[ testrec ] shows the ` raw ' spectrum obtained . a comparison with the known spectra shows similar numerical noise and artefacts in all the cases , especially , the rise at very low and very high wave numbers .
as shown in appendix [ refspec ] , the main artifacts at the low and high @xmath15 ends can be understood and modeled analytically
. however we find it easier to remove them constructing a numerically generated template reference spectrum that also takes care of small features which appear due to changes in the @xmath15 spacing . to remove these numerical artifacts , we generate a synthetic @xmath0 using a harrison - zeldovich spectrum and then apply the deconvolution method to recover a ` reference ' spectrum @xmath58 .
we divide every ` raw ' spectrum by @xmath58 to obtain a ` differenced ' spectrum that does not have numerical artifacts as seen in middle column ( `` diff . '' ) of fig [ testrec ] .
the difference spectra are a noisy version of the test spectra .
hence , the exercise with the test spectra suggests that the differenced spectrum needs to be ` suitably ' smoothed to recover the true power spectrum . a simple smoothing procedure with simple fixed width window function leads to a satisfactory recovery of the test spectrum . ( a ` bowler hat ' window constructed with hyperbolic tangent functions is used for smoothing . )
the last column in fig [ testrec ] shows the remarkably successful recovery of the test spectra in the top two case .
as expected , the @xmath1 is recovered best in the @xmath15 space ( within the vertical lines in bottom row of fig [ testrec ] ) where the kernel @xmath6 is significant ( marked by the vertical lines in fig .
[ glk ] and [ testrec ] ) . in the bottom row ,
the feature in the test spectrum is outside the range of the kernel @xmath6 , and , as expected , the recovery process misses it completely . for the top row ,
if we ignore the region where there is no power for @xmath6 the recovered spectrum is well matched with the test spectrum .
the match may be further improved by using more elaborate , adaptive smoothing procedure . for application to real data , the phrase ` suitable smoothing procedure ' may appear ambiguous .
but the smoothing is in fact very well defined for real data by demanding that the smoothed @xmath1 produces a theoretical @xmath0 that has higher likelihood given the data .
we find that this approach works extremely well .
the deconvolution algorithm uses the binned @xmath0 data , hence the theoretical @xmath0 corresponding to the recovered @xmath1 fits the binned data very well .
however , the wmap likelihood of the theoretical @xmath0 suffers owing to spurious oscillations in the differenced spectra .
the wmap likelihood improves as the differenced spectra , @xmath59 is smoothed .
we smooth the differenced spectrum so that wmap likelihood of the corresponding theoretical @xmath0 is maximized .
although it is difficult to establish that the final result is the unique solution with maximum likelihood , in practice , our simple scheme does lead to a well defined result ( no distinct degenerate solutions were found ) .
since our smoothing procedure is simple minded , possible avenues for improvement with more elaborate smoothing procedure remain open .
work is in progress to employ wavelet decomposition for the smoothing procedure .
we apply the method described in the previous section to the angular power spectrum obtained with the first year of wmap data publicly released in february 2003 @xcite to recover the primordial power spectrum . in section
[ wmapdat ] , the publicly available wmap data and how it is used in our work is discussed .
we also describe the choice of the ` base ' cosmological model .
the recovered primordial power spectrum for this model is presented in section [ primspec ] .
the next section [ paramvar ] presents the effect of varying the cosmological parameters ( within @xmath60 error bars ) on the recovered primordial power spectrum .
we also present the primordial spectrum for a set of cosmological models with large optical depths ( @xmath61 ) corresponding to possible the early reionization scenarios suggested by the wmap temperature - polarization ( te ) cross - spectrum .
accurate measurements of the angular power spectrum of cmb anisotropy was derived from the first year wmap data recently @xcite .
the spectrum obtained by averaging over @xmath62 cross - channel power measurements is essentially independent of the noise properties of individual radiometers .
the power at each multipole ranging from @xmath63 to @xmath30 was estimated together with the covariance are uncorrelated for an ideal full - sky map , but in practice , the covariances between neighboring multipole arise due to non - uniform / incomplete sky coverage , beam non - circularity , etc .. ] .
the instrumental errors are smaller than the cosmic variance up to @xmath64 , and the signal to noise per mode is above unity up to @xmath65 .
the angular power spectrum estimate and covariance matrix are publicly available at the lambda data archive .
the wmap team has also made available a suite of f90 codes that computes the likelihood for a given theoretical @xmath0 spectrum given the full angular power spectrum included the covariance measured by wmap .
we use the tt likelihood code for computing the likelihood of @xmath0 obtained from the recovered power spectrum and refer to these numbers as the ` wmap likelihood ' in the paper .
in addition , the wmap team has also obtained a binned angular power spectrum where an average @xmath0 is defined over bins in multipole space .
the binned @xmath0 estimates can be treated as independent data points since the covariance between binned estimates is negligible .
we use this binned spectrum as @xmath26 in the deconvolution of eq .
( [ clsum ] ) .
( we are aware of but do not consider the revised estimates of the low multipoles made by other authors after the wmap results @xcite ) the variance of @xmath0 measurements is given by _ l^2 = ^2 , [ cosvar ] where @xmath66 is the noise per pixel , @xmath67 the angular pixel size , @xmath68 is the fraction of sky covered and @xmath69 is the transform of the experimental beam @xcite .
it is important to note that contribution to the error from cosmic variance is proportional to the underlying theoretical / true @xmath70 spectrum . to obtain the total error bars @xmath53 used in the irl deconvolution eq .
( [ rlerr ] ) , we should add the cosmic variance for the theoretical @xmath51 to the the statistical error bars given with the binned data .
however , @xmath51 rapidly iterates to @xmath26 within the error bars in the irl method and the simpler option of using @xmath26 instead @xmath51 for computing the cosmic variance works equally well since @xmath53 in eq .
( [ rlerr ] ) simply regulates the convergence of irl for the best fit @xmath71-cdm model with a power law primordial spectrum .
our error bars are computed using @xmath26 . ] .
we consider a flat @xmath71-cdm universe , with hubble constant , @xmath72 , baryon density , @xmath73 , a cosmological constant corresponding to @xmath74 , with the remaining balance of matter to critical density in cold dark matter .
this is the ` concordance ' cosmological model suggested by the wmap parameter estimation .
while we mainly focus attention on a cosmological model without early reionization ( optical depth to reionization , @xmath75 ) we do present in section [ paramvar ] the recovered primordial spectra for models with early reionization with opacity going up to @xmath76 . the case for a large optical depth
@xmath77 comes from the temperature - polarization cross correlation . for simplicity
, we limit ourselves to the temperature anisotropy spectrum and avoid undue attention to the cosmology suggested by the yet incomplete polarization data . the polarization spectrum is expected to be announced by wmap soon .
it is then easy to extend our method to include both the temperature and polarization anisotropy spectra .
@xmath57 \includegraphics[scale=0.32 , angle=-90]{cl00mainnonsmooth.ps } \includegraphics[scale=0.32 , angle=-90]{cl00mainnonsmooth2.ps } \mbox{\bf ( a ) } \end{array}$ ] @xmath78 \includegraphics[scale=0.32 , angle=-90]{cl00mainsmooth.ps } \includegraphics[scale=0.32 , angle=-90]{cl00mainsmooth2.ps } \mbox{\bf ( b ) } \end{array}$ ]
we apply the method described in the previous section to the wmap data .
[ pk00steps ] shows the raw deconvolved spectrum , the reference spectrum , the differenced spectrum and the final recovered spectrum after smoothing obtained at each step of our method outlined in section [ meth ] .
we discuss the advantage of using improved richardson - lucy ( irl ) in appendix [ irl ] .
the effect of irl method is also evident in fig .
[ cl00steps ] .
the @xmath0 from the recovered @xmath1 matches the binned @xmath26 only within the error - bars . the @xmath0 spectrum corresponding to the differenced and smoothed @xmath1 is shown in fig . [ cl00steps ] .
although , the former has better @xmath55 for the binned data , the wmap likelihood of the latter is much better .
the poor likelihood of the differenced spectrum arises because there is no check on fluctuations in @xmath0 at intermediate multipoles between the bin centers arising from the spurious numerical effects of @xmath15-space and @xmath25-space sampling .
better wmap likelihood is more relevant than good @xmath55 for the binned data since the former incorporates the estimation of each @xmath0 and the full error covariance .
thus the smoothing step our method is carried out with a well defined goal of maximizing the wmap likelihood .
the primordial spectra recovered from wmap data is again shown in fig .
[ pk00main ] .
the dark solid line is the primordial spectrum that has the best wmap log - likelihood of @xmath79 for the base cosmological model described in the previous section .
the other lines have likelihood @xmath80 , i.e. , roughly within @xmath81 of the best one . for comparison , the same cosmological model with a scale invariant ( harrison - zeldovich ) primordial spectrum
has @xmath82 , and , with a tilted scale free primordial spectrum @xmath83 , the likelihood improves to @xmath84 .
( a comparison of likelihood numbers for various primordial spectra is given in table [ thlike ] . ) the improvement in likelihood for the recovered spectra is striking .
the most prominent feature of the recovered spectrum is the infra - red cutoff in the power spectrum remarkably close to the horizon scale ( @xmath85 ) chiefly in response to the low quadrupole measured by wmap .
another notable point is the slight tilt ( @xmath86 ) of the plateau in the recovered spectrum at large @xmath15 which is consistent with the best fit @xmath87 obtained by wmap for power law primordial spectra .
after the wmap results , model power spectra with infra - red cutoffs of the form p(k)=a_sk^1-n_s [ expcut ] invoked to explain the suppressed low multipoles of wmap have had limited success @xcite . while the effective @xmath88 improved by at least @xmath89 over a power law model , the model infra - red cutoff eq .
( [ expcut ] ) could improve the @xmath90 merely by @xmath91 @xcite .
figure [ pk00comp ] compares the recovered spectrum with these model infra - red spectra .
the power law spectrum shown in the figure also highlights the small tilt ( @xmath86 ) recovered by our method .
@xmath57 \includegraphics[scale=0.32 , angle=-90]{pk00mainh.ps } \mbox{\bf ( a ) } \includegraphics[scale=0.32 , angle=-90]{pk00mainov.ps } \mbox{\bf ( b ) } \end{array}$ ] @xmath78 \includegraphics[scale=0.32 , angle=-90]{pk00mainob.ps } \mbox{\bf ( c ) } \includegraphics[scale=0.32 , angle=-90]{pk00maindiffparam2.ps } \mbox{\bf ( d ) } \end{array}$ ] the primordial spectrum is recovered for a set of ` best fit ' cosmological parameters defining the ` base ' model .
we have varied each cosmological parameter by the quoted @xmath60 error - bars of the wmap estimates @xcite .
figure [ vps ] shows the dependence of the recovered power spectrum on small changes to cosmological parameters varied one at a time keeping the others fixed at their central value .
we find that recovered spectrum is insensitive to variations hubble constant , @xmath92 , modulo an overall shift due to the change in the horizon size , @xmath11 .
the wmap likelihood of the corresponding @xmath0 are @xmath93 and @xmath94 at @xmath95 and @xmath96 , respectively .
the variation of baryon density around its middle value @xmath97 yields almost identical primordial power spectra with a minor change in likelihood @xmath98 and @xmath99 for @xmath100 and @xmath101 , respectively .
the variation of @xmath102 ( or equivalently , @xmath103 , for flat models ) modulo a shift due to change in @xmath11 , affects only the amplitude of the infra - red cut - off .
the wmap likelihood of the corresponding @xmath0 are @xmath104 and @xmath105 for @xmath106 and @xmath107 , respectively .
we note that the likelihood at the central values are not necessarily the best .
this suggests that , as discussed in the introduction , one should explore the entire space of cosmological parameters and compute the likelihood after optimizing with ` best ' recovered primordial spectrum .
such an analysis is certainly possible , but has large demands on computational resources .
hence , we defer it to a future publication .
finally , we also consider the effect of a cosmological model with significant optical depth to reionization , @xmath108 , such as suggested by the angular power spectrum of the temperature - polarization ( @xmath109 ) cross correlation from wmap @xcite .
it has been pointed out that the estimated @xmath77 increases when the primordial spectrum has an infra - red cut - off .
we compute the primordial power spectrum for base cosmological models with @xmath110 .
the standard rl deconvolution performs poorly for these models , and , the improved rl method was crucial for these models .
[ tau ] compares the recovered primordial spectra for early reionization models which all show an infra - red cutoff at the horizon scale with different smoothing . ] .
the direct recovery of the primordial power spectrum has revealed an infra - red cutoff of a very specific form .
model spectra with monotonic infra - red cutoff such as that in eq .
( [ expcut ] ) do not improve the wmap likelihood significantly . while , to match the low value of the quadrupole , a very sharp cut - off ( such as @xmath111 , see fig [ pk00comp ] ) is required , such a steep monotonic cut - off tends to pull down the power in the next few higher multipoles above the quadrupole and octopole as well .
our recovered spectrum has a compensating excess which allows a steep cut - off to match the low quadrupole and octopole without suppressing the higher multipoles .
naively , one would think that a designer infra - red cutoff would ` cost ' in the language of bayesian evidence due to the introduction of extra parameters .
that is is not necessarily so .
an infra - red cut - off of the form we recover does not necessarily have more parameters than an infra - red cut - off of the form in eq .
( [ expcut ] ) .
moreover , it is striking that the location of the cut - off is close to a well known scale the horizon scale . in this section
we show that infra - red cut - off of the form we recover arises from very simple scenarios in inflation .
we explicitly mention two of them .
starobinsky @xcite has shown that a kink ( sharp change in the slope ) in the inflaton potential can modulate the underlying primordial power spectrum @xmath112 with a step like feature at a wavenumber @xmath113 p(k ) & = & p_o(k ) d(k , k_c , r ) + & = & a_sk^1-n_s[1 - 3(r-1)((1-)2y+2y ) + & + & ( r -1)^2 ( ( 1+)2y-2y ) ] [ staro ] where we assume a power law @xmath114 , @xmath115 and @xmath116 is the ratio of the slope @xmath117 before and after kink in the inflaton potential .
the power spectrum @xmath1 in eq .
( [ staro ] ) has a step up ( going to larger @xmath15 ) for @xmath118 and a step down feature for @xmath119 .
an infra - red cut off is created with @xmath118 .
[ thpk ] shows a spectrum with a starobinsky step ( eq .
[ staro ] ) that can not only mimic the sharp infra - red cutoff but also produces the required bump after it .
table [ thlike ] shows that a introducing an appropriate starobinsky step gives a very good wmap likelihood compared to eq .
( [ expcut ] ) . besides the location of the break , @xmath113
, the starobinsky step spectrum has only another parameter @xmath116 that fixes the slope and the depth of the cut - off as well as the size of the bump .
we have not systematically searched through the @xmath120 parameter space to arrive at a ` best - fit ' model .
hence , it may be possible to get even better match to the wmap data with starobinsky breaks .
similar scenarios have been studied earlier @xcite and has also been pointed to in the post - wmap literature @xcite .
multiple scalar field inflation provide ample scope for generating features in the primordial spectrum @xcite and has been been invoked to model a sharp cut - off at horizon scale ( see eg . ,
more exotic origin of an infra - red cut off in the scalar spectrum have also been investigated @xcite . another compelling theoretical scenario for generating
a feature of the form we have recovered is well - known .
it is well known that radiation , or matter dominated era prior to inflation does affect the primordial power spectrum on scales that ` exit the horizon ' soon after the onset of inflation . for a pre - inflationary radiation dominated epoch
the power spectrum was given by vilenkin and ford ( vf ) @xcite p(k ) = a_sk^1-n_s| e^-2 i y ( 1 + 2 i y ) -1 - 2 y^2|^2 [ vf ] where @xmath115 .
[ thpk ] shows that the vf spectrum ( eq .
[ vf ] ) can also provide an infra - red cutoff with required bump after it .
table [ thlike ] shows that a vf spectrum can give better wmap likelihood compared to model spectra of the form eq .
( [ expcut ] ) .
the infra - red cut - off ( @xmath121 ) here is not very sharp .
however , if the epoch prior to inflation is dominated by matter with some other equation of state , the slope would be different .
a more complete analysis may give rise to spectra closer to the kind we have recovered .
the scale @xmath113 is set by the hubble parameter at the onset of inflation . for this scenario to applicable to our results ,
the @xmath122-mode corresponding to the horizon scale must have crossed the hubble radius very close to the onset of inflation . in a single scalar field inflation
this would happen naturally @xcite .
another possibility is that the infra - red cutoff arises due to non - trivial topology of the universe .
a dodecahedral universe model that matches the low multipoles of wmap angular spectrum has been proposed @xcite . in future work
we would like to check whether the recovery of a discrete initial spectrum with our method appears similar to the spectrum we recovered here .
however , non - trivial cosmic topology is expected to also violate the statistical isotropy of the cmb anisotropy and give rise of correlation features which are potentially detectable @xcite .
the cmb anisotropy is usually expected to be statistically isotropic and gaussian @xcite . in that case
, the angular power spectrum of cmb anisotropy encodes all the information that may be obtained from the primary cmb anisotropy , in particular , the estimation of cosmological parameters .
it is very important to note that cosmological parameters estimated from cmb anisotropy ( and other similar observations of the perturbed universe ) usually assume a simple parametric form of spectrum of primordial perturbations . it is clear , however , that estimation of cosmological parameters depends on the extent and nature of parameterization of the primordial ( initial ) perturbations included into the parameter space considered @xcite .
we proceed on a complimentary path of determining the primordial power spectrum directly from the cmb anisotropy for a set of cosmological parameters . assuming the best fit cosmological parameters from wmap , our method applied to the angular power spectrum measured by wmap yields interesting deviations from scale invariance in the recovered primordial power spectrum .
the recovered spectrum shows an infrared cut - off that is robust to small changes in the assumed cosmological parameters .
the recovered spectrum points to the form of infra - red cut off that matches the low multipoles of wmap .
we also show that the such forms of infra - red cut - off can arise from simple well - known effects with inflation .
it is important to recall that the angular power spectrum from the ` full ' sky cmb anisotropy measurement by cobe - dmr @xcite also indicated an infra - red cutoff @xcite .
although we mostly emphasize the infrared cutoff , the final recovered spectrum shows a damped oscillatory feature after the infra - red break ( ` ringing ' ) .
it has been pointed out recently that such oscillations improve @xmath123 and could be possibly a signature of trans - planckian effects @xcite .
we have not assessed the significance and robustness of the features at large @xmath124 that may be consistent with features deduced in the analysis of recent redshift surveys @xcite .
lcc power spectrum & @xmath125 & @xmath126 + & & ( @xmath127 + direct recovered & 956.76 & 0.71 + flat harrison zeldovich & 1007.28 & + power law & 978.60 &
+ ( @xmath4 ) & & + exponential cutoff & 978.08 & 0.64 + ( @xmath4 , @xmath128 ) & & + exponential cutoff & 977.84 & 0.64 + ( @xmath4 , @xmath129 ) & + starobinsky break & 973.86 & 0.32 + ( @xmath4 , @xmath130 ) & & + vilenkin & ford ( vf - i ) & 976.88 & 0.43 + ( @xmath4 ) & & + vilenkin & ford ( vf - ii ) & 978.66 & 0.96 + ( @xmath4 ) & & + + [ thlike ] here we have considered a ` best fit ' cosmology and some finite number of variations around it .
however , with large but reasonable computational resource , it is possible to explore a multi - dimensional space of cosmological parameters to compute a likelihood at each point for an optimal ( recovered ) primordial spectrum . in this work ,
we have used only the angular power spectrum of wmap measured temperature fluctuations ( tt ) .
once the e - polarization power spectrum ( ee ) is announced by the wmap team , our method can be extended to include te and ee angular power spectra in obtaining the primordial power spectrum .
it is also possible extend our recovery of the primordial spectrum to larger wave - numbers by using the matter density power spectrum measured by large scale redshift surveys such as the sloan digital sky survey sdss and @xmath37degree field survey , measurement from ly-@xmath5 absorption , and possibly , weak gravitational lensing in the near future @xcite . in the absence of a definitive and precise scenario for the generation of primordial perturbations ,
a direct inversion of the primordial spectrum is an extremely appealing approach made possible by the remarkable quality of the recent cosmological data , in particular , the anisotropy in the cosmic microwave background temperature .
we greatly benefited from the exploratory attempt to solve the problem using a non - iterative method by v. gopisankararao as part of the visiting student programme at iucaa in the summer of 2001 .
we acknowledge very useful discussions with s. sridhar and n. sambhus regarding the rl deconvolution method . as thanks iucaa for use of its facilities during his master thesis work .
in this appendix , we provide support and justification for some steps in our method for recovering the primordial power spectrum . in section [ initguess ] we demonstrate the robustness of the iterative richardson - lucy deconvolution to changes in the initial guess .
the next section ( [ irl ] ) discusses the advantage of the improved richardson - lucy method used in our work . in the last section ( [ refspec ] ) , we model the broad features of reference spectrum analytically and
show that it is well understood .
the effect of the initial guess is negligible in the richardson - lucy method of deconvolution for our problem .
we find that the result for various different initial guesses come very close to each other after a few iterations . here
we demonstrate the robustness of our method by deconvolving using the @xmath0 from a test spectrum ( with a step ) using different initial guesses . in fig .
[ guess1 ] deconvolved raw @xmath1 arising from different forms of the initial guess spectra are shown .
the effect of the initial guess is absent in the portion of the @xmath15 space probed well by the kernel .
the deconvolved spectra are almost identical for all these different initial guesses in that region of @xmath15 space .
as discussed in section [ refspec ] , the dependence of deconvolved spectrum at the small and large @xmath15 ends is well understood and get largely removed when divided by the reference spectrum . in a previous work
we have checked a large variety of initial guesses and concluded the rl method applied to our problem of recovering the primordial spectrum from the cmb anisotropy is independent of the initial guess .
@xcite . in this section
we show the advantage of the improved richardson lucy algorithm given by eq .
( [ rlerr ] ) over the standard method given by eq .
( [ rlstd ] ) .
we compare the raw power spectrum @xmath45 and the corresponding angular spectrum @xmath51 obtained from the wmap binned angular power spectrum , @xmath26 using rl and irl algorithm as the iterations progress . in fig .
[ chisqiter ] , the left panel shows the @xmath55 of @xmath51 w.r.t @xmath26 .
as expected , the rl method leads to a lower ( better ) @xmath55 than the irl method .
however , we are interested in a recovering a @xmath1 which give @xmath0 that has better likelihood with respect to wmap data . the right panel of fig .
[ chisqiter ] plots the wmap likelihood given in terms of @xmath131 .
this figure justifies the ` improved ' label to the irl method . for the rl case ,
the value of @xmath123 is seen to bounce and increase at large iterations .
( this bounce is more pronounced and happens at a lower iteration for certain cosmological model , e.g. , one with high optical depth to reionization , @xmath77 . )
this is a reflection of the problem of the rl method fitting the noise in the data .
in contrast , in the irl case the @xmath123 converges with iterations and to significantly better ( lower ) value that for the rl case . in this subsection
, we show that the reference spectrum used to remove numerical artifacts from the raw power spectrum recovered by the deconvolution is analytically well understood .
the reference spectrum reflects the sensitivity of the kernel @xmath132 to the @xmath15 space .
( here , the over - bar in @xmath133 alludes to the binned version related to @xmath134 following eq .
( [ clsum ] ) ) . in the regions of @xmath15 space where the kernel probes @xmath1 weakly
, there is scope for changing the power spectrum without changing the @xmath0 . in such degenerate regions , the richardson lucy method ( both rl and irl ) pushes the power spectrum up to the extent possible changing the @xmath0 .
[ refspec_anal ] shows that the strong features at the low @xmath15 and high @xmath15 are well fitted by analytically obtained power law forms . at the low wavenumbers ( @xmath135 ) ,
the cmb anisotropy is dominated by the sachs - wolfe effect .
the sensitivity of the cmb anisotropy to the power at very small wavenumbers is dominated by the lowest multipole as @xmath136 .
this is the well known grishchuk - zeldovich effect @xcite .
when the quadrupole is included , @xmath137 is the strongest probe of the @xmath1 at low wavenumbers .
hence , there is no change to @xmath0 ( here , chiefly the quadrupole ) caused from this region of @xmath15 space from the initial guess @xmath42 onward if @xmath138 for @xmath139 . in our work
we have used an initial guess of the form @xmath140 .
hence , the slope of @xmath1 at low @xmath15 is driven to the @xmath141 form shown in fig .
[ refspec_anal ] . at the large @xmath15 end
, the @xmath132 is governed by the power slope of the tail in the sensitivity of the highest few multipole bins shown plotted in fig .
[ refspec_anal ] .
the slope of these tails are well fitted by a power law form @xmath142 . using the same argument
, it is clear that there will be no change to the @xmath0 caused from this region of @xmath15 space from the initial guess onward if @xmath143 . for the initial
guess @xmath144 , the rise at high @xmath15 end is driven to a @xmath145 form .
the division of reference spectrum used here is a simple numerical recipe for estimating and removing the artefacts of the deconvolution method for a given kernel .
we have shown that the gross features of the @xmath58 can be understood analytically .
the finer features of @xmath58 are likely linked to the details of the structure of @xmath132 and will be explored in future work .
a. a. starobinsky , pis ma zh .
. fiz . * 42 * , 124 ( 1985 ) [ jetp lett . * 42 * , 152 ( 1985 ) ] ; l. a. kofmann , a. d. linde and a. a. starobinsky , phys . lett . * b157 * , 361 ( 1985 ) ; l. a. kofmann and a. d. linde , nucl . phys .
* b 282 * , 555 ( 1987 ) ; j. silk and m. s. turner , phys . rev . *
d 35 * , 419 ( 1987 ) ; l. a. kofmann and d. y. pogosyan , phys . lett . *
b214 * , 508 ( 1988 ) ; j. r. bond and d. salopek , phys . rev .
* d 40 * , 1753 , ( 1989 ) ; h. m. hodges , g. r. blumenthal , l. kofman , j. primack , nucl.phys .
* b335 * , ( 1990 ) ; v.f . mukhanov and m.i .
zelnikov , phys.lett . *
b263 * , 169,1991 ; d. polarski and a. a. starobinsky , nucl . phys . * b385 * , 623 ( 1992 ) ; j. lesgourgues , d. polarski and a. a. starobinsky , mon . not .
. soc . * 308 * , 281 ( 1999 ) ; j. lesgourgues , s. prunet and d. polarski , mon . not .
. soc . * 303 * , 45 ( 1999 ) .
t. souradeep , j. r. bond , l. knox , g. efstathiou and m.s .
turner , _ proc .
cosmo-97 international workshop on ` particle physics and early universe ' sept .
15 - 19 , 1997 , ambleside , u.k .
_ , ( ed l. roszkowski , world scientific , 1998 ) , astro - ph/9802262 .
b. ratra and j. peebles , phys .
d * 52*,1837 , 1995 ; k. yamamoto , m. sasaki , t. tanaka , phys .
d * 54 * , 1996 ; a. a. starobinsky , _ proc . of cosmo-94
klopov , m. e. prokhorov , a. a. starobinsky , and j. tran thanh van , ( editions frontieres , 1996 ) ; j. r. bond , d. pogosyan , and t. souradeep , phys .
d 62 , 043005 ( 2000 ) .
a. a. starobinsky , sov .
11 * , 133 , ( 1985 ) ; r. davis et al .
lett . * 69 * , 1856 , ( 1992 ) ; t. souradeep and v. sahni , mod .
a , * 7 * , 3541 , ( 1992 ) ; r. crittenden et al . ,
* 71 * , 324 , ( 1993 ) . | the observed angular power spectrum of the cosmic microwave background temperature anisotropy , @xmath0 , is a convolution of a cosmological radiative transport kernel with an assumed primordial power spectrum of inhomogeneities .
exquisite measurements of @xmath0 over a wide range of multipoles from the wilkinson microwave anisotropy probe ( wmap ) has opened up the possibility to deconvolve the primordial power spectrum for a given set of cosmological parameters ( base model ) .
we implement an improved ( error sensitive ) richardson - lucy deconvolution algorithm on the measured angular power spectrum from wmap assuming the concordance cosmological model .
the most prominent feature of the recovered @xmath1 is a sharp , infra - red cut off on the horizon scale .
the resultant @xmath0 spectrum using the recovered spectrum has a likelihood far better than a scale invariant , or , ` best fit ' scale free spectra ( @xmath2 _ w.r.t . _
harrison zeldovich , and , @xmath3 _ w.r.t .
_ power law with @xmath4 ) .
the recovered @xmath1 has a localized excess just above the cut - off which leads to great improvement of likelihood over the simple monotonic forms of model infra - red cut - off spectra considered in the post wmap literature .
the recovered @xmath1 , in particular , the form of infra - red cut - off is robust to small changes in the cosmological parameters . we show that remarkably similar form of infra - red cutoff is known to arise in very reasonable extensions and refinements of the predictions from simple inflationary scenarios .
our method can be extended to other cosmological observations such as the measured matter power spectrum and , in particular , the much awaited polarization spectrum from wmap . |
the goal of identifying and studying the nature of ultra high - redshift galaxies remains one of the most important challenges in observational cosmology , and holds the key to furthering our understanding of the earliest stages of galaxy evolution and unveiling the nature of the sources responsible for cosmic reionization .
observational constraints provided by the gunn - peterson trough in the spectra of high - redshift quasars ( e.g fan et al .
2006 ) suggest that reionization was coming to an end at @xmath13 ( becker et al .
moreover , optical polarization measurements from the wmap experiment indicate that reionization began at @xmath14 if it is assumed to be a single , rapid event ( dunkley et al .
consequently , it is now apparent that to improve our understanding of cosmic reionization , and to unveil the earliest epoch of galaxy formation , it is necessary to extend studies of high - redshift galaxies into the redshift @xmath15 rgime ( e.g. robertson et al . 2010 ) . given our existing knowledge of the evolution of the galaxy luminosity function in the redshift interval @xmath16 ( e.g. bouwens et al .
2007 ; mclure et al . 2009 ) it is clear that achieving this aim requires ultra - deep near - ir imaging , reaching detection limits of @xmath17 . at the bright end of this range , wide - field
, ground - based imaging has a unique contribution to make , and has recently allowed the luminosity function , the clustering properties and the stellar populations of luminous lyman - break galaxies ( lbgs : mclure et al .
2009 ; ouchi et al . 2009 ; grazian et al .
2010 ) and lyman alpha emitters ( laes : ouchi et al . 2010 ; ota et al . 2010 ; ono et al . 2010 ; nakamura et al .
2010 ) to be studied in detail .
indeed , the importance of ground - based imaging and spectroscopy has recently been highlighted by the spectroscopic confirmation of two lbgs at @xmath18 and @xmath19 by vanzella et al .
however , in advance of 30-m class ground - based telescopes , it is clear that routinely identifying and studying sub - l@xmath20 galaxies at @xmath21 is only possible using space - based imaging .
consequently , the unparalleled near - ir sensitivity provided by the new wfc3 camera , installed on the hubble space telescope ( hst ) in late 2009 , has proven to be a crucial breakthrough in high - redshift galaxy studies . indeed , despite only covering an area of @xmath22 sq .
arcmin , the unprecedented depth ( @xmath23 , @xmath24 ) of the first tranche of wfc3/ir imaging of the hubble ultra deep field ( hudf ; go-11563 ) led to a raft of early science papers investigating the number densities , luminosity functions , stellar masses and stellar populations of @xmath25 galaxies ( e.g. bouwens et al .
2010ab ; oesch et al .
2010 ; mclure et al .
2010 ; yan et al . 2010 ; labb et al .
2010 ; finkelstein et al . 2010 ; bunker et al .
interestingly , lehnert et al . (
2010 ) have recently claimed the tentative detection of ly@xmath26 emission at @xmath27 in a wfc3/ir candidate in the hudf , originally identified by bouwens et al .
( 2010a ) and mclure et al .
although the large rest - frame equivalent width ( ew@xmath28 ) of the ly@xmath26 emission line suggests that , if confirmed , this object must be a decidedly atypical example of a @xmath0 lbg ( stark , ellis & ouchi 2010 ) , the location of the claimed ly@xmath26 emission line is in good agreement with the original photometric redshifts derived by mclure et al .
( 2010 ) and finkelstein et al .
( 2010 ) ; @xmath29 and @xmath30 respectively .
in addition to the wfc3/ir imaging of the hudf , another key dataset has been the wfc3/ir imaging taken as part of the early release science extra - galactic programme ( ers ; go-11359 ) which , although substantially shallower than the hudf wfc3/ir imaging ( @xmath31 , @xmath24 ) , covers an area approximately ten times larger . by combining the hudf and ers datasets to obtain greater dynamic range in uv luminosity ,
bouwens et al .
( 2010b ) and labb et al . (
2010 ) investigated the relationship between the uv spectral slope ( @xmath32 ) and uv luminosity .
both studies find a correlation , with @xmath33 changing from @xmath34 ( a typical value for lower - redshift starburst galaxies ) at @xmath35 , to extremely blue values of @xmath36 at @xmath37 . as discussed by bouwens et al .
( 2010b ) , although dust - free , low metallicity models can produce slopes of @xmath38 , they can only do so under the assumption that the ionising photon escape fraction is high ( @xmath39 ) and that , correspondingly , the contribution from nebular continuum emission is low .
based on stacking the acs+wfc3/ir+irac photometry of @xmath40 lbg candidates in the hudf and ers datasets , labb et al ( 2010 ) find the same correlation between @xmath41 and spectral slope @xmath33 as bouwens et al .
( 2010b ) . however , labb et al . (
2010 ) conclude that it is not possible to reproduce both the blue spectral slopes and significant @xmath42 spectral breaks displayed by the faintest @xmath40 lbg candidates , without recourse to episodic starformation histories and/or a significant contribution from nebular line emission . indeed , ono et al .
( 2010 ) also conclude that nebular line emission may be necessary to reproduce the observed @xmath43m@xmath44 colour in a stack of @xmath13 lae photometry . despite the uncertainties , one observational result that has received significant attention recently is the apparent relationship between star - formation rate and stellar mass . both labb et al .
( 2010 ) and gonzlez et al .
( 2010 ) find an approximately linear correlation between stellar mass and star - formation rate at @xmath40 , consistent with the results derived by stark et al .
( 2009 ) for lbgs in the redshift range @xmath45 . as a result , labb et al ( 2010 ) and gonzlez et al .
( 2010 ) conclude that the specific star - formation rate ( ssfr ) of @xmath40 lbgs is remarkably constant ( ssfr @xmath46 gyr@xmath12 ) , and consistent with the value of ssfr @xmath47 gyr@xmath12 observed in star - forming galaxies at @xmath48 ( e.g. daddi et al .
2007 ; magdis et al . 2011 ) . as previously discussed by stark et al .
( 2009 ) , a natural explanation of this observation would be to invoke a star - formation rate which exponentially increases with time although , as shown by finlator , oppenheimer & dav ( 2011 ) , star - formation histories of this type may have difficulty in reproducing some of the most extreme balmer breaks reported in the literature at @xmath0 .
the majority of previous studies which have investigated the high - redshift galaxy population using wfc3/ir imaging have relied on traditional colour - cut , or `` drop - out '' , selection techniques .
in contrast , the principal motivation for this paper is to investigate what can be learned about the @xmath4 galaxy population by fully exploiting the excellent multi - wavelength ( acs+wfc3/ir+irac ) data which is now available over an area of @xmath49 sq .
arcmin . rather than applying standard `` drop - out '' criteria , in this work we continue to pursue the strategy we have previously adopted ( mclure et al . 2006 ; 2009 ; 2010 ) and employ a template - fitting , photometric redshift analysis to select our final high - redshift galaxy sample . a key new element in this strategy is our development of a deconfusion algorithm capable of providing the robust irac photometry necessary for improved photometric redshift and stellar - mass estimates . in principle
this technique should have several advantages over the standard lbg `` drop - out '' selection .
firstly , by employing all of the available multi - wavelength data , including the irac photometry , it is possible to make optimal use of the available information . secondly , by avoiding any colour pre - selection this approach should be less biased towards simply selecting the very bluest galaxies at high - redshift .
this second point is potentially crucial within the context of investigating the claims of ultra - blue uv spectral slopes for lbg candidates at @xmath4 ( see section 4 ) .
finally , an sed fitting analysis also provides an estimate of the photometric redshift probability density function , @xmath50 , and therefore allows the prevalence , and significance , of competing photometric redshift solutions at low redshift to be transparently investigated .
the primary motivation of this paper is therefore to construct the most robust sample possible using the available data and techniques reviewed above , in order to critically address some of the claims about the properties of the @xmath51 population newly - found with hst .
the structure of the paper is as follows .
in section 2 we describe the available data in each of the three fields , including a brief description of our irac deconfusion algorithm . in section 3
we describe our initial candidate selection , photometric redshift analysis and the construction of the final catalogue of robust candidates . in sections 4 and 5
we investigate the uv spectral slopes , stellar masses and star - formation rates of the final robust sample . in section 6
we perform a detailed comparison of our final robust sample with samples previously derived in the literature , exploring the reasons behind any apparent discrepancies . in section 7
we provide a summary of our main conclusions . in the appendix we provide a full description of our irac deconfusion procedure , full photometry and grey - scale postage stamp images for each high - redshift candidate and individual plots illustrating the results of our sed fitting .
all magnitudes are quoted in the ab system ( oke & gunn 1983 ) and all calculations assume @xmath52 and @xmath53 kms@xmath12mpc@xmath12 .
the analysis in this paper relies on the publicly available optical , near - ir and mid - ir imaging data covering the hudf , hudf09 - 2 and ers fields . in this section
we briefly describe the relevant details of the various imaging datasets , and the depth analysis which was performed in order to attribute accurate error estimates to the candidate photometry .
the basic properties of the three fields are listed in table 1 .
the wfc3/ir imaging of both the hudf and the hudf09 - 2 fields was taken as part of the public treasury programme go-11563 ( p.i.=illingworth ) and consists of single pointings ( @xmath22 sq .
arcmin ) of the wfc3/ir camera in the f105w , f125w and f160w filters ( hereafter referred to as @xmath54 and @xmath55 ) .
the wfc3/ir dataset in the ers field was taken as part of the public programme go-11359 ( p.i.=oconnell ) and consists of a mosaic of 10 pointings of the wfc3/ir camera in the f098 m ( @xmath56 ) , @xmath57 and @xmath55 filters ( windhorst et al .
2011 ) .. ] the wfc3/ir data were calibrated using calwf3 and subsequently combined using multidrizzle ( koekemoer et al . 2002 ) as summarized in mclure et al .
2010 ) ; full details are presented in koekemoer et al .
the final mosaics have psfs with fwhm in the range @xmath58@xmath59depending on the filter , and were drizzled onto a final grid of 0.06/pix . in the case of the hudf and ers mosaics ,
the final astrometry was matched to that of the publicly available reductions of the optical acs imaging of the udf ( beckwith et al .
2006 ) and goods - s ( goodsv2.0 ; giavalisco et al .
2004 ) respectively .
the astrometry for the final mosaics of the hudf09 - 2 field was matched to the @xmath43band imaging of goods - s taken as part of the musyc survey ( cardamone et al .
2010 ) , with a typical r.m.s .
accuracy of @xmath60 .
the wfc3/ir imaging of the ers analysed in this paper consists of the data comprising the completed programme .
however , for the hudf09 - 2 and hudf fields we make use of the epoch 1 observations , which consist of the data publicly available as of february 2010 and august 2010 respectively . for the hudf and ers
fields the acs data used in this study consists of the publicly available reductions of the f435w , f606w , f775w and f850lp ( hereafter @xmath61 , @xmath62 and @xmath63 ) imaging of the hudf ( beckwith et al .
2006 ) and goods - s ( v2.0 ; giavalisco et al .
the acs imaging covering the hudf09 - 2 field is our own reduction ( based on calacs and multidrizzle ) of the @xmath64 and @xmath63 imaging obtained as part of the udf05 programme ( oesch et al .
all of the optical acs imaging was re - sampled to a 0.06/pix grid to match the wfc3/ir data , and the astrometry of the acs data covering hudf09 - 2 was also registered to match the musyc imaging of goods - s . for the hudf and ers fields we make use of the publicly available reductions ( v0.30 ) of the @xmath7 m and @xmath65 m irac imaging obtained as part of the goods survey ( proposal i d 194 , dickinson et al . ; in preparation ) .
the irac data covering the ers consists of approximately 23 hours of on - source integration in both the @xmath7 m and @xmath65 m bands . for the hudf we performed an inverse variance weighted stack of the overlapping region of the epoch 1 and epoch 2 imaging , producing final mosaics consisting of approximately 46 hours of on - source integration at @xmath7 m and @xmath65 m . for the hudf09 - 2 we re - registered and stacked the mopex reductions of the @xmath7 m imaging obtained via proposal i d 30866 ( p.i .
r. bouwens ) producing a final @xmath7 m mosaic with an on - source integration time of approximately 33 hours .
the consistency of the irac photometry was checked via reference to the simple ( damen et al .
2010 ) imaging at @xmath7 m and @xmath65 m of the e - cdfs , which overlaps all three fields .
the astrometry of the irac imaging in all three fields was registered to that of the corresponding @xmath55 wfc3/ir imaging . a key new feature of the analysis undertaken in this paper is the inclusion of the irac photometry in the candidate high - redshift galaxy selection procedure . due to its depth , and
comparatively broad psf ( fwhm @xmath66 ) , the @xmath67 m irac imaging covering the three fields of interest is heavily confused , making the process of obtaining aperture photometry matched to the optical / nearir hst imaging non trivial . in order to achieve this aim
it is therefore necessary to pursue some form of deconfusion process , which allows the irac imaging to be utilised beyond the natural confusion limit .
although there are several techniques which can be used to deconfuse irac imaging ( see appendix a ) we have developed our own software which uses the wfc3/ir imaging data to provide normalized templates for each object in the field and then , via a transfer function , produces synthetic irac images on the native 0.6/pix plate scale . through a matrix inversion procedure , the amplitude ( or total flux ) of each template
can be simultaneously fitted to produce the optimal reproduction of the observed irac image ( see fig .
1 ) . as a result of this procedure
it is effectively possible to extract accurate aperture photometry from the @xmath7 m and @xmath65 m irac imaging at the spatial resolution of the wfc3/ir imaging .
an additional advantage of this approach is that it naturally provides robust uncertainties on the delivered flux measurements , which depend both on the signal - to - noise ratio of the irac detection _ and _ the local level of confusion in the irac image . | we present the results of a photometric redshift analysis designed to identify redshift @xmath0 galaxies from the near - ir hst imaging in three deep fields ( hudf , hudf09 - 2 & ers ) covering a total area of 45 sq . arcmin . by adopting a rigorous set of criteria for rejecting low - redshift interlopers , and by employing a deconfusion technique to allow the available ultra - deep irac imaging to be included in the candidate selection process ,
we have derived a robust sample of 70 lyman - break galaxies ( lbgs ) spanning the redshift range @xmath1 . based on our final sample
we investigate the distribution of uv spectral slopes ( @xmath2 ) , finding a variance - weighted mean value of @xmath3 which , contrary to some previous results , is not significantly bluer than displayed by lower - redshift starburst galaxies .
we confirm the correlation between uv luminosity and stellar mass reported elsewhere , but based on fitting galaxy templates featuring a range of star - formation histories , metallicities and reddening we find that , at @xmath4 , the range in mass - to - light ratio ( @xmath5 ) at a given uv luminosity could span a factor of @xmath6 . focusing on a sub - sample of twenty - one candidates with irac detections at @xmath7 m
we find that @xmath8 lbgs at @xmath9 have a median stellar mass of @xmath10 ( chabrier imf ) and a median specific star - formation rate ( ssfr ) of @xmath11 gyr@xmath12 . using the same sub - sample we have investigated the influence of nebular continuum and line emission , finding that for the majority of candidates ( 16 out of 21 ) the best - fitting stellar masses are reduced by less than a factor of 2.5 .
however , galaxy template fits exploring a plausible range of star - formation histories and metallicities provide no compelling evidence of a clear connection between star - formation rate and stellar mass at these redshifts .
finally , a detailed comparison of our final sample with the results of previous studies suggests that , at faint magnitudes , several high - redshift galaxy samples in the literature are significantly contaminated by low - redshift interlopers . #
1to 0pt#1 # 1 # 1 to size # 1 to size = cmr10 scaled2 = cmbx10 scaled2 = cmti10 scaled2 ps . 2 = 6truein galaxies : high - redshift - galaxies : evolution - galaxies : formation |
the multiferroic materials , which simultaneously exhibit a long - range magnetic order and a spontaneous electric polarization , have attracted a great deal of attention due to their potential applications in the next - generation electronic devises as well as the fundamental interest in the origin of magnetoelectric ( me ) coupling.@xcite among them , there is a very important subclass of materials , which are called `` improper multiferroics '' , where the magnetic order does not simply coexist with the spontaneous polarization , but breaks the inversion symmetry by itself and , thus , becomes primarily responsible for the ferroelectric ( fe ) activity . because of this intrinsic interconnection between polarization and magnetism ,
such materials are expected to exhibit a strong me coupling .
there is a large number of theoretical studies , which introduce and emphasize the importance of different mechanisms of the me coupling , associated with the spin current;@xcite nonrelativistic magnetostriction;@xcite inverse dzyalishinskii - moriya ( dm ) mechanism , which is another type of magnetostriction , caused by the relativistic spin - orbit ( so ) coupling;@xcite and spin - dependent @xmath6-@xmath7 hybridization.@xcite currently , most of these theories have a phenomenological status , as each of them is typically oriented on the description of properties , observed in some narrow group of materials . presumably , the most striking example is the properties of orthorhombic manganites , which are typically interpreted from two completely different standpoints : the fe activity in the twofold periodic systems ( such as homno@xmath3 and ymno@xmath3 ) is ascribed solely to the nonrelativistic magnetostriction mechanism , whereas in materials with longer magnetic periodicity ( such as tbmno@xmath3 ) it is believed to have a purely relativistic origin , associated with either the spin current or the inverse dm mechanism . in this work
we continue to develop the double exchange ( de ) theory of the me coupling,@xcite which is oriented on the wide class of multiferroic manganites with different types of the crystallographic and magnetic structure .
manganites play one of the key roles in the materials science engineering of novel multiferroic compounds .
there are two reasons for it : ( i ) the orbital ordering , which , in the combination with other factors , assists the antisymmetric charge transfer and the formation of the spontaneous polarization;@xcite ( ii ) the high spin state of the mn@xmath8 ions , which is driven by intraatomic hund s coupling and plays a crucial role in the magnetic inversion symmetry breaking : since the nonmagnetic state of the mn@xmath8 sites would lead to a gigantic loss of the intraatomic hund s energy , in certain magnetic structures with competing magnetic interactions , it is more favorable energetically to keep these sites magnetic , but to abandon the inversion symmetry.@xcite the basic electronic and magnetic properties of manganites are described by the de model.@xcite although the concept of the double exchange was originally introduced for the analysis of the metallic state , realized in hole - doped manganites , today it is understood much more generally as a generic property of high spin compounds with the partially filled majority - spin states . from a mathematical point of view , the ` high spin state ' means that the intraatomic exchange splitting between the majority- and minority - spin states is so large that the effect of the latter states on the considered properties can be neglected . the reorientation of spins in the de model may lead to a dramatic change of the electronic structure and even open the band gap.@xcite therefore , in order to understand the behavior of the fe polarization in the de model , it is very important to link it to the change of the electronic structure .
this can be achieved by using the berry phase theory of polarization , which can be reformulated in terms of the occupied wannier functions in the real space.@xcite for compounds with the centrosymmetric crystal structure , such procedure naturally gives us the electronic polarization , induced by a noncentrosymmetric magnetic order .
hopefully , the electronic structure of insulating manganites is characterized by another large parameter @xmath9 , which is the splitting between the occupied and unoccupied states with the same spin .
it is caused by the jahn - teller distortion and additionally enhanced by the screened on - site coulomb repulsion .
this enables us to use the perturbation theory in order to evaluate the asymmetric spin - dependent change of the wannier functions in the first order of @xmath10.@xcite this change will automatically gives the spin dependence of the electronic polarization .
needless to say that this mechanism of the me coupling is essentially nonrelativistic one : the relativistic spin - orbit interaction can play an important role in stabilizing noncentrosymmetric magnetic structures .
however , once it is known , the fe polarization can be described by the nonrelativistic de theory .
therefore , depending on the type of the magnetic structure , this mechanism can produce a large fe polarization . in any model analysis , in order to describe the properties of realistic materials and to elucidate the differences between these materials
, it is very important to make a link to the first - principles calculations . in our case
all parameters of the de model are derived from the first - principles electronic structure calculations , by constructing some effective hamiltonians for the magnetically active states in the basis of wannier functions.@xcite such a procedure typically gives us very reliable description of multiferroic and other properties of transition - metal oxides at a semi - quantitative level.@xcite in our previous work ( ref . ) , we have applied such de model for the analysis of ferroelectric ( fe ) activity in one particular type of manganites , crystallizing in the orthorhombic structure and forming twofold periodic magnetic structure in the ground state .
we have argued that by using the de model one can indeed successfully rationalize many aspects of the fe activity in this type of systems . moreover , even at a quantitative level , it reproduces results of more general mean - field hartree - fock calculations for the effective model , which were used as the starting point for the construction of the de model .
these results are also in a good agreement with available first - principles electronic structure calculations.@xcite in this work we will systematically apply the de model for the analysis of the wide class of multiferroic manganites .
we will show that the double exchange is indeed the key microscopic mechanism , which explains the basic aspects of the fe activity , related to the interplay between magnetic and crystallographic symmetries in various types of manganites .
the rest of the article is organized as follows . in sec .
[ sec : method ] , we will introduce the de model of the me coupling in multiferroic manganites . in sec .
[ sec : results ] , we will discuss applications of this model for different types of manganites .
particularly , we will consider the orthorhombic systems ( the space group @xmath11 ) , forming twofold and fourfold periodic magnetic structures in the ground state ( sec .
[ subsec : ortho ] ) , the monoclinic @xmath12 phase of bimno@xmath3 ( sec .
[ subsec : bimno3 ] ) , and hexagonal manganites , crystallizing in the @xmath13 structure , using ymno@xmath3 as an example ( sec .
[ subsec : hexo ] ) . finally , in sec .
[ sec : conc ] , we will present a summary of our work .
our strategy consists of the following steps : \(i ) we assume that fe and magnetic properties in the ground state of considered systems can be described reasonably well at the level of the mean - field theories
. it can be the kohn - sham density functional theory or its refinements,@xcite which are widely used in the first - principles electronic structure calculations . in our case
, we focus on the behavior of magnetically active @xmath0 bands of manganites and replace the first - principles calculations for this part by the solution of the realistic hubbard - type model , which was rigorously constructed using the technique of wannier functions.@xcite more specifically , we start with the electronic band structure in the local - density approximation ( lda ) , construct the wannier functions for the @xmath0 bands , and calculate the matrix elements of the lda hamiltonian in the basis of these wannier functions . such a construction gives us the proper one - electron part of the model hamiltonian .
then , we calculate the screened on - site coulomb interactions for the @xmath0 bands , using the simplified version of the constrained random - phase approximation,@xcite as explained in ref . .
after that , we solve the effective hubbard - type model in the mean - field hartree - fock ( hf ) approximation .
the solution gives us the mean - field hamiltonian of the form : @xmath14 here , @xmath15 $ ] is the proper one - electron part of the hubbard model between sites @xmath16 and @xmath17 in the basis of wannier orbitals @xmath18 (
@xmath19 ) @xmath20 @xmath21 , @xmath22 , @xmath23@xmath24@xmath25 , @xmath26 , and @xmath27@xmath24@xmath28 , and @xmath29 is the self - consistent hf potential at the site @xmath16 , which is constructed from the screened coulomb interactions and the density matrix.@xcite the wannier functions themselves were constructed using the projector - operator technique ( refs . and ) and the orthonormal linear muffin - tin orbitals ( lmto ) ( ref . ) as the trial wave functions .
as the lmto basis is already well localized , such procedure allows us to generate well localized wannier functions .
therefore , the obtained transfer integrals ( @xmath30 for @xmath31 ) are typically restricted by the nearest neighbors , while other contributions are substantially smaller .
since the lda band structure is nonmagnetic and we do not consider explicitly the so coupling , the matrix @xmath32 does not depend on the spin indices ( @xmath33 and @xmath34@xmath20 @xmath35 or @xmath36 ) . without so interaction ,
@xmath29 is diagonal with respect to @xmath33 and @xmath34:@xmath37 where each @xmath38 is the @xmath39@xmath40@xmath39 matrix in the orbital subspace .
\(ii ) we assume that , to a good approximation , @xmath41 can be replaced by @xmath42 , where @xmath43 is the averaged exchange splitting between the majority- and minority - spin states . due to hund s interactions ,
the splitting @xmath43 is large in manganites .
therefore , the details of @xmath41 in the unoccupied part of the spectrum becomes relatively unimportant on the energy scale of @xmath43 .
moreover , one can consider the limit @xmath44 , and replace @xmath45 by the de hamiltonian:@xcite @xmath46 which operates in the subspace of @xmath35-spin states , in the local coordinate frame , specified by the directions of spins @xmath47 .
the prefactor @xmath48 in eq .
( [ eqn : de ] ) is given by the well known expression:@xcite @xmath49 \(iii ) the next step is the calculation of the electronic polarization using the berry - phase method.@xcite for our purpose , it is convenient to use the real - space formulation of this method , in terms of the occupied wannier functions @xmath50 .
then , the electronic polarization is given by : @xmath51 where @xmath24@xmath52 ( @xmath53 ) is the electron charge , @xmath54 is the volume of _ magnetic _ unit cell ( with @xmath55 being the volume of the crystallographic cell and @xmath56 being the number of such cells ) , and the summation @xmath57 runs over the occupied bands .
alternatively , one can sum up unoccupied bands .
this should give us @xmath58 .
it is important to note that we treat @xmath1 as a nonrelativistic quantity , which does not explicitly depend on the so coupling .
nevertheless , the latter can still define the directions @xmath59 of spins in the ground state .
the spin dependence of @xmath1 in eq .
( [ eqn : pw ] ) is accumulated in @xmath60 .
then , for each @xmath57 , one can arbitrarily shift the origin of integration in eq .
( [ eqn : pw ] ) .
since each @xmath60 is normalized , the shifted integral and the original one will differ by some vector , which depends on the shift , but does not depend on the spin variables .
therefore , since we are interested only in the spin dependence of @xmath1 , we shift the origin of each integral in eq .
( [ eqn : pw ] ) to the position of that atomic site for which @xmath50 was constructed and drop all concomitant terms , which do not depend on the spin degrees of freedom .
\(iv ) our next observation is that many manganites exhibit the jahn - teller distortion , which splits the occupied and unoccupied @xmath61 orbitals .
this splitting is additionally enhanced by the on - site coulomb repulsion .
thus , there is another large parameter @xmath9 , which characterizes the electronic structure of manganites .
we will define it as the intraatomic energy splitting between the center of gravity of occupied manifold , consisting of three @xmath62 and one @xmath61 levels , and the unoccupied @xmath61 level ( see fig .
[ fig : idea ] ) .
( color online ) schematic view , explaining the de model for fe polarization in manganites : due to specific magnetic alignment , the spins can form different angles in the bonds lying on the right ( @xmath63 ) and on the left ( @xmath64 ) relative to some center site . in the de model ,
this leads to different scaling of the transfer integrals operating in these two bonds , which is described by the factors @xmath65 and @xmath66 , respectively .
even if the central site is located in the inversion center , this magnetic alignment breaks the inversion symmetry . using the berry phase theory,@xcite
the polarization can be related to the spin - dependent asymmetric transfer of the weights of the occupied wannier functions to the neighboring sites , which are given by transfer integrals between occupied and unoccupied orbitals ( shown by dashed arrows ) . since the unoccupied @xmath61 orbital splits off from the occupied ones by large crystal field and the coulomb repulsion ( the corresponding splitting is denoted by @xmath9 )
, these transfer integrals can be treated as a perturbation in the leading order of @xmath10.,height=226 ] the large @xmath9 allows us to use the perturbation theory for the occupied wannier functions @xmath50 and evaluate their change in the first order of @xmath10 .
practically , @xmath9 is obtained form the diagonalization of the site - diagonal part @xmath67 of the de hamiltonian ( [ eqn : de ] ) , which specifies the so - called crystal - field representation ( if there are two types of mn sites , like in bimno@xmath3 , we use the averaged value of @xmath9 ) .
note that the occupied states in manganites are typically not well separated from each other and form one broad band . from this point of view , when we consider the energy splitting , it is more reasonable to use only one energy for all occupied states and take it in the center of gravity of these states .
the transfer integrals can be also transformed to the crystal - field representation : @xmath68 .
then , one can start with the atomic limit , where all @xmath50 are fully localized on their atomic sites , and consider the transfer of weight of @xmath50 to the neighboring sites in the first order of @xmath69 . since the transfers within occupied states correspond to some unitary transformation of @xmath50 , they will not change the physical properties .
thus , it is sufficient to consider only the transfer integrals , connecting the occupied and unoccupied orbitals in the crystal - field representation . as was pointed out above , alternatively
, one can consider the change of the unoccupied states and the transfer integrals from the unoccupied to occupied orbitals , which is more convenient for our purposes .
since , at each mn site , there is only one unoccupied orbital , one can drop the index @xmath57 in the notations of @xmath50 and replace it by the site index @xmath16 .
furthermore , we adopt the lattice model and assume that all weights of @xmath70 are localized in the lattice points : i.e. , if @xmath71 is centered at the site @xmath16 , its weight can be presented in the form @xmath72 where @xmath73 is the position of the site @xmath17 relative to the site @xmath16 .
then , in the atomic limit , the weight is accumulated at the central site ( @xmath74 and @xmath75 for @xmath76 and @xmath77 , respectively ) .
then , the weights at the neighboring sites can be obtained in the first order of the perturbation theory for @xmath71 as @xmath78 where the summation runs over four occupied orbitals . since @xmath79 , the weights @xmath80 and @xmath81 can be obtained from the same matrix of transfer integrals .
\(v ) finally , in the de model , the transfer integrals @xmath82 should be additionally modulated by @xmath48 , which depends on the relative orientation of spins at the sites @xmath16 and @xmath17 .
thus , even though the crystal structure itself is centrosymmetric and the mn sites are located in the inversion centers ( like in orthorhombic manganites , crystallizing in the @xmath11 structure ) , the multipliers @xmath48 can make some bonds , connecting the central mn site with its neighbors , inequivalent .
for example , if the right bond @xmath83 in fig .
[ fig : idea ] is transformed to the left bond @xmath84 by the inversion operation , the spin alignment yielding @xmath85 will make these bonds inequivalent .
this will break the inversion symmetry and produce some ` dipole ' , associated with the site @xmath16 , which will contribute to the electronic polarization as @xmath86 ( note that the sign was changed because here we consider the change of the _ unoccupied _ band and the electron transfer from the _ unoccupied _ @xmath61 orbital @xmath39 to the occupied orbitals @xmath87-@xmath88 ) . in order to obtain the total polarization , one should sum up all inequivalent dipoles , induced by the magnetic symmetry breaking . moreover , it is straightforward to show that @xmath89 therefore , the spin dependence of @xmath90 in the de model is given by the isotropic correlation functions , @xmath2 , between directions of spins .
this behavior should also specify the temperature dependence of the polarization , associated with the spin disorder .
the spin dependence of @xmath1 has the same form as for the phenomenological magnetostriction mechanism,@xcite which is frequently used for the analysis of magnetoelectric ( me ) coupling in manganites.@xcite nevertheless , the new point of our analysis is that this dependence is natural result of the de physics and is not necessary related to the magnetically driven fe displacements .
formally speaking , the proposed de mechanism can take place even in a centrosymmetric crystal structure without any magnetostriction , although these two effects can coexist : once the inversion symmetry is broken by the magnetic order , there will be magnetostrictive forces , which will move the atoms away from their centrosymmetric positions .
this will activate the magnetostriction mechanism . in terms of the modern berry phase theory of polarization,@xcite
one can say that the de and magnetostriction mechanism give rise to , respectively , electronic and ionic parts of the polarization .
according to the first - principles calculations , the electronic contribution in manganites is at least equally important as the ionic one and can not be neglected.@xcite by summarizing this section : in the de model , the inversion symmetry is broken by noncentrosymmetric modulation of the transfer integrals by @xmath48 .
however , this modulation will also interplay with the crystallographic symmetry of manganites , which is reflected in the behavior of @xmath91 and @xmath82 .
in the next section , we will illustrate how this interplay will work for different types of manganites , giving rise to finite me coupling .
the connection between fe polarization and magnetic structure of orthorhombic ( @xmath11 ) manganites was discussed in our previous article ( ref . ) , also on the level of the de model . in this section
we present a more general and more comprehensive analysis of the problem .
we assume that the manganites form the perfect antiferromagnetic ( afm ) order along the orthorhombic @xmath5 axis , that is indeed consistent with the experimental data.@xcite then , in the de model , all @xmath92 planes become effectively decoupled , and it is sufficient to consider the single plane .
moreover , it is assumed that the magnetic structure is periodic along the @xmath93 axis and may have arbitrary periodicity along the @xmath4 axis ( see fig .
[ fig : orthoschem ] ) , being again in total agreement with the experimental situation.@xcite ( color online ) schematic view on the crystal and magnetic structures in the @xmath92 plane of orthorhombic manganites .
mn atoms are indicated by circles .
two types of bonds , which are transformed to themselves by the symmetry operation @xmath94@xmath95@xmath96 , are shown by solid and broken lines .
the magnetic structure is periodic along the orthorhombic @xmath93 axis and may have arbitrary periodicity along the @xmath4 axis.,height=226 ] the most known examples are the twofold periodic e - type afm structure , realized in homno@xmath3 and ymno@xmath3,@xcite and the nearly fourfold periodic `` spiral '' magnetic structure , realized in tbmno@xmath3.@xcite however , it was also suggested that both types of magnetic structures are deformed by relativistic interactions and this deformation has a profound effect on the value of the fe polarization.@xcite the crystal structure in the @xmath92 plane can be generated by two symmetry operation : @xmath94@xmath95@xmath96 ( the @xmath97 rotation around the orthorhombic @xmath93 axis , combined with the translation by @xmath98@xmath95@xmath99 ) and the inversion @xmath100 .
moreover , the mn sites are located in the inversion centers . then , we take an arbitrary mn site ( ` 0 ' in fig . [
fig : orthoschem ] ) and evaluate its contribution to the fe polarization , using eq .
( [ eqn : ppt ] ) . in this case
, we should sum up the contributions of four bonds : @xmath75-@xmath87 , @xmath75-@xmath101 , @xmath75-@xmath102 , and @xmath75-@xmath103 , which correspond to @xmath104 @xmath98@xmath95@xmath99 , @xmath24@xmath98@xmath95@xmath99 , @xmath24@xmath98@xmath24@xmath99 , and @xmath98@xmath24@xmath99 , respectively .
the inversion @xmath100 transforms the bonds @xmath75-@xmath102 and @xmath75-@xmath103 to the bonds @xmath75-@xmath87 and @xmath75-@xmath101 , respectively , and the symmetry operation @xmath94@xmath95@xmath96 transforms the bond @xmath75-@xmath101 to the bond @xmath87-@xmath75 .
hence , we have @xmath105 , @xmath106 , and @xmath107 .
moreover , the periodicity of the magnetic structure along @xmath93 imposes the following constraints : @xmath108 and @xmath109 , where the notations @xmath110 and @xmath111 stand for the bonds spreading in the positive and negative directions of the @xmath4 axis , starting from the site @xmath75 .
then , the vector @xmath112 can be presented in the form : @xmath113 , where @xmath114 is the fe polarization of the e - type afm phase ( @xmath115),@xcite @xmath116 since all mn sites are located strictly in the @xmath92 plane and , in the de model , all transfer integrals to the neighboring planes are suppressed by the afm order , there will be no polarization parallel to the orthorhombic @xmath5 axis .
then , we can repeat this procedure and evaluate the contributions associated with the neighboring sites @xmath87 and @xmath103 in the magnetic cell ( see fig . [
fig : orthoschem ] ) .
clearly , half of these contributions will involve the same bonds @xmath75-@xmath87 , @xmath75-@xmath101 , @xmath75-@xmath102 , and @xmath75-@xmath103 .
moreover , since the sites @xmath87 and @xmath103 are obtained from the site @xmath75 by the symmetry operation @xmath94@xmath95@xmath96 , in the corresponding expressions for @xmath117 and @xmath118 we will have to interchange @xmath119 and @xmath120 .
furthermore , we note that @xmath121 and @xmath122
. then , one can see that @xmath123 .
therefore , all such contributions will be canceled out . on the other hand
, the contributions parallel to @xmath93 will satisfy the condition @xmath124@xmath125 .
these contributions can be regrouped so that the total polarization parallel to @xmath93 can be presented in the following form : @xmath126 where the summation runs over inequivalent sites of the magnetic unit cell ( note also that the crystallographic cell of orthorhombic manganites contains two mn sites in the @xmath92 plane ) .
thus , the polarization will be parallel to the @xmath93 axis .
similar result was obtained in our previous work,@xcite where we considered the magnetic structures , which respect the symmetry operation @xmath94@xmath95@xmath96 .
the present work indicates that this result is more general and does not require any specific symmetry of the magnetic structure , apart from its periodicity along the @xmath93 axis and the afm coupling along the @xmath5 axis .
another important point is that the fe polarization in all magnetic structures can be obtained by scaling the one in the e - type afm state .
the scaling factor depends on the relative orientation of spins .
below , we will study it more in details .
then , eq . ( [ eqn : xiasee ] ) can be rewritten for our purposes as @xmath127 . by substituting it in eq .
( [ eqn : portho1 ] ) and noting that only the direction - dependant part of @xmath128 will contribute to @xmath129 , one can find that @xmath130 this formula has the following consequences : \(i ) in the perfect e - type afm structure , @xmath131 is equal to @xmath95@xmath87 and @xmath24@xmath87 for the even and odd @xmath16 , respectively .
however , in the latter case , the minus sign will be additionally changed due to the prefactor @xmath132 .
therefore , all terms in the sum will be equal to @xmath87 and we will indeed obtain that @xmath133 . one can also consider the deformation of the e - type afm state , where the odd sublattice is additionally rotated relative to the even one by the angle @xmath134 ( see , e.g. , fig . 3 of ref . ) .
such a deformation is caused by the so interaction and was obtained in the hf calculations for the realistic hubbard - type model.@xcite in this case , we have @xmath135 and , therefore , @xmath136 thus , for small @xmath134 , the first correction to @xmath137 appears only in the second order of @xmath134 .
this explains that the fe polarization in the e - phase is relatively robust against the small canting of spins .
however , @xmath129 vanishes in the spin - spiral state , corresponding to @xmath138 .
\(ii ) the previous claim appears to be more general and can be reformulated as follows : any homogeneous arrangement of spins , which is characterized by the same values of @xmath131 for all bonds in the @xmath92 plane , does not break the inversion symmetry .
all these states have zero electric polarization , that directly follows from eq .
( [ eqn : portho2 ] ) .
such a situation is realized , for instance , in the ferromagnetic ( fm ) state or in the _ homogeneous spin - spiral state _
, despite a widespread believe.@xcite therefore , in order to obtain the finite polarization , it is essential to _ deform the spin spiral_. let us consider such a deformed spin - spiral structure in the @xmath92 plane , for which @xmath139 and @xmath140 .
namely , the phase @xmath141 describes the propagation of the homogeneous spin spiral and the small parameters @xmath142 , satisfying the condition @xmath143 , describe its deformation .
the general geometry of the magnetic structure implies that @xmath144 and @xmath145 , where @xmath57 is an integer .
then , since @xmath146@xmath24@xmath147 , we will have @xmath148 in the first order of @xmath149 , this expression can be further transformed to @xmath150 thus , the spin - spiral inhomogeneity contributes to @xmath129 in the first order of @xmath151 .
if the phases @xmath151 are small , the corresponding polarization is also expected to be small .
\(iii ) eq . ( [ eqn : pe ] ) clearly shows the similarities and differences between the fe polarization and interatomic superexchange interactions . indeed ,
the expression @xmath152 is nothing but the ( minus ) energy gain caused by the fm alignment of spins in the bond @xmath75-@xmath87 , which contributes to the superexchange interaction.@xcite thus , this superexchange interaction is given by the same parameters @xmath153 and @xmath154 .
the basic difference is that the superexchange interaction is given by the symmetric part of @xmath153 and @xmath154 , while the polarization depends on the antisymmetric part .
below , we present some numerical estimates , using results of previous calculations for the magnetic ground state and fe polarization of orthorhombic manganites.@xcite namely , we take the magnetic structures , obtained in the unrestricted hf calculations with the so coupling , and analyze the behavior of the fe polarization in terms of the simplified expressions ( [ eqn : p2 ] ) and ( [ eqn : p4 ] ) , obtained in the de model for the deformed magnetic structures of the afm e - type and spin - spiral type , respectively .
in the hf calculations , we typically obtain two types of magnetic structures with the twofold ( @xmath155 ) and fourfold ( @xmath156 ) periodicity .
for homno@xmath3 , they are shown in fig .
[ fig : homno3 ] .
( color online ) twofold periodic ( a ) and fourfold periodic ( b ) spin structures , as obtained in the mean - filed hartree - fock calculations for the effective hubbard - type model , constructed for the @xmath11 phase of homno@xmath3 .
the manganese atoms are indicated by the big red ( dark ) spheres and the oxygen atoms are indicated by the small green ( grey ) spheres . the same numbering of mn atoms is used in table [ tab : spiral ] .
other numbers stand for the angles @xmath157 between spin magnetic moments in the mn - o - mn bonds.,height=226 ] similar behavior was obtained for tbmno@xmath3 and ymno@xmath3.@xcite the magnetic structures with larger periodicity typically include the fragments of low periodic structures and the domain wall - defects ( see fig .
4 of ref . ) .
as was explained above , the twofold periodic structure can be viewed as the deformed e - type afm state .
the deformation is caused by the single - ion anisotropy .
for the @xmath11 phases , the angle @xmath134 , characterizing this deformation , varies from @xmath158 , in the case of tbmno@xmath3,@xcite till @xmath159 in the case of ymno@xmath3 and homno@xmath3 ( ref . and fig .
[ fig : homno3 ] ) .
then , @xmath129 can be estimated using eq .
( [ eqn : p2 ] ) as @xmath160 , which is perfectly consistent with results of the hf calculations
. indeed , the values of @xmath137 , obtained in the hf calculations for the collinear e - type afm phase without the so coupling , are @xmath161 , @xmath162 , and @xmath163 @xmath164c/@xmath165 for tbmno@xmath3 , homno@xmath3 , and ymno@xmath3 , respectively . the same calculations , but with the so coupling , yield @xmath166 @xmath167 , @xmath168 , and @xmath169 @xmath164c/@xmath165 for tbmno@xmath3 , homno@xmath3 , and ymno@xmath3 , respectively .
thus , in all three examples , the above relationship @xmath160 works very well and this tendency is nicely explained by the de model , where the so interaction is used in order to obtain the direction of spins in the ground state , while the fe polarization is calculated as a nonrelativistic quantity for the given distribution of spins .
similar analysis can be done for the fourfold periodic structure , which can be viewed as a deformed spin spiral , propagating along the @xmath4 axis . in this case , the spin spiral is stabilized by isotropic magnetic interactions and deformed by anisotropic and dm interactions.@xcite first , we take the angles @xmath170 , characterizing the directions of spins in the actual hf calculations with the so coupling ( see fig . [
fig : homno3 ] ) , and decompose them into the homogeneous ( @xmath141 ) and inhomogeneous ( @xmath142 ) parts .
then , @xmath171 is the averaged value of @xmath172@xmath24@xmath173 in the magnetic supercell and @xmath174@xmath24@xmath175@xmath24@xmath176 .
for all three compounds , we obtain @xmath177 ( in units of reciprocal lattice translation @xmath178 ) , which is close to the equilibrium values @xmath179 , obtained in the spin - spiral calculations without the so coupling.@xcite the parameters @xmath174@xmath24@xmath180 , characterizing the deformation of the spin spiral , are summarized in table [ tab : spiral ] .
.parameters @xmath181@xmath24@xmath142 , characterizing deformation of the spin - spiral state ( in degrees ) , as obtained in the unrestricted hartree - fock calculations with the spin - orbit coupling for the @xmath11 phase of tbmno@xmath3 , homno@xmath3 , and ymno@xmath3 .
the atomic positions are explained in fig .
[ fig : homno3 ] . [ cols="^,^,^,^",options="header " , ] using these parameters , the nearest - neighbor contribution to @xmath182 ( parallel yo @xmath183 ) can be estimated as @xmath24@xmath88 @xmath164c / m@xmath184 , which is small due to the strong cancelation between two inequivalent types of bonds .
the contribution of next - nearest neighbors is @xmath24@xmath185 @xmath164c / m@xmath184 , which is consistent with the value @xmath24@xmath186 @xmath164c / m@xmath184 , obtained in the hf calculations without additional approximations.@xcite thus , the value of @xmath182 is mainly determined by the wannier transfer between next - nearest neighbors in adjacent plane .
this is consistent with other magnetic properties of hexagonal manganites .
for instance , the type of the magnetic coupling between the planes is also controlled by the nnn interactions , while the contribution of the nearest neighbors is small or can be of the opposite sign.@xcite
in this work , we extend the de theory of the me coupling and systematically apply it to the wide class of multiferroic manganites , exhibiting different crystallographic and magnetic structures . for all considered materials , we are able to present a transparent physical picture of how the fe polarization is induced and controlled by the magnetic order .
this picture is based on the de theory , which was formulated for the effective low - energy model , derived from the first - principles calculations .
our basic idea is that for the analysis of electronic properties of manganites one can use two physical limits .
the first one is the de limit , which means that the intraatomic exchange splitting between the majority- and minority - spin states , driven by hund s interactions , is so large that the contribution of the latter states to the electronic polarization can be neglected .
the second one is the limit of large intraatomic splitting @xmath9 between the occupied and unoccupied orbitals with the majority spin , which is driven by the jahn - teller distortion and the screened on - site coulomb repulsion .
the second limit allows us to use the perturbation theory for the occupied wannier functions in the first order of @xmath10 , which can be incorporated in the general berry - phase theory of polarization .
thus , the electronic polarization can be described in term of the asymmetric transfer of the weights of the wannier functions to the neighboring sites , which , in the de model , are additionally modulated by the spin - dependent factors .
the de model allows to greatly simplify the analysis of the polarization and present it in the transparent form , explaining the interplay between the crystallographic and magnetic structures .
moreover , it provides the simple analytical dependence of the fe polarization on the relative directions of spins .
as expected for the nonrelativistic theory of the fe polarization , this dependence is given by the ` isotropic correlators ' @xmath2 .
it has the same form as for the phenomenological magnetostriction mechanism .
nevertheless , this is a new mechanism , which is not directly related to the magnetostriction and can take place without magnetostriction , even in a centrosymmetric crystal structure .
first , we have systematically applied the de theory to the orthorhombic @xmath11 manganites and generalized results of our previous work ( ref . ) .
our present result is valid for any noncollinear magnetic structure , of an arbitrary periodicity , propagating along the orthorhombic @xmath4 axis and antiferromagnetically coupled along the @xmath5 axis . for this type of magnetic structures ,
we have argued that the fe polarization should be parallel to the orthorhombic @xmath93 axis and can be obtained by scaling the one of the collinear e - type afm state with the scaling factor depending exclusively on the relative direction of spins .
the fe polarization vanishes in the homogeneous spin - spiral state , which preserves the inversion symmetry of the de hamiltonian .
therefore , the only possibility to obtain the finite polarization is to deform the homogeneous spin spiral and to produce some inhomogeneity in the distribution of spins . in multiferroic manganites ,
such a deformation is caused by the relativistic so interaction .
this picture works equally well for manganites with the twofold and fourfold periodic magnetic structures , which typically attributed to homno@xmath3 and tbmno@xmath3 , respectively .
the basic difference is that , even despite some spin canting and deviation from the collinear e - type afm alignment , the twofold periodic magnetic structure remains strongly inhomogeneous , that leads to the large polarization . on the contrary
, the fourfold periodic magnetic structure can be viewed as a distorted homogenous spin spiral . therefore , if the distortion is small , the polarization is also small .
next , we have studied the microscopic origin of the fe polarization , caused by the magnetic inversion symmetry breaking in the @xmath12 phase of bimno@xmath187 .
the uniqueness of this situation is that the magnetic ground state of bimno@xmath3 contains both afm @xmath188 component , which breaks the inversion symmetry , and the fm magnetization , caused by the canting of spins .
thus , this is a rare case , where the fe polarization indeed coexists with the fm magnetization , that is very important from the viewpoint of the mutual control of ferroelectricity and magnetism . according to the mean - field hf calculations ,
the afm magnetization lies in the monoclinic @xmath26 plane , while the fm one is parallel to the @xmath189 axis.@xcite we have modeled this magnetic structure in our de analysis in order to find a quantitative relationship between the fm magnetization and the fe polarization .
as expected , these two quantities ` anticorrelate ' with each other : by enforcing the fm magnetization , one can decreases the polarization .
the latter can be obtained by scaling the one of the collinear @xmath188 afm state .
the scaling factor is given again by the correlation function between directions of neighboring spins , but since the fm moments do not contribute to the polarization , this correlation function includes only afm components of the magnetization .
the polarization in this case lies in the @xmath26 plane ( so as the afm magnetization ) .
moreover , the @xmath190 component of the polarization is substantially larger than the @xmath183 one .
we have found this behavior to be closely related to the orbital ordering , realized in bimno@xmath187 : while the @xmath190 component is very robust and can be expected even in the perfect cubic lattice ( provided that it supports the particular type of the orbital ordering , realized in bimno@xmath3 ) , the weak @xmath183 component is the measure of deviation from the perfect cubic environment , which crucially depends on the details of the monoclinic distortion . finally , we have explained the origin of the me coupling , associated with the reversal of spins in every second @xmath21 plane of hexagonal manganites .
such a reversal can be indeed expected in realistic materials if one can induce the change of the magnetic structure from @xmath191 to @xmath192 ( or between any two types of the magnetic structures , in which the magnetic moments in the neighboring planes are transformed , respectively , by the native symmetry operations of the @xmath193 space group and by the same symmetry operations , combined with the time reversal ) .
although the @xmath193 space group has no inversion symmetry and , therefore , the system is expected to be ferroelectric , irrespectively on the magnetic order , this change of the magnetic structure produced a finite change of the fe polarization .
we have derived an analytical expression for the spin - dependent part of the polarization and evaluated different contributions to it , associated with the transfer of the weights of the wannier functions to different groups of sites in the adjacent planes .
we have found that the main contribution comes from next - nearest neighbors , while the ones from the nearest sites are small due to the strong cancelation , which occurs between two inequivalent types of bonds . in conclusion ,
the de mechanism of the me coupling plays a very important role in physics of multiferroic manganites and explains many basics aspects of the fe activity in these systems on a unified ground .
thus , this is the key microscopic mechanisms , which should be considered in the analysis of multiferroic properties of manganites and related compounds .
w. eerenstein , n. d. mathur , and j. f. scott , nature * 442 * , 760 ( 2006 ) ; t. kimura , annu .
res . * 37 * , 387 ( 2007 ) ; s .- w .
cheong and m. mostovoy , nature materials * 6 * , 13 ( 2007 ) ; d. khomskii , physics * 2 * , 20 ( 2009 ) ; y. tokura and s. seki , adv .
* 22 * , 1554 ( 2010 ) .
the magnetic structure of many multiferroic manganites is organized in such a way that there is certain group of mn sites , which are transformed to themselves by the regular inversion .
this typically contradicts to the symmetry properties of some other group of mn atoms , which , due to specific form of interatomic magnetic interactions , should be transformed to themselves by the inversion combined with the time reversal operation . in order to satisfy simultaneously these two symmetry constraints ,
one of atomic groups should become nonmagnetic .
however , it would correspond to a gigantic loss of intraatomic hund s energy . therefore
, it is more favorable energetically to preserve the high spin state of mn@xmath8 , but to break the inversion symmetry . c. zener , phys .
82 * 440 ( 1951 ) ; p. w. anderson and h. hasegawa , phys
. rev . * 100 * 675 ( 1955 ) ; p .- g . de gennes , phys
. rev . * 118 * 141 ( 1960 ) ; k. kubo and n. ohata , j. phys .
. jpn . * 33 * 21 ( 1972 ) .
i. v. solovyev and z. v. pchelkina , pisma zh .
. fiz . * 89 * , 701 ( 2009 ) [ jetp lett . * 89 * , 597 ( 2009 ) ] ; i. v. solovyev and z. v. pchelkina , phys .
b * 82 * , 094425 ( 2010 ) .
note that theoretical values of the fe polarization , reported in the last paper ( and also in ref . ) , were oversetimated by about @xmath194 because of numerical error .
w. kohn and l. j. sham , phys .
rev . * 140 * , a1133 ( 1965 ) ; r. g. parr and w. yang , _ density - functional theory of atoms and molecules _ ( clarendon press , oxford , 1989 ) ; w. kohn , rev .
phys . * 71 * , 1253 ( 1999 ) .
a. a. belik , s. iikubo , t. yokosawa , k. kodama , m. igawa , s. shamoto , m. azuma , m. takano , k. kimoto , y. matsui , and e. takayama - muromachi , j. am .
soc . * 129 * , 971 ( 2007 ) ; a. a. belik , j. solid state chem . *
195 * , 32 ( 2012 ) . | the double exchange ( de ) model is systematically applied for studying the coupling between ferroelectric ( fe ) and magnetic orders in several prototypical types of multiferroic manganites .
the model itself was constructed for the magnetically active mn @xmath0 bands in the basis of wannier functions and include the effect of screened on - site coulomb interactions in the hartree - fock approximation .
all model parameters were derived from the first - principles electronic structure calculations .
the essence of our approach for the fe polarization is to use the berry phase theory , formulated in terms of occupied wannier functions , and to evaluate the asymmetric spin - dependent change of these functions in the framework of the de model .
this enables us to quantify the effect of the magnetic symmetry breaking and derive several useful expressions for the electronic polarization @xmath1 , depending on the relative directions of spins .
the spin - dependence of @xmath1 in the de model is given by the isotropic correlation functions @xmath2 between directions of neighboring spins . despite formal similarity with the magnetostriction mechanism , the magnetoelectric ( me ) coupling in the proposed de theory
is not related to the magnetically driven fe atomic displacements and can exist even in compounds with the centrosymmetric crystal structure , if the spacial distribution of @xmath2 does not respect the inversion symmetry .
the proposed theory is applied to the solution of three major problems : ( i ) the magnetic - state dependence of @xmath1 in hexagonal manganites , using ymno@xmath3 as an example ; ( ii ) the microscopic relationship between canted ferromagnetism and @xmath1 in monoclinic bimno@xmath3 ; ( iii ) the origin of fe activity in orthorhombic manganites .
particularly , we will show that for an arbitrary noncollinear magnetic structure , propagating along the orthorhombic @xmath4 axis and antiferromagnetically coupled along the @xmath5 axis , @xmath1 is induced by an _
inhomogeneous _ distribution of spins and can be obtained by scaling the one of the e - type antiferromagnetic ( afm ) phase with the prefactor depending only on the relative directions of spins and being the measure of this spin inhomogeneity .
this picture works equally well for the twofold ( homno@xmath3 ) and fourfold ( tbmno@xmath3 ) periodic manganites .
the basic difference is that , even despite some spin canting of the relativistic origin and deviation from the collinear e - type afm aligment , the twofold periodic magnetic structure remains strongly inhomogeneous , that leads to large @xmath1 . on the contrary
, the fourfold periodic magnetic structure can be viewed as a moderately distorted homogeneous spin spiral , which corresponds to much weaker @xmath1 . |
in their landmark paper @xcite , lascoux , leclerc and thibon conjectured that the decomposition matrices of the iwahori - hecke algebras of type @xmath2 could be computed using the canonical bases of the _ level _ one fock spaces .
ariki @xcite generalised , and proved , this conjecture for the cyclotomic hecke algebras of arbitrary level in type @xmath2 . unlike in level one , the canonical bases of higher level fock spaces are quite difficult to compute . building on @xcite ,
fayers @xcite gave a more efficient algorithm for computing the canonical bases of the higher level fock spaces .
as a result of his calculations he made a natural conjecture for the maximal degrees of the llt - polynomials , which are certain parabolic kazhdan - lusztig polynomials that are computed by fayers algorithm . this conjecture ,
if true , further improves the efficiency of fayers algorithm .
this paper started as a project to prove fayers conjecture but the machinery that we develop to do this has several other applications .
fayers conjecture is a purely combinatorial statement that gives an upper bound on the degrees of certain polynomials .
as we will show , the representation theoretic significance of his conjecture gives information about the socle of the weyl modules .
to prove fayers conjecture , we take advantage of recent advances by stroppel - webster @xcite , maksimau @xcite , brundan - kleshchev @xcite and the authors @xcite in the graded representation theory of the cyclotomic hecke and schur algebras to connect the higher level llt - polynomials with the graded decomposition numbers of these algebras .
this machinery allows us to reduce fayers conjecture to understanding the socle of a weyl module .
this paper starts by analysing the endomorphism algebra @xmath3 of a _
schur pair _ @xmath4 , where @xmath2 is a self - injective algebra and @xmath5 is a faithful @xmath2-module . in this very general
setting , which includes the cyclotomic hecke algebras @xmath6 and the schur algebras @xmath7 as special cases , we classify the projective - injective @xmath8-modules and show that the modules that can appear in the socle of the projective @xmath8-modules are labelled by the simple @xmath2-modules .
we then proceed to apply this general theory to the cyclotomic hecke algebras and schur algebras .
our main results are : * we prove that the simple modules appearing in the socle of a weyl modules are indexed by _ kleshchev multipartitions _ ( [ t : deltasocle ] ) .
* we prove fayers conjecture ( [ t : fayers ] ) .
* we give a cyclotomic generalisation of the carter - lusztig theorem @xcite ( [ t : carterlusztig ] ) . along the way we prove a number of useful new results about the cyclotomic hecke and schur algebras , such as a new description of the simple @xmath6-modules ( [ l : simplesubmodules ] ) and an interpretation of ringel duality at the level of hecke algebras ( [ t : ringelduality ] ) .
all of the results in this paper apply to what are traditionally called the degenerate and non - degenerate cyclotomic hecke algebras and schur algebras ( but not to the ariki - koike algebras with @xmath9 ) .
this is because , following @xcite , we use a slightly different presentation ( [ d : hecke ] ) of the cyclotomic hecke algebras that simultaneously captures the degenerate and non - degenerate cyclotomic hecke algebras .
both authors were supported by the australian research council .
the first author was also supported by the national natural science foundation of china ( no . 11525102 ) .
partly inspired by the extensive literature of schur algebras acting on _ tensor space _ , such as @xcite , this section studies the endomorphism algebra of a finite dimensional faithful module of a finite dimensional self - injective algebra .
our main aim is to understand , in this general setting , what simple modules can appear in the socle of `` tensor space '' . throughout this paper
we fix a field @xmath10 .
if @xmath2 is a @xmath10-algebra let @xmath11 and @xmath12 be the categories of finite dimensional left and right @xmath2-modules , respectively
. recall that a * trace form * on an algebra @xmath2 is a linear map @xmath13 such that @xmath14 , for all @xmath15 .
the algebra @xmath2 is * symmetric * if it has a non - degenerate symmetric bilinear form @xmath16 such that @xmath17 the algebra @xmath2 is * self - injective * if the regular representation of @xmath2 is injective as an @xmath2-module . as is well - known , symmetric algebras are self - injective .
[ d : schurpair ] a * schur pair * is an ordered pair @xmath4 , where @xmath2 is a self - injective finite dimensional @xmath10-algebra and @xmath5 is a faithful finite dimensional right @xmath2-module .
if @xmath4 is a schur pair let @xmath18 be the algebra of @xmath2-module endomorphisms of @xmath5 . by extending @xmath10 , if necessary
, we will always assume that @xmath2 and @xmath8 are both split over @xmath10 .
an explicit example to keep in mind is the case where @xmath2 is the iwahori - hecke algebra of the symmetric group and @xmath5 is tensor space ( * ? ? ?
* and 2 ) , so that @xmath8 is the @xmath1-schur algebra .
more generally , as we show in [ p : schurpairs ] , we can take @xmath2 to be a cyclotomic hecke algebra of type @xmath2 , in which case there are two natural choices for @xmath5 . for the remainder of this section
we assume that @xmath4 is a schur pair and that @xmath19 .
our aim is to understand the interplay between the representation theory of the algebras @xmath2 and @xmath8 .
our starting point is the following easy fact .
[ p : equivalence ] suppose that @xmath4 is a schur pair . 1 .
there exists an integer @xmath20 and an @xmath2-module @xmath21 such that @xmath22 as @xmath2-modules .
2 . if @xmath20 then @xmath23 is a schur pair and the algebras @xmath24 and @xmath25 are morita equivalent . first consider ( a ) . since @xmath5 is a faithful finite dimensional @xmath2-module we can find an integer @xmath26 such that @xmath2 is isomorphic to an @xmath2-submodule of @xmath27 as a right @xmath2-module . indeed , if @xmath28 is a basis of @xmath5 then the map @xmath29 given by @xmath30 is injective .
hence , @xmath2 is isomorphic to a submodule of @xmath27 whenever @xmath31 .
fix such an integer @xmath32 .
since @xmath2 is self - injective the map @xmath33 splits , so @xmath34 as an @xmath2-module .
this completes the proof of part ( a ) .
part ( b ) follows directly from the definitions since @xmath5 and @xmath35 have the same indecomposable summands . in terms of the representation theory of @xmath8 , [ p : equivalence ] says that there is no harm in assuming that @xmath2 is a direct summand of @xmath5 whenever @xmath4 is a schur pair
. this will be convenient in several of the arguments below .
the algebra @xmath8 acts from the left on @xmath5 as an algebra of endomorphisms , with @xmath36 , for @xmath37 and @xmath38 .
moreover , since @xmath2 acts faithfully on @xmath5 , there is an algebra embedding @xmath39 , where we identify @xmath40 with the endomorphism @xmath41 given by @xmath42 , for @xmath38 . by construction , @xmath43 hence
, the left and right actions of @xmath8 and @xmath2 on @xmath5 commute with each other and @xmath44 is a subalgebra of @xmath45 . the next result will let us show that @xmath46 .
[ l : doubelcentralizer ] suppose that @xmath4 is a schur pair and @xmath47
. then @xmath48 .
the proof is elementary but we give the argument because we need this idea below .
there is an embedding @xmath49 given by @xmath50 , where @xmath51 is the endomorphism of @xmath35 given by @xmath52 , for @xmath53 . for @xmath54 let @xmath55 be projection onto the @xmath56th component of @xmath35 and
let @xmath57 be the map that swaps the @xmath56th and @xmath58th components .
then @xmath59 . therefore ,
if @xmath60 and @xmath61 then @xmath62 .
hence , @xmath63 maps the @xmath56th component of @xmath35 into the @xmath56th component so that @xmath64 for suitable @xmath65 and @xmath66 .
finally , if @xmath67 then @xmath68 .
this implies that @xmath69 , for all @xmath38 .
hence , the map @xmath70 given above is surjective and the lemma is proved .
the next result shows that @xmath2 and @xmath8 enjoy a _
double centralizer property _ in the sense that @xmath71 and @xmath72 .
this result is well - known and appears as .
we give a self contained proof of this result both because it is central to all of the results in this section and because the proof is quite short .
[ t : doublecentraliser ] let @xmath4 be a schur pair
. then @xmath46 . by [ p : equivalence ] and [ l : doubelcentralizer ]
, we can assume that @xmath73 , for some @xmath2-module @xmath21 .
this reduces the proof of the theorem to the claim that @xmath74 , where @xmath75 .
recall that @xmath76 , where @xmath77 is the left regular representation of @xmath2 . with this observation
the claim now follows easily using a similar argument to the proof of [ l : doubelcentralizer ] ; see for more details .
[ c : schurfunctor ] let @xmath4 be a schur pair such that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map
. then @xmath79 , @xmath80 as an @xmath8-module , @xmath81 . by definition
, @xmath82 is an endomorphism of @xmath5 that commutes with the action of @xmath2 so we can consider @xmath82 to be an element of @xmath8 .
then @xmath83 as an left @xmath8-module .
explicitly , the isomorphism is given by @xmath84 , where @xmath85 is the identity element of @xmath2 .
hence , by [ t : doublecentraliser ] , @xmath86 .
suppose that @xmath4 is a schur pair .
define a functor @xmath87 by @xmath88 , for any @xmath89 and @xmath90 for any @xmath91 .
note that @xmath92 is a right @xmath2-module with @xmath2-action given by @xmath93 , for all @xmath94 .
moreover , if @xmath2 is a direct summand of @xmath5 with @xmath78 be the natural projection map , then @xmath95 .
[ c : mprojective ] the module @xmath5 is projective as a left @xmath8-module . in view of [ p : equivalence ] , it is sufficient to consider the case when @xmath2 is a direct summand of @xmath8
. then @xmath80 as an @xmath8-module by [ c : schurfunctor ] , so @xmath5 is a projective @xmath8-module .
[ c : fullyfaithful ] the functor @xmath96 is exact and fully faithful on projective @xmath8-modules .
the functor @xmath96 is exact because @xmath5 is projective as an @xmath8-module by [ c : mprojective ] . to prove that @xmath96 is fully faithful on projectives it is enough to consider the case when @xmath2 is a direct summand of @xmath5 by [ p : equivalence ]
. then @xmath97 and @xmath98 by [ c : schurfunctor ] and the discussion above [ c : mprojective ] , for any @xmath8-module @xmath99 .
hence , @xmath96 is fully faithful on projectives by [ t : doublecentraliser ] . by [ c : schurfunctor ] , if @xmath2 is a direct summand of @xmath5 then @xmath81 and @xmath100 , for some idempotent @xmath79 .
therefore , in order to understand the connection between @xmath8-modules and @xmath2-modules we can apply auslander s theory of quotient functors , or schur functors , as in @xcite or @xcite .
let @xmath101 be a complete set of pairwise non - isomorphic simple @xmath8-modules in @xmath102 , where @xmath103 is an indexing set . for each @xmath104
, define the @xmath2-module @xmath105 and set @xmath106 . by the general theory of quotient functors , such as in (
* ( 6.2 g ) ) , @xmath107 is a complete set of pairwise non - isomorphic simple right @xmath2-modules . by [ p : equivalence ]
this is true whenever @xmath19 and @xmath4 is a schur pair ( that is , we do not need to assume that @xmath2 is a direct summand of @xmath5 ) . [
d : youngmodules ] let @xmath108 .
let @xmath109 be the * projective cover * of @xmath110 and define the * young module * @xmath111 .
@xmath109 and @xmath110 are @xmath8-modules and @xmath112 and @xmath113 are @xmath2-modules . in the special case when @xmath8 is the schur algebra and @xmath2 is the group algebra of the symmetric group , @xmath113 is the usual young module . [
p : youngmodules ] suppose that @xmath114 . then the following hold . 1 .
the young module @xmath113 is an indecomposable @xmath2-module .
2 . there is an isomorphism @xmath115 if and only if @xmath116 .
the young module @xmath113 is projective if and only if @xmath117 .
4 . if @xmath117 then @xmath113 is the projective cover of @xmath112 . by [ c : fullyfaithful ] ,
the functor @xmath118 is fully faithful on projective @xmath8-modules .
therefore , @xmath119 is a local @xmath10-algebra since @xmath109 is indecomposable .
therefore , @xmath113 is an indecomposable @xmath2-module .
further , @xmath120 , so @xmath115 if and only if @xmath116 . we have established parts ( a ) and ( b ) , so it remains to prove ( c ) and ( d ) . by [ c : schurfunctor ]
we may assume that @xmath2 is a direct summand of @xmath5 , as an @xmath2-module , so that @xmath81 for some idempotent @xmath79 .
therefore , @xmath121 as right @xmath2-modules . therefore , every indecomposable projective right @xmath2-module is isomorphic to @xmath122 , for some @xmath123 .
fix @xmath124 and let @xmath125 be the projective cover of the simple @xmath2-module @xmath126 .
by the last paragraph , @xmath127 for some @xmath104 .
suppose , by way of contradiction , that @xmath128 .
as @xmath129 is the projective cover of @xmath130 , by applying @xmath96 there is a surjective map @xmath131 since @xmath132 .
let @xmath133 be the multiplicity of @xmath126 in the head of @xmath134 .
as @xmath135 is indecomposable we can find a non - zero map @xmath136 such that @xmath126 does not appear in the head of @xmath137 .
hence , by [ c : fullyfaithful ] , there exists a non - zero map @xmath138 such that @xmath130 does not appear in the head of @xmath139
. however , this is a contradiction because @xmath129 has simple head @xmath130 .
hence , @xmath116 and @xmath140 as we wanted to show .
this completes the proof of parts ( c ) and ( d ) .
[ t : sdecomposition ] suppose that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map . 1 . as a right @xmath8-module , @xmath141
2 . as a right @xmath2-module , @xmath142 by [ c : schurfunctor ] , @xmath81 for some idempotent @xmath79 . first consider ( a ) .
there exist non - negative integers @xmath143 such that @xmath144 therefore , using [ c : schurfunctor ] for the second equality , @xmath145 the direct summands of @xmath2 are necessarily projective @xmath2-modules , so @xmath146 if and only if @xmath124 and @xmath134 is the projective cover of @xmath126 by [ p : youngmodules](d ) .
moreover , since @xmath2 is @xmath10-split by assumption , @xmath147 for all @xmath148 .
this completes the proof of ( a ) .
now consider ( b ) . as a right @xmath8-module , @xmath149 since @xmath8
is split over the field @xmath10 . by definition , @xmath150 . in view of [ c : schurfunctor ] , @xmath151 as a right @xmath2-module . by [ p : youngmodules ]
, @xmath152 is a complete set of pairwise non - isomorphic indecomposable projective right @xmath2-modules . as @xmath2 is self - injective
, @xmath152 is also a complete set of pairwise non - isomorphic indecomposable injective right @xmath2-modules .
the next few results assume that @xmath81 as in [ c : schurfunctor ] . for these results
we identify the algebras @xmath2 and @xmath153 using this isomorphism . for the next result ,
say that a right @xmath8-submodule @xmath10 of @xmath154 is * generated * by @xmath2 if @xmath155 .
[ l : ideals ] suppose that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map .
then : 1 . the map @xmath156 defines an inclusion preserving bijection between the set of right @xmath2-submodules of @xmath153 and the set of right @xmath8-submodules of @xmath154 generated by @xmath2 .
if @xmath157 and @xmath158 are right @xmath2-submodules of @xmath153 then @xmath159 3 . if @xmath157 and @xmath158 are right @xmath2-submodules of @xmath153 then @xmath160 as @xmath2-modules if and only if @xmath161 as @xmath8-modules .
let @xmath157 be a right @xmath2-submodule of @xmath153 .
we claim that @xmath162 .
certainly , @xmath163 .
conversely , if @xmath164 then @xmath165 and we can write @xmath166 , for some @xmath167 . therefore , @xmath168 since @xmath169 and @xmath157 is a right @xmath2-module .
therefore , @xmath162 as claimed .
hence , the ideal @xmath170 is generated by @xmath2 .
moreover , if @xmath157 and @xmath158 are right @xmath2-submodules of @xmath153 then @xmath171 if and only if @xmath172 .
we have now proved part ( a )
. as ( c ) follows immediately from ( b ) , it remains to prove ( b ) .
let @xmath157 and @xmath158 be @xmath2-submodules of @xmath153 . by the last paragraph , any homomorphism @xmath173 restricts to a well - defined @xmath2-module homomorphism @xmath174 since @xmath81 .
conversely , since @xmath2 is self - injective any homomorphism between two right ideals of @xmath2 is given by left multiplication by .
that is , if @xmath175 is an @xmath2-module homomorphism then there exists an @xmath176 such that @xmath177 , for all @xmath178 .
therefore , there is a well - defined @xmath8-module homomorphism @xmath173 given by @xmath179 , for all @xmath180 . by construction , @xmath181
is uniquely determined by @xmath182 , and it restricts to @xmath182 , so ( b ) follows .
suppose that the algebra @xmath2 comes equipped with an anti - involution @xmath183 .
if @xmath99 is any right @xmath2-module then the dual module @xmath184 becomes a right @xmath2-module with action @xmath185 , for @xmath40 , @xmath186 and @xmath187 .
[ a : selfdual ] a * self - dual schur pair * is a schur pair @xmath4 where @xmath2 is equipped with an anti - involution @xmath183 such that @xmath188 as @xmath2-modules . except for the last two results in this section
, we now assume that @xmath4 is a self - dual schur pair and we fix an @xmath2-module isomorphism @xmath189 . following ( * ? ? ?
* ( 1.5 ) ) , define an anti - isomorphism @xmath190 where @xmath191 , for @xmath192 and @xmath193 .
then the anti - isomorphism @xmath183 on @xmath2 induces an anti - isomorphism on @xmath3 which is given by @xmath194 , for all @xmath195 . by definition , if @xmath195 and @xmath38 then @xmath196 so that @xmath197 or , equivalently , @xmath198
. hence , @xmath199 is also an @xmath8-module isomorphism .
[ l : msselfdual ] suppose that @xmath4 is a self - dual schur pair .
then @xmath5 is self - dual and projective - injective as an @xmath8-module . in the last paragraph we observed that @xmath199 is an @xmath8-module homomorphism , so @xmath188 as an @xmath200-bimodule . in particular
, @xmath5 is self - dual as an @xmath8-module . by [ c : mprojective ] ,
@xmath5 is projective as an @xmath8-module so @xmath201 is injective giving the remaining claim .
suppose that @xmath4 is a self - dual schur pair .
for any @xmath8-module @xmath99 let @xmath202 be the right @xmath8-module such that @xmath203 as a @xmath10-vector space and @xmath204 , for @xmath195 and @xmath205 .
suppose that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map . by definition and [ c : schurfunctor ] , @xmath206 .
the next lemma , which is part of the motivation for introducing self - dual schur pairs , allows us to replace @xmath207 with @xmath154 .
note that @xmath154 is a left @xmath2-module since @xmath81 .
[ l : ee * ] suppose that @xmath4 is a self - dual schur pair such that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map .
then @xmath208 and @xmath183 induces the isomorphisms : @xmath209 by [ c : schurfunctor ] we can identify @xmath2 with @xmath153 and @xmath5 with @xmath207 .
we have fixed an @xmath2-module isomorphism @xmath210 and , by definition , if @xmath38 then @xmath211 so that @xmath212 . for @xmath195 , the map @xmath213 defines a right @xmath8-module isomorphism @xmath214 . in particular , @xmath215 .
if @xmath216 is a right @xmath217-module for an algebra @xmath217 then the * socle * of @xmath216 , @xmath218 , is the maximal semisimple ( right ) submodule of @xmath216 .
dually , the * head * of @xmath216 , @xmath219 , is its maximal semisimple ( right ) quotient module .
when the algebra @xmath217 is clear then we simply write @xmath220 and @xmath221 .
[ l : keystep ] suppose that @xmath4 is a self - dual schur pair such that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map .
let @xmath99 be a non - zero right @xmath8-submodule of @xmath154 .
then @xmath222 is a non - zero right @xmath2-submodule of @xmath153 .
it is easy to see that @xmath222 is a right @xmath2-submodule of @xmath153 .
we need to prove that @xmath223 . before we start
the proof observe that we can regard @xmath154 as a left @xmath2-module because @xmath81 .
let @xmath224 and let @xmath225 be the injective hull of @xmath226 . as @xmath2 is self - injective @xmath227 is also projective as a left @xmath2-module .
therefore , we can find an integer @xmath228 such that @xmath227 embeds into @xmath229 as a left @xmath2-module . recalling that @xmath227 is the injective hull of @xmath224
, this implies that there exists a left @xmath2-module homomorphism @xmath230 such that @xmath199 restricts to the identity map on @xmath226 .
now are now ready to prove the lemma .
fix a non - zero element @xmath187 .
then @xmath231 so we can find a non - zero element @xmath232 such that @xmath233 because @xmath234 .
therefore , @xmath235 . by composing @xmath199 with a suitable projection of @xmath229 on to @xmath2 , and then using the inclusion @xmath236
, we obtain a left @xmath2-module homomorphism @xmath237 such that @xmath238 and @xmath239 .
consequently , @xmath240 since @xmath241 .
write @xmath242 , for some @xmath243
. then @xmath244 , so @xmath245 . in particular
, @xmath246 is a non - zero element of @xmath2 . to complete the proof recall that @xmath247 by [ l : ee * ] .
therefore , we may assume that @xmath248 and consider the endomorphism @xmath249 of @xmath207 . since @xmath99 is an @xmath8-module , @xmath250 .
that is , @xmath251 is a non - zero element of @xmath222 , so @xmath223 as we wanted to show
. we can now prove the key result of this section .
recall that @xmath252 is a complete set of irreducible @xmath8-modules and that @xmath253 is a complete set of irreducible @xmath2-modules .
[ t : socle ] suppose that @xmath254 is a self - dual schur pair such that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map .
let @xmath148 .
the simple @xmath8-module @xmath130 is isomorphic to a submodule of @xmath154 if and only if @xmath124 .
suppose that @xmath255 is an irreducible right @xmath8-submodule of @xmath154
. then @xmath256 by [ l : keystep ] .
therefore , @xmath257 is a non - zero @xmath8-submodule of @xmath255 so that @xmath258 since @xmath255 is irreducible . in view of [ l : ideals](a ) , @xmath226 is an irreducible right @xmath2-submodule of @xmath153 .
moreover , since @xmath259 for all @xmath260 it follows that @xmath261 since @xmath262 .
therefore , @xmath263 since @xmath226 is a non - zero @xmath2-submodule of @xmath264 and @xmath264 is irreducible
. that is , @xmath265 . consequently , if @xmath266 then @xmath267 so that @xmath124 .
conversely , if @xmath124 then there exists an irreducible submodule @xmath226 of @xmath153 such that @xmath268 as a right @xmath2-module . by [ l : ideals](a ) , @xmath258 is an irreducible right @xmath8-submodule of @xmath154 . by the argument of the last paragraph @xmath269 .
hence , @xmath266 and @xmath154 has a submodule isomorphic to @xmath130 for all @xmath124 .
note that [ t : socle ] would follow from [ t : sdecomposition ] if @xmath154 were self - dual as an @xmath8-module .
recall that @xmath270 is the socle of @xmath154 as an @xmath8-module .
[ t : socle ] says that @xmath130 appears in @xmath270 if and only if @xmath124 .
[ c : socle ] let @xmath114
. then @xmath130 appears in the socle of @xmath109 only if @xmath271 . by definition
, @xmath154 is a faithful right @xmath8-module .
therefore , as in the proof of [ c : schurfunctor ] , @xmath8 is isomorphic to a submodule of @xmath272 , for some @xmath273 .
it follows that @xmath109 is isomorphic to a submodule of @xmath272 .
in particular , the socle of @xmath109 is isomorphic to a submodule of @xmath274 .
hence , @xmath130 appears in @xmath275 only if @xmath124 by [ t : socle ]
. a * block * of an algebra @xmath217 is a two - sided ideal .
every algebra has a unique decomposition @xmath276 into a direct sum of _ indecomposable _ blocks .
a @xmath217-module @xmath99 * belongs * to the block @xmath277 if @xmath278 .
it is easy to see that indecomposable modules belong to a unique block . in particular
, the block decomposition of @xmath217 induces an equivalence relation on the simple @xmath217-modules where two simples are equivalent if they belong to the same block . as is well - known , this equivalence relation coincides with the * linkage classes * of simple @xmath217-modules , where the equivalence relation for the linkage classes is generated by @xmath279 if @xmath280 , where @xmath281 and @xmath282 are the projective covers of the simple modules @xmath226 and @xmath216 . [ c : blocks ] suppose that @xmath254 is a self - dual schur pair .
the functor @xmath96 induces a one - to - one correspondence between the indecomposable blocks of @xmath8 and the indecomposable blocks of @xmath2 . without loss of generality
, we can assume that @xmath2 is a direct summand of @xmath5 .
let @xmath283 and @xmath284 be the decompositions of @xmath2 and @xmath8 into indecomposable blocks .
let @xmath285 be the block idempotents of @xmath2 .
set @xmath286 , for @xmath287 .
then @xmath288 for @xmath287 since @xmath5 is faithful as an @xmath2-module , so @xmath289 is the decomposition of @xmath5 into its block components . in particular , @xmath290 if @xmath291 .
therefore , @xmath292 is a decomposition of @xmath8 into ( not necessarily indecomposable ) blocks .
in particular , @xmath293 . on the other hand ,
as recalled before the corollary , the linkage classes of simple modules are determined by the non - vanishing hom - spaces between the projective indecomposable modules . by [ c : fullyfaithful ] , if @xmath294 then @xmath295 , so @xmath110 and @xmath130 are linked whenever @xmath112 and @xmath126 are linked .
furthermore , by [ c : socle ] , every simple @xmath8-module @xmath110 with @xmath104 is linked with a simple @xmath8-module @xmath130 with @xmath124 . therefore , there are at least as many linkage classes of simple @xmath2-modules as there are linkage classes of simple @xmath8-modules .
that is , @xmath296 . by the last two paragraphs , @xmath297 ,
so that @xmath2 and @xmath8 have the same number of indecomposable blocks . finally , note that the blocks of @xmath2 and @xmath8 are direct sums of indecomposable projective modules .
hence , @xmath96 decomposes as a direct sum of functors between the module categories of the blocks of @xmath8 and @xmath2 and it induces a bijection between the indecomposable blocks of @xmath8 and @xmath2 .
[ l : yselfdual ] suppose that @xmath298 , for all @xmath124 .
then @xmath299 as @xmath2-modules for all @xmath124 . since @xmath298 for all @xmath124 this follows from the general formalism for symmetric algebras given by ( * ?
* theorem 3.1 ) .
alternatively , a direct proof can be given by adapting the argument of ( * ? ? ?
* theorem 6 ) .
the next result shows that the projective - injective @xmath8-modules are indexed by the simple @xmath2-modules .
the assumption that the simple @xmath2-modules are self dual is used to ensure that the head and socle of @xmath109 are isomorphic whenever @xmath109 is self - dual . in applications @xmath2 is usually a cellular algebra , in the sense of @xcite , in which case this assumption is automatic . [
t : equivalence ] suppose that @xmath4 is a self - dual schur pair such that @xmath298 , for all @xmath124 .
let @xmath104 .
then the following are equivalent : 1 .
@xmath117 2 .
@xmath300 3 .
@xmath110 is a right submodule of @xmath301 4 .
@xmath109 is a direct summand of @xmath301 5 .
@xmath109 is a projective - injective right @xmath8-module 6 .
@xmath109 is self - dual . using [ p : equivalence ]
, we can assume without loss of generality that @xmath2 is a direct summand of @xmath5 .
let @xmath78 be the natural projection map and we identify @xmath2 with @xmath302 and @xmath5 with @xmath207
. then @xmath303 and @xmath215 by [ l : ee * ] by definition , @xmath117 if and only if @xmath300 .
further , ( a ) and ( c ) are equivalent by [ t : socle ] and ( a ) and ( d ) are equivalent by [ t : sdecomposition ] .
hence , ( a ) , ( b ) , ( c ) and ( d ) are equivalent . to complete the proof
we show that ( a ) @xmath304 ( f ) @xmath304 ( e ) @xmath304 ( d ) .
first suppose that ( a ) holds so that @xmath117 . by definition , @xmath109 has a simple head @xmath110 .
then @xmath109 is an indecomposable direct summand of @xmath5 by [ t : sdecomposition ] .
therefore , @xmath305 is an indecomposable summand of @xmath301 by [ a : selfdual ] , so @xmath306 for some @xmath124 by [ t : sdecomposition ] .
hence , @xmath129 has simple socle @xmath307 which implies that there exists a non - zero homomorphism @xmath308 from @xmath309 to @xmath129 such that @xmath310 . as the functor @xmath96 is fully faithful there exists a non - zero homomorphism @xmath311 from @xmath312 to @xmath134 such that @xmath313 . by [ l : yselfdual ] , @xmath134 is self - dual so it has simple socle @xmath126 .
therefore , @xmath314 so that @xmath315 is self - dual .
hence , ( a ) @xmath304 ( f ) . now suppose that ( f ) holds so that @xmath109 is self - dual . by definition , @xmath316 is projective so it is also injective and ( e ) holds .
finally , suppose that ( e ) holds so that @xmath109 is a projective - injective @xmath8-module .
since @xmath301 is a faithful right @xmath8-module , @xmath8 embeds into @xmath317 , for some @xmath228 , by the argument of [ p : equivalence ] .
fix such a @xmath318 .
since @xmath305 is projective it embeds into @xmath317 . taking duals
there is a surjection @xmath319 .
as @xmath109 is projective this map splits , so @xmath109 is isomorphic to an indecomposable submodule of @xmath317 .
hence , @xmath109 is isomorphic to an indecomposable submodule of @xmath301 .
hence , ( e ) @xmath304 ( d ) as we wanted to show .
if @xmath8 is a quasi - hereditary algebra then [ t : equivalence](e ) says that the projective indecomposable module @xmath129 is also the injective hull of @xmath130 and the indecomposable ( partial ) tilting module corresponding to @xmath320 ; see [ s : schuralgebras ] .
[ c : kleshsocle ] suppose that @xmath99 is a right @xmath8-module such that @xmath321\ne0 $ ] only if @xmath117 .
then @xmath322=[\soc \fun(x):d^\lambda]\qquad \text { for all } { { \boldsymbol\lambda}}\in\k.\ ] ] in particular , if @xmath323 is a projective @xmath8-module then @xmath324=[\soc \fun(p):d^\lambda]$ ] , for all @xmath325 .
if @xmath326 is projective then @xmath324\ne0 $ ] only if @xmath117 by [ c : socle ] .
suppose , then , that @xmath99 is a right @xmath8-module such that @xmath321\ne0 $ ] only if @xmath117 . by exactness ,
we need to show that the reverse inclusion @xmath327 holds . by assumption ,
@xmath328 for some non - negative integers @xmath329 .
let @xmath330 be the injective hull of @xmath331 .
then @xmath332 since @xmath129 is self - dual for all @xmath124 by [ t : equivalence ] .
therefore we have injections @xmath333 applying the functor @xmath96 gives injections @xmath334 in particular , the socle of @xmath92 is contained in @xmath335 .
hence , @xmath336 and the corollary follows .
we conclude this section by relating hom - spaces of certain @xmath2-modules and @xmath8-modules .
these results do not assume that @xmath254 is a self - dual schur pair .
if @xmath217 is an algebra and @xmath99 is a subset of a left or right @xmath217-module , respectively , define the left and right * annihilators * of @xmath99 to be @xmath337 set @xmath338 and @xmath339 . if @xmath340 write @xmath341 and similarly for @xmath342 , @xmath343 and @xmath344 .
[ l : annihilators ] suppose that @xmath2 is a direct summand of @xmath5 and that @xmath38 . 1 .
if @xmath99 is a left @xmath8-submodule of @xmath5 then @xmath345 as vector spaces .
2 . if @xmath346 is a right @xmath2-submodule of @xmath5 then @xmath347 as vector spaces .
\(a ) if @xmath348 then the map @xmath349 is an @xmath8-module homomorphism .
conversely , if @xmath350 then @xmath351 since @xmath352 whenever @xmath353 .
it is straightforward to check that these maps are mutually inverse isomorphisms .
\(b ) the argument is almost identical to part ( a ) in that if @xmath354 then the map @xmath355 , for @xmath40 , is an @xmath2-module homomorphism . for the converse , however , we need to work harder because if @xmath356 then it is not immediate that @xmath357 since @xmath308 is only an @xmath2-module homomorphism and not necessarily an @xmath8-module homomorphism
. let @xmath358 .
we need to show that @xmath359 whenever @xmath360 .
that is , @xmath361 whenever @xmath362 .
fix @xmath360 and let @xmath363 be the endomorphism of @xmath5 given by @xmath364 , for @xmath365
. then @xmath366 , by [ t : doublecentraliser ] , so there exists an @xmath40 such that @xmath367 , for all @xmath365 .
therefore , @xmath368 .
[ p : ahoms ] suppose that @xmath2 is a direct summand of @xmath5 and let @xmath78 be the natural projection map .
let @xmath369 and @xmath370 .
then : 1 . there is a vector space isomorphism @xmath371 .
2 . there is an injection @xmath372 .
this map is an isomorphism if and only if @xmath373 .
\(a ) by [ l : annihilators](a ) , there is a vector space isomorphism @xmath374 .
we claim that @xmath375 .
suppose that @xmath376 .
then @xmath377 whenever @xmath378 so @xmath379 .
therefore , @xmath380 . on the other hand , since the algebra @xmath2 is self - injective , @xmath381 by . hence ,
@xmath382 as claimed .
\(b ) by part ( a ) , @xmath371 , whereas @xmath383 by [ l : annihilators](b ) . by definition , @xmath384 so there is a natural injection @xmath372 which is an isomorphism if and only if @xmath385 .
the main results of this paper apply the results of [ s : schurpairs ] to the cyclotomic schur algebras @xmath7 , which were introduced in @xcite . in order to do this
we first need to produce a schur pair @xmath4 that defines the cyclotomic schur algebras . using the notation of [ d : schurpair ]
, the algebra @xmath2 will be a cyclotomic hecke algebra @xmath6 and @xmath5 will be a direct sum of permutation modules for @xmath6 . this section recalls the representation theory of the cyclotomic hecke algebras of type @xmath2 and proves some new results about the simple modules of these algebras that will be needed later .
fix a non - negative integer @xmath386 .
let @xmath387 be the symmetric group on @xmath386 letters and let @xmath388 be the standard set of coxeter generators for @xmath387 , where @xmath389 for @xmath390 . if @xmath391 then the * length * of @xmath392 is @xmath393
. fix a field @xmath10 and @xmath394 . if @xmath395 then the * @xmath396-quantum integer * is @xmath397=[k]_\xi = \begin{cases } \phantom{-}\xi+\xi^3+\dots+\xi^{2k-1},&\text{if } k\ge0,\\ -\xi^{-1}-\xi^{-3}-\dots-\xi^{2k+1},&\text{if } k<0 . \end{cases}\ ] ] so , @xmath398_\xi=-[-k]_{\xi^{-1}}$ ] and if @xmath399 then @xmath398=(\xi^{2k}-1)/(\xi-\xi^{-1})$ ] .
the * quantum characteristic * of @xmath396 is the smallest positive integer @xmath82 such that @xmath400_\xi=0 $ ] , where we set @xmath401 if @xmath398\ne0 $ ] for all @xmath402 .
notice that @xmath396 and @xmath403 have the same quantum characteristic and if @xmath404 then the quantum characteristic of @xmath396 is the characteristic of @xmath10 .
finally fix an integer @xmath405 and a * multicharge * @xmath406 .
[ d : hecke ] the * cyclotomic hecke algebra * of type @xmath2 with hecke parameter @xmath396 and multicharge @xmath407 is the unital associative @xmath10-algebra @xmath408 with generators @xmath409 , @xmath410 and relations @xmath411)&=0 , & ( t_r-\xi)(t_r+\xi^{-1})&=0 , \\
l_rl_t&=l_tl_r , & t_rt_s&=t_st_r & \text{if } |r - s|>1,\\ t_st_{s+1}t_s&=t_{s+1}t_st_{s+1 } , & t_rl_t&=l_tt_r,&\text{if } t\ne r , r+1,\\ \span\span l_{r+1}=t_rl_rt_r+t_r,\span\span\span \end{aligned}\ ] ] where @xmath390 , @xmath412 and @xmath413 .
these algebras are almost the same as the ariki - koike algebras introduced in @xcite except that the presentation of [ d : hecke ] changes the algebras when @xmath9 .
this allows the so - called _ degenerate _ ( @xmath9 ) and _ non - degenerate _ ( @xmath399 ) cases , to be treated simultaneously .
see @xcite for more details .
[ d : hecke ] is a renormalisation of the presentation of the cyclotomic hecke algebras given in @xcite .
explicitly , if @xmath414 and @xmath415 are the generators of the algebra @xmath416 given in ( * ? ? ?
* definition 2.2 ) then @xmath417 and @xmath418 , for @xmath390 and @xmath419 .
( in the notation of @xcite , the cyclotomic parameters of @xmath416 are @xmath420 , for @xmath421 . ) if @xmath391 has reduced expression @xmath422 , so that @xmath423 , then set @xmath424
. then @xmath425 depends only on @xmath392 and @xmath426 is a @xmath10-basis of @xmath6 by the argument of @xcite .
let @xmath427 be the unique anti - isomorphism of @xmath6 that fixes each of the generators .
then @xmath428 and @xmath429 , for all @xmath391 and @xmath430 .
[ t : symmetricalgebra ] the algebra @xmath6 is a symmetric algebra with non - degenerate trace form @xmath431 . in particular , @xmath6 is a self - injective algebra .
the paper @xcite proves this result when @xmath399 whereas @xcite treats the case when @xmath9 . in both cases the trace form @xmath431 is described explicitly , however , dual bases are known only when @xmath432 .
a * partition * of an integer @xmath433 is a weakly decreasing sequence @xmath434 of non - negative integers such that @xmath435 .
a * multipartition * , or @xmath436-partition , of @xmath386 is an ordered @xmath436-tuple @xmath437 of partitions such that @xmath438 . the * diagram * of @xmath439 is the set @xmath440=\set{(k , r , c)|r\ge 1 , 1\le c\le\lambda^{(k)}_r \text { and } 1\le k\le \ell}.\ ] ] a multipartition is uniquely determined by its diagram .
a * node * is any element of the diagram of some multipartition .
let @xmath441 be the set of multipartitions of @xmath386 .
if @xmath442 then @xmath439 * dominates * @xmath443 , written @xmath444 , if @xmath445 dominance defines a partial order on @xmath441 . if @xmath99 is a set then an @xmath99-valued @xmath439-tableau is a function @xmath446}x$ ] .
if @xmath447 is a @xmath439-tableau write @xmath448 . for convenience
we identify @xmath449 with a labeling of the diagram @xmath450 $ ] by elements of @xmath99 in the obvious way .
thus , we can talk of the components , rows and columns of @xmath447 .
a * standard @xmath439-tableau * is a map @xmath451}\{1,2,\dots , n\}$ ] such that for @xmath452 the entries in each row of @xmath453 increase from left to right and the entries in each column of @xmath453 increase from top to bottom .
let @xmath454 be the set of standard @xmath439-tableaux and set @xmath455 if @xmath456 and @xmath430 let @xmath457 if @xmath433 appears in the @xmath458-th component of @xmath459 .
the * conjugate * of @xmath439 is the multipartition @xmath460 whose diagram is obtained from @xmath450 $ ] by reversing the order of the components and then swapping rows and columns .
thus , @xmath461=\set{(k , r , c)|(\ell - k+1,c , r)\in[{{\boldsymbol\lambda}}]}$ ] .
the * conjugate * of @xmath459 is the @xmath460-tableau @xmath462 obtained by reversing the order of the components in @xmath459 and then transposing the tableau in each component ( that is , swapping rows and columns ) .
let @xmath463 be the standard @xmath439-tableau such that the entries in @xmath463 increase from left to right along the rows of @xmath464 in order .
similarly , let @xmath465 be the standard @xmath439-tableau with entries increase from top to bottom down the columns of @xmath466 in order
. then @xmath467 , for all @xmath468 .
if @xmath459 is a standard @xmath439-tableau let @xmath469 be the unique permutations such that @xmath470 and @xmath471 .
let @xmath472 .
it is easy to see that @xmath473 and @xmath474 , for all @xmath475 .
if @xmath459 is a tableau with entries in @xmath476 let @xmath477 and @xmath478 be the subgroups of @xmath387 that stabilise the rows and columns of @xmath459 , respectively . if @xmath468 let @xmath479 be the corresponding * young subgroup * , or * parabolic subgroup * , of @xmath387 . in particular , @xmath480 and @xmath481 , where @xmath482 and @xmath483 are the natural subgroups of @xmath387 associated with the multipartitions @xmath439 and @xmath460 . for @xmath468
define @xmath484 , where @xmath485 ) \quad\text{and}\quad x^{{\boldsymbol\lambda}}=\sum_{w\in\row(\tlam)}\xi^{\ell(w)}t_w.\ ] ] let @xmath486 and set @xmath487 .
below we define the cyclotomic schur algebra to be the endomorphism algebra of @xmath5 .
fix @xmath468 and define @xmath488 , for @xmath489 . by ( * ? ? ?
* theorem 3.26 ) , @xmath490 is a cellular basis of @xmath6 , where @xmath441 is ordered by dominance .
consequently , if @xmath491 is the @xmath10-subspace of @xmath6 spanned by @xmath492 , then @xmath491 is a two - sided ideal of @xmath6 .
invoking the theory of cellular algebras @xcite , for @xmath468 there is a right cell module @xmath493 , called a * specht module*. by definition , @xmath494 .
set @xmath495 , for @xmath475 .
then @xmath496 is a @xmath10-basis for @xmath493 .
the specht module comes equipped with an associative bilinear form @xmath497 that is determined by @xmath498 let @xmath499 .
then @xmath500 is an @xmath6-submodule of @xmath493 .
we need a parallel construction for the dual specht modules
. for @xmath468 set @xmath501 , where @xmath502 ) \quad\text{and}\quad
x_{{\boldsymbol\lambda}}=\sum_{w\in\col(\tllam)}(-\xi)^{-\ell(w)}t_w.\ ] ] for @xmath489 set @xmath503 . by (
* ( 2.7 ) ) , @xmath504 is a second cellular basis of @xmath6 , where @xmath441 is ordered by _ reverse _ dominance . for each multipartition
@xmath468 the theory of cellular algebras gives us a * dual specht module * @xmath505 . here , @xmath506 is the two - sided ideal of @xmath6 with basis @xmath507 much as before , set @xmath508 .
[ r : nbasislabels ] the labelling that we are using for the @xmath386-basis is conjugate to that used in @xcite so @xmath509 should be replaced with @xmath510 , and @xmath511 with @xmath512 , when comparing with these papers .
our notation reflects that fact that the elements @xmath513 and @xmath514 come from looking at the row and column stabilizers of @xmath463 and @xmath465 , respectively .
similarly , our labelling of the dual specht modules follows the same convention , which is in agreement with the papers @xcite . in @xcite the module @xmath515 is written as @xmath516 . implicitly , the definitions of the elements @xmath513 , @xmath514 , @xmath517 and @xmath511 all depend upon the choice of hecke parameter @xmath396 and the multicharge @xmath407
. this remark will be important below when we vary these parameters . for future use we summarise the properties of these modules that follow directly from the general theory of cellular algebras .
[ t : cellularalgebras ] suppose that @xmath10 is a field and @xmath518 . 1 .
the @xmath519 is a complete set of pairwise non - isomorphic irreducible @xmath6-modules .
moreover , @xmath520 .
the @xmath521 is a complete set of pairwise non - isomorphic irreducible @xmath6-modules .
moreover , @xmath522 .
for the main results in this paper we need to describe the specht modules and dual specht modules as submodules , which is already known , and we need to determine the isomorphisms between the two sets of simple modules given by [ t : cellularalgebras ] . as remarked above , if @xmath523 then no pairs of dual bases for @xmath6 are known .
the following fundamental result implies that the two bases @xmath524 and @xmath525 are dual bases of @xmath6 `` modulo higher terms '' .
[ p : mndual ] suppose that @xmath526 are pairs of tableaux of the same shape .
then @xmath527 and @xmath528 only if @xmath529 .
similarly , @xmath530 and @xmath530 only if @xmath531 . for @xmath468 set @xmath532 and @xmath533 . observe that @xmath534 since @xmath535 .
moreover , @xmath536 and @xmath537 by [ p : mndual ] .
[ l : spechtsubmodules ] suppose that @xmath468 .
then @xmath538 and @xmath539 as right @xmath6-modules . to prove this
observe that , because @xmath540 , there is a well - defined @xmath6-module homomorphism @xmath541 given by @xmath542 , for @xmath543 . to complete the proof it remains to check that @xmath544 is a basis of @xmath545 .
see ( * ? ? ?
* theorem 2.9 ) or ( * ? ? ? * proposition 3.13 ) for details .
the proof that @xmath546 is similar . as their names suggest , the specht modules and dual specht modules are dual to each other .
the proof of this requires the trace form @xmath431 from [ t : symmetricalgebra ] and a strenthening of [ p : mndual ] .
[ t : nonvashingtrace ] suppose that @xmath468 .
then @xmath547 we can now prove that @xmath493 and @xmath515 are dual to each other .
[ c : dualspechts ] let @xmath468 .
then @xmath548 and @xmath549 .
this result is proved in @xcite but we give a proof of the isomorphism @xmath549 because we need the details below . for @xmath550
let @xmath551 be the linear map determined by @xmath552 , for all @xmath475 .
now define a linear map @xmath553 by @xmath554 , for all @xmath550 . then @xmath199 is a vector space isomorphism by [ p : mndual ] and [ t : nonvashingtrace ] . for @xmath543 write @xmath555 , for @xmath556 .
fix @xmath475 . using [ p : mndual ] ,
@xmath557 hence , @xmath199 is an @xmath6-module homomorphism , completing the proof . in [ s : tilting ] below we need an analogue of [ c : dualspechts ] relating the simple modules @xmath558 and @xmath559 , for @xmath560 . to establish this we use the following characterisation of the simple @xmath6-modules together with the `` dual '' algebras @xmath561 that are introduced below .
this description of the simple @xmath6-modules as submodules of @xmath6 generalises a remark made by james @xcite for the symmetric groups .
[ l : simplesubmodules ] suppose that @xmath562 . 1 .
the simple module @xmath563 if and only if @xmath564 .
the simple module @xmath565 if and only if @xmath566 . moreover , if @xmath563 then @xmath567 and if @xmath565 then @xmath568 . we consider only the claims for @xmath558 since those for @xmath569 can be proved in the same way . as remarked after [ l : spechtsubmodules ] , the map @xmath570 determined by @xmath571 , for @xmath572 , is an isomorphism .
if @xmath573 then using the definitions @xmath574 where the third equality follows from [ e : bilinearform ] because @xmath575 by [ p : mndual ] .
hence , @xmath563 if and only if @xmath564 , proving ( a ) .
for the second claim , suppose that @xmath563 .
we need to prove that @xmath567 . combining [ l : spechtsubmodules ] and [ c : dualspechts ] , there are @xmath6-module isomorphisms @xmath576 where @xmath577 , @xmath578 and @xmath579 , for @xmath580 and @xmath581 , where @xmath582 is defined in the proof of [ c : dualspechts ]
. moreover , @xmath583 and @xmath199 are both isomorphisms .
let @xmath584 .
we claim that , up to a non - zero scalar , @xmath585 is the map @xmath586 induced by the inner product @xmath587 on @xmath588 . to see this fix @xmath580 and write @xmath589 as @xmath590 is a cellular basis , with @xmath441 ordered by reverse dominance
, it follows that @xmath591 only if @xmath592 with equality only if @xmath593 . if @xmath594 set @xmath595 . using [ p : mndual ] for the third equality , @xmath596 therefore , if @xmath572 then , using [ p : mndual ] and [ e : bilinearform ] again , @xmath597 by [ t : nonvashingtrace ] , @xmath598 is a non - zero element of @xmath10 .
hence , up to an invertible scalar , the map @xmath585 coincides with the natural map induced by the bilinear form on @xmath588 . in particular , if @xmath563 then @xmath599 . by definition , @xmath600 so this completes the proof . to identify the isomorphic simple modules @xmath558 and @xmath559 we need to shift to a `` dual '' hecke algebra . as we make precise in [ t : ringelduality ] below , this is a shadow of ringel duality in the hecke algebra world .
let @xmath601 be the cyclotomic hecke algebra with hecke parameter @xmath403 and multicharge @xmath602 . to help distinguish between the elements of the algebras @xmath6 and @xmath561
let @xmath603 be the generators of @xmath561 .
more generally , we decorate all elements of @xmath561 with an appropriately placed @xmath604 . for example , @xmath605 and @xmath606 both belong to @xmath561 . similarly , there are cellular bases @xmath607 and @xmath608 that give rise to specht modules @xmath609 and dual specht modules @xmath610 , for @xmath468 .
let @xmath611 and @xmath612 .
the next result , which does not appear to be in the literature , generalises the sign automorphism on the group algebra of a symmetric group .
we leave the proof as an exercise to the reader because it follows by simply inspecting the relations in the two algebras @xmath6 and @xmath561 .
( recall from after [ e : qint ] that @xmath398_\xi=-[-k]_{\xi^{-1}}$ ] , for all @xmath395 . )
[ l : hashinvolution ] there is a unique algebra isomorphism @xmath613 such that @xmath614 moreover , @xmath615 , @xmath616 , @xmath617 , @xmath618 , for all @xmath468 and all @xmath489 .
the isomorphism @xmath619 induces an equivalence of categories @xmath620 .
to describe the effect of @xmath621 on the simple @xmath561-modules define @xmath622 recall from after [ e : qint ] that @xmath396 and @xmath403 both have quantum characteristic @xmath82 . by the results in @xcite , @xmath623 if and only if @xmath439 is a * kleshchev * , or * restricted * , multipartition with respect to @xmath624 and @xmath625 if and only if @xmath439 is kleshchev with respect to @xmath626 .
the kleshchev multipartitions index the crystal graphs of certain integrable highest weight modules , so their definition is explicit but recursive @xcite .
the crystal graph combinatorics shows that there is a crystal isomorphism , known as the * mullineux map * , @xmath627 ( using the set @xmath157 defined in [ s : fayersconjecture ] , if @xmath628 labels a path from @xmath629 to @xmath443 in the crystal graph @xmath630 then @xmath631 labels a path from @xmath632 to @xmath633 in the crystal graph @xmath634 .
this property determines the map @xmath635 uniquely
. see ( * ? ? ?
* theorem 4.12 ) or @xcite for more details . )
the next result connects the mullineux map @xmath635 with the representation theory of @xmath6 .
[ p : mullineux ] suppose that @xmath468 and @xmath636 .
then @xmath637 as right @xmath561-modules , for graded specht modules the first two isomorphisms are proved in ( * ? ? ?
* theorem 8.5 ) , although it is not completely clear that the isomorphism in [ l : hashinvolution ] agrees with the homogeneous sign isomorphism considered in @xcite .
fortunately , these two isomorphisms follow directly from the definitions because @xmath638 and @xmath639 by [ l : hashinvolution ] .
the isomorphism @xmath640 , together with the modular branching rules @xcite and a standard argument due to kleshchev , now imply that @xmath641 .
for the corresponding result for the graded simple modules see ( * ? ? ?
* theorem 3.6.6 ) and ( * ? ? ? * ( 3.53 ) ) .
we can now prove the promised comparison results for the simple @xmath6-modules .
[ c : simpleisomorphism ] suppose that @xmath562 .
then @xmath565 if and only if @xmath642 . moreover , if @xmath636 then @xmath643 as @xmath6-modules . by [ l : simplesubmodules](b ) and [ l : hashinvolution ] , @xmath565 if and only if @xmath644 , so that @xmath565 if and only if @xmath642 . hence , using [ p : mullineux ] twice , if @xmath636 then @xmath645 hence , @xmath646 as required .
[ c : hdmu ] suppose that @xmath562 .
then @xmath647 if and only if @xmath648 . moreover , if @xmath648 then @xmath649 as @xmath561-modules . by [ l : simplesubmodules](a ) and [ l : hashinvolution ] , @xmath650 if and only if @xmath651 , so @xmath652 if and only if @xmath648 .
now suppose that @xmath648 .
then @xmath653 where the second isomorphism follows by [ p : mullineux ] .
[ c : simplesocles ] suppose that @xmath636
. then @xmath654 .
if @xmath636 then @xmath655 .
moreover , @xmath656 by [ c : simpleisomorphism ] . now take duals using [ c : dualspechts ] ( and [ t : cellularalgebras ] ) .
we are now ready to introduce the cyclotomic schur algebras , which are one of the main objects of study in this paper .
our first goal is to describe which simple modules appear in the socle of the weyl modules , tilting modules and the projective indecomposable modules and , as an application describe the socles of the young modules . in the last section we saw that the specht modules and dual specht modules came from two different cellular bases for @xmath6 and we needed to introduce the algebra @xmath561 to understand the difference in the labelling of the simple modules coming from these two bases .
similarly , in this section we introduce two variations of the cyclotomic schur algebras which we will use to translate information about the socle of a tilting module to the socle of a projective module .
as a prelude to defining the schur algebras , for @xmath468 let @xmath657 then @xmath658 , @xmath659 , @xmath5 and @xmath21 are @xmath6-modules and @xmath660 , @xmath661 , @xmath662 and @xmath663 are @xmath561-modules . [ p : schurpairs ] suppose that @xmath518 .
then @xmath664 , @xmath665 , @xmath666 and @xmath667 are self - dual schur pairs . by [ t : symmetricalgebra ] , @xmath6 and @xmath561 are both symmetric algebras and so , in particular , they are self - injective . moreover , since @xmath668 as right @xmath6-modules , @xmath5 and @xmath21 are both faithful @xmath6-modules . therefore , @xmath664 and @xmath665 are both schur pairs , similarly , @xmath666 and @xmath667 are both schur pairs . as algebras @xmath6 and @xmath561
are both cellular they have an anti - isomorphism , which is the unique anti - isomorphism that fixes each of the generators of @xmath6 and of @xmath561 , respectively .
moreover , by [ t : cellularalgebras ] , the simple @xmath6-modules and the simple @xmath561-modules are self - dual with respect to these involutions . to show that @xmath664 , @xmath665 , @xmath666 and @xmath667 are self - dual schur pairs it is enough to show that
each of the modules @xmath658 , @xmath659 , @xmath660 and @xmath661 is self - dual , for @xmath468 .
this is proved in ( * ? ? ?
* proposition 5.13 ) .
( the careful reader will notice that ( * ? ? ?
* proposition 5.13 ) assumes that the cyclotomic parameters @xmath669 are invertible . as explained in @xcite , because of [ d : hecke ] this condition translates into the vacuous requirement that @xmath670(\xi-\xi^{-1})+1=\xi^{2\kappa_r}$ ] is invertible in @xmath10 , for @xmath421 .
hence , the requirement from @xcite that @xmath669 be invertible is satisfied . )
consequently , all of the results in [ s : schurpairs ] apply to the endomorphism algebras @xmath24 , where @xmath4 is one of the schur pairs given in [ p : schurpairs ] .
notice that , by definition , @xmath6 is a direct summand of both @xmath5 and @xmath21 since @xmath671 as right @xmath6-modules .
similarly , @xmath561 is a direct summand of @xmath662 and of @xmath663 .
[ d : schur ] the * cyclotomic schur algebra * and the * twisted cyclotomic schur algebra * of @xmath6 are the endomorphism algebras @xmath672 the cyclotomic schur algebras include , as special cases , the classical schur algebras studied by green @xcite and the @xmath1-schur algebras introduced by dipper and james @xcite . just as our definition of @xmath6 differs slightly from the definition given by ariki and koike @xcite , the algebra @xmath673 is isomorphic to a cyclotomic @xmath396-schur algebra of @xcite when @xmath399 and it is isomorphic a _ degenerate _
cyclotomic schur algebra @xcite when @xmath9 .
we are mainly interested in the cyclotomic schur algebra @xmath674 , however , we need the algebras @xmath675 to prove some of our results about @xmath7-modules . given [ p : schurpairs ] it is also natural to consider the algebras @xmath676 and @xmath677 .
in fact , these algebras give no additional information because [ p : mullineux ] readily implies the following .
[ l : schurequivs ] if @xmath468 then @xmath678 and @xmath679 as right @xmath6-modules .
therefore , @xmath680 and @xmath681 .
we leave the proof of [ l : schurequivs ] to the reader .
in fact , as we recall below , @xmath682 and @xmath675 are quasi - hereditary algebras with weight posets @xmath441 and @xmath683 , respectively . similarly , @xmath684 and
@xmath685 are quasi - hereditary with weight posets @xmath683 and @xmath441 , respectively . the isomorphisms @xmath686 and @xmath687 of [ l : schurequivs ] are both isomorphisms of quasi - hereditary algebras .
the algebras @xmath7 and @xmath688 are finite dimensional cellular and quasi - hereditary @xmath10-algebras with simple modules labelled by @xmath441 . to describe cellular bases of @xmath7 and @xmath688 we need some more combinatorics .
fix @xmath562 and @xmath468 .
a @xmath439-tableau of * type * @xmath443 is a map @xmath446}\set{(k , r)|1\le k\le\ell\text { and } r\ge1}$ ] such that @xmath689|\bt(\alpha)=(k , r)}$ ] , for all @xmath690 and @xmath691 .
order the pairs @xmath692 lexicographically .
following ( * ? ? ?
* definition 4.4 ) , a @xmath439-tableau @xmath447 of type @xmath443 is * row semistandard * if 1 .
the entries in @xmath447 are weakly increasing along rows , 2 .
the entries in @xmath447 are strictly increasing down columns , 3 . if @xmath693 $ ] then @xmath694 .
let @xmath695 be the set of semistandard @xmath439-tableaux of type @xmath443 .
there is a bijection @xmath696 given by replacing each entry @xmath433 in a tableau with @xmath697 , where @xmath698 .
let @xmath699 be the set of * column semistandard * @xmath439-tableaux of type @xmath443 , where conjugation reverses the order of components , swaps rows and columns and replaces each entry @xmath700 with @xmath701 .
( we need these tableau in [ s : homomorphisms ] . )
if @xmath702 then the entries in @xmath447 weakly increase down columns , strictly increase along rows and if @xmath703 $ ] then @xmath704 where , throughout , the pairs @xmath692 are ordered by @xmath705 if @xmath706 or @xmath707 and @xmath708 . by construction , if @xmath702 then @xmath447 is a tableau of type @xmath443 .
observe that @xmath709 only if @xmath444 .
similarly , @xmath710 only if @xmath711 .
let @xmath712 and @xmath713 .
for @xmath475 let @xmath714 be the tableau of type @xmath443 obtained from @xmath459 by replacing each entry @xmath433 in @xmath459 with @xmath700 if @xmath433 appears in _ row _
@xmath32 of the @xmath458th component of @xmath715 and define @xmath716 similarly except that we use the _ column _ index of @xmath433 in @xmath717 instead .
it is not hard to see that every semistandard tableau of type @xmath443 is equal to @xmath718 and to @xmath716 , for some @xmath489 , although the converse is not true in general .
let @xmath719 and @xmath720 , respectively , be the unique row and column semistandard @xmath439-tableau of type @xmath439 . as a quick exercise in these definitions , @xmath721 . by ( * ?
* theorem 4.14 ) , @xmath722 has @xmath10-basis @xmath723 , where @xmath724 extending this idea , following @xcite , if @xmath725 and @xmath726 define @xmath727 then @xmath728 by the remarks above .
therefore , the map @xmath729 defined by @xmath730 , for all @xmath543 , is an @xmath6-module homomorphism .
we consider @xmath731 as an element of @xmath7 . by
( * ? ? ? * theorem 6.6 ) , @xmath732 is a cellular basis of @xmath7 .
for each @xmath468 the algebra @xmath7 has a * weyl module * @xmath733 , which is the corresponding right cell module of @xmath7 . to make this more explicit ,
let @xmath734 be the two - sided ideal of @xmath7 with basis the @xmath731 where @xmath735 and @xmath447 are semistandard @xmath443-tableau with @xmath736 . then @xmath737 . if @xmath738 then set @xmath739 .
then @xmath740 is a @xmath10-basis of @xmath741 .
exactly as for a specht module , the weyl module @xmath741 comes equipped with a bilinear form , also written as @xmath742 , determined by @xmath743 let @xmath744 where @xmath745 is the radical of the form @xmath742 . by definition
, @xmath746 is the identity map on @xmath658 , so @xmath747 and , consequently , @xmath748 . in a similar way we can define elements @xmath749 and homomorphisms @xmath750 , for semistandard tableaux @xmath751 .
hence , we can define weyl modules @xmath752 and simple modules @xmath753 for @xmath688 .
the cellular algebra involutions on @xmath6 and @xmath561 induce involutions on @xmath7 and @xmath688 with the property that @xmath754 and @xmath755 .
for @xmath468 let @xmath756 and @xmath757 be * costandard modules * for @xmath7 and @xmath688 , respectively .
[ t : cellularsalgebras ] suppose that @xmath518 . 1 .
the algebra @xmath7 is a quasi - hereditary cellular algebra with weight poset @xmath441 , cellular basis @xmath758 standard modules @xmath759 , costandard modules @xmath760 and pairwise non - isomorphic self - dual simple modules @xmath761 .
the algebra @xmath688 is a quasi - hereditary cellular algebra with weight poset @xmath683 , cellular basis @xmath762 standard modules @xmath763 , costandard modules @xmath764 and pairwise non - isomorphic self - dual simple modules @xmath765 .
although we will not need them , there are analogous definitions of elements @xmath766 and @xmath767 and maps @xmath768 and @xmath769 , where @xmath735 and @xmath447 are _ column _ semistandard tableaux .
then [ l : schurequivs ] and[t : cellularsalgebras ] imply that @xmath770 and @xmath771 are cellular bases for the quasi - hereditary algebras @xmath684 and @xmath685 , respectively .
let @xmath82 be the natural projection from @xmath5 onto @xmath6 . recall that @xmath664 is a schur pair . by the results in [ s : schurpairs ] , we have that @xmath772 in particular
, we can regard @xmath773 as a right @xmath774-module .
similar results hold for the schur pair @xmath775 . in what follows
we consider @xmath776 to be an element of @xmath777 and @xmath778 as an element of @xmath779 .
the next result shows how the classical definitions of weyl modules extend to the cyclotomic case . in the semisimple case
this follows directly from [ l : ideals ] and [ l : spechtsubmodules ] but in general we need to work harder . in the special case
when @xmath780 this is due to dipper and james @xcite .
[ p : weylsubmodules ] suppose that @xmath468 . 1 . as @xmath7-modules , @xmath781 .
in particular , @xmath741 is ( isomorphic to ) a submodule of @xmath773 .
2 . as @xmath688-modules , @xmath782 . in particular , @xmath752 is ( isomorphic to ) a submodule of @xmath783 .
we prove ( a ) and leave ( b ) for the reader .
recall that @xmath784 .
define an @xmath7-module homomorphism by @xmath785 fix semistandard tableaux @xmath786 , for some @xmath562
. then @xmath787 by [ p : mndual ] , if @xmath573 then @xmath788 only if @xmath444 .
consequently , @xmath789 if @xmath790 .
in particular , @xmath791 since @xmath792 is a basis of @xmath734 . on the other hand , if @xmath793 then @xmath794 . in view of [ l : spechtsubmodules ] , @xmath795 is linearly independent ( see also ( * ? ? ? * ( 2.10 ) ) or ( * ? ? ? * proposition 5.9 ) ) . therefore , @xmath796 as required . the last result in this section
explains the significance of the algebras @xmath797 in the representation theory of @xmath7 ( and hence why we need them in this paper ) . to do this
we first recall the definition of the ringel dual of a quasi - hereditary algebra .
let @xmath8 be a quasi - hereditary algebra with standard modules @xmath798 and costandard modules @xmath799 , where @xmath56 runs over a poset @xmath800 .
let @xmath801 be the full subcategory of @xmath802-filtered @xmath8-modules.thus , @xmath803 if and only if @xmath99 has a filtration with each subquotient isomorphic to a weyl module @xmath798 , for @xmath804 .
similarly , let @xmath805 be the full subcategory of @xmath806-filtered @xmath8-modules . if @xmath803 let @xmath807 be the number of subquotients of @xmath99 isomorphic to @xmath798 .
define @xmath808 in the same way when @xmath809 .
since @xmath8 is quasi - hereditary the multiplicities @xmath807 and @xmath808 are independent of the choices of @xmath802 and @xmath806 filtrations . an * @xmath8-tilting module * is any @xmath8-module in @xmath810 . as @xmath8 is quasi - hereditary , by (
* a4 ) for each @xmath804 there is a unique indecomposable tilting module @xmath811 for @xmath8 such that @xmath812 and @xmath813 only if @xmath814 .
moreover , up to isomorphism @xmath815 is a complete set of pairwise non - isomorphic indecomposable tilting modules .
a * full tilting module * is a tilting module that contains @xmath816 as a summand .
a * ringel dual * of the algebra @xmath8 is any algebra @xmath817 , where @xmath818 is any full tilting module .
then @xmath819 is quasi - hereditary with respect to the opposite poset @xmath820 . by construction ,
the ringel dual is unique up to morita equivalence .
there is an exact functor @xmath821 given by @xmath822 .
returning now to the cyclotomic schur algebras , let @xmath823 and @xmath824 be the tilting modules for @xmath7 and @xmath688 , for @xmath468 .
let @xmath825 and @xmath826 be the projective covers of @xmath827 and @xmath827 , respectively .
let @xmath828 and @xmath829 be the schur functors , defined in [ s : schurpairs ] . [
t : ringelduality ] the twisted cyclotomic schur algebra @xmath688 is ringel dual to @xmath7 .
moreover , there is an exact functor @xmath830 such that the following diagram of functors commutes : ( m)[matrix of math nodes , row sep=1cm , column sep=16 mm ] & + _ n & _
n. + ; ( m-1 - 1)node[above]@xmath831(m-1 - 2 ) ; ( m-2 - 1)node[below]@xmath621(m-2 - 2 ) ; ( m-1 - 1)node[left]@xmath832(m-2 - 1 ) ; ( m-1 - 2)node[right]@xmath833(m-2 - 2 ) ; the functor @xmath831 is determined by @xmath834 , for all @xmath468 .
moreover , @xmath835 as @xmath7-modules . by (
* corollary 7.3 ) , there is a full tilting module @xmath216 for @xmath7 such that @xmath836 . by [ l : schurequivs ] , @xmath837 so @xmath688 is a ringel dual of @xmath7 .
note that @xmath838 by [ l : ee * ] .
the isomorphism @xmath839 induces an equivalence of categories @xmath840 , so ringel duality give an exact functor @xmath841 that sends indecomposable tilting modules to indecomposable projective modules .
by induction on the dominance order it follows that @xmath842 , for all @xmath468 .
since @xmath831 is exact on @xmath843 , it is uniquely determined by its action on the costandard modules .
in particular , @xmath835 since @xmath844 is a quotient of @xmath823 . it remains to check that the diagram above commutes . by exactness
it is sufficient to check commutativity on the costandard @xmath7-modules . by [ c : schurfunctor ] , [ l : spechtsubmodules ] and [ p : weylsubmodules ] , if @xmath468 then @xmath845 and @xmath846 .
therefore , as @xmath6-modules , @xmath847 where the third isomorphism comes from [ p : mullineux ] .
it is worth mentioning that the ringel duality functor @xmath831 , when considered a functor from @xmath848 to @xmath849 , is _ left _ only exact .
we restrict to the subcategory @xmath850 in [ t : ringelduality ] only because this ensures that @xmath831 is exact . following @xcite , and [ d : youngmodules ] , we make the following definition .
the * young modules * and the * twisted young modules * are the @xmath6-modules @xmath851 and @xmath852 , respectively , for @xmath468 . similarly , define young modules for @xmath561 by setting @xmath853 and @xmath854 .
we need ringel duality for the next corollary , which we will use in [ s : tilting ] to prove results about the socles of young modules for @xmath6 and projective modules for @xmath7 .
[ c : equivyoungmodules ] suppose that @xmath468 .
then @xmath855 as @xmath561-modules . using the definitions and [ t : ringelduality ] , as @xmath561-modules , @xmath856 as required
. let @xmath857 be the injective hull of @xmath827 , for @xmath468 . then @xmath858 .
[ c : ] suppose that @xmath468
. then @xmath859 .
the schur functor commutes on the dualities on @xmath860 and @xmath861 and @xmath862 by ( * ? ? ?
* corollary 5.14 ) ( when @xmath623 this follows from [ l : yselfdual ] ) .
moreover , ringel duality sends injective modules to tilting modules .
therefore , @xmath863 the final result in this section is part of the folklore for @xmath6 but as far as we are aware the result is not proved in the literature .
suppose that @xmath562 .
then : 1 . if @xmath636 then @xmath864 is the projective cover of @xmath558 .
2 . if @xmath642 then @xmath865 is the projective cover of @xmath866 .
3 . if @xmath636 then @xmath867 as @xmath6-modules . by [ p : youngmodules ]
, @xmath864 is projective if and only if @xmath636 , in which case it is the projective cover of @xmath558 .
this proves ( a ) .
similarly , @xmath868 is projective if and only if @xmath642 , in which case it is the projective cover of @xmath869 .
hence , ( b ) follows by [ c : equivyoungmodules ] and [ c : simpleisomorphism ] .
part ( c ) is now automatic from parts ( a ) and ( b ) .
we now come to the first main result of this paper , which motivated much of the development of [ s : schurpairs ] . in more detail
, we describe the simple @xmath7-modules that can appear in the socles of the weyl modules , tilting modules and projective indecomposable modules and give similar results for the socles of the ( twisted ) young modules .
before we begin we note the following immediate consequence of [ p : schurpairs ] and [ t : equivalence ] .
[ t : snequivalence ] suppose that @xmath104 .
then the following are equivalent : 1 .
@xmath117 , 2 .
@xmath300 , 3 .
@xmath110 is a right submodule of @xmath301 , 4 .
@xmath109 is a direct summand of @xmath301 , 5 .
@xmath109 is a projective - injective right @xmath8-module , 6 .
@xmath870 is an indecomposable tilting model , 7 .
@xmath109 is self - dual .
there are analogous results for @xmath688-modules .
the first of these results generalises a classical result of james ( * ? ? ?
* theorem 2.8 ) in level one ( when @xmath9 and @xmath780 ) .
this result will be used to prove fayers conjecture in [ t : fayers ] below .
[ t : deltasocle ] suppose that @xmath442 . 1 .
the simple module @xmath871 is a submodule of @xmath733 only if @xmath636 .
2 . the simple module @xmath872 is a submodule of @xmath873 only if @xmath874 .
both parts can be proved in the same way so we consider only ( a ) . by [ p : schurpairs ] , @xmath875 is a schur pair .
therefore , by [ t : socle ] , @xmath876 is an @xmath7-submodule of @xmath5 if and only if @xmath563 , which is if and only if @xmath636 . on the other hand
, @xmath741 is isomorphic to an @xmath7-module of @xmath5 by [ p : weylsubmodules ] , so @xmath877 .
hence , @xmath876 is a submodule of @xmath741 only if @xmath636 as claimed .
[ c : ptsocle ] suppose that @xmath878 and @xmath442 .
then @xmath876 is a submodule of @xmath99 only if @xmath623 . in particular
, @xmath876 is a submodule of @xmath879 only if @xmath636 . by [ p : weylsubmodules ] and [ l : ideals ] , or working directly with the definitions , @xmath880 for all @xmath468 .
hence , applying the schur functor to [ t : deltasocle ] gives the following .
[ c : soc5 ] suppose that @xmath442 .
then @xmath881=[\soc s^{{\boldsymbol\lambda}}:d^{{\boldsymbol\mu } } ] \qquad\text{and}\qquad [ \soc t^{{\boldsymbol\lambda}}:l^{{\boldsymbol\mu}}]=[\soc y_{{\boldsymbol\lambda}}:d^{{\boldsymbol\mu}}].\ ] ] in particular , these two multiplicities are non - zero only if @xmath636 .
recall from [ p : schurpairs ] that @xmath664 is a self - dual schur pair .
therefore , the result follows by [ c : ptsocle ] and [ c : kleshsocle ] .
. then @xmath882 .
equivalently , @xmath883 .
combine [ c : soc5 ] and [ c : simplesocles ] for the first isomorphism and take duals for the second .
in this section , as our second application of the results in [ s : schurpairs ] , we prove fayers conjecture @xcite .
this is one of the main result of this paper .
we first give a precise statement of fayers conjecture and then recall some recent results that we need from the graded representation theory of @xmath6 and @xmath7 . throughout this section
we assume that @xmath884 and we fix @xmath885 , an element of quantum characteristic @xmath82 ( that is , a primitive @xmath886th root of unity in @xmath887 ) , and a multicharge @xmath888 . in fact , we will work with @xmath602 because our argument uses results of stroppel and webster @xcite who worked with the twisted cyclotomic schur algebra @xmath837 , by [ l : schurequivs ] .
set @xmath889 , where we adopt the convention that @xmath890 when @xmath401 ( so that @xmath891 when @xmath401 ) .
let @xmath1 be an indeterminate over @xmath892 .
let @xmath893 be the * quantised enveloping algebra * of the * kac - moody algebra * @xmath894 .
the algebra @xmath893 is a hopf algebra with coproduct @xmath802 determined by @xmath895 for @xmath804 .
let @xmath896 be the set of * fundamental weights * and @xmath897 the * simple roots * for @xmath894 and set @xmath898 and @xmath899 then @xmath893 is the @xmath900-algebra generated by elements @xmath901 subject to the well - known quantised relations @xcite .
the * bar involution *
@xmath902{.6em}{.1ex}}$ ] is the @xmath892-linear ring automorphism of @xmath893 determined by @xmath903 for @xmath804 .
the ( combinatorial ) * fock space * @xmath904 is the @xmath900-vector space @xmath905 the * residue * of a node @xmath906 $ ] is @xmath907 .
if @xmath908 then @xmath2 is an * @xmath56-node*. following hayashi @xcite and misra and miwa @xcite , the action of @xmath893 on @xmath904 can be described explicitly using the combinatorics of addable and removable @xmath56-nodes . at the categorical ,
level the @xmath893-action corresponds to graded @xmath56-induction and @xmath56-restriction for the cyclotomic hecke algebras @xcite . as we do not need the precise details we refer interested reader to @xcite or @xcite .
let @xmath909 , where @xmath910 , and set @xmath911
. then @xmath912 and @xmath913 is isomorphic to the integrable highest weight @xmath893-module of high weight @xmath914 .
the bar involution @xmath902{.6em}{.1ex}}$ ] induces a unique @xmath892-linear bar involution @xmath902{.6em}{.1ex}}$ ] on @xmath915 such that @xmath916 and @xmath917 , for all @xmath918 .
brundan and kleshchev ( * ? ? ?
* theorem 3.26 ) show that the bar involution on @xmath915 extends to an involution @xmath902{.6em}{.1ex}}$ ] on @xmath904 and hence that the following holds .
[ t : uglov ] suppose that @xmath562 .
then there is a unique bar - invariant element @xmath919 such that @xmath920 for some polynomials @xmath921 $ ] . using the coproduct @xmath802 on @xmath893 , it is straightforward to show that @xmath922 as @xmath893-modules .
uglov @xcite has proved a more general version of these results where the fock space does not necessarily satisfy this tensor product decomposition .
fayers @xcite used the tensor product decomposition of @xmath904 to give an algorithm for computing the elements @xmath923 whenever @xmath924 where and the partition @xmath925 is @xmath82-restricted , for @xmath926 .
this set of multipartitions contains @xmath927 .
( a partition @xmath320 is * @xmath82-restricted * if @xmath928 , for @xmath929 . ) given @xmath468 define @xmath930}\alpha_{\res a}\in q^+$ ] .
the combinatorial classification of the blocks of @xmath561 and @xmath688 given in @xcite is equivalent to the statement that the specht modules , or weyl modules , indexed by @xmath439 and @xmath443 are in the same block if and only if @xmath931 .
following @xcite define the * defect * of @xmath439 by @xmath932 by the remarks above , the defect is a block invariant .
the defect is easily seen to be equivalent to the combinatorial definition of the * weight * @xmath933 of @xmath439 given by @xcite .
we can now state fayers conjecture @xcite .
[ con : fayers ] suppose that @xmath10 is a field of characteristic zero and that @xmath442 then @xmath934 and , moreover , @xmath935 only if @xmath636 .
to prove this conjecture , we work in a graded setting where we can interpret @xmath936 as a graded decomposition number . building on work of khovanov and lauda
@xcite and rouquier @xcite , brundan and kleshchev @xcite showed that ( each block of ) @xmath6 is a @xmath937-graded algebra . extending this result , stroppel and webster @xcite and the authors ( when @xmath401 ) @xcite showed that the cyclotomic schur algebras admit a @xmath937-grading .
following stroppel and webster @xcite , we will work with the twisted cyclotomic schur algebra @xmath688 .
let @xmath938 and @xmath939 be the basic graded algebras of @xmath688 and @xmath561 , respectively , and let @xmath940 and @xmath941 be the corresponding categories of finite dimensional graded modules with homogeneous maps of degree zero .
( unlike in the ungraded setting , graded basic algebras are not uniquely determined up to isomorphism , as graded algebras .
there is , however , a unique grading on @xmath938 such that [ t : koszul ] below holds and by applying the graded schur functor this fixes the grading on @xmath939 .
more explicitly , @xmath938 is the graded endomorphism algebra of @xmath942 , where @xmath943 is the graded projective cover of the corresponding self - dual graded simple module . )
the algebra @xmath938 comes equipped with graded standard modules @xmath944 and graded simple modules @xmath945 , where we fix the grading on these modules by requiring that @xmath945 is graded self - dual and that there is a homogeneous surjection @xmath946 , for @xmath468 .
similarly , the graded algebra @xmath939 has graded analogues of the specht modules @xmath947 and graded simple modules @xmath948 where we again fix the gradings requiring that @xmath948 is graded self - dual and that @xmath949 surjects onto @xmath948 , for @xmath468 and @xmath636 . in particular
, this implies that @xmath950 and @xmath951 are both one dimensional modules concentrated in degree zero .
if @xmath952 is a @xmath937-graded module let @xmath953 be the shift functor so that @xmath954 is the @xmath937-graded module with homogeneous component of degree @xmath955 being @xmath956 .
then the * graded decomposition numbers * of @xmath938 and @xmath939 are the laurent polynomials @xmath957_q=\sum_{d\in{\mathbb z } } [ { \rlap{\textcolor{black!80!white}{$\delta$}}\hspace*{0.6truemm}\delta_\prime}^{{\boldsymbol\lambda}}:{\mathbb{l}_\prime}^{{\boldsymbol\mu}}\<d\>]q^d \quad\text{and}\quad [ { \mathbb{s}_\prime}^{{\boldsymbol\lambda}}:{\mathbb{d}_\prime}^{{\boldsymbol\mu}}]_q=\sum_{d\in{\mathbb z}}[{\mathbb{s}_\prime}^{{\boldsymbol\lambda}}:{\mathbb{d}_\prime}^{{\boldsymbol\mu}}\<d\>]q^d,\ ] ] for @xmath442 .
abusing notation slightly , as in [ s : schurpairs ] there is a graded schur functor @xmath958 such that @xmath959 and @xmath960 which , of course , is zero if @xmath961 .
[ t : stroppelwebster ] suppose that @xmath884 and let @xmath442 . then @xmath957_q = d_{{{\boldsymbol\lambda}}{{\boldsymbol\mu}}}(q).\ ] ] moreover , if @xmath636 then @xmath962_q = d_{{{\boldsymbol\lambda}}{{\boldsymbol\mu}}}(q)$ ] .
in particular , @xmath963 $ ] , for all @xmath442 .
the graded decomposition numbers of @xmath939 were first computed by brundan and kleshchev @xcite .
we need one property of these polynomials , which was part of fayer s motivation for [ con : fayers ] .
[ l : defect ] suppose that @xmath884 , @xmath468 and @xmath874 .
. moreover , the following are equivalent : 1 .
@xmath935 , 2 .
@xmath965 , 3 .
@xmath966 . using properties of the polynomials
@xmath967 from [ t : uglov ] , fayers ( * ? ? ?
* corollary 2.4 ) proved that if @xmath874 then there is a unique multipartition @xmath468 such that @xmath968 , where @xmath969 is his weight function .
combining fayers result with [ l : defect ] gives a representation theoretic proof that @xmath970 .
we need one more ( deep ) result from the graded representation theory of @xmath938 .
we refer the reader to @xcite for the definition of a _ koszul _ algebra .
[ t : koszul ] suppose that @xmath884 . then the basic twisted cyclotomic schur algebra @xmath938 is a koszul algebra .
the papers @xcite actually prove this result for the blocks of @xmath938 , however , this implies [ t : koszul ] because the direct sum of two koszul algebras is again koszul .
notice that [ t : stroppelwebster ] implies that @xmath938 and @xmath939 are both positively graded algebras ( strictly , non - negatively graded algebras ) .
moreover , since @xmath950 and @xmath948 are both concentrated in degree zero , this implies that each of the modules @xmath944 , @xmath945 , @xmath947 and @xmath971 is positively graded since @xmath963 $ ] .
if @xmath972 is a positively graded @xmath938-module then its * grading filtration * is @xmath973 is given by @xmath974 . since @xmath938 is a positively graded algebra
, @xmath975 is an @xmath938-module for @xmath976 . in particular
, @xmath944 has a grading filtration . for @xmath442 write @xmath977
let @xmath978 be the radical filtration of @xmath944 .
[ c : radical ] suppose that @xmath884 and let @xmath468
. then @xmath979 .
consequently , @xmath980_q = d_{{{\boldsymbol\lambda}}{{\boldsymbol\mu}}}^{(r)}q^r , \qquad\text { for all } r\ge0.\ ] ] by ( * ? ? ?
* corollary 2.3.3 ) , any koszul algebra is quadratic . therefore ,
since @xmath981 is simple and concentrated in degree zero , the radical filtration of @xmath944 coincides with the grading filtration of @xmath944 by ( * ? ? ?
* proposition 2.4.1 ) .
we can now prove a stronger version of [ con : fayers ] ( and of [ l : defect ] ) .
[ t : fayers ] suppose @xmath884 and that @xmath442 .
then @xmath934 with equality only if @xmath874 .
moreover , the following are equivalent : 1 .
@xmath935 , 2 .
@xmath965 , 3 .
@xmath966 , 4 .
@xmath982 , 5 .
@xmath983 .
first observe that by the comments before [ e : defect ] , if @xmath984 then @xmath985 .
fix @xmath562 such that @xmath984 and @xmath986 is maximal in the sense that @xmath987 for all @xmath988 . by the maximality of @xmath955 , @xmath989 and @xmath990 is a summand of @xmath991 , where the last equality comes from [ c : radical ] .
consequently , @xmath990 is a summand of the socle of @xmath944 . forgetting the gradings
, this implies that @xmath992 is contained in socle of @xmath752 .
therefore , @xmath874 by [ t : deltasocle ] . as @xmath874 and @xmath993 whenever @xmath986 is maximal the theorem now follows by applying [ l : defect ] .
notice that by interchanging the roles of @xmath407 and @xmath994 , [ t : fayers ] becomes a result about @xmath7-modules . by [ t : fayers ] if @xmath995 then the socle of @xmath944 will not have a component in degree @xmath996 and , _ a priori _ , @xmath997 is not necessarily homogeneous . on the other hand , by [ t : deltasocle ] and [ t : fayers ] , if @xmath998 and @xmath990 appears in @xmath997 then @xmath999 and @xmath874 .
in our final section we return to ungraded representation theory of the cyclotomic schur algebras @xmath7 and we prove some results relating the hom - spaces between specht modules and weyl modules .
the main result in this section is a cyclotomic analogue of the classical carter - lusztig theorem ( * ? ? ?
* theorem 3.7 ) .
the @xmath1-analogue of this result in level one was proved by dipper and james @xcite .
since @xmath832 is exact it suffices to consider the case when @xmath1002 and @xmath1003 , for some @xmath442 .
suppose that @xmath1004 and that @xmath1005 .
it follows that @xmath1006 now , if @xmath1007 then @xmath1008 so there exists @xmath636 such that @xmath1009\ne0 $ ] by [ t : deltasocle ]
. therefore , @xmath1010 , showing that @xmath1011 if and only if @xmath1012 .
the main result of this section gives sufficient conditions for the map of [ l : injection ] to be an isomorphism of vector spaces . in the language of rouquier (
* definition 4.37 ) , our result gives sufficient conditions for @xmath7 to be a _
@xmath1013-faithful cover _ of @xmath6 . by [ l : injection ] there is an injection @xmath1014 , for all @xmath442 . using [ c : dualspechts ] to take duals gives an isomorphism @xmath1015 .
therefore , [ l : injection ] also follows directly from [ p : ahoms](b ) in view of [ l : spechtsubmodules ] and [ p : weylsubmodules ] .
we work homomorphisms between dual specht modules because this better fits the framework developed in [ s : schurpairs ] .
fix @xmath442 . by [ p : weylsubmodules ] and [ l : spechtsubmodules ]
we can make the identifications @xmath1016 and @xmath1017 , respectively .
therefore , @xmath1018 by [ p : ahoms](b ) , these hom - spaces are isomorphic as vector spaces if and only if @xmath1019 . to investigate
when this happens we use another description of the ( left ) weyl modules .
first , however , we need the following fact that does not seem to be in the literature . in level one , this is a result of dipper and james ( * ? ? ?
* lemma 4.1 ) . by (
* theorem 4.14 ) , and as discussed in [ s : schuralgebras ] , @xmath1023 has basis @xmath1024 .
therefore , @xmath1020 only if there exist tableaux @xmath1025 and @xmath1026 , for some @xmath1027 , such that @xmath1028 .
hence , @xmath1029 by [ p : mndual ]
. consequently , @xmath1030 , so that @xmath1031 , as required .
recall the element @xmath1032 from [ s : heckealgebras ] .
let @xmath1033 be the natural embedding of @xmath1034 into @xmath1035 and set @xmath1036 .
then @xmath1037 is a left @xmath7-submodule of @xmath1038 .
( the @xmath7-module @xmath1039 is the full tilting module underpinning the proof of [ t : ringelduality ] ; see ( * ? ? ?
* theorem 6.18 ) . ) for the proof of the next result recall from [ s : schuralgebras ] that @xmath1040 and @xmath1041 , respectively , are the sets of row and column semistandard @xmath1042-tableau of type @xmath443 . if @xmath1043 let @xmath1044 be the unique standard @xmath1042-tableau such that @xmath1045 and @xmath1046 whenever @xmath459 is standard and @xmath1047 . mirroring the definition of the @xmath1048-basis of @xmath7 , for @xmath1049 and @xmath1050 , where @xmath1027 , define @xmath1051 by @xmath1052 , for @xmath543 and @xmath1053 , where the sum is over all standard tableaux such that @xmath1054 and @xmath1055 . then [ t : cellularsalgebras ] and [ p : mullineux ] imply that @xmath1056 is a cellular basis of @xmath684 . by (
* proposition 6.4 ) , a basis of @xmath1037 is given by the maps @xmath1060 where @xmath1061 , for @xmath1062 . as noted in ( * ? ? ?
* theorem 6.5 ) this implies that @xmath1037 has a filtration by left weyl modules such that @xmath1063 . in particular ,
@xmath1064 only if @xmath1065 . by ( * ?
* proposition 7.1 ) , the hom - space @xmath1066 has basis @xmath1067 where @xmath1068 for @xmath1069 .
( as in [ r : nbasislabels ] , the notation used in @xcite is conjugate to what we use here , and we are working with left @xmath7-modules rather than right @xmath7-modules , so some care must be taken when comparing our results with those of @xcite . )
armed with these facts we can prove the lemma .
fix @xmath1070 , @xmath1071 , @xmath1049 and @xmath1050 .
then the @xmath6-module homomorphism @xmath1072 is completely determined by @xmath1073 in particular , @xmath1074 only if @xmath1075 by [ l : spechtsubmodules ] .
therefore , @xmath1076 only if @xmath1077 by [ l : mnideals ] , and hence only if @xmath1078 .
notice that @xmath1079 .
therefore , as a submodule of @xmath1037 , the left weyl module @xmath1080 has basis @xmath1081 .
hence , by the last paragraph ( with @xmath1082 and @xmath1083 ) , if @xmath1084 and @xmath1085 then @xmath1086 contains @xmath1080 , for all @xmath1085 .
that is , @xmath1087 .
conversely , if @xmath1088 and there exist tableaux @xmath1070 and @xmath1071 then @xmath1089 and @xmath1090 since @xmath1091 .
hence , @xmath1092 by the calculation above and [ p : mndual ] .
that is , @xmath1093 .
this calculation is independent of @xmath735 and , in fact , it shows that @xmath1094 , whenever @xmath1095 , not all zero . more generally , suppose that @xmath1096 , for some scalars
@xmath1097 such that @xmath1098 only if @xmath1099 .
fix @xmath1100 such that @xmath1101 and where @xmath447 is minimal in the sense that @xmath1102 whenever @xmath1098 for some @xmath1103 . combining [ p : mndual ] with what we have just shown , @xmath1104 .
we have @xmath1106 since @xmath1107 . on the other hand ,
if @xmath1084 then , as is evident from the proof of [ l : kernel ] , every homomorphism @xmath1085 is given by right multiplication by some element @xmath1108 of @xmath6 .
as @xmath1109 by [
l : kernel ] , @xmath1110 , so if @xmath1111 then @xmath1112 .
therefore , @xmath1113 , for all @xmath1085 . the corollary now follows by [ l : kernel ] . by [ c : simpleisomorphism ] , @xmath1115 and hence @xmath1116 .
in particular , @xmath566 , so there exists @xmath543 and @xmath1117 such that @xmath1118 , or equivalently , @xmath1119 .
suppose that @xmath1120 .
then @xmath1121 .
hence @xmath1122 , so that @xmath1123 , since @xmath1124 .
hence , @xmath1125 as required . in view of [ e : carterlusztig ] and [ p : ahoms](b ) , it is enough to show that @xmath1133 , under the assumptions of the theorem . by [ c : weylintersection ] , @xmath1134 ,
so it is enough to show that @xmath1135 . if @xmath642 then this is immediate from [ l : kleshannihiator ] . on the other hand ,
if the conditions in ( b ) hold then @xmath1136 by [ l : nnuannihilator ] , but @xmath1137 , since @xmath1107 , so again the result follows . by [ l : injection ] , the schur functor induces an injective map @xmath1014 . under the assumptions of [ t : carterlusztig ] this map is an isomorphism . if neither of the conditions in ( a ) and ( b ) in [ t : carterlusztig ] hold then it is not difficult to find examples where these two hom - spaces are not isomorphic .
see , for example , ( * ? ? ?
* remark 6.10 ) . under the assumptions in part
( b ) the isomorphism of [ t : carterlusztig ] is implied by ( * ? ? ?
* theorem 6.6 ) . as here
, rouquier s argument relies on results from @xcite and , in fact , the conditions in [ t : carterlusztig](b ) come from ( * ? ? ?
* theorem 6.9 ) .
we state the full result here because almost no extra effort is required to prove [ t : carterlusztig ] under the assumptions in parts ( a ) and ( b ) . ,
http://dx.doi.org/10.1017/cbo9780511623592 [ _ local representation theory _
] , cambridge studies in advanced mathematics , * 11 * , cambridge university press , cambridge , 1986 .
modular representations as an introduction to the local representation theory of finite groups . height 2pt depth -1.6pt width 23pt , http://dx.doi.org/dx.doi.org/10.1112/plms/pdv007[_quiver schur algebras for linear quivers _ ] , proc .
london math .
soc . , * 110 * ( 2015 ) , 13151386 .
http://arxiv.org/abs/1110.1699[arxiv:1110.1699 ] . height 2pt depth -1.6pt width 23pt , http://dx.doi.org/10.1007/s00208-015-1242-8[_seminormal forms and cyclotomic quiver hecke algebras of type @xmath2 _ ] , mathematische annalen , 2015 , in press .
http://arxiv.org/abs/1304.0906[arxiv:1304.0906 ] . , http://dx.doi.org/10.1112/plms/pds019[_universal graded specht modules for cyclotomic hecke algebras _ ] , proc .
( 3 ) , * 105 * ( 2012 ) , 12451289 .
http://arxiv.org/abs/1102.3519[arxiv:1102.3519 ] . height 2pt depth -1.6pt width 23pt , http://dx.doi.org/10.1016/j.jalgebra.2004.07.021[_matrix units and generic degrees for the ariki - koike algebras _ ] , j. algebra , * 281 * ( 2004 ) , 695730 .
http://arxiv.org/abs/math/0108164[arxiv:math/0108164 ] . height 2pt depth -1.6pt width 23pt , http://dx.doi.org/10.1142/9789814651813_0005[_cyclotomic quiver hecke algebras of type a _ ] , in modular representation theory of finite and @xmath0-adic groups , g. w. teck and k. m. tan , eds . ,
national university of singapore lecture notes series , * 30 * , world scientific , 2015 , ch . 5 , 165266 .
http://arxiv.org/abs/1310.2142[arxiv:1310.2142 ] . ,
_ canonical bases of higher - level @xmath1-deformed fock spaces and kazhdan - lusztig polynomials _ , in physical combinatorics ( kyoto , 1999 ) , progr . math .
, * 191 * , boston , ma , 2000 , birkhuser boston , 249299 . | gordon james proved that the socle of a weyl module of a classical schur algebra is a sum of simple modules labelled by @xmath0-restricted partitions .
we prove an analogue of this result in the very general setting of `` schur pairs '' . as an application
we show that the socle of a weyl module of a cyclotomic @xmath1-schur algebra is a sum of simple modules labelled by kleshchev multipartitions and we use this result to prove a conjecture of fayers that leads to an efficient llt algorithm for the higher level cyclotomic hecke algebras of type @xmath2 .
finally , we prove a cyclotomic analogue of the carter - lusztig theorem . |
the spatial distribution of galaxies , _ large scale structure _ ( lss ) , provides fundamental knowledge about the formation and evolution of the density structure in our universe .
a recent measurement of the cmb ( cosmological microwave background ) anisotropy by wmap ( wilkinson microwave anisotropy probe ) puts a stringent constraint on the primordial density fluctuations , and favors the @xmath2-dominated spatially flat cold dark matter ( lcdm ) model with a gaussian initial condition @xcite .
one of the goals in the analysis of galaxy distribution is to test cosmological models in an independent and complementary manner to the cmb analysis .
another goal is to understand galaxy clustering , with particular emphasis on galaxy biasing , which is a statistical relation of the clustering between galaxies and the underlying dark matter .
recent wide - field galaxy surveys , such as sdss ( sloan digital sky survey ) and 2dfgrs ( two degree field galaxy redshift survey ) , indeed enable a detailed study of galaxy clustering with unprecedented accuracy .
more conventional statistics for analyzing the galaxy distribution is the two - point correlation function , or the power spectrum in fourier space .
two - point statistics have been extensively studied for more than 35 years since the pioneer work by @xcite , and have been applied to various redshift surveys , including sdss ( e.g. , @xcite for sdss galaxies ; @xcite for sdss quasars ) .
they , however , can not fully describe the statistical nature of the galaxy distribution , because lss is highly non - gaussian due to the nonlinear gravitational evolution and nonlinearity in the galaxy biasing . to characterize the non - gaussian properties of the galaxy distribution
requires higher order statistics beyond two - point statistics .
motivated by this requirement , various statistics have been introduced as complementary tools to two - point statistics for the analysis of galaxy clustering : higher order correlation functions ( e.g. , peebles 1980 ) , genus statistics @xcite , minkowski functionals @xcite , minimum spanning trees @xcite , void statistics @xcite , and so on . while these statistics capture different parts of non - gaussian features , they are not completely independent of the two - point statistics .
the fourier transform , @xmath10 , of a density fluctuation field is separately written in terms of the amplitude and the phase , @xmath11 , as @xmath12 the gaussian fields have a uniform distribution of the fourier phases over @xmath13 . therefore , characterizing the correlation of phases is expected to be a direct means to explore the non - gaussian features .
furthermore , fourier phases are statistically independent information of the power spectrum , which is defined by @xmath14 .
nevertheless , finding useful statistics of the fourier phases is not easy , mainly because of the cyclic property of the phase .
for example , the one - point phase distribution turns out to be essentially uniform , even in a strongly non - gaussian field @xcite .
for this reason , previous studies of the fourier phase have been mainly devoted to the evolution of phase shifts in individual modes @xcite , or the phase differences between the fourier modes @xcite .
@xcite recently proposed , as a new measure of phase correlations , the distribution function of the `` phase sum '' , @xmath15 , where the corresponding wavevectors satisfy @xmath16 .
although a connection between the higher order statistics and the phase correlations was suggested earlier @xcite , he discovered an important analytic relation between the distribution function of the phase sum and the polyspectra using perturbation theory .
a subsequent numerical analysis by @xcite explained the behavior of the phase - sum distribution with respect to the density structure in real space . when a prominent density peak exists in a given sampling volume , the fourier phases of various modes are synchronized to have nearly zero values at the position of peak , and thus the phase sum distributes around zero .
on the other hand , when several peaks with comparable heights exist , the synchronization of phases is diluted , and thus the phase - sum distribution becomes almost uniform .
the phase - sum distribution is sensitive to the non - gaussian feature of the density field , especially the relative strength of the most high - density peak to other density peaks .
the nonlinear gravitational evolution and the nonlinearity in the galaxy biasing statistically changes the relative strength of density peaks in galaxy distribution , and therefore the phase - sum distribution is a useful tool to probe the nonlinear effects .
the present analysis applies , for the first time , the above phase statistics to the sdss galaxy catalogs to quantify the non - gaussianity in galaxy distributions .
there are three possible sources of the non - gaussianity in the present galaxy density field : primordial density field , nonlinear gravitational evolution , and nonlinear galaxy biasing .
recently , wmap showed that the primordial density field is well approximated by gaussian statistics .
thus , the primordial non - gaussianity can be safely ignored in the following analysis , and the prominent source of non - gaussianity in the galaxy distribution is nonlinear gravitational evolution . in our analysis , we use the phase statistics as a cosmological tool independent of the two - point statistics .
we compare the phase correlations of sdss galaxies with those of mock samples constructed from @xmath0-body simulations based on various cosmological models . for a fair comparison with observations , we consider the observational systematic effects including the survey geometry , the redshift distortion and the shot - noise due to the sparse sampling .
we find that the sdss galaxy data agree very well with the lcdm model predictions with @xmath3 , which is consistent with the results of wmap .
cccccccccc & & & & @xmath17 & & @xmath18 & @xmath19 & range of @xmath20 + [ 0pt]@xmath21 & [ 0pt]@xmath22 & [ 0pt]@xmath23 & [ 0pt]@xmath24 & [ @xmath25 & [ 0pt]@xmath26 & [ @xmath27mpc ] & [ @xmath27mpc ] & [ @xmath28mpc ) ] + @xmath29 & @xmath30 & @xmath31 & @xmath32 & @xmath33 & @xmath34 & @xmath35 & @xmath36 & @xmath37@xmath38 + @xmath30 & @xmath39 & @xmath40 & @xmath41 & @xmath42 & @xmath43 & @xmath44 & @xmath45 & @xmath46@xmath47 + @xmath39 & @xmath48 & @xmath49 & @xmath31 & @xmath50 & @xmath51 & @xmath52 & @xmath53 & @xmath54@xmath55 + @xmath48 & @xmath56 & @xmath57 & @xmath40 & @xmath58 & @xmath59 & @xmath60 & @xmath61 & @xmath62@xmath63 + + [ tab : vollim ] another possible source for non - gaussianity , galaxy biasing , is the uncertainty of the statistical relation between galaxies and the underlying mass , which originates from complicated processes of galaxy formation .
it is known that galaxy clustering depends sensitively on various properties of galaxies , such as the luminosity , color , morphology , and environment ( e.g. , @xcite ) .
the two - point statistics using sdss galaxies by @xcite and @xcite clearly exhibits the luminosity and morphology dependence of the galaxy biasing ( see also @xcite for the dependence in three - point correlation functions ) .
if the density fluctuation of galaxies , @xmath64 , is related to the mass density fluctuation , @xmath65 , as @xmath66 , their fourier transforms also have the linear relationship @xmath67 . decomposing the fourier wavevectors to the parts of amplitudes and phases ,
the amplitudes change proportionally to @xmath7 regardless of the scale ; however , the phases are completely preserved under linear biasing .
therefore , the agreement of the observed phase - sum distribution with the simulated predictions suggests that the galaxy biasing is well approximated by the linear relation if a correct cosmological model is assumed . for definiteness
, we adopt a quadratic deterministic biasing and put a constraint on the degree of the nonlinearity of galaxy biasing in a weakly nonlinear regime ( @xmath68mpc ) .
this paper is organized as follows . in section 2
we describe the sdss data sets and our simulation mock catalogs .
section 3 summarizes the perturbative formula of phase - sum distribution by @xcite and explains the method of measuring phase correlations using the distribution function of phase sum .
the results of observed phase correlations are also presented .
section 4 is focused on the bispectrum analysis of @xmath69 from comparison between observations and their mock catalogs .
finally section 5 is devoted to a summary of our results and further discussion .
our present analysis is based on a subset of the sdss galaxy redshift data , ` large - scale structure sample 15 ' @xcite .
this sample includes the spectroscopic data of @xmath70 galaxies and covers the sky area of @xmath71 square degrees .
the angular selection function of the survey is written in terms of spherical polygons @xcite .
details of the sdss can be found in the following literature : @xcite provide an overview of the sdss .
technical articles providing details of the sdss include descriptions of the photometric camera @xcite , photometric analysis @xcite , the photometric system and photometric calibration @xcite , the photometric pipeline @xcite , astrometric calibration @xcite , selection of the galaxy spectroscopic samples @xcite , and spectroscopic tiling @xcite .
the details of publicly released data are summarized in @xcite for early data release and abazajian et al .
( 2003 , 2004 , 2005 ) for data release one , two and three , respectively .
in our analysis we use observational data in the northern hemisphere , except for the distant area from the other observed areas , which is designed to overlap with the spitzer space telescope first look survey .
we also omit the area of three stripes in the southern hemisphere , which has a survey geometry not appropriate for the analysis of galaxy clustering in three - dimensional space .
the range of the @xmath72-band apparent magnitude ( @xmath73 ) is set to be a conservative range of ( @xmath74 ) after correction for galactic reddening using the maps of @xcite .
we construct four volume - limited samples with a unit width of the absolute magnitude , which cover the range of the @xmath72-band absolute magnitude from @xmath75 to @xmath76 ( table [ tab : vollim ] ) .
the redshift range of each volume - limited sample ( @xmath77 ) is determined by the following distance modulus relation : @xmath78 \\ & -&k(z_{\rm min / max } ) , \nonumber\end{aligned}\ ] ] where @xmath79 is the comoving distance at a redshift of @xmath80 and @xmath81 is a quadratic fitting formula of the averaged @xmath82-correction as a function of @xmath83 @xcite , @xmath84 the details of @xmath82-correction is described in @xcite .
table [ tab : vollim ] summarizes the properties of our volume - limited subsamples . for a fair comparison with the observation
, we construct a set of mock simulation data , including observational effects of the survey geometry , the number density , and the redshift distortion ( hikage et al . 2002 , 2003 ) . for constructing mock samples , we use p@xmath85 m @xmath0-body simulations , provided by @xcite .
the simulation employs @xmath86 particles in a periodic comoving box with a length of @xmath87 or @xmath88 mpc using gaussian initial conditions and a cold dark matter ( cdm ) transfer function @xcite .
we use the @xmath89 snapshot simulation data ( for simplicity we neglect the light - cone effect ) in various cdm models with the parameters listed in table [ tab : modelpara ] ; the dimensionless matter - density parameter , @xmath90 , the dimensionless cosmological constant , @xmath91 , the shape parameter , @xmath92 of the cdm transfer function @xcite , and the r.m.s .
density - fluctuation amplitude smoothed by a top - hat filter with a scale of @xmath93mpc , @xmath94 .
we extract about @xmath95 realizations of wedge samples for each volume - limited sample out of the full simulation cube so that they have the same sample - shape and number of particles ( averaged over mock samples ) as each volume - limited sample . to construct mock samples that extend beyond the simulation box size ,
we duplicate particles using the periodic boundary conditions .
we create mock data in redshift space by adding the line - of - sight component of the peculiar - velocity to the distance of each particle for calculating redshift .
the mock sample simply assumes that each mass particle represents a simulated galaxy , and neglects the effect of galaxy biasing , which are further discussed in section [ sec : results ] .
.simulation model parameters [ cols="^,^,^,^,^",options="header " , ] [ tab : modelpara ]
@xcite derived an analytical relation between the polyspectra and the distribution of the phase sum in perturbation theory . in the lowest - order approximation ,
the probability density function ( pdf ) of the phase sum , @xmath96 , over closed wavevectors @xmath97 , and @xmath98 reduces to @xcite : @xmath99 where @xmath17 is the sampling volume and @xmath69 is defined by the bispectrum , @xmath100 , and the power spectrum , @xmath101 , as follows : @xmath102 in the lowest - order approximation , the pdf of the phase sum over three wavevectors is fully determined by the parameter @xmath69 . if the hierarchical clustering ansatz is valid , @xmath103 and thereby @xmath69 is approximately given by @xmath104 .
therefore , the lowest - order approximation [ equation ( [ eq : matsubara ] ) ] breaks down at a scale @xmath20 satisfying the condition @xmath105 .
note that this is different from the conventional condition of the gravitational nonlinear clustering , @xmath106 .
in fact , a subsequent numerical analysis confirmed the validity of the perturbative formula , even in a fairly nonlinear regime of clustering @xcite . the dependence of the phase - sum distribution on the sampling volume can be naturally explained as follows : the synchronization of phases becomes weak when multiple density peaks with comparable heights exist ( section [ sec : intro ] ) . because the sampling volume is larger , the number of peaks increases , and thus the phase - sum distribution approaches being uniform .
when @xmath17 is sufficiently large , the higher order terms above @xmath69 become negligible and then equation ( [ eq : matsubara ] ) is applicable , even in nonlinear regimes , in the sense that @xmath107 . on the contrary ,
decreasing @xmath17 by dividing the original survey volume into several sub - volumes would provide information about the higher - order spectrum . dividing the original sample
, however , leads to decreasing the number of objects per sub - sample .
thus , a significant fraction of the available fourier modes suffers from the contamination of shot noise . in the present galaxy sample
, it is quite difficult to extract higher order information beyond @xmath69 from the distribution function of the three - point phase sum . in the present work ,
therefore , we do not divide the survey volume of each volume - limited sample .
we compute the distribution function of the phase sum for density contrast fields from mock samples and sdss galaxies .
we use the cloud - in - cell interpolation to assign galaxies or dark matter particles to mesh densities .
then , the density contrast field is fourier - transformed on cubic grids with a mesh number @xmath108 per side .
we calculate the sum of the fourier phases for the closed set of three wavevectors , and then compute the distribution function of the phase sum within each binning range of triangle configurations . for simplicity ,
we focus on the configuration of wavevectors to be a nearly isosceles triangle , where the absolute values of two wavevectors , @xmath109 and @xmath110 , are within the same binning range of scale @xmath20 .
the scale @xmath20 is divided into @xmath111 bins with equal width in logarithmic scale in the range of @xmath112 .
the nyquist frequency @xmath113 is equal to @xmath114 , where @xmath19 is a side of a cubic box for a fourier transformation .
table [ tab : vollim ] lists @xmath19 for each volume - limited sample , which is chosen for the sample volume to be barely covered .
we limit the range of scale to be @xmath115 and @xmath116 so that the observational systematics due to the complicated survey geometry and the discreteness effect may be neglected compared to the sample variance ; at smaller @xmath20 , the poisson error increases due to the limited number of independent combinations of modes forming closed triangles , and furthermore the convolution effect with the survey mask is more serious . at larger @xmath20 ,
the shot noise becomes more serious .
table [ tab : vollim ] lists the range of scale @xmath20 for each volume - limited sample .
the unit of scale @xmath20 in fourier space is set to be @xmath117 , and thereby the corresponding scale - length in real space is just the inverse of @xmath20 in units of @xmath27mpc .
the angle @xmath118 between @xmath119 and @xmath120 , i.e. , @xmath121 $ ] with @xmath122 , is binned to have an equal width of @xmath123 .
( 80mm,80mm)figure1a.eps ( 80mm,80mm)figure1b.eps figure [ fig : obsphasesum ] plots the distribution function of the phase sum for each volume - limited sample in symbols . for comparison ,
we plot in lines the lowest - order approximation [ equation ( [ eq : matsubara ] ) ] using @xmath69 computed from the bispectrum and the power spectrum ( [ eq : p3 ] ) . in the left panels , the triangle configurations are nearly equilateral shape ( @xmath124 ) for different @xmath20 . in the right panels ,
the triangle configurations are nearly isosceles for different angles , @xmath118 .
the phase - sum distributions are found to be well approximated by the lowest - order approximation for all of the volume - limited samples , regardless of the triangle configurations .
the dependence of @xmath69 on @xmath17 and @xmath20 can be understood from the nature of perturbative parameter , @xmath69 , which is roughly proportional to @xmath125^{1/2}$ ] under the hierarchical clustering ansatz ; volume - limited samples for fainter galaxies have a smaller @xmath17 ( increase @xmath69 ) , but a smaller @xmath101 ( decrease @xmath69 ) because the measurable range of scale @xmath20 is shifted to larger values due to the smaller @xmath126 .
the redshift distortion due to the random motion of galaxies also smears the power at small scales , and thus @xmath69 reaches up to @xmath127 at a maximum , much smaller than unity .
figure [ fig : obsphasesum ] shows that @xmath69 calculated from the amplitude of the phase - sum distribution using equation ( [ eq : matsubara ] ) is nearly equal to that from the combination of the bispectrum and the power spectrum [ equation ( [ eq : p3 ] ) ] . in what follows , we will use equation ( [ eq : p3 ] ) to compute @xmath69 instead of fitting the pdf to equation ( [ eq : matsubara ] ) .
figures [ fig : obsp3_scale ] and @xmath128 show @xmath69 for sdss galaxies in different volume - limited samples ( plotted in symbols ) against the wavenumber @xmath20 ( figure [ fig : obsp3_scale ] ) and the angle @xmath118 ( figure [ fig : obsp3_shape ] ) . for comparison ,
the lcdm predictions with @xmath3 are plotted in lines .
clearly , the scale - dependence of @xmath69 is well approximated by a power - law of @xmath20 .
this can be understood again from equation ( [ eq : p3 ] ) ; because @xmath129 and @xmath101 may be approximated as power - laws at those scales , we expect @xmath130 .
the right panels show the ratio of the observed @xmath69 to the corresponding mock estimations in the left panels .
the overall agreement between observations and lcdm predictions with @xmath3 is very good .
figures [ fig : obsp3_scale_cdm ] and [ fig : obsp3_shape_cdm ] show comparisons of the observed @xmath69 with various mock predictions to examine the cosmology dependence .
figure [ fig : obsp3_scale_cdm ] focuses on the scale dependence , while figure [ fig : obsp3_shape_cdm ] focuses on the shape dependence ( the angle @xmath118 of the triangles ) .
the right panels show the ratio of @xmath69 between observations and the corresponding mock samples in the left panels .
again , lcdm predictions with @xmath3 show the best agreement with the observations among our mock samples ; the scdm model predicts a smaller @xmath69 than the other models over the whole range of scales , mainly because of the smaller value of @xmath94 ( see table [ tab : modelpara ] ) .
the ocdm model has the same value of @xmath94 as the lcdm models , but their spectral shape is different : the small - scale fluctuation in ocdm is slightly larger due to the larger value of @xmath92 , and thus ocdm predicts higher values of phase correlations than lcdm .
figures [ fig : obsp3_scale_sig8 ] and [ fig : obsp3_shape_sig8 ] are same as figures [ fig : obsp3_scale_cdm ] and [ fig : obsp3_shape_cdm ] , but for the @xmath94 dependence using mock samples based on the same cosmology of lcdm .
a systematic increase of @xmath69 as @xmath94 is clearly found because the phase correlations become strong as the gravitational evolution proceeds .
( 80mm,80mm)figure2a.eps ( 80mm,80mm)figure2b.eps ( 80mm,80mm)figure3a.eps ( 80mm,80mm)figure3b.eps ( 80mm,80mm)figure4a.eps ( 80mm,80mm)figure4b.eps ( 80mm,80mm)figure5a.eps ( 80mm,80mm)figure5b.eps ( 80mm,80mm)figure6a.eps ( 80mm,80mm)figure6b.eps ( 80mm,80mm)figure7a.eps ( 80mm,80mm)figure7b.eps galaxy biasing is another source of uncertainty in comparing the observations and the mock predictions . if the galaxy biasing is linear , i.e. , @xmath66 , the amplitude changes in proportion to the linear coefficient , @xmath7 , but the phases are invariant .
thus , the consistency of the lcdm model remains valid as long as the galaxy biasing is well approximated by the linear relation in _ redshift _ space ( note that this does not strictly correspond to the linear biasing in _ real _ space ) .
the analysis of the bispectrum for 2dfgrs finds that galaxy biasing is consistent to be linear under lcdm model in the range of scales from @xmath131mpc to @xmath132mpc @xcite .
indeed , the combined analysis of three - point correlation functions and two - point correlation functions shows a significant deviation from linear biasing at _ nonlinear _
scales from @xmath133mpc to @xmath134mpc @xcite .
in contrast , the scales that we probe in the present analysis is much larger , i.e. , @xmath135mpc .
we therefore parameterize the galaxy biasing by the local and quadratic relation between the density fluctuation of galaxies , @xmath136 , and that of mass , @xmath137 , as follows ; @xmath138 where @xmath7 and @xmath6 are biasing parameters . in a weakly nonlinear regime , for example at scales larger than @xmath132mpc , the locality of the galaxy biasing is valid because the scale of galaxy formation is much smaller ( @xmath139 mpc ) .
also the density fluctuation is less than unity , and thereby higher order terms of the density fluctuation are negligible . under the above quadratic biasing , @xmath140 for galaxies and @xmath141 for mass
are related as @xmath142 @xmath143 where @xmath144 , and @xmath145 .
equation ( [ eq : p3bias ] ) implies that the difference of @xmath146 from @xmath147 is proportional to the ratio @xmath8 alone .
figure [ fig : phase_bias_mean ] plots @xmath8 calculated from the difference between @xmath146 and @xmath147 for each cosmology using equation ( [ eq : p3bias ] ) .
the plotted values of @xmath8 are averaged over scales larger than @xmath132mpc ( @xmath148mpc ) ] ) . in constraining @xmath8
, we do not use the phase sum for the configuration of triangle wavevectors with the opening angle , @xmath149 , because we find that they strongly suffer from the observational systematic due to the survey geometry beyond the sample variance .
we note that the ratio of the biasing parameters in figure [ fig : phase_bias_mean ] is calculated from the density field in redshift space and thus their values are quantitatively different in real space .
( 80mm,80mm)figure8.eps ( 80mm,80mm)figure9.eps the results in figure [ fig : phase_bias_mean ] suggest that the nonlinearity of the galaxy biasing , @xmath8 , is generally rather small and simply dependent on @xmath94 , regardless of the cosmology and galaxy luminosity .
we thereby plot the relation between @xmath94 and @xmath8 averaged over all of four volume - limited samples in figure [ fig : sigma8vsbias ] .
indeed , we find that @xmath94 mainly determines the resulting @xmath8 regardless of cosmology at scales @xmath68mpc , and that their relation is well fitted by the following linear relation : @xmath150 the linear relation is valid if @xmath151 and the power spectrum of the mass fluctuation is mainly determined by the square of @xmath94 at @xmath68mpc ; both @xmath141 and @xmath152 ( equation [ eq : p3bias ] ) are roughly given by @xmath153 , which is proportional to @xmath94 in a nearly linear regime .
thus , @xmath8 is linearly related to @xmath94 as long as @xmath154 is much smaller than unity . when we adopt a value of @xmath155 , measured from the combined analysis of recent wmap results and the power spectrum of sdss galaxies @xcite , @xmath156 .
therefore , we conclude that the galaxy biasing is well approximated by a linear relation .
complementary measurements of @xmath94 from other observations which are independent of the galaxy biasing , including weak gravitational lensing and galaxy clusters , will put additional constraints on the nonlinearity of the galaxy biasing .
we have performed the first measurement of the distribution function of the phase sum of galaxies .
we applied the phase statistics to volume - limited samples of the latest sdss galaxy catalogs .
we found that the distribution functions of the phase sum for all of the sdss galaxy samples are in good agreement with the lowest - order perturbation formula .
since the latter is characterized by @xmath69 , we explored various properties of @xmath69 . for a quantitative comparison with observations , we constructed realistic mock catalogs from @xmath0-body simulations .
we found that the observational phase correlations agree best with the lcdm model with @xmath3 if galaxy biasing is linear ; the scdm model predicts weaker correlations of phases than observations overall , mainly due to the small @xmath94 .
actually , we found a systematic increase of @xmath69 as @xmath94 increases with the same parameters of the lcdm model . assuming that galaxy biasing is expressed by a quadratic deterministic relation in a weakly nonlinear regime ( @xmath157 $ ] ) , we placed constraints on a linear combination of the nonlinearity in the galaxy biasing @xmath8 and @xmath94 as @xmath9 .
using a recent measurements of @xmath155 from a combined analysis of wmap and sdss , the galaxy biasing was found to be well approximated by a linear relation . in the present analysis , we have focused on the behavior of @xmath69 , which carries phase information contained in the bispectrum @xcite . in this sense ,
our analysis is equivalent to an analysis of the bispectrum , which is also a first attempt using sdss data .
higher order information beyond bispectrum can be obtained by measuring the distribution function of the phase sum over more than four numbers of wavevectors which form closed polygons .
another approach is to decrease @xmath17 , because the perturbative parameter , @xmath69 , is inversely proportional to @xmath158 .
however , extracting higher order information is limited mainly by the shot noise , as discussed in section [ sec : phasesum ] . therefore , deeper and denser surveys than sdss are required to carry out this attempt in reality .
we deeply appreciate i. kayo and y. p. jing who kindly provided a large set of @xmath0-body simulation data . c. h. acknowledges support from a jsps ( japan society for the promotion of science ) fellowship .
t. m. acknowledges the support from the ministry of education , culture , sports , science , and technology , grant - in - aid for encouragement of young scientists ( no .
15740151 ) .
the research of y.s . was supported in part by grants - in - aid for scientific research from the japan society for promotion of science ( nos.14102004 and 16340053 ) .
numerical computations were carried out at adac ( the astronomical data analysis center ) of the national astronomical observatory , japan ( project i d : yys08 ) , and also at computer facilities at the university of tokyo supported by the special coordination fund for promoting science and technology , ministry of education , culture , sport , science and technology .
funding for the sloan digital sky survey ( sdss ) 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 , the max planck society , and the hefce .
the sdss web site is http://www.sdss.org/. the sdss is a joint project of the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the korean scientist group , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
lupton , r. , gunn , j. e. , ivezi@xmath159 , @xmath160 . , knapp , g. r. , kent , s. , & yasuda , n. 2001 , in asp conf .
238 , astronomical data analysis software and systems x , ed .
f. r. harnden , jr . , f. a. primini , and h. e. payne ( san
francisco : astr .
soc . pac . ) | we present a first analysis of the clustering of sdss galaxies using the distribution function of the sum of fourier phases .
this statistical method was recently proposed by one of the authors as a new probe of the phase correlations of cosmological density fields . since the fourier phases are statistically independent of the fourier amplitudes , the phase statistic plays a complementary role to the conventional two - point statistics of galaxy clustering . in particular ,
we focus on the distribution functions of the phase sum over three closed wavevectors as a function of the triangle configuration .
we find that the observed distribution functions of the phase sum are in good agreement with the lowest - order approximation from perturbation theory .
for a direct comparison with observations , we construct mock catalogs from @xmath0-body simulations taking account of the survey geometry , the redshift distortion , and the discreteness due to the limited number of data .
indeed the observed phase correlations for the galaxies in the range of the absolute magnitude , @xmath1 , agree well with those for @xmath2-dominated spatially flat cold dark matter predictions with @xmath3 evolved from the gaussian initial condition .
this agreement implies that the galaxy biasing is approximately linear in redshift space . instead , assuming that the galaxy biasing is described by a quadratic deterministic function at @xmath4[@xmath5
, we can constrain the ratio of the quadratic biasing parameter , @xmath6 , to the linear biasing parameter , @xmath7 , from the difference of phase correlations between observations and mock predictions .
we find that the resulting @xmath8 is well fitted by @xmath9 and is almost insensitive to the cosmology and luminosity in those ranges .
indeed , @xmath8 is nearly zero when @xmath3 . |
polymers with contour length @xmath2 much larger than the persistence length @xmath3 , which is the correlation length for the tangent - tangent correlation function along the polymer and is quantitative measure of the polymer stiffness , are flexible and are described by using the tools of quantum mechanics and quantum field theory @xcite-@xcite .
if the chain length decreases , the chain stiffness becomes an important factor .
many polymer molecules have internal stiffness and can not be modelled by the model of flexible polymers developed by edwards @xcite .
the standard coarse - graining model of a worm - like polymer was proposed by kratky and porod @xcite .
the essential ingredients of this model is the penalty for bending energy , and the local inextensibility .
the latter make the treatment of the model much more difficult .
there is a substantial amount of studies of the kratky - porod model in the last half century @xcite-@xcite ( and the references therein ) . in recent years
there is an increasing interest in theoretical description of semiflexible polymers @xcite-@xcite ( and the references therein ) .
a reason for this interest is due to potential applications of semiflexible polymers in biology and in researches on semi - crystalline polymers . in this article
we present results of the study of the behaviour of the worm - like chain for short lengths .
the consideration is based on our recent work @xcite,@xcite , where the fourier - laplace transform @xmath4 of the end - to - end distribution function @xmath0 ( @xmath5 is the laplace conjugate to @xmath1 ) was represented as the matrix element of the infinite order matrix @xmath6 with matrices @xmath7 and @xmath8 related to the spectrum of the quantum rigid rotator .
a truncation of @xmath9 by matrix of order @xmath10 gives the end - to - end distribution function @xmath4 as a rational function being an infinite series in powers of @xmath11 , i.e. it contains all moments of the end - to - end distribution function , and describes the first @xmath12 moments exactly . in context of eigenstates of the quantum rigid rotator , the truncation at order @xmath10 takes into account the eigenstates with quantum number of the angular momentum up to the value @xmath13 .
the moment expansion of @xmath4 can be represented as double series in powers of @xmath11 and @xmath14 .
alternatively , one can expand @xmath15 in double series in powers of @xmath11 and @xmath1 . analyzing the series of known moments ( we have analytically calculated the first @xmath16 moments )
, we have established that the coefficients in subseries in powers of @xmath11 have a simple analytical structure , which enables one to perform a partial resummation of the moment expansion of @xmath4 or @xmath17 ( see below ) .
this resummation procedure results in an expansion of @xmath18 around the stiff rod limit , where the end - to - end distribution function is given by the expression , @xmath19 , so that because @xmath20 is a distribution , so derived short - chain expansion is not a taylor expansion but rather an expansion in the space of distributions .
the knowledge of @xmath15 enables one to compute directly the scattering function @xmath21 and the extension - force relation @xmath22 . in this paper
we present the results of the calculation of terms of the short - chain expansions of @xmath0 , @xmath21 , and @xmath22 by taking into account the deviations of the moments from their stiff rod behaviour to the seven order in chain length @xmath1 .
the procedure can be extended to higher orders .
the short - chain expansions besides the interest in their own can be used for comparisons with approximative treatments , and also in studies of the behaviour of short semiflexible polymers .
the present paper is organized as follows .
section [ formalism ] introduces to the description of the worm - like chain using the formalism of quantum rigid rotator .
section [ short - ch ] explains the idea of the derivation of the short - chain expansion , and presents the short - chain expansions for the distribution function of the end - to - end distance , scattering function , and extension - force relation .
the fourier transform of the distribution function of the end - to - end polymer distance of the continuous kratky - porod model @xcite @xmath23 is expressed by the path integral as follows @xmath24where @xmath25 is the persistence length , and @xmath26 is the tangent vector at the point @xmath27 along the contour of the polymer . the product over @xmath27 in eq.([w1 ] ) takes into account that the polymer chain is locally inextensible . in the following the arc length of the polymer @xmath2 will be measured in units of @xmath28 and will be denoted by @xmath1 .
we now will consider the green s function @xmath29 associated with eq.([w1 ] ) .
the differential equation for @xmath30 is@xmath31where @xmath32 is the potential energy of the rigid rotator in an external field @xmath33 , where @xmath34 is measured in units of @xmath35 .
the end - to - end distribution function is obtained from @xmath36 as follows @xmath37the differential equation ( [ w4 ] ) can be rewritten as an integral equation as follows@xmath38where the bare green s function @xmath39 reads @xmath40with @xmath41 being the spherical harmonics , and @xmath42 and @xmath43 are the quantum numbers of the angular momentum . due to the convolution character of the expression ( [ w5 ] ) with respect to the integration over the contour length ( @xmath44 depends on the difference @xmath45 ) ,
the laplace transform of @xmath36 in eq.([w5 ] ) with respect to @xmath1 permits to get rid of integrations over the contour length .
thus , in the following we will consider the laplace transform of @xmath15 with respect to @xmath1 .
it was shown in @xcite,@xcite that the solution of eq.([w5 ] ) results in the following expression of the fourier - laplace transform of the end - to - end distribution function as the matrix element of an infinite order square matrix @xmath46with @xmath47where the square matrices @xmath48 and @xmath7 are defined by@xmath49with @xmath50 , and @xmath51respectively .
the superscript @xmath27 specifies that the quantities @xmath52 and @xmath48 are square matrices .
the quantity @xmath53 denotes the @xmath54 matrix element of the infinite order square matrix @xmath52 .
since the @xmath54 matrix element corresponds to the expectation value of the quantum rigid rotator in the ground state with the quantum number @xmath55 , we prefer to use the above notation .
summations over the magnetic quantum number in the intermediate states in the expression of @xmath4 obtained by solving eq.([w5 ] ) can be eliminated @xcite , so that summations in the intermediate states occur only over the eigenvalues of the angular momentum @xmath56 .
this is the reason why the calculation of @xmath4 reduces to the computation of the matrix element of an infinite order square matrix . the quantity @xmath57 plays the key role in the theory similar to the bare propagator in common quantum field theories .
the end - to - end distribution function @xmath4 is simply the matrix element @xmath58 , the scattering function of the polymer is the inverse laplace transform of @xmath59 multiplied by @xmath60 , the partition function of the stretched polymer is @xmath61 etc @xcite .
to derive the short - chain expansion for the kratky - porod chain we insert the expansion of @xmath7 given by eq.([w10 ] ) according to@xmath62into the moment expansion of @xmath4@xmath63and obtain as a result @xmath4 as double series in powers of ( @xmath64 and @xmath14 as follows@xmath65the analysis of @xmath16 exactly computed moments of @xmath66 , which are proportional to the coefficients @xmath67 in eq.([mom - exp ] ) , shows that the coefficients @xmath68 are polynomials in powers of @xmath10 of the order @xmath69 , i.e. they have the form @xmath70 ( @xmath71 ) .
we determine the coefficients @xmath72 at given @xmath27 using @xmath69 known terms of the moment expansion .
further we test the correctness of so obtained @xmath68 from the comparison with the remaining known terms in the moment expansion , which were not used to determine @xmath72 .
the exact knowledge of @xmath16 moments enables one to determine @xmath68 for @xmath73 . the coefficients @xmath74 we have obtained in this way are@xmath75we have also computed the coefficients @xmath76 and @xmath77 , but do not write them here to save space .
the determination of higher coefficients demands the knowledge of more moments of the end - to - end distribution function and can be obtained in a similar way .
after the determination of the coefficients @xmath68 we have checked the equivalence of ( w12 ) with ( [ mom - exp ] ) to the corresponding order .
unfortunately , we did not succeed in deriving the expression for the coefficients @xmath68 for arbitrary @xmath27 . with known coefficients @xmath68 one can sum the series over @xmath10 in ( w12 ) . restricting ourselves to the leading order @xmath78 ( @xmath79 is the number of terms in the sum over @xmath27 in ( [ w12 ] ) ) we obtain @xmath15 @xmath80and the distribution function of the end - to - end distance @xmath81 in the stiff rod limit .
taking into account the next term in the sum over @xmath27 in eq.([w12 ] ) gives the correction to the stiff rod limit of the end - to - end distribution function .
thus , the resummation of the moment expansion of @xmath4 according to ( [ w12 ] ) yields the short - chain expansion for kratky - porod chain . the above derivation does not allow to make claims on the convergence of the short - chain expansion .
the comparison of results of computations of quantities under consideration ( for example the scattering function or deformation - force relation ) for different @xmath82 gives a criterion determining the quality of the short - chain expansion .
notice that alternatively one can first carry out the inverse laplace transform of ( [ mom - exp ] ) , and then expand the moments in taylor series in powers of @xmath1 . as a result one will arrive at double series similar to ( [ w12 ] ) with @xmath14 replaced by @xmath1 .
this makes more clear the meaning of the resummation procedure . to leading order ( @xmath78 )
one replaces the moments by its stiff rod behaviour , @xmath83 . in the next to
leading order ( @xmath84 ) one takes into account the next - order corrections to the stiff rod behaviour of all moments , and so on . due to the fact that the scattering function of the semiflexible polymer @xmath85 is the inverse laplace transform of @xmath59 multiplied with the factor @xmath60 , the short - chain expansion of @xmath4 enables one to get the short - chain expansion of the scattering function .
in fact , hermans and ullman @xcite derived the stiff rod limit of the scattering function using the stiff rod limit of the moments . in following subsections
we will consider separately the short - chain expansion of the end - to - end distribution function , the scattering function , and the extension - force relation , which also can be obtained from @xmath17 .
the series over @xmath10 in ( [ w12 ] ) for known coefficients @xmath68 can be easily expressed through the derivatives of the geometric series .
carrying out the inverse laplace transform over @xmath5 gives the short - chain expansion of @xmath15 .
we have computed @xmath15 by taking into account eight terms in the sum over @xmath27 in ( [ w12 ] ) i.e. @xmath86 .
to save place we present below the result for @xmath87@xmath88the subscript at @xmath89 and at quantities below is @xmath79 .
since @xmath90 tends to one for @xmath91 , the end - to - end distribution function is normalized .
it is easily to see that the terms in ( [ gkn ] ) can be represented as derivatives of @xmath92 with respect to @xmath1 i.e. of @xmath93 .
the short - chain expansion of the distribution function of the end - to - end distance is then immediately obtained from ( [ gkn ] ) as @xmath94 where @xmath95 denotes the kth derivative of the dirac s delta - function .
note that @xmath0 is a distribution function with respect to @xmath96 , while the contour length @xmath1 is a parameter .
this should be taken into account in using eq.([grn ] ) to compute different mean values , for example the moments of the end - to - end distribution function of a short chain .
similar to the corresponding expansion of the function @xmath97 in powers of @xmath98 , the short - chain expansion of @xmath0 can not be interpreted as a taylor series but is rather a generalized taylor expansion in space of distributions .
the expansions ( [ gkn]-[grn ] ) can be used to test approximative expressions of the end - to - end distribution function .
we did not succeed so far to convert the short - chain expansion of the distribution function to a close expression .
the scattering function of a semiflexible polymer is defined by @xmath99expressing @xmath100 in ( [ w17 ] ) through the tangent vectors , @xmath101 , and representing the average in ( [ w17 ] ) using the formalism of the quantum rigid rotator yields that the scattering function of the semiflexible polymer @xmath102 is the inverse laplace transform of @xmath103 multiplied with the factor @xmath60 @xcite,@xcite .
thus , the short - chain expansion of @xmath4 enables one to derive in a straightforward way the short - chain expansion of the scattering function .
taking into account the first four terms in the sum over @xmath27 in ( [ w12 ] ) results in the following expression for the scattering function @xmath104where @xmath105 , and @xmath106is the scattering function of a stiff rod .
the plot of the scattering function multiplied by @xmath107 for different values of @xmath79 is shown in fig.[fig1 ] .
for chain length @xmath108 .
dots : stiff rod ; continuous : @xmath87 ; dashes : @xmath86 .
, width=240 ] the accuracy of the computations is determined by the values of @xmath105 , where curves corresponding to different @xmath79 begin to diverge .
the fig.[fig1 ] shows that for @xmath108 the continuous and dashed curves begin to diverge for @xmath109 .
the partition function of a semiflexible polymer with one end fixed and force @xmath110 applied to the another end @xmath111can be expressed through the distribution function of the end - to - end distance as follows @xmath112using the definition of the free energy @xmath113 the extension - force relation can be expressed through the partition function as @xmath114thus , the short - chain expansion of the extension - force relation is directly obtained from that of the fourier transform of the distribution function of the end - to - end distance ( [ gkn ] ) . taking into account the first @xmath115 terms in the series in ( [ w12 ] ) results in the following expression for the extension - force relation @xmath116 where @xmath117 .
the first two terms on the right - hand side of ( [ r - f ] ) gives the extension - force relation for stiff rod .
the log - log plot of @xmath118 for different @xmath79 and for stiff rod is shown in fig.[fig4 ] . .
dots : stiff rod ; continuous : @xmath115 ; dashes : @xmath119.,width=240 ] the slope for the stiff rod is @xmath120 , while the slope for the worm - like chain approaches the value @xmath121 , which is the asymptotic result for finite worm - like chain at large forces @xcite .
fig.[fig4 ] shows that the convergence of the short - chain expansion for the extension - force relation is worse than that of the scattering function .
to conclude , using the exact computation of a large number of moments of the end - to - end distribution function @xmath0 of the worm - like chain , which are obtained from the representation of the distribution function as a matrix element of the infinite order matrix @xcite,@xcite , we have established the analytical form of the coefficients in taylor expansions of the moments for short @xmath1 .
the knowledge of these coefficients enabled us to resummate the moment expansion of @xmath0 by taking into account consecutively the deviations of the moments from their stiff rod limit . within this procedure
we have derived the short - chain expansion for the distribution function of the end - to - end polymer distance , the scattering function , and the extension - force relation , by taking into account the deviations of the moments from their stiff rod limit to the seventh order in @xmath1 .
the procedure can be extended to higher orders .
the short - chain expansion could be useful in studies of the behaviour of short polymers , where the deviation from stiff rod is small .
99 s. f. edwards , proc . phys . soc . * 85 * , 613 ( 1965 ) . | using the exact computation of a large number of moments of the distribution function of the end - to - end distance @xmath0 of the worm - like chain , we have established the analytical form of the coefficients in taylor expansions of the moments for short chain lengths @xmath1 .
the knowledge of these coefficients enabled us to resummate the moment expansion of @xmath0 by taking into account consecutively the deviations of the moments from their stiff rod limit . within this procedure
we have derived the short - chain expansion for @xmath0 , the scattering function , and the extension - force relation , which take into account the deviations of the moments from their stiff rod limit to the seventh order in @xmath1 . |
the the relativistic mean - field theory ( rmf ) @xcite , which is one of the main applications of quantum hadrodynamics ( qhd ) , is successful in nuclear structure studies @xcite . the nonlinear ( nl ) walecka model ( nlwm ) , based on the rmf approach , has been extensively used to study the properties of nuclear and neutron matter , @xmath1-stable nuclei , and then extended to the drip - line regions @xcite . in recent years some authors @xcite have stressed the importance of including the scalar isovector virtual @xmath2 field in hadronic effective field theories when asymmetric nuclear matter is studied .
the inclusion of the @xmath0 meson leads to the structure of relativistic interactions , where a balance between an attractive ( scalar ) and a repulsive ( vector ) potential exists .
the @xmath0 meson plays the role in the isospin channel and mainly affects the behavior of the system at high density regions and so is of great interest in nuclear astrophysics .
the properties of nuclear matter at high density play a crucial role for building models of neutron stars ( ns ) .
neutron stars are objects of highly compressed matter . the structure of a compact star is characterized by its mass and radius , which are obtained from appropriate equation of state ( eos ) at high densities .
the eos can be derived either from relativistic or potential models . in order to describe the medium dependence of nuclear interactions
, a density dependent relativistic hadron ( ddrh ) field theory has been recently proposed @xcite .
recently the authors @xcite used the density dependent coupling models with the @xmath0 meson being included to study the neutron stars .
they found that the introduction of the @xmath0 meson in the constant coupling model leads to heavier neutron stars in a nucleon - lepton picture .
the neutron star masses in the density dependent models can be reduced when the @xmath0 meson is taken into account .
the in - medium modification to the masses of @xmath3 , @xmath4 , and @xmath5 mesons has been studied in experiments and theoretical approaches for a decade @xcite .
recently the authors of ref .
@xcite investigated the effect of in - medium meson masses on the properties of the nuclear matter in the walecka model and the effective masses of @xmath3 and @xmath4 mesons in the nuclear medium were calculated by taking into account the effects of the vacuum fluctuation ( vf ) . in this work
we want to see the vf effects on asymmetric matter and neutron stars .
we also want to clarify the density dependence of in - medium nucleon and meson masses .
the vf effects are naturally introduced by considering loop corrections to the self - energies of in - medium nucleons and mesons .
the effective masses of nucleons and mesons ( @xmath3 , @xmath4 , @xmath5 , and @xmath0 ) in the nuclear medium will be calculated in the vf - rmf model .
the vf effects on asymmetric matter and neutron stars will be studied .
this paper is organized as follows . in sec .
ii , we derive the in - medium effective masses of nucleons and mesons and the eos for nuclear matter in vf - rmf model . in sec .
iii , we compare our results with those of the nl - rmf model . in sec .
iv , a brief summary is presented .
the relativistic lagrangian density of the interacting many - particle system consisting of nucleons , isoscalar ( scalar @xmath3 , vector @xmath4 ) , and isovector ( scalar @xmath0 , vector @xmath5 ) mesons used in this work is @xmath6\psi \nonumber\\ & & + \frac{1}{2}(\partial_{\mu}\phi\partial^{\mu}\phi - m_{\sigma}^{2}\phi^2)-u(\phi ) + \frac{1}{2}m_{\omega}^{2}\omega_{\mu}\omega^{\mu } \nonumber\\ & & + \frac{1}{2}m_{\rho}^{2}\vec{b}_{\mu}\cdot\vec{b}^{\mu } + \frac{1}{2}(\partial_{\mu}\vec{\delta}\partial^{\mu}\vec{\delta } -m_{\delta}^{2}\vec{\delta}^2 ) \nonumber\\ & & -\frac{1}{4}f_{\mu\nu}f^{\mu\nu}-\frac{1}{4}\vec{g}_{\mu\nu}\vec{g}^{\mu\nu } + \delta{\cal l},\end{aligned}\ ] ] where @xmath7 , @xmath8 , @xmath9 , and @xmath10 represent @xmath3 , @xmath4 , @xmath5 , and @xmath0 meson fields , respectively , @xmath11 is the nonlinear potential of the @xmath3 meson , @xmath12 , and @xmath13 . in order to remove the divergence in the loop calculation , the counterterm for the lagrangian density , @xmath14 , included in eq .
( 1 ) reads @xmath15 where the parameters @xmath16 , @xmath17 , and @xmath18 ( @xmath19=1 , 2 , 3 , 4 ; @xmath20=@xmath3 , @xmath4 , @xmath5 , @xmath0 ) are fixed by the renormalization methods suggested in refs .
@xcite .
the field equations in the rmf approximation are @xmath21\psi&=&0,\label{eq:6 - 1}\nonumber\\ m_{\sigma}^{2}\phi+a\phi^2+b\phi^3&=&g_{\sigma}\rho^{s},\label{eq:6 - 2}\nonumber\\ m_{\omega}^{2}\omega_{0}&=&g_{\omega}\rho,\label{eq:6 - 3}\nonumber\\ m_{\rho}^{2 } b_{0}&=&g_{\rho}\rho_{3},\label{eq:6 - 4}\nonumber\\ m_{\delta}^{2 } \delta_{3}&=&g_{\delta}\rho_{3}^{s},\end{aligned}\ ] ] where @xmath22 and @xmath23 , where @xmath24 and @xmath25 ( the index ( subscript or superscript ) @xmath19 denotes proton or neutron throughout this paper ) are the nucleon and scalar densities , respectively .
the nucleon and the scalar densities are given by , respectively , @xmath26 and @xmath27 where @xmath28 is the fermi momentum of the nucleon and @xmath29 is the nucleon propagator in the vf - rmf model : @xmath30\nonumber\\ & \equiv&g^{i}_{f}(k)+g^{i}_{d}(k),\end{aligned}\ ] ] where @xmath31 , @xmath32 is the nucleon effective masses , and @xmath33 is infinitesimal . here @xmath34 describes the free propagation of positive and negative energy quasi - nucleons .
@xmath35 describes quasi - nucleon holes inside the fermi sea and corrects the propagation of positive energy quasi - nucleons by the pauli exclusion principle .
denotes nucleon and @xmath36 is the four momentum of the meson . ]
the nucleon effective mass without the @xmath0 field in the rmf approach is @xmath37 . when the vf effects are considered , the loop - diagram corrections to the self - energy of nucleons and mesons shown in fig .
1 are naturally introduced .
the nucleon effective mass without the @xmath0 field in the rmf approach including the vacuum fluctuations can be calculated from fig .
thus , we have @xmath38 + \frac{a}{m_{\sigma}^{\star2}g_{\sigma}}(m^{\star}-m)^2 -\frac{b}{m_{\sigma}^{\star2}g_{\sigma}^{2}}(m^{\star}-m)^3\nonumber\\ & = & m-\frac{g_{\sigma}^{2}}{2\pi^2m_{\sigma}^{\star2 } } \bigl[k_{f_{p}}m^{\star}e_{f_{p}}^{\star}-m^{\star3}{\rm ln}(\frac{k_{f_{p}}+e_{f_{p}}^{\star}}{m^{\star}})\bigr ] \nonumber\\ & & -\frac{g_{\sigma}^{2}}{2\pi^2m_{\sigma}^{\star2 } } \bigl[k_{f_{n}}m^{\star}e_{f_{n}}^{\star}-m^{\star3}{\rm ln}(\frac{k_{f_{n}}+e_{f_{n}}^{\star}}{m^{\star}})\bigr ] \nonumber\\ & & + \frac{g_{\sigma}^{2}}{\pi^2m_{\sigma}^{\star2 } } \bigl[m^{\star3}\ln(\frac{m^{\star}}{m})-m^2(m^{\star}-m)-\frac{5}{2}m(m^{\star}-m)^2-\frac{11}{6}(m^{\star}-m)^3\bigr]\nonumber\\ & & + \frac{a}{m_{\sigma}^{\star2}g_{\sigma}}(m^{\star}-m)^2 -\frac{b}{m_{\sigma}^{\star2}g_{\sigma}^{2}}(m^{\star}-m)^3,\end{aligned}\ ] ] where the second term in the right hand side of the first line in eq . ( 7 ) is given by fig .
1(a ) , and the third and fourth terms are the contributions from the nonlinear potential of the @xmath3 meson .
as well known , the @xmath0 meson leads to a definite splitting of proton and neutron effective masses , the nucleon effective masses with the @xmath0 meson in the rmf approach are given by , respectively , @xmath39 and @xmath40 the nucleon effective masses with the @xmath0 meson in the rmf approach including the vacuum fluctuations can be calculated from fig .
1(a ) , @xmath41\nonumber\\ & & + \frac{ig_{\delta } ^{2}}{m_{\delta}^{\star 2}}\int \frac{{\rm d } ^4k}{(2\pi)^4}[{\rm tr}g^{p}(k)-{\rm tr}g^{n}(k)]\nonumber\\ & & -a\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^2}{(-2g_{\sigma})^3 } -b\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^3}{(-2g_{\sigma})^4},\end{aligned}\ ] ] and @xmath42\nonumber\\ & & + \frac{ig_{\delta } ^{2}}{m_{\delta}^{\star 2}}\int \frac{{\rm d } ^4k}{(2\pi)^4}[{\rm
tr}g^{n}(k)-{\rm tr}g^{p}(k)]\nonumber\\ & & -a\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^2}{(-2g_{\sigma})^3 } -b\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^3}{(-2g_{\sigma})^4}.\end{aligned}\ ] ] thus , we get after the momentum integral @xmath43 \nonumber\\ & & + \frac{1}{2\pi^2}(\frac{g_{\sigma}^{2}}{m_{\sigma}^{\star2 } } + \frac{g_{\delta}^{2}}{m_{\delta}^{\star2 } } ) \bigl[m^{\star3}_{p}\ln(\frac{m^{\star}_{p}}{m})-m^2(m^{\star}_{p}-m)\nonumber\\ & & -\frac{5}{2}m(m^{\star}_{p}-m)^2-\frac{11}{6}(m^{\star}_{p}-m)^3\bigr]\nonumber\\ & & -\frac{1}{2\pi^2}(\frac{g_{\sigma}^{2}}{m_{\sigma}^{\star2 } } -\frac{g_{\delta}^{2}}{m_{\delta}^{\star2 } } ) \bigl[k_{f_{n}}m^{\star}_{n}e_{f_{n}}^{\star}-m_{n}^{\star3}{\rm ln}(\frac{k_{f_{n}}+e_{f_{n}}^{\star}}{m_{n}^{\star}})\bigr ] \nonumber\\ & & + \frac{1}{2\pi^2}(\frac{g_{\sigma}^{2}}{m_{\sigma}^{\star2 } } -\frac{g_{\delta}^{2}}{m_{\delta}^{\star2 } } ) \bigl[m^{\star3}_{n}\ln(\frac{m^{\star}_{n}}{m})-m^2(m^{\star}_{n}-m)\nonumber\\ & & -\frac{5}{2}m(m^{\star}_{n}-m)^2-\frac{11}{6}(m^{\star}_{n}-m)^3\bigr]\nonumber\\ & & -a\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^2}{(-2g_{\sigma})^3 } -b\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^3}{(-2g_{\sigma})^4},\end{aligned}\ ] ] and @xmath44 \nonumber\\ & & + \frac{1}{2\pi^2}(\frac{g_{\sigma}^{2}}{m_{\sigma}^{\star2 } } -\frac{g_{\delta}^{2}}{m_{\delta}^{\star2 } } ) \bigl[m^{\star3}_{p}\ln(\frac{m^{\star}_{p}}{m})-m^2(m^{\star}_{p}-m)\nonumber\\ & & -\frac{5}{2}m(m^{\star}_{p}-m)^2-\frac{11}{6}(m^{\star}_{p}-m)^3\bigr]\nonumber\\ & & -\frac{1}{2\pi^2}(\frac{g_{\sigma}^{2}}{m_{\sigma}^{\star2 } } + \frac{g_{\delta}^{2}}{m_{\delta}^{\star2 } } ) \bigl[k_{f_{n}}m^{\star}_{n}e_{f_{n}}^{\star}-m_{n}^{\star3}{\rm ln}(\frac{k_{f_{n}}+e_{f_{n}}^{\star}}{m_{n}^{\star}})\bigr ] \nonumber\\ & & + \frac{1}{2\pi^2}(\frac{g_{\sigma}^{2}}{m_{\sigma}^{\star2 } } + \frac{g_{\delta}^{2}}{m_{\delta}^{\star2 } } ) \bigl[m^{\star3}_{n}\ln(\frac{m^{\star}_{n}}{m})-m^2(m^{\star}_{n}-m)\nonumber\\ & & -\frac{5}{2}m(m^{\star}_{n}-m)^2-\frac{11}{6}(m^{\star}_{n}-m)^3\bigr]\nonumber\\ & & -a\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^2}{(-2g_{\sigma})^3 } -b\frac{2g_{\sigma}^{2}}{m_{\sigma}^{\star2}}\frac{(m_{p}^{\star}+m_{n}^{\star}-2m)^3}{(-2g_{\sigma})^4},\end{aligned}\ ] ] where @xmath45 are
the in - medium meson masses .
the effective mass of the meson ( or in - medium meson mass ) is defined as the pole position of the meson propagator at zero three - momentum ( on - shell ) or zero four - momentum ( off - shell ) .
we calculate the in - medium meson masses in the random phase approximation ( rpa ) @xcite .
the corresponding diagrams are given in fig .
thus , we obtain the in - medium masses of scalar mesons : @xmath46 where @xmath47 the expressions of @xmath48 ( on - shell and off - shell ) are , respectively , @xmath49\nonumber\\ & & -\frac{g^{2}_{j}}{2\pi^2}\sum\limits_{i = n , p}\biggl\{\frac{3}{2}(m_{i}^{\star2}-m^{2})\int\limits_{0}^{1 } { \rm d}x \ln\bigl(1-\frac{m_{j}^{2}}{m^{2}}x(1-x)\bigr)\nonumber\\ & & + \frac{3}{2}\int\limits_{0}^{1}{\rm d}x\bigl[(m_{i}^{\star2}-m_{j}^{\star2}x(1-x ) ) \ln\bigl(\frac{m_{i}^{\star2}-m_{j}^{\star2}x(1-x)}{m^{2}-m_{j}^{2}x(1-x)}\bigr)\bigr]\nonumber\\ & & -\frac{m_{j}^{2}-m_{j}^{\star2}}{4}-3m(m_{i}^{\star}-m ) -\frac{9}{2}(m_{i}^{\star}-m)^2\biggr\},\end{aligned}\ ] ] and @xmath50\nonumber\\ & & -\frac{3g_{j}^{2}}{4\pi^2}\sum\limits_{i = n , p}\bigl[2m_{i}^{\star2 } \ln\bigl(\frac{m_{i}^{\star}}{m}\bigr)-m^{2}+4mm_{i}^{\star}-3m_{i}^{\star2}\bigr].\end{aligned}\ ] ] the effective masses of vector mesons can be obtained from fig .
1(b ) @xmath51 where @xmath52 is the transverse part of the polarization tensor as follows : @xmath53 so the expressions of @xmath52 ( on - shell and off - shell ) are , respectively , @xmath54\nonumber\\ & & -\frac{g_{j}^{2}m_{j}^{\star2}}{2\pi^2}\sum\limits_{i = n , p}\int\limits_{0}^{1}{\rm d}xx(x-1 ) \ln\bigl[\frac{m_{i}^{\star2}-m_{j}^{\star2}x(1-x)}{m^{2}-m_{j}^{2}x(1-x)}\bigr],\end{aligned}\ ] ] and @xmath55 we note that the vf contributions are included in the second summations of @xmath56 and @xmath57 ( eqs .
( 16 ) , ( 17 ) and ( 20 ) ) , the vf contribution in eq .
( 21 ) equals zero .
the in - medium meson propagator is significantly modified by the interaction of the mesons with the nucleons .
clearly , this modification of the meson propagators will change the nucleon effective mass as well as the energy density .
the meson propagators in the tadpole diagram are calculated at zero four momentum transfer .
so we must replace the meson mass appearing in the nucleon effective mass and the energy density by the meson effective mass defined as @xcite : @xmath58 and @xmath59 therefore , the energy - momentum tensor in the vf - rmf model can be expressed as @xmath60.\end{aligned}\ ] ] thus the eos for nuclear matter in the vf - rmf model can be obtained .
the energy density is given by @xmath61+u(\phi ) \nonumber\\ & & -\frac{1}{8\pi^2}\sum\limits_{i = n , p}\bigl[m_{i}^{\star4}{\rm ln}(\frac{m_{i}^{\star}}{m})+m^{3}(m - m_{i}^{\star } ) -\frac{7}{2}m^2(m - m_{i}^{\star})^2\nonumber\\ & & + \frac{13}{3}m(m - m_{i}^{\star})^3-\frac{25}{12}(m - m_{i}^{\star})^4\bigr],\end{aligned}\ ] ] and the pressure is @xmath62-u(\phi)\nonumber\\ & & + \frac{1}{8\pi^2}\sum\limits_{i = n , p}\bigl[m_{i}^{\star4}{\rm ln}(\frac{m_{i}^{\star}}{m})+m^{3}(m - m_{i}^{\star } ) -\frac{7}{2}m^2(m - m_{i}^{\star})^2\nonumber\\ & & + \frac{13}{3}m(m - m_{i}^{\star})^3-\frac{25}{12}(m - m_{i}^{\star})^4\bigr].\end{aligned}\ ] ] the density dependence of the nuclear symmetry energy , @xmath63 , is one of the basic properties of asymmetric nuclear matter for studying the structure of neutron stars .
empirically , we have information on @xmath63 only at the saturation point , where it ranges from 28 to 35 mev according to the nuclear mass table @xcite .
the nuclear symmetry energy is defined through the expansion of the binding energy in terms of the asymmetry parameter @xmath64 @xcite : @xmath65 where the binding energy density is defined as @xmath66 , and the asymmetry parameter @xmath67 .
the nuclear symmetry energy is defined by @xmath68 according to the definition , an explicit expression for the symmetry energy in the vf - rmf model is obtained as @xmath69,\end{aligned}\ ] ] where @xmath70,\ ] ] and @xmath71 as discussed in refs .
@xcite , the incompressibility @xmath72 is one of the important ingredients for the nuclear eos .
the incompressibility of nuclear matter is defined by @xcite @xmath73 where @xmath74 is the baryon chemical potential and @xmath75 is the saturation density .
we can easily obtain the incompressibility from the eos in the vf - rmf model as @xmath76 where @xmath77 and @xmath78 the final outcome of a supernova explosion can be a neutron star or a black hole .
the neutron star is believed to evolve from an initially hot protoneutron star .
the matter in cold neutron stars is in the ground state in nuclear equilibrium .
matter in equilibrium concerning weak interactions is called as @xmath1-equilibrium matter .
the composition of @xmath1-equilibrium system is determined by the request of charge neutrality and chemical - potential equilibrium @xcite .
the balance processes for @xmath1-equilibrium ( @xmath79 ) system are the following weak reactions : @xmath80 the chemical - potential equilibrium condition for ( @xmath79 ) system can be written as @xmath81 where the electron chemical - potential @xmath82 , @xmath83 is the electron momentum at the fermion level and @xmath84 is the electron mass .
the charge neutrality condition is @xmath85 where @xmath86 is the electron density , and the proton fraction @xmath87 . in the ultra - relativistic limit for noninteracting electrons ,
the electron density can be expressed as a function of its chemical potential @xmath88 then , we can obtain the relation between the proton fraction @xmath89 and the nuclear symmetry energy @xmath63 @xmath90^{3}=0.\ ] ] the eos for @xmath1-equilibrium ( @xmath79 ) matter can be estimated by using the values of @xmath89 , which can be obtained by solving eq .
the properties of the neutron stars can be finally studied by solving tolmann - oppenheimer - volkov ( tov ) equations @xcite with the derived nuclear eos as an input .
in order to make a comparison with the nl - rmf model @xcite , the same saturation properties of nuclear matter and hadron masses are listed in table i , which are used to determine the model parameters .
the obtained model parameters are presented in table ii with the nl - rmf model parameters together for a comparison .
the coupling constants are defined as @xmath91 ( @xmath92 , @xmath4 , @xmath5 , @xmath0 ) in refs .
@xcite . the parameters of self - interacting terms in table ii are defined as @xmath93 and @xmath94 .
* table i. * saturation properties of nuclear matter and hadron masses . [ cols="^,^,^",options="header " , ] we note that the coupling constant @xmath91 in the nl - rmf model is a constant fixed by the saturation properties of nuclear matter @xcite . in order to make a comparison , we define @xmath95 ( @xmath20=@xmath3 , @xmath4 , @xmath5 , @xmath0 ) in our model , here @xmath96 is the off - shell meson mass .
we are interested in the vf effects on the coupling constants .
it is well known that the coupling constants in the ddrh model are density dependent ( see @xcite ) , whereas the meson masses are constant . in the present model ,
the coupling constants are constant , but the meson masses are density dependent . in order to distinguish , we define @xmath97 is the coupling constant ) for the ddrh@xmath5 case
. we present a comparison between @xmath98 and @xmath99 in fig . 7 .
we see from fig .
7 that @xmath98 decreases with the increase of the baryon density for both @xmath3 and @xmath4 mesons , which are due to the increase of the off - shell meson masses in high density regions .
it shows that the variance trends of @xmath98 and @xmath99 are roughly the same for @xmath3 , @xmath4 , and @xmath5 mesons .
if we attribute the density dependence of the coupling constants in the ddrh model to the vf effects in our model , the density dependence of @xmath98 and @xmath99 should have the same physical origin . and @xmath99.,title="fig : " ] -3.5 cm
the vf corrections are investigated in the framework of the rmf approximation by using the relativistic lagrangian density with the @xmath0 field in this work . by taking into account the loop corrections to the self - energies of the in - medium nucleons and mesons ,
the vf effects are naturally introduced into the rmf model . in order to make a comparison with the nl - rmf model ,
the same saturation properties of nuclear matter are used to determine the parameters of the vf model .
we calculate the contributions from self - energy diagrams to the masses of the nucleons and mesons .
the effective masses of the nucleons and mesons , especially the @xmath5 and @xmath0 mesons , in the nuclear medium are obtained .
we find that the nucleon effective masses decrease more slowly with the increase of the baryon density comparing with the nl - rmf case .
we also find that the dependence of the off - shell meson masses on the density in the medium is different from that of the on shell meson masses ( see fig .
2 ) . the effects of the in - medium hadron masses on the nuclear eos and the properties of the neutron stars are studied .
the vf effects lead to softness of the symmetry energy and the eos of asymmetric matter . due to such softness
, the neutron star masses are reduced quite a lot .
this indicates that the vf effects on the neutron stars are important .
we see from fig .
2(b ) that the off - shell in - medium meson masses increase with the increase of the baryon density .
the off - shell in - medium meson masses can be used to calculate @xmath98 .
the variance trends of @xmath98 for @xmath3 , @xmath4 , and @xmath5 mesons are roughly consistent with @xmath99 in the ddrh@xmath5 case @xcite .
the density dependence behavior of the off - shell in - medium meson masses is very interesting indeed . in the present work , we work in the hartree approximation in which we only consider the dominant contributions from tadpole diagrams to the nucleon self - energy when we investigate the vf effects .
in fact , the exchange diagram contributions can only provide small corrections to the eos in the rmf approach at high densities @xcite .
it is well known that the symmetries of the infinite nuclear matter system can simplify the mean - field lagrangian considerably . as pointed out in ref .
@xcite , when translational and rotational invariance of the nuclear matter is taken into account , the expectation values of all three - vector fields must vanish .
therefore , the expectation value of @xmath100 in eq .
( 1 ) is zero and nonlinear @xmath5 meson interactions ( three or four @xmath5 vertices ) do not appear .
furthermore , since we only take into account the tadpole diagrams , only the neutral iso - vector meson ( @xmath101 ) is involved in the nucleon self - energy diagrams even in the asymmetric nuclear matter ( @xmath102 could contribute to the exchange diagrams which are ignored in the hartree approximation ) .
it is suggested that the tensor coupling to the nucleon should be taken into account for the study of the vector mesons in nuclear medium @xcite . in the present work ,
since we focus our attention on the vf effects on the properties of neutron stars , we do not include the tensor coupling effects in the calculations for the effective masses of the vector mesons
. this will be studied in the future work , especially for the study of the @xmath5 meson masses in the medium .
furthermore , the future study of the vf effects on the meson - nucleon coupling constants will be carried out .
in addition , more careful study of the high density behavior of the meson - nucleon effective couplings , especially @xmath103 , is important and attractive . this project is supported by the national natural science foundation of china ( project nos .
10675022 and 10875160 ) , the key project of chinese ministry of education ( project no .
106024 ) , and the special grants from beijing normal university .
walecka , ann .
( n.y . ) * 83 * , 491 ( 1974 ) .
serot , j.d .
walecka , adv .
* 16 * , 1 ( 1986 ) ; int .
e * 6 * , 515 ( 1997 ) .
y. sugahara and h. toki , nucl .
a * 579 * , 557 ( 1994 ) .
lalazissis , j. knig , and p. ring , phys .
c * 55 * , 540 ( 1997 ) .
sharma , a.r .
farhan , and s. mythili , phys .
c * 61 * , 054306 ( 2000 ) .
j. boguta and a.r .
bodmer , nucl .
a * 292 * , 413 ( 1977 ) .
bodmer , nucl .
a * 526 * , 703 ( 1991 ) .
h. meller and b.d .
serot , phys .
c * 52 * , 2072 ( 1995 ) .
furnstahl and b.d .
serot , nucl .
a * 671 * , 447 ( 2000 ) .
h. meller and b.d .
serot , nucl .
phys . a * 606 * , 508 ( 1996 ) .
b. liu , v. greco , v. baran , m. colonna , and m.d .
toro , phys .
c * 65 * , 045201 ( 2002 ) .
b. liu , h. guo , m.d .
toro , and v. greco , eur .
j. a * 25 * , 293 ( 2005 ) .
d.p . menezes and c. provid@xmath104ncia , phys .
c * 70 * , 058801 ( 2004 ) .
t. gaitanos , m.d .
toro , s. typel , v. baran , c. fuchs , v. greco , and h.h .
wolter , nucl .
a * 732 * , 24 ( 2004 ) . v. baran , m. colonna , v. greco , and m.d .
toro , phys .
rep . * 410 * , 335 ( 2005 ) .
b. liu , m.d .
toro , v. greco , c.w .
shen , e.g. zhao , and b.x .
sun , phys .
c * 75 * , 048801 ( 2007 ) . c. fuchs , h. lenske , and h.h .
wolter , phys .
c * 52 * , 3043 ( 1995 ) .
f.d . jong and h. lenske , phys .
c * 57 * , 3099 ( 1998 ) ; f. hofmann , c.m .
kell , and h. lenske , ibid .
* 64 * , 034314 ( 2001 ) .
s. typel and h.h .
wolter , nucl .
a * 656 * , 331 ( 1999 ) .
brown and m. rho , phys .
lett . * 66 * , 2720 ( 1991 ) .
t. hatsuda and s.h .
lee , phys .
c * 46 * , r34 ( 1992 ) .
s. sarkar _
et al . _ ,
a * 634 * , 206 ( 1998 ) .
et al . _ ,
rev . lett . * 86 * , 5019 ( 2001 ) .
d. trnka _ et al . _
( cbelsa / taps ) , phys .
lett . * 94 * , 192303 ( 2005 ) . b. krusche , int .
a * 22 * , 406 ( 2007 ) .
r. nasseripour _ et al . _
( clas ) , phys .
* 99 * , 262302 ( 2007 ) .
a. bhattacharyya , s.k .
ghosh , and s.c .
phatak , phys .
c * 60 * , 044903 ( 1999 ) .
h. kurasawa and t. suzuki , nucl .
a * 490 * , 571 ( 1988 ) .
w.d . myers and w.j .
swiatecki , ann .
* 55 * , 395 ( 1969 ) .
n.k . glendenning , _ compact stars , nuclear physics , particle physics and general relativity _
( springer - verlag , new york , berlin , heidelberg , 2000 ) .
h. kouno , _ et al .
c * 51 * , 1754 ( 1995 ) .
lattimer and m. prakash , science * 304 * , 536 ( 2004 ) .
tolman , phys . rev . * 55 * , 364 ( 1939 ) ; j.r .
oppenheimer and g.m .
volkoff , phys . rev . * 55 * , 374 ( 1939 ) . r.s .
hayano and t. hatsuda , nucl - ex/08121702 .
m. post , s. leupold and u. mosel , nucl .
a * 741 * , 81 ( 2004 ) . c. maieron , m. baldo , g.f .
burgio , and h.j .
schulze , phys .
d * 70 * , 043010 ( 2004 ) .
r. machleidt , k. holinde and c. elster , phys . rep .
* 149 * , 1 ( 1987 ) . | the vacuum fluctuation ( vf ) effects on asymmetric nuclear matter are investigated .
masses of nucleons and mesons are modified in the nuclear medium by calculating the loop - diagram corrections and the density dependence of hadron masses is obtained . the relativistic lagrangian density with the isovector scalar @xmath0 meson is used to calculate the nuclear equation of state ( eos ) in the framework of the relativistic mean - field ( rmf ) approach ,
the effects of the in - medium hadron masses on the properties of neutron stars are finally studied . with the inclusion of the vf corrections ,
the nuclear eos becomes softer and the neutron star masses are reduced . |
recently , optical properties of composites have attracted great interest . in order to open new possibilities in the information processing and transmission
, a large optical nonlinearity may be desirable@xmath1 .
composite materials , especially , small metal particles embedded in a dielectric host and metal clusters on the nanometer scale , exhibit a strong nonlinear optical enhancement through the inhomogeneous local - field and geometric - response effect@xmath2 .
it is also known that these composites give rise to an anomalously large absorption in the infrared spectrum@xmath3 . to analyze the optical responses , it is more convenient to adopt the spectral representation@xmath4 . for various anisotropic composites ,
the spectral density was calculated as a function of volume fractions @xmath5 and @xmath6 by the effective - medium approximation ( ema)@xmath7 .
it was found a large nonlinear optical enhancement , as well as the separation of the absorption peak from the nonlinear enhancement peak .
by ema , the fluctuation of local field has been averaged out .
so it is difficult to find the physical origins of the optical enhancement and the separation of optical peaks . in this connection , greens - function formalism ( gff )
was developed to deal with the optical responses of the arbitrary - shaped metallic clusters embedded in the infinite dielectric networks at the quasistatic limit@xmath8 . by the formalism
, the resonance spectrum and local field distribution for each eigenmode can be analytically obtained .
the aim of this paper is to investigate the linear and nonlinear optical responses throuth the dilute anisotropic networks in view of the local field distribution . in the following ,
a binary dilute anisotropy network is considered .
we use the random generator to produce the geometric anisotropy .
the metallic bonds parallel to the applied field are assigned with the fraction @xmath5 , and metallic bonds perpendicular to the applied field with @xmath6 . in the dilute systems ,
the condition , @xmath9 , is satisfied .
when we change @xmath5 from @xmath10 to @xmath11 , the corresponding @xmath6 varies from @xmath10 to @xmath12 . in section
ii , the green s functions for the effective linear response @xmath13 and effective nonlinear response @xmath14 are derived . in section iii , the inverse participation ratios ( ipr ) with @xmath0 are used to represent the localized and extended eigenstates in anisotropic networks . in section vi , the spectra of the absorption and the third order nonlinear enhancement are illustrated .
when the applied field is parallel to the anisotropy of the networks , the overlap of the absorption peak with nonlinear enhancement peak is found within the interval @xmath15 .
in contrast , for the perpendicular applied field , the separation of the absorption peak from nonlinear enhancement peak is enhanced when the anisotropy @xmath5 is increased .
these results are explained in section iii , numerically confirmed in section vi , and concluded in section v.
consider an infinite binary network as shown in fig .
1 , where the impurity bonds with admittance @xmath16 are employed to replace the bonds in an otherwise homogeneous network of identical admittance @xmath17 .
the admittance of each bond is generally complex and frequency - dependent .
all the impurity bonds construct the clusters subspace .
when resonance happens , the potential of the jointing points can be computed by the gff . instead of the point source @xmath18 ,
the source term is replaced by @xmath19 at the point @xmath20 .
hence in the subspace , @xmath21 is a linear combination of right eigenvector @xmath22 s of greens - matrix m , @xmath23and for the site @xmath24 outside the cluster , @xmath25 becomes @xmath26 in the uniform field @xmath27 along @xmath28 direction , @xmath21 reads @xmath29for a binary network with @xmath30 square lattices , in the quasistatic limit , the displacement @xmath31 of the @xmath32th bond is related to the local field @xmath33 by the relation @xmath34 , where @xmath35 is the dielectric constant and @xmath36 is the third order nonlinear susceptibility of the bond .
@xmath37 are set to be @xmath38 for the impurity bonds , and @xmath39 for the matrix bonds respectively . by the finite difference transformation , the effective linear response along the applied field is @xmath40with the magnitude of field @xmath41 and lattice constant @xmath42 , where @xmath43 means that @xmath20 and @xmath44 are the nearest neighbors in the same cluster while @xmath45 means @xmath46 and @xmath47 are the nearest neighbors but not in the same cluster .
when we employ @xmath48 , @xmath49 can be expressed by the points within the clusters subspace as , @xmath50here @xmath51 , so we have @xmath52because @xmath20 and @xmath44 are neighbors , the values of @xmath53 can be only @xmath54 or @xmath55 . physically , the absorption along the applied field is larger than that along the perpendicular direction of the applied field . in the following calculations , @xmath13 always represents the absorption along the applied field . for simplicity , we let @xmath56 , @xmath57 and @xmath58 , the imaginary part of @xmath59 corresponding to the absorption is @xmath60@xmath61 , referred to as the cross section , obeys the sum rule@xmath62 , @xmath63where @xmath64 is the number of horizontal bonds along the applied field . since the local fields are determined completely , our formulas can be used to calculate the nonlinear response . by relating the total electrostatic energy to the effective coefficients , the third order nonlinearity @xmath65
is defined by@xmath66@xmath67,\]]where @xmath68 is the space averaged electric field . for the infinite networks , @xmath69 and the nonlinear response function
is given by@xmath70 .
@xmath71where the summation is over all bonds and @xmath72 is the(general complex ) potential difference across the bond @xmath32 and for two dimensional case , @xmath73 .
when all the bonds have nonlinear term , i.e. , @xmath74 , with the lattice constant @xmath42 , the effective nonlinear response is written as @xmath75the third order nonlinearity @xmath14 is expressed as the summation of the forth moment of local electric field .
so the fluctuation of local electric field enhances the nonlinear optical responses well .
first , the ipr of eigenvectors of greens - matrix @xmath76 in the eigensystem with @xmath77 jointing points are defined .
the @xmath78th normalized right eigenvector is @xmath79with @xmath80 and @xmath81 .
the ipr of @xmath82 is written as@xmath83 , @xmath84here @xmath0 .
the calculation of ipr@xmath85 is limited in the nontrivial eigenstates .
the number of ipr@xmath85 is equal to or less than @xmath77 . in the above eqs .
( 1 ) and ( 2 ) , it is found that the right eigenvectors of @xmath76 are closely related to the local fields of the impurity cluster in the subspace .
so the ipr can be used to represent the localization of the eigenstates .
the ipr amplify the profiles of eigenstates , namely , the localized states become more pronounced and extended states become smoother .
hence , the larger values of ipr are always corresponding to the stronger optical responses .
this will be verified in the following sections . to consider the size effect ,
the ipr in the dilute isotropic composites with @xmath86 and @xmath87 are shown in fig .
2 . in this figure ,
two samples are in size @xmath88 and @xmath89 .
the peaks represent the localized states and the valleys correspond to the extended states . comparing the distribution of peaks and valleys in two samples , we find that the size effect is not very obvious .
it is also seen that the density of states are larger around @xmath90 than that around @xmath91 or @xmath92 .
the localized states incline to accumulate @xmath91 or @xmath92 , while , the extended states are near @xmath90 .
so there exists a duality about @xmath93 .
the high values of ipr imply the strong optical responses at @xmath91 or @xmath92 . @xmath94 fig .
3 displays the ipr of the dilute anisotropic systems .
we have known that the resonance spectrum is very sensitive to the microstructure@xmath95 .
for each case , comparing two samples , we see that the peaks of ipr are very stable though their microstructure is completely different .
so in the following , for the definite parameters @xmath5 and @xmath6 , it is reasonable to investigate the optical responses by only one sample . when the anisotropy is increased from fig .
3(a ) to fig .
3(g ) , the localized states incline towards @xmath96 or @xmath97 .
the distributions of localized and extended states have roughly the same feature as that in the isotropic case , i.e. , we find more localized states at @xmath91 and @xmath92 than at @xmath90 , but denser at @xmath90 .
the distribution of ipr is caused purely by the morphology of the sample . in the next section
, we will discuss how the different applied fields act on the optical responses when the geometric anisotropy is increased .
the optical responses can be properly described by the absorption spectrum @xmath98 and the third order nonlinear enhancement spectrum as shown in fig . 4 and fig .
when the applied field is parallel to the direction of anisotropy , it affects the optical properties well .
the envelopes of the absorption and third order nonlinear enhancements become narrower and the intensity is more enhanced with the increasing anisotropy .
the main optical peaks are ranged in the interval @xmath99 $ ] , which correspond to the dipolar approximation . in this figure
, we can not find the separation of the absorption peak from the third order nonlinear enhancement peak .
the results of single sample conflict with those of previous works@xmath95 .
the reason is that the ema has averaged out the fluctuation of local field , which plays the central role for the effective optical responses . at @xmath90 , the density of localized states
are much smaller than that of extended states .
the optical properties of the specific sample are determined by its own eigenstates , not by `` effective medium '' .
so it is not reasonable for all of the cases to express the optical properties of one specific sample only by the same macroscopic parameters @xmath5 and @xmath100 . in fig .
5 , when the applied field is perpendicular to the anisotropy , the optical responses of the dilute anisotropic samples are plotted .
it is seen that the peaks of absorption are red - shifted while the peaks of the third order nonlinear enhancement are blue - shifted when the anisotropy is added .
there is a larger peak separation with increasing geometric anisotropy .
for all the applied fields , it is natural to find a large third order nonlinear enhancement and a broad infared absorption because the localized states exist in the whole resonant area . only for the perpendicular applied field ,
the separation of optical peaks is found .
@xmath94 @xmath94 then , the drude free electronic model is employed to calculate the optical properties of the dilute anisotropic composites .
the admittance of the impurity metallic bonds is @xmath101where @xmath102 is the plasma frequency , and @xmath103 a damping constant . for metal , the plasma frequency @xmath104 ,
being in the ultraviolet .
we choose @xmath105 , which is the typical value of a metal , and @xmath106 , which is the dielectric constant of water for model calculations .
the range of optical responses is @xmath107 .
optical responses of various anisotropic networks are illustrated in fig .
the nonlinear enhancement @xmath108(or @xmath14 ) is divided by @xmath109 for the comparison with the linear absorption .
the direction of anisotropy in figs .
6(a ) , ( c ) , ( e ) and ( g ) is parallel to the applied field .
the peak at @xmath110 corresponds to the dipolar approximation of the whole system . in these figures ,
the separation of absorption peak and nonlinear enhancement peak is not found , as observed in the above fig .
4 . note that this result is different from that of the previous works@xmath111 .
however , when the applied field is perpendicular to the anisotropy , as shown in fig .
6(b ) , ( d ) , ( f ) and ( h ) , in each of which , both the separation of linear and nonlinear peaks , and the third order nonlinear enhancement are seen .
in this paper , optical responses of dilute binary anisotropic networks are investigated numerically by the gff .
green s functions for the effective linear and nonlinear responses are derived .
the ipr with @xmath0 are given to describe the localization and extension of eigenstates .
the spectra of the absorption and the third order nonlinear enhancement are exhibited for the different applied fields .
we also compute the optical responses for the drude model . for the dilute anisotropic networks
, we conclude that : 1 .
the size effect of optical responses is not obvious .
2 . the ipr with @xmath0 can be used to represent the localization and extension of eigenstates .
the wide optical absorption and large nonlinear enhancement are caused by the geometric anisotropy .
4 . the peaks of the absorption and nonlinear enhancement overlap when the applied field is parallel to the anisotropy . while , for the perpendicular applied field , the absorption peak is separated from the nonlinear enhancement peak . 5 .
the structure sensitivity does not affect the main optical properties , but the details of optical responses .
> from above 2 , 3 , 4 and the previous works@xmath7 , geometric anisotropy is the main reason to enhance the optical responses . for a specific sample ,
when the different fields are applied , the positions of optical responses do not vary , but the local field distributions are quite different . therefore the optical properties of resonant composites are determined by the geometric anisotropy , as well as by the applied sources .
gu y acknowledges the useful discussions with prof .
yang zhan ru . | we investigate the linear and nonlinear optical responses of dilute anisotropic networks using the greens - function formalism ( gff)[gu y et al .
1999 _ phys . rev .
b _ * 59 * 12847 ] . for the different applied fields ,
numerical calculations indicate that a large third order nonlinear enhancement and a broad infrared absorption arise from the geometric anisotropy .
it is also shown the overlap and separation between the absorption peak and nonlinear enhancement peak when the applied field is parallel , perpendicular to the anisotropy respectively . in terms of the inverse participation ratios ( ipr ) with @xmath0 and spectral distribution of optical responses ,
the results can be understood . |
peer review is the fundamental process used by the scientific community to ensure the quality of academic publications ( cf .
e.g. , @xcite ) .
several generations of scientists have contributed high - quality reviews , while only authorship has been credited for academic career .
it is not easy to rationalize why researchers provide impartial reviews and constructive advice voluntarily , as they need to sacrifice time that could be used for their own research activities @xcite . in consequence , it is a puzzle how the system of peer review can be sustainable at all .
this puzzle can be described as a double social dilemma game @xcite where scientists can choose levels of efforts for both manuscripts and reviews .
given the presence of costs in terms of time and effort , no contribution ( sloppy review ) is the best reply strategy for reviews .
when all scientists play according to their best reply strategy , the resulting outcome is no scientific control on the quality of submitted papers .
the dominant strategy equilibrium of low quality reviews mean that submissions need not be of high quality . due to the immense costs of producing high quality work , authors are best off by submitting poor manuscripts . in short , in the lack of explicit sanctions and incentives , low - quality submissions and low - quality reviews from the dominant strategy equilibrium in the social dilemma of scientific production .
although social dilemmas of this kind are difficult to resolve in general , certain theoretical solutions have already been proposed that can be applied to the context of peer review .
the most evident improvement might come from shifting the payoffs in favor of cooperation .
for instance , the reward for overall cooperation can be introduced by attaching higher importance on scientific quality through the introduction or increased emphasis on journal metrics , such as impact factor .
another straightforward solution is the application of selective incentives @xcite : allocating additional benefits for authors of high - quality papers ( e.g. , promotion and grants , which is current practice ) and for reviewers writing high - quality reports ( which is rare in current practice ) .
direct @xcite and indirect reciprocity @xcite are potential solutions to social dilemmas , if the chance of repetition is high and the opportunities for retaliation are in foresight . in the context of peer review ,
indirect reciprocity can be facilitated , for instance , by rotating the roles of authors , reviewers , and editors @xcite .
the social embeddedness of scientific production further enhances the chance of solving the social dilemma efficiently @xcite .
the small world aspect of working in specific fields , the intertwined network of co - authorship , participation in international project consortia , and hangouts at conferences all reduce the competitiveness and improve the pro - social character of reviewing .
furthermore , once informal communication , gossip , reputation , image scoring , and stratification enters into the structure of the social dilemma , cooperation might emerge and be maintained @xcite . recording , keeping , and relying on reputational scores has become an efficient guideline for cooperation in many areas of life @xcite .
earlier achievement and reputational information certainly plays an important role in current practice both for editorial and reviewer decisions @xcite .
social incentives that are associated with direct and indirect reciprocity and the structural embeddedness of peer review might be key mechanisms that rationalize cooperative behavior of reviewers .
the importance of social incentives is highlighted also by surveys asking for the motivations of reviewers @xcite .
in fact , monetary incentives might be in conflict with or drive out social incentives when applied to reviewers @xcite .
these solutions for social dilemmas might help us to understand how peer review can work and be sustained .
once the fundamental mechanisms are studied rigorously , they can also lead to policy recommendations on improvements of the current system and the design of new solutions ( cf .
@xcite ) .
the rest of this paper is divided in three sections . in section [ themodel ]
, we introduce our agent - based model of scientific work and peer review .
we report simulations results in section [ results ] and we conclude in section [ conclusion ] .
we model the production of scientific work and its peer review by a simple agent - based model . in this paper
, we build on the view that the aim of peer review is to ensure scientific production and evaluation @xcite .
the model emphasizes the costly character of high - quality manuscript submissions and reviewer contributions .
we take it as granted that high - quality submissions lead to better science , but reviews impact scientific quality only indirectly .
our model makes radical simplifications on the practical aspects of the peer review system intentionally .
this way , we aim at providing a straightforward assessment of the institutional conditions and editorial policies under which authors are motivated to produce high - quality submissions and reviewers are motivated to provide high quality reviews . hence , we are primarily interested in the emergence of cooperation that results in scientific quality .
the model contains @xmath0 scientists as agents who write single - author papers . at each discrete period @xmath1
, each scientist performs the task of an author - by producing one article - and the task of a reviewer .
authors and reviewers comprise of an identical set of agents . for the sake of simplicity
, we consider a single journal with a single editor , who is not an author or a reviewer in the journal .
authors can produce low or high quality contributions that they submit to the editor .
the editor selects @xmath2 ( set to two in our simulations ) reviewers for each paper choosen uniformly at random with an upper limit @xmath3 of reviews for each reviewer ( set to four for the sake of our simulations ) .
we assume that reviewers always accept requests , but they do not necessarily invest high effort in performing reviews .
it is of their strategic decision to produce a low quality review at low cost ( normalized to zero ) or a high quality review at high cost .
the former is the best reply strategy of the reviewer if no further incentives are provided .
the review is translated into a binary recommendation ( accept or reject ) which is passed on to the editor .
we assume that the recommendation is random with fifty - fifty percent chance of accept or reject in case of low reviewer effort and reflects the quality of the submission perfectly in case of high reviewer effort .
the editor s decision is based on incoming recommendations and it has a binary outcome : accept or reject .
we consider a single round of review .
acceptance benefits the author , but benefits the editor ( the journal and scientific development in general ) only if the submission was of high quality .
the incentive structure and the strategy space of the game are defined as follows .
the editor wants to maximize high - quality scientific output in the journal and would like to minimize the number of low - quality articles appearing in the journal .
this means that similarly to the model of @xcite , papers that are accepted and of bad quality are the most harmful for the journal .
these values are used as output measures to evaluate the performance of the entire system in our simulations .
the situation is of asymmetric information @xcite , in which the editor is unable to assess the true quality of submissions or the true effort of reviewers .
the true quality of accepted papers is revealed probabilistically after publication ( that we will fix to @xmath4 in the simulations reported ) and there is no way for the editor to assess the true quality of rejected submissions .
the crucial parts of editorial policy are therefore the selection of reviewers who can be trusted for their recommendations and the extent of reliance on reviewer recommendations .
several strategies could possibly be used in order to arrive at proper conclusion .
we vary these editorial strategies in between simulations , because these could be easily translated into policy recommendations . keeping a reputational account of scientists
is part of all such strategies .
editorial reputations of scientists are improved largely by high quality publications ( named @xmath5 ) , degraded even more by low quality publications ( @xmath6 ) , worsened by rejected submissions ( @xmath7 ) , improved by good reviews ( @xmath8 ) and worsened by bad reviews ( @xmath9 ) .
good reviews are those where the paper is revealed of the same quality of what the reviewer said .
bad reviews are those where there is a difference , i.e. where the paper is revealed as bad while the reviewer said that it was good or where the paper was revealed as good , but the reviewer recommended rejection ( in case of conflicting reviews ) . in summary ,
the editorial reputation of scientist @xmath10 is given as : @xmath11 where @xmath12 ( the sum of bad and good rejected papers ) and @xmath13 .
please note the asymmetric character of reviewer bias : as the high quality of rejected papers is never revealed , rejection is a safer strategy for reviewers ( a rejection recommendation can only turn out to be a bad review in case other reviewers recommended publication , the paper has been published , and its true high quality has been revealed ) .
parameter @xmath14 represents the relative detrimental effect for the journal reputation of accepting a low quality paper .
incoming reviews are assumed to lead to editorial conclusions according to an editorial policy which is assumed to be fixed and not updated within a simulation .
we manipulate editorial strategies , between the simulations in order to compare the effectiveness of these policies .
editorial strategies differ with regard to reviewer selection , the handling of conflicting reviews , and relying on author reputations in desk rejection and acceptance .
the last element will only be added to the extended model .
we consider four editorial strategies with regard to reviewer selection and handling of conflicting reviews .
for all of them , if all referees agree , then the editor follows the unanimous advice . in case of _ disagreement _ , the editor can use one of the following strategies : * * ap * : reject ; * * 1p * : accept ; * * er * : follow the advice of one of the referees chosen at random probability proportional to the relative editorial reputations of the referees ; * * mr * : follow the advice of the most reputed referee .
authors have perfect information about the quality of their own submissions . similarly to the model of @xcite , authors decide to submit a paper of low quality ( at low cost ) or at high quality ( at high cost ) .
they are best off with the publication of their low - quality papers .
thus , in terms of obtained payoffs : @xmath15 . for the simulations we assume that all agents receive a unit of endowment @xmath16 ( research time ) in each period , which they lose by investing high effort in writing a high - quality paper ( @xmath17 ) and high - quality reviews ( @xmath18 , where @xmath19 ) .
we assume no gain from not being published ( @xmath20 ) .
being published yields a positive return @xmath21 .
the numerical payoffs in our simulations are : @xmath22 when in the role of reviewers , scientists first accept editorial requests to review submissions .
we assume that this does not entail any significant costs : the time spent on pushing the `` accept '' button in an editorial managerial system is negligible and the social costs of being committed to reviewing are counterbalanced by gaining access to papers before their publication . once papers are assigned to reviewers , they decide to invest low effort ( no loss from the endowment of valuable time ) or high effort in performing the task .
a high reviewing effort implies a deduction from the endowment @xmath18 that is proportional to the ratio between number of good reviews and the number of assigned reviews .
low effort is the best reply strategy of reviewers .
altogether , writing high quality papers as well as writing high quality reviews entails sacrificing valuable research time for scientists . for the sake of simplicity
, we assume that endowments are not transferable between the two kind of activities . in this way , unlike @xcite , we disregard the potential time conflict between writing high - quality papers and reviews . in their model , scientists need to allocate time between submissions and reviewing . in our model ,
both activities are costly if they are performed well . in line with the current duality of practices , two cases
are compared : single blind and double blind reviews @xcite . in case of single blind reviews , reviewers can condition their recommendations on the reputation of the author . in case of double blind reviews ,
only the editor is able to make decisions based on the reputation of authors . in case of a single blind review system , the reviewer strategy can be conditional on the public reputation of the author .
unlike the editorial reputation of authors , the public reputation of an author @xmath10 depends only on _ published _ papers and does not severely punishes bad publications : @xmath23 as reviewers are also authors , author and reviewer strategies are bundled and are characterized by the following elements : * a decision for the production of manuscripts can be either : * * @xmath24 : produce manuscripts of high quality ( cooperation ) * * @xmath25 : produce manuscripts of low quality ( defection ) * a decision for the production of reviews that can be either : * * @xmath26 : produce reviews at high effort ( cooperation ) * * @xmath27 : produce reviews at low effort ( defection ) * * @xmath28 : exercise high effort with a likelihood that is proportional to the public reputation of the author of the paper ( cooperation conditional on reputation ) .
individual strategies are assigned to scientists at the start randomly and in equal numbers for each combination .
this results in four bundled strategies for the double blind case and in six bundled strategies for the single blind case .
authors decide according to their strategy types to submit a paper of low quality or at high quality ( at high cost ) , which is sent out for review according to the rules determined by the editorial policy .
reviewers act according to their strategies and provide recommendations of accept or reject to the editor .
papers are selected for publication as a result of the recommendations of the reviewers and the editorial policy .
the evolution of author / reviewer strategies is modeled with a replicator dynamics rule adjusted to a finite population .
scientists adopt strategies that ensured higher average payoffs in the previous time period in the population , while disregard strategies that resulted in lower payoffs .
specifically , at a given time @xmath1 each individual has a probability @xmath29 of being selected for updating his strategy .
if this happens , then his new strategy is selected randomly with a probability proportional to the difference of the given strategy to the average expected payoff , weighted for the current strategy frequencies .
formally , the expected new population size for strategy @xmath30 is given by : @xmath31 where @xmath32 is the average payoff of strategy @xmath30 , while @xmath33 is the average payoff in the whole population and @xmath34 is the speed of evolution .
we constrain the finite replicator dynamics process such that all strategies have an integer number of representations in the population and no @xmath35 .
note that the strategies of agents evolve and not the agents are replaced .
this means that agents accumulate editorial and public reputation throughout the simulation . in order to keep the model simple and to concentrate on some key mechanisms , other details
are fixed to some natural values and some possible elements can be activated upon choice , which we list below .
[ [ limit - to - the - number - of - publications . ] ] limit to the number of publications .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the number of publications is limited to a fixed proportion @xmath36 of submissions .
if the referee process produces too many accepted papers given the current editorial policy of the editor , then all accepted contributions are ranked according to editorial reputation @xmath37 of the authors and only the first @xmath38 papers are published . throughout the paper , for simulations reported , we consider a medium level of competition with @xmath39 . [ [ journal - impact - factor . ] ] journal impact factor .
+ + + + + + + + + + + + + + + + + + + + + + in recognition of the fact that publishing in a reputed journal produces a payoff that depends also on the quality of past published papers , agents who publish a paper may receive an increase to their payoff equal to : @xmath40 at each given time step @xmath1 , the public good benefit of journal impact factor ( jif ) is given to all authors who get their papers published , irrespective of quality .
the higher the proportion of high - quality papers , the higher the jif is .
note that the introduction of jif increases payoffs for publications produced at high effort , but increases free rider rewards for those who are able to publish low - quality work to the same extent .
we assume that the journal impact factor is a public good for those who published with a linear production function , where the increment @xmath41 describes the public value of a single scientific contribution .
the lack of a journal impact factor and a linear public good with @xmath42 describes a situation that is worse than the prisoner s dilemma : even full cooperation does not compensate for the entailed costs ( @xmath43 ) .
considering the journal impact factor and @xmath44 , translates the game into a true linear public good game , which is still extremely difficult to solve . for the sake of the simulations of this paper we assume @xmath45 .
[ [ desk - rejections - and - speeding - up - publication . ] ] desk - rejections and speeding up publication . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + with increased time - pressure and burden of reviewers , it is common practice that not all submissions are sent out for review .
some submissions are desk - rejected and others have an easy , speedy route for publication .
editorial decisions behind are very much based on the reputation of authors .
we implement desk - rejection and desk - acceptance as an additive feature compared to the baseline model .
desk - rejection and acceptance introduces a bias in favour of individuals with higher reputation and in damage of individuals with lower editorial reputation compared to the baseline model .
essentially , we introduce the possibility for the editor to desk - reject or accept submissions proportional to the relative editorial reputation of the author , regardless of recommendations by referees .
please note that as producing low or high - quality reviews is also part of editorial reputations , in this way , reviewers receive some compensation for their low or high reviewing efforts .
the rule added to editorial decisions is as follows : * compute the minimal editorial reputation @xmath46 , the maximal editorial reputation @xmath47 and the median editorial reputation @xmath48 of agents . *
if @xmath49 , then with probability @xmath50 the paper is desk rejected , and with probability @xmath51 , it is sent to the referees . * if @xmath52 , then with probability : @xmath53 the paper is desk accepted , and with probability @xmath54 , it is sent out for review .
we first demonstrate that in the baseline model in which scientists face costs for producing high - quality manuscripts and costs for producing high - quality reviews , there is no chance of any cooperation .
this is not surprising , because low effort in writing papers as well as low effort in writing reviews is the dominant strategy in the baseline game .
figures [ baseline_doubleblind ] , [ baseline_singleblind ] , and [ stats_baseline ] report that this is the case both for double blind and for single blind peer review systems considering a neutral editorial policy ( ap ) . in all cases ,
the strategy implying the production of low quality papers and reviews ( dd ) overtakes the entire population . to examine the failure of the scientific peer review process more closely , consider that high - quality review does not return any benefits in any case , therefore every reviewer is better off by choosing d. in a population with only dd and cd strategies , dd yields higher average payoffs since , if people review randomly , there is a 50% chance of getting a low - quality paper published , which is exactly the same for high - quality submissions . as cd strategies do not benefit anything from peer review , but they entail higher costs for the author , they die out . without any feedback loop that would help to ensure the production of scientific quality , science ends up as an empty exercise . ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=24.0% ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=24.0% ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=24.0% ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=24.0% ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=24.0% ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=24.0% ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=24.0% ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",scaledwidth=25.0% ] [ [ baseline - with - reputation - weighted - consideration - of - reviews . ] ] baseline with reputation - weighted consideration of reviews .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + figure [ baseline_doubleblind ] and figure [ baseline_singleblind ] contain also cases in which the editorial policy takes into account the editorial reputations of reviewers and attach higher weights to the opinion of higher reputed referees .
these are editorial policies mr and er . the strategy dd still gains overwhelming dominance under these policies .
this is because such policies do not ensure the occurrence of correlated equilibria : the larger weight to opinions does not mean at all that these opinions would favor high - quality contributions more than others .
low effort in reviewing is a dominant strategy for every author .
the opinion of those who started with cooperation receive higher weights , but they still underscore defectors with regard to payoffs .
the initial population that is equally divided among different types of strategies goes through a quick and progressive elimination of cooperative strategies . in this process , having a relatively bad reputation does not matter for payoffs as the chances of publishing a paper become equivalent for good and bad papers over time . [ [ baseline - with - the - entry - of - reputational - concerns . ] ] baseline with the entry of reputational concerns .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + somewhat surprisingly , the editorial policy that ensures a low , but stable level of cooperation is 1p .
this is an editorial policy , which accepts all papers , if there is at least one positive recommendation from the reviewers .
seemingly , this policy is neutral to reputation .
but in fact , this is not the case . as there is a constraint on how many papers can be published
, the friendly 1p editorial policy generates the largest surplus of acceptable manuscripts . in case of a surplus
, the editor ranks the papers based on the editorial reputation of the authors .
hence , a direct feedback on reputation exists , which is sufficient to guarantee in some simulation runs a stochastic mixed strategy equilibrium with the survival of cooperation .
as we demonstrated , accounting on reputation of reviewers for making judgment on manuscripts is insufficient to trigger the production of high - quality reviews and high - quality papers .
we should see if a more direct consideration of editorial reputation leads to higher efforts and as a result to better science .
note that the model extension in this direction is related intentionally to popular discussions whether editors should be unbiased or they could rely on reputation signals of authors from the past .
should they catalyze the matthew effect in science , in which the successful get even more success ?
should they contribute to the maintenance of the old - boyism bias ? if they do , does it hurt or help scientific development ?
let us first introduce an editorial bias in favour of authors with high @xmath37 and against authors with low one .
we assume that authors with an editorial reputation higher than the median has a chance of desk acceptance that increases linearly with their reputation ( equation [ deskrejection ] ) .
similarly , we assume that authors with an editorial reputation lower than the median has a chance of desk rejection that is in negative linear association with their reputation ( equation [ deskacceptance ] )
. this modification does not rule out peer review , but concentrates its decisive character to the middle range , where no clear reputational judgment can be expected from the editor .
results with this extension show no major breakthrough for cooperation : dd dominates the outcomes ( figure [ reputationeditorialalone ] ) . either with double blind or single blind peer review , all agents become of dd type .
this indicates that a direct editorial bias in desk acceptance and rejection in itself is insufficient to trigger a large extent of cooperation .
this kind of editorial bias , however , is able to support the survival of conditional cooperation of reviewers ( figure [ stats_reputation ] ) . when a strict editorial acceptance policy is applied ( ap )
, the lack of publishable material leads to the need of selecting submissions based on reputation . the high effort in producing scientific material , however , does not pay off because of the difficulties of acceptance .
scientists therefore follow the easier path of gaining higher reputation and might place high effort in reviewing others .
reviewing efforts are profitable when they are most likely to provide reputation benefits .
the public reputation helps the referees to get the best out of their reputation - based conditional strategy : when the public reputation of an author @xmath55 is high , then it is more likely that his paper gets published , and therefore it is more likely that a review of high quality will ensure positive returns in terms of editorial reputation . as a result ,
the cooperative reviewer strategy that is conditional on the reputation of the author might survive in case of single blind peer review .
this happens because the editor might provide a differential treatment for individuals with higher reputations earned strictly by high - quality reviews . at the opposite ,
when the author s reputation is low , then reviewers with a strategy conditional on author reputation do not bother and follow the cheap strategy of providing random advice . in this case , their payoff is not different from agents who never put high effort in reviewing .
the introduction of journal impact factor implies that a public good bonus is added to the payoff of each agent publishing a paper .
the size of the public good is proportional to the performance of the journal in terms of good papers published .
large public good benefits in the presence of some reputational motives allow for strategies producing high quality papers to survive and disseminate ( figure [ jif_alone ] ) .
furthermore , the analysis of the population evolution shows that when jif is active , most strategies producing low - quality papers disappear from the population .
this means that if a journal publishes high - quality papers , it ensures that submissions are also of high quality .
this is good news given the fact that the public good reward of jif as a payoff supplement does not erase the social dilemma structure of the game .
defection is still the best reply strategy both for authors and for reviewers .
still , cooperation evolves ; thanks to the editorial account of author reputations and to the large initial share of cooperative strategies that survive the early phase of the simulation .
full cooperation is among those who disappear relatively late ( figure [ jif_alone ] ) , which assists the dominance of the high - effort - in - writing and low - effort - in - reviewing strategy . as a consequence ,
the rise of good papers at the end is not accompanied by good reviews ( figure [ stats_jif ] ) .
still , the scientific development is maintained at the best and results in the highest possible jif .
this means that only high quality papers are published .
peer review just adds a random noise for the publication process and it is meaningless anyway because everyone contributes with high - quality submissions . [ [ journal - impact - factor - together - with - strong - reputation - concerns . ] ] journal impact factor together with strong reputation concerns .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + public good rewards that supplement the original payoff structure largely improve the opportunities for scientific development and lead to the overall success of the cd strategy .
we have also seen that under certain editorial policies that take account of reputation , a low level of full cooperation ( cc ) can be sustained also without the public good reward .
when we introduced desk rejections and acceptance based on reputation to the baseline , then some agents gained reputation successfully with a conditional reviewing strategy drep .
it is therefore interesting to observe which strategies are successful if both jif and strong reputation concerns are accounted for .
the results show that the strongest determinant of the evolution is the journal impact factor ( figures [ jifandreputation ] and [ stats_jifandreputation ] ) . when it matters , even under strong reputation concerns , high - effort publication strategies gain dominance with low - effort reviews .
this is very much meaningful once there is a reward for reputation and the most reputational gains can be obtained by high - quality publications . ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",width=188 ] ; papers are assigned in random order to randomly chosen referees ; each referee can review up to 4 papers and a maximum of 30% of the papers are publishable in each period.,title="fig:",width=188 ] ) .,title="fig:",width=188 ] ) .,title="fig:",width=188 ] ) .,title="fig:",width=188 ] ) .,title="fig:",width=188 ] ) .,title="fig:",width=188 ] ) .,title="fig:",width=188 ] [ [ tacit - agreements ] ] tacit agreements + + + + + + + + + + + + + + + + to complete the story , we still miss a mechanism that makes writing as well as reviewing papers plausible and sustainable without radically altering the payoff structure of the game . a realistic possibility is to consider `` tacit agreements '' that work against the impartiality of peer review in practice .
tacit agreements are based on direct reciprocity and might work better in case of single blind reviews , in which at least one side of the informational asymmetry is relaxed .
a `` nice '' tacit agreement strategy could start with high effort in the first @xmath56 rounds , recognizes previous reviewers of own papers with @xmath57 probability and retaliates them not with high effort / low effort , but with acceptance / rejection recommendation .
once there is a coordination device that brings high quality submissions in the hand of highly reputed reviewers , then acceptance recommendations would match true quality . in this case , the editorial selection policy will matter , because this would create the possibility of low cost reciprocation for high performing scientists .
this is in line with the conclusion of similar empirical and simulation work @xcite .
many believe that this is the game that is played by the top researchers in top journals .
it is important to note , however , the possible drawbacks of self - emerged network - based practices , including old - boyism , partiality , and a conservative bias @xcite .
it is puzzling why scientists devote considerable time and effort for writing reviews that decreases their time spent on their own research .
once everyone acts according to self - interest , reviews are all of low quality and they can not be adequately used to judge scientific quality . as a consequence ,
scientists submit low quality work in the hope of passing through randomly judging reviewers .
we have labeled this puzzle as the miracle of peer review and scientific development .
we investigated some potential mechanisms that might resolve this puzzle by agent based modeling .
we have modeled scientific production accordingly : with an incentive structure in which low efforts in writing papers as well as in writing reviews is the dominant strategy of agents .
we applied a replicator dynamics rule to the population of scientists , allowing for the reproduction of strategies that result in higher payoffs .
not surprisingly , low effort in writing papers as well as in writing reviews have spread in the population and scientific practice has become an empty exercise in our baseline model .
next , we assumed that editors might rely on the reputations of authors in their choices . in our model with a single journal , the editor took perfect account of high- and low - quality publications of authors , the number of their rejected papers , and if their reviewer recommendations were in line with the true quality of the paper or not .
we examined different editorial policies that took account of the reputations of scientists .
we showed that if reputations are used in reviewer selection , then it does not save science from low quality submissions and low quality reviews .
a bit more surprisingly , easing the route for publication by desk acceptance for highly reputed authors alone has not changed anything either .
all this indicated that the emergence of cooperation in the form of high efforts is an extremely difficult puzzle .
some cooperation has resulted from a friendly editorial policy that categorized submissions as publishable if at least one reviewer recommended publication .
this policy led to an oversupply of publishable material , which called for the ranking of submissions based on author reputations .
this direct feedback has made the investment in reputations profitable .
consequently , high - effort author strategies survived in a mixed equilibrium together with low - effort author strategies , but nobody invested effort in reviewing . also when a strict editorial acceptance policy was applied , in which only papers with unanimous reviewer support are published , some cooperation has emerged . in this case , however , the lack of publishable material was responsible for the worth of reputation .
as the investment in reputation via writing papers was risky due to the difficulties of publishing , scientists profited more from the investment in reputation via reviews . a strategy that conditioned high reviewing efforts on the author s reputation was able to gain a notable share in the population .
reputations worked to some extent , but public good benefits worked clearly better for scientific development .
once we introduced the journal impact factor as a public good benefit , which meant the distribution of an additional payoff for all authors who published in the journal ( either a good or a bad paper ) , cooperation has become the most successful strategy of authors . in this case , editorial reputations became correlated with actual contributions to the provision of the pubic good .
but as the production of high quality papers was still much more important for reputations than high quality reviews , the cooperative strategy that emerged as successful was investing high effort only in manuscript writing and not in reviewing . as a consequence
, cooperation has been observed in scientific production , but peer review has just added random noise to this development , which raises doubts of its use @xcite and concerns about the use of public money . at the end , we were successful in demonstrating in a simple model the puzzling motivational problem of peer review .
we highlighted that it is not easy to find the way out of this puzzle .
we showed that a high - value of the public good of science maintains scientific development .
reputational systems that are heavily building on author contributions might be partially sucessful , especially if the reputational hierarchy is directly used for selecting between similarly rated submissions . these mechanisms
, however , will not help to sustain the efficiency of peer review .
paradoxically , mechanisms that are able to induce some level of high - quality reviews are building on reciprocity , and in practice they are often associated with impartiality , old - boyism , the emergence of invisible colleges , the matthew effect , the conservative bias , and the stratification of science .
we plan to extend our simple model towards studying multiple journals that compete for success with each other .
this extension allows for the evolution of editorial strategies in a straightforward way and in parallel to theoretical studies that highlight how group selection can ensure higher cooperation @xcite , is expected to lead to better reviewer performance .
barrera , d. and buskens , v. 2009 `` third - party effects on trust in an embedded investment game . '' in cook , k. , snijders , c. , buskens , v. and cheshire ( eds ) , trust and reputation , new york , russell sage , 37 - 72 .
chetty , r. , saez , e. , and sndor , l. `` what policies increase prosocial behavior ?
an experiment with referees at the journal of public economics . ''
journal of economic perspectives , ( 2014 ) 28(3 ) : 169 - 188 .
milinski , manfred , dirk semmann , and h. krambeck .
`` donors to charity gain in both indirect reciprocity and political reputation . '' proceedings of the royal society of london b : biological sciences 269.1494 ( 2002 ) : 881 - 883 .
milinski , manfred , et al .
`` cooperation through indirect reciprocity : image scoring or standing strategy ? '' proceedings of the royal society of london b : biological sciences 268.1484 ( 2001 ) : 2495 - 2501 .
sobkowicz , pawel .
`` innovation suppression and clique evolution in peer - review - based , competitive research funding systems : an agent - based model . ''
journal of artificial societies and social simulation 18.2 ( 2015 ) : 13 .
sommerfeld , ralf d. , hans - juergen krambeck , and manfred milinski .
`` multiple gossip statements and their effect on reputation and trustworthiness . ''
proceedings of the royal society of london b : biological sciences 275.1650 ( 2008 ) : 2529 - 2536 . | it is not easy to rationalize how peer review , as the current grassroots of science , can work based on voluntary contributions of reviewers .
there is no rationale to write impartial and thorough evaluations .
consequently , there is no risk in submitting low - quality work by authors . as a result , scientists face a social dilemma : if everyone acts according to his or her own self - interest , low scientific quality is produced .
still , in practice , reviewers as well as authors invest high effort in reviews and submissions .
we examine how the increased relevance of public good benefits ( journal impact factor ) , the editorial policy of handling incoming reviews , and the acceptance decisions that take into account reputational information can help the evolution of high - quality contributions from authors .
high effort from the side of reviewers is problematic even if authors cooperate : reviewers are still best off by producing low - quality reviews , which does not hinder scientific development , just adds random noise and unnecessary costs to it .
we show with agent - based simulations that tacit agreements between authors that are based on reciprocity might decrease these costs , but does not result in superior scientific quality .
our study underlines why certain self - emerged current practices , such as the increased importance of journal metrics , the reputation - based selection of reviewers , and the reputation bias in acceptance work efficiently for scientific development .
our results find no answers , however , how the system of peer review with impartial and thorough evaluations could be sustainable jointly with rapid scientific development .
_ keywords : peer review ; evolution of cooperation ; reputation ; agent based model _ |
binggeli & jerjen ( 1998 ) conclude that the shape of a dwarf - elliptical galaxy s surface - brightness profile ( as quantified by the curvature parameter @xmath0 from srsic s ( 1968 ) @xmath1 law ) is not a useful distance indicator .
their conclusion is based on their finding that the scatter on the relevant correlations ` can be reduced ... never below @xmath2 mag .
, at least for the virgo cluster . ' should the intrinsic scatter on the relevant correlations be as large as 0.7 mag .
, the profile - shape indicator would indeed be of only limited value , and there would be strong grounds for believing that virgo dwarf ellipticals define a single cluster of galaxies of small depth .
should however , the intrinsic scatter be about 0.5 mag . or lower
, profile shape would be a valuable indicator of distance and there would be strong grounds for believing that the virgo cluster s depth is significant .
note however , that the latter interpretation does not require the existence of a ` prolate structure ' as presumed by binggeli & jerjen .
there are of course alternative models , notably the substructure model we favoured in young & currie ( 1995 ) ( hereunder yc95 ) .
the central issue in this debate is therefore whether binggeli & jerjen have demonstrated that the intrinsic scatter in the scaling relationships is , as they claim , of the order of 0.7@xmath3 mag . , or whether or not they have at least put forward strong circumstantial evidence in support of their case .
as binggeli & jerjen have mis - interpreted certain aspects of the luminosity-@xmath0 ( @xmath4-@xmath0 ) and scalelength-@xmath0 ( @xmath5-@xmath0 ) distance indicators of young & currie ( 1994 ) ( hereunder yc94 ) and yc95 , it would probably be pertinent to re - emphasize the scope of the indicators and how they are related to one another .
the indicators can be interpreted as follows .
dwarf and intermediate elliptical galaxies of the same @xmath0 are approximately the same physical size [ i.e. they have similar @xmath6 . only therefore ,
when such galaxies have very similar stellar populations can they be expected to have similar central and mean surface brightnesses , whence similar luminosities ) and @xmath0 .
it could , however , exhibit a bright - end cut off defined by those objects with particularly luminous stellar populations . ] .
it follows that galaxies of different colours can be expected to have different stellar populations and therefore can not be compared directly using the @xmath4-@xmath0 method .
the converse is not always true however , as objects of the same overall colour may have different stellar populations .
the @xmath4-@xmath0 relationship appears to be most useful for those dwarf ellipticals with colours of @xmath7 , because other dwarfs are generally bluer ( i.e. @xmath8 ) regardless of whether they are higher or lower surface - brightness objects ( yc95 ) .
another important caveat is that if the stellar populations within a galaxy are not well mixed , the surface - brightness profile shape may deviate significantly from that which one might expect on the basis of its size .
the @xmath0 based distance indicators should therefore , ideally , not be applied to galaxies with internal colour gradients .
although the absence of any colour gradient within a galaxy does strictly not imply that its stars are well mixed , in such cases it is probably safe to assume that they are .
this is because a conspiracy of many different factors would be required in order to balance the colour gradients that would otherwise inevitably arise from segregated stellar populations . with the above in mind
, the local group early - type dwarf , ngc 205 , should not be used as an absolute - distance calibrator for @xmath0-based scaling laws if those target galaxies possessing colour gradients can be screened out .
however , it can be used as a calibrator if , as in yc94 and yc95 , colour - gradient information is not available for target galaxies . since the distance indicators were first presented , graham et al .
( 1996 ) and gerbal et al .
( 1997 ) have found that the correlations on which they are based probably apply not only to dwarf and intermediate ellipticals , but also to classical ones , including the brightest cluster ellipticals .
also , binggeli & jerjen have also shown that the correlations probably apply to dwarf lenticulars as well , while de jong ( 1996 ) has demonstrated that even the bulges of spiral galaxies appear to exhibit a continuous range of @xmath0 values .
it should be remembered however , that colour gradients are more common and often much larger in classical ellipticals and lenticulars than in dwarfs .
when dealing with samples of classical early - type galaxies , it is therefore even more important to screen them for objects with colour gradients as such objects can not be expected to conform to the @xmath5-@xmath0 relationship [ or the @xmath4-@xmath0 relationship ] .
binggeli & jerjen presented srsic profile parameters for 128 dwarf elliptical and dwarf lenticular galaxies , which they derived from the surface photometry of binggeli & cameron ( 1991 & 1993 ) .
the distribution of their objects on the sky is shown in fig .
[ coverage ] , from which it is evident that only to the south - west of the cluster core direction ( defined collectively by m84 , m86 and m87 ) is there detailed coverage . turning now to binggeli & jerjen s total magnitudes ,
we were surprised to discover that the @xmath9 values they quoted were in fact those of binggeli & cameron rather than the systemic magnitudes ( obtained by integrating srsic s law to @xmath10 ) that we would have expected . binggeli & cameron s photometric zero points were based on the total - magnitude scales of de vaucouleurs & pence ( 1979 ) ; brngen ( 1980 & 1984 ) and the _ virgo cluster catalog _ , hereunder vcc , of binggeli et al .
ichikawa et al.s ( 1986 ) total - magnitude scale was probably also used for the calibration of two or three plates , but binggeli & cameron were ambiguous on this point .
although , binggeli & cameron were quite modest about the limitations of their photometry , binggeli & jerjen allowed for a photometric error of only 0.2 mag . in their correlation analyses .
there are several reasons why the real photometric errors must be very much larger than this .
these reasons are outlined below . from comparisons with our
independently calibrated _ virgo photometry catalogue _
( young & currie , 1998)-system total magnitudes in the @xmath11 and @xmath12 bands , as well as @xmath13 and @xmath14 colours . the @xmath15 system is described in detail by young et al .
( in press ) , and its application to the vpc is described by young & currie ( 1998 ) .
] , hereunder vpc , it is clear that binggeli & cameron s adopted magnitude - scale standards do not define a single mutually consistent magnitude scale .
this point is evident from fig.s [ bo - vpc ] , [ iwo - vpc ] and [ vcc - vpc ] , in which serious scale discrepancies between the different sources are also noticeable .
we are confident that our vpc magnitude scales are not to blame for these scale discrepancies , as amongst other reasons , our photometry was calibrated with many hundreds of photoelectric aperture - photometry as well as simulated aperture - photometry measurements derived from ccd images .
furthermore , the agreement between vpc magnitude measurements and those of durrell ( 1997 ) , which were based on deep ccd photometry of relatively faint virgo dwarfs , is better than 0.04 mag .
we are also confident that the zero point of the vpc s @xmath12-band total magnitude ( @xmath16 ) scale is accurate to several percent ( note that it is independent of the general transformation we adopted to calibrate our @xmath11 plates with @xmath12 and @xmath17-band photoelectric photometry measurements ) . for detailed comparisons with existing photoelectric photometry for vpc galaxies ,
see sect .
8 of young & currie ( 1998 ) .
another [ albeit related ] reason why binggeli & jerjen must have severely under - estimated their photometric errors is that binggeli & cameron did not calibrate their photometry directly .
instead , they first calibrated their extrapolated total - magnitude scale with existing total - magnitude scales .
the other srsic profile parameters derived by binggeli & jerjen were presumably derived on the basis of these magnitude - scale calibrations . the obvious weakness in this approach
is that even if the different sources of standard objects had been accurately calibrated , there would be systematic differences between them on account of the different extrapolation ( or in the cases of brngen s and the vcc datasets , visual total - magnitude estimation ) procedures .
a further problem is likely to be the scarcity of calibrators in certain fields .
binggeli & cameron s plate 1 was for example only calibrated with one galaxy .
if one compares fig.s [ bc - vpc ] with fig.s [ bo - vpc ] , [ iwo - vpc ] and [ vcc - vpc ] , it is clear that binggeli & cameron s magnitude - scale does not , as a whole , bear much resemblance to any of the scales invoked for calibration purposes .
binggeli & jerjen must therefore have severely under - estimated the errors in both their total magnitude values and their central surface brightness , @xmath18 , measurements .
young ( 1994 & 1997 ) has already presented some preliminary findings on the sizes of and origins of systematic errors in the faint ends of existing magnitude scales for virgo galaxies .
a much more detailed paper on this subject , covering the whole magnitude scale and dealing with the ramifications of the zero - point and scale errors uncovered , will be presented by young et al .
( in preparation ) . in spite of the calibration problems described above , the srsic scalelength , @xmath19 , and the shape parameter , @xmath0 , should be independent of zero point , so we would expect binggeli & jerjen s measurements of these quantities to be useful .
binggeli & jerjen noted ` quite good ' agreement with durrell s @xmath0 values , with a rms ( 1@xmath20 ) scatter of 0.10 .
note that one should not be alarmed by the much larger scatter between these authors log@xmath19 values , because this quantity is a strong function of @xmath0 , assuming that the @xmath5-@xmath0 correlation is genuine ( which even binggeli & jerjen do nt question though they believe that it has a large intrinsic scatter ) . in summary
then , binggeli & jerjen s galaxy sample is not a complete sample of galaxies down to a well - defined total magnitude limit .
also , its coverage of the virgo cluster direction is very patchy .
however , it does contain a large number of dwarf galaxies and is therefore useful on the basis of its size .
unfortunately the photometric zero points adopted for different plates are not mutually consistent , thereby rendering the @xmath9 and @xmath18 values of little use .
however , this should not affect the @xmath0 or @xmath19 values , which are probably more accurate than the yc95 values because they are based on higher - resolution photometry .
binggeli & jerjen investigated the following four correlations : @xmath9 versus @xmath21 , @xmath9 versus @xmath22 , @xmath9 versus @xmath18 , and @xmath23 versus @xmath21 for virgo galaxies .
they observed rms scatters in these correlations of 0.92 , 0.73 , 0.76 and 0.85 mag . respectively , and asserted that : ` a scatter of 0.7 mag .
is what one can already get from the relation between the mean effective surface brightness @xmath24 and total magnitude ' . as is evident from fig .
[ bc - vpc ] there is a significant and not necessarily linear scale error in their magnitude scale for galaxies that lie within the vpc survey area ( corresponding to their plates 17 , 18 and 26 but with two objects on their plate 4 ) .
the sense of this error is such that the luminosities of their fainter galaxies were over - estimated with respect to their brighter objects . in the case of the outlying fields
their scale errors are almost certainly even larger as the only calibrators used were vcc galaxies with total - magnitude values taken from either the vcc or de vaucouleurs & pence ( 1979 ) . as mentioned in sect . 3
, there are very large systematic errors in both of these sources of magnitudes .
in fact the preliminary work of young ( 1994 & 1997 ) finds that these sources over - estimate luminosities by about 0.7 mag . at the faint end . as already demonstrated in sect . 3 ,
binggeli & jerjen s photometry was based on differentially zero - pointed plates ( i.e. objects on each of the 13 different plates received different absolute calibrations ) . furthermore , their @xmath9 values were not systemic ones ( i.e. obtained by integrating srsic s function through 360@xmath25 to @xmath10 ) , but those of binggeli & cameron ( 1993 ) , which were obtained using a different extrapolation procedure and including the nuclear light contribution when present .
the effects of both of these limitations in their reduction procedures would be to increase the observed scatter in the @xmath9 versus @xmath21 , the @xmath9 versus @xmath22 and @xmath9 versus @xmath18 correlations .
of these three correlations , the first would be affected the most .
this is because the @xmath18 term in the other two correlations can to a certain extent compensate for the errors in the @xmath9 values adopted ( even if neither the measured @xmath18 nor the measured @xmath9 bear much resemblance to the actual values ) . also , the @xmath9 versus @xmath21 correlation is the one most susceptible to increased scatter when , as by binggeli & jerjen , applied indiscriminately to objects of different stellar populations in the absence of galaxy - colour information .
we therefore find that binggeli & jerjen s dataset is useful only for investigating the @xmath23 versus @xmath21 correlation , assuming of course that binggeli & cameron s ( 1993 ) background subtraction procedures were adequate .
we are therefore confronted with an observed scatter of 0.85 mag . in a scaling relationship based on a sample of 128 virgo galaxies .
clearly , even if the measurement errors in the parameters @xmath19 and @xmath0 introduced a random component as high as 0.30 mag .
, we are still left with a scatter of 0.80 mag . to explain .
binggeli & jerjen attribute this remaining component mainly to intrinsic scatter , while we would attribute a large part of it to spatial depth .
binggeli & jerjen make a big issue of the mutual dependence between the residuals in magnitude space with respect to their @xmath9 versus @xmath0 correlation and the residuals in angular - distance space with respect to their @xmath19 versus @xmath0 correlation .
they plot these residuals in their fig .
9 . while they are correct in pointing out that there must be some dependence between the two sets of residuals , whether this dependence is significant enough to affect our previous findings
is another matter .
we have re - plotted their fig . 9 here as fig .
[ res - bj ] , this time using equal axis scales .
they claim that in the absence of any dependence between the residuals , fig .
[ res - bj ] should be devoid of any correlation . however , their test for dependence is fatally flawed because it is based on the prior assumption of negligible depth as illustrated by the following example .
imagine that we have five galaxies , which collectively constitute a complete sample of galaxies devoid of any malmquist bias .
the nearest galaxy is at @xmath26 while the farthest is at @xmath27 , and the spatial separation between each object is @xmath28 .
the mean distance modulus of these five galaxies [ in log(distance ) space ] is therefore @xmath29 .
now , let us imagine that we have two perfect distance indicators based on two completely independent scaling laws which we shall denote sl1 and sl2 . both indicators can measure the distances of these objects precisely because both methods are perfect . if we were now to construct a diagram analogous to fig .
[ res - bj ] , we would end up with a plot like fig .
[ res - hypo ] .
the rms scatter in the residuals with respect to sl1 would be identical to that with respect to sl2 , and both of these quantities would be equal to @xmath30 mag .
now , according the binggeli & jerjen , for two such ` independent but equivalent ' measurements , we would expect the scatter with respect to the equality line on fig .
[ res - hypo ] to be 2 mag .
however , because the distance indicators are perfect , the actual scatter with respect to the equality line is _ zero _ [ regardless of which axis it is measured parallel to ] .
the reason for this is that while the two different measurements for an individual galaxy are ` equivalent ' ; the measurements for different objects are not , simply because each object is at a different distance . in spite of the above , we accept that in sect .
4 of yc95 , we did indeed under - estimate our formal internal distance errors , because there must be some [ non - distance related ] dependence between the distance estimates based on the different scaling laws .
however , even if our formal internal errors [ including both intrinsic scatter and photometric errors ] were as high as , say , 0.6 mag .
0.47 mag . as quoted in yc95 ] , that would still leave room for a cluster depth of 0.54 mag .
, as the observed scatter per relationship in yc95 was 0.81 mag .
note that a cluster with a depth of 0.54 mag . would be half as deep as it is distant , with a further one - third of its objects lying even further out from its centroid .
we should also like to emphasize that our ` independent information on the intrinsic scatter ' was not ` seized ... by applying both the @xmath0-@xmath31 and @xmath0-@xmath32 relation at the same time ' .
the independent information was in fact , the much smaller scatter found in our samples of fornax and local group galaxies . as long as our fornax and local - group samples are representative and as long as they contain galaxies structurally similar to their counterparts in virgo , the depth interpretation still holds .
binggeli & jerjen state that : ` if the intrinsic dispersion of the @xmath0-@xmath31 or the @xmath0-log@xmath19 relation is much smaller for fornax dwarfs than for virgo dwarfs as it appears ( which , however , might be caused by the incompleteness of yc s fornax sample ) we are in need of an explanation for this difference ' . in response to their criticisms that we excluded three suspected non - cluster members when investigating the scatter in our @xmath5-@xmath0 relationship
, we have re - measured the scatter in our @xmath5-@xmath0 correlation without excluding any outlier . for a polynomial of the form @xmath33 , which has the advantage over equation 1 in yc95 of being monotonic , the scatter in @xmath5 based on all 26 of the objects listed in table 2 of yc95 is [ in terms of magnitudes ] : 0.55 mag .
allowing for a conventional fornax - cluster depth of 0.15 mag . , but
not making any allowance for possible foreground or background objects , places an upper limit on the intrinsic scatter of 0.53 mag .
this is very much lower than the scatter found in the same relationship for virgo galaxies ( see sect . 4 ) .
if , as binggeli & jerjen maintain , the much larger scatter observed for virgo galaxies were intrinsic , we would therefore indeed be in need of an explanation as to why these fornax dwarf galaxies differ so radically from their counterparts in virgo .
as binggeli & jerjen concede , king ( 1966 ) profiles do not fit virgo dwarf - elliptical galaxy profiles well .
this suggests that tidal truncation is not a significant contributor to the luminosity profile shapes at the radii of interest .
we therefore consider it unlikely that tidal effects could offer the explanation .
furthermore , on the basis of the colour information presented by caldwell & bothun ( 1987 ) and yc95 , it is clear that most of the brighter fornax and virgo dwarf ellipticals have very similar colours , suggesting that they may well have very similar stellar populations and histories .
we therefore remain to be convinced of binggeli & jerjen s suggestion that there is probably a dichotomy between virgo and fornax dwarfs .
this relationship was cited by binggeli & jerjen as being of comparable value to the profile - shape parameter , @xmath0 , as a distance indicator .
we do not deny that it is a reasonably useful relationship .
in fact , it is related to the @xmath4-@xmath0 and @xmath5-@xmath0 relationships , and probably a direct consequence of them .
however , it can be expected to be significantly harder than the @xmath4-@xmath0 and @xmath5-@xmath0 relationships to measure accurately , because it invokes the effective surface - brightness parameter , which is a tertiary parameter ( unlike @xmath0 and @xmath19 which are primary parameters and total magnitude which is a secondary one ) . in order to measure effective surface brightness accurately ,
a model profile must first be fitted , then the profile must be extrapolated to obtain a total - light estimate and then the profile model must be integrated to the half - light radius .
clearly , an extra stage is involved .
we therefore can not accept binggeli & jerjen s assertion that no profile modeling is required in the measurement of either total magnitude or effective surface brightness .
young ( 1997 ) and young et al.(in press ) have already demonstrated that for dwarf galaxies in particular , total magnitude values ( and therefore effective parameters too ) are critically dependent on the profile model adopted .
binggeli & jerjen expected to find a ` well - defined velocity - distance relation ' based on their ` fairly large ' subsample of 43 objects with known velocities , if there were significant depth in the spatial distribution of virgo dwarfs .
they cited the lack of such a relationship based on their @xmath4-@xmath0 and @xmath4-@xmath18 ` pseudo - distance ' estimates , as evidence against the depth interpretation .
the most fundamental problem with their argument is that in order to generate relative distances based on each of two different scaling relationships , they have already assumed negligible depth when they estimate these relationships directly from the residuals with respect to the best - fitting curves to their data .
the crucial point here is that should there be significant depth , the mean distance of their high-@xmath0 objects must be lower than the mean distance of their low-@xmath0 objects , due to malmquist bias .
the relative distance scales they construct for each relationship should therefore not be based on a best fit to data for virgo galaxies , but on a curve defined by a best fit to data from either a sample of objects known to be at similar distances ( e.g. fornax - cluster galaxies ) and/or a sample of objects whose distances are known ( e.g. local - group galaxies ) . in fig .
[ malmquist - n ] the differential frequency of galaxies should increase monotonically with increasing @xmath0 if their galaxy sample were unbiased . as this is not what is observed , we can conclude that their galaxy sample suffers from bias against high @xmath0 objects .
should there be depth in the spatial distribution of their sample galaxies , the mean distance of their lower @xmath0 objects must be greater than the mean distance of their higher @xmath0 objects .
their residuals can , at best , therefore only yield meaningful relative distances for galaxies within very small ranges in @xmath0 for which the degree of the malmquist bias can be assumed to be constant .
furthermore , as already demonstrated in sect . 3 , the ` pseudo - distances ' derived by binggeli & jerjen for their fig.s 10 and 11 , must indeed be highly inaccurate on account of the errors in their photometry . also , they apply the @xmath9 versus @xmath0 and @xmath9 versus @xmath18 relationships to all early - type dwarfs indiscriminately , in the absence of e.g. colour information .
the existence of so many outliers ( two of which are objects with negative radial velocities ) on binggeli & jerjen s fig.s 10 & 11 is therefore not surprising .
however , as previously suggested in yc95 , there may well be significant line - of - sight substructure in the spatial distribution of virgo galaxies , complicating the kinematics of the galaxy populations present .
significant spatial depth therefore need not necessarily imply the ` quiet ' velocity - distance relationship presumed by binggeli & jerjen .
we have been interested in the kinematics of the dwarf - galaxy populations in virgo for a number of years now , and a major programme to measure large numbers of redshifts for early - type dwarf - candidates is already well underway .
the virgo galaxies targeted by drinkwater et al .
( 1996 ) with the multi - fibre spectrograph on the united kingdom schmidt telescope ( ukst ) , were , on account of the observing constraints at the time , generally of high surface brightness .
consequently , most of them were found to be in the background .
however , 8 objects were confirmed to be dwarf or intermediate early - type galaxies . in 1997 ,
early - type objects of low surface - brightness were targeted by drinkwater et al .
( in preparation ) , again using the ukst , and 67 velocity measurements were obtained .
further measurements made in 1997 with a different telescope and and future ones ( from two separate telescope - time allocations in 1998 ) will be presented in subsequent papers . a detailed investigation into the virgo - dwarf velocity field
will then be based on these new data .
even without any new datum , we still have one potentially decisive test that might be able to help us decide whether the large scatter observed in binggeli & jerjen s @xmath23@xmath34 relationship is due primarily to intrinsic scatter or depth effects .
this test involves looking for departures from uni - modality and/or normality in the differential frequency distribution of scale - length residuals with respect to the best - fit curve for the data ( not in this case a curve defined by galaxies from an external galaxy sample ) .
should the distribution not be consistent with a uni - modal gaussian [ measured in log(distance ) space to be rigorous ] , we can say that the scatter is not consistent with the intrinsic scatter origin hypothesis . in fig .
[ bf - rn - res ] , we find some evidence for a tri or quad - modal distribution .
such a distribution could arise if early - type dwarfs exist in three or four discrete size ranges within the same cluster .
however , not only does this seem most unlikely , but it would also be very hard to reconcile such a scenario with the theoretical work on the subject ( hjorth & masden 1995 , gerbal et al .
1997 , prugniel & simien 1997 & ciotti & lanzoni 1997 ) .
it is perhaps more likely that any multi - modality present is due to line - of - sight substructure in the galaxy sample s spatial distribution , albeit smoothed considerably due to the relative distance scales being different for different ranges in @xmath0 .
we have performed several non - parametric statistical tests as described by lucey et al .
( 1986 ) and references therein . comparing the fig .
[ bf - rn - res ] data against 100,000 monte - carlo realisations drawn from a gaussian distribution having the standard deviation of the residuals , yielded probabilities that the fig .
[ bf - rn - res ] distribution could arise by chance .
for the kolmogorov - smirnov ( lillefors ) test and geary s a - test ( mean absolute deviation / standard deviation ) , the probabilities were 0.255 and 0.222 respectively ; which were suggestive but not by any means conclusive .
however , for the skewness and kurtosis tests and the u - test ( data range / standard deviation ) the probabilities were 0.041 , 0.011 and 0.005 respectively .
we take these results as significant evidence against binggeli & jerjen s intrinsic scatter interpretation .
the findings presented in yc95 were based on the assumption that the relatively small scatters observed in two scaling relationships found for fornax and local group dwarfs , are also applicable to virgo dwarfs .
although , we still consider this assumption to be a most reasonable one , we concede that we should have stated it explicitly in yc95 . unlike binggeli & jerjen
, we ( yc95 & this work ) did not make any prior assumption as to the actual depth of virgo when making our case . also , contrary to the impression given by binggeli & jerjen , our case is not undermined by any dependence between the scaling laws which merely means that we slightly under - estimated our internal errors in yc95 . in this paper , we have also presented further statistical evidence to support our case based on binggeli & jerjen s own dataset .
binggeli & jerjen claim to have presented evidence both that ( 1 ) profile shape is not a useful distance indicator because scaling laws based on it have large intrinsic scatters and that ( 2 ) the spatial depth of the virgo cluster must be small .
however , all that they have in fact achieved in their 9-page paper , is to point out that if one first assumes that the depth of the cluster is small enough , then there must be large intrinsic scatters in the scaling laws in question for virgo galaxies . in order to explain away the smaller intrinsic scatters found in fornax and local group samples of galaxies , they suggest [ in the absence of any supporting evidence ] that either there is a dichotomy between virgo and non - virgo dwarfs or that our samples of non - virgo dwarfs are not representative .
they therefore can not legitimately claim to have presented evidence for both ( 1 ) and ( 2 ) simultaneously ; or to have presented evidence for ( 1 ) or ( 2 ) without first assuming ( 2 ) or ( 1 ) respectively .
the virgo cluster is not a suitable target for investigating the reliability of the curvature - based indicators .
reliable photometry of galaxies in other clusters and groups , where there is no controversy concerning line - of - sight depth , are required for this purpose .
cky gratefully acknowledges a pdrf from the national postdoctoral fellowship office of china and use of the computing facilities of the qso & observational cosmology group at beijing astronomical observatory . 99 binggeli b. , jerjen h. , 1998 , a&a , in press binggeli b. , cameron l.m . , 1991 , a&a , 252 , 27 binggeli b. , cameron l.m . , 1993 , a&as , 98 , 297 binggeli b. , sandage a. , tammann g.a . , 1985 , aj , 90 , 1681 ( vcc ) brngen f. , 1980 , astr .
, 301 , 305 brngen f. , 1983 , astr .
nachr . , 304 , 41 caldwell n. , bothun g.d . , 1987 ,
aj , 94 , 1126 ciotti l. , lanzoni b. , 1997 , a&a , 321 , 724 de jong r.s .
, 1996 , a&a , 313 , 45 de vaucouleurs g. , pence w.d . , 1979 , apjs , 40 , 425 drinkwater m.j . ,
currie m.j .
, young c.k .
, hardy e. , yearsley j.m .
, 1996 , mnras , 279 , 595 durrell p.r . , 1997 , aj , 113 , 531 gerbal d. , lima neto g.b . , marquez i. , verhagen h. , 1997 , mnras , 285 , l41 graham a. , lauer t.r . , colless m. , postman m. , 1996 , apj , 465 , 534 hjorth j. , madsen j. , 1995 , apj , 445 , 55 ichikawa s .- i . , wakamatsu k .-
, okamura s. , 1986 , apjs , 60 , 475 king i.r .
, 1966 , aj , 71 , 64 lucey j.r . ,
currie m.j .
, dickens r.j .
, 1986 , mnras , 221 , 453 prugniel p. , simien f. , 1997 , a&a , 321 , 111 srsic j.l . , 1968 ,
atlas de galaxias australes , observatorio astronomica , cordoba young c.k . , 1994 , d.phil
. thesis , university of oxford young c.k .
, 1997 , galaxies . in : k.s . cheng & k.l .
chan ( eds . ) proceedings of 21st century chinese astronomy conference , world scientific , singapore young c.k .
, currie m.j . , 1994 ,
mnras , 268 , l11 ( yc94 ) young c.k . , currie m.j .
, 1995 , mnras , 273 , 1141 ( yc95 ) young c.k . ,
currie m.j . , 1998 ,
a&as , 127 , 367 ( vpc ) young c.k . , metcalfe n. , zhu j. , wu h. , chen j.s . , a&as , in press | in a recent paper , binggeli & jerjen ( 1998 ) question the value of the extragalactic distance indicators presented by young & currie ( 1994 & 1995 ) and state that they have refuted ` the claim that the virgo des [ dwarf - elliptical galaxies] ... are distributed in a prolate structure stretching from 8 to 20 mpc distance ( young & currie 1995 ) . ' even though no such claim was ever made . in this paper , we examine binggeli & jerjen s claims that intrinsic scatter rather than spatial depth must be the main cause of the large scatters observed in the relevant scaling relationships for virgo galaxies .
we investigate the accuracy of binggeli & jerjen s photometric parameters and find that while their profile curvature and scalelength measurements are probably useful , their total magnitude and central surface - brightness measurements are not useful for the purpose of investigating scaling laws because they suffer from serious systematic and random errors .
we also investigate binggeli & jerjen s criticisms of our ( 1995 ) analysis .
we demonstrate that their test for strong mutual dependence between distance estimates based on the two different scaling laws is invalid because of its prior assumption of negligible cluster depth .
we further demonstrate that the [ relative ] distance estimates on which their kinematical arguments are based can not be meaningful , not only because of the seriousness of the photometric errors , but also because they are undermined by the prior assumption that depth effects can again be neglected .
interestingly , we also find that binggeli & jerjen s own dataset does itself contain evidence for large depth . using the observed correlation between scale - length and profile - curvature , ( the only correlation that can be investigated meaningfully using their dataset ) , we find that the frequency distribution of residuals with respect to the best fitting curve deviates significantly from that expected from a uni - modal gaussian distribution . clearly ,
if as binggeli & jerjen claim , the very large scatter observed in this scaling relationship for virgo galaxies ( which is not observed for fornax or local group ones ) were intrinsic , one would expect a uni - modal gaussian distribution . |
in this paper , we study homeomorphisms and diffeomorphisms of one - manifolds , i.e. , @xmath9 $ ] , or @xmath10 . for convenience , we will always assume they are orientation - preserving .
from a dynamical standpoint , a homeomorphism of the line is not very interesting .
it will have some closed ( possibly empty ) set of fixed points , on the complement of which points wander to the left or to the right .
the dynamics of a homeomorphism of the circle _ is _ interesting ; this theory is due to poincar .
he defined a quantity called the _ rotation number _ @xmath11 of a homeomorphism @xmath12 , and showed that this quantity contains valuable dynamical information about @xmath12 .
namely , if @xmath13 is zero ( rational ) , @xmath12 has fixed ( periodic ) points . on the other hand , if @xmath13 is irrational , @xmath12 is topologically semi - conjugate to a rotation of the circle by the angle @xmath13 . in this situation
there are two sub - possibilities : ( 1 ) @xmath12 is in fact conjugate to a rotation , so its dynamics is minimal ( every orbit is dense ) , or ( 2 ) there is a cantor set on which @xmath12 is minimal , the complement of which consists of `` wandering intervals '' that never return to themselves .
see @xcite for more information .
a homeomorphism of a manifold @xmath0 can be viewed as giving an action of @xmath14 on @xmath0 , via @xmath15 . in this light , it is natural to consider actions of other groups , such as abelian groups , or more generally nilpotent groups .
formally , by an _ action _ of @xmath1 on @xmath0 we mean a homomorphism @xmath16 , the group of orientation - preserving homeomorphisms of @xmath0 .
( we deal only with discrete groups in this paper , so continuity of @xmath17 is not an issue . )
we say the action is _ faithful _ if @xmath17 is one - to - one . for example , one can get faithful @xmath18 actions on @xmath10 ( resp .
@xmath19 ) by taking @xmath20 rationally independent rotations ( resp .
translations ) .
notice that these actions will be minimal for any @xmath21 on @xmath10 , and for any @xmath22 on @xmath19 .
there is a close interplay between group actions on the circle and on the line .
first , given an abelian action on the circle with a _ global fixed point _ ( point fixed by the whole group ) , one can remove the fixed point , and what remains is an action on the line .
conversely , given an action on @xmath19 we can add a point at infinity to get an action on @xmath10 . more interestingly ,
suppose we are given @xmath20 commuting homeomorphisms @xmath23 of the line .
it turns out that if there are no global fixed points for the action of @xmath23 , then some @xmath24 must have no fixed points .
we can quotient @xmath19 by the action of @xmath24 , yielding a circle .
since the other homeomorphisms commute with @xmath24 , they induce homeomorphisms @xmath25 of the quotient circle @xmath26 .
thus , an action of @xmath18 on @xmath19 with no global fixed points yields an action of @xmath27 on @xmath10 .
the converse is also true : given an action of @xmath27 on @xmath28 , we can lift it to a group of @xmath29 homeomorphisms of @xmath19 commuting with translation by one , giving a @xmath18 action on @xmath19 .
it is a folklore theorem that any finitely - generated torsion - free nilpotent group admits faithful actions by homeomorphisms of the line ( and hence also the circle ) .
this is equivalent to the fact that such a group admits left - invariant total orders @xcite .
explicitly , following farb and franks we can construct actions of these groups as follows .
malcev s work @xcite ( see also @xcite ) shows that every finitely - generated torsion - free nilpotent group @xmath1 embeds in some @xmath30 .
thus , to define an action of @xmath1 by homeomorphisms of the line , we need only define an action of @xmath31 .
we take lower - triangular matrices because they are convenient in the following way : the standard action of @xmath31 on @xmath18 ( on the left ) preserves the lexicographic order on @xmath18 .
it is possible to choose a collection of disjoint open subintervals of @xmath19 with dense union in @xmath19 , indexed by @xmath18 and with the property that @xmath32 is to the left of @xmath33 if and only if @xmath34 in the lexicographic order .
we can define an action of @xmath31 on @xmath35 by letting @xmath36
send @xmath37 to @xmath38 , say in an affine way .
since we chose the @xmath37 to have dense union in @xmath19 , these maps extend uniquely to homeomorphisms of the line . in the examples of nilpotent actions we have seen so far ,
there have been two basic types of behavior .
one is minimality , as in at least one irrational rotation of the circle or at least two independent translations of the line .
the other is discrete intervals with dense union being permuted in some way .
this was the case for the general construction of nilpotent actions we gave , and the reader can check it also holds for any single homeomorphism of a one - manifold not conjugate to an irrational rotation of the circle .
our main result says that these two behaviors , in some combination , are all we can see for nilpotent actions on one - manifolds .
theoremstructuretheorem [ structuretheorem ]
let @xmath0 be a one - manifold , and @xmath39 be a finitely - generated virtually nilpotent group .
there exist ( countably many ) open sets @xmath40 and @xmath41 such that the following hold : * @xmath42 unless @xmath43 , and @xmath44 for all @xmath45 . *
each @xmath40 and @xmath41 is @xmath1-invariant .
* let @xmath3 be the open intervals comprising @xmath40 .
for any @xmath46 , there exists @xmath47 such that @xmath48 . if @xmath49 , then @xmath50 . * for any @xmath51 , the action of @xmath1 on @xmath41 is minimal . if @xmath52 and the action of @xmath1 on @xmath10 is minimal , @xmath1 is abelian .
if some @xmath41 is composed of open intervals @xmath4 then the group @xmath53 is abelian .
* @xmath54 is dense . in figure 1 , we show an example of what this could look like for an action of the heisenberg group @xmath55 = h \text { and } [ f , h ] = [ g , h ] = id\rangle$ ] on @xmath19 .
the action of @xmath12 is by translation , as shown .
the elements @xmath56 and @xmath57 act minimally on the wavy intervals , while in the other intervals , @xmath57 acts trivially and a @xmath56-orbit is depicted . for this action , there is only one @xmath58 and only one @xmath59 , as shown .
note that the action of @xmath60 on @xmath59 is _ not _ abelian , but its restriction to any of the intervals in @xmath59 is .
there two key nontrivial assertions in our theorem , which depend on fact that finitely - generated virtually nilpotent groups have polynomial _ growth _ ( see below for the definition ) .
the first is that , on a @xmath1-invariant interval on which the action is minimal , the action is abelian .
this assertion holds for any finitely - generated group of sub - exponential growth .
however , it is not true for solvable groups . for instance , consider the group of homeomorphisms of @xmath19 generated by @xmath61 and @xmath62 .
this action is minimal : the orbit of @xmath63 contains every dyadic rational . but clearly , the group is not abelian .
the second nontrivial assertion is that , in regions where the action is not minimal , there are no dense orbits .
this fails in general for groups not having polynomial growth .
grigorchuk - maki gave an example of a group @xmath1 of intermediate ( super - polynomial , sub - exponential ) growth acting faithfully by homeomorphisms of the line .
this group action has dense orbits , but also has ( a dense set of ) orbits that are not dense .
we show three of these orbits in figure 2 .
all the dots of a given size lie in the same orbit ; for clarity , we have shown only part of the densest orbit . this second assertion is proved for finitely - generated virtually nilpotent groups by an inductive procedure , looking at finer and finer intervals .
the procedure must end after finitely many steps , since these groups have polynomial growth , and each time we restrict to a finer interval the growth degree becomes smaller .
the inductive procedure yielding the desired intervals is a conjugacy invariant : if we have two topologically conjugate actions of a finitely - generated virtually nilpotent group @xmath1 , then the structure of the intervals we get will agree .
we have the following immediate corollary : let @xmath0 be a 1-manifold , and @xmath1 a
finitely - generated virtually nilpotent group .
two actions @xmath64 are conjugate if and only if there is a homeomorphism of @xmath0 sending @xmath65 and @xmath66 to @xmath67 and @xmath68 and intertwining the @xmath17 and @xmath69 actions of @xmath1 on these intervals .
in particular , the action of @xmath1 on @xmath66 by @xmath17 and the action of @xmath1 on @xmath68 by @xmath69 are conjugate groups of translations whose elements must have the same translation numbers relative to each other .
let @xmath0 be a manifold , and @xmath71 a homeomorphism .
it is natural to ask , what is the smoothest diffeomorphism of @xmath0 topologically conjugate to @xmath12 ? on surfaces , it is easy to construct homeomorphisms not conjugate to @xmath5 diffeomorphisms .
for example , let @xmath12 have a fixed point @xmath72 , and rotate points faster and faster as one approaches @xmath72 , with the amount of rotation tending to infinity .
no matter what conjugacy one applies , there will still be a `` bad '' point with infinitely fast rotation , so the result will not be a @xmath5 diffeomorphism .
in fact , harrison @xcite has shown the following : on any @xmath20-manifold ( @xmath73 or @xmath74 ) , there exist
@xmath75 diffeomorphisms not conjugate to @xmath76 diffeomorphisms for any real numbers @xmath77 ( if these are not integers , the fractional parts are interpreted as hlder exponents ) .
she shows moreover that if @xmath0 is compact , such _
unsmoothable _ diffeomorphisms are dense in @xmath78 in the @xmath70 topology . in dimension one
the possibilities for dynamics , and correspondingly for unsmoothability , are much more limited .
the line @xmath19 is particularly simple : as we have seen , a homeomorphism @xmath79 is characterized by its set of fixed points and the direction points in @xmath80 move .
one can easily check , using this , that any homeomorphism of @xmath19 is topologically conjugate to a @xmath81 diffeomorphism ; one simply needs to use gentle enough `` bump functions '' in the intervals of @xmath80 .
the circle @xmath10 is not quite so simple , since there can be nontrivial recurrence . from the point of view of smoothability
, there are two possibilities .
the first is that the homeomorphism @xmath82 either has rational rotation number or is conjugate to an irrational rotation . in this case
, @xmath12 is conjugate to a @xmath81 diffeomorphism of @xmath10 .
the second is that @xmath12 is semi - conjugate , but not conjugate , to an irrational rotation ( i.e. , it has an exceptional minimal set ) . in this case , by the following theorem of denjoy , @xmath12 can not be topologically conjugated to @xmath83 : if @xmath82 is a @xmath83 diffeomorphism with irrational rotation number @xmath84 , then @xmath12 is topologically conjugate to the rotation @xmath85 . in this paper
, we show that a homeomorphism of the second type is still always conjugate to a @xmath6 diffeomorphism for any @xmath86 .
thus there is a threshold of regularity for circle diffeomorphisms ; one class of diffeomorphisms can be smoothed only up to this threshold , while the other can be smoothed past it .
more generally , we prove the following , with the help of theorem [ structuretheorem ] : [ smoothability ] let @xmath0 be a 1-manifold , and let @xmath39 be a finitely - generated , virtually nilpotent subgroup .
then there exists a homeomorphism @xmath87 such that @xmath88 for any @xmath89 , where @xmath90 is the degree of polynomial growth of @xmath1 .
this result strengthens the following theorem of farb and franks : let @xmath0 be a 1-manifold . every finitely - generated , torsion - free nilpotent group is isomorphic to a subgroup of @xmath91 .
farb and franks s theorem says that for finitely - generated torsion - free nilpotent groups there _ exist _ faithful actions ; our theorem says that _ every _ action can be conjugated to @xmath5 ( and a little more ) .
the following comes easily from our theorem : [ corollary ] let @xmath92 be finitely - generated and virtually nilpotent . given any fixed generating set for @xmath1 , it is possible to conjugate @xmath1 to a subgroup of @xmath93 such that the generators are as @xmath5 close to rotations as desired .
the same is true for @xmath94 , with `` rotations '' replaced by `` the identity . ''
this gives another method for proving theorem b from @xcite . in this section
, we explore the extent to which our theorem is sharp .
we first observe that we can not in general conjugate the group action to @xmath95 , even if @xmath96 .
to see this , take a homeomorphism @xmath12 of the circle with exceptional minimal set .
lift @xmath12 to a homeomorphism @xmath97 commuting with translation by 1 , @xmath98 .
the group generated by @xmath99 and @xmath98 can not be conjugated to @xmath95 .
next , we note that our theorem can not be strengthened to include the class of solvable groups .
it is shown in and that there exist actions of the baumslag - solitar group @xmath100 on @xmath10 which can not be conjugated to @xmath5 .
this example makes use of the fact that @xmath101 is exponentially distorted ( see below ) in @xmath102 .
bonatti , monteverde , navas , and rivas @xcite give a large class of solvable examples ( including this one ) that can not be conjugated to @xmath5 . by gromov
s theorem @xcite , our results apply to any group of polynomial growth
. it would be interesting to understand actions of groups with sub - exponential growth .
in particular , it seems natural to ask : is the grigorchuk - maki group action on the line conjugate to a @xmath5 action ?
navas @xcite has shown that it is _ semi - conjugate _ to a @xmath5 action : by adding `` buffer '' regions on which the action is trivial , he constructs a faithful @xmath5 action of this group .
another necessary assumption in our results is finite generation of the group .
it is easy to construct uncountable abelian groups of homeomorphisms of @xmath19 which are not conjugate to lipschitz , let alone @xmath5 . on the other hand
, a remarkable result of deroin , kleptsyn , and navas @xcite says that any countable group of homeomorphisms in dimension one is conjugate to a group of bi - lipschitz homeomorphisms .
this leads us to the following question : if @xmath0 is a 1-manifold , and @xmath39 is a countable nilpotent ( or even abelian ) subgroup , is @xmath1 conjugate to a subgroup of @xmath91 ?
a final line of questions involves the optimal regularity for finitely - generated nilpotent actions .
we show in this paper that any finitely - generated virtually nilpotent subgroup @xmath39 is conjugate to a subgroup of @xmath103 for any @xmath104 less than the degree of polynomial growth of @xmath1 .
this leaves open the following questions ( the first one due to navas ) : for a finitely - generated virtually nilpotent group @xmath1 , what is the smoothest faithful action of @xmath1 on [ 0 , 1 ] ? what is the least smoothable action ?
let @xmath105 ( @xmath106 ) be the group of @xmath107 upper - triangular matrices with 1s on the diagonal .
farb and franks exhibited a @xmath5 action of this group on @xmath108 $ ] ; castro , jorquera , and navas @xcite showed that this action can be smoothed to any differentiability class less than @xmath109 ( and subsequently navas @xcite showed @xmath109 is not possible ) .
note that the group @xmath105 has growth degree @xmath110 , so @xmath111 .
thus , for these groups one can always achieve greater smoothness than our results suggest .
in fact , jorquera @xcite showed that there exist groups of arbitrarily high nilpotence degree that embed in @xmath112)$ ] for any @xmath86 .
* acknowledgments .
* i would like to thank andrs navas for referring me to several important articles , including the one of cantwell - conlon and the one of guelman - liousse , for pointing out corollary [ corollary ] and its relation to his article @xcite , and for his enthusiasm for this research .
i would like to thank tobias hartnick and uri bader for helpful conversations .
i am grateful for the support of uri shapira and a fellowship at the technion from the lady davis foundation .
the next two results are involved in showing that , for nilpotent group actions , minimal implies abelian .
the following proposition is easy .
see e.g. @xcite .
if @xmath92 is any amenable subgroup ( not necessarily finitely - generated ) , there a @xmath1-invariant borel probability measure @xmath113 . therefore , rotation number is a homomorphism on @xmath1 , so each element of @xmath114 $ ] has a fixed point , and for any @xmath115 , @xmath116 } \operatorname{fix}(g)$ ] in particular , @xmath114 $ ] has a global fixed point .
radon measure _ on @xmath19 to be a borel measure which is finite on compact subsets ( note that the other usual requirements for a radon measure are automatically satisfied on @xmath19 ) .
the following result is due to plante @xcite ; see also @xcite , theorem 2.2.39 .
[ radonmeasure ] if @xmath1 is a finitely - generated virtually nilpotent subgroup of @xmath117 , there is a @xmath1-invariant radon measure on the line .
therefore , translation number is a homomorphism on @xmath1 , so each element of @xmath114 $ ] has a fixed point , and for any @xmath115 , @xmath116 } \operatorname{fix}(g)$ ] in particular , @xmath114 $ ] has a global fixed point .
it is an easy fact that for any @xmath0 and @xmath118 , @xmath119)$ ] is a @xmath1-invariant set .
we can see this as follows .
let @xmath120 $ ] and @xmath47 ; we want to show that for any @xmath121 $ ] , @xmath122 . but @xmath114 $ ] is a normal subgroup of @xmath1 , so @xmath123 $ ] ; thus , @xmath124 , so @xmath125 , as desired . before proving theorem [ structuretheorem ] , we need the following notions .
let @xmath126 be monotone increasing
. we will write @xmath127 if there exists a constant @xmath128 such that @xmath129 for all @xmath130
. we will write @xmath131 if @xmath127 and @xmath132 .
this defines an equivalence relation .
if @xmath133 $ ] and @xmath134 $ ] are equivalence classes , we may write @xmath133 \leq [ g]$ ] or @xmath133 = [ g]$ ] if @xmath127 or @xmath131 , respectively . by abuse of notation ,
@xmath133 = g$ ] will mean @xmath133 = [ g]$ ] .
we define the _ growth function _ of @xmath1 with respect to the generating set @xmath138 to be @xmath139 though this function depends on the choice of generating set , if @xmath140 is any other choice we will have @xmath141 , so we can talk unambiguously about the equivalence class of these , which we will denote @xmath142 .
it is immediate that if @xmath60 is a quotient of @xmath1 , @xmath143 .
it is also easy to see that if @xmath144 is a finitely - generated subgroup then @xmath143 .
we say that @xmath1 has _ polynomial growth _ if for some integer @xmath8 , @xmath145 .
let @xmath146 be a finitely - generated nilpotent group , and @xmath147 the @xmath148 subgroup in the lower central series for @xmath146 ( with @xmath149 ) .
let @xmath150 ( the torsion - free rank )
. then @xmath151 .
the quantity @xmath152 is also referred to as the _ homogeneous dimension _ of @xmath146 ; see e.g. @xcite . if @xmath1 is a finitely - generated virtually nilpotent group , and @xmath146 is a ( necessarily finitely - generated ) nilpotent subgroup of finite index , then @xmath153 .
it is a fact , not hard and proved in bass @xcite , is that if @xmath1 is a finitely - generated nilpotent group and @xmath144 a subgroup of infinite index , the @xmath154 .
of course , the same holds if @xmath1 is virtually nilpotent and @xmath144 infinite index . indeed ,
if @xmath155 is a finite - index subgroup , then @xmath156 has finite index in @xmath60 , so @xmath157 ; also , @xmath156 has infinite index in @xmath146 , so by bass , @xmath158 let us write @xmath159 for the set of @xmath40 , and @xmath160 for the set of @xmath41 .
we have a bit of freedom in constructing @xmath159 , but no control over @xmath160 . indeed , if @xmath41 and @xmath161 are minimal open sets for the action of @xmath146 , we claim they are equal or disjoint
. suppose neither of these holds ; then we can find @xmath162 , and without loss of generality @xmath163 .
a small open neighborhood of @xmath72 must be contained in @xmath164 , since @xmath164 is open .
since the orbit of @xmath165 is dense in @xmath41 and @xmath166 , the orbit of @xmath165 must enter this small neighborhood of @xmath72 , and hence enter @xmath164 .
but since @xmath161 is a @xmath1-invariant set and @xmath167 , the orbit of @xmath165 can not enter @xmath161 , a contradiction .
first suppose that @xmath52 , and that the action of @xmath1 is minimal .
we claim that @xmath1 must be abelian . by proposition 2.1 , @xmath119)$ ]
is nonempty .
since this set is @xmath1-invariant , by minimality it is dense in @xmath10 ; therefore , @xmath119 ) = s^1 $ ] .
thus @xmath1 is abelian .
if the action of @xmath1 on @xmath10 is not minimal , let @xmath168 be a point whose @xmath1-orbit is not dense . since @xmath169 is a @xmath1-invariant set , we may restrict attention to the open intervals in @xmath170 .
if , given an arbitrary maximal open interval @xmath171 , we can find the desired sets @xmath40 and @xmath41 for the group @xmath172 , then we can certainly find these sets for the @xmath1-action on @xmath10 .
we may therefore assume without loss of generality that the manifold on which @xmath1 acts is an interval , say @xmath96 .
suppose that @xmath41 is some minimal open @xmath1-set , and @xmath173 is a maximal open interval .
again , let @xmath172 .
as above , by applying theorem [ radonmeasure ] , we can see that @xmath174 is abelian .
if @xmath175 is dense , then the theorem is proved .
therefore , we assume that @xmath176 has nonempty interior .
let @xmath177 be an arbitrary maximal open interval .
let @xmath174 be as before . if we can find a set of intervals @xmath40 for the action of @xmath174 on @xmath178 having the requisite properties ( including dense union in @xmath178 ) , then we automatically get the desired intervals for the action of @xmath1 on the @xmath1-orbit of @xmath178
. therefore , it suffices to restrict attention to @xmath178 ; we may assume @xmath1 acts on @xmath19 with no open minimal sets .
we will show , by induction on @xmath8 , that for a group @xmath1 of homeomorphisms of @xmath19 with no open minimal subsets we can find the desired permuted intervals . if @xmath179 , then since @xmath180
, @xmath1 is torsion - free , and therefore trivial .
thus the result is trivial in this case ; we let @xmath159 simply contain @xmath19 .
if @xmath181 , it is not hard to see that @xmath182 .
indeed , @xmath1 contains a nilpotent subgroup of finite index @xmath146 such that @xmath183 , and a torsion - free nilpotent group with @xmath183 must be isomorphic to @xmath14 ( any larger nilpotent group would contain a copy of @xmath184 , and @xmath185 ) . thus @xmath1 is a finite torsion - free extension of a group isomorphic to @xmath14 , so @xmath1 itself is isomorphic to @xmath14 .
therefore , @xmath1 is generated by a single element @xmath56 .
if @xmath186 has nonempty interior , we may add the maximal open intervals in @xmath186 to @xmath159 .
thus we may remove the fixed points of @xmath56 , and show that we get the desired interval structure on the remaining open intervals @xmath178 .
let @xmath72 be an arbitrary element of such an @xmath178 , and let @xmath187 be the interval @xmath188 .
this gives us the desired interval structure .
now let @xmath8 be arbitrary , and assume the result holds for finitely - generated virtually nilpotent subgroups of @xmath117 with no open minimal regions having growth degree smaller than that of @xmath1 . we may remove any global fixed points of the action of @xmath1 on @xmath19 , putting any open intervals contained therein into @xmath159 .
thus we may assume that @xmath1 has no global fixed points .
obviously , @xmath189 ) ) \cup \operatorname{int}(\operatorname{fix}([g , g ] ) ) \subset { \mathbb{r}}$ ] is dense , and these two sets are @xmath1-invariant .
therefore , it suffices to find the desired intervals in @xmath190)$ ] and in @xmath191))$ ] . to find the necessary intervals for @xmath190)$ ] , it suffices to do this for an arbitrary maximal open interval @xmath192)$ ] .
as before , form @xmath174 by taking the subgroup of @xmath1 sending @xmath178 to itself , followed by the quotient of this subgroup that we get by restriction to @xmath178 . we have taken a quotient of a subgroup of infinite index , since if @xmath193 then @xmath194 for all @xmath21 . therefore , as noted above , we have that @xmath195 .
so by inductive hypothesis we can find the desired intervals for the action of @xmath174 on @xmath178 .
similarly , if @xmath178 is a maximal open interval in @xmath191))$ ] then the group @xmath174 satisfies @xmath195 , completing the argument by induction , _ unless _
@xmath1 is abelian ( so @xmath178 is all of @xmath19 ) . in that case
we have @xmath196 for some @xmath197 .
we have assumed that @xmath1 has no global fixed points , so some @xmath198 is conjugate to a translation and we can consider @xmath199 to be a subgroup of @xmath200 isomorphic to @xmath201 .
@xmath202 does not act minimally on @xmath10 , or else @xmath1 would have acted minimally on @xmath19 .
therefore , @xmath202 has a minimal invariant closed set @xmath203 which is either finite or a cantor set . for @xmath178 a maximal open interval in @xmath204 , we can take @xmath205 , the quotient of a subgroup of @xmath202 acting on @xmath178 .
we have @xmath206 , so by inductive hypothesis we can find the desired interval structure for the action of @xmath205 on @xmath178 and hence for @xmath1 on @xmath19 .
now our goal is to prove theorem [ smoothability ] with the aid of theorem [ structuretheorem ] .
we will assume that we are not in the situation of a minimal abelian action on @xmath10 , because actions of that type can always be conjugated to a group of rotations .
therefore , we have intervals @xmath3 and @xmath4 ; we must construct a conjugacy to adjust the @xmath1-action on these intervals and the lengths of the intervals so that the result is @xmath6 .
it will be useful to us to define a family of diffeomorphisms @xmath207 which is _ equivariant _ , i.e. such that @xmath208 .
( thus we have the structure of a groupoid , whose objects are intervals and whose morphisms are the diffeomorphisms . )
following farb - franks and navas @xcite , who credit j. c. yoccoz for the idea , for any @xmath209 we define @xmath210 by @xmath211
. then we define @xmath212 by @xmath213 .
it is immediate from this definition that the diffeomorphisms @xmath214 are equivariant .
they have very well - behaved differentiability properties , namely that @xmath215 , and @xmath216 is uniformly close to 1 provided that @xmath217 is close to 1 .
in fact , navas ( @xcite , lemma 4.1.25 ) shows that for any hlder exponent @xmath104 , if @xmath209 and @xmath218 are such that @xmath219 , @xmath220 , and @xmath221 then we get a bound on the @xmath222 _ norm _ of @xmath216 : it is these properties which make this family useful to us . if @xmath224 , @xmath225 , we will abuse notation and write @xmath226 where we really mean @xmath227 . we will similarly abuse notation for intervals lying on @xmath10 .
[ defineshomeo ] suppose we are given a set of finite disjoint open intervals @xmath228 such that @xmath229 is dense .
let @xmath230 be the length of @xmath40 for each @xmath51 .
let @xmath231 be positive real numbers such that the following condition is satisfied : whenever @xmath232 is such that @xmath233 is contained in some bounded interval , @xmath234 .
in particular , this holds if @xmath235 . then there exists a homeomorphism @xmath236 such that the length of @xmath237 is @xmath231 , and @xmath238 , where @xmath239 and @xmath240 is the lebesgue measure .
assume @xmath96 ; the other cases are almost the same .
define @xmath241 . without loss of generality , @xmath242 .
let @xmath243 ( the case @xmath244 is similar ) .
if @xmath245 , let @xmath246 . if @xmath247 , let @xmath248
first , this really does define a function on the whole real line since by assumption @xmath249 .
this function @xmath17 is monotone : if @xmath250 , then since @xmath251 is dense there is ( at least part of ) an interval @xmath40 lying between @xmath72 and @xmath165 .
we claim it is also continuous . if @xmath72 lies in an interval , continuity at @xmath72 is obvious .
suppose @xmath245 , and @xmath252 is a strictly increasing sequence of points approaching @xmath72 from the left with @xmath253 .
note that @xmath254 , since @xmath255 implies @xmath256 for some @xmath20 .
but the right - hand side is just @xmath257 .
therefore , @xmath17 is a monotone - increasing , continuous function , hence a homeomorphism onto its image .
notice that @xmath238 , since for @xmath250 , @xmath258 depends only on intervals lying between @xmath72 and @xmath165 .
recall that we write @xmath3 ( resp .
@xmath4 ) for the open intervals making up the sets @xmath40 ( resp .
@xmath41 ) in @xmath159 ( resp .
@xmath160 ) .
the numbering is arbitrary . after applying a topological conjugacy
, we want these intervals ( which by abuse of notation we will still write @xmath3 or @xmath4 ) to have certain lengths , which we now define .
let @xmath259 . recall that we are trying to conjugate @xmath1 to a group of @xmath6 diffeomorphisms , where @xmath260 .
let us consider @xmath265 , the other sum being similar .
fix @xmath51 .
note that @xmath266 .
therefore , @xmath267 , since @xmath268 implies that there is @xmath56 of length @xmath20 such that @xmath269 , but there may be more than one such @xmath56 .
thus modulo constants , the sum @xmath270 is bounded by @xmath271 .
this sum clearly converges , since it has the form @xmath272 and @xmath273 .
in fact , the reader can check that the value to which it converges decays exponentially as @xmath274 , so @xmath275 .
we apply a further topological conjugacy , after which the action of @xmath1 has the following form . the unique map from @xmath276 to @xmath3 under the action of @xmath1
is defined to be @xmath277 ( so the map from @xmath3 to @xmath278 is @xmath279 ) . for each @xmath280 , we arbitrarily choose @xmath281 such that @xmath282 ( so @xmath283 is an element of minimum length among those sending @xmath284 to @xmath4 ) .
we define @xmath283 to act via @xmath285 .
finally , we must describe how @xmath1 acts on @xmath284 after conjugacy .
let @xmath286 .
recall from before that @xmath287 is abelian .
we ask that every element of @xmath287 have the form @xmath288 .
we claim that the conjugated action of @xmath1 as we have now described it is of class @xmath6 . to prove this , we need the following lemma , which will tell us that we need only check the action is @xmath6 on the intervals @xmath3 and @xmath4 . [ patching ]
let @xmath289 be a family of closed intervals in @xmath108 $ ] ( resp . in @xmath10 ) having disjoint interiors and such that the complement of their union has zero lebesgue measure .
suppose that @xmath17 is a homeomorphism of @xmath108 $ ] ( resp .
@xmath10 ) such that its restrictions to each interval @xmath187 are @xmath6 diffeomorphisms which are @xmath5-tangent to the identity at the endpoints of @xmath187 and whose derivatives have @xmath104-norms bounded from above by a constant @xmath290 .
then @xmath17 is a @xmath6 diffeomorphism of the whole interval @xmath108 $ ] ( resp .
@xmath10 ) , and the @xmath104-norm of its derivative is less than or equal to @xmath291 . before we finish the proof of theorem [ smoothability ]
, we need a definition .
let @xmath1 be a finitely - generated group , and @xmath60 a finitely - generated subgroup .
let @xmath292 be finite generating sets of @xmath1 and @xmath60 , respectively .
we define the _ distortion _ to be @xmath293 where @xmath294 is the ball of radius @xmath20 about the identity in @xmath1 measured in the word metric coming from @xmath138 , and @xmath295 is the diameter measured in the word metric coming from @xmath296 .
although this definition depends on the chosen generating sets @xmath138 and @xmath296 , its growth type does not ; as for growth , if @xmath140 and @xmath297 are different generating sets of @xmath1 and @xmath60 respectively , then @xmath298 , so we can speak unambiguously about the equivalence class @xmath299 . for background , see e.g. @xcite . if @xmath1 is a finitely - generated virtually nilpotent group , and @xmath60 is any subgroup , we have @xmath300 , with @xmath301 only if @xmath1 is finite or has @xmath14 as a finite - index subgroup .
otherwise , @xmath1 contains a finite - index torsion - free nilpotent subgroup @xmath146 which is not trivial or @xmath14 .
it is elementary to show that the distortion of @xmath144 is equivalent to the distortion of @xmath302 . if @xmath303 , for any nontrivial subgroup @xmath304 we will have @xmath305 , which is equivalent to @xmath306 only if @xmath307
otherwise , let @xmath197 be the nilpotence class of @xmath146 .
we have @xmath151 .
it is a well - known fact ( see for example 5.2.6 . of @xcite ) that the abelianization of an infinite nilpotent group is infinite ; if @xmath146 is finitely - generated , then the torsion - free rank of @xmath308 $ ] must be at least 1 .
on the other hand , the torsion - free rank of @xmath309 is at least 1 since @xmath146 is torsion - free and @xmath309 is nontrivial .
therefore , @xmath310 . on the other hand , for any @xmath304 , @xmath311 ( for an exact expression of @xmath312
, see @xcite ) . on the @xmath3 ,
this is easy
. indeed , by construction @xmath314 can differ from @xmath315 by at most 1 .
the reader can check , with the interval lengths as we have defined them , together with equation ( 1 ) , that we do get a uniform bound on the @xmath222 norm of @xmath316 on @xmath3 .
now suppose there are some minimal intervals @xmath4 , so in particular @xmath1 is not finite and does not have a copy of @xmath14 as a finite - index subgroup .
consider the action of @xmath313 on an interval @xmath4 .
say @xmath317 .
note that , if @xmath283 and @xmath318 are our chosen maps from @xmath284 to @xmath4 and @xmath319 respectively , then @xmath320 here , @xmath321 sends @xmath284 to itself .
it is a word of length at most @xmath322 , since @xmath323 and @xmath324 .
recall from above that the degree of distortion for subgroups of @xmath1 is at most the nilpotence class @xmath325 of a finite - index torsion - free nilpotent subgroup of @xmath1 , which is strictly less than @xmath259 .
the length of @xmath321 in terms of a generating set for @xmath287 will be at most of order @xmath326 , and hence @xmath327 will be translation by an amount of order at most @xmath326 .
therefore , @xmath313 has the form @xmath328 , where @xmath329 @xmath330 being a constant that depends on the speed of translation of the generators of @xmath287 on @xmath284 . by applying the appropriate conjugacy
, we can make the translation speeds and hence the @xmath330 as small as we like .
consider the function @xmath335 on @xmath4 . as we vary @xmath51 and @xmath280 ,
we vary the constants @xmath336 , @xmath101 , and @xmath337 ; observe that for all @xmath72 and for all @xmath45 these functions are uniformly bounded , provided that we make the @xmath330 small enough .
this is because whenever @xmath338 or @xmath339 is large we will have @xmath340 and @xmath341 .
obviously , the sine and cosine functions in the numerator have magnitude at most 1 , so we only need to consider their coefficients as we vary @xmath51 and @xmath280 , these being ( disregarding a factor of @xmath342 ) @xmath343 and @xmath344 . we claim that there exists a constant @xmath290 such that these are both bounded above by @xmath345 for all @xmath45 . as before , for large @xmath338 or @xmath339 , @xmath358 is close to 1 , so we can disregard it .
therefore , we are left with @xmath359 the reader can check that as @xmath360 , the quantity @xmath361 approaches @xmath362 , so for large enough @xmath290 , we will have @xmath363 for all @xmath45 .
m. gromov , _ asymptotic invariants of infinite groups _ , vol . 2 of geometric group theory : proceedings of the symposium at sussex university ( 1991 ) , g. niblo and m. roller eds . , london math .
lecture note ser .
* 182 * , cambridge university press , cambridge , 1993 . | let @xmath0 be a connected one - manifold , and let @xmath1 be a finitely - generated nilpotent group of homeomorphisms of @xmath0 .
our main result is that one can find a collection @xmath2 of open disjoint intervals with dense union in @xmath0 , such that the intervals are permuted by the action of @xmath1 , and the restriction of the action to any @xmath3 is trivial , while the restriction of the action to any @xmath4 is minimal and abelian .
it is a classical result that if @xmath1 is a finitely - generated , torsion - free nilpotent group , then there exist faithful continuous actions of @xmath1 on @xmath0 .
farb and franks showed that for such @xmath1 , there always exists a faithful @xmath5 action on @xmath0 .
as an application of our main result , we show that _ every _ continuous action of @xmath1 on @xmath0 can be conjugated to a @xmath6 action for any @xmath7 , where @xmath8 is the degree of polynomial growth of @xmath1 . |
in crystalline insulators , the sole value of polarization is meaningless since its value depends on the state of the surface .
the meaningful one , which can be compared with experiments , is the change of polarization induced by external factors such as electric field or stress .
so a rigorous derivation of polarization is based on the analysis on the variation of the system with respect to the external electric field .
denote the expected value of @xmath95 as @xmath96 it is a physical quantity that is well defined in crystalline insulators .
then the physical energy of the system can be written as : @xmath97 where @xmath3 is the physical polarization whose expression is unknown at this point . for a ground ( stable ) state , the energy functional reaches its minimum , i.e. , @xmath98 for any infinitesimal variation of wavefunctions .
now , considering the variation of wavefunctions in a virtual process @xmath99 , from eqs .
( a2 ) and ( a3 ) we have @xmath100 so @xmath101 it relates @xmath3 to the well - defined quantity @xmath75 .
now define the berry - phase polarization as : @xmath102 it is not a physical quantity since it changes with gauge transformation . from eq .
( 5 ) , we have @xmath103 conducting differential on the above equation with respect to @xmath14 , we have @xmath104 combined with eqs .
( 10 ) , ( a5 ) and ( a6 ) , we reach @xmath105 so , the polarization defined in berry - phase expression , @xmath106 , is equal to the physical polarization @xmath3 except a constant .
under finite electric field , there are correlations between wavefunctions @xmath44 with different @xmath5 values .
however , it can be shown that @xmath44 are normalized and orthogonal as the case of @xmath39 .
compared with the schrodinger s equation @xmath107 , it is recognized that the role of @xmath5 in eq .
( 5 ) is similar to that of @xmath6 .
a consequence of this similarity is the conservation of possibility in the @xmath5-space : @xmath108 which is a physical requirement for the evolution equation .
so the wavefunction can be normalized as @xmath59 or @xmath109 .
the difference between @xmath5 and @xmath6 exists in the boundary condition . in the schrodinger s equation
, no boundary condition is required in @xmath6-space , while in eq .
( 5 ) @xmath5 and @xmath110 label the same state ( with a possible phase difference ) , so that the solutions to eq . ( 5 ) are discrete ( band structure ) . in the conventional band theory without electric field ,
the wave - functions coming from different bands are orthogonal , i.e. , @xmath111 where @xmath61 and @xmath62 denote different bands . since eq .
( 12 ) is a conventional eigen - problem , the orthogonality gives @xmath112 at finite fields .
in the first glance , eq .
( b2 ) is no longer valid .
nevertheless , to one s surprise , eq . ( b2 ) keeps valid in this case . to prove this point
, we transform @xmath113 into @xmath114 ( @xmath57 can be any lattice vector ) , which is also an eigen wavefunction of eq .
( 12 ) , thus @xmath115 by applying the knowledge of fourier s transformation on eq . (
18 ) , we arrive at eq .
( b2 ) with electric fields .
the width of the peak can be evaluated by considering the degenerate perturbation analysis on two bands @xmath116 and @xmath117 .
expanding the wavefunction on the bands @xmath61 and @xmath62 of @xmath39 , the hamiltonian becomes : @xmath118,\ ] ] where @xmath119eq .
( c1 ) describes the transition between bands @xmath61 and @xmath62 , where the transition range ( width of the peak ) is : @xmath120the degenerate condition requires that @xmath121which gives @xmath122from eq .
( 15 ) , @xmath123 is approximated as @xmath124combining eqs .
( c3,c5,c6 ) , we arrive @xmath86^{1/2}.\ ] ]
in the above main text , it is shown that our theory provides the basis for existing computational schemes@xcite .
based on the theory , other computational schemes may also be developed .
the calculation in the main text is conducted by solving the conventional eigen problem eq .
( 12 ) with subroutine zheevx in the program packet lapack , which is applicable for any finite electric fields but requires more computer capacity . here , we demonstrate a second scheme to solve eq .
the computational scheme is based on the following iteration : @xmath125@xmath5-space is discretized into 99 points ( @xmath126 ) .
iteration stops when the discrepancy is smaller than the convergence criterion : @xmath127 .
a computation example on the three - band model is shown in fig 4 .
it can be seen that the convergence is obtained at @xmath128 for the first band ( fig .
4(a ) ) , which exceeds the critical field given by ref .
9 ( @xmath129 ) by about one order .
the analytic solution for the static susceptibility at zero field can be computed from the kubo formula . in the following
we show that the susceptibility derived from the theory in this paper is consistent with what is given by kubo formula .
according to the reference ( h. ehrenreich and m. h. cohen , phys .
rev * 115 * , 786 ) , the analytical expression of the static susceptibility at zero field is given as : @xmath130 where @xmath131 is the fermi distribution and @xmath132 at zero temperature ( @xmath133 ) , @xmath134 for the current theory , with an infinitesimal variation @xmath31 , the linear term of eq .
( 5 ) at @xmath39 is : @xmath135 expand @xmath136 as @xmath137 , thus @xmath138 integral the above equation with @xmath139 yielding @xmath140 so the dielectric susceptibility is @xmath141d^3\mathbf{k } \\ & = & \frac{2e^2}{(2\pi)^3\hbar } { \rm re } \sum_{l ' } \dint \frac { \langle u_{\mathbf{k}l ' } | \nabla_{\mathbf{k } } | u_{\mathbf{k}l } \rangle ^ * \langle u_{\mathbf{k}l ' } | \nabla_{\mathbf{k } } | u_{\mathbf{k}l } \rangle } { \omega_{l'l } } d^3\mathbf{k}. \end{array}\ ] ] eq .
( e9 ) can be further simplified .
multipling @xmath142 with eq .
( 5 ) gives : @xmath143 multipling with @xmath139 yields @xmath144 so the dielectric susceptibility is @xmath145 considering the difference of the units used ( si vs. gauss ) , eqs .
( e12 ) and ( e5 ) are consistent .
essentially , the zener tunneling is a kinetic effect in which the transition rate between different states should be calculated .
therefore , zener tunneling can not be directly solved in the framework of this paper or the existing computational schemes@xcite or perturbation theories@xcite whose purpose is to obtain the eigen states of the system .
eigen states are unchanged with time and no transition will occur if no extra term is introduced into the hamiltonian . secondly , we consider the motion of the wave packet .
consider a wave packet @xmath152 at @xmath153 .
then at time @xmath6 , the @xmath5 will evolve into @xmath154 , and @xmath155 so the @xmath156 at finite fields should be used to replace the quantity at zero field to analyze the zener tunneling . | based on the conventional energy band theory , an approach is presented to describe the electronic structure of crystalline insulators in the presence of a finite homogeneous electric field .
the expression of polarization is derived which extends the `` berry - phase '' theory of polarization to nonzero fields .
the characteristics of the solutions are studied in details and the associations with the existing perturbation theories and computational schemes are discussed . in recent years , much attention was paid to the theory of macroscopic polarization and the effect of homogeneous electric field in crystalline solids.@xcite although the energy band theory and the density - functional theory ( dft ) have provided a solid basis in understanding the equilibrium properties of crystals , the problem of crystalline insulators in the presence of a homogeneous macroscopic electric field was not fully solved.@xcite the obstacle comes from the fact that the electric potential @xmath0 is a linear term in the spatial coordinates , which violates the periodicity condition underlying bloch s theorem . moreover , strictly speaking , there is no ground state for a crystalline system under an external field .
on the other hand , crystalline insulators are experimentally stable when the applied field is not too strong , and the field - dependent response of the systems ( such as dielectric and piezoelectric properties ) is well defined in the measurements .
this looks like a paradox .
it appears that some works are needed to bridge the gap between the theory and the experiment .
some progress within the framework of perturbation theory has been achieved in this direction,@xcite nevertheless finite electric fields can not be handled directly in such theories due to the weakness of the perturbation approach .
computational schemes applicable at finite electric fields were developed.@xcite in such schemes , the non - periodic electric potential term @xmath1 in the energy functional is replaced by @xmath2 where the polarization @xmath3 is expressed in the berry - phase approach that holds the periodicity condition .
they are effective in numerical computing , but the physical basis is not very clear . on the basis of a band - like electronic - structure theory under finite electric fields ,
this paper aims at providing basic theoretical understandings on the subject .
the expression of macroscopic polarization is derived to extend the `` berry - phase '' theory of polarization into the cases of nonzero fields .
it is shown that the formalism can be transferred into a conventional eigen problem in @xmath4-space .
the properties of the solutions are identified , and the connections with the existing perturbation theories and computational schemes are discussed .
when the system evolves from the state ( @xmath5 , @xmath6 ) to ( @xmath7 , @xmath8 ) . if we assume the phase keeps unchanged when the system transit from @xmath5 to @xmath9 ( process of dash line ) , the total variation of phase is @xmath10 .
, width=321 ] the theory can be developed as what were done in refs . 5 and 11 , or by use of the technique of the crystal momentum representation@xcite .
here we present a simplified derivation that keeps the insights in physics .
our approach is based on the conventional energy band theory . according to the bloch s theorem
, the eigen - state of an electron moving in a periodic potential @xmath11 can be expressed as @xmath12 where @xmath13 is a periodic function .
when there exists an external electric field @xmath14 , the hamiltonian of the system is given as : @xmath15 in this case , @xmath5 is no longer the good quantum number to specify the eigen - state . in the general solid state
physics,@xcite the response of solids to an electric field is described as the constant - velocity movement of the wave vector @xmath5 , i.e. , @xmath16the bands are either fully filled or empty in insulators , so a static equilibrium state is reached regardless of the movement of @xmath5 . since @xmath5 varies with the time @xmath6 , the differential of eq .
( 1 ) with respect to @xmath6 yields : @xmath17a key point in deriving the above equation is that we assume the phase @xmath18 keeps unchanged when the system transits between different @xmath19 states , so @xmath20 is ignored in the analysis ( see fig . 1 ) .
it can be demonstrated more clearly as follows : if the system transits from an initial wave - function @xmath21 to a final one @xmath22 , then for a second initial wave - function @xmath23 , the system will transit into the final one @xmath24 , i.e. , the initial phase keeps unchanged . such a process is invariant for time translation .
in contrast , if the term @xmath25 in fig . 1 ( dash line )
is included , the symmetry of the system for time translation will be violated .
based on the above analysis and the schrdinger s equation @xmath26 , we get the main equation of crystalline insulators under electric field as : @xmath27where @xmath28it can be seen that the introduction of an electric field causes interactions between different @xmath5 states , and @xmath29 can be no longer solved independently .
however , the periodicity condition in @xmath30-space can be safely applied upon eq .
( 5 ) , while eq .
( 2 ) is not periodic , thus it overcomes the difficulties mentioned above .
essentially , eq . ( 5 ) is equivalent to what is obtained in refs . 5 and 11 , but it is more convenient for the following analyzing . to one s surprise , the calculation of polarization in electronic structure theory remained a challenge until the berry - phase theory of polarization was developed in the early 1990s.@xcite although such polarization theory has been successfully applied in many cases , its validity under nonzero electric field was not proven in any publications . with eq .
( 5 ) , the expression of polarization at finite fields can be derived easily@xcite .
consider an infinitesimal variation of electric field , @xmath31 , in the system .
the linear terms of eq .
( 5 ) with respect to @xmath32 yields @xmath33multiplying the above equation by @xmath34 and integrating it , one gets @xmath35if the periodic condition is adopted in @xmath5-space as @xmath36where @xmath37 is any reciprocal lattice vector .
so the macroscopic polarization is achieved as @xmath38which is consistent with the result given by the berry - phase theory.1,2 although the original derivation of the polarization s formula in the berry - phase approach assumed @xmath39 , the above analysis showed that the same formula keeps valid for @xmath40 .
gauge transformation plays an important role in eq .
( 5 ) .
it is noted that eq .
( 5 ) is invariant under the following transformation : @xmath41 , \\
\widetilde{e}_{\mathbf{k}}=e_{\mathbf{k}}-e\mathbf{\cal e } \cdot \nabla _ { \mathbf{k } } \beta ( \mathbf{k}),% \end{array}% \right.\ ] ] where @xmath42 is an arbitrary function .
such phase arbitrariness can not be removed even if the periodic boundary condition [ eq . ( 9 ) ] is adopted .
this reveals that the polarization in eq .
( 10 ) is a multivalued quantity .
gauge transformation will not change the physical phase of the system . within an infinitesimal time interval @xmath43 ,
the extra phase of @xmath44 cased by gauge transformation is @xmath45 while that of @xmath46 is @xmath47 . with ( 3 )
we arrive @xmath48 .
so the gauge transformation just moves a part of the phase from @xmath44 to @xmath46 while keep the total phase of @xmath49 unchanged .
it does not cause any measurable effects .
furthermore , such an arbitrariness can be utilized to simplify some analyses . by choosing @xmath42 appropriately ,
@xmath50 can be transformed into a constant independent of @xmath5 , i.e. , @xmath51 , and thus eq .
( 5 ) becomes a conventional eigen - problem in ( @xmath52)-space : @xmath53whose properties can be well understood based on the existing knowledge of eigen problem .
some properties of the solution are listed in the following : ( 1 ) if there are @xmath54 bands for @xmath39 , then eq . ( 12 ) has @xmath54 physically independent solutions for @xmath40 , each of which is associated with a group of physically equivalent solutions as @xmath55 and @xmath56 where @xmath57 is any lattice site ; ( 2 ) @xmath58 so that @xmath44 can be normalized as @xmath59 , which is not self - evident for @xmath40 because different @xmath5 states are correlated by eqs ( 5 ) or ( 12 ) ; ( 3 ) the wavefunctions coming from different bands and with the same @xmath5 value are orthogonal , i.e. , @xmath60 , where @xmath61 and @xmath62 denote different bands . , ( b ) polarization @xmath63 and ( c ) usual energy with polarization s contribution excluded , @xmath64 , of three bands for the tigh - binding model ( see eq . ( 14 ) ) as functions of electric field .
labels @xmath65 , @xmath66 and @xmath67 in fig .
2(c ) are used to specify the index @xmath67 of peaks between different bands ( see eq . ( 16 ) ) .
, width=321 ] eq .
( 12 ) is equivalent to the variational analysis on the energy functional : @xmath68with @xmath3 in eq .
( 10 ) , which is the basis of the perturbation theories and computational schemes in references@xcite . to learn more characteristics of the solution
, we conduct calculations on a 1d tight - binding hamiltonian with three bands:@xcite @xmath69 \right\},\ ] ] where the site energy is given as @xmath70 .
@xmath61 , @xmath6 and @xmath71 are three parameters of the hamiltonian .
the position operator is set to be @xmath72 since there are three atoms in each unit cell . in fig . 2 ,
the properties of three bands are reported as functions of electric field for the parameter @xmath73 and @xmath74 .
the most extraordinary feature at finite electric fields is that there are many peaks in the curves of polarization @xmath63 and the usual energy with polarization s contribution excluded , @xmath75 ( @xmath76 ) .
correspondingly , the curves of band energy @xmath77 are discontinuous with many gaps .
the peaks are caused by the overlap of the energy of different bands . according to eq .
( 12 ) , the dielectric susceptibility of the band @xmath61 is given as : @xmath78 note that the band energy is a multivalued quantity with a term @xmath79 where @xmath80 is the crystal lattice in the current one - dimensional case .
so if the electric field satisfies the following relation : @xmath81 the dielectric susceptibility diverges and its sign changes when the electric field goes across such points , which results in the peaks of the polarization .
the revealed properties of polarization at finite fields here are helpful in understanding the convergency of the perturbation theories .
the electric filed associated with the peaks of polarization is a kind of `` singular point '' . from eq .
( 16 ) it is known that there are more and more such singular points when the electric filed is getting smaller .
especially , the zero field @xmath82 is a non - isolated singular point of which there are always other singular points within any small neighborhoods .
therefore , if the properties of the system are expanded in the form of power series in the electric field , the series are not convergent within any finite neighborhood of radius centered about @xmath82 .
that is the reason why the perturbation theories in continuum formulation are valid only for infinitesimal fields.@xcite when the band energy @xmath77 has a minimal value , the solution of ground state can be obtained by minimizing the energy functional eq .
( 13 ) .
unfortunately , the minimal value of @xmath77 does not exist in continuum formulation due to the term @xmath79 .
a technique to remove such difficulty is to perform a discretization on the @xmath5-space , i.e. , to restrict the maximal value of the number @xmath67 , which has been adopted in numerical schemes.@xcite however , when the electric field is large enough to shift the higher band to below the lower band , i.e. , @xmath83 where @xmath84 is the discretization number in @xmath5-space , the minimizing procedure will fail in obtaining the ground state .
that is the phenomena of collapse appeared in the numerical schemes.@xcite it is clear that such phenomena are associated with the breaking of the numerical schemes , but not with the breaking of the system such as zener tunneling@xcite .
although there are infinite peaks in the calculated curve of polarization around @xmath82 , the properties of the system is smooth enough in experiments .
so it is necessary to have a short discussion on the width of the peaks . after a simple analysis of degenerate perturbation around the peak
, it is shown that the width of the peaks , @xmath85 , is approximately estimated as : @xmath86^{1/2},\ ] ] where @xmath87 and @xmath88 are the average band energy when there is no electric field . for a numerical estimate
, we set @xmath89 , @xmath90 and @xmath91 .
then we get @xmath92^{1/2}= 3.3\times 10^{-11}$ ] ( in si unit ) .
it is obvious that such narrow peaks at small electric fields are too difficult to be observed experimentally .
therefore , the properties obtained in experiments reflect an average effect where the peaks are smoothed out . in fig .
3 we depict the curves of the average polarization and dielectric susceptibility after the peaks are removed .
nonlinear effect can be observed .
the dielectric susceptibility is non - monotonic , which means some high - order susceptibility terms are negative .
it is noted that some issues discussed in this paper , such as the multivalued property of the polarization and the absence of ground state at finite electric fields , are caused by the periodic boundary conditions imposed on the insulators .
if we consider a finite piece of a solid and surface effects , such problems will disappear .
that is what should be kept in mind when one considers the effect of electric fields on insulators .
we would like to thank dr .
jian - she liu for helpful discussions .
this work was supported by the chinese state key program of basic research development ( grant no.tg2000067108 ) , the national high technology research and development program ( grant no.2002aa311153 ) and the national natural science foundation of china .
r. d. king - smith and d. vanderbilt , phys .
rev .
b * 47 * , 1651 ( 1993 ) .
r. resta , rev .
mod .
phys . * 66 * , 899 ( 1994 ) .
n. sai , k. m. rabe , and d. vanderbilt , phys .
rev .
b * 66 * , 104108 ( 2002 ) .
r. w. nunes and x. gonze , phys .
rev .
b * 63 * , 155107 ( 2001 ) .
g. h. wannier , phys .
rev .
* 117 * , 432 ( 1960 ) .
s. baroni , p. giannozzi , and a. testa , phys .
rev .
lett . * 58 * , 1861 ( 1987 ) .
x. gonze and j. p. vigneron , phys .
rev .
b * 39 * , 13120 ( 1989 ) .
r. w. nunes and d. vanderbilt , phys .
rev .
lett . * 73 * , 712 ( 1994 ) .
i. souza , j. iniguez , and d. vanderbilt , phys .
rev .
lett . * 89 * , 117602 ( 2002 ) .
umari and a. pasquarello , phys .
rev .
lett . * 89 * , 157602 ( 2002 ) .
see , for instance , c. kittel , _ introduction to solid state physics _
( wiley , new york , 1986 ) , chap .
9 .
e. i. blout , in _ solid state physics _ , edited by f.
seitz and d. turnbull ( academic , new york , 1962 ) , vol . 13 , p. 305 . r. m. martin , phys .
rev .
b * 9 * , 1998 ( 1974 ) . a longer but
more rigorous derivation is based on the formula @xmath93 where @xmath94 , which is not given in this paper . c. zener , proc .
roy .
soc .
( london ) * a145 * , 523 ( 1934 ) . *
note : the following appendix are provided for the referees reference , but not for publication . * |
many - electron states in quantum dots have been theoretically investigated by various approaches , e.g. , the hartree - fock method @xcite , the density - functional theory @xcite , the quantum monte carlo simulation @xcite , the variational monte carlo method and the pad approximation @xcite , a numerical diagonalization of the hamiltonian in a finite - dimensional basis @xcite .
the electron - phonon interaction was not taken into account in these investigations , although it can contribute significantly to both equilibrium and non - equilibrium properties of quantum dots .
for instance , the effects due to the electron - phonon interaction play a key role in the optical spectra of some quantum dots ( see ref . @xcite and references therein ) .
some characteristics revealed in the mid - infrared region of the experimentally observed optical absorption spectra of high-@xmath2 cuprates @xcite were assigned @xcite to the polaron optical absorption at intermediate values of the electron - phonon coupling constant @xmath1 .
the infrared optical absorption band in the neodymium - cerium cuprate nd@xmath3ce@xmath4cuo@xmath5 ( ncco ) , which was studied experimentally @xcite as a function of the electron density , has been associated with the polaron optical absorption .
also some peculiarities of the infrared optical absorption spectra of cuprates and pr@xmath6nio@xmath7 were interpreted in terms of a mixture of large and small polarons or bipolarons @xcite .
the theory of the optical conductivity of arbitrary - coupling single polarons has been developed in refs .
@xcite ( see also the review @xcite and references therein ) within the memory - function method based on the path - integral formalism @xcite . recently , the optical absorption of a gas of interacting polarons at weak coupling was investigated @xcite on the basis of a variational many - particle approach @xcite .
the optical absorption of interacting polarons in bulk semiconductors at arbitrary coupling strength has been treated in ref .
@xcite within the random - phase approximation , using variational parameters obtained from feynman s _ single - polaron _
model @xcite , but this treatment does not seem to be a self - consistent approach to the many - polaron problem .
in contrast to the polaron mechanism of optical absorption in bulk ( see refs . @xcite and references therein ) , the polaron optical absorption in quantum dots , to the best of our knowledge , has not yet been widely studied . in order to investigate the ground state and the optical response of a system of interacting polarons in a quantum dot , it is crucial to take into account the fact that the system contains a _
finite _ number of identical particles ( electrons or holes ) .
indeed , the thermodynamic properties of systems with a finite number of identical particles might substantially deviate from those obtained within the grand - canonical formalism ( see , e.g. , @xcite ) .
the variational path - integral method for identical particles @xcite provides a useful tool for investigating interacting quantum many - body systems with a fixed ( few or many ) number of @xmath0 particles .
an outline of the method is given in ref .
@xcite , where we sketched the calculation of the ground - state energy of a fixed number of interacting polarons , which form a closed - shell system in a quantum dot . in the present paper the calculation of the ground state and of the optical conductivity
is performed for both closed - shell and open - shell systems of interacting polarons in a quantum dot . in section [ sec :
freeenergy ] , we derive an upper bound to the free energy of a finite number of interacting polarons confined in a parabolic quantum dot in three dimensions and in two dimensions . in section [ sec : groundstate ] we discuss the numerical results for the ground - state energy and for the addition energy of this system . in section [ sec : optconductivity ] , the optical conductivity of interacting polarons in a quantum dot is derived on the basis of the memory - function method .
the numerical results for the optical conductivity are discussed in section [ sec : resultsconductivity ] .
the last section [ sec : conclusions ] contains conclusions .
we consider a system of @xmath0 electrons with mutual coulomb repulsion and interacting with the lattice vibrations .
the system is assumed to be confined by a parabolic potential characterized by the frequency parameter @xmath8 .
the total number of electrons is represented as @xmath9 where @xmath10 is the number of electrons with the spin projection @xmath11 .
the electron 3d ( 2d ) coordinates are denoted by @xmath12 with @xmath13 the bulk phonons ( characterized by 3d wave vectors @xmath14 and frequencies @xmath15 ) are described by the complex coordinates @xmath16 which possess the property @xcite @xmath17 the full set of the electron and phonon coordinates are denoted by @xmath18 and @xmath19 throughout the sections [ sec : freeenergy ] and [ sec : groundstate ] , the euclidean time variable @xmath20 is used , where @xmath21 is the real time variable . in this representation
the lagrangian of the system is @xmath22 where @xmath23 is the lagrangian of an electron with band mass @xmath24 in a quantum dot : @xmath25 @xmath26 is the potential energy of the electron - electron coulomb repulsion in the medium with the high - frequency dielectric constant @xmath27 : @xmath28 @xmath29 is the lagrangian of free phonons : @xmath30 further , @xmath31 is the lagrangian of the electron - phonon interaction : @xmath32 where @xmath33 is the fourier transform of the electron density operator:@xmath34 @xmath35 is the amplitude of the electron - phonon interaction . in this paper
, we only consider electrons interacting with the long - wavelength longitudinal optical ( lo ) phonons with a dispersionless frequency @xmath36 , for which the amplitude @xmath35 is @xcite@xmath37 where @xmath1 is the electron - phonon coupling constant and @xmath38 is the volume of the crystal .
we treat a _ canonical _
ensemble , where the numbers @xmath10 are fixed .
the partition function @xmath39 of the system can be expressed as a path integral over the electron and phonon coordinates : @xmath40 } , \label{z}%\ ] ] where @xmath41 $ ] is the `` action '' functional : @xmath42 = -\frac{1}{\hbar}\int_{0}^{\hbar\beta}l\left ( \mathbf{\dot{\bar{x}}% } , \dot{\bar{q}};\mathbf{\bar{x}},\bar{q}\right ) d\tau.\label{s}%\ ] ] the parameter @xmath43 is inversely proportional to the temperature @xmath44 . in order to take the fermi - dirac statistics into account ,
the integral over the electron paths @xmath45 in eq .
( [ z ] ) contains a sum over all permutations @xmath46 of the electrons with the same spin projection , and @xmath47 denotes the parity of a permutation @xmath46 .
the action functional ( [ s ] ) is quadratic in the phonon coordinates @xmath48 therefore , the path integral over the phonon variables in @xmath49 can be calculated analytically @xcite . as a result ,
the partition function of the electron - phonon system ( [ z ] ) factorizes into a product @xmath50 of the partition function of free phonons with a partition function @xmath51 of interacting polarons , which is a path integral over the electron coordinates only:@xmath52 } .\label{zp}%\ ] ] the functional @xmath53 & = -\frac{1}% { \hbar}\int_{0}^{\hbar\beta}\left [ l_{e}\left ( \mathbf{\dot{\bar{x}}}\left ( \tau\right ) , \mathbf{\bar{x}}\left ( \tau\right ) \right ) + v_{c}\left ( \mathbf{\bar{x}}\left ( \tau\right ) \right ) \right ] d\tau\nonumber\\ & -\sum_{\mathbf{q}}\frac{\left| v_{\mathbf{q}}\right| ^{2}}{2\hbar^{2}}% \int\limits_{0}^{\hbar\beta}d\tau\int\limits_{0}^{\hbar\beta}d\tau^{\prime } \frac{\cosh\left [ \omega_{\mathrm{lo}}\left ( \left| \tau-\tau^{\prime } \right| -\hbar\beta/2\right ) \right ] } { \sinh\left ( \beta\hbar \omega_{\mathrm{lo}}/2\right ) } \rho_{\mathbf{q}}\left ( \tau\right ) \rho_{-\mathbf{q}}\left ( \tau^{\prime}\right ) \label{sp}%\end{aligned}\ ] ] results from the elimination of the phonon coordinates and contains the `` influence phase '' of the phonons ( the last term in the right - hand side ) .
it describes the phonon - induced retarded interaction between the electrons , including the retarded self - interaction of each electron .
the free energy of a system of interacting polarons @xmath54 is related to their partition function ( [ zp ] ) by the equation : @xmath55 at present no method is known to calculate the non - gaussian path integral ( [ zp ] ) analytically . for _
distinguishable _ particles , the jensen - feynman variational principle @xcite provides a convenient approximation technique .
it yields a lower bound to the partition function , and hence an upper bound to the free energy .
the formulation of a variational principle for the free energy for a system of _ identical _ _ particles _ is a _ _ _ _ non - trivial problem .
however , it can be shown @xcite that the path - integral approach to the many - body problem for a fixed number of identical particles can be formulated as a feynman - kac functional on a state space for @xmath0 indistinguishable particles , by imposing an ordering on the configuration space and by the introduction of a set of boundary conditions at the boundaries of this state space .
the path integral ( in the imaginary - time variable ) for identical particles was shown to be _ positive _ within this state space .
this implies that a many - body extension of the jensen - feynman inequality was found , which can be used for interacting identical particles ( ref.@xcite , p. 4476 ) .
a more detailed analysis of this variational principle for both local and retarded interactions can be found in ref .
it is required that the potentials are symmetric with respect to all permutations of the particle positions , and that both the exact propagator and the model propagator are antisymmetric ( for fermions ) with respect to permutations of any two electrons at any moment in time .
this means that these propagators have to be defined on the same configuration space . keeping in mind these constraints
, the variational inequality for identical particles takes the same form as the jensen - feynman variational principle : @xmath56 where @xmath57 is a model action with corresponding free energy @xmath58 .
the angular brackets mean a weighted average over the paths@xmath59 } } { \sum_{p}\frac{\left ( -1\right ) ^{\mathbf{\xi}_{p}}}{n_{1/2}!n_{-1/2}!}\int d\mathbf{\bar{x}}\int _ { \mathbf{\bar{x}}}^{p\mathbf{\bar{x}}}d\mathbf{\bar{x}}\left ( \tau\right ) e^{-s_{0}\left [ \mathbf{\bar{x}}\left ( \tau\right ) \right ] } } .\label{aver}%\ ] ] we consider a model system consisting of @xmath0 electrons with coordinates @xmath18 and @xmath60 fictitious particles with coordinates @xmath61 in a harmonic confinement potential with elastic interparticle interactions as studied in ref .
the lagrangian of this model system takes the form@xmath62 the frequencies @xmath63 @xmath64 @xmath65 the mass of a fictitious particle @xmath66 and the force constant @xmath67 are variational parameters . clearly , this lagrangian is symmetric with respect to electron permutations .
performing the path integral over the coordinates of the fictitious particles @xcite , the partition function @xmath68 of the model system of interacting polarons becomes a path integral over the electron coordinates:@xmath69 } , \label{z0}%\ ] ] with the action functional @xmath70 $ ] given by@xmath71 & = \frac{1}{\hbar } \int_{0}^{\hbar\beta}\sum_{\sigma}\sum_{j=1}^{n_{\sigma}}\frac{m_{b}}% { 2}\left [ \mathbf{\dot{x}}_{j,\sigma}^{2}\left ( \tau\right ) + \omega ^{2}\mathbf{x}_{j,\sigma}^{2}\left ( \tau\right ) \right ] d\tau\nonumber\\ & -\frac{1}{\hbar}\int_{0}^{\hbar\beta}\sum_{\sigma,\sigma^{\prime}}\sum _ { j=1}^{n_{\sigma}}\sum_{l=1}^{n_{\sigma^{\prime}}}\frac{m_{b}\omega^{2}}% { 4}\left [ \mathbf{x}_{j,\sigma}\left ( \tau\right ) -\mathbf{x}% _ { l,\sigma^{\prime}}\left ( \tau\right ) \right ] ^{2}d\tau\nonumber\\ & -\frac{k^{2}n^{2}n_{f}}{4m_{f}\hbar\omega_{f}}\int\limits_{0}^{\hbar\beta } d\tau\int\limits_{0}^{\hbar\beta}d\tau^{\prime}\frac{\cosh\left [ \omega _ { f}\left ( \left| \tau-\tau^{\prime}\right| -\hbar\beta/2\right ) \right ] } { \sinh\left ( \beta\hbar\omega_{f}/2\right ) } \mathbf{x}\left ( \tau\right ) \cdot\mathbf{x}\left ( \tau^{\prime}\right ) , \label{s0}%\end{aligned}\ ] ] where @xmath72 is the center - of - mass coordinate of the electrons,@xmath73 the details of the analytical calculation of the model partition function ( [ z0 ] ) are described in appendix a. after substituting the model action functional ( [ s0 ] ) into the right - hand side of the variational inequality ( [ jf ] ) , we obtain an upper bound to the free energy @xmath74@xmath75 \nonumber\\ & + \frac{k^{2}n^{2}n_{f}}{4m_{f}\beta\hbar\omega_{f}}\int\limits_{0}% ^{\hbar\beta}d\tau\int\limits_{0}^{\hbar\beta}d\tau^{\prime}\frac{\cosh\left [ \omega_{f}\left ( \left| \tau-\tau^{\prime}\right| -\hbar\beta/2\right ) \right ] } { \sinh\left ( \beta\hbar\omega_{f}/2\right ) } \left\langle \mathbf{x}\left ( \tau\right ) \cdot\mathbf{x}\left ( \tau^{\prime}\right ) \right\rangle _ { s_{0}}\nonumber\\ & -\sum_{\mathbf{q}}\frac{\left| v_{\mathbf{q}}\right| ^{2}}{2\hbar^{2}% \beta}\int\limits_{0}^{\hbar\beta}d\tau\int\limits_{0}^{\hbar\beta}% d\tau^{\prime}\frac{\cosh\left [ \omega_{\mathrm{lo}}\left ( \left| \tau -\tau^{\prime}\right| -\hbar\beta/2\right ) \right ] } { \sinh\left ( \beta\hbar\omega_{\mathrm{lo}}/2\right ) } g\left ( \mathbf{q},\tau -\tau^{\prime}|\left\ { n_{\sigma}\right\ } , \beta\right ) .\label{varfun}%\end{aligned}\ ] ] here , @xmath76 is the two - point correlation function for the electron density operators:@xmath77 both the free energy and the correlation functions of the model system can be calculated analytically using the generating function technique @xcite .
in the zero - temperature limit @xmath78 the variational free energy ( [ varfun ] ) becomes an upper bound @xmath79 to the ground - state energy @xmath80 of the system of interacting polarons .
the details of the calculation of the correlation functions are given in the appendix b.
for the numerical calculations , we use effective atomic units , where @xmath81 the electron band mass @xmath24 and @xmath82 have the numerical value of 1 .
this means that the unit of length is the effective bohr radius @xmath83 , while the unit of energy is the effective hartree@xcite @xmath84 .
these units allow to present results for quantum dots with and without the electron - phonon interaction on the same scale .
therefore , for confined polarons they are more convenient than the usual polaron units , where the unit of length is @xmath85 ^{1/2},$ ] and the energy is measured in units of the lo - phonon energy @xmath86 in terms of the dimensionless parameters @xmath1 and @xmath87where @xmath88 is the static dielectric constant , the following relations exist between both systems of units:@xmath89 in fig .
1 , the total spin @xmath90 of a system of interaction polarons in their ground state is plotted as a function of the number of electrons in a 3d quantum dot for different values of the confinement frequency @xmath91 , of the electron - phonon coupling constant @xmath1 and of the parameter @xmath92 as distinct from few - electron systems without the electron - phonon interaction , three types of spin polarization are possible for the ground state , which should be distinguishable from each other using , e.g. , capacity measurements .
\(i ) except for the strong - coupling case and for the low - density case , the filling in the ground state is as follows : in an open shell , with less - than - half filling , each new electron is added with one and the same spin , so that the total spin ( in the shell under consideration ) is maximal in accordance with hund s rule @xcite .
as soon as half - filling is achieved with electrons possessing a certain spin , each new electron is added with the spin opposite to that in the group of electrons providing the aforementioned half - filling . when the number of electrons corresponds to the number of states in the shell under consideration , the shell becomes closed , and the total spin is zero .
this mode of filling is referred to as hund s rule for a quantum dot .
hund s rule means , that the electrons in a partly filled upper shell build up a minimal possible number of pairs in order to minimize the electron - electron repulsion . for a quantum dot with @xmath93 @xmath94 at @xmath95 and at @xmath96
, the shell filling always obeys hund s rule , as shown in fig .
1(_a _ ) .
\(ii ) with decreasing confinement frequency @xmath91 at a given number of electrons , the electron density lowers . for densities smaller than a certain value
, it can happen that the ground state is a state with a maximal total ( in all shells ) spin . in this state
, the electrons are filling consecutively all the single - electron states with one and the same spin and are referred to as spin - polarized .
the examples are the states at @xmath95 for @xmath97 and @xmath98 in fig .
a spin - polarized ground state precedes the formation of a wigner crystal when further lowering the density @xcite .
\(iii ) in the strong - coupling case ( @xmath99 and @xmath100 ) , it can happen that the ground state is a state with a minimal total spin ( 0 for even number of electrons and @xmath101 for odd number of electrons ) .
this is the case when due to the phonon - mediated electron - electron attraction pairing of electrons with opposite spins occurs , analogous to a singlet bipolaron ground state in bulk .
the examples are the states at @xmath102 and @xmath103 for @xmath0 in the range from 4 to 6 in fig .
1(_a _ ) and for @xmath0 in the range from 4 to 10 in fig .
this trend to minimize the total spin is a consequence of the electron - phonon interaction , presumably due to the fact that the phonon - mediated electron - electron attraction overcomes the coulomb repulsion . with an increasing number of electrons , at a certain value of @xmath0 , such states with a minimal total spin cease to form the ground state , and the shell filling abruptly returns to that prescribed by hund s rule
[ see a jump in the spin at @xmath102 when @xmath0 changes from 10 to 11 in fig .
1(_b _ ) ] .
this jump is analogous to a transition from states with paired electrons ( like superconducting states ) to another type of states with unpaired electrons ( like normal states ) .
the addition energy @xmath104 , which is the variation of the chemical potential when putting an extra electron into a quantum dot , is defined as @xcite@xmath105 fig . 2
presents the addition energy in a 3d quantum dot as a function of the number of electrons .
the structure of @xmath104 clearly manifests the shell structure of a quantum dot .
the most pronounced peaks in the addition energy occur for closed - shell systems with @xmath106 . the peaks in @xmath104 at @xmath107 and @xmath108 obtained within the present approach for relatively weak electron - phonon coupling correspond to the systems with the half - filled upper shell ( see fig . 2(_a _ ) for @xmath95 and @xmath96 ) .
in these cases the total spin for the upper shell takes its maximal possible value , in accordance with hund s rule . at sufficiently large values of @xmath109
the electron - phonon interaction substantially modifies the addition energy . in the strong - coupling case ,
the peaks corresponding to half - filled shells become less pronounced , while those corresponding to closed - shell systems become more prominent as compared to the weak - coupling case . to the best of our knowledge , the addition energy for parabolic quantum dots was obtained using the density functional theory ( dft ) ( see , e.g. , refs.@xcite ) only without the electron - phonon interaction .
our results for the addition energy for a 3d quantum dot as a function of @xmath0 in the particular case @xmath95 are very close for @xmath110 to those calculated within the dft @xcite with an optimized effective potential and a self - interaction correction @xcite .
the panels _ a _ and _ b _ in fig . 3 represent , respectively , the total spin and the addition energy for interacting polarons in a 2d parabolic gaas quantum dot with the confinement parameter @xmath111 mev . the pronounced peaks in @xmath104 at @xmath112 correspond to the closed - shell systems , for which the total spin equals zero .
in accordance with hund s rule , the upper shell is filled in such a way that the total spin of electrons in this shell takes the maximal possible value .
therefore for the half - filled upper shell ( at @xmath113 ) maxima of the total spin occur as a function of @xmath114 at these electron numbers , the addition energy manifests peaks , which are less pronounced than those corresponding to closed shells .
the inset to fig .
3(_b _ ) shows the experimental data for the addition energy in a cylindric gaas quantum dot @xcite . as seen from fig .
3 , the peak positions for the addition energy of interacting polarons in a 2d parabolic quantum dot agree well with the experimental results for the addition energies of cylindrical quantum dots .
the height of the calculated peaks of the addition energy falls down as the shell number increases , which is qualitatively consistent with the experimentally observed behavior @xcite .
the peaks in @xmath104 corresponding to the half - filled shells are weaker than those for the closed shells both in the experiment@xcite and in our theory .
for a system of interacting polarons in a parabolic confinement potential , we calculate the real part of the optical conductivity within the memory - function approach . for a single polaron at arbitrary coupling strength
it was developed in refs .
@xcite . for a polaron gas in the weak - coupling limit , this technique was applied in ref .
@xcite . in the present paper
we extend the memory - function approach to a system of arbitrary - coupling interacting polarons .
since the optical conductivity relates the current @xmath115 per electron to a time - dependent uniform electric field @xmath116 in the framework of linear response theory , we have to return to the _ real - time _ representation in the path integrals .
the fourier components of the electric field are denoted by @xmath117 @xmath118 and the similar denotations are used for other time - dependent quantities .
the electric current per electron @xmath115 is related to the mean electron coordinate response @xmath119 by @xmath120 and hence@xmath121 within the linear - response theory , both the electric current and the coordinate response are proportional to @xmath122:@xmath123 where @xmath124 is the conductivity per electron . because we treat an isotropic electron - phonon system
, @xmath125 is a scalar function .
it is determined from the time evolution of the center - of - mass coordinate : @xmath126 the symbol @xmath127 denotes an average in the _ real - time _
representation for a system with action functional @xmath90:@xmath128 } \left ( \bullet\right ) \left .
\left\langle \mathbf{\bar{x}}_{0}\left| \hat{\rho}\left ( t_{0}\right ) \right|
\mathbf{\bar{x}}_{0}^{^{\prime}}\right\rangle \right| _ { t_{0}% \rightarrow-\infty},\label{average}%\ ] ] where @xmath129 is the density matrix before the onset of the electric field in the infinite past @xmath130 .
the corresponding action functional is @xcite@xmath131 = \int\limits_{-\infty}^{t}\left [ l_{e}\left ( \mathbf{\dot{\bar{x}}}\left ( t\right ) , \mathbf{\bar{x}}\left ( t\right ) , t\right ) -l_{e}\left ( \mathbf{\dot{\bar{x}}}^{\prime}\left ( t\right ) , \mathbf{\bar{x}}^{\prime}\left ( t\right ) , t\right ) \right ] dt^{\prime } -i\hbar\phi\left [ \mathbf{\bar{x}}\left ( t\right ) , \mathbf{\bar{x}% } ^{\prime}\left ( t\right ) \right ] , \label{s1}%\ ] ] where @xmath132 is the lagrangian of @xmath0 interacting electrons in a time - dependent uniform electric field @xmath116 @xmath133 the influence phase of the phonons ( see , e.g. , ref .
@xcite ) @xmath134 & = -\sum_{\mathbf{q}}\frac{\left| v_{\mathbf{q}% } \right| ^{2}}{\hbar^{2}}\int\limits_{-\infty}^{t}ds\int\limits_{-\infty}% ^{s}ds^{\prime}\left [ \rho_{\mathbf{q}}\left ( s\right ) -\rho_{\mathbf{q}% } ^{\prime}\left ( s\right ) \right ] \nonumber\\ & \times\left [ t_{\omega_{\mathbf{q}}}^{\ast}\left ( s - s^{\prime}\right ) \rho_{\mathbf{q}}\left ( s^{\prime}\right ) -t_{\omega_{\mathbf{q}}}\left ( s - s^{\prime}\right ) \rho_{\mathbf{q}}^{\prime}\left ( s^{\prime}\right ) \right ] \label{phi}%\end{aligned}\ ] ] describes both a retarded interaction between different electrons and a retarded self - interaction of each electron due to the elimination of the phonon coordinates .
this functional contains the free - phonon green s function : @xmath135 the equation of motion for @xmath119 can be derived by analogy with that described in ref .
@xcite:@xmath136 where @xmath137 is the average force due to the electron - phonon interaction , @xmath138 the two - point correlation function @xmath139 should be calculated from eq .
( [ average ] ) using the exact action ( [ s1 ] ) , but like for the free energy above , this path integral can not be calculated analytically .
instead , we perform an approximate calculation , replacing the two - point correlation function in eq .
( [ f ] ) by @xmath140 where @xmath141 $ ] is the action functional with the optimal values of the variational parameters for the model system considered in the previous section in the presence of the electric field @xmath116 .
the functional @xmath142 $ ] is quadratic and describes a system of coupled harmonic oscillators in the uniform electric field @xmath143 .
this field enters the term @xmath144 in the lagrangian , which only affects the center - of - mass coordinate .
hence , a shift of variables to the frame of reference with the origin at the center of mass @xmath145{l}% \mathbf{x}_{n}\left ( t\right ) = \mathbf{\tilde{x}}_{n}\left ( t\right ) + \mathbf{r}\left ( t\right ) , \\ \mathbf{x}_{n}^{\prime}\left ( t\right ) = \mathbf{\tilde{x}}_{n}^{\prime } \left ( t\right ) + \mathbf{r}\left ( t\right ) , \end{array } \right .
\label{tr1}%\ ] ] results in @xcite @xmath146 } .\label{tr2}%\ ] ] this result ( [ tr2 ] ) is valid for any _ quadratic _ model action @xmath147 the applicability of the parabolic approximation for @xmath148 is confirmed by the fact , that for the polaron ground - state energy , the results of the feynman approach @xcite are very close to the values obtained using other reliable methods @xcite .
thus , a self - induced polaronic potential , created by the polarization cloud around an electron , is rather well described by a parabolic potential whose parameters are determined by a variational method . for @xmath149 the
lowest known values of the bipolaron ground - state energy are provided by the path - integral variational method with parabolic potentials both in bulk @xcite and for confined systems @xcite for realistic values of @xmath1 .
the aforesaid approximation for the right - hand side of eq .
( [ f ] ) is a direct generalization of the all - coupling approach @xcite to a many - polaron system . for weak coupling ,
our variational method is at least of the same accuracy as the perturbation theory , which results from our approach at a special choice of the variational parameters . for strong coupling , an interplay of the electron - phonon interaction and the coulomb correlations within a confinement potential can lead to the assemblage of polarons in multi - polaron systems . as shown in refs .
@xcite for a system with @xmath150 , the presence of a confinement potential strongly favors the bipolaron formation .
our choice of the model variational system is reasonable because of this trend , apparently occurring in a many - polaron system with arbitrary @xmath0 for a finite confinement strength .
the correlation function @xmath151 corresponds to the model system in the absence of an electric field . for @xmath152 this function
is related to the imaginary - time correlation function @xmath153 described in the previous section:@xmath154 ) and the relation ( [ tr3 ] ) , one readily obtains@xmath155 } g\left ( \mathbf{q},i\left ( t - s\right ) |\left\ { n_{\sigma}\right\ } , \beta\right ) ds.\label{f1}%\ ] ] within the framework of the linear - response theory , the external electric field @xmath116 is a small perturbation , so that @xmath119 is a linear functional of @xmath156 \right| _ { t^{\prime}\leqslant t}.$ ] expanding the function @xmath157 } $ ] in the right - hand side of eq . ( [ f1 ] ) in powers of @xmath158 $ ] up to the first - order term
, we obtain the fourier component @xmath159 of the force due to the electron - phonon interaction which is proportional to @xmath160 as a result , the optical conductivity can be expressed in terms of the memory function @xmath161 ( cf .
@xcite),@xmath162 ^{2}+\left [ \operatorname{im}\chi\left ( \omega\right ) \right ] ^{2}% } , \label{kw}%\ ] ] where @xmath161 is given by @xmath163 .\label{hi}%\ ] ] it is worth noting that the optical conductivity ( [ kw ] ) differs from that for a translationally invariant polaron system both by the explicit form of @xmath161 and by the presence of the term @xmath164 in the denominator . for @xmath165
the optical conductivity tends to a @xmath166-like peak at @xmath167@xmath168 for a translationally invariant system @xmath169 , and this weak - coupling expression ( [ limit ] ) reproduces the `` central peak '' of the polaron optical conductivity @xcite .
in the zero - temperature limit , the memory function of eq .
( [ hi ] ) is derived in the analytical form of eq .
( [ mf ] ) in appendix b for 3d and 2d interacting polarons .
due to the confinement , the electron motion in a quantum dot is fully quantized .
hence , when a photon is absorbed , the electron recoil can be transferred only by discrete quanta . as a result ,
the optical conductivity spectrum of a system of interacting polarons in a quantum dot is a series of @xmath166-like peaks as distinct from the optical conductivity spectrum of a bulk polaron @xcite .
these peaks are related to the internal polaron excitations . because @xmath170 for all frequencies except for a discrete set of combinatorial frequencies ( [ oklm ] ) , the peaks in the optical conductivity ( [ kw ] ) are positioned at the frequencies which are given by the roots of the equation@xmath171 which are denoted as @xmath172 .
one of these roots is close to the variational parameter @xmath173 , which is the eigenfrequency of the motion of a polaron as a whole .
it satisfies the inequality @xmath174 because the polaron effective mass is larger than that of a bare electron .
@xmath173 is close to @xmath175 in the weak - coupling case and decreases with increasing @xmath176 hence , it tends to zero in the limit @xmath177 the peak in @xmath178 corresponding to this root can be considered as the _ zero - phonon _ line , which is an analogue of the `` central peak '' of the polaron optical conductivity @xcite .
the peaks of @xmath179 determined by the other roots of eq .
( [ peaks ] ) can be attributed to transitions into excited states of the many - polaron system .
the changes of the shell filling schemes , which occur when varying the confinement frequency , manifest themselves in the spectra of the optical conductivity . in fig.4 , optical conductivity spectra for @xmath180 polarons are presented for a quantum dot with the parameters of cdse : @xmath181 @xmath182 @xcite and with different values of the confinement energy @xmath183 . in this case , the spin - polarized ground state changes to the ground state satisfying hund s rule with increasing @xmath183 in the interval @xmath184 . in the inset to fig.4 , the first frequency moment of the optical conductivity @xmath185 as a function of @xmath183 shows a _
, at the value of the confinement energy corresponding to the change of the shell filling schemes from the spin - polarized ground state to the ground state obeying hund s rule .
this discontinuity should be observable in optical measurements . in fig .
5 , the first frequency moment ( [ moment ] ) is plotted as a function of the number of electrons for a cdse quantum dot with @xmath186 ( corresponding to @xmath187 @xmath94 ) .
the total spin of the system as a function of @xmath0 is shown in the inset . as a general trend
, @xmath188 decreases with increasing @xmath189 with kinks corresponding to the ground - state transitions from states obeying hund s rule with @xmath190 9 , and 18 , into spin - polarized states with @xmath191 10 , and 19 , respectively . in fig .
6 , optical conductivity spectra are plotted for several values of the confinement frequency for @xmath98 polarons in a quantum dot with @xmath192 @xmath193 .
these values of @xmath1 and @xmath194 are typical for the high - t@xmath195 superconducting cuprates of the ncco family @xcite . in the `` weak - confinement '' region ( @xmath196 and @xmath197 )
the zero - phonon peak is expressively dominant over the other peaks . when the confinement frequency parameter passes through the value @xmath198 , the so - called `` confinement - phonon resonance '' @xcite occurs . in this case , the peaks at @xmath199 have comparable oscillator strengths .
the position @xmath200 of the second peak is substantially shifted from the lo phonon frequency @xmath201 moreover , the intensities of the phonon - assisted transitions increase as compared to the `` weak - confinement '' case .
this resonance has a clear analogy with the magneto - phonon resonance ( see , e.g. , ref .
@xcite ) , as far as the energy levels of an electron in a parabolic confinement are similar to the landau levels of an electron in a magnetic field . with further increasing @xmath202
when @xmath203 the dominant part of the optical conductivity spectrum shifts to higher frequencies .
for instance , at @xmath204 the most intensive peak is that with @xmath205 .
the intensities of the peaks , beginning with the second peak , increase in comparison with the intensities of their `` weak - confinement '' analogs .
the positions of the zero - phonon line and the subsequent peaks are substantially shifted from the `` weak - confinement '' values towards higher frequencies .
these effects are a manifestation of the mixing of the zero - phonon state with different excited states of the many - polaron system .
a similar behavior of the optical absorption spectra at and above the magneto - phonon resonance is explained by the mixing of zero - phonon and one - phonon quantum states @xcite .
the shell structure for a system of interacting polarons in a quantum dot is clearly revealed when analyzing the addition energy and the first frequency moment of the optical conductivity in parallel . in figs . 7 and 8 , we show both the function @xmath206 and the addition energy @xmath104 for interacting polarons in different 3d quantum dots .
as seen from figs.7 and 8 for quantum dots of cdse and with @xmath207@xcite , respectively , distinct peaks appear in @xmath208 and @xmath104 at the `` magic numbers '' corresponding to closed - shell configurations at @xmath209 for the states obeying hund s rule in panels @xmath210 and to half - filled - shell configurations at @xmath211 for the spin - polarized states in panels @xmath212 of fig.8 . in the case when the shell filling scheme is the same for different @xmath0 ( see panels @xmath210 in figs.7,8 , where the filling obeys hund s rule ) , each of the peaks of @xmath213 corresponds to a peak of the addition energy . in the case when the shell filling scheme changes with varying @xmath0 ( panels @xmath212 in figs.7,8 )
, the function @xmath213 exhibits pronounced minima for @xmath0 corresponding to the change of the filling scheme from the states , obeying hund s rule , to the spin - polarized states .
it follows that measurements of the addition energy and the first frequency moment of the optical absorption as a function of the number of polarons in a quantum dot can reflect the difference between open - shell and closed - shell configurations .
in particular , the closed - shell configurations may be revealed through peaks in the function @xmath213 .
the filling patterns for a many - polaron system in a quantum dot can be determined from the analysis of the first moment of the optical absorption for different numbers of polarons .
the appearance of minima in the function @xmath213 will then indicate a transition from the states which are filled according to hund s rule to the spin - polarized states .
we presented a formalism for calculating the ground - state energy and the optical conductivity spectra of a system of @xmath0 interacting polarons in a parabolic confinement potential for arbitrary electron - phonon coupling strength .
the path integral treatment of the quantum statistics of indistinguishable particles @xcite allows us to find an upper bound @xcite to the ground - state energy of a finite number of polarons .
the parameters from the variational procedure are used as input for the calculation of the optical conductivity spectrum of the system .
two types of transitions were found for @xmath0 polarons confined in a parabolic potential , with the corresponding ground states characterized by different values of the total spin . in the weak - confinement regime ,
the polaron system is in the spin - polarized state .
when increasing the confinement frequency @xmath91 , the system goes into a state obeying hund s rule at a specific value of @xmath91 . for a strongly coupled system of interacting polarons ,
a third type of state appears , for which the total spin takes its minimal value .
the analysis is performed for both closed - shell and open - shell systems .
the calculations of the optical conductivity spectra for a finite number of polarons in a quantum dot are based on the memory - function approach .
the dependence of the optical conductivity spectra on the confinement parameter @xmath91 reveals a resonant behavior for @xmath214 .
transitions between states with different values of the total spin manifest themselves through discontinuous changes of the optical conductivity spectra and of the addition energy as a function of the number of electrons .
the first frequency moment of the optical conductivity as a function of the number of electrons clearly shows the transition between the spin - polarized ground state of interacting polarons in a quantum dot and the ground state obeying hund s rule , and it also can be used to discriminate between open - shell and closed - shell configurations .
optical measurements are therefore suggested as possible tools for examining the shell structure of a system of interacting polarons .
discussions with v. n. gladilin are gratefully acknowledged .
this work has been supported by the goa bof ua 2000 , iuap , fwo - v projects g.0306.00 , g.0274.01n , g.0435.03 , the wog wo.025.99n ( belgium ) and the european commission growth programme , nanomat project , contract no .
g5rd - ct-2001 - 00545 .
this work is performed in the framework of the fwo - v projects nos .
g.0435.03 , g.0306.00 , the w.o.g .
project wo.025.99n , and the goa bof ua 2000 project .
in this appendix we discuss the analytical calculation of the partition function @xmath215 [ eq .
( [ z0 ] ) ] for the model system of interacting polarons .
it can be expressed in terms of the partition function @xmath216 of the model system of interacting electrons and fictitious particles with the lagrangian @xmath217 [ eq . ( [ lm ] ) ] as follows:@xmath218 where @xmath219 is the partition function of fictitious particles,@xmath220 with@xmath221 and d=3(2 ) for 3d(2d ) systems .
the partition function @xmath222 is the path integral for both the electrons and the fictitious particles:@xmath223 } \label{zm}%\end{aligned}\ ] ] with the `` action '' functional @xmath224 = -\frac{1}{\hbar}\int_{0}^{\hbar\beta}l_{m}\left ( \mathbf{\dot{\bar{x}}},\mathbf{\dot{\bar{y}}};\mathbf{\bar{x}},\mathbf{\bar { y}}\right ) d\tau,\label{sm}%\ ] ] where the lagrangian is given by eq .
( [ lm ] ) .
let us consider an auxiliary `` ghost '' subsystem with the lagrangian @xmath225 with two frequencies @xmath226 and @xmath227 where @xmath226 is given by@xmath228 the partition function @xmath229 of this subsystem @xmath230 \right\ } , \label{zg1}%\ ] ] with the `` action '' functional@xmath231 = -\frac{1}{\hbar}\int\limits_{0}^{\hbar\beta}% l_{g}\left ( \mathbf{\dot{x}}_{g},\mathbf{x}_{g},\mathbf{\dot{y}}% _ { g},\mathbf{y}_{g}\right ) \,d\tau\label{sg}%\ ] ] is readily calculated : @xmath232 ^{d}}\frac{1}{\left [ 2\sinh\left ( \frac{\beta\hbar w_{f}}{2}\right ) \right ] ^{d}}.\label{zg2}%\ ] ] the product @xmath233 of the two partition functions @xmath229 and @xmath234 is a path integral in the state space of @xmath0 electrons , @xmath60 fictitious particles and two `` ghost '' particles with the coordinate vectors @xmath235 and @xmath236 the lagrangian @xmath237 of this system is a sum of @xmath217 and @xmath238@xmath239 the `` ghost '' subsystem is introduced because the center - of - mass coordinates in @xmath237 can be explicitly separated more easily than in @xmath217 .
this separation is realized by the linear transformation of coordinates,@xmath145{l}% \mathbf{x}_{j,\sigma}=\mathbf{x}_{j,\sigma}^{\prime}+\mathbf{x}-\mathbf{x}% _ { g},\\ \mathbf{y}_{j\sigma}=\mathbf{y}_{j\sigma}^{\prime}+\mathbf{y}-\mathbf{y}_{g } , \end{array } \right . \label{trans}%\ ] ] where @xmath72 and @xmath240 are the center - of - mass coordinate vectors of the electrons and of the fictitious particles , correspondingly:@xmath241 before the transformation ( [ trans ] ) , the independent variables are @xmath242 with the center - of - mass coordinates @xmath72 and @xmath240 determined by eq .
( [ xy ] ) .
after this transformation the independent variables can be considered to be @xmath243 where the coordinates
@xmath244 obey the equations@xmath245 a substitution of eq .
( [ xy ] ) into eq .
( [ lmt ] ) results in the following 3 terms:@xmath246 where @xmath247 and @xmath248 are lagrangians of non - interacting identical oscillators with the frequencies @xmath226 and @xmath227 respectively,@xmath249 , \label{lwa}\\ l_{w_{f}}\left ( \mathbf{\dot{\bar{y}}}^{\prime},\mathbf{\bar{y}}^{\prime } \right ) & = -\frac{m_{f}}{2}\sum_{j=1}^{n_{f}}\left [ \left ( \mathbf{\dot { y}}_{j,\sigma}^{\prime}\right ) ^{2}+w_{f}^{2}\left ( \mathbf{y}_{j,\sigma } ^{\prime}\right ) ^{2}\right ] .\label{lwb}%\end{aligned}\ ] ] the lagrangian @xmath250 describes the combined motion of the centers - of - mass of the electrons and of the fictitious particles,@xmath251 with@xmath252 it is reduced to a diagonal quadratic form in the coordinates and the velocities by a standard transformation for two interacting oscillators : @xmath253 with the coefficients @xmath254 ^{1/2},\quad a_{2}=\left [ \frac{1-\chi}{2}\right ] ^{1/2},\label{a12}\\ \chi & \equiv\frac{\tilde{\omega}^{2}-\tilde{\omega}_{f}^{2}}{\left [ \left ( \tilde{\omega}^{2}-\tilde{\omega}_{f}^{2}\right ) ^{2}+4\gamma^{2}\right ] ^{1/2}},\quad\gamma\equiv k\sqrt{\frac{nn_{f}}{m_{b}m_{f}}}.\label{c}%\end{aligned}\ ] ] the eigenfrequencies of the center - of - mass subsystem are then given by the expression @xmath145{c}% \omega_{1}=\sqrt{\frac{1}{2}\left [ \tilde{\omega}^{2}+\tilde{\omega}_{f}% ^{2}+\sqrt{\left ( \tilde{\omega}^{2}-\tilde{\omega}_{f}^{2}\right ) ^{2}+4\gamma^{2}}\right ] } , \\
\omega_{2}=\sqrt{\frac{1}{2}\left [ \tilde{\omega}^{2}+\tilde{\omega}_{f}% ^{2}-\sqrt{\left ( \tilde{\omega}^{2}-\tilde{\omega}_{f}^{2}\right ) ^{2}+4\gamma^{2}}\right ] } .
\end{array } \right .
\label{o12}%\ ] ] as a result , four independent frequencies @xmath255 @xmath256 @xmath226 and @xmath257 appear in the problem .
three of them ( @xmath255 @xmath256
@xmath226 ) are the eigenfrequencies of the model system .
@xmath258 is the frequency of the relative motion of the center of mass of the electrons with respect to the center of mass of the fictitious particles ; @xmath173 is the frequency related to the center of mass of the model system as a whole ; @xmath226 is the frequency of the relative motion of the electrons with respect to their center of mass .
the parameter @xmath257 is an analog of the second variational parameter @xmath226 of the one - polaron feynman model .
further , the lagrangian ( [ lc ] ) takes the form @xmath259 leading to the partition function corresponding to the combined motion of the centers - of - mass of the electrons and of the fictitious particles @xmath260 ^{d}}\frac{1}{\left [ 2\sinh\left ( \frac{\beta\hbar\omega_{2}}% { 2}\right ) \right ] ^{d}}.\label{zc}%\ ] ] taking into account eqs .
( [ zg2 ] ) and ( [ zc ] ) , we obtain finally the partition function of the model system for interacting polarons @xmath261 ^{d}% \mathbb{\tilde{z}}_{f}\left ( \left\ { n_{\sigma}\right\ } , w,\beta\right ) .\label{z02}%\ ] ] here@xmath262 is the partition function of @xmath263 non - interacting fermions in a parabolic confinement potential with the frequency @xmath264 the analytical expressions for the partition function of @xmath10 spin - polarized fermions @xmath265 were derived in ref . @xcite .
the two - point correlation function ( [ tpcf ] ) is represented as the following path integral:@xmath266 } \rho_{\mathbf{q}}\left ( \tau\right ) \rho_{-\mathbf{q}}\left ( 0\right ) .\label{g1}%\end{aligned}\ ] ] we observe that @xmath267 can be rewritten as an average within the model `` action '' @xmath268 $ ] of interacting electrons and fictitious particles:@xmath269 } \nonumber\\ & \times\rho_{\mathbf{q}}\left ( \tau\right ) \rho_{-\mathbf{q}}\left ( 0\right ) .\label{g2}%\end{aligned}\ ] ] indeed , one readily derives that the elimination of the fictitious particles in ( [ g2 ] ) leads to ( [ g1 ] ) .
the representation ( [ g2 ] ) allows one to calculate the correlation function @xmath270 in a much simpler way than through eq .
( [ g1 ] ) , using the separation of the coordinates of the centers of mass of the electrons and of the fictitious particles .
this separation is performed for the two - point correlation function ( [ g2 ] ) by the same method as it has been done for the partition function ( [ zm ] ) . as a result ,
one obtains@xmath271 \right\rangle _ { s_{c}}}{\left\langle \exp\left [ i\mathbf{q\cdot}\left ( \mathbf{x}% _ { g}\left ( \tau\right ) -\mathbf{x}_{g}\left ( \sigma\right ) \right ) \right ] \right\rangle _ { s_{g}}},\label{fact2}%\ ] ] where @xmath272 is the time - dependent correlation function of @xmath0 non - interacting electrons in a parabolic confinement potential with the frequency @xmath226 , @xmath273 the action functional @xmath274 $ ] is related to the lagrangian @xmath275 [ eq . ( [ lwa])]@xmath276 = \frac{1}{\hbar}\int \limits_{0}^{\hbar\beta}l_{w}\left ( \mathbf{\dot{\bar{x}}},\mathbf{\bar{x}% } \right ) \,d\tau.\label{sw}%\ ] ] the averages in ( [ fact2 ] ) are calculated using feynman s method of generating functions @xcite , @xmath277 \right\rangle _ { s_{c } } & = \exp\left\ { -\frac{\hbar q^{2}}{nm_{b}}\left [ \sum_{i=1}% ^{2}a_{i}^{2}\frac{\sinh\left ( \frac{\omega_{i}\left| \tau-\sigma\right| } { 2}\right ) \sinh\left ( \frac{\omega_{i}\left ( \hbar\beta-\left| \tau-\sigma\right| \right ) } { 2}\right ) } { \omega_{i}\sinh\left ( \frac{\beta\hbar\omega_{i}}{2}\right ) } \right ] \right\ } , \\
\left\langle \exp\left [ i\mathbf{q\cdot}\left ( \mathbf{x}_{g}\left ( \tau\right ) -\mathbf{x}_{g}\left ( \sigma\right ) \right ) \right ] \right\rangle _ { s_{g } } & = \exp\left [ -\frac{\hbar q^{2}}{nm_{b}}% \frac{\sinh\left ( \frac{w\left| \tau-\sigma\right| } { 2}\right ) \sinh\left ( \frac{w\left ( \hbar\beta-\left| \tau-\sigma\right| \right ) } { 2}\right ) } { w\sinh\left ( \frac{\beta\hbar w}{2}\right ) } \right ] .\end{aligned}\ ] ] the two - point correlation function @xmath278 is derived using the generating - function technique for identical particles @xcite . after the path integration , the following expression is obtained:@xmath279 \nonumber\\ & \times\left [ f_{1}\left ( n,\sigma|\left\ { n_{\sigma}\right\ } , \beta\right ) -f_{2}\left ( n,\sigma;n^{\prime},\sigma|\left\ { n_{\sigma } \right\ } , \beta\right ) \right ] , \label{g4}%\end{aligned}\ ] ] where @xmath280 is the one - electron matrix element,@xmath281 for a 3d quantum dot , @xmath282 is the eigenfunction of a 3d oscillator with the frequency @xmath226 [ see , e.g. , ref .
the index @xmath283 denotes the set @xmath284 where @xmath285 is the number of the energy level @xmath286 , @xmath287 is the quantum number of the orbital angular momentum and @xmath288 is the quantum number of the @xmath289 projection of the orbital angular momentum . similarly , for a 2d quantum dot , @xmath290 is the eigenfunction of a 2d oscillator with the frequency @xmath264 the one - electron distribution function @xmath291 is the average number of electrons with the spin projection @xmath292 at the @xmath285-th energy level , while the two - electron distribution function @xmath293 is the average product of the numbers of electrons with the spin projections @xmath292 and @xmath294 at the levels @xmath285 and @xmath295 these functions are expressed through the following integrals ( cf . ref .
@xcite):@xmath296{c}% \frac{1}{2\pi\mathbb{z}_{f}\left ( n_{\sigma},w,\beta\right ) } \int \limits_{-\pi}^{\pi}f\left ( \varepsilon_{n},\theta\right ) f\left ( \varepsilon_{n\mathbf{^{\prime}}},\theta\right ) \phi\left ( \theta , \beta , n_{\sigma}\right ) d\theta,\;\text{if } \sigma^{\prime}=\sigma;\\ f_{1}\left ( n,\sigma|n_{\sigma},\beta\right ) f_{1}\left ( n,\sigma^{\prime } |n_{\sigma^{\prime}},\beta\right ) , \;\text{if } \sigma^{\prime}\neq\sigma \end{array } \right .
\label{num2b}%\end{aligned}\ ] ] with the notations@xmath297 , \label{fin}%\]]@xmath298 the function @xmath299 formally coincides with the fermi - dirac distribution function of the energy @xmath300 with the `` chemical potential '' @xmath301 from here on we consider the zero - temperature limit , for which the integrals ( [ num2a ] ) and ( [ num2b ] ) can be calculated analytically .
the result for the one - electron distribution function is@xmath302{cc}% 1 , & n < l_{\sigma};\\ 0 , & n > l_{\sigma};\\ \frac{n_{\sigma}-n_{l_{\sigma}}}{g_{l_{\sigma } } } , & n = l_{\sigma}. \end{array } \right . \label{p1}%\ ] ] according to ( [ p1 ] ) , @xmath303 is the number of the lowest open shell , and @xmath304{ccc}% \frac{1}{2}\left ( n+1\right ) \left ( n+2\right ) & & \left ( 3d\right ) , \\
n+1 & & \left ( 2d\right ) . \end{array } \right.\ ] ] is the degeneracy of the @xmath285-th shell .
@xmath305 is the number of electrons in all the closed shells with the spin projection @xmath306@xmath307{ccc}% \frac{1}{6}l_{\sigma}\left ( l_{\sigma}+1\right ) \left ( l_{\sigma}+2\right ) & & \left ( 3d\right ) , \\
\frac{1}{2}l_{\sigma}\left ( l_{\sigma}+1\right ) & & \left ( 2d\right ) .
\end{array } \right . \label{nls}%\ ] ] the two - electron distribution function @xmath308 at @xmath309 takes the form@xmath310{cc}% \left .
f_{1}\left ( n,\sigma|\beta , n_{\sigma}\right ) \right| _ { \beta \rightarrow\infty}\left .
f_{1}\left ( n^{\prime},\sigma^{\prime}% n^{\prime}\;\mathrm{or}\;\sigma\neq\sigma^{\prime}\\ 1 , & \sigma=\sigma^{\prime}\;\mathrm{and}\;n = n^{\prime}<l_{\sigma};\\ 0 , & \sigma=\sigma^{\prime}\;\mathrm{and}\;n = n^{\prime}>l_{\sigma};\\ \frac{n - n_{l_{\sigma}}}{g_{l_{\sigma}}}\frac{n - n_{l_{\sigma}}-1}{g_{l_{\sigma } } -1 } , & \sigma=\sigma^{\prime}\;\mathrm{and}\;n = n^{\prime}=l_{\sigma}. \end{array } \right .
\label{p2}%\end{aligned}\ ] ] finally , using the two - point correlation function eq .
( [ me ] ) , the one - electron ( [ p1 ] ) and the two - electron ( [ p2 ] ) distribution functions , the memory function of eq .
( [ hi ] ) can be represented in the unified form for 3d and 2d interacting polarons : @xmath311 \right| _ { \beta\rightarrow\infty}\right .
\nonumber\\ & \times\left ( \begin{array } [ c]{c}% \frac{1}{\omega-\omega_{\mathrm{lo}}-\left [ p_{1}\omega_{1}+p_{2}\omega _ { 2}+\left ( p_{3}-m+n\right ) w\right ] + \mathrm{i}\varepsilon}% -\frac{1}{\omega+\omega_{\mathrm{lo}}+p_{1}\omega_{1}+p_{2}\omega_{2}+\left ( p_{3}-m+n\right ) w+\mathrm{i}\varepsilon}\\ + \mathcal{p}\left ( \frac{2}{\omega_{\mathrm{lo}}+p_{1}\omega_{1}+p_{2}% \omega_{2}+\left ( p_{3}-m+n\right ) w}\right ) \end{array } \right ) \nonumber\\ & \times\sum\limits_{l=0}^{m}\sum\limits_{k = n - m+l}^{n}\frac{\left ( -1\right ) ^{n - m+l+k}\gamma\left ( p_{1}+p_{2}+p_{3}+k+l+\frac{3}{2}\right ) } % { k!l!}\left ( \frac{1}{wa}\right ) ^{l+k}\nonumber\\ & \left .
\times\binom{n+d-1}{n - k}\binom{2k}{k - l - n+m}\right ] \nonumber\\ & + \left [ \left ( \begin{array } [ c]{c}% \frac{1}{\omega-\omega_{\mathrm{lo}}-\left ( p_{1}\omega_{1}+p_{2}\omega _ { 2}+p_{3}w\right ) + \mathrm{i}\varepsilon}-\frac{1}{\omega+\omega _ { \mathrm{lo}}+p_{1}\omega_{1}+p_{2}\omega_{2}+p_{3}w+\mathrm{i}\varepsilon}\\ + \mathcal{p}\left ( \frac{2}{\omega_{\mathrm{lo}}+p_{1}\omega_{1}+p_{2}% \omega_{2}+p_{3}w}\right ) \end{array } \right ) \right .
\nonumber\\ & \times\sum\limits_{m=0}^{\infty}\sum\limits_{n=0}^{\infty}\sum \limits_{\sigma,\sigma^{\prime}}\left .
f_{2}\left ( n,\sigma;m,\sigma ^{\prime}|\left\ { n_{\sigma}\right\ } , \beta\right ) \right| _ { \beta \rightarrow\infty}\nonumber\\ & \times\sum\limits_{k=0}^{n}\sum\limits_{l=0}^{m}\frac{\left ( -1\right ) ^{k+l}\gamma\left ( p_{1}+p_{2}+p_{3}+k+l+\frac{3}{2}\right ) } { k!l!}\left ( \frac{1}{wa}\right ) ^{k+l}\nonumber\\ & \left .
\times\binom{n+d-1}{n - k}\binom{m+d-1}{m - l}\right ] \right\ } , \label{mf}%\end{aligned}\ ] ] where @xmath312 is the dimensionality of the space , @xmath313 denotes the principal value , @xmath314 is defined as @xmath315 /n$ ] , @xmath316 and @xmath226 are the eigenfrequencies of the model system , @xmath317 and @xmath318 are the coefficients of the canonical transformation which diagonalizes the model lagrangian ( [ lm ] ) derived in appendix a. the typical spectra of the real and imaginary parts of the memory function @xmath319 are plotted in figs .
9(_a _ ) and 9(_b _ ) , respectively .
according to eq .
( [ mf ] ) , the poles of @xmath320 and the @xmath166-like peaks of @xmath321 $ ] are positioned at the combinatorial frequencies @xmath322 , which are linear combinations of the lo - phonon frequency and three eigenfrequencies@xmath323 with integer coefficients
@xmath324 @xmath325 @xmath326 each combinatorial frequency @xmath322 corresponds to a phonon - assisted transition to an excited state of the model system .
the roots of the equation ( [ peaks ] ) , which provide the peaks in the optical conductivity , @xmath172 , are indicated in fig .
_ by the vertical arrows . for the chosen parameters ,
the peak at @xmath327 is the zero - phonon line . fig .
9__a _ _ also reveals peaks of @xmath328 with frequencies in between two neighboring discrete values of @xmath329 , e.g. , at @xmath200 . following the physical interpretation of the memory function in refs .
@xcite , these peaks can be related to transitions into excited states of the many - polaron system .
3 . the total spin ( _ a _ ) and the addition energy ( _ b _ ) of a system of interacting polarons in a 2d parabolic gaas quantum dot as a function of the number of electrons for @xmath93@xmath94 .
inset : the experimentally observed addition energy vs number of electrons in a cylindrical gaas quantum dot for two values of the diameter @xmath333 m and @xmath334 m @xcite .
optical conductivity spectra of @xmath180 interacting polarons in cdse quantum dots with @xmath335 , @xmath182 for different confinement energies close to the transition from a spin - polarized ground state to a ground state obeying hund@xmath336s rule .
_ inset _ : the first frequency moment @xmath188 of the optical conductivity as a function of the confinement energy .
5 . the first frequency moment @xmath188of the optical conductivity as a function of the number of electrons for systems of interacting polarons in cdse quantum dots with @xmath335 , @xmath182 and @xmath337 ( @xmath338 ) .
open squares denote the spin - polarized ground state ; full dots denote the ground state , obeying hund s rule ; open triangles denote the ground state of the third type , with more than one partly filled shells , which is not totally spin - polarized .
_ inset _ : the total spin of the system of interacting polarons as a function of @xmath114 fig .
optical conductivity spectra of @xmath98 interacting polarons in a quantum dot with @xmath339 , @xmath193 for several values of the confinement frequency from @xmath196 to @xmath340 the spectrum for @xmath341 corresponds to the confinement - phonon resonance .
the function @xmath213 and the addition energy @xmath104 for systems of interacting polarons in cdse quantum dots with @xmath335 , @xmath182 for @xmath331 ( panels @xmath210 ) and for @xmath342 ( panels @xmath212 ) .
open squares denote the spin - polarized ground state ; full dots denote the ground state , obeying hund s rule ; open triangles denote the ground state of the third type , with more than one partly filled shells , which is not totally spin - polarized .
the function @xmath213 and the addition energy @xmath104 for systems of interacting polarons in quantum dots with @xmath343 @xmath344 and @xmath345 ( panels @xmath210 ) and with @xmath343 @xmath346 and @xmath347 ( panels @xmath212 ) .
real ( _ a _ ) and imaginary ( _ b _ ) parts of the memory function @xmath348 for a system of interacting polarons in a quantum dot for @xmath349 @xmath192 @xmath350 and @xmath351 .
the dashed line in panel _ a _ represents @xmath352 the vertical arrows in panel _ a _ indicate the roots of eq .
( [ peaks ] ) . the height of peaks in panel _
b _ represents the relative intensity of the @xmath166-like peaks of @xmath353 $ ] . | the ground - state energy , the addition energies and the optical absorption spectra are derived for interacting polarons in parabolic quantum dots in three and two dimensions .
a path integral formalism for identical particles is used in order to take into account the fermion statistics .
the approach is applied to both closed - shell and open - shell systems of interacting polarons . using a generalization of the jensen - feynman variational principle ,
the ground - state energy of a confined @xmath0-polaron system is analyzed as a function of @xmath0 and of the electron - phonon coupling constant @xmath1 .
as distinct from the few - electron systems without the electron - phonon interaction , _ three _ types of spin polarization are possible for the ground state of the few - polaron systems : ( i ) a spin - polarized state , ( ii ) a state where the spin is determined by hund s rule , ( iii ) a state with the minimal possible spin .
a transition from a state fulfilling hund s rule , to a spin - polarized state occurs when decreasing the electron density . in the strong - coupling limit
, the system of interacting polarons turns into a state with the minimal possible spin .
these transitions should be experimentally observable in the optical absorption spectra of quantum dots . |
ballistic transport through quantum billiards has been extensively studied in recent years due to its relevance for quantum chaos and the possibility of physical applications .
realizations of ballistic billiards include structured two - dimensional electron gases in semiconductor heterostructures@xcite and , exploiting the analogy between quantum and wave mechanics , microwave cavities@xcite .
various experiments have been designed to test theoretical ideas on conductance fluctuations @xcite , weak localization @xcite and the signatures of classical integrability .
the main theoretical tool for making the connection between the quantum and classical properties is the semiclassical expansion @xcite .
this intuitive and powerful approach has been tested numerically for the transport through circular cavities @xcite .
in particular , the identification of the most relevant trajectories for transmission and reflection has been accurately demonstrated ( analogously to the relationship between the density of states and periodic orbits of closed systems @xcite ) .
moreover , the semiclassical approach has been extended by the inclusion of diffraction effects at the entrance and exit of the cavities @xcite . in this work
we further extend the applicability of semiclassical methods in open systems to treat the case where tunneling takes place .
the modification of the trace formula in a closed system by the inclusion of a potential step has recently been addressed for a circular billiard @xcite in the context of ray splitting . there
the possibility of electrons entering in the region of higher potential has to be taken into account .
our work shows that a tunnel barrier within a cavity can very simply be incorporated in a semiclassical description and changes the relative importance of different classical trajectories in a non - trivial manner .
our interest in tunneling inside a cavity stems from a fundamental point of view as well as from the fact that experiments with a high potential barrier within a quantum dot ( pacman ) have already been performed @xcite .
the inclusion of a dielectric slab within a microwave cavity would also result in a barrier .
starting from the well - studied circular billiard @xcite , we introduce a thin barrier placed symmetrically between the two leads and extending from the edge to the center of the circle as shown by the dashed line in fig .
[ fig : pacgeom ] .
the barrier height @xmath0 is variable and allows to interpolate between a circle and the billiard studied in ref .
@xcite . of special interest will be the regime where the electron energy is of the order of the barrier height so that tunneling becomes relevant . in the following ,
we therefore refer to this billiard as the circular tunneling billiard .
we will consider phase coherent and ballistic transport through the cavity which is attached to two hard - wall leads of width @xmath1 .
for a fixed energy @xmath2 of the incident particles there exists only a finite number of transverse modes @xmath3 , given by the largest integer smaller than @xmath4 , which contribute to the transport . here , @xmath5 is the energy of the lowest transverse mode in the leads and @xmath6 is the mass of the particles .
= 0.3 within the landauer formalism@xcite the two - probe conductance @xmath7 through the structure is just proportional to the total transmission coefficient @xmath8 at the fermi energy @xmath9 @xmath10 the transmission amplitude connecting the incoming mode @xmath11 to the outgoing mode @xmath12 is given by the projection@xcite @xmath13 of the retarded green function @xmath14 of the structure onto the transverse modes @xmath15 according to @xmath16 where we have discarded an unimportant phase factor .
the longitudinal wave vector is given by @xmath17 the expression ( [ eq : projgreen ] ) has to be evaluated with @xmath18 in the incoming and @xmath19 in the outgoing lead .
for the amplitude of reflection one finds the corresponding expression @xmath20 with @xmath18 and @xmath19 in the incoming lead . the transmission and reflection amplitudes can be obtained numerically by means of the recursive green function method@xcite .
this method uses a discretized version of the cavity and calculates the green function by starting from the exact green function in one of the leads and successively building up the solution by means of the dyson equation .
alternatively , a semiclassical approach to transport can be developed from the semiclassical path - integral form of the green function leading to a transmission amplitude @xcite @xmath21 given as the sum over classical trajectories @xmath22 between the entrance and exit cross sections with incoming and outgoing angles @xmath23 and @xmath24 such that @xmath25 and @xmath26 ( @xmath27 , @xmath28 ) .
the reduced action @xmath29 is the legendre transform of the action integral @xmath30 , @xmath31 where the starting and end points of the trajectory , @xmath32 and @xmath33 respectively , are determined by the angle quantization . for billiards @xmath34 , where @xmath35 is the length of the trajectory .
the amplitude is @xmath36 and @xmath37 is the maslov index .
a similar expression holds for the reflection amplitude , with the difference that now the trajectories start and end in the same lead .
direct comparison between the semiclassical amplitudes ( or the conductance ) and the exact counterparts is rather difficult since the expansion ( [ eq : scltmn ] ) includes an infinite number of terms associated with an exponentially large number of contributing classical trajectories .
however , as in the case of the trace formula , the validity of the semiclassical approach can be established by identifying the fourier components of @xmath38 with classical trajectories .
this has been done in ref .
@xcite and we verify it in our system since it gives the starting point of our analysis . in sec .
[ sec : identification ] we present the numerical calculation of the reflection amplitude and its interpretation in terms of classical trajectories without and with a high barrier . in sec .
[ sec : paths ] we consider the influence of a tunnel barrier on the different trajectories contributing to the reflection and transmission amplitudes and find a nonmonotonic behaviour .
this behaviour is modeled in sec .
[ sec : sclctb ] within a modification of the semiclassical transmission amplitudes that includes tunneling in a very simple way . in sec .
[ sec : trtot ] we consider a similar analysis for the transmission and reflection coefficients , and show that the analysis becomes considerably more involved since we now have to deal with pairs of trajectories . in the final section we present our conclusions and discuss the extension of our work to the case with magnetic field .
in fig . [ fig : r ] we present the ( exact ) total reflection coefficient together with the contribution from the lowest mode @xmath39 .
obviously , for @xmath40 the two quantities coincide since there @xmath41 . in order to identify the classical trajectories we carried out a discrete fourier transformation of 600 values of the complex reflection amplitude @xmath42 calculated over a momentum interval ranging from 1 to 7 @xmath43 . the ratio between the radius @xmath44 of the circle and the width @xmath1 of the leads was in all calculations taken to be @xmath45 .
= 0.45 = 0.45 the semiclassical form ( [ eq : scltmn ] ) of the transmission amplitude implies that its fourier transform with respect to the momentum @xmath46 should exhibit peaks at lengths @xmath35 corresponding to the contributing classical trajectories .
similarly , the reflection amplitudes can be interpreted in terms of classical paths . for a detailed analysis
we now consider the power spectrum of @xmath42 with respect to length .
this quantity , which is shown in fig .
[ fig : r11]a , is given by the squared modulus of the fourier transform of @xmath42 and will be denoted as @xmath47 in the following . in complete agreement with ref .
@xcite , we can establish a correspondence between peaks of @xmath48 and the classical trajectories including their repetitions .
the first peak is not a classical trajectory contributing to reflection , but corresponds to diffraction off the lead mouths @xcite .
this effect can be interpreted in terms of a trajectory that gets reflected back at the right lead ( ghost path ) . for larger lengths
@xmath35 , we can identify a triangular path , a five - star path , a seven - star path , and so on . in agreement with the semiclassical quantization of the initial and final angle , the star - shaped paths are the most important ones for small mode numbers which favor the forward direction .
the resolution of the length spectrum is limited by the width of the momentum interval used for the fourier transformation . while this can easily be controlled ,
there are also intrinsic effects restricting the resolution . within a semiclassical picture , at given energy and mode numbers the angle quantization selects the appropriate paths .
depending on the width of the leads , the transverse position of the starting and end points in the leads are variable and therefore a given type of trajectories exists in a certain momentum interval . in the fourier transformation these trajectories will contribute with different weights since the corresponding action will depend on the momentum , thus yielding a finite resolution .
already the fact that a type of trajectory effectively contributes only in a finite momentum interval may limit the resolution more strongly than the finite interval imposed by the numerics . in this respect
it is also important to note that quantum mechanical diffraction effects at the lead mouths @xcite influence the effective momentum interval and may be relevant for the resolution . placing a sufficiently high barrier into the cavity yields a reflection coefficient ( not shown here ) uncorrelated to that presented in fig .
[ fig : r ] . on the other hand
, we expect that individual trajectories would be greatly altered by the barrier and thus we analyze the length spectrum @xmath49 as a function of the barrier height .
for very high barriers one can see in fig .
[ fig : r11]b that new length scales have appeared rendering the identification more involved as compared to the case of vanishing barrier .
some features are easily explained like the appearance of a peak at length @xmath50 , while in the absence of a barrier the minimum length is @xmath51 . in the presence of a high barrier , the direct path may get reflected at the barrier thus leading to a peak at half of the previous minimal length .
the peak at @xmath51 now consists of two contributions , namely the direct path which is reflected at the right lead and twice the direct path reflected at the barrier which involves a reflection at the left lead mouth .
the hierarchy continues with a smaller peak at about @xmath52 which stems from three repetitions of the direct path reflected at the barrier . in principle , it is not clear that an analysis in terms of classical paths is applicable for arbitrary barrier height since tunneling necessarily implies non - classical trajectories .
however , we will show that such an analysis is still possible and helpful towards the understanding of the transport problem .
for instance , comparison between figs .
[ fig : r11]a and b shows a large suppression of the harmonic coming from the triangular path , while the five - star path component is much less affected .
simple arguments given in the next section explain this difference in behaviour .
for a semiclassical analysis of the energy - dependent transmission and reflection spectra , we first have to discuss how the classical paths are modified by the barrier .
the cases where well - defined classical paths exist , are those of the circular billiard ( @xmath53 ) and the circular billiard with a very high barrier ( @xmath54 ) . while postponing a more detailed analysis to sec .
[ sec : sclctb ] , we expect that at intermediate barrier heights , the transmission and reflection amplitudes for the billiard should be given by both classes of trajectories properly weighted according to the transmission and reflection coefficients of the barrier .
the length spectrum of the reflection amplitude , referred to as reflection spectrum in the following , of the circular billiard shown in fig .
[ fig : r11]a displays distinct peaks which can be associated with a triangle , a five- , and a seven - star . in the following discussion
we will focus on these three trajectories .
we start with the triangular path as the simplest case . in the absence of a barrier
, the trajectory just follows the triangle as shown in fig . [
fig : trianglepac]a . as the barrier height @xmath0
is increased the transmission probability through the barrier decreases and for high barriers the original triangle is no longer a possible path .
accordingly , the peak in the reflection spectrum corresponding to the triangle will decrease in amplitude with increasing barrier height .
for sufficiently large @xmath0 , the possibility of reflection at the barrier has to be taken into account . as a consequence of this reflection the path will no longer continue on the original triangle shown as dotted line in fig .
[ fig : trianglepac ] but follow , at least for sufficiently thin barriers , the dashed - dotted line obtained as mirror image with respect to a vertical line through the barrier .
it is important that this path has the same length as the original path .
however , now the end point no longer lies in the entrance lead but in the opposite lead and thus the path reflected at the barrier will contribute to the transmission through the billiard .
correspondingly , the triangle will become more important in the transmission spectrum as the barrier height is increased .
this qualitative discussion is confirmed by the numerical results for the reflection and transmission spectra , @xmath48 and @xmath55 , shown in fig .
[ fig:3peakh ] as diamonds and triangles , respectively .
= 0.45 = 0.45 while the triangular path can not appear in the reflection spectrum for very high barriers , it is interesting to note that there is a peak at a length corresponding to two repetitions of the triangular path .
this can readily be verified by comparing figs .
[ fig : r11]a and b. while one reflection at the barrier changes the exit lead from the left to the right , an additional reflection restores the left lead as exit lead .
therefore , the peak can be identified with two repetitions of the path shown in fig .
[ fig : trianglepac]b including a reflection at the right lead due to diffraction at the lead mouth .
the behaviour of the five - star trajectory is more complex due to the fact that it crosses the barrier twice . making use of the same geometrical arguments as for the triangle
, we may distinguish four different classes of trajectories shown in fig .
[ fig : fivepac ] which correspond to two transmissions at the barrier ( a ) , one reflection and one transmission ( b and c ) , and two reflections ( d ) .
in fact , since the trajectory may either start into the upper or lower = 0.45 half of the billiard , these four classes correspond to eight different trajectories which can be obtained by reading the diagrams in the forward and backward direction . like for the case of the triangle , a reflection at the barrier changes the sense of rotation in the circle thereby changing the exit lead . as can be seen from fig .
[ fig : fivepac ] , the paths with an even number of reflections ( a and d ) contribute to the reflected paths through the billiard while the paths with an odd number of reflections ( b and c ) contribute to the transmission through the billiard .
let us first consider the trajectories contributing to the reflection .
the trajectory ( a ) will only contribute for small barrier heights because it has to be transmitted through the barrier twice .
on the other hand , the trajectory ( d ) will appear only for rather high barriers since it requires two reflections at the barrier . as a consequence
, we expect that the peak in the reflection spectrum corresponding to the five - star trajectory will exhibit a minimum at intermediate barrier heights where none of the trajectories ( a ) and ( d ) contribute significantly .
the other trajectories ( b and c ) appear in the transmission spectrum at intermediate barrier heights because they have to be reflected as well as transmitted once at the barrier .
we conclude from the discussion of the five - star trajectory that in general the dependence on barrier height of the peak heights in the transmission and reflection spectra should be nonmonotonic .
this is confirmed by the numerical data shown as diamonds ( @xmath48 ) and triangles ( @xmath55 ) in fig .
[ fig:5peakh ] .
the lines shown there are results of a semiclassical analysis which will be discussed in detail in the next section . for this nonmonotonic behaviour
it is crucial that = 0.45 the reflection of a classical path at the barrier does not change the length of the trajectory .
only then a change in the barrier height will not affect the position of the peak and contributions of different paths have to be coherently superposed .
the behaviour of the peak heights becomes more complex as the trajectories encounter the barrier more often .
this will become clear from our final example , the seven - star trajectory . in this case
, the trajectory encounters the barrier three times giving rise to eight classes of trajectories .
we may classify these trajectories according to their behaviour at the barrier by assigning a t or an r for each transmission or reflection , respectively .
then , the paths contributing to the reflection spectrum are those containing an even number of r , namely ttt , rrt , rtr , and trr . while the first trajectory will contribute for very small barriers , the other trajectories appear only for sufficiently high barriers . at very high barriers none of these paths
is allowed .
accordingly , the peak in the reflection spectrum associated with the seven - star will initially decrease with increasing barrier height , exhibit a minimum followed by a maximum and then go to zero as the barrier height becomes very large .
this behaviour can readily be verified by comparison with the numerical results for @xmath48 shown as diamonds in fig .
[ fig:7peakh ] .
the behaviour of the transmission spectrum @xmath55 shown in this figure as triangles can be understood along the same line of reasoning .
the trajectories contributing to this spectrum are those with an odd number of reflections at the barrier , i.e. rtt , trt , ttr , and rrr .
since there is at least one reflection at the barrier , seven - star trajectories may contribute to the transmission spectrum only for finite barrier heights .
with increasing height there will be a maximum followed by a minimum and at = 0.45 = 0.45 very high barriers the rrr - trajectory will contribute .
so far , we have concentrated on trajectories which in the absence of a barrier contribute to the reflection .
the behaviour of classical paths connecting two different leads , i.e. paths contributing to the transmission , as a function of the barrier height is slightly more complex . in this case
, one has to distinguish the paths first going into the upper and lower half of the billiard . as an example , we consider the paths shown in fig . [ fig : eight ] which represent one half of an eight - star . from this figure
it becomes clear that depending on the incident angle of the path there will be an even or odd number of contacts with the barrier .
accordingly , the path shown in fig .
[ fig : eight]a will behave very much like the five - star which also encounters the barrier twice . on the other hand , the path shown in fig .
[ fig : eight]b has just one contact with the barrier and its dependence on the barrier height resembles that of the triangle . as a consequence , while both trajectories will contribute to the transmission for low barriers , only one of them ( fig .
[ fig : eight]a ) will do so at high barriers .
the other path will turn into a reflected path ( right part of fig .
[ fig : eight]b ) instead .
this line of reasoning generally applies to trajectories which are transmitted in the absence of a barrier . even though the analysis now becomes more complicated
, these paths also lead to nonmonotonic reflection and transmission spectra as a function of barrier height . again
an increasing number of contacts with the barrier will lead to an increasing number of extrema .
an example will be shown in sec .
[ sec : trtot ] where the spectra of the total transmission are discussed ( fig .
[ fig : vat ] ) .
the previous qualitative discussion allowed us to understand the effect of a tunnel barrier by simple consideration of classical trajectories . in the next section
we will show that a quantitative agreement with the exact calculations can be obtained within a semiclassical approach where the possibility of transmission or reflection at the barrier is incorporated .
we now want to include tunneling into the semiclassical picture while remaining rather close to the expression for the semiclassical green function ( [ eq : scltmn ] ) in terms of classical paths . to this end
, we multiply the contribution of the classical paths by amplitudes @xmath56 or @xmath57 accounting for each transmission or reflection of the classical path at the barrier . the treatment of a barrier of finite length in the circular tunneling billiard represents a rather complicated two - dimensional problem .
however , we may approximately describe the behaviour of an electron at the barrier as plane wave encountering a barrier of infinite length .
then the problem may be separated into the directions perpendicular and parallel to the barrier and the only parameter describing the scattering geometry is the incident angle @xmath58 . at this point
it is important to note that the sequence of transmissions and reflections at the barrier matters .
for example the five - star trajectories shown in figs .
[ fig : fivepac]b and c which , if read from left to right , correspond to tr and rt , respectively , have different incident angles for the transmission and reflection events .
for an infinitely long barrier the relevant momentum component is the one perpendicular to the barrier @xmath59 we now may use the standard results for one - dimensional barrier penetration to approximate the transmission and reflection amplitudes by @xmath60 respectively , where @xmath61 and @xmath62 is the width of the barrier .
expressions ( [ eq : alphat ] ) and ( [ eq : alphar ] ) reduce to the usual one - dimensional expressions for incident angle @xmath63 .
we emphasize that it would not be sufficient to take the modulus of @xmath56 and @xmath57 since in general the contributions of different paths have to be added up coherently .
in fact , destructive interference of paths is quite important for the interpretation of the barrier height dependence of the length spectra . as can be seen from ( [ eq : alphar ] ) , the relative phase factor between the reflected and the transmitted path is always @xmath64 . since changing
a reflection at the barrier into a transmission and vice versa will change the exit lead , two classical paths going to the same exit lead differ by their behaviour at an even number of barrier encounters .
therefore , the relative phase factor will be @xmath65 .
in addition , a reflection at the barrier does not change the classical amplitude @xmath66 , so that the two paths either interfere perfectly constructive or destructive . as an example we consider the minimum in the peak height of the reflection spectrum corresponding to the five - star shown in fig .
[ fig:5peakh ] .
there is such a pronounced minimum only because the two contributing paths , tt and rr ( cf . figs .
[ fig : fivepac]a and d ) , interfere destructively . on the other hand
, @xmath48 is not vanishing at the minimum .
this is due to the fact that the fourier transformation has to be taken over a finite energy interval .
since the barrier height at which the minimum occurs is energy - dependent , the minimum of @xmath48 will be smeared out . for finite barrier width
, the phase factors appearing in the transmission and reflection amplitudes have an additional effect .
the contribution of the barrier region to the total action of the path will depend on the barrier height which will result in an effective change of the length of the trajectory .
this becomes more important as the width and height of the barrier are increased .
however , for the parameters used here , the barrier is thin enough so that the change in length is below the resolution of the discrete fourier transformation .
nevertheless , the peak height is affected .
this may become important for sufficiently high barriers and cause the decrease in the reflection spectrum of the five - star ( fig .
[ fig:5peakh ] ) at large @xmath0 . for high barriers of finite width slight changes in the scattering geometry with respect to the ideal geometry for a thin barrier may also affect the peak height .
we now turn to a more detailed discussion of the quantum mechanical results and those obtained from the semiclassical approach just introduced .
the data are compared in fig .
[ fig:5peakh ] for the five - star trajectory and in fig .
[ fig:7peakh ] for the seven - star . in both cases
the diamonds and triangles correspond to the quantum mechanical results for @xmath48 and @xmath55 , respectively , while the solid and dotted lines are the corresponding semiclassical results with the modifications described above .
the expressions for the reflection and transmission amplitudes ( [ eq : alphat ] ) and ( [ eq : alphar ] ) depend on the momentum component @xmath67 perpendicular to the barrier and the height @xmath0 and width @xmath62 of the barrier .
the geometry of the classical path together with the fermi energy determines @xmath67 . in the quantum calculations the barrier was implemented by increasing on three lattice points the potential to @xmath0 . for the curves shown in figs .
[ fig:5peakh ] and [ fig:7peakh ] we used an effective barrier width of 3.5 lattice spacings . from figs .
[ fig:5peakh ] and [ fig:7peakh ] we find , that at not too high barriers the agreement between the quantum mechanical data and the semiclassical theory modified for tunneling is very good . on the other hand , for high barriers deviations do occur .
this seems to be at odds with the fact that in the limit of infinite barrier the modified theory becomes equivalent to the usual semiclassical expansion .
a qualitative deviation appears for rather high barriers in the transmission spectrum of the five - star and the reflection spectrum of the seven - star where the quantum mechanical data saturate at a finite value .
this has been checked for barriers as high as @xmath68 . on the other hand ,
the semiclassical result decreases to zero since at least one barrier transmission is needed in order to get a path which describes transmission through the billiard and has the length of a five - star .
the same holds for paths which describe reflection at the billiard and have the same length as a seven - star .
this discrepancy may be explained by paths with length close to those of the five- or seven - star . as an example we consider the five - star for which the reflection spectrum in the absence of a barrier and the transmission spectrum for very high barrier are shown in fig .
[ fig : peakver ] as dashed and solid lines , respectively .
the highest peak of the dashed line indicates the length of the five - star .
as expected , there is no peak at this position in the transmission spectrum .
however , there exists a peak at somewhat smaller length which is broad enough to yield a contribution at the length of the five - star . the corresponding classical trajectory is shown in fig . [ fig : halfeight ] .
it involves three reflections at the barrier as well as a reflection at the exit lead , a quantum mechanical effect @xcite already mentioned above .
these arguments also provide a partial explanation of the quantitative deviations found for high barriers in the reflection spectrum of the five - star trajectory and the transmission spectrum of the seven - star .
in addition , as discussed in the previous section the discreteness of the fourier transformation in combination with the change = 0.45 = 0.3 in path length as a function of barrier height results in another source of discrepancy .
while the agreement between the quantum mechanical and the semiclassical calculation demonstrates that for prominent peaks like the ones corresponding to the five- and seven - star the interpretation in terms of classical paths is possible , the deviations just discussed show that nevertheless the identification is not necessarily straightforward and requires a certain amount of caution .
the dependence of peaks in the length spectrum on the barrier height may be of help in the identification since the number of extrema is related to the number of barrier encounters of the corresponding classical trajectory .
the scattering amplitudes discussed above should be accessible to microwave experiments where the `` fermi energy '' can be varied by changing the input frequency .
a barrier could be introduced by placing a dielectric slab into the cavity . on the other hand , in transport experiments on ballistic microstructures
it is the conductance which is measured . according to the landauer formula , eq .
( [ eq : landauer ] ) , the conductance through a cavity is proportional to the total transmission coefficient .
recently , measurements of transport through billiards with variable fermi energy have been performed @xcite so that it appears feasible to experimentally determine the length spectrum of the total transmission for the circular tunneling billiard .
since the semiclassical approaches developed so far provide a basis for the understanding of quantum transport , we would like to analyze now the effect of a barrier on the scattering probabilities .
the analysis for the transmission and reflection probabilities is more difficult than for the amplitudes because taking the squared modulus makes @xmath69 depend on pairs of trajectories .
accordingly , a fourier transformation yields peaks at lengths which correspond to differences between the lengths of two classical paths .
additionally , in order to calculate the total reflection or transmission , one has to take into account all scattering amplitudes @xmath70 or @xmath38 where the mode number is restricted by the fermi energy via the maximum mode number @xmath3 introduced in sec .
[ sec : intro ] .
this will increase the number of different paths to be considered since at higher mode numbers paths enclosing a larger angle with the lead axis become relevant .
moreover , when a new mode opens up it starts by being completely reflected , and the reflection coefficient jumps by one .
this staircase effect will introduce high harmonics in the reflection coefficient which are not related to classical trajectories .
therefore we will focus our attention on the total transmission coefficient @xmath8 . as an example
, the length spectrum of @xmath8 for the circular billiard without barrier is shown in fig .
[ fig : ft ] . due to the large amount of possible length differences
the spectrum is quite complex .
most of the peaks correspond not only to one pair of trajectories but to a combination of several pairs of approximately the same length difference .
for example the highest peak , which is found at @xmath71 , contains contributions of four different trajectories , namely a half of an eight - star ( cf .
[ fig : eight ] ) , of a twelve - star , a sixteen - star , and a twenty - star . from these trajectories ,
three length differences , @xmath72 , @xmath73 , and @xmath74 , can be constructed which are all quite close to @xmath71 . in principle , higher - order stars lead to further length differences in the same range . however , their contributions are negligible .
since the resolution of our discrete fourier transformation is @xmath75 and the typical peak width is twice as large ( cf .
[ fig : r11 ] ) , the contributions from the different combinations of trajectories can not be resolved .
given the complexity of the length spectrum , we may ask the question of whether the inclusion of a barrier can be treated as a perturbation that simply randomizes the phases of the contributing trajectories as do shape distortions @xcite or magnetic field changes @xcite .
our analysis in the previous sections suggests that this is not = 0.45 = 0.45 the case .
the inclusion of a barrier has the effect of suppressing the contribution from certain trajectories while increasing that of the symmetry related ones , resulting in the nonmonotonic behaviour of the peaks in the length spectrum as a function of the barrier height which _ can not be interpreted as conductance fluctuations . _ in fig .
[ fig : vat ] we present the dependence of the length spectrum on the barrier height for the total transmission coefficient as well as two individual probabilities , @xmath76 and @xmath77 . in the following , these spectra
are referred to as @xmath78 , @xmath79 , and @xmath80 , respectively .
the data are given for the peak at @xmath71 to which several paths contribute as discussed above .
the nonmonotonic behaviour resulting from the dependence of the individual trajectories ( of the contributing pairs ) on @xmath0 illustrates the interplay between semiclassics and tunneling . the fact that @xmath78 as well as @xmath81 and @xmath82 roughly exhibit the same structure implies that the same pairs of paths contribute to various transmission probabilities provided the momentum interval is large enough
of course , predicting the peak height dependence as we did for the scattering amplitudes is more difficult since the relative weight and the phases of the trajectories of the contributing pairs become relevant .
we therefore do not attempt to attain the agreement of sec .
[ sec : sclctb ] .
instead , we will give a qualitative explanation for the difference between @xmath81 and @xmath82 . with growing number of reflections at the circle , the incident angle of the half - star trajectories decreases , thus suppressing the contribution of the higher - order stars , especially to @xmath82 .
since the higher - order stars lead to more extrema in the @xmath0 dependence , the number of extrema of @xmath80 should be smaller than that of @xmath79 .
this is consistent with the numerical results of fig .
[ fig : vat ] .
the number of extrema suggests that the dominant contribution to @xmath80 stems from the combination of one half of an eight- and of a twelve - star .
one half of a sixteen - star should yield a contribution with a second minima for which there is no clear indication . on the other hand
, @xmath79 displays a second broad maximum which is probably a combination of all the extrema from higher stars .
obviously , the experimental resolution is limited and it would not be possible to observe a length spectrum as structured as that of fig . [
fig : ft ] . in mesoscopic
systems temperature plays two roles . on one hand
, it controls the inelastic processes like electron - electron and electron - phonon interactions .
this results in a cut - off length given by the inelastic mean - free path @xmath83 beyond which no trajectory should contribute in a semiclassical expansion .
therefore , length differences between trajectories which are longer than @xmath83 should not be considered , rendering in practice a length spectrum smoother than that of fig .
[ fig : ft ] , thus allowing for a simpler analysis . in the language of ref .
@xcite , the inclusion of @xmath83 corresponds to the `` smoothing - then squaring '' process , since it is the scattering amplitudes which loose their high harmonics .
the other effect of temperature is the rounding of the fermi surface that , within our semiclassical approach , cuts the large length differences .
that is , we should not consider the high harmonics of fig . [ fig : ft ] beyond the thermal length @xmath84 .
microwave cavities are not sensitive to @xmath84 , but have a finite @xmath83 , which can reach very large values in state - of - the - art experiments @xcite .
in this work we have studied quantum mechanically and semiclassically the effect of a tunneling barrier on quantum transport through ballistic cavities .
this effect is most evident in the length spectra , i.e. the fourier transform of the energy - dependent transmission and reflection amplitudes of the cavity .
we have shown that the peak heights in the quantum mechanical length spectrum exhibit a nonmonotonic variation upon increasing the barrier height .
this behaviour is quantitatively reproduced by combining a semiclassical approach to conductance with a simple model for barrier reflection and transmission of paths .
the model provides furthermore an intuitive physical picture of the underlying process leading to the variations in the peak heights : they reflect the superposition of coherent contributions from paths being reflected at or transmitted through the barrier .
this mechanism is a clear - cut manifestation of tunneling orbits in the conductance of quantum billiards , which should be in principle observable in experiment .
the effect of the tunnel barrier on individual peaks in the _ length _ spectrum manifests itself in the energy - dependent transmission and reflection coefficients .
we have also performed a corresponding analysis of the effect of a tunnel barrier on the _ area _ spectrum , the fourier transform of the magnetic field dependent transmission at fixed fermi energy . however , in that case a semiclassical analysis of the quantum mechanical results is more involved for two reasons : firstly , the areas of the trajectories are no longer unchanged by a reflection at the barrier . increasing the barrier height leads to a shift of the spectrum to small areas @xcite .
secondly , in order to obtain well - resolved peaks in the area spectrum one has to perform the fourier transformation over a rather large range of magnetic fields .
then , the condition that the cyclotron radius is much larger than the system size is no longer fulfilled in our numerical calculations and the shapes of the trajectories become field - dependent .
this leads to broadening and thus to a severe restriction of resolution . for high magnetic fields
one may even find splitting of the peaks in the area spectrum .
hence in most cases it will be hard to unambiguously identify peaks in the area spectrum .
low - frequency structure has been obtained in the ( non - averaged ) experimental area spectra of ballistic microstructures @xcite .
the observed peaks correspond to areas close to those of the shortest periodic orbits .
however , a clear identification between peaks and trajectories has not been possible to establish .
as discussed above the analysis becomes quite difficult when pairs instead of single trajectories are involved .
thus , a certain amount of caution has to be exerted for a detailed interpretation in terms of classical trajectories .
we have shown that the length spectrum admits a simple semiclassical analysis , _ even in the presence of tunneling_. in view of microwave experiments @xcite as well as recent work on cavities in two - dimensional electron gases @xcite which has demonstrated the possibility of measuring length spectra , it seems feasible to experimentally verify the nonmonotonic dependence on the barrier height discussed here . c. w. j. beenakker and h. van houten , in _ solid state physics _ , vol .
44 , ed . by h.
ehrenreich and d. turnbull ( academic press , san diego , 1991 ) .
s. datta , _ electronic transport in mesoscopic systems _ ( cambridge university press , cambridge , 1995 ) ; y. imry , _ introduction to mesoscopic physics _ ( oxford university press , new york , 1997 ) ; t. dittrich , p. hnggi , g .-
ingold , b. kramer , g. schn , and w. zwerger , _ quantum transport and dissipation _ ( wiley - vch , weinheim , 1997 ) .
stckmann and j. stein , phys .
* 64 * , 2215 ( 1990 ) ; h .- d .
grf , h. l. harney , h. lengeler , c. h. lewenkopf , c.rangacharyulu , a. richter , p. schardt , and h. a. weidenmller , phys.rev .
lett . * 69 * , 1296 ( 1992 ) ; m. kollmann , j. stein , u. stoffregen , h .- j .
stckmann , and b. eckhardt , phys .
e * 49 * , r1 ( 1994 ) . c. m. marcus , a. j. rimberg , r. m. westervelt , p. f.hopkins , and a. c. gossard , phys .
* 69 * , 506 ( 1992 ) ; c. m. marcus , r. m. westervelt , p. f. hopkins , and a. c. gossard , chaos * 3 * , 643 ( 1993 ) .
m. w. keller , o. millo , a. mittal , d. e. prober , and r. n. sacks , surf .
sci . * 305 * , 501 ( 1994 ) .
m. j. berry , j. h. baskey , r. m. westervelt , and a. c. gossard , phys .
b * 50 * , 8857 ( 1994 ) .
m. w. keller , a. mittal , j. w. sleight , r. g. wheeler , d. e. prober , r. n. sacks , and h. shtrikmann , phys .
b * 53 * , r1693 ( 1996 ) .
a. m. chang , h. u. baranger , l. n. pfeiffer , and k. w. west , phys .
lett . * 73 * , 2111 ( 1994 ) .
y. lee , g. faini , and d. mailly , chaos , solitons&fractals * 8 * , 1325 ( 1997 ) ; y. lee , g. faini and d. mailly , phys .
b , in press .
r. p. taylor , r. newbury , a. s. sachrajda , y. feng , p. t. coleridge , c. dettmann , n. zhu , h. guo , a. delage , p. j. kelly , and z. wasilewski , phys
. rev .
* 78 * , 1952 ( 1997 ) .
m. c. gutzwiller , _ chaos in classical and quantum mechanics _
( springer , new york , 1990 ) ; m. brack and r. k. bhaduri , _ semiclassical physics _ ( addison - wesley , reading , 1997 ) .
r. a. jalabert , h. u. baranger , and a. d. stone , phys .
lett . * 65 * , 2442 ( 1990 ) ; h. u. baranger , r. a. jalabert , and a. d. stone , chaos , * 3 * , 665 ( 1993 )
. h. ishio and j. burgdrfer , phys . rev .
b * 51 * , 2013 ( 1995 ) .
w. a. lin and r. v. jensen , phys .
b * 53 * , 3638 ( 1996 ) . c. d. schwieters , j. a. alford , and j. b. delos , phys .
b * 54 * , 10652 ( 1996 ) .
r. e. prange , e. ott , t. m. antonsen , jr .
, b. georgeot , and r. blmel , phys .
e * 53 * , 207 ( 1996 ) ; r. blmel , t. m. antonsen , jr . , b. georgeot , e.ott , and r. e. prange , phys .
e * 53 * , 3284 ( 1996 ) ; r. blmel , t. m. antonsen , b. georgeot , e. ott , and r. e. prange , phys .
* 76 * , 2476 ( 1996 ) .
m. j. berry , j. a. katine , r. m.westervelt , and a. c. gossard , phys .
b * 50 * , 17721 ( 1994 ) .
d. s. fisher and p. a. lee , phys .
b * 23 * , 6851 ( 1981 ) .
a. d. stone and a. szafer , ibm j. res .
, 384 ( 1988 ) .
p. a. lee and d. s. fisher , phys .
* 47 * , 882 ( 1981 ) .
h. u. baranger , d. p. divincenzo , r. a. jalabert , and a. d. stone , phys .
b * 44 * , 10637 ( 1991 ) .
i. v. zozoulenko , r. schuster , k .- f .
berggren , and k.ensslin , phys .
b * 55 * , r10209 ( 1997 ) .
h. bruus and a. d. stone , phys .
b * 50 * , 18275 ( 1994 ) . i. h. chan , r. m. clarke , c. m. marcus , k. campman , and a. c. gossard , phys .
* 74 * , 3876 ( 1995 ) . | we study the influence of a tunnel barrier on the quantum transport through a circular cavity .
our analysis in terms of classical trajectories shows that the semiclassical approaches developed for ballistic transport can be adapted to deal with the case where tunneling is present .
peaks in the fourier transform of the energy - dependent transmission and reflection spectra exhibit a nonmonotonic behaviour as a function of the barrier height in the quantum mechanical numerical calculations .
semiclassical analysis provides a simple qualitative explanation of this behaviour , as well as a quantitative agreement with the exact calculations .
the experimental relevance of the classical trajectories in mesoscopic and microwave systems is discussed . 2 |
in this letter we shall concentrate on the 1d non - linear @xmath1-sigma model : ( a ) the space - time is the ( 1 + 1)-dimensional @xmath2 minkowski space . @xmath3 ( @xmath4 ) denotes a point in @xmath2 .
we choose the metric in the form @xmath5 and the dalembertian reads : @xmath6 .
( b ) the target ( internal ) space is the @xmath0-sphere .
this is in contrast with the original gell - mann / lvy model where the target space is @xmath7 @xcite .
three scalar fields @xmath8 , @xmath9 , define a map @xmath10 if they are constrained to the surface in @xmath11 : @xmath12 the action @xmath13 seems to be simple but together with the constraint ( [ sphere ] ) governs the complicated non - linear dynamics of two goldstone bosons with coupling constant @xmath14 .
in one spatial dimension , however , goldstone particles do not exist @xcite .
the infrared asymptotics of the ( [ sphere])-([action2 ] ) system induces a potential energy density that we choose as : @xmath15 giving masses @xmath16 and @xmath17 to the massless excitations . with no loss of generality
, we assume that : @xmath18 , and we define the non - dimensional parameter : @xmath19 , @xmath20 , measuring the ratio between particle masses .
we also re - scale the space - time coordinates to address non - dimensional variables : @xmath21 . solving the constraint
the action reads @xmath22 despite the potential energy density being quadratic , there are two homogeneous minima of the action ( vacua ) : the north and south poles : @xmath23 .
thus , the discrete symmetry of the action([action3 ] ) @xmath24 generated by @xmath25 , @xmath9 is spontaneously broken to @xmath26 ( generated by @xmath27 , @xmath28 ) . if the two masses were equal , this unbroken symmetry would become the @xmath29 rotation group around the north - south pole axis .
using spherical coordinates , @xmath30 , @xmath31 , @xmath32 , in the chart @xmath33 of @xmath0 the energy of static configurations : @xmath34 , @xmath35 , @xmath36 , and the potential energy density read : @xmath37+v(\theta,\varphi)\right\}\ ] ] @xmath38 the configuration space of the system @xmath39 is formed by four disconnected sectors according to the tendency of every finite energy configuration towards either the north or south pole at the extremes of the spatial line @xmath40 .
solutions for which the temporal dependence is of the form @xmath41 for some velocity @xmath42 , are solitary or traveling waves .
lorentz invariance provides all the solitary waves from the static solutions of the field equations @xmath43 on the orbits @xmath44 ( half meridians ) the system ( [ field - eq1])-([field - eq2 ] ) becomes the ode of the pendulum and the separatrix trajectories between bounded and unbounded motion ; @xmath45 are the solitary waves or kinks ( at their center of mass ) of finite energy : @xmath46 , interpolating between the north @xmath47 and south @xmath48 poles .
thus , these kinks belong to a different sector in configuration space than the vacua , an evident fact in cartesian coordinates ; see figures 1 - 2:mpg/ _ ( mathematica tools ) there is a mathematica file containing several extra figures and formulas concerning this work . ]
@xmath49},\ \pm { \rm tanh}[\sigma(x - x_0)]\right)\ ] ] and @xmath50 ( @xmath51 ) kink orbits .
b ) @xmath52 ( blue ) and @xmath50 ( red ) kink energy densities , title="fig:",height=94 ] and @xmath50 ( @xmath51 ) kink orbits .
b ) @xmath52 ( blue ) and @xmath50 ( red ) kink energy densities , title="fig:",height=94 ] kinks .
b ) perspective from one component of the boundary of the target sphere @xmath53 the spatial line infinite cylinder : @xmath54 .
the meridian @xmath55 , @xmath56 is plotted as orthogonal to the spatial line.,title="fig:",height=94 ] kinks .
b ) perspective from one component of the boundary of the target sphere @xmath53 the spatial line infinite cylinder : @xmath54 .
the meridian @xmath55 , @xmath56 is plotted as orthogonal to the spatial line.,title="fig:",height=94 ] the half - meridians @xmath57 or @xmath58 are also good trial orbits in the sense that they provide new kinks of finite energy @xmath59 @xmath60 via finite action solutions of another pendulum equation .
these solitary waves live in the same sector of the configuration space as the previous ones and look similar in cartesian coordinates ; see figures 1 - 3 : @xmath61 kinks .
b ) perspective from one component of the boundary of the infinite cylinder : @xmath54 . the meridian @xmath62 , @xmath63 , is plotted aligned with the spatial line.,title="fig:",height=94 ] kinks .
b ) perspective from one component of the boundary of the infinite cylinder : @xmath54 .
the meridian @xmath62 , @xmath63 , is plotted aligned with the spatial line.,title="fig:",height=94 ] at the @xmath64 limit , all the half meridians @xmath65 are good trial orbits and there is a one - parametric family of solitary waves : @xmath66 , @xmath67 degenerated in energy : @xmath68 , @xmath69 .
we could now try to search for more kinks even in the case @xmath70 of distinct masses by using rajaraman s trial orbit method @xcite , but instead we shall profit from the fact that the ode system ( [ field - eq1])-([field - eq2 ] ) is integrable using elliptic coordinates in the @xmath0 sphere . in @xmath0
we fix the two points : @xmath71 , @xmath72 , @xmath73 .
the distance between them is : @xmath74 , see figure 4 . given a point @xmath75 , let us consider the distances @xmath76 $ ] and @xmath77 $ ] from @xmath78 to @xmath79 and @xmath80 .
the elliptic coordinates of @xmath78 are half the sum and half the difference of @xmath81 and @xmath82 : @xmath83 .
the formulae @xmath84 allow one to pass from elliptic to cartesian coordinates . here
, a simplified notation is used : @xmath85 , @xmath86 , @xmath87 , @xmath88 , etc . the differential arc - length in @xmath1 in this system of coordinates
is @xmath89 choosing the foci in such a way that @xmath90 , the energy density of our system in elliptic coordinates reads : @xmath91&=&\frac{1}{2}\left[\frac{\su^2-\sv^2}{\su^2-\sf^2}\left(\frac{du}{dx}\right)^2 + \frac{\su^2-\sv^2}{\sf^2-\sv^2}\left(\frac{dv}{dx}\right)^2\right ] \nonumber \\ & & -\frac{f(u)+g(v)}{\su^2-\sv^2 } \label{separation}\end{aligned}\ ] ] @xmath92 , @xmath93 .
the mechanical analogy demands that we think of @xmath94 as the lagrangian , @xmath95 as the time , @xmath96=-v[u(x),v(x)]$ ] as the mechanical potential energy , and the target manifold @xmath0 as the configuration space .
the structure of @xmath97 $ ] is such that we are dealing with a type i liouville model @xcite on the sphere , i.e. , a dynamical system which is hamilton - jacobi separable in elliptic coordinates .
the kink orbits ( finite action trajectories ) and the kink profiles ( time " schedules of these trajectories ) are given in the hamilton - jacobi framework @xcite-@xcite via the quadratures : ( @xmath98 , @xmath99 ) @xmath100 @xmath101
in this way we find a family of non - topological kink ( ntk ) orbits by integrating the first quadrature ( they start and end either at the north or the south pole , they live in the vacua sectors of the configuration space ) parametrized by the integration constant @xmath102 : @xmath103^{{\rm sg}p_u } \left [ \frac{|\tan \frac{v}{2r}|}{\left|\tan \frac{v - f}{2r } \tan \frac{v+f}{2r}\right|^{\frac{1}{2\cf } } } \right]^{{\rm sg}p_v}\ ] ] the kink profiles of these non - topological solitary waves given by the integration of the second quadrature @xmath104 depend on one integration constant , @xmath105 , which sets the center of each kink , see figures 5 and 6 . ntk kink .
b ) perpendicular cross section of the field variation.,title="fig:",height=94 ] ntk kink .
b ) perpendicular cross section of the field variation.,title="fig:",height=94 ] : 1 ) @xmath106 , highest peak on the left ( blue ) 2 ) @xmath107 , symmetrical peaks ( green ) 3 ) @xmath108 highest peak on the right ( red).,title="fig:",height=94 ] : 1 ) @xmath106 , highest peak on the left ( blue ) 2 ) @xmath107 , symmetrical peaks ( green ) 3 ) @xmath108 highest peak on the right ( red).,title="fig:",height=94 ] the hamilton - jacobi method also provides the energy of the ntk .
the hamilton characteristic function ( the solution of the hamilton - jacobi stationary equation ) for zero mechanical energy is : @xmath109 from this function we compute the energy of the ntk kinks : @xmath110 all the ntk kinks have the same energy and satisfy the kink mass sum rule : @xmath111 @xmath52 and @xmath50 are singular kinks that arise at the @xmath112 limit of the ntk kink moduli space .
their orbits lie on the boundary of the elliptic rectangle and the @xmath113 axis , see figure 7 .
the @xmath52 kink orbit is the straight line @xmath114 .
the @xmath50 kink orbit , however , is a three - step trajectory : the @xmath115 , @xmath116 , and @xmath117 edges of the rectangle . observe that all the ntk orbits starting from the north pole @xmath47 ( south pole @xmath48 ) meet at one of the antipodal foci @xmath118-@xmath119 ( foci @xmath79-@xmath80 ) , which are thus conjugate points to the poles . according to the morse index theorem
, these kinks are unstable , see @xcite , @xcite
. the @xmath50 orbits also cross the foci and only the kinks of type @xmath52 are stable .
the wess - zumino action @xmath120\frac{\partial\phi_a}{\partial t}(t , x)\ ] ] produces the euler - lagrange equation : @xmath121\cdot\frac{\partial\phi_b}{\partial t}\ ] ] we choose the non - null components of the vector potential `` in the north and south hemispheres of the target space in the form ( using @xmath122 , @xmath123 ) : @xmath124=-\sum_{\beta=1}^2\varepsilon_{\alpha\beta}\frac{\phi_\beta(t , x)}{2r(\phi_3(t , x)\pm r ) } \ , \ , , \ , \ , \alpha=1,2\ ] ] a magnetic monopole field '' arises in the target space : @xmath125=\frac{\phi_a(t , x)}{r^3}$ ] .
the combined euler - lagrange equations for @xmath126 are : @xmath127 at the long wavelength limit @xmath128 , system ( [ ll ] ) becomes the landau - lifhsitz equations for ferromagnetism with a dispersion relation : @xmath129 .
the connection between the semi - classical ( high - spin ) limit of the heisenberg model with the quantum non - linear sigma model is well established @xcite .
thus , our kinks , which are also static finite energy solutions of ( [ ll ] ) , are solitary spin waves in the low - energy regime of quantum ferromagnets , although the symmetry is contracted from lorentzian to galilean .
in this letter we have reached the following conclusions about the kink manifold of the 1d massive non - linear @xmath0-sigma model : 1 .
if the masses of the pseudo - goldstone particles are equal , there exist a @xmath130-family ( fixed the kink center of mass ) of topological kinks degenerated in energy living on all the half - meridians of the @xmath0-sphere .
when the masses differ , only two pairs of topological kinks survive , each pair of kinks having distinct energy .
even if the masses of the pseudo - goldstone particles are different , we have shown that there exist a one - parametric family ( for fixed cm ) of non - topological kinks degenerated in energy by using elliptic coordinates on the @xmath0-sphere .
it is also shown that there is a curious kink mass sum rule between the non - topological and topological kinks and that only one of the topological kink pairs is formed by stable kinks .
4 . finally
, we have noticed by adding a wess - zumino - type term to the action that our kinks are solitary spin waves in the long wavelength limit of ferromagnetic materials .
we are grateful to mariano santander for enlightening explanations about elliptic coordinates on the sphere and other cayley - klein two - dimensional spaces . | we describe the kink solitary waves of a massive non - linear sigma model with an @xmath0 sphere as the target manifold .
our solutions form a moduli space of non - relativistic solitary waves in the long wavelength limit of ferromagnetic linear spin chains . |
lately , outstanding achievements have been made in development of a novel class of uncooled microbolometer ir focal plane arrays ( fpas ) , the ones based on si - on - insulator diodes as temperature sensors , whose format has reached 2 megapixels with a noise equivalent temperature difference ( netd ) of 60 mk at the frame rate of 15 hz and the f - number of 1 ; the same group has also demonstrated a vga fpa with outstanding netd of 21 mk ( at f/1 , 30 hz ) [ see , e.g. , ref.@xcite and earlier articles cited therein ] .
this success , as well as previous achievements in this field @xcite , stimulates search for simple cmos compatible technological solutions based on diode bolometers which would be suitable for mass production of ir fpas with low cost and netd figures sufficient for many civil applications @xcite .
one of such solutions consists in utilization of metal / poly - si schottky barriers for formation of sets of temperature sensors on bolometer membranes @xcite .
schottky barrier bolometer arrays seem to be first proposed theoretically for very sensitive cooled bolometers @xcite . in this article ,
nickel silicide schottky diodes formed on polycrystalline si@xmath0p@xmath1 films are proposed as thermosensitive elements of monolithic uncooled microbolometer ir fpas . the possibility of integration of technological process of the silicide based schottky diode structure formation into the standard cmos technology of vlsi manufacturing @xcite as well as the possibility of cascade connection of schottky diodes to increase the temperature sensitivity of bolometer elements of fpa and the use of layers of the diode structures as absorbing coatings in bolometers are advantages of these structures .
schottky barriers were formed on commercial single - crystalline cz silicon wafers ( @xmath6 cm , ( 100 ) , @xmath7-type ) coated by a @xmath4600 nm thick layer of sio@xmath3 formed by thermal oxidation and a @xmath4180 nm thick layer of pyrolytic si@xmath2n@xmath8 ( the dielectric layers simulated a design of the supporting membranes of the previously tested bolometer cells @xcite ) .
films of polycrystalline si@xmath0p@xmath1 with the thicknesses of @xmath4150 nm were deposited by thermal decomposition of monosilane at the substrate temperature @xmath9 ; then they were doped by phosphorus by ion implantation ( @xmath10=35kev ) to the dose of @xmath11@xmath12 and annealed at 700 for 30 min . after wafer cleaning in a boiling ammonia - peroxide mixture solution ( nh@xmath8oh : h@xmath3o@xmath3:h@xmath3o=1:1:4 , 10 min ) and surface hydrogenation ( hf : h@xmath3o=1:10 , 30 sec at room temperature ) , ni - silicide / poly - si schottky diodes were formed by thermal deposition of a nickel film ( @xmath445 nm thick , @xmath13k , the residual gas pressure @xmath14torr ) from a tungsten crucible followed by annealing at 400 in nitrogen for 30 min .
al contacts to poly - si were formed by thermal deposition from tungsten crucible in vacuum ( @xmath15torr , @xmath13k ) and annealing at 450 in nitrogen for 15 min . aluminium contacts to the structure top
were deposited in the same way but without annealing .
golden wires were welded to the contact pads .
structural perfectness and chemical composition of the layers were explored by means of transmission electron microscopy ( tem ) .
test elements for electrical measurements were formed by contact lithography and had the sizes of @xmath16 mm .
_ i v _ characteristics of the schottky diodes were measured in darkness at different temperatures varied in the range from 20 to 70 and at the temperature of 80k .
photovoltage ( @xmath17 ) spectra were obtained as described in ref.@xcite ; for each photon energy ( @xmath18 ) the photoresponse value @xmath17 was normalised to the number of incident photons .
uncoated satellites were used for measurement of sheet resistance ( @xmath19 ) of the poly - si films .
the wsxm software @xcite was used for tem image processing .
a typical tem micrograph of the resultant structure ( fig.[fig : tem ] ) represents images of polycrystalline ni silicide and polysilicon layers between si@xmath2n@xmath8 and al films .
the ni silicide film is seen to be composed by a number of phases : at least two phases with the grains close in sizes and comparable volume fractions are distinctly observed by tem .
bright inclusions are also observed at ni - silicide / poly - si interface ; we presumably interpret them as residual silicon oxide particles .
it is also seen in fig.[fig : tem ] that after formation of the ni - silicide / poly - si film , the average thicknesses of the ni silicide and poly - si layers became 60 and 135 nm , respectively .
using the mass conservation law this allows us to estimate the density of the silicide film as @xmath47 g/@xmath20 ( we adopt the density of poly - si to be 2.33 g/@xmath20 and the density of the initial poly - ni film to be 8.9 g/@xmath20 ) .
this in turn allows us to roughly evaluate the composition of the silicide layer ( the required densities of ni silicides can be found , e.g. , in refs.@xcite ) .
if we postulate that the silicide film consists of only two phases , as it is stated in ref.@xcite , then they might be ni@xmath3si and nisi ( the process temperature did not exceed 450 and mainly was 400 or lower ; it is known however that nisi@xmath3or slightly more silicon rich compound @xcite , according to ref .
, ni@xmath21si@xmath22forms at @xmath23 ( or even @xmath24 @xcite ) , whereas nisi and ni@xmath3si form at @xmath25 and 200 , respectively ; according to ref.@xcite , appearance of these two low - temperature phases of ni silicides after annealing in vacuum would evidence that the original ni film has been completely ( or nearly completely ) consumed by the growing ni@xmath3si phase).@xmath26 in this case the volume fraction of ni@xmath3si : nisi@xmath2785:15 ( taking into account all uncertainties , the maximum estimate yields 100% of ni@xmath3si ) ; the mass fraction of ni@xmath3si exceeds 88% .
this obviously contradicts our tem observations and makes us assume the presence of the heaviest of the ni silicides , ni@xmath2si @xcite , which also may form at low temperatures , especially taking into account the possible presence of oxygen in the metal film that , according to refs.@xcite , impedes diffusion of ni atoms to ni / ni@xmath3si interface and , in our opinion , may result in simultaneous formation of ni@xmath3si and ni@xmath2si phases in the silicide film .
if our assumption is true the silicide film might be composed , by a rough estimate , by 20 to 40% of ni@xmath2si , 30 to 60% of ni@xmath3si and 10 to 30% of nisi in respective proportions to give a total of 100% of the silicide film volume .
the lightest ( the least dense ) silicide phase having a si - rich stoechiometry ( disilicide ) may also be available in the form of a thin diffusion layer at the ni - silicide / poly - si interface ( this does not contradict our observations ) @xcite ; it may affect the barrier height of the whole silicide layer , however @xcite .
@xmath28@xmath29 characteristics of the structures ( fig.[fig : i - v-300k]a , b ) with low - resistance poly - si ( @xmath30 ) , which forms in our process , manifest a diode behavior with the rectification ratios changing from @xmath4100 to @xmath420 for the temperature varied from 22 to 70 ( fig.[fig : i - v-300k]c ) . at liquid nitrogen temperature
the rectification becomes more pronounced and exceeds 1000 at biases exceeding 2 v ( fig.[fig : i - v-80k ] ) .
it should be noticed that at forward biasing the negative lead was set on the silicide top contact pad whereas the positive one was set on the contact pad to the polysilicon film .
photo - emf spectra obtained at 300 and 80k ( fig.[fig : emf ] ) demonstrate photoresponse for photons with energies greater and less than the si bandgap width ( @xmath31 ) as well as the presence of a number of potential barriers in the diode film .
room temperature measurements with and without a silicon filter have revealed the only barrier with the height @xmath32ev ( remark that the negative pole of the photodiode was on the contact pad to the silicide when the diode was illuminated by the white light ) .
a richer collection of barriers has been revealed at 80k .
the highest one nearly coincides in energy with @xmath31 ( @xmath33ev with 95% confidence limits of 1.08 and 1.14ev ) .
a lower one @xmath34 ( with the 95% confidence limits of 0.66 and 0.78ev ) is close to the values ascribed in the literature to all ni silicide barriers with @xmath35-type si @xcite ( equality of barrier heights of all nickel silicides was explained by the presence of similar diffusion layers in all nickel - silicide / silicon interfaces @xcite ) .
estimation of the lowest one yields a figure of @xmath36ev ( the 95% confidence band is from 0.48 to 0.54ev ) ; a barrier of this height , to our knowledge , has never been connected with a ni - silicide / si transition in the literature.@xmath37 however , we attribute all the above barriers to the ni - silicide / poly - si interface .
our reasoning is as follows .
the band structure of a polysilicon film is known to be spatially inhomogeneous : a strong potential relief is associated with grain boundaries @xcite . in @xmath35-si , even in the heavily doped @xmath38 one , there may exist depleted or even @xmath7-type spatial domains @xcite which , on the one hand , as a result of band - to - band transitions may be sources of electron - hole pairs which , in turn , are separated by the potential relief and generate the photo - emf of the observed polarity because , despite that the potential peaks should be more or less symmetrical and the electron - hole pairs should arise with close likelihoods on both their slopes , a part of electrons escapes from the si film accumulating in silicide whereas holes are localized at the grain boundaries .
this process may give rise to the photovoltage under irradiation by photons with energies @xmath39 .
in addition to charge separation on opposite sides of the film , this process also increases the potential relief . on the other hand
, grain boundaries may serve as potential barriers for electrons localized in @xmath38-si grains segregating them from the ni - silicide film and producing the photo - emf of the observed polarity due to electron injection into the silicide under the effect of photons with @xmath40 . @xmath38-si potential valleys adjoining the ni - silicide film form ohmic contacts .
this argumentation explains the presence of the only barrier @xmath41 detected at room temperature as well as the observed polarity of both the resultant photovoltage and the forward current .
a model of processes taking place at liquid nitrogen temperature is some more tricky . as it follows from the @xmath28@xmath29 characteristics ( fig.[fig : i - v-80k ] ) , the free electrons are partially frozen out in the structure and the fermi level moves down that increases the barriers for electron photoinjection into the silicide film .
it makes @xmath41 move to the right in energy to appear in the photovoltage spectra as @xmath42 .
two processes can be mixed in this conditions , band - to - band transition with separation of electron - hole pairs and electron injection into the silicide over the potential barrier , both generating photo - emf . in addition
, a reduction of @xmath35 may increase barriers at the interface @xcite ; a usual ni - silicide barrier ( @xmath43 ev ) may be completely restored at some domains or be still reduced ( @xmath44 ev ) at different places . hole injection into the silicide layer from polysilicon grain boundaries may become more probable over reduced barriers to holes .
this statement finds confirmation in the spectra plotted in fig.[fig : emf+ir ] which have been obtained under irradiation of a diode by a wide - band ir radiation of a tungsten bulb filtered by a polished si wafer [ @xmath45 ) ] .
it is seen in the spectra that the higher the power density of the incident radiation on the sample , the stronger the curves bow in the high - energy part of the graph and the less values of the photo - emf are detected .
it may be caused by injection of holes from potential wells at grain boundaries of poly - si into the silicide film because of additional wide - band ir lighting of the sample resulting in charge reduction of both the silicide and polysilicon layers .
thus , a set of competing processes becomes possible at 80k .
non - uniformity of the spatial potential throughout the ni - silicide / poly - si interface may locally act in favour of one of these competing processes . as a consequence , impact of several barriers
is observed in the photoresponse spectra in the order of magnitude of contribution of processes associated with them to the resultant photo - emf in different spectral ranges .
investigating the temperature dependences of the @xmath28@xmath29 characteristics close and above the room temperature we have found the thermal sensitivity of the diodes to be sufficiently high to consider them as a potential elements of uncooled bolometers .
panels ( a ) and ( b ) in fig.[fig : tsc - i ] demonstrate temperature dependences of the forward and reverse currents of the diodes @xmath28 for fixed ( and stabilized ) voltages @xmath46 .
temperature coefficient of the sensor current tcs@xmath47/{\rm d}t$ ] , where @xmath48 , derived from the graphs presented in the panels ( a ) and ( b ) as a function of bias voltage ( fig.[fig : tsc - i]c ) vary from @xmath49 to @xmath50%/ for the forward bias and remains nearly constant around 2.5%/ for the reverse bias .
notice that at small values of the forward bias , tcs is positive but rapidly drops with the growth of the absolute bias and equals 0 at @xmath51v .
we think that the negative tcs may result from the metallic behavior of the poly - si film as a function of temperature .
temperature dependences of the voltage drop across the diode @xmath46 for fixed ( and stabilized ) forward and reverse currents @xmath28 are shown in fig.[fig : tsc - u]a , b .
the temperature coefficient of voltage tcs ( @xmath52 ) derived from the graphs depicted in fig.[fig : tsc - u]a , b ( the curves in the panel ( b ) are linearized over an interval from 20 to 60 ) vary from 0.3 to 0.6%/ for forward biasing and from @xmath53 to @xmath54%/ for reverse biasing ( fig.[fig : tsc - u]c , d ) .
as of now , we foresee two ways of improvement of electrical properties of the structure .
the first of them consists in modification of the schottky barrier formation process proposed in ref.@xcite which enables production of poly - si / ni - polycide schottky diodes with rectification ratios as high as @xmath55 .
the other possibility is to replace poly - si by @xmath56-si : h and to apply the metal induced crystallization to form diodes nearly as perfect as those produces on the basis of single - crystalline si @xcite .
each of these alternatives in principle could enable the development of high - performance monolithic schottky - diode microbolometer ir fpas.@xmath57
in summary , nickel silicide schottky diodes formed on polycrystalline si@xmath0p@xmath1 films are proposed as temperature sensors of monolithic uncooled microbolometer ir focal plane arrays . structure and chemical composition of the schottky diodes
have been examined by tem .
the ni silicide has been identified as multi - phase mixture composed by 20 to 40% of ni@xmath2si , 30 to 60% of ni@xmath3si and 10 to 30% of nisi with probable minor content of nisi@xmath3 at the silicide / poly - si interface .
@xmath28@xmath29 characteristics of the diodes studied at different temperatures demonstrate the rectification ratios varying from @xmath420 to @xmath4100 when the temperature changes from 70 to 22 and exceeding 1000 at 80k .
a barrier of @xmath5 ev has been found to control the photovoltage spectra at room temperature .
three barriers with approximate heights from 1.08 to 1.14ev , from 0.66 to 0.78 and from 0.48 to 0.54 ev have been observed in photo - emf spectra at 80k and associated with the ni - silicide / poly - si interface .
absolute values of temperature coefficients of voltage and current have been found to vary from 0.3 to 0.6%/ for the forward biased structures and around 2.5%/ for the reverse biased ones .
cmos , complementary metal - oxide semiconductor ; cz , czochralski or grown by the czochralski method ; emf , electromotive force ; fpa , focal plane array ; ir , infrared ; netd , noise equivalent temperature difference ; tem , transmission electron microscopy .
the authors declare that they have no competing interests .
kvc participated in the design of the study , carried out the experiments , performed data analysis and participated in the discussions and interpretation of the results .
vac participated in the design of the study , took part in the discussions and interpretation of the results ; he also supervised the researches performed by young scientists and students .
vpk participated in the design of the study , took part in the discussions and interpretation of the results .
vyr performed the tem studies and took part in the discussions and interpretation of the results .
mss investigated the photo - emf spectra ; he carried out the experiments , performed data analysis and took part in discussions and interpretation of the results .
vay conceived of the study and designed it , performed data analysis and took part in the discussions and interpretation of the results ; he also supervised the research project .
all authors read and approved the final manuscript .
kvc is a junior research fellow , vac is a leading research fellow and mss is a phd student at laboratory of nanophotonics , department of applied thermography , prokhorov general physics institute , russian academy of sciences .
vyr is a senior research fellow and vpk is a head of laboratory of medium ir - range crystalline lasers at department of applied thermography , prokhorov general physics institute ; vpk is also a co - founder and a board member of technopark of gpi ras and a co - founder and a partner in thermographic systems ltd .
vay is a head of department of applied thermography and a head of laboratory of nanophotonics at prokhorov general physics institute ; he is also a co - founder and a board member of technopark of gpi ras and a co - founder and a partner in thermographic systems ltd .
the equipment of the center for collective use of scientific equipment of gpi ras was used for this study .
we acknowledge the technological support of our work .
we thank ms .
n. v. kiryanova for her valuable contribution to arrangement and management of this research .
we express our appreciation to mr .
v. p. korolkov and mr .
g. a. rudakov for performing the technological processes .
we are grateful to ms .
l. a. krylova for carrying out chemical treatments of the experimental samples .
@xmath26we can not discriminate between @xmath58 and @xmath59 phases of ni@xmath3si@xcite and , following ref.@xcite , suppose that only the @xmath58 phase is present ; the experimental value of its density , taken from ref.@xcite , makes 7.23 g/@xmath20 whereas its x - ray density ( 7.405 g/@xmath20 ) coincides in various sources @xcite .
@xmath37a barrier of this height is attributed to the ni / si interface in ref.@xcite ; yet , we have not observed a direct contact of ni to si by tem after the silicide film formation .
@xmath57notice also that there is an additional advantage of the considered structures with schottky barriers : they may be applied both as temperature sensors of bolometers for detection in mid - ir or far - ir and as photonic sensors for detection in near - ir and visible spectral ranges .
\(1 ) the si@xmath2n@xmath8 substrate film ; the diode film consists of ( 2 ) poly - si , ( 3 ) nickel silicide and ( 4 ) al contact layers .
\(a ) forward and ( b ) reverse biasing ; ( c ) rectification ratio vs the applied voltage .
\(a ) @xmath28@xmath29 characteristics and ( b ) rectification ratio as a function of the applied bias .
@xmath41 is an estimate of the barrier height derived from the photo - emf spectral measurements at 300k with and without a si filter ; @xmath42 , @xmath60 and @xmath61 are barrier heights estimated from the photo - emf spectral response at 80k ( solid lines show line fits , dotted ones set 95% confidence bands ) .
the power density of light with @xmath62 on the diode is shown in the legend in mw/@xmath63 ; dashed lines are guides to the eye .
\(a ) forward and ( b ) reverse currents ( the legend represents the applied bias in volts for each line ) ; ( c ) temperature coefficient of current vs fixed voltage on the structure ; negative and positive values of @xmath46 in panel ( c ) correspond to forward and reverse biasing , respectively .
\(a ) forward and ( b ) reverse biases ( the legends represent the currents in @xmath64a for each line ) ;
( c , d ) temperature coefficient of voltage for each branch of _ i v _ characteristics vs fixed current through the structure ; to derive the graph ( d ) , the curves in the panel ( b ) were linearized in the interval from 20 to 60 ; negative and positive values of @xmath28 in panels ( c ) and ( d ) correspond to forward and reverse biasing , respectively .
tem images demonstrate a schottky - diode film composed by three layers on si@xmath2n@xmath8 : ( 1 ) the si@xmath2n@xmath8 substrate film ; the diode film consists of ( 2 ) poly - si , ( 3 ) nickel silicide and ( 4 ) al contact layers . , title="fig:"](a ) tem images demonstrate a schottky - diode film composed by three layers on si@xmath2n@xmath8 : ( 1 ) the si@xmath2n@xmath8 substrate film ; the diode film consists of ( 2 ) poly - si , ( 3 ) nickel silicide and ( 4 ) al contact layers . ,
title="fig:"](b ) @xmath28@xmath29 characteristics of the ni - silicide / poly - si structure and its rectification ratios at different temperatures : ( a ) forward and ( b ) reverse biasing ; ( c ) rectification ratio vs the applied voltage .
, title="fig : " ] @xmath28@xmath29 characteristics of the ni - silicide / poly - si structure and its rectification ratios at different temperatures : ( a ) forward and ( b ) reverse biasing ; ( c ) rectification ratio vs the applied voltage .
, title="fig : " ] + + @xmath28@xmath29 characteristics of the ni - silicide / poly - si structure and its rectification ratios at different temperatures : ( a ) forward and ( b ) reverse biasing ; ( c ) rectification ratio vs the applied voltage . , title="fig : " ] @xmath28@xmath29 characteristics of the ni - silicide / poly - si structure its rectification ratios at liquid nitrogen and room temperatures : ( a ) @xmath28@xmath29 characteristics and ( b ) rectification ratio as a function of the applied bias .
, title="fig : " ] @xmath28@xmath29 characteristics of the ni - silicide / poly - si structure its rectification ratios at liquid nitrogen and room temperatures : ( a ) @xmath28@xmath29 characteristics and ( b ) rectification ratio as a function of the applied bias .
, title="fig : " ] photovoltage spectra obtained at room and liquid nitrogen temperatures : @xmath41 is an estimate of the barrier height derived from the photo - emf spectral measurements at 300k with and without a si filter ; @xmath42 , @xmath60 and @xmath61 are barrier heights estimated from the photo - emf spectral response at 80k ( solid lines show line fits , dotted ones set 95% confidence bands ) . ]
photovoltage spectra obtained at 80k ; the diode is irradiated by lite of a tungsten lamp through a si filter : the power density of light with @xmath62 on the diode is shown in the legend in mw/@xmath63 ; dashed lines are guides to the eye . ]
temperature dependences of current for fixed voltages on a ni - silicide / poly - si schottky diode and temperature coefficient of signal ( current ) for each branch of the _ i v _ characteristics : ( a ) forward and ( b ) reverse currents ( the legend represents the applied bias in volts for each line ) ; ( c ) temperature coefficient of current vs fixed voltage on the structure ; negative and positive values of @xmath46 in panel ( c ) correspond to forward and reverse biasing , respectively .
, title="fig : " ] temperature dependences of current for fixed voltages on a ni - silicide / poly - si schottky diode and temperature coefficient of signal ( current ) for each branch of the _ i v _ characteristics : ( a ) forward and ( b ) reverse currents ( the legend represents the applied bias in volts for each line ) ; ( c ) temperature coefficient of current vs fixed voltage on the structure ; negative and positive values of @xmath46 in panel ( c ) correspond to forward and reverse biasing , respectively .
, title="fig : " ] + + temperature dependences of current for fixed voltages on a ni - silicide / poly - si schottky diode and temperature coefficient of signal ( current ) for each branch of the _ i v _ characteristics : ( a ) forward and ( b ) reverse currents ( the legend represents the applied bias in volts for each line ) ; ( c ) temperature coefficient of current vs fixed voltage on the structure ; negative and positive values of @xmath46 in panel ( c ) correspond to forward and reverse biasing , respectively .
, title="fig : " ] temperature dependences of voltage for fixed currents through a ni - silicide / poly - si schottky diode and temperature coefficient of signal ( voltage ) for each branch of _ i v _ characteristics : ( a ) forward and ( b ) reverse biases ( the legends represent the currents in @xmath64a for each line ) ; ( c , d ) temperature coefficient of voltage for each branch of _ i v _ characteristics vs fixed current through the structure ; to derive the graph ( d ) , the curves in the panel ( b ) were linearized in the interval from 20 to 60 ; negative and positive values of @xmath28 in panels ( c ) and ( d ) correspond to forward and reverse biasing , respectively . , title="fig : " ] temperature dependences of voltage for fixed currents through a ni - silicide / poly - si schottky diode and temperature coefficient of signal ( voltage ) for each branch of _ i v _ characteristics : ( a ) forward and ( b ) reverse biases ( the legends represent the currents in @xmath64a for each line ) ; ( c , d ) temperature coefficient of voltage for each branch of _ i v _ characteristics vs fixed current through the structure ; to derive the graph ( d ) , the curves in the panel ( b ) were linearized in the interval from 20 to 60 ; negative and positive values of @xmath28 in panels ( c ) and ( d ) correspond to forward and reverse biasing , respectively .
, title="fig : " ] + + temperature dependences of voltage for fixed currents through a ni - silicide / poly - si schottky diode and temperature coefficient of signal ( voltage ) for each branch of _ i v _ characteristics : ( a ) forward and ( b ) reverse biases ( the legends represent the currents in @xmath64a for each line ) ; ( c , d ) temperature coefficient of voltage for each branch of _ i v _ characteristics vs fixed current through the structure ; to derive the graph ( d ) , the curves in the panel ( b ) were linearized in the interval from 20 to 60 ; negative and positive values of @xmath28 in panels ( c ) and ( d ) correspond to forward and reverse biasing , respectively .
, title="fig : " ] temperature dependences of voltage for fixed currents through a ni - silicide / poly - si schottky diode and temperature coefficient of signal ( voltage ) for each branch of _ i v _ characteristics : ( a ) forward and ( b ) reverse biases ( the legends represent the currents in @xmath64a for each line ) ; ( c , d ) temperature coefficient of voltage for each branch of _ i v _ characteristics vs fixed current through the structure ; to derive the graph ( d ) , the curves in the panel ( b ) were linearized in the interval from 20 to 60 ; negative and positive values of @xmath28 in panels ( c ) and ( d ) correspond to forward and reverse biasing , respectively .
, title="fig : " ] | nickel silicide schottky diodes formed on polycrystalline si@xmath0p@xmath1 films are proposed as temperature sensors of monolithic uncooled microbolometer ir focal plane arrays .
structure and composition of nickel silicide / polycrystalline silicon films synthesized in a low - temperature process are examined by means of transmission electron microscopy .
the ni silicide is identified as multi - phase compound composed by 20 to 40% of ni@xmath2si , 30 to 60% of ni@xmath3si and 10 to 30% of nisi with probable minor content of nisi@xmath3 at the silicide / poly - si interface .
rectification ratios of the schottky diodes vary from @xmath4100 to @xmath420 for the temperature increasing from 22 to 70 ; they exceed 1000 at 80k .
a barrier of @xmath5 ev is found to control the photovoltage spectra at room temperature .
a set of barriers is observed in photo - emf spectra at 80k and attributed to the ni - silicide / poly - si interface .
absolute values of temperature coefficients of voltage and current are found to vary from 0.3 to 0.6%/ for forward biasing and around 2.5%/ for reverse biasing of the diodes .
[ 1995/12/01 ] |
by applying the physical - layer network coding @xcite or analogue network coding ( anc ) @xcite in two - way relaying ( twr ) , only two time slots are required for one complete information exchange using twr . in the first time slot ,
both source nodes transmit simultaneously to the relay . in the second time slot ,
the relay broadcasts the common message which is obtained by combining the received messages . since both source nodes know their own transmitted signals , each of their self - interference can be completely canceled prior to decoding .
the twr has been studied in @xcite to @xcite for the case of single source pair .
the beamformer design for af - based multiple - input - multiple - output ( mimo ) twr is studied in @xcite@xcite , in which the receive and transmit beamforming are derived separately and then combined to form the relay beamformer .
furthermore , some strategies to enhance the performance of twr can be found in @xcite to @xcite and references therein .
the non - anc - based twr for multiple source pairs is studied in @xcite to @xcite .
unlike the single source pair case , in multiple source pairs scenario , additional inter - pair interference exists between different source pairs , which degrades the twr performance . in @xcite ,
a relay network with multiple source pairs and multiple relay nodes is studied , where all sources and relay stations have multiple antennas .
the multiuser twr is proposed and studied in @xcite , where multiple source pairs are communicating via multiple relays . in @xcite @xcite , the mimo twr where multiple wireless node pairs are communicating via a single decode - and - forward ( df ) relay
is studied . due to
the presence of inter - pair interference , previous beamforming solutions for twr with single source pair is no longer useful and new solutions are required for the case of multiple source pairs . in almost all the above works @xcite to @xcite , the inter - pair interference are canceled using zero - forcing ( zf)-based approach . in this paper ,
two new non - zf - based beamforming schemes or beamformers are proposed for anc - based twr . instead of completely canceling the inter - pair interference for all source pairs by zf - based methods as was done in previous works ,
we propose joint grouping and beamforming scheme that divides a given large number of source pairs into smaller subgroups , and then apply the proposed beamformers to each subgroup . to the best of our knowledge , this approach has not been studied in previous works .
simulation results are presented to compare the performance of the proposed schemes .
we study wireless twr with a multi - antenna relay and a total number of @xmath0 single - antenna sources , where @xmath0 is an even number . due to the fact that the inter - pair interference contains the desired signals of all the other sources ,
these desired signals are also suppressed by any suboptimal beamformer which causes significant loss in the sinr , especially when @xmath0 is large .
we propose to overcome that this shortcoming by first dividing a large number of @xmath0 source pairs into @xmath1 subgroups , each with a smaller number of @xmath2 source pairs , where @xmath2 is an even number .
then , by using time division approach , the relay performs non - zf - based beamforming on each subgroup of users at one time , and take turn to serve all source pairs , to achieve a better throughput performance .
next , we consider a given subgroup of @xmath2 sources , and derive twr beamformers for these @xmath2 sources . without loss of generality , we assume the @xmath3-th source node @xmath4 , @xmath5 , is to exchange information with another source @xmath6 , where @xmath7 , @xmath8 .
each @xmath3-th source have single antenna , whereas the relay station r is equipped with @xmath9 antennas .
let the @xmath10 vectors @xmath11 and @xmath12 denote , respectively , the channel response matrices from @xmath4 to r , and that from @xmath13 to r. we assume that the elements of @xmath11 and @xmath12 follow the distribution of circularly symmetric complex gaussian with zero mean and unity variance , which is denoted as @xmath14 .
let @xmath15 and @xmath16 denote respectively , the transmitted symbols from @xmath4 to @xmath13 , and from @xmath13 to @xmath4 .
we assume the optimal gaussian codebook is used at each @xmath4 , and therefore @xmath15 s are independent random variable each is distributed as @xmath14 .
assume tdd is used and two time slots are needed for information exchange using analogue network coding ( anc)@xcite , @xcite . in the first time slot , all active source nodes transmit their signals simultaneously , the received baseband signal vector @xmath17 at r is given by , @xmath18 where @xmath19 is symbol index , @xmath20 is the transmit power at @xmath21 , @xmath22 is the received additive noise vector , and without loss of generality , it is assume that @xmath22 follows the distribution of @xmath23 , where @xmath24 is an identity matrix . throughout this paper
, we assume that the @xmath20 s are given or fixed . at the relay , we consider af relaying using linear beamforming which is represented by a @xmath25 matrix @xmath26 .
the transmit signal @xmath27 at r can be expressed in terms of its inputs @xmath17 as @xmath28 .
we assume channel reciprocity for uplink and downlink transmission through the relay . in the second time slot , when @xmath27 is transmitted from r ,
the channels from r to @xmath21 become @xmath29 , @xmath30 .
the total transmit power at r , denoted as @xmath31 , can be shown as , @xmath32 where @xmath33 denotes the trace of @xmath34 .
the anc is adopted as follows .
we assume that using training and estimation , @xmath35 and @xmath36 are perfectly known at @xmath21 , @xmath37 prior to signal transmission .
each of the @xmath21 can first cancel its self - interference and then coherently demodulate for @xmath38 .
this yields @xmath39 for @xmath37 .
we assume that the received noise @xmath40 is distributed as @xmath41 , and @xmath22 are independent of @xmath42 .
@xmath43 , and @xmath44 is distributed as @xmath45 . at each @xmath21 , coherent signal detection can then be used to recover @xmath46 from @xmath47 .
the signal - to - interference - plus - noise ratio ( sinr ) for the @xmath3-th destination node .
@xmath37 , can be expressed as @xmath48 +
we define the uplink ( ul ) channel gain matrix @xmath49 $ ] and denote the singular value decomposition ( svd ) of @xmath50 as @xmath51 , where the @xmath52 matrix @xmath53 , and the @xmath54 matrix @xmath55 , are with orthogonal column vectors , and @xmath56 is singular value matrix with dimension @xmath57 where @xmath58_{\ell,\ell } = \sigma_{\ell}$ ] for @xmath59 .
we have the following new result . _
proposition : _ the optimal beamforming matrix to achieve maximum sinr in ( [ eq : sinrk ] ) has the following structure : @xmath60 where @xmath61 is a @xmath54 matrix .
_ proof : _ the above new result can be proven by extending the proof of @xcite@xcite , which considers the twr for the case of single source pair who are exchanging information . for the case of multiple source pairs ,
each source is also subject to interference from the other source pairs .
this so - called inter - pair interference term also depends only on @xmath62 as @xmath63 , which also spans the total signal subspace of @xmath62 .
therefore , we obtain @xmath61 .
the beamforming matrix @xmath62 can be solved as shown next .
+ let @xmath64 , @xmath37 , represent the effective channel from @xmath21 to r , with @xmath65 given in ( [ eq : optimalbaf ] ) .
similarly , let @xmath66 represent the effective channel from r to @xmath21 .
the sinr formula in ( [ eq : sinrk ] ) can be written in terms of @xmath67 . throughout this paper ,
the optimization problems are formulated in terms of the beamforming matrix @xmath62 and the effective channels @xmath68 . the minimum power ( mp ) beamformer
is derived by minimizing the total relay transmit power with respect to the relay beamforming matrix @xmath67 ( or equivalently , @xmath65 ) , subject to sinr constraints @xmath69 , @xmath30 , @xmath70 for given transmit powers @xmath71 , @xmath30 , we can use the same approach as in @xcite , to develop an efficient algorithm based on the second - order cone programming ( socp ) @xcite to solve the problem in .
we derive the suboptimal minimum interference ( mi ) beamformer that does not require computations via optimization technique .
the tradeoff between the desired signals and interference is taken into account by minimizing the sum of inter - pair interference plus af noise at the output of the beamformer , subject to the constraints that the desired signal gain for each @xmath3-th receiver is equal to a constant @xmath72 .
the additive gaussian noise @xmath73 at each @xmath3-th source , which is not affected by the beamforming matrix @xmath67 , and is neglected in the optimization . again in this case , the joint design of both receive and transmit beamforming is considered .
the interference minimization problem is formulated as , @xmath74 where @xmath75 the above constraints in ( [ eq : minimizeinterference ] ) are introduced to preserve the desired signal components at each source , so as to minimize the inter - pair interference plus af noise component that point into any undesired direction . in ( [ eq : interference ] ) , @xmath76 represents the sum of inter - pair interference ( from other source pairs ) plus af noise that is imposed on @xmath21 . to solve for @xmath67 , the ul and dl channel response of each @xmath3-th source
is assumed to be known at the relay .
we assume that the @xmath20 s are given , and write @xmath77 , @xmath78 , and @xmath79 , where @xmath80 , and @xmath81 denotes a @xmath82 equivalent beamforming weights vector which is generated by the rule of ( [ eqn : vector ] ) as , @xmath83.\ ] ] the problem ( [ eq : minimizeinterference ] ) can be written in terms of @xmath81 as , @xmath84 we further define @xmath85 , @xmath86 , where both @xmath87 , and @xmath88 are @xmath89 matrices . by writing @xmath90 , @xmath91 , and @xmath92 , the problem ( [ eq : minimizeinterference2 ] )
can be written as , @xmath93 where @xmath94 $ ] is the @xmath95 constraint matrix , and @xmath96 has been defined in the paragraph after eq.([eqn : vector ] ) .
each of the column of @xmath97 imposes a constraint on the equivalent beamforming weight vector @xmath81 .
the @xmath98 response vector @xmath99^t$ ] contains the k scalar constraints values @xmath100 , which are chosen to satisfy the sinr requirements of each source .
the beamforming vector solution , denoted as @xmath101 , that satisfies ( [ eq : lcmv ] ) is solved as @xcite , @xmath102 the mi relay beamformer , denoted as @xmath103 , can then be obtained from @xmath101 as @xmath104 , where @xmath105 denotes the inverse operation of @xmath106 defined in .
it can be shown that the solution @xmath103 maximizes the sinr of each source , where the constant @xmath107 is used to control the total relay power @xmath31 in ( [ eq : relay power ] ) .
the iterative steps to search for @xmath107 are presented next .
we propose the following joint grouping and beamforming scheme that divides a given large number of source pairs into smaller subgroups , and then apply the above beamformers to each subgroup . for simplicity ,
here we consider that the grouping is done arbitrarily . by doing
so we can reduce the feedback of channel state information and beamforming calculation during each relaying .
* divide the total number of @xmath0 source pairs into @xmath1 subgroups , @xmath108 , and let @xmath109 denote the number of source in each subgroup . * for each @xmath19-th subgroup , @xmath110 , compute the mi beamforming matrix as follows : * * given * @xmath111 $ ] , @xmath112^{t}$ ] .
* * initialize * @xmath113 , @xmath114 . *
compute the svd of the ul channel matrix @xmath115 .
* estimate the correlation matrix @xmath116 .
* compute the constraint matrix @xmath97 and the response vector @xmath117 . *
compute the beamforming weights using @xmath118 @xmath119 . * obtain the mi relay beamforming matrix by arranging the beamforming weights of @xmath101 into the matrix form , @xmath120 . *
* repeat * * * set @xmath121 . *
* update @xmath107 by the bisection method @xcite : if @xmath31 is less than a given power constraint , set @xmath122 ; otherwise , @xmath123 . * * until * @xmath124 , where the small positive constant @xmath125 is chosen to ensure sufficient accuracy . * compute the mi relay beamforming matrix as @xmath126 for each @xmath19-th subgroup , @xmath110 . *
select the number of subgroup @xmath1 and its corresponding @xmath127 which results in the largest achievable sum - rate .
in this section , the sum rate performance of the proposed twr beamformers are presented . for simplicity ,
the channel correlations between the @xmath128-th and the @xmath3-th sources are set to be equal , that is , @xmath129 , @xmath130 . in fig .
[ fig : rate comp no cor ] , two pairs of single - antenna source ( @xmath131 ) and a multi - antenna relay ( @xmath132 ) are considered in the simulations .
the transmit power at each source @xmath133 , @xmath134 , @xmath135 and @xmath136 are fixed as @xmath137 , and the relay transmit power is set to be @xmath138 . both sources within each pair
are set to have identical sinr requirement . with no channel correlation between different sources ,
the proposed mi beamformer achieves the same achievable rate region as the optimal mp beamforming scheme .
however , when the channel correlations between different source pairs , and within each source pair , are set to be higher , the proposed mi beamformer achieves a smaller rate region ( not shown here due to space constraint ) . this is because when
the channels of different source pairs are correlated , the desired signal components that point in the direction of the inter - pair interference will also be suppressed by the mi beamformer , which in turn reduces the achievable rate for each source . , @xmath131 , fixed @xmath139 , @xmath140 , and @xmath141 . ] , @xmath142 , @xmath141 . ] in fig .
[ fig : sum - rate no cor ] , we present the achievable sum - rate versus snr at the relay , which is defined as @xmath143 , for the proposed mi beamformer with various number of subgroups .
four pairs of single - antenna source ( @xmath142 ) and a multi - antenna relay ( @xmath132 ) are considered , and the channels between different sources are uncorrelated .
time division is used to serve different subgroup .
we observe that either the smaller subgroups or large group transmission does not always perform the best in twr . with low snr at the relay ,
the case of four groups ( each with one source pair ) performs the best , whereas for large snr , the case of a single group ( with four source pairs ) performs the best .
this is because with either large number of source pairs or small relay transmit power , the mi beamformer tends to suppress the interference more for the case of single group ( with more interfering users ) , which results in sinr loss . to overcome this shortcoming ,
joint grouping and beamforming scheme can be used to reduce sinr loss as follows . for small snr
, we should apply the proposed beamformers to four different subgroups , and for large snr , we should apply the proposed beamformers to a single group ( with four source pairs ) . the improved sum - rate performance by joint grouping and mi beamforming scheme is highlighted by the solid line in fig .
[ fig : sum - rate no cor ] .
the optimal grouping and selection of the snr thresholds correspond to different number of subgroups are interesting subjects for further study .
new optimal and suboptimal beamformers for anc - based twr with multiple source pairs are derived by taking into account the tradeoff between the desired signals and the inter - pair interference . for low snr
, a better sum - rate performance can be achieved by first diving a large number of source pairs into smaller subgroups , and then apply beamforming to each subgroup using time division .
this research is partly supported by the singapore university technology and design ( grant no .
srg - epd-2010 - 005 ) .
s. katti , s. gollakota , and d. katabi , embracing wireless interference : analog network coding " , _ computer science and artificial intelligence laboratory technical report _ , mit - csail - tr-2007 - 012 , feb .
23 , 2007 .
i. hammerstr@xmath144 m , m. kuhn , c. esli , j. zhao , a. wittneben , and g. bauch , mimo two - way relaying with transmit csi at the relay " , in _ proc .
ieee signal proc . adv .
wireless comm .
( spaw ) _ , jun .
t. unger and a. klein , linear transceive filters for relay stations with multiple antennas in the two - way relay channel " , _
16th ist mobile and wireless communications summit _ , budapest , hungary , jul .
2007 .
r. f. wyrembelski , t. j. oechtering , i. bjelakovi@xmath145 .
, c. schnurr , and h. boche , capacity of gaussian mimo bidirectional broadcast channels " , in _ proc .
inf . theory ( isit ) _ , jul .
2008 .
r. zhang , y. c. liang , c. c. chai , and s. cui , optimal beamforming for two - way multi - antenna relay channel with analogue network coding " , _ ieee journal on selected areas in communication _
5 , october 2009 , pp .
3988 - 3999 .
l. song , y. li , h. peng , b. jiao , and a. v. vasilakos , differential modulation for bidirectional relaying with analog network coding " , _ ieee transactions on signal processing _
58 , no . 7 , pp .
3933 - 3938 , jun . 2010 .
l. song , g. hong , b. jiao , and m. debbah , joint relay selection and analog network coding using differential modulation in two - way relay channels " , _ ieee transactions on vehicular technology _ ,
59 , no . 6 , jul .
t. abe , h. shi , t. asai and h. yoshino ,
relay techniques for mimo wireless networks with multiple source and destination pairs " , in _ eurasip journal on wireless communications and networking _ , pp
. 1 - 9 , 2006 . | we study amplified - and - forward ( af)-based two - way relaying ( twr ) with multiple source pairs , which are exchanging information through the relay .
each source has single antenna and the relay has multi - antenna .
the optimal beamforming matrix structure that achieves maximum signal - to - interference - plus - noise ratio ( sinr ) for twr with multiple source pairs is derived .
we then present two new non - zero - forcing based beamforming schemes for twr , which take into consideration the tradeoff between preserving the desired signals and suppressing inter - pair interference between different source pairs .
joint grouping and beamforming scheme is proposed to achieve a better signal - to - interference - plus - noise ratio ( sinr ) when the total number of source pairs is large and the signal - to - noise ratio ( snr ) at the relay is low .
analogue network coding ( anc ) , two - way relaying ( twr ) , multiple source pairs , information exchange , analogue relaying , optimal beamforming .
+ [ section ] [ section ] [ section ] [ section ] [ section ] |
the solution of inverse problems provides a rich source of applications of the bayesian nonparametric methodology .
it encompasses a broad range of applications from partial differential equations ( pdes ) @xcite , where there is a well - developed theory of classical , non - statistical , regularization @xcite . on the other hand ,
the area of nonparametric bayesian statistical estimation and in particular the problem of posterior consistency has attracted a lot of interest in recent years ; see for instance @xcite . despite this
, the formulation of many of these pde inverse problems using the bayesian approach is in its infancy @xcite .
furthermore , the development of a theory of bayesian posterior consistency , analogous to the theory for classical regularization , is under - developed with the primary contribution being the recent paper @xcite .
this recent paper provides a roadmap for what is to be expected regarding bayesian posterior consistency , but is limited in terms of applicability by the assumption of simultaneous diagonalizability of the three linear operators required to define bayesian inversion . our aim in this paper is to make a significant step in the theory of bayesian posterior consistency for linear inverse problems by developing a methodology which sidesteps the need for simultaneous diagonalizability .
the central idea underlying the analysis is to work with precision operators rather than covariance operators , and thereby to enable use of powerful tools from pde theory to facilitate the analysis .
let @xmath0 be a separable hilbert space , with norm @xmath1 and inner product @xmath2 , and let @xmath3 be a known self - adjoint and positive - definite linear operator with bounded inverse .
we consider the inverse problem to find @xmath4 from @xmath5 , where @xmath5 is a noisy observation of @xmath6 .
we assume the model , @xmath7 where @xmath8 is an additive noise .
we will be particularly interested in the small noise limit where @xmath9 .
a popular method in the deterministic approach to inverse problems is the generalized tikhonov - phillips regularization method in which @xmath4 is approximated by the minimizer of a regularized least squares functional : define the tikhonov - phillips functional @xmath10 where @xmath11 are bounded , possibly compact , self - adjoint positive - definite linear operators .
the parameter @xmath12 is called the regularization parameter , and in the classical non - probabilistic approach the general practice is to choose it as an appropriate function of the noise size @xmath13 , which shrinks to zero as @xmath9 , in order to recover the unknown parameter @xmath4 @xcite . in this paper
we adopt a bayesian approach for the solution of problem ( [ eq : int1 ] ) , which will be linked to the minimization of @xmath14 via the posterior mean .
we assume that the prior distribution is gaussian , @xmath15 , where @xmath16 and @xmath17 is a self - adjoint , positive - definite , trace class , linear operator on @xmath0 .
we also assume that the noise is gaussian , @xmath18 , where @xmath19 is a self - adjoint positive - definite , bounded , but not necessarily trace class , linear operator ; this allows us to include the case of white observational noise .
we assume that the , generally unbounded , operators @xmath20 and @xmath21 have been maximally extended to self - adjoint positive - definite operators on appropriate domains .
the unknown parameter and the noise are considered to be independent , thus the conditional distribution of the observation given the unknown parameter @xmath4 ( termed the likelihood ) is also gaussian with distribution @xmath22 define @xmath23 and let @xmath24 in finite dimensions the probability density of the posterior distribution , that is , the distribution of the unknown given the observation , with respect to the lebesgue measure is proportional to @xmath25 this suggests that , in the infinite - dimensional setting , the posterior is gaussian @xmath26 , where we can identify the posterior covariance and mean by the equations @xmath27 and @xmath28obtained by completing the square .
we present a method of justifying these expressions in section [ sec : justification ] .
we define @xmath29 and observe that the dependence of @xmath30 on @xmath31 and @xmath32 is only through @xmath12 . since @xmath33 the posterior mean also depends only on @xmath12 : @xmath34 .
this is not the case for the posterior covariance @xmath35 , since it depends on @xmath31 and @xmath32 separately : @xmath36 . in the following ,
we suppress the dependence of the posterior covariance on @xmath12 and @xmath31 and we denote it by @xmath35 .
observe that the posterior mean is the minimizer of the functional @xmath37 , hence also of @xmath38 that is , the posterior mean is the tikhonov - phillips regularized approximate solution of problem ( [ eq : int1 ] ) , for the functional @xmath14 with @xmath39 . in @xcite and @xcite ,
formulae for the posterior covariance and mean are identified in the infinite - dimensional setting , which avoid using any of the inverses of the prior , posterior or noise covariance operators .
they obtain @xmath40and @xmath41 which are consistent with formulae ( [ eq : int4 ] ) and ( [ eq : int7 ] ) for the finite - dimensional case . in @xcite this
is done only for @xmath19 of trace class while in @xcite the case of white observational noise was included .
we will work in an infinite - dimensional setting where the formulae ( [ eq : int4 ] ) , ( [ eq : int7 ] ) for the posterior covariance and mean can be justified .
working with the unbounded operator @xmath30 opens the possibility of using tools of analysis , and also numerical analysis , familiar from the theory of partial differential equations . in our analysis
we always assume that @xmath20 is regularizing , that is , we assume that @xmath20 dominates @xmath30 in the sense that it induces stronger norms than @xmath42 this is a reasonable assumption since otherwise we would have @xmath43 ( here @xmath44 is used loosely to indicate two operators which induce equivalent norms ; we will make this notion precise in due course ) .
this would imply that the posterior mean is @xmath45 , meaning that we attempt to invert the data by applying the , generally discontinuous , operator @xmath46 ( * ? ? ?
* proposition 2.7 ) .
we study the consistency of the posterior @xmath47 in the frequentist setting . to this end
, we consider data @xmath48 which is a realization of @xmath49 where @xmath50 is a fixed element of @xmath0 ; that is , we consider observations which are perturbations of the image of a fixed true solution @xmath50 by an additive noise @xmath51 , scaled by @xmath52 . since the posterior depends through its mean on the data and also through its covariance operator on the scaling of the noise and the prior , this choice of data model gives as posterior distribution the gaussian measure @xmath53 , where @xmath35 is given by ( [ eq : int4 ] ) and @xmath54 we study the behavior of the posterior @xmath55 as the noise disappears ( @xmath9 ) .
our aim is to show that it contracts to a dirac measure centered on the fixed true solution @xmath50 . in particular
, we aim to determine @xmath56 such that @xmath57 where the expectation is with respect to the random variable @xmath58 distributed according to the data likelihood @xmath59 . as in the deterministic theory of inverse problems , in order to get convergence in the small noise limit , we let the regularization disappear in a carefully chosen way , that is , we will choose @xmath60 such that @xmath61 as @xmath9 . the assumption that @xmath20 dominates @xmath30
, shows that @xmath30 is a singularly perturbed unbounded ( usually differential ) operator , with an inverse which blows - up in the limit @xmath61 .
this together with equation ( [ eq : int7 ] ) , opens up the possibility of using the analysis of such singular limits to study posterior contraction : on the one hand , as @xmath61 , @xmath62 becomes unbounded ; on the other hand , as @xmath9 , we have more accurate data , suggesting that for the appropriate choice of @xmath60 we can get @xmath63 . in particular , we will choose @xmath32 as a function of the scaling of the noise , @xmath64 under the restriction that the induced choice of @xmath65 , is such that @xmath61 as @xmath9 .
the last choice will be made in a way which optimizes the rate of posterior contraction @xmath56 , defined in ( [ eq : main1 ] ) . in general
there are three possible asymptotic behaviors of the scaling of the prior @xmath66 as @xmath9 , @xcite : 1 .
@xmath67 ; we increase the prior spread , if we know that draws from the prior are more regular than @xmath50 ; 2 .
@xmath66 fixed ; draws from the prior have the same regularity as @xmath50 ; 3 .
@xmath68 at a rate slower than @xmath69 ; we shrink the prior spread , when we know that draws from the prior are less regular than @xmath70 the problem of posterior contraction in this context is also investigated in @xcite and @xcite . in @xcite , sharp convergence rates
are obtained in the case where @xmath71 and @xmath72 are simultaneously diagonalizable , with eigenvalues decaying algebraically , and in particular @xmath73 , that is , the data are polluted by white noise . in this paper
we relax the assumptions on the relations between the operators @xmath71 and @xmath72 , by assuming that appropriate powers of them induce comparable norms ( see section [ sec : assumptions ] ) . in @xcite
, the non - diagonal case is also examined ; the three operators involved are related through domain inclusion assumptions .
the assumptions made in @xcite can be quite restrictive in practice ; our assumptions include settings not covered in @xcite , and in particular the case of white observational noise . in the following section we present our main results which concern the identification of the posterior ( theorem [ justth ] ) and the posterior contraction ( theorems [ pdecor1 ] and [ pdecor2 ] ) . in section [ sec : assumptions ] we present our assumptions and their implications .
the proofs of the main results are built in a series of intermediate results contained in sections [ sec : prmean]-[sec : main ] . in section [ sec : prmean ] , we reformulate equation ( [ eq : int7 ] ) as a weak equation in an infinite - dimensional space . in section [ sec :
justification ] , we present a new method of identifying the posterior distribution : we first characterize it through its radon - nikodym derivative with respect to the prior ( theorem [ prop0 ] ) and then justify the formulae ( [ eq : int4 ] ) , ( [ eq : int7 ] ) for the posterior covariance and mean ( proof of theorem [ justth ] ) . in section [ sec : normbounds ] , we present operator norm bounds for @xmath62 in terms of the singular parameter @xmath12 , which are the key to the posterior contraction results contained in section [ sec : main ] and their corollaries in section [ sec : mainresults ] ( theorems [ pdeth1 ] , [ pdeth2 ] and [ pdecor1 ] , [ pdecor2 ] ) . in section [ sec : ex ] , we present some nontrivial examples satisfying our assumptions and provide the corresponding rates of convergence . in section [ sec : diag ] , we compare our results to known minimax rates of convergence in the case where @xmath71 and @xmath72 are all diagonalizable in the same eigenbasis and have eigenvalues that decay algebraically .
finally , section [ sec : conclusions ] is a short conclusion .
the entire paper rests on a rich set of connections between the theory of stochastic processes and various aspects of the theory of linear partial differential equations .
in particular , since the green s function of the precision operator of a gaussian measure corresponds to its covariance function , our formulation and analysis of the inverse problem via precision operators is very natural . furthermore , estimates on the inverse of singular limits of these precisions , which have direct implications for localization of the green s functions , play a key role in the analysis of posterior consistency .
in this section we present our main results . we postpone the rigorous presentation of our assumptions to the next section and the proofs and technical lemmas
are presented together with intermediate results of independent interest in sections [ sec : prmean ] - [ sec : main ] . recall that we assume a gaussian prior @xmath74 and a gaussian noise distribution @xmath75 .
our first assumption concerns the decay of the eigenvalues of the prior covariance operator and enables us to quantify the regularity of draws from the prior .
this is encoded in the parameter @xmath76 ; smaller @xmath77 implies more regular draws from the prior .
we also assume that @xmath78 and @xmath79 , for some @xmath80 , where @xmath44 is s used in the manner outlined in section [ sec : intro ] , and defined in detail in section [ sec : assumptions ] .
finally , we assume that the problem is sufficiently ill - posed with respect to the prior .
this is quantified by the parameter @xmath81
which we assume to be larger than @xmath82 ; for a fixed prior , the larger @xmath83 is , the more ill - posed the problem .
our first main theorem identifies the posterior measure as gaussian and justifies formulae [ eq : int4 ] and [ eq : int7 ] .
this reformulation of the posterior in terms of the precision operator is key to our method of analysis of posterior consistency and opens the route to using methods from the study of partial differential equations ( pdes ) .
these methods will also be useful for the development of numerical methods for the inverse problem .
[ justth ] under the assumptions [ a2 ] , the posterior measure @xmath84 is gaussian @xmath26 , where @xmath35 is given by ( [ eq : int4 ] ) and @xmath85 is a weak solution of ( [ eq : int7 ] ) .
we now present our results concerning frequentist posterior consistency of the bayesian solution to the inverse problem .
we assume to have data @xmath86 as in ( [ eq : int10 ] ) , and examine the behavior of the posterior @xmath87 , where @xmath88 is given by ( [ eq : int11 ] ) , as the noise disappears ( @xmath9 ) .
the first convergence result concerns the convergence of the posterior mean @xmath88 to the true solution @xmath50 in a range of weighted norms @xmath89 induced by powers of the prior covariance operator @xmath17 .
the spaces @xmath90 are rigorously defined in the following section .
the second result provides rates of posterior contraction of the posterior measure to a dirac centered on the true solution as described in ( [ eq : main1 ] ) . in both results ,
we assume a priori known regularity of the true solution @xmath91 and give the convergence rates as functions of @xmath92 .
[ pdecor1 ] assume @xmath91 , where @xmath93 and let @xmath94 , where @xmath95 $ ] . under the assumptions [ a2 ]
, we have the following optimized rates of convergence , where @xmath96 is arbitrarily small : 1 . if @xmath97,$ ] for @xmath98 @xmath99 2 . if @xmath100 , for @xmath101 @xmath102 3 .
if @xmath103 and @xmath104 for @xmath105 @xmath106if @xmath103 and @xmath107 then the method does not give convergence .
[ pdecor2]assume @xmath91 , where @xmath93 . under the assumptions [ a2 ] , we have the following optimized rates for the convergence in ( [ eq : main1 ] ) , where @xmath96 is arbitrarily small : 1 . if @xmath108 $ ] for @xmath98 @xmath109 2 .
if @xmath100 for @xmath101 + @xmath110 to summarize , provided the problem is sufficiently ill - posed and the true solution @xmath50 is sufficiently regular we get the convergence in ( [ eq : main1 ] ) for @xmath111 . in blue are the sharp convergence rates obtained in the diagonal case in @xcite , while in green the rates predicted by our method , which applies to the more general non - diagonal case , width=340 ] our rates of convergence agree , up to @xmath96 arbitrarily small , with the sharp convergence rates obtained in the diagonal case in @xcite across a wide range of regularity assumptions on the true solution ( figure [ fig ] ) ; yet , our rates cover a much more applicable range of non - simultaneously diagonalizable problems .
( the reason for the appearance of @xmath112 is that in the assumed non - diagonal setting we can only use information about the regularity of the noise as expressed in terms of the spaces @xmath113 ( cf .
lemma [ l2 ] ) , rather than the explicit representation of the noise . )
the rates we obtain are not as strong as in the simultaneously diagonalizable case when the true solution is too regular ; in particular our rates saturate earlier as a function of increasing regularity , and we require a certain degree of regularity of the true solution in order to secure convergence . it is not known if our results can be improved but it would be interesting to try . both of the two discrepancies are attributed to the fact that our method relies on interpolating between rates in a strong and a weak norm of the error @xmath114 ; on the one hand the rate of the error in the weak norm saturates earlier , and on the other hand the error in the strong norm requires additional regularity in order to converge ( cf . section [ sec : diag ] ) .
in this section we present the setting in which we formulate our results .
first , we define the spaces in which we work , in particular , we define the hilbert scale induced by the prior covariance operator @xmath17 . then we define the probability measures relevant to our analysis . furthermore , we state our main assumptions , which concern the decay of the eigenvalues of @xmath17 and the connections between the operators @xmath17 , @xmath19 and @xmath115 , and present regularity results for draws from the prior , @xmath116 , and the noise distribution , @xmath75
. finally we briefly overview the way in which the hilbert scale defined in terms of the prior covariance operator @xmath17 , which is natural for our analysis , links to scales of spaces defined independently of any prior model .
we start by defining the hilbert scale which we will use in our analysis .
recall that @xmath0 is an infinite - dimensional separable hilbert space and @xmath117 is a self - adjoint , positive - definite , trace class , linear operator . since @xmath118 is injective and self - adjoint we have that @xmath119 .
this means that @xmath120 is a densely defined , unbounded , symmetric , positive - definite , linear operator in @xmath0 .
hence it can be extended to a self - adjoint operator with domain @xmath121 ; this is the friedrichs extension @xcite .
thus , we can define the hilbert scale @xmath122 , with @xmath123 @xcite , where@xmath124 the bounded linear operator @xmath125 is assumed to be self - adjoint , positive - definite ( but not necessarily trace class ) ; thus @xmath126 can be extended in the same way to a self - adjoint operator with domain @xmath127 finally , recall that we assume that @xmath128 is a self - adjoint and positive - definite , linear operator with bounded inverse , @xmath129 .
we assume that we have a probability space @xmath130 .
the expected value is denoted by @xmath131 and @xmath132 means that the law of the random variable @xmath51 is the measure @xmath133 .
let @xmath134 and @xmath135 be the prior and noise distributions respectively .
furthermore , let @xmath136 denote the measure constructed by taking @xmath4 and @xmath137 as independent gaussian random variables @xmath138 and @xmath139 respectively : @xmath140 where @xmath141 .
we denote by @xmath142 the measure constructed by taking @xmath4 and @xmath5 as independent gaussian random variables @xmath138 and @xmath143 respectively : @xmath144 let @xmath145 be orthonormal eigenpairs of @xmath17 in @xmath0 .
thus , @xmath146 are the singular values and @xmath147 an orthonormal eigenbasis . since @xmath17 is trace class we have that @xmath148 .
in fact we require a slightly stronger assumption see assumption [ a2]([a1 ] ) below .
we are now ready to present our assumptions .
the first assumption enables us to quantify the regularity of draws from the prior whereas the rest of the assumptions regard interrelations between the three operators @xmath17 , @xmath19 and @xmath72 ; these assumptions reflect the idea that @xmath149 for some @xmath150 , where @xmath44 is used in the same manner as in section [ sec : intro ] .
this is made precise by the inequalities presented in the following assumption , where the notation @xmath151 means that there exist constants @xmath152 such that @xmath153 .
[ a2]suppose there exist @xmath76 , @xmath154 and constants @xmath155 such that 1 .
[ a1]@xmath156 is trace class for all @xmath157 ; 2 .
[ a2delta ] @xmath158 , where @xmath159 3 .
[ a2i]@xmath160 4 .
[ a2ii ] @xmath161 5 .
[ a2iii]@xmath162;$ ] 6 .
[ a2iv]@xmath163;$ ] 7 .
[ a2v]@xmath164.$ ] notice that , by assumption [ a2]([a2delta ] ) we have @xmath165 which , in combination with assumption [ a2]([a2i ] ) , implies that @xmath166 capturing the idea that the regularization through @xmath17 is indeed a regularization .
in fact the assumption @xmath167 connects the ill - posedness of the problem to the regularity of the prior .
we exhibit this connection in the following example : assume @xmath168 and @xmath17 are simultaneously diagonalizable , with eigenvalues having algebraic decay @xmath169 , @xmath170 and @xmath171 , respectively , for @xmath172 and @xmath173 so that @xmath17 is trace class
. then assumptions ( [ a1]),([a2i])-([a2v ] ) are trivially satisfied with @xmath174 @xmath175 and @xmath176 .
the assumption ( [ a2delta ] ) @xmath167 is then equivalent to @xmath177 .
that is , for a certain degree of ill - posedness ( encoded in the difference @xmath178 ) we have a minimum requirement on the regularity of the prior ( encoded in @xmath179 ) .
put differently , for a certain prior , we require a minimum degree of ill - posedness .
we refer the reader to section [ sec : ex ] for nontrivial examples satisfying assumptions [ a2 ] . in the following ,
we exploit the regularity properties of a white noise to determine the regularity of draws from the prior and the noise distributions using assumption [ a2]([a1 ] ) .
we consider a white noise to be a draw from @xmath180 , that is a random variable @xmath181 . even though the identity operator is not trace class in @xmath0 , it is trace class in a bigger space @xmath182 , where @xmath183 is sufficiently large .
[ l1]under the assumption [ a2]([a1 ] ) we have : 1 . let @xmath184 be a white noise .
then @xmath185 for all @xmath157 .
2 . let @xmath186 .
then @xmath187 @xmath188-a.s .
for every @xmath157 . 1 .
we have that @xmath189 , thus @xmath190 is equivalent to @xmath156 being of trace class . by the assumption
@xmath191 it suffices to have @xmath157 .
we have @xmath192 , where @xmath184 is a white noise , therefore using part ( i ) we get the result . [ remark1 ]
note that as @xmath77 changes , both the hilbert scale and the decay of the coefficients of a draw from @xmath116 change .
the norms @xmath193 are defined through powers of the eigenvalues @xmath194 . if @xmath195 , then @xmath17 has eigenvalues that decay like @xmath196 , thus an element @xmath197 has coefficients @xmath198 , that decay faster than @xmath199 .
as @xmath77 gets closer to zero , the space @xmath200 for fixed @xmath201 , corresponds to a faster decay rate of the coefficients . at the same time , by the last lemma , draws from @xmath202 belong to @xmath203 for all @xmath157 .
consequently , as @xmath77 gets smaller , not only do draws from @xmath116 belong to @xmath203 for smaller @xmath204 , but also the spaces @xmath203 for fixed @xmath204 reflect faster decay rates of the coefficients .
the case @xmath205 corresponds to @xmath17 having eigenvalues that decay faster than any negative power of @xmath206 .
a draw from @xmath116 in that case has coefficients that decay faster than any negative power of @xmath206 . in the next lemma
, we use the interrelations between the operators @xmath207 to obtain additional regularity properties of draws from the prior , and also determine the regularity of draws from the noise distribution and the joint distribution of the unknown and the data .
[ l2 ] under the assumptions [ a2 ] we have : 1 . @xmath208-a.s .
for all @xmath209 2 .
@xmath210-a.s.@xmath211 3 . @xmath212-a.s .
for all @xmath213 4 .
@xmath214 @xmath215-a.s . for all @xmath216 . 1 .
we can choose an @xmath112 as in the statement by the assumption [ a2]([a2delta ] ) .
by lemma [ l1](ii ) , it suffices to show that @xmath217 .
indeed , @xmath218 2 . under assumption [ a2]([a2i ] )
it suffices to show that @xmath219 .
indeed , by lemma [ l1](ii ) , we need to show that @xmath220 , which is true since @xmath76 and we assume @xmath221 , thus @xmath222 3 . it suffices to show it for any @xmath223 . noting that @xmath224 is a white noise , using assumption [ a2]([a2ii ] )
, we have by lemma [ l1](i ) @xmath225 since @xmath226 .
4 . by ( ii )
we have that @xmath6 is @xmath116-a.s . in the cameron - martin space of the gaussian measures @xmath227 and @xmath228 , thus the measures @xmath227 and @xmath228 are @xmath116-a.s . equivalent ( * ? ? ?
* theorem 2.8 ) and ( iii ) gives the result . the theory is naturally developed in the scale of hilbert spaces defined via the prior .
however application of the theory may be more natural in a different functional setting .
we explain how the two may be connected .
let @xmath229 be an orthonormal basis of the separable hilbert space @xmath0 .
we define the spaces @xmath230 as follows : for @xmath201 we set @xmath231 and the spaces @xmath232 are defined by duality , @xmath233 for example , if we restrict ourselves to functions on a periodic domain @xmath234^d$ ] and assume that @xmath229 is the fourier basis of @xmath235 , then the spaces @xmath236 can be identified with the sobolev spaces of periodic functions @xmath237 , by rescaling : @xmath238 ( * ? ? ?
* proposition 5.39 ) . in the case
@xmath195 , as explained in remark [ remark1 ] we have algebraic decay of the eigenvalues of @xmath17 and in particular @xmath194 decay like @xmath196 . if @xmath17 is diagonalizable in the basis @xmath229 , that is , if @xmath239 , then it is straightforward to identify the spaces @xmath200 with the spaces @xmath240 .
the advantage of this identification is that the spaces @xmath241 do not depend on the prior so one can use them as a fixed reference point for expressing regularity , for example of the true solution . in our subsequent analysis
, we will require that the true solution lives in the cameron - martin space of the prior @xmath242 , which in different choices of the prior ( different @xmath77 ) is a different space .
furthermore , we will assume that the true solution lives in @xmath243 for some @xmath93 and provide the convergence rate depending on the parameters @xmath244 .
the identification @xmath245 and the intuitive relation between the spaces @xmath236 and the sobolev spaces , enable us to understand the meaning of the assumptions on the true solution .
we can now formulate the following guidelines for applying the theory presented in the present paper : we work in a separable hilbert space @xmath0 with an orthonormal basis @xmath229 and we have some prior knowledge about the true solution @xmath50 which can be expressed in terms of the spaces @xmath236 . the noise is assumed to be gaussian @xmath75 , and the forward operator is known ; that is , @xmath19 and @xmath72 are known .
we choose the prior @xmath246 , that is , we choose the covariance operator @xmath17 , and we can determine the value of @xmath77 .
if the operator @xmath17 is chosen to be diagonal in the basis @xmath229 then we can find the regularity of the true solution in terms of the spaces @xmath200 , that is , the value of @xmath92 such that @xmath91 , and check that @xmath93 which is necessary for our theory to work .
we then find the values of @xmath247 and @xmath248 and calculate the value of @xmath83 appearing in assumption [ a2 ] , checking that our choice of the prior is such that @xmath167 .
we now have all the necessary information required for applying the theorems [ pdecor1 ] and [ pdecor2 ] presented in section [ sec : mainresults ] to get the rate of convergence .
observe that in the above mentioned example of periodic functions , we have the identification @xmath249 , thus since @xmath250 we have that the assumption @xmath251 implies that @xmath252 , for @xmath253 . by the sobolev embedding theorem ( * ? ? ?
* theorem 5.31 ) , this implies that the true solution is always assumed to be continuous .
however , this is not a disadvantage of our method , since in many cases a gaussian measure which charges @xmath254 with probability one , can be shown to also charge the space of continuous functions with probability one ( * ? ? ?
* lemma 6.25 )
we now make sense of the equation @xmath255 weakly in the space @xmath242 , under the assumptions presented in the previous section . to do so , we define the operator @xmath30 from ( [ eq : int6 ] ) in @xmath242 and examine its properties . in section [ sec : justification ]
we demonstrate that ( [ eq : int4 ] ) and ( [ eq : int7 ] ) do indeed correspond to the posterior covariance and mean .
consider the equation @xmath256 where @xmath257 define the bilinear form @xmath258 , @xmath259 [ weak ] let @xmath260 .
an element @xmath261 is called a weak solution of ( [ eq : pm1 ] ) , if @xmath262 [ laxmil ] under the assumptions [ a2]([a2delta ] ) and ( [ a2i ] ) , for any @xmath260 , there exists a unique weak solution @xmath261 of ( [ eq : pm1 ] ) .
we use the lax - milgram theorem in the hilbert space @xmath242 , since @xmath263 1 .
@xmath264 is coercive : @xmath265 2 .
@xmath264 is continuous : indeed by the cauchy - schwarz inequality and the assumptions [ a2]([a2delta ] ) and ( [ a2i ] ) , @xmath266 [ yosida ] the lax - milgram theorem defines a bounded operator @xmath267 , such that @xmath268 for all @xmath269 , which has a bounded inverse @xmath270 such that @xmath271 for all @xmath269 .
henceforward , we identify @xmath272 and @xmath273 . furthermore , note that in proposition [ laxmil ] , lemma [ i d ] below , and the three propositions in section [ sec : normbounds ] , we only require @xmath274 and not the stronger assumption @xmath167 .
however , in all our other results we actually need @xmath275 [ i d ] suppose the assumptions [ a2]([a2delta ] ) and ( [ a2i ] ) hold .
then the operator @xmath276 is identical to the operator @xmath277 , where @xmath278 is defined weakly in @xmath279 .
the lax - milgram theorem implies that @xmath280 is bounded .
moreover , @xmath281 is bounded , thus the operator @xmath282 is also bounded and satisfies @xmath283 define @xmath278 weakly in @xmath279 , by the bilinear form @xmath284 given by @xmath285 by assumption [ a2]([a2i ] ) , @xmath286 is coercive and continuous in @xmath279 , thus by the lax - milgram theorem , there exists a uniquely defined , boundedly invertible , operator @xmath287 such that @xmath288 for all @xmath289 .
we identify @xmath278 with the bounded operator @xmath290 . by assumption [ a2]([a2delta ] ) we have @xmath274 hence @xmath291 that is , @xmath292 is bounded . by the definition of @xmath293 and @xmath294 , this implies that @xmath295 .
[ meanlem ] under the assumptions [ a2]([a1]),([a2delta]),([a2i]),([a2ii]),([a2v ] ) , there exists a unique weak solution , @xmath296 of equation ( [ eq : int7 ] ) , @xmath136-almost surely .
it suffices to show that @xmath297 @xmath136-almost surely .
indeed , by lemma [ l2](iv ) we have that @xmath298 @xmath136-a.s .
for all @xmath216 , thus by the assumption [ a2]([a2v ] ) @xmath299 since @xmath300 , which holds by the assumption [ a2]([a2delta ] ) .
suppose that in the problem ( [ eq : int1 ] ) we have @xmath301 and @xmath18 , where @xmath4 is independent of @xmath51 .
then we have that @xmath302 .
let @xmath47 be the posterior measure on @xmath303 . in this section
we prove a number of facts concerning the posterior measure @xmath47 for @xmath303 .
first , in theorem [ prop0 ] we prove that this measure has density with respect to the prior measure @xmath116 , identify this density and show that @xmath47 is lipschitz in @xmath5 , with respect to the hellinger metric .
continuity in @xmath5 will require the introduction of the space @xmath304 , to which @xmath4 drawn from @xmath116 belongs almost surely .
secondly , we prove theorem [ justth ] , where we show that @xmath47 is gaussian and identify the covariance and mean via equations ( [ eq : int4 ] ) and ( [ eq : int7 ] ) .
this identification will form the basis for our analysis of posterior contraction in the following section .
[ prop0]under the assumptions [ a2]([a1]),([a2delta]),([a2i]),([a2ii]),([a2iii]),([a2iv ] ) , the posterior measure @xmath47 is absolutely continuous with respect to @xmath116 and @xmath305 where @xmath306 and @xmath307 is the normalizing constant
. furthermore , the map @xmath308 is lipschitz continuous , with respect to the hellinger metric : let @xmath309 , @xmath310 ; then there exists @xmath311 such that for all @xmath312 with @xmath313 @xmath314 consequently , the @xmath47-expectation of any polynomially bounded function + @xmath315 where @xmath316 is a banach space , is locally lipschitz continuous in @xmath5 .
in particular , the posterior mean is locally lipschitz continuous in @xmath5 as a function @xmath317 .
the proofs of theorem [ prop0 ] and theorem [ justth ] are presented in the next two subsections . each proof is based on a series of lemmas . in this subsection
we prove theorem [ prop0 ] .
we first prove several useful estimates regarding @xmath318 defined in ( [ eq : phi ] ) , for @xmath319 and @xmath320 , where @xmath321 $ ] .
observe that , under the assumptions [ a2]([a1]),([a2delta]),([a2i]),([a2ii ] ) , for @xmath309 where @xmath96 sufficiently small , the lemma [ l2 ] implies on the one hand that @xmath319 @xmath322-almost surely and on the other hand that @xmath320 @xmath136-almost surely .
[ asslem ] under the assumptions [ a2]([a1]),([a2i]),([a2iii]),([a2iv ] ) , for any @xmath321 $ ] , the potential @xmath318 given by ( [ eq : phi ] ) satisfies : 1 . for every @xmath323 and @xmath324 there exists an @xmath325 such that for all @xmath319 and all @xmath320 with @xmath326 @xmath327 2 . for every @xmath324
there exists a @xmath328 such that for all @xmath319 and @xmath320 with @xmath329 @xmath330 3 . for every @xmath324 there exists an @xmath331 such that for all @xmath332 and @xmath320 with @xmath333 @xmath334 4 . for every @xmath323 and @xmath324 there
exists a @xmath335 such that for all @xmath336 with @xmath337 and for all @xmath338 @xmath339 @xmath340 1 . by first using the cauchy - schwarz inequality , then the assumptions [ a2 ] ( [ a2iii ] ) and ( [ a2iv ] ) , and then the cauchy with @xmath341 inequality for @xmath342 sufficiently small
, we have @xmath343 2 . by the cauchy - schwarz inequality and the assumptions [ a2]([a2i]),([a2iii ] ) and ( [ a2iv ] ) , we have since @xmath344 @xmath345 3 . by first using the assumptions [ a2 ] ( [ a2iii ] ) and ( [ a2iv ] ) and the triangle inequality , and then the assumption [ a2]([a2i ] ) and the reverse triangle inequality
, we have since @xmath344 @xmath346@xmath347@xmath348@xmath349@xmath350 4 . by first using the cauchy - schwarz inequality and then the assumptions [ a2]([a2iii ] ) and ( [ a2iv ] )
, we have @xmath351 under the assumptions [ a2]([a1]),([a2delta]),([a2i]),([a2iii]),([a2iv])@xmath352 for all @xmath353 where @xmath354 . in particular ,
if in addition the assumption [ a2]([a2ii ] ) holds , then @xmath355 @xmath215-almost surely .
fix @xmath320 and set @xmath356 .
gaussian measures on separable hilbert spaces are full ( * ? ? ?
* proposition 1.25 ) , hence since by lemma [ l2](i ) @xmath357 we have that @xmath358 . by lemma [ asslem](ii )
, there exists @xmath359 such that @xmath360 recalling that , under the additional assumption [ a2]([a2ii ] ) , by lemma [ l2](iv ) we have @xmath320 @xmath215-almost surely for all @xmath157 , completes the proof .
we are now ready to prove theorem [ prop0 ] : recall that @xmath361 and @xmath362 by the cameron - martin formula ( * ? ? ?
* corollary 2.4.3 ) , since by lemma [ l2](ii ) we have @xmath363-a.s .
, we get for @xmath116-almost all @xmath4 @xmath364
thus we have for @xmath116-almost all @xmath4 @xmath365 by ( * ? ? ?
* lemma 5.3 ) and corollary [ c1 ] we have the relation ( [ eq : rn ] ) . + for the proof of the lipschitz continuity of the posterior measure in @xmath5 , with respect to the hellinger distance , we apply ( * ? ? ? * theorem 4.2 ) for @xmath366 , using lemma [ asslem ] and the fact that @xmath367 , by lemma [ l2](i ) .
we first give an overview of the proof of theorem [ justth ] .
let @xmath368 and @xmath186 .
then by proposition [ meanlem ] , there exists a unique weak solution , @xmath296 , of ( [ eq : int7 ] ) , @xmath136-almost surely .
that is , with @xmath136-probability equal to one , there exists an @xmath369 such that @xmath370 where the bilinear form @xmath371 is defined in section [ sec : prmean ] , and @xmath372 . in the following
we show that @xmath26 , where @xmath373the proof has the same structure as the proof for the identification of the posterior in @xcite .
we define the gaussian measure @xmath374 , which is the independent product of a measure identical to @xmath375 in the finite - dimensional space @xmath376 spanned by the first @xmath377 eigenfunctions of @xmath17 , and a measure identical to @xmath116 in @xmath378 .
we next show that @xmath374 converges weakly to the measure @xmath47 which as a weak limit of gaussian measures has to be gaussian @xmath379 , and we then identify @xmath380 and @xmath381 with @xmath85 , @xmath35 respectively .
fix @xmath5 drawn from @xmath215 and let @xmath382 be the orthogonal projection of @xmath0 to the finite - dimensional space @xmath383 where as in section [ sec : assumptions ] , @xmath147 is an orthonormal eigenbasis of @xmath17 in @xmath0 .
let @xmath384 .
we define @xmath385 by @xmath386 where @xmath387 and @xmath388 [ lemproj ] we have @xmath389 , where @xmath390 @xmath391 and @xmath392 .
let @xmath393 . since @xmath394 we have by @xmath395 @xmath396 the right hand side is @xmath377-dimensional gaussian with density proportional to the exponential of the following expression @xmath397 which by completing the square we can write as @xmath398 where @xmath399 is the covariance matrix and @xmath400 the mean . by equating with expression ( [ star ] ) , we find that @xmath401 and @xmath402 , thus on @xmath376 we have that @xmath403 . on @xmath378 ,
the radon - nikodym derivative in @xmath395 is equal to 1 , hence @xmath404 .
[ prop2 ] under the assumptions [ a2]([a1]),([a2delta]),([a2i]),([a2ii]),([a2iii]),([a2iv ] ) , for all @xmath320 , @xmath309 , where @xmath405 , the measures @xmath385 converge weakly in @xmath0 to @xmath47 , where @xmath47 is defined in theorem [ prop0 ] .
in particular , @xmath385 converge weakly in @xmath0 to @xmath406-almost surely .
fix @xmath320 .
let @xmath407 be continuous and bounded . then by @xmath408 and lemma [ l2](i ) , we have that @xmath409 and @xmath410 let @xmath319 and set @xmath411 to get , by lemma [ asslem](iii ) , that @xmath412 since @xmath413 .
by lemma [ asslem](i ) , for any @xmath414 for @xmath415 , there exists @xmath416 such that @xmath417 where the right hand side is @xmath116-integrable for @xmath418 sufficiently small by the fernique theorem ( * ? ? ? * theorem 2.8.5 ) .
hence , by the dominated convergence theorem , we have that @xmath419 , as @xmath420 , where we get the convergence of the constants @xmath421 by choosing @xmath422 .
thus we have @xmath423 .
recalling , that @xmath424-almost surely completes the proof .
we are now ready to prove theorem [ justth ] : by proposition [ prop2 ] we have that @xmath385 converge weakly in @xmath0 to the measure @xmath47 , @xmath215-almost surely . since by lemma [ lemproj ] , the measures @xmath385 are gaussian , the limiting measure @xmath47 is also gaussian . to see this we argue as follows .
the weak convergence of measures implies the pointwise convergence of the fourier transforms of the measures , thus by levy s continuity theorem ( * ? ? ?
* theorem 4.3 ) all the one dimensional projections of @xmath385 , which are gaussian , converge weakly to the corresponding one dimensional projections of @xmath47 . by the fact that the class of gaussian distributions in @xmath425 is closed under weak convergence ( * ? ? ?
* chapter 4 , exercise 2 ) , we get that all the one dimensional projections of the @xmath47 are gaussian , thus @xmath47 is a gaussian measure in @xmath0 , @xmath426 for some @xmath427 and a self - adjoint , positive semi definite , trace class linear operator @xmath381 .
it suffices to show that @xmath428 and @xmath429 .
we use the standard galerkin method to show that @xmath430 in @xmath0 . indeed ,
since by their definition @xmath431 solve ( [ eq : int7 ] ) in the @xmath377-dimensional spaces @xmath376 , for @xmath432 , we have that @xmath433 . by the coercivity and the continuity of @xmath371
( see proposition [ laxmil])@xmath434 choose @xmath435 to obtain @xmath436 where as @xmath420 the right hand side converges to zero since @xmath296 .
on the other hand , by ( * ? ? ?
* example 3.8.15 ) , we have that @xmath437 in @xmath0 , hence we conclude that @xmath428 , as required . for the identification of the covariance operator , note that by the definition of @xmath438 we have @xmath439 recall that @xmath440 are the eigenfunctions of @xmath17 and fix @xmath441 .
then , for @xmath442 and any @xmath443 , we have that @xmath444@xmath445 where the right hand side converges to zero as @xmath420 , since @xmath443 .
this implies that @xmath446 converges to @xmath447 weakly in @xmath0 , as @xmath420 and this holds for any @xmath441 . on the other hand by ( *
* example 3.8.15 ) , we have that @xmath448 in @xmath0 , as @xmath420 , for all @xmath441 .
it follows that @xmath449 , for every @xmath206 and since @xmath440 is an orthonormal basis of @xmath0 , we have that @xmath429 .
the following propositions contain several operator norm estimates on the inverse of @xmath30 and related quantities , and in particular estimates on the singular dependence of this operator as @xmath61 .
these are the key tools used in section [ sec : main ] to obtain posterior contraction results . in all of them
we make use of the interpolation inequality in hilbert scales , ( * ? ? ?
* proposition 8.19 ) . recall that we consider @xmath30 defined on @xmath242 , as explained in remark [ yosida ] .
[ pdelem1 ] let @xmath94 , where @xmath95 $ ] .
under the assumption [ a2]([a2i ] ) the following operator norm bounds hold : there is @xmath450 independent of @xmath451 such that @xmath452 and @xmath453 in particular , if @xmath454 , interpolation of the two bounds gives @xmath455 where @xmath456 $ ] .
let @xmath457 .
then by proposition [ laxmil ] for @xmath458 , there exists a unique weak solution of ( [ eq : pm1 ] ) , @xmath459 . by definition [ weak ] , for @xmath460
, we get @xmath461using the assumption [ a2]([a2i ] ) , and the cauchy - schwarz inequality , we get @xmath462 we interpolate the norm on @xmath463 appearing on the right hand side between the norms on @xmath463 appearing on the left hand side , then use the cauchy with @xmath112 inequality , and then young s inequality for @xmath464 , to get successively , for @xmath450 a changing constant @xmath465 @xmath466 @xmath467 by choosing @xmath96 small enough we get , for @xmath450 independent of @xmath468 , @xmath469 replacing @xmath470 gives the result . [ pdelem3 ] let @xmath471 , where @xmath95 $ ] and @xmath321 $ ] , where @xmath76 as defined in assumption [ a2]([a1 ] ) . under the assumptions [ a2]([a2delta ] ) and ( [ a2i ] ) , the following norm bounds hold : there is @xmath450 independent of @xmath451 such that@xmath472and @xmath473 in particular , @xmath474.\nonumber\ ] ] let @xmath475 .
then @xmath476 since @xmath274 , thus @xmath477 .
by proposition [ laxmil ] for @xmath478 , there exists a unique weak solution of ( [ eq : pm1 ] ) , @xmath479 . since for @xmath480
we have that @xmath481 , we conclude that for any @xmath480 @xmath482where @xmath483 .
choosing @xmath484 we get @xmath485 by the assumption [ a2]([a2i ] ) and the cauchy - schwarz inequality , we have @xmath486 we interpolate the norm of @xmath463 appearing on the right hand side between the norms of @xmath463 appearing on the left hand side , to get as in the proof of proposition [ pdelem1 ] , for @xmath450 independent of @xmath468 and @xmath204 @xmath487 replacing @xmath488 gives the first two rates .
+ for the last claim , note that we can always choose @xmath489 , since @xmath250 and @xmath274 .
using the first two estimates , for @xmath490 , that is @xmath491,$ ] we have that @xmath492and@xmath493 let @xmath494 .
then , for any @xmath201 , we have the decomposition@xmath495 where @xmath440 are the eigenfunctions of @xmath17 and @xmath496 since @xmath497 and @xmath498 , we have @xmath499@xmath500 @xmath501@xmath502 the first term on the right hand side is increasing in @xmath503 , while the second is decreasing , so we can optimize by choosing @xmath504 making the two terms equal , that is @xmath505 to obtain the claimed rate .
in this section we employ the developments of the preceding sections to study the posterior consistency of the bayesian solution to the inverse problem . that is , we consider a family of data sets @xmath86 given by ( [ eq : int10 ] ) and study the limiting behavior of the posterior measure @xmath506 as @xmath9 . intuitively we would hope to recover a measure which concentrates near the true solution @xmath50 in this limit . following the approach in @xcite , @xcite , @xcite and @xcite
, we quantify this idea as in ( [ eq : main1 ] ) . by the markov inequality we have
@xmath507 so that it suffices to show that @xmath508 in addition to @xmath509 , there is a second small parameter in the problem , namely the regularization parameter , @xmath39 , and we will choose a relationship between @xmath31 and @xmath12 in order to optimize the convergence rates @xmath56 . we will show that determination of optimal convergence rates follows directly from the operator norm bounds on @xmath62 derived in the previous section , which concern only @xmath12 dependence ; relating @xmath31 to @xmath12 then follows as a trivial optimization .
thus , the @xmath12 dependence of the operator norm bounds in the previous section forms the heart of the posterior contraction analysis .
the relationship between @xmath12 and @xmath31 will induce a relationship between @xmath32 and @xmath31 , where @xmath32 being the scaling parameter in the prior covariance is the relevant parameter in the current bayesian framework .
we now present our convergence results . in theorem [ pdeth1 ]
we study the convergence of the posterior mean to the true solution in a range of norms , while in theorem [ pdeth2 ] we study the concentration of the posterior near the true solution as described in ( [ eq : main1 ] ) .
the proofs of theorems [ pdeth1 ] and [ pdeth2 ] are provided later in the current section .
the two main convergence results , theorems [ pdecor1 ] and [ pdecor2 ] follow as direct corollaries of remark [ r1 ] and theorems [ pdeth1 ] and [ pdeth2 ] respectively . [ pdeth1 ] let @xmath251 . under the assumptions [ a2 ] , we have that , for the choice @xmath510 and for any @xmath95 $ ] @xmath511 where @xmath94 .
the result holds for any @xmath512 $ ] , chosen so that @xmath513 , for @xmath514 , where @xmath515 [ pdeth2]let @xmath251
. under the assumptions [ a2 ] , we have that , for @xmath510 , the convergence in ( [ eq : main1 ] ) holds with @xmath516 the result holds for any @xmath512 $ ] , chosen so that @xmath517 , for + @xmath518 , where @xmath515 [ r1 ] 1 . to get convergence in the pde method we need @xmath519 for a @xmath520 . under the a priori information
that @xmath91 , we need @xmath521 for some @xmath522 $ ] .
thus the minimum requirement for convergence is @xmath103 in agreement to our assumption @xmath251 . on the other hand , to obtain the optimal rate ( which corresponds to choosing @xmath523 as small as possible ) we need to choose @xmath524 .
if @xmath525 then the right hand side is negative so we have to choose @xmath526 , hence we can not achieve the optimal rate .
we say that the method saturates at @xmath527 which reflects the fact that the true solution has more regularity than the method allows us to exploit to obtain faster convergence rates .
2 . to get convergence we also need @xmath528 for a @xmath529 . by lemma [ l2](iii ) , it suffices to have @xmath530 .
this means that we need @xmath531 , which holds by the assumption [ a2]([a2delta ] ) , in order to be able to choose @xmath529 . on the other hand , since @xmath274 and @xmath532
, we have that @xmath533 thus we can always choose @xmath534 in an optimal way , that is , we can always choose @xmath535 where @xmath96 is arbitrarily small .
if we want draws from @xmath116 to be in @xmath536 then by lemma [ l1](ii ) we need @xmath537 .
since the requirement for the method to give convergence is @xmath538 while @xmath539 , we can never have draws exactly matching the regularity of the prior . on the other hand
if we want an undersmoothing prior ( which according to @xcite in the diagonal case gives asymptotic coverage equal to 1 ) we need @xmath540 , which we always have .
this , as discussed in section [ sec : intro ] , gives an explanation to the observation that in both of the above theorems we always have @xmath541 as @xmath9 .
when @xmath542 in theorem [ pdeth2 ] and in theorem [ pdecor2 ] below , we get suboptimal rates .
the reason is that our analysis to obtain the error in the @xmath0-norm is based on interpolating between the error in the @xmath279-norm and the error in the @xmath242-norm . when @xmath542 interpolation is not possible since the @xmath0-norm is now weaker than the @xmath279-norm
however , we can at least bound the error in the @xmath0-norm by the error in the @xmath279-norm , thus obtaining a suboptimal rate .
note , that the case @xmath543 does not necessarily correspond to the well posed case : by lemma [ l2 ] we can only guarantee that a draw from the noise distribution lives in @xmath544 , while the range of @xmath72 is formally @xmath545 .
hence , in order to have a well posed problem we need @xmath546 , or equivalently @xmath547 . this can happen despite our assumption @xmath167 , when @xmath548 and for appropriate choice of @xmath248 and @xmath247 . in this case , regularization is unnecessary .
note that , since the posterior is gaussian , the left hand side in ( [ eq : main2 ] ) is the square posterior contraction @xmath549 which is the sum of the mean integrated squared error ( mise ) and the posterior spread .
let @xmath251 .
by lemma [ i d ] , the relationship ( [ eq : int10 ] ) between @xmath50 and @xmath58 and the equation ( [ eq : int11 ] ) for @xmath88 , we obtain @xmath550 where the equations hold in @xmath551 , since by a similar argument to the proof of proposition [ meanlem ] we have @xmath552 . by subtraction
we get @xmath553 therefore @xmath554 as an equation in @xmath242 .
using the fact that the noise has mean zero and the relation ( [ eq : int6 ] ) , equation ( [ eq : main5 ] ) implies that we can split the square posterior contraction into three terms@xmath555 provided the right hand side is finite .
a consequence of the proof of theorem [ justth ] is that @xmath62 is trace class .
note that for @xmath184 a white noise , we have that @xmath556@xmath557 which for @xmath157 since by lemma [ l1 ] we have that @xmath185 , provides the bound @xmath558 where @xmath450 is independent of @xmath12 . if @xmath559 are chosen sufficiently large so that @xmath560 and @xmath561 then we see that @xmath562 where @xmath450 is independent of @xmath12 and @xmath31 . thus identifying @xmath56 in ( [ eq : main1 ] ) can be achieved simply through properties of the inverse of @xmath30 and its parametric dependence on @xmath12 . in the following
, we are going to study convergence rates for the square posterior contraction , ( [ eq : main6 ] ) , which by the previous analysis will secure that @xmath563 for @xmath564 at a rate almost as fast as the square posterior contraction .
this suggests that the error is determined by the mise and the trace of the posterior covariance , thus we optimize our analysis with respect to these two quantities . in
@xcite the situation where @xmath71 and @xmath46 are diagonalizable in the same eigenbasis is studied , and it is shown that the third term in equation ( [ eq : main6 ] ) is bounded by the second term in terms of their parametric dependence on @xmath12 .
the same idea is used in the proof of theorem [ pdeth2 ] .
we now provide the proofs of theorem [ pdeth1 ] and theorem [ pdeth2 ] . since @xmath51 has zero mean , we have by ( [ eq : main5 ] ) @xmath565and @xmath566 using proposition [ pdelem1 ] and assumption [ a2]([a2v ] ) , we get @xmath567and @xmath568 since the common parenthesis term , consists of a decreasing and an increasing term in @xmath32 , we optimize the rate by choosing @xmath569 such that the two terms become equal , that is , @xmath570 .
we obtain , @xmath571 by interpolating between the two last estimates we obtain the claimed rate . recall equation ( [ eq : main6])@xmath572 the idea is that the third term is always dominated by the second term . combining equation ( [ eq : main7 ] ) with proposition [ pdelem3 ] , we have that @xmath573.\nonumber\ ] ] 1 .
suppose @xmath574 so that by proposition [ pdelem1 ] we have , where @xmath456 $ ] , using assumption [ a2]([a2v ] ) @xmath575and @xmath576 note that @xmath534 is chosen so that @xmath528 , that is , by lemma [ l2](iii ) , it suffices to have @xmath577 .
noticing that by choosing @xmath204 arbitrarily close to @xmath578 we can have @xmath579 arbitrarily close to @xmath580 , and since @xmath581 , we deduce that the third term in equation ( [ eq : main6 ] ) is always dominated by the second term . combining
, we have that @xmath582 2 .
suppose @xmath543 .
using proposition [ pdelem1 ] and assumption [ a2]([a2v ] ) we have @xmath583 and @xmath584@xmath585where as before @xmath586 the third term in equation ( [ eq : main6 ] ) is again dominated by the second term , since on the one hand @xmath577 and on the other hand , since @xmath543 , we can always choose @xmath587 to get @xmath588 combining the three estimates we have that @xmath589 in both cases , the common term in the parenthesis consists of a decreasing and an increasing term in @xmath32 , thus we can optimize by choosing @xmath569 making the two terms equal , that is , @xmath590 , to get the claimed rates .
we now present some nontrivial examples satisfying assumptions [ a2 ] .
let @xmath591 be a bounded and open set .
we define @xmath592 , where @xmath593 is the dirichlet laplacian which is the friedrichs extension of the classical laplacian defined on @xmath594 that is , @xmath595 is a self - adjoint operator with a domain @xmath596 dense in @xmath597 @xcite . for @xmath598
sufficiently smooth we have @xmath599 .
it is well known that @xmath595 has a compact inverse and that it possesses an eigensystem @xmath600 where the eigenfunctions @xmath601 form a complete orthonormal basis of @xmath0 and the eigenvalues @xmath602 behave asymptotically like @xmath603 @xcite . in subsections [ ex1 ] and [ ex2 ] , we consider the inverse problem to find @xmath4 from @xmath5 , where @xmath604for @xmath463 solving the partial differential equation @xmath605 that is , @xmath606 , where @xmath607 is a nonnegative real function of certain regularity .
we choose prior and noise distributions with covariance operators which are not simultaneously diagonalizable with the forward operator . later on , in subsection [ ex3
] , we consider more complicated examples and in particular , we consider fractional powers of the dirichlet laplacian in the forward operator , as well as more general choices of prior and noise covariance operators . our general strategy for proving the validity of our norm equivalence assumptions is : 1 .
if needed , use proposition [ inter ] below to reduce the range of spaces required to check an assumption s validity to a finite set of spaces ; 2 . reformulate the assumptions as statements regarding the boundedness of operators of the form considered in lemma [ prolem ] below .
the statement of proposition [ inter ] , which is a well known result from interpolation theory , and the statement and proof of lemma [ prolem ] are postponed to subsection [ interpolation ] .
we study the bayesian inversion of the operator @xmath608 where @xmath609 is the multiplication operator by a nonnegative function @xmath610 .
we assume that the observational noise is white , so that @xmath73 , and we set the prior covariance operator to be @xmath611 .
the operator @xmath17 is trace class .
indeed , let @xmath612 be its eigenvalues .
then they behave asymptotically like @xmath613 and @xmath614 for @xmath615 .
furthermore , we have that @xmath616 provided @xmath617 , that is , the assumption [ a2]([a1 ] ) is satisfied with @xmath618 we define the hilbert scale induced by @xmath611 , that is , @xmath619 , for @xmath620 , where@xmath621 observe , @xmath622 .
our aim is to show that @xmath623 and @xmath624 , where @xmath625 and @xmath626 , in the sense of the assumptions [ a2 ] .
we have @xmath627 .
since for @xmath628 we have @xmath629 , the assumption [ a2]([a2delta ] ) is satisfied .
moreover , note that since @xmath73 the assumptions [ a2]([a2ii ] ) and ( [ a2iii ] ) are trivially satisfied . we now show that assumptions [ a2 ] ( [ a2i ] ) , ( [ a2iv ] ) , ( [ a2v ] ) are also satisfied . in this example
the three assumptions have the form 1 .
@xmath630 2 . @xmath631 $ ] ; 3 .
@xmath632 $ ] .
observe that assumption ( [ a2iv ] ) is implied by assumption ( [ a2v ] ) .
[ proass]the assumptions [ a2 ] are satisfied in this example .
we only need to show that assumptions ( [ a2i ] ) and ( [ a2v ] ) hold . 1 .
the assumption is equivalent to @xmath633 and @xmath634 being bounded in @xmath0 . since @xmath635 which is bounded in @xmath0
, we only need to show that @xmath636 is bounded .
indeed , @xmath637 , which is bounded by lemma [ prolem ] applied for @xmath638 .
2 . by proposition
[ inter ] , it suffices to show @xmath639 we have already shown that @xmath640 .
for @xmath641 note that it is equivalent to @xmath642 , which holds by lemma [ prolem ] applied for @xmath643 . finally , for @xmath644 ,
note that it is equivalent to @xmath645 , which holds by lemma [ prolem ] applied for @xmath638
. we can now apply theorem [ pdecor1 ] and theorem [ pdecor2 ] to get the following convergence result .
let @xmath646 .
then , for @xmath647 , the convergence in ( [ eq : main1 ] ) holds with @xmath648 , where @xmath649for @xmath96 arbitrarily small and where @xmath650 is the dimension .
furthermore , for @xmath651 , for the same choice of @xmath32 , we have @xmath652 where @xmath653 for @xmath654 the above rate holds provided @xmath655 .
as in example [ ex1 ] , we study the bayesian inversion of the operator @xmath656 for a nonnegative @xmath610 .
we assume that the observational noise is gaussian with covariance operator @xmath657 , where @xmath658 is the multiplication operator by a nonnegative function @xmath659 .
as before , we set the prior covariance operator to be @xmath611 , thus the assumption [ a2]([a1 ] ) is satisfied with the same @xmath77 and we work in the same hilbert scale @xmath660 we show that @xmath661 and @xmath662 , where @xmath663 and @xmath626 , in the sense of the assumptions [ a2]([a2i])-([a2v ] ) .
first note that we have @xmath664 for @xmath628 , so that the assumption [ a2]([a2delta ] ) is satisfied .
the rest of the assumptions have the form 1 .
@xmath665 2 .
@xmath666 3 .
@xmath667;$ ] 4 . @xmath668 $ ] ; 5 . @xmath669 $ ] . the assumptions [ a2 ] are satisfied in this example .
we have already seen that the first two assumptions are satisfied . 1 .
we need to show that @xmath670 and @xmath671 are bounded operators in @xmath0 .
indeed , @xmath672 which is bounded by lemma [ prolem ] applied for @xmath673 and @xmath674 .
for @xmath671 we have , @xmath675 , which again by lemma [ prolem ] is the composition of two bounded operators .
2 . since @xmath676 for @xmath628
respectively , it suffices to show that it holds for all @xmath677 $ ] . by proposition [ inter ]
it suffices to show that @xmath678 .
this is equivalent to showing that @xmath679 and @xmath680 are bounded in @xmath0 , which holds by lemma [ prolem ] .
3 . by proposition [ inter ]
it suffices to show that @xmath681 .
indeed , @xmath682 . on the other hand , to show @xmath683 it is equivalent to show that @xmath684 .
indeed , @xmath685 which is bounded by lemma [ prolem ] .
4 . by proposition [ inter ]
it suffices to show that @xmath686 .
indeed , we have already shown in part @xmath687 of the current proof that @xmath688 . to show @xmath689 it is equivalent to show that @xmath690 .
indeed , @xmath691 which by lemma [ prolem ] is the composition of two bounded operators in @xmath0 .. 5 . by proposition [ inter ]
it suffices to show that @xmath692 .
we start by showing @xmath688 .
indeed , we have @xmath693 , which by lemma [ prolem ] , is the composition of three bounded operators . for showing @xmath683
it is equivalent to show that @xmath694 .
indeed , @xmath695 , which by lemma [ prolem ] , is the composition of three bounded operators . finally , we show that @xmath689 or equivalently @xmath696 .
indeed we have @xmath697 , which again by lemma [ prolem ] , is the composition of three bounded operators
. we can now apply theorem [ pdecor1 ] and theorem [ pdecor2 ] to get the following convergence result .
let @xmath646 .
then , for @xmath698 , the convergence in ( [ eq : main1 ] ) holds with @xmath648 , where @xmath699for @xmath96 arbitrarily small and where @xmath650 is the dimension .
furthermore , for @xmath700 , for the same choice of @xmath32 , we have @xmath652 where @xmath701 for @xmath654 the above rate holds provided @xmath655 .
the same methodology can be applied to more general examples , for instance , in the case where @xmath702 , @xmath703 and @xmath704 , for nonnegative functions @xmath705 and @xmath706 , where @xmath707 and @xmath708 such that @xmath709 .
the values of @xmath710 are chosen as sufficiently large even integers depending on the values of @xmath711 .
note that we require @xmath707 for our compactness arguments to work , however , the cases @xmath625 and/or @xmath712 also work using a slightly modified proof .
the proof is omitted for brevity but the interested reader may consult @xcite for details .
let @xmath713 be the hilbert scale induced by a self - adjoint positive definite linear operator @xmath714 ( cf .
section [ sec : assumptions ] ) .
the following result holds ( * ? ? ?
* theorems 4.36 , 1.18 , 1.6 ) : [ inter ] for any @xmath201 , the couples @xmath715 and @xmath716 are interpolation couples and for every @xmath95 $ ] we have @xmath717 and @xmath718 . in particular , for any @xmath719 , if @xmath720 then @xmath721 for any @xmath95 $ ] .
let @xmath722 be a nonnegative function and define the multiplication operator @xmath723 .
note that by the hlder inequality the operator @xmath724 is bounded .
the last proposition , implies the following lemma .
[ prolem ] for any @xmath725 , @xmath726 is a bounded operator in @xmath0 , provided @xmath727 .
furthermore , for any @xmath183 the operators @xmath728 and @xmath729 are compact in @xmath0 and @xmath730 are bounded in @xmath0 .
we begin by showing that @xmath731 for @xmath732 $ ] . by the last proposition applied for @xmath733 and since @xmath724 is bounded , it suffices to show that @xmath734 and @xmath735 are bounded in @xmath0 .
in fact it suffices to show that @xmath735 is bounded since @xmath736 indeed , since @xmath737 @xmath738@xmath739 for general @xmath725 , let @xmath740 , then as before it suffices to show that @xmath741 is bounded in @xmath0 .
again , using the fact that @xmath742 , we have by the product rule for derivatives that @xmath741 is bounded , provided @xmath743 .
the operators @xmath744 are compact in @xmath0 , since they are compositions between the compact operator @xmath745 and the bounded operator @xmath726 .
positivity of the operator @xmath595 and nonnegativity of the operator @xmath724 show that @xmath746 can not be an eigenvalue of @xmath744 , so that by the fredholm alternative ( * ? ? ?
* , theorem 7 ) we have that @xmath730 are bounded in @xmath0 .
in the case where @xmath71 and @xmath46 , are all diagonalizable in the same eigenbasis our assumptions are trivially satisfied , provided @xmath167 . in @xcite ,
sharp convergence rates are obtained for the convergence in ( [ eq : main1 ] ) , in the case where the three relevant operators are simultaneously diagonalizable and have spectra that decay algebraically ; the authors only consider the case @xmath73 since in this diagonal setting the colored noise problem can be reduced to the white noise one .
the rates in @xcite agree with the minimax rates provided the scaling of the prior is optimally chosen , @xcite . in figure
[ fig ] ( cf .
section [ sec : mainresults ] ) we have in green the rates of convergence predicted by theorem [ pdecor2 ] and in blue the sharp convergence rates from @xcite , plotted against the regularity of the true solution , @xmath91 , in the case where @xmath747 and @xmath17 has eigenvalues that decay like @xmath748 . in this case @xmath749 and @xmath750 , so that @xmath167 .
as explained in remark [ r1 ] , the minimum regularity for our method to work is @xmath103 and our rates saturate at @xmath527 , that is , in this example at @xmath751 .
we note that for @xmath752 $ ] our rates agree , up to @xmath96 arbitrarily small , with the sharp rates obtained in @xcite , for @xmath753 our rates are suboptimal and for @xmath754 the method fails . in @xcite
, the convergence rates are obtained for @xmath755 and the saturation point is at @xmath756 , that is , in this example at @xmath757 . in general the pde method
can saturate earlier ( if @xmath758 ) , at the same time ( if @xmath759 ) , or later ( if @xmath760 ) compared to the diagonal method presented in @xcite .
however , the case @xmath760 in which our method saturates later , is also the case in which our rates are suboptimal , as explained in remark [ r1](iv ) .
the discrepancies can be explained by the fact that in proposition [ pdelem1 ] , the choice of @xmath451 which determines both the minimum requirement on the regularity of @xmath50 and the saturation point , is the same for both of the operator norm bounds .
this means that on the one hand to get convergence of the term @xmath761 in equation ( [ eq : main6 ] ) in the proof of theorem [ pdeth2 ] , we require conditions which secure the convergence in the stronger @xmath242-norm and on the other hand the saturation rate for this term is the same as the saturation rate in the weaker @xmath279-norm .
for example , when @xmath762 the saturation rate in the pde method is the rate of the @xmath0-norm hence we have the same saturation point as the rates in @xcite .
in particular , we have agreement of the saturation rate when @xmath763 , which corresponds to the problem where we directly observe the unknown function polluted by white noise ( termed the _ white noise model _ ) .
we have presented a new method of identifying the posterior distribution in a conjugate gaussian bayesian linear inverse problem setting ( section [ sec : mainresults ] and section [ sec : justification ] ) .
we used this identification to examine the posterior consistency of the bayesian approach in a frequentist sense ( section [ sec : mainresults ] and section [ sec : main ] ) .
we provided convergence rates for the convergence of the expectation of the mean error in a range of norms ( theorem [ pdeth1 ] , theorem [ pdecor1 ] ) .
we also provided convergence rates for the square posterior contraction ( theorem [ pdeth2 ] , theorem [ pdecor2 ] ) .
our methodology assumed a relation between the prior covariance , the noise covariance and the forward operator , expressed in the form of norm equivalence relations ( assumptions [ a2 ] ) .
we considered gaussian noise which can be white . in order for our methods to work we required a certain degree of ill - posedness compared to the regularity of the prior ( assumption [ a2]([a2delta ] ) ) and for the convergence rates to be valid a certain degree of regularity of the true solution . in the case where the three involved operators are all diagonalizable in the same eigenbasis , when the problem is sufficiently ill - posed with respect to the prior , and for a range of values of @xmath92 , the parameter expressing the regularity of the true solution , our rates agree ( up to @xmath96 arbitrarily small ) with the sharp ( minimax ) convergence rates obtained in @xcite ( section [ sec : diag ] ) .
our optimized rates rely on rescaling the prior depending on the size of the noise , achieved by choosing the scaling parameter @xmath66 in the prior covariance as an appropriate function of the parameter @xmath13 multiplying the noise .
however , the relationship between @xmath32 and @xmath31 depends on the unknown regularity of the true solution @xmath92 , which raises the question how to optimally choose @xmath32 in practice .
an attempt to address this question in a similar but more restrictive setting than ours is taken in @xcite , where an empirical bayes maximum likelihood based procedure giving a data driven selection of @xmath32 is presented .
a different approach is taken in @xcite in the simultaneously diagonalizable case . as discussed in @xcite , for a fixed value of @xmath32 independent of @xmath31 ,
the rates are optimal only if the regularity of the prior exactly matches the regularity of the truth . in @xcite ,
an empirical bayes maximum likelihood based procedure and a hierarchical method are presented providing data driven choices of the regularity of the prior , which are shown to give optimal rates up to slowly varying terms .
we currently investigate hierarchical methods with conjugate priors and hyperpriors for data driven choices of both the scaling parameter of the prior @xmath32 and the noise level @xmath13 .
the methodology presented in this paper is extended to drift estimation for diffusion processes in @xcite .
future research includes the extension to an abstract setting which includes both the present paper and @xcite as special cases .
other possible directions are the consideration of nonlinear inverse problems , the use of non - gaussian priors and/or noise and the extension of the credibility analysis presented in @xcite to a more general setting .
25 natexlab#1#1[2]#2 , , ( ) .
, , , , volume of _ _ , , , . .
, , , volume of _ _ , , , .
, , volume of _ _ , , , .
, , universitext , , , . .
, , , ( ) . . , , , , volume of _ _ , , , . , , , ( ) . , , , , ( ) . , , , ( ) . , , , , ( ) . , , north - holland series in applied mathematics and mechanics , vol . 6 , , , . , , probability and its applications ( new york ) , , , . , , , , , http://arxiv.org/abs/1209.3628 , .
, , , , ( ) . , , pure and applied mathematics ( new york ) , , , . , , , , ( ) .
, , appunti .
scuola normale superiore di pisa ( nuova serie ) .
[ lecture notes .
scuola normale superiore di pisa ( new series ) ] , , edition , .
, , , , ( ) .
, , cambridge texts in applied mathematics , , , . .
, , , ( ) . , , ( ) . , , , ( ) .
, , , ( ) . | we consider a bayesian nonparametric approach to a family of linear inverse problems in a separable hilbert space setting with gaussian noise .
we assume gaussian priors , which are conjugate to the model , and present a method of identifying the posterior using its precision operator .
working with the unbounded precision operator enables us to use partial differential equations ( pde ) methodology to obtain rates of contraction of the posterior distribution to a dirac measure centered on the true solution .
our methods assume a relatively weak relation between the prior covariance , noise covariance and forward operator , allowing for a wide range of applications .
posterior consistency , posterior contraction , gaussian prior , posterior distribution , inverse problems 62g20 , 62c10 , 35r30 , 45q05 |
given an initial quantum state @xmath0 , that may be either a pure state or a density matrix , and assuming factorizability of the total hilbert space to which it belongs , @xmath1 , we first define the reduced density matrix @xmath2 as @xmath3 , where @xmath4 is the complementary of @xmath5 in @xmath6 .
the entanglement entropy ( ee ) is the von neumann entropy of @xmath2 : @xmath7 it is easy to show that @xmath8 corresponds to some measure of the entanglement between degrees of freedom inside @xmath5 and degrees of freedom outside , hence the name . a simple formula exists for the holographic computation of ee , if factorizability of the total hilbert space is induced by dividing the space ( or more covariantly a space - like surface ) into non intersecting subspaces @xmath9 , @xmath10 , @xmath11 , @xmath12 , each one defining the corresponding hilbert space @xmath5 , @xmath13 , @xmath14 , @xmath12 for the corresponding local degrees of freedom , and if an holographic description of the quantum mechanical theory is possible in terms of einstein gravity , with @xmath0 represented by a classical metric solution .
ryu and takayanagi for the static case at fixed boundary time @xcite , and later hubeny , rangamani and takayanagi ( hrt ) for the covariant generalization @xcite , proposed the following : @xmath15 in the above equation @xmath16 is the newton constant in the gravitational @xmath17 dimensional theory , and @xmath18 is the area of the extremal codimension two space - like surface @xmath19 homologous to @xmath9 and such that , at the boundary of the gravitational manifold , @xmath20 .
if the gravitational geometry is static , the above formula simplifies in the euclidean signature by using a minimal surface . ) and the corresponding surface @xmath19 as ryu - takayanagi or hrt depending if used for respectively static spacetimes in euclidean signature or generically lorentzian time dependent . ] in the static case a proof for the ryu - takayanagi formula was provided by @xcite , but still lacks for the time dependent covariant case .
nonetheless we will assume , throughout the paper , that ( [ ryutak ] ) is valid .
one interesting thing about holography , and in particular when dealing with formulas like ( [ ryutak ] ) , is the interplay between quantum mechanics and classical general relativity ; working with ee we can compare properties obeyed by @xmath8 , deeply related to entanglement at the very foundations of quantum mechanics , with measures of areas of surfaces probing classical geometries of the dual space - time !
the literature somehow related to this is vast , and covers the question of what quantum conditions ( inequalities ) can be derived from holography @xcite , @xcite and @xcite , the opposite problem of what gravitational restrictions are imposed from quantum inequalities @xcite and @xcite , studies the appearance of gravity equations of motions and dynamics from ee @xcite , @xcite , @xcite,@xcite and @xcite , and the reverse @xcite and @xcite , and even reconstructs the metric from boundary ee data @xcite .
it is well known that ee satisfies a series of inequalities , among which the two most restricting go under the name of strong subadditivity . given three hilbert spaces that factorize the total one @xmath1 , the two inequalities read : @xmath21 where @xmath22 refers to the ee computed by using the reduced density matrix living on the product of the two hilbert spaces @xmath23 , and so on . if factorizability can be achieved by considering spacial non intersecting regions @xmath9 , @xmath10 and @xmath11 , then the same result holds for the corresponding ees and @xmath22 is the ee corresponding to the spatial region @xmath24 and so on . note that these two equations are equivalent , as will be shown in the next section .
even if quantum mechanical proofs of ( [ ssa1 ] ) and ( [ ssa2 ] ) exist , see for example @xcite , it is a nontrivial question to ask if the ee computed holographically by the ryu - takayanagi ( or covariantly hrt ) formula satisfies them or not . why the question is nontrivial is due to three possible pitfalls .
the first problem is the non induced factorizability of the total hilbert space from dividing the physical space into subregions .
it is well known for example for gauge theories , where gauge invariance at the boundary between a region @xmath9 and its complementary @xmath25 does not allow a factorization of the total hilbert space of physical gauge invariant states into @xmath5 and @xmath26 . or , said in another way , the degrees of freedom of gauge theories are not point - like but rather nonlocal ( wilson loops ) , and consequently when considering a certain region of space they do not construct a physical ( gauge invariant ) hilbert space .
the literature is wide , see for example the lists of papers @xcite for a more theoretical approach and @xcite for attempts to formulate a lattice definition . while a formulation of ee may still be possible for gauge theories , for example by appropriately enlarging the hilbert space or by correctly defining the path integral replica trick , whenever the total hilbert space does not factorize we are in general not guaranteed that inequalities like ( [ ssa1 ] ) and ( [ ssa2 ] ) hold .
in fact this argument may apply even to ee computed non holographically , as long as the reduced density matrices are defined with respect to a certain region of space .
a pure cft example of strong subadditivity violation without passing through the holographic description , but rather considering ee computed using the replica trick in the cft , will be provided in section [ cft ] . there
we will consider two dimensional cfts with lorentz anomaly , by using the results of @xcite where it was noted that the ee , as expected , is not lorentz invariant .
this means that the ee transforms under boosts and , if these are applied to space regions entering inequalities ( [ ssa1 ] ) and ( [ ssa2 ] ) , the two sides can change so that to possibly lead to a violation .
said otherwise , strong subadditivity may be respected in some fixed reference frame in which the quantization of the theory has been implemented , but if we consider hilbert spaces corresponding to space regions that have been differently boosted with respect to this original frame , and the theory has a lorentz anomaly , violation can occur .
this is the second problem and again it applies even without passing through an holographic computation . on the issue of ee and its relation to anomalies
see also the recent papers @xcite @xcite @xcite .
the third problem arises only in holographic ee : while einstein s equations can be solved using any energy momentum tensor as a source , it is not guaranteed that any choice is actually physically meaningful . to constrain the energy momentum tensor various energy conditions
have been proposed , see for example @xcite for a review of these conditions and of their effects . without imposing any condition
it is not improbable that the resulting classical geometry may not be dual to same actual physical quantum system .
in particular we are no longer guaranteed on the validity of any quantum mechanically - derived inequality .
we will discuss this further in the next section in relationship with the proof of the equivalence between ( [ ssa1 ] ) and ( [ ssa2 ] ) , and in the conclusions .
note that , if any of the above pitfalls affects the proof of either ( [ ssa1 ] ) or ( [ ssa2 ] ) , that is we have violation of only one of the two inequalities , then similar reasons should necessary lead to the violation of the proof for the equivalence between ( [ ssa1 ] ) and ( [ ssa2 ] ) that we will discuss in the next section . extremely simple holographic proofs of ( [ ssa1 ] ) and ( [ ssa2 ] ) exist when the three connected adjacent regions are at the same constant boundary time ( or its boosted version ) and the bulk geometry is static , somehow in contrast with the complication of the standard quantum mechanical proofs
. however the argument fails when more generic configurations of regions and/or time dependent backgrounds are considered . because of this some works considered dropping one or both of these restrictions in order to check if the strong subadditivity inequalities were still satisfied or not , and if some conditions should be eventually applied to the geometry in order to enforce their validity . in particular @xcite and @xcite considered static backgrounds with generic space - like boundary regions @xmath9 , @xmath10 and @xmath11 , and found that requiring ( [ ssa2 ] ) leads to a integrated version of the null curvature condition ( ncc ) ( [ intncc ] ) .
further a simple time dependent geometry , asymptotically ads in 2 + 1 dimensions which is the vaidya metric was used by @xcite , @xcite and @xcite ; depending on the choice of a sign the vaidya geometry may be selected to satisfy or not the local ncc ( [ ncc ] ) , and what was found is that the ncc is a sufficient requirement to respect ( [ ssa2 ] ) , while ( [ ssa1 ] ) is always satisfied .
finally a proof for ( [ ssa2 ] ) was provided by @xcite for dual geometries satisfying the ncc and generic connected adjacent space - like intervals .
the main goals of this paper are three .
first of all to review in an organized way the present understanding of strong subadditivity inequalities in holographic theories and their connection with energy conditions .
second to enlighten the geometrical part of the above problem studying what is the holographic counterpart of the violation of strong subadditivity and the role of energy conditions .
third to fill some gaps in the literature and discuss further discoveries as , notably , the violation of ( [ ssa2 ] ) by two dimensional cfts with lorentz anomaly , the development of time - like distances between geodesics entering ( [ ssa2 ] ) whenever the inequality does not hold and some new proofs along the way . the paper is organized by discussing strong subadditivity in set - ups of crescent complication , always in 2 + 1 dimensions in the bulk ( although some results may be generalized ) and boundary regions @xmath9 , @xmath10 and @xmath11 chosen to be adjacent , starting with static backgrounds with collinear intervals then moving to generic space - like configurations and finally to time dependent geometries , discussing both the purely quantum mechanical problem and its holographic version .
two appendices contain respectively the generic result and proof of what configurations for the boundary intervals @xmath9 , @xmath10 and @xmath11 create the strongest bound on ee by strong subadditivity , and various computations for vaidya metric that will be used in the last sections as specific examples for the more generic discussion .
we begin with the purely quantum mechanical proof of the equivalence between the two strong subadditivity inequalities , ( [ ssa1 ] ) and ( [ ssa2 ] ) @xcite .
let us introduce an auxiliary hilbert space @xmath27 such that partial tracing some pure state @xmath28 over it , reproduces the reduced density matrix @xmath29 @xcite : @xmath30 then @xmath31 and we can convert ( [ ssa1 ] ) into ( [ ssa2 ] ) @xmath32 a discussion on why this proof may fail in certain circumstances can be found inside @xcite .
the argument is that , purely quantum mechanically we are always guaranteed to find @xmath27 .
if the theory has zero entropy it is just the complementary of @xmath33 inside the total hilbert space , @xmath34 , while if the theory is described by some density matrix it may not be the hilbert space of degrees of freedom of the theory itself .
still the proof applies if we limit ourselves to ee computed quantum mechanically and not holographically .
the only pitfall in this case is the already discussed issue of factorizability of the total hilbert space that we will not repeat .
holographically instead , if the bulk theory contains a black hole , the total entropy of the system in non zero and the boundary theory is not in a pure state but rather in a thermal density matrix .
the proof then holds for the holographic entanglement entropy only if the bulk actually describes some quantum system and we can thus use the above derivation .
if the bulk is sufficiently unphysical we may instead have troubles ; this is in fact just another way of recasting the argument that generic bulk geometries may not be dual to quantum mechanical systems , as discussed in the introduction .
finally we will soon see a purely cft violation of ( [ ssa2 ] ) in the following sections while preserving ( [ ssa1 ] ) for two dimensional theories with lorentz anomaly , so the above proof should break down in this case for analogous reasons .
the argument here is as in the introduction , and the problem arises when we consider differently boosted space regions starting from the reference frame that was used to quantize the theory and in which strong subadditivity holds ; lorentz anomaly implies that all the expressions written in the original frame will transform and relationships between them are no longer guaranteed to hold .
similarly simple proofs do not exist for the two strong subadditivity inequalities , that from now on we will name ssa1 for ( [ ssa1 ] ) and ssa2 for ( [ ssa2 ] ) , that instead require some amount of work to be verified .
simplicity is restored when we consider their holographic description as ryu - takayanagi surfaces when considering collinear boundary intervals in static spacetimes . for the rest of the paper ,
unless otherwise stated , the 2 + 1 dimensional bulk theories will be in the euclidean signature when static , and lorentzian when dynamic .
the easiest case is that of static geometry with the three boundary intervals collinear , either at fixed time or belonging to a straight space - like line ; also we assume lorentz invariance .
it is quite surprising that the proof for both ( [ ssa1 ] ) and ( [ ssa2 ] ) just amounts to look at picture [ figuraspc ] .
the one dimensional minimal surfaces @xmath35 and @xmath36 intersect at the point p , thus defining @xmath37 , @xmath38 , @xmath39 and @xmath40 ; due to the minimality of @xmath19 and @xmath41 , it is immediate the result @xmath42 and @xmath43 , proving ( [ ssa1 ] ) .
analogously we can obtain ( [ ssa2 ] ) from @xmath44 and @xmath45 .
the above proofs unfortunately only apply to the case described . if @xmath9 , @xmath10 and @xmath11 for example are not collinear , and/or if the geometry is time dependent so that the curves bend in the time direction , then we will not generically have any intersection between @xmath35 and @xmath36 .
further , the time dependent case brings one additional complication as the hrt surfaces are in this case extremals , so even if we could actually find a way to compare areas , we would not be able to write down inequalities as if minimal surfaces were involved .
when lorentz invariance is preserved the ee is just a function of the proper length of the interval @xmath46 ; ssa1 implies monotonicity for the function @xmath46 while for collinear intervals ssa2 leads instead to concavity .
monotonicity is immediately proven by considering the special case of collinear intervals with proper lengths @xmath47 and
@xmath48 , then ( [ ssa1 ] ) just gives @xmath49 in fact it is also obviously true the opposite , that monotonicity implies ssa1 and the two conditions are equivalent .
when considering ee computed holographically by the area of ryu - takayanagi surfaces , monotonicity can in fact be obtained directly without passing through ssa1 , by showing that the proper length of a minimal surface ( or more generally any curve that minimize some fixed bulk functional ) is monotonically increasing as a function of the proper length of the boundary interval to which it is attached .
concavity is just barely more complicated to derive from ssa2 , see for example @xcite for the proof of the equivalence between concavity and ssa2 for collinear intervals .
again this property can be obtained directly , as for monotonicity , when dealing with the holographic minimal surfaces @xmath50 , whenever we vary the proper length @xmath51 of the boundary interval along the fixed space - like direction determined by the boundary end points .
this is obtained with a slightly different construction then the one of figure [ figuraspc ] ; let us consider a boundary straight space - like line and pick @xmath52 ( @xmath53 ) minimal curves ending on intervals of length @xmath54 that belongs to such line , displaced by a distance @xmath55 , as in figure [ figuramonmin ] ( @xmath51 does not need to be an integer multiple of @xmath55 ) .
we have that the proper length of @xmath56 , @xmath57 , obeys the inequality : @xmath58 which is proven as in figure [ figuramonmin ] , just noticing that the black curve has necessarily higher or equal proper length ( or whatever functional we are minimizing ) than @xmath59 , while each of the @xmath60 unions of a blue and a red arc is bigger than the proper length of what would be the corresponding minimal curve @xmath50 . in the limit of @xmath61 , while keeping @xmath62 at some finite value , the above formula reads @xmath63 that is just concavity for @xmath64 .
the case of non collinear intervals is more interesting . in appendix
[ a ] it is shown that , as a function of the slopes of the three intervals @xmath65 , the stricter bound on the ee @xmath66 from ssa1 comes from the case @xmath67 , and from ssa2 @xmath68 or @xmath69 ( or their parity transformed counterparts ) .
let us show what conditions on @xmath66 the strong subadditivity inequalities corresponds to for these configurations .
the ssa2 inequality has already been considered in @xcite for a configuration slightly less general then the one we are using , and we will follow a similar procedure . for reference for the parameterization consider the first picture of figure [ figuravssa ] ( the situation with @xmath69 is completely analogous ) ; defining @xmath70 , @xmath71 and @xmath72 , @xmath73 , and a function @xmath74 such that @xmath75 , by computing the proper lengths for the intervals appearing in ssa2 , ( [ ssa2 ] ) reads for this case @xmath76 as we always have @xmath77 ( as explained in figure [ figuravssa ] caption ) and obviously @xmath78 , the above inequality is just concavity for @xmath79 , so @xmath80 that is stronger than simple concavity for @xmath66 , due to the monotonicity required from ssa1 : @xmath81 . we can now follow a similar path for ssa1 applied to the configuration with slopes @xmath67 .
in addition , as the interval b appears only on the bigger side of the inequality , an even stronger bound can be obtained by further minimizing the proper lengths of @xmath24 and @xmath82 by considering an infinitesimal space coordinate length @xmath83 .
parameterizing this configuration as in appendix [ a ] , and defining @xmath84 , @xmath85 ( and correspondingly for @xmath86 ) and @xmath87 , we obtain the following result from ( [ ssa1 ] ) ( using @xmath83 to rewrite @xmath88 ) @xmath89 which is just monotonicity for @xmath90 .
however in this case , as a first derivative is involved , the condition implies nothing more than the usual monotonicity for @xmath66 : @xmath91 that ( [ con2 ] ) becomes stricter than concavity for non collinear intervals while ( [ con1 ] ) remains the usual monotonicity will soon have its counterpart : in various cases , holographically and not , for generic adjacent interval configurations ee will still satisfy ssa1 while the stricter ssa2 will be generically violated .
in the holographic description we will see as respecting ssa2 shall require certain geometrical conditions to be satisfied by the background . still studying static bulk geometries , with all the nice properties listed in the previous sections but this time in the lorentz signature , let us see what happens in the bulk when considering geodesics bounded to non collinear adjacent intervals , as for instance in figure [ figuraspnc ] .
the first clear issue is that geodesics ending on @xmath24 and @xmath82 do not generically intersect , and consequently try to prove ssa1 and ssa2 as in figure [ figuraspc ] does not work .
ssa1 however has an alternative proof derivable from the simple relation , @xcite and @xcite : @xmath92 where @xmath93 is the conformal scale factor of asymptotically ads metrics , reached by the geodesics at its vertex ( coordinates chosen so that we have large @xmath94 when close to the boundary ) . as it will be useful later ,
let us review briefly the derivation of ( [ r0 ] ) , following @xcite with some modification .
the assumptions are a conformal two dimensional boundary minkowski metric and translation invariance along a boundary space - like direction with associated killing vector @xmath95 .
the geodesic extremizes the action @xmath96 where the interval @xmath97 is taken to be at fixed boundary time @xmath98 .
we vary only the position of the final end point `` f '' by a purely spatial translation @xmath99 .
the variation of the action is just the boundary term @xmath100 with @xmath101 .
then @xmath102 as this quantity is conserved along all the geodesic we can evaluate it at its vertex , where @xmath103 simplifies by writing @xmath104 .
a brief computation leads to @xmath105 this equation is in fact true with or without time translation invariance .
if time translation is a symmetry of the bulk , we can further boost the above equation ( [ r0st ] ) to obtain ( [ r0 ] ) .
the important point here is that @xmath106 always , and thus @xmath66 is monotonically increasing .
one may then try to use this formula to check ( [ con2 ] ) by exploiting the dependence @xmath107 ; the result obtained by @xcite ( see also @xcite ) is that ssa2 in its stronger bound ( [ con2 ] ) is equivalent to the condition on the bulk geometry @xmath108 where @xmath56 is our geodesic ending on the interval of proper length @xmath51 and @xmath109 is a null vector perpendicular to @xmath95 .
, whenever einstein s equations are holding . ] our goal is to find a more geometrically transparent proof for the emergence of ( [ intncc ] ) , enhancing the difference between collinear and non collinear intervals and the role of @xmath110 , where the original derivation uses a metric ansatz and einstein s equation to show the equivalence between the explicit formulae for ( [ intncc ] ) and ( [ con2 ] ) . and
its intersection @xmath111 with the ( not shown ) achronal slice.,scaledwidth=100.0% ] so let us start from figure [ figuraspnc ] ; the first issue to solve is that the two curves @xmath35 and @xmath36 do not generically intersect .
let us then shoot out from @xmath35 ( that in our example has been chosen to be nowhere in the past of @xmath36 , if otherwise inverting the roles ) a congruence of null geodesics in the past boundary direction , that is then a codimension one surface @xmath112 .
given translation invariance in the boundary space direction , this null geodesics are chosen to be perpendicular to the corresponding killing vector , this in order to ensure always intersection between @xmath112 and @xmath36 at a point @xmath113 .
now consider any achronal slice ( with points either space - like or null separated ) that contains both the boundary interval @xmath24 and the point @xmath113 , and intersects @xmath112 along some curve ; this curve goes from the left end point of @xmath9 to @xmath113 , lets call this piece @xmath37 , and continues from @xmath113 to the right end point of @xmath10 , that we call @xmath38 as shown in figure [ figuraspnc ] .
the point @xmath113 also splits up @xmath36 into @xmath39 to the left and @xmath40 to the right .
here enters the raychaudhuri equation @xmath114 where , applied to our case , @xmath115 is the affine parameter along @xmath109 , @xmath116 is the shear and @xmath117 represents the variation of the line element of @xmath35 along @xmath115 , divided by the line element itself . as @xmath35 is a geodesic , @xmath118 on it ; then , as on the right hand side of ( [ ray ] ) all the quantities are negative definite but @xmath110 , the total proper length of @xmath35 decreases along @xmath115 if the integral of @xmath110 on @xmath35 is higher or equal then zero , which is just the integrated ncc condition ( [ intncc ] ) . , but also on all the evolution curves created along the flow by @xmath115 .
however we can require only the integral on @xmath56 if we restrict to boundary intervals with @xmath9 and @xmath11 of infinitesimal coordinate length .
this is also what was done in @xcite , although by a different road , and there it was also shown as having ssa2 respected for this infinitesimal configuration implies ssa2 valid for generic cases .
so only integration along @xmath56 is actually required .
] we finally obtain @xmath119 where the first inequality is guaranteed by the raychaudhuri equation coupled with ( [ intncc ] ) ; the second inequality instead , is the usual argument of figure [ figuraspc ] , but this time applied to extremal surfaces restricted to common achronal slices , one containing @xmath38 , @xmath39 and @xmath120 , and the other @xmath37 , @xmath40 and @xmath121 . on such a slice
extremal surfaces are minimal ( as will also happen when considering the maximin construction of @xcite ) , and that such slices exist on the present static case is what makes the proof applicable here but not on the corresponding time dependent situation , where the work will be harder .
this proves that the integrated ncc implies ( [ intncc ] ) for any adjacent interval configurations in static spacetimes . the reverse is true only for configurations maximizing the ssa2 bound , like the ones discussed in the previous section or in appendix [ a ] .
it is now an interesting question to ask if we can find a purely cft example of violation of strong subadditivity , using a time independent state @xmath0 ( the vacuum ) , as to my knowledge so far all the cases found in the literature rely on the holographic description , often with time dependent metrics . by analysing its holographic counterpart
we can isolate , inside the mechanics of strong subadditivity violation in the bulk , the dual of a genuine boundary cft violation rather then issues of the holographic formula of ee for possibly unphysical backgrounds .
in fact such an example exists and it is when a two dimensional cft has different left and right central charges . the computation of ee for the vacuum was done in @xcite , together with their proposal for the holographic counterpart that we will discuss in the next section , giving as a result for the ee in lorentzian signature ( otherwise the result is complex ) @xmath123 where @xmath124 is the uv regulator and @xmath125 ( or @xmath126 as in the notation of appendix [ a ] ) is the slope of the interval @xmath9 with respect to the constant time line , or equivalently the rapidity of the lorentz boost from a constant time interval to the actual @xmath9 .
let us start with the inequality ssa2 .
we expect the highest amount of violation , if any , by some of the interval configurations that maximize the bound from ssa2 , as explained in appendix [ a ] .
in fact , as the term proportional to @xmath127 inside ( [ clcree ] ) respects ssa2 as it is the usual two dimensional ee result from the vacuum of ( non - lorentz violating ) cfts , to have ssa2 violation we have better to manage an interval configuration that saturates the inequality for that term , or is close enough ( note that the uv regulator @xmath124 simplifies inside the strong subadditivity inequalities ) . among these configurations
the simplest one has the interval @xmath10 at constant time @xmath128 , and @xmath9 and @xmath11 of the same proper length @xmath129 with light - like slopes @xmath130 , @xmath131 ( or @xmath132 , @xmath133 ) . using ( [ clcree ] ) however , it is immediately clear that it respect ssa2 because @xmath134 and @xmath135 .
we then consider some deformations of it , either by studying a generic bound - maximizing structure like the first one in figure [ figuravssa ] , or one where the segment @xmath9 and @xmath11 are made space - like , as in the second picture of figure [ figuravssa ] .
let us start from the first configuration ; writing down the proper lengths of all the intervals as functions of the parameters @xmath136 and @xmath137 as appearing in the picture , ssa2 reads as follows @xmath138 @xmath139 are always positive in the second picture , while may be negative in the first with the obvious bounds for positive coordinate length for @xmath11 , @xmath140 , and for space or light - like interval @xmath10 , @xmath141 .
, scaledwidth=100.0% ] the terms proportional to the sum of the central charges and containing the dependence of the ee on the interval s lengths simplify leaving an inequality that depends only on the angles @xmath142s .
as the tangent of an angle between @xmath143 and @xmath144 is a monotonically increasing function , this inequality may be rewritten taking the tangent of the sums @xmath145 and @xmath146 , and using the usual formula for @xmath147 we obtain @xmath148 that in terms of @xmath149 reads @xmath150 @xmath151 every configuration out of this parameter range violates ssa2 . the second configuration instead leads to an ssa2 inequality that reads as follow : @xmath152 @xmath153 the @xmath154 parameter region that respects the above inequality is shown in figure [ figurasecssa2 ] for a specific example .
violation is obvious for a region increasing as @xmath155 becomes smaller than @xmath156 .
parameter values respecting the inequality ( [ 2ssa2 ] ) for @xmath157 and @xmath158 respectively ; the region decreases as @xmath155 becomes smaller .
for @xmath159 the inequality always holds in this example .
, scaledwidth=50.0% ] finally we inspect ssa1 for a generic ( adjacent ) interval configuration parameterized as in figure [ figuracldcr ] of appendix [ a ] .
the resulting inequality depends on the parameters @xmath160 and @xmath65 and it contains both logarithms and arcotangents .
we may split it in two , independent inequalities , proportional to @xmath156 and @xmath155 that can be analysed separately ( respecting both is a sufficient condition for respecting the complete equation ) .
we could nt provide a proof , but numerical analysis for various values of @xmath160 and the complete range of @xmath65 shows that no violation of ssa1 occurs . in the introduction and section [ init ]
we have discussed the quantum mechanical mechanism that may lead to the strong subadditivity violation in lorentz anomalous theories , as observed in the previous section .
also a detailed discussion on ee in anomalous theories can be found in @xcite , @xcite and @xcite .
here we want to understand how this mechanism acts when looking at the dual holographic theory .
the holographic description of cfts with @xmath122 is provided by a theory called topological massive gravity , tmg , with an action which is the sum of einstein - hilbert ( with possible cosmological constant ) and gravitational chern - simons with a real relative coefficient of @xmath161 , see @xcite and references within .
the holographic description of ee for these theories has been proposed to be given by a curve @xmath56 , with the usual boundary conditions and holonomy properties , but extremising a functional which is not its proper length but instead , in lorentzian signature , appears in front of the framing term and the functional becomes complex , as the corresponding boundary result ( [ clcree ] ) . ]
@xmath162 where @xmath163 and @xmath164 , with @xmath165 and @xmath166 the boundary coordinates .
the coefficient @xmath167 entering ( [ nfun ] ) is connected to the boundary difference between central charges as @xmath168 there are more solutions for @xmath56 than just geodesics , but geodesics are always a solution ; similarly metric solutions to the tmg s equations of motion are broader than proper einstein gravity solutions , but these are always solutions of tmg .
as the analytic continuation is not clear in the broadest context , we will restrict as in @xcite to usual geodesics in einstein metrics .
if we try to prove ssa2 as we have done for static spacetimes and general intervals in section [ faec ] and repeat the steps leading to ( [ sss ] ) , we immediately face a complication : even requiring the integrated ncc ( [ intncc ] ) , the raychaudhuri equation , that allowed us to minimize the proper length along the null geodesics flow of the congruence , does not constrain anymore the growth of the full functional ( [ nfun ] ) as it is no longer simply the proper length of the curve @xmath56 .
even try to enforce some modified energy condition to minimize ( [ nfun ] ) along the null congruence flow and so keep ssa2 valid appears difficult .
this because the term proportional to @xmath161 is essentially a boundary contribution .
we can speculate this being a sign of the essential difference between a violation of ssa2 due to using unphysical backgrounds , and thus entirely avoidable by appropriate bulk conditions , and a violation due to a pure cft mechanism with its dual holographic description .
finally let me discuss briefly on ssa1 .
as in the static case we have here that ssa1 remains valid ; the point is that ssa1 is generically a weaker condition than ssa2 . in the static case we showed that its bound on the ee did not change by varying the interval configuration , remaining monotonicity the condition on @xmath66 . here
the ee is no longer a function of the sole proper length of the interval , and the discussion and the proof do not apply any longer . still the moral appears to be the same , and an interesting development would be to try to generalize them to the present case of @xmath122 two dimensional cfts .
it is time to study time dependent situations , that for the cft means a time dependent quantum state @xmath0 while for the bulk a time dependent metric . for what concerns
the boundary side we unfortunately have not much to say that does not come from holography .
the reason being first the missing knowledge of the state for computable backgrounds ( as the vaidya example we will soon introduce ) , and second the non invariance by time translation that spoils the usual dependence of the ee by only the proper length of the interval .
this means that any ee computation , attempting to verify ssa1 and ssa2 , should pass through the understanding of the bulk description .
this we will now discuss in much more detail . when a time dependent background is considered , additional complications emerge in understanding the validity or not of strong subadditivity
. first of all hrt surfaces @xmath19 bend in the time direction and thus , even for collinear intervals , intersection does not generically happen .
second , inside the formula ( [ ryutak ] ) , @xmath19 does not refer any longer to minimal surfaces as in the euclidean static case , but to extremal ones .
consequently , even if we had intersection , we could not straightforwardly construct area inequalities as so far done ( unless we could `` project '' the extremal surfaces @xmath35 and @xmath36 to a certain common achronal slice containing @xmath120 and @xmath121 , where they become minimal .
we will soon see that non existence of this slice is the key point for failure of ssa2 . ) .
a first understanding on how and when ssa1 and ssa2 are valid comes from using formula ( [ r0st ] ) , that is the specific case of ( [ r0 ] ) when the boundary interval and its variation are at fixed time . in this case , as we have previously derived , ( [ r0st ] ) is valid even without requirement of time translation invariance .
further in this case we still can write the ee as a a function of the interval length @xmath169 , @xmath170 , with the dependence on the boundary time @xmath98 decoupling when studying inequalities at the same @xmath98 ( while for generic space - like intervals , without time translation invariance , the ee is a generic function of all the coordinate end points , not only the proper length ) .
the positivity of the conformal factor at the vertex @xmath93 is just the monotonicity property that guarantees the validity of ssa1 , but what about ssa2 ? clearly to have the ee @xmath170 concave as a function of @xmath169 , we should have that @xmath171 ( at fixed time @xmath98 ) is monotonically decreasing or , by inverting the function for given @xmath98 , that @xmath172 is monotonically decreasing . is close to the boundary and small @xmath94 to the center of the bulk .
] this means that violation of concavity and ssa2 happens for geodesics whose vertex moves towards the boundary when extending the size @xmath169 of the interval , for fixed @xmath98 . to understand better this point we can use a theorem by wall , @xcite theorem 17 , that says that if ncc is valid , two geodesics ( maximin surfaces in the paper as we will discuss later ) ending on space - like boundaries one contained in the other , @xmath173 and @xmath174 in our case , will always be at space - like distance one from the other with the smaller one inside , towards the boundary , with respect to the bigger .
so ssa2 is respected . instead giving up ncc
not only time - like distances are possible but , in order to have violation of ssa2 the narrower geodesic should have its vertex extending in the bulk further then the larger one . to be more concrete
let us consider an example of time dependent , asymptotically ads background , where analytic computations are possible .
this is the vaidya metric , representing the collapse of a mass shell that interpolates between an ads metric and a btz black hole .
this example has been extensively studied in the past by @xcite , @xcite and @xcite , where the first connection between the ncc and the violation of ssa2 was established .
the metric is @xmath175 which is an ads solution with the addition of a local energy momentum tensor whose only nonzero component is @xmath176 the case we consider is when @xmath177 is a step function on @xmath178 and @xmath179 becomes a delta . if the delta is positive ( @xmath180 ) the metric will be ads inside for @xmath181 and btz outside for @xmath182 ; if instead the delta is negative ( @xmath183 ) we have btz inside for @xmath181 and ads outside for @xmath182 .
the first case satisfies the ncc , the second violates it , as can be trivially checked . from now on @xmath184 .
geodesics that starts and ends on the boundary and cross the mass shell ( otherwise they are simply contained in the static ads or btz depending on the choice for @xmath177 ) depend on two parameters that , following @xcite , we choose to be @xmath185 and @xmath186 , the bulk radius at which the geodesic crosses the shell and a conserved momentum for the space translation symmetry that turns out to correspond to the radius of the vertex .
we start with backgrounds respecting ncc .
the goal is to verify the monotonically decreasing behaviour of @xmath172 , where @xmath187 as explained in appendix [ b ] and where relevant formulas can be found ; @xmath169 is there called @xmath188 and is a function of @xmath185 and @xmath186 .
we solve for @xmath185 to give , for the geodesic with a certain value of @xmath186 , the chosen boundary time @xmath98 , and then plot @xmath189 for the given value of @xmath98 .
some sample curves are represented in figure [ figuratra2 ] where we can check the monotonically decreasing behaviour , as expected .
, for vaidya background respecting ncc.,scaledwidth=100.0% ] more interesting is the case for ncc violating geometries . here
formulas depend on the range of value for the parameters , with three possible cases .
case 1 : @xmath190 ; case 2 : @xmath191 and @xmath192 ; case 3 : @xmath193 and @xmath194 with @xmath195 a certain function of @xmath185 and @xmath186 whose formula and meaning is in appendix [ b ] .
some sample plots are in figure [ figuratra ] , where we can see both monotonically decreasing curves , for geodesics belonging to case 1 and thus respecting ssa2 , and monotonically increasing for geodesics in cases 2 and 3 , so violating ssa2 . , for vaidya background violating ncc .
the pictures refer to case 1 , case 2 and case 3 respectively . , scaledwidth=120.0% ] let us now discuss the most general scenario , with time dependent backgrounds and generic adjacent interval configurations .
the starting point for the discussion is the paper by wall @xcite , where the author introduces the concept of maximin surfaces , as an equivalent description of the extremal surfaces @xmath19 .
a codimension two maximin surface is defined starting with a generic codimension one achronal surface @xmath196 in the bulk containing the boundary of the interval @xmath9 ; then a maximin surface is the minimal codimension one surface on @xmath196 having maximal area when varied over all the possible @xmath196 . in brief
the extremality is achieved by minimizing and maximizing in the space and time directions respectively .
the power of the maximin construction is exploited in the theorem that states ( under generic assumptions for the bulk spacetime ) that ssa2 is valid if the ncc condition @xmath197 is respected .
the original proof in @xcite uses ncc and the raychaudhuri equation for constructing inequalities between areas on different achronal slices with the same boundary condition .
this result may be obtained as well just by using the maximin construction alone , so in order to single out the places where the ncc necessarily enters as a condition for ssa2 , and to be naturally lead to the main claim of this section , we slightly modify the proof . skipping some details , that may be found in the original paper ,
it is : proof : theorem 4 of @xcite states that , given two null congruences of geodesics @xmath198 and @xmath199 , in our case obtained by shooting out null curves from maximin surfaces , with @xmath199 nowhere in the past of @xmath198 and touching at a point @xmath113 belonging to some achronal slice @xmath196 , then there exists a sufficiently small neighbourhood of @xmath113 , @xmath200 , such that either @xmath201 or @xmath198 and @xmath199 do coincide there . given this general result , theorem 14 states that two maximin surfaces with space - like boundary condition ( attached to @xmath120 and @xmath121 in our case ) , are always at space - like distance if ncc holds ; the idea is simply that , as maximin surfaces are extremal , on @xmath120 and @xmath121 we have @xmath202 ; then starting from a situation where the two curves are always at a space - like distance between them , not only close to the boundary but all the way through the bulk , let us suppose we can continuously deform the curves , for example enlarging @xmath203 while keeping @xmath10 fixed , to a situation where somewhere the proper distance approaches the null value ; this means that two points @xmath204 and @xmath205 , one for each curve , are connected by a null geodesics ( for symmetry there is either a single @xmath113 corresponding to the vertex or two symmetric points on the right and left hand side of the vertex ) .
this is shown in the first picture of figure [ figuratld ] . and
@xmath121.,scaledwidth=90.0% ] let us choose @xmath121 to be nowhere in the past of @xmath120 and focus on @xmath204 ; theorem 4 says that in a neighbourhood of @xmath204 the null congruence @xmath206 . as ncc is valid , @xmath207 and this contradicts the extremality condition .
not being able to continuously approach a light - like distance between any two points , it is then implied that the two curves can not be deformed to develop time - like distances either , as shown in the second picture of figure [ figuratld ] .
this point is crucial for proving theorem 17 that shows that both @xmath120 and @xmath121 always belong to the same achronal slice @xmath196 .
finally the key passage is that , using the maximin construction , there exists a representative for both @xmath35 and @xmath36 on @xmath196 of smaller area ( proper length ) then the maximin surfaces , @xmath208 ( and analogously for @xmath36 ) .
thus on @xmath196 we have @xmath209 where the last inequality is just the usual static argument ( that can now be used as on @xmath196 the surfaces @xmath120 and @xmath121 are minimal and @xmath210 and @xmath211 intersects . ) .
this formulation of the theorem makes evident that a necessary condition for violation of ssa2 is to develop non space - like distances between @xmath120 and @xmath121 .
this is understood as ncc enters the above proof only once , in constraining @xmath120 and @xmath121 to belong to the same achronal slice @xmath196 while remaining at space - like distances .
the importance of this result , that is the non space - like distance between two surfaces @xmath212 and @xmath212 , with the domain of dependence of @xmath213 , @xmath214 , containing @xmath215 , @xmath216 , resides in being a counter example to statements sometimes used in the past literature , for instance conjecture c2 of @xcite .
we would like to emphasize the difference between the local ncc energy condition obtained in the present section , and the integrated ncc that we obtained for static spacetimes .
the former one is clearly more restricting than the second , as respecting the local ncc obviously implies the integrated ncc , but not the opposite . in fact we could have used the maximin construction for proving that local ncc implies ssa2 for static backgrounds ( which is true ) , but the maximin construction requires the local ncc to be applicable ( for example in proving the equivalence with hrt surfaces ) .
so in order to have the theorem as strong as possible , by requiring the weakest energy condition , we proceeded there without this powerful tool .
given the generic discussion of the past section , let us construct a concrete example , where we show the developing of either null or time - like distances between the two geodesics @xmath120 and @xmath121 , when ncc does not hold and ssa2 is violated , while in general maintaining space - like distances when ncc is valid ( although it is possible to have ssa2 respected in the former case as the condition is necessary but not sufficient ) . in this section
we will study only situations where both the @xmath120 and @xmath121 geodesics belong to the same parameter range ( case 1,2,3 as previously introduced , see also appendix [ b ] ) , even though in appendix [ b ] formulae are provided for the most general scenario
. our goal will be to probe the distance between the vertices of @xmath120 and @xmath121 to check if they are at space - like distance or not , depending on the value of the parameters @xmath185 and @xmath186 of both curves
. we will not consider distances between generic points on the geodesics as it would be excessively complicated to derive corresponding inequalities and ultimately unnecessary , as non space - like distances between the two curves is just a necessary but not sufficient condition for ssa2 violation .
as the boundary conditions force space - like distance between the end points of @xmath120 and @xmath121 , the vertices are likely to be the most prone to develop either null or time - like distances between them , so we restrict to this case .
further to simplify an otherwise too complicated computation we will consider only collinear intervals at the same boundary time @xmath98 and symmetrically disposed around the central point of @xmath10 , that is @xmath9 and @xmath11 are taken to be of the same length .
in appendix [ b ] we derived inequalities for the parameters @xmath185 and @xmath186 of two geodesics @xmath120 and @xmath121 that , when respected , correspond to space - like distances between the vertices .
the constrain that comes from having both curves end points at the same boundary time @xmath98 eliminates one of the four parameters and further partially restricts the available parameter space that comes from the other three .
the results are the following .
+ * metric that violates the ncc * : + we here generically name `` 1 '' and `` 2 '' the labels for the two geodesics attached to the two parameters @xmath185 and @xmath186 . as the goal is just to show
if the distance between the vertices is space - like or not , it is not really relevant to distinguish which one corresponds to the geodesic attached to the largest boundary interval @xmath121 , and which to @xmath120 .
however , as the intervals are collinear , we know that for case 1 smaller value of @xmath186 corresponds to bigger value for @xmath217 , so that parameter belong to @xmath121 . for case 2 and 3 instead , the smaller @xmath186 belong to @xmath120 .
* case 2 : @xmath218 + the inequalities defining space - like distance between the vertices are ( see appendix [ b ] ): + @xmath219 + where we have defined @xmath220 in a parameter space spanned by @xmath221 with @xmath222 and @xmath223 .
@xmath224 is a function of the other three parameters , as explained in appendix [ b ] .
we can sweep this parameter space to look for what volume satisfies the inequalities ( [ in21a ] ) and ( [ in22b ] ) and what violates them .
the result is that they are always violated as shown in figure [ figura5 ] and thus geodesics always develop non space - like distances .
+ ) and ( [ in22b ] ) , thus giving null or time - like distances between the vertices ; the second and third show the ( empty ) region that satisfies ( [ in21a ] ) and ( [ in22b ] ) respectively , thus giving space - like distances .
the small angle missing is due to the requirement @xmath225 . ,
scaledwidth=80.0% ] * case 3 : @xmath226 + the inequalities are as previously : + @xmath227 + inside a parameter space @xmath228 with the additional constrain @xmath229 and again @xmath224 being a function of the other three parameters . again always violation , as shown in figure [ figura5b ] + ) and ( [ in32b ] ) .
note that the parameter space is different from the previous section not only in the values of the domains for @xmath230 but also in shape . , scaledwidth=80.0% ] * case 1 : @xmath190 + inequalities here are slightly different : + @xmath231 + with the parameter range for @xmath230 limited by the constraints @xmath232 and @xmath233 , and @xmath224 again a function to match the boundary time @xmath98 . inside this volume
we separately deal with the subspaces @xmath234 and @xmath235 ; now we can apply ( [ in11a ] ) and ( [ in12b ] ) on the relevant space and see if they are satisfied or not .
it turns out that when @xmath235 , @xmath236 is always respected , while @xmath237 is respected in some region and violated in its complementary , as is shown in figure [ figura7 ] . thus here
we face geodesics that both present vertices at space - like and non space - like distances , depending on the values of their parameters .
+ case : agreement and violation of @xmath237 divide the total parameter space in two complementary subspaces .
as the other inequality of ( [ in11a ] ) is always respected , we have space - like distances in the first case , and null or time - like in the second.,scaledwidth=60.0% ] + correspondingly when @xmath234 @xmath238 is always respected , while @xmath239 is respected in some region and violated in its complementary ( with respect to the total space ) .
this is shown in figure [ figura8 ] .
+ : agreement and violation of @xmath239 divide the total parameter space in two complementary subspaces . as the other inequality of ( [ in12b ] )
is always respected , we have space - like distances in the first case , and null or time - like in the second.,scaledwidth=60.0% ] * metric that respects the ncc * : + the space - like condition for distances between vertices is here given by the inequalities @xmath240 with only @xmath230 independent and @xmath235 . the parameter space defined by the usual condition for the boundary time and either ( [ csl1 ] ) or ( [ csl2 ] )
may be obtained numerically , in chosen parameter domains .
the result is that ( [ csl1 ] ) and ( [ csl2 ] ) are always respected implying vertices are at space - like distance one from the other .
we have discussed in detail the holographic description of the two strong subadditivity inequalities , from static backgrounds to time dependent , with collinear boundary intervals or general configurations .
we have seen that , for static geometries , ssa1 and ssa2 are always respected for collinear intervals , that the second requires an integrated ncc for non collinear intervals and is generically violated when we have lorentz anomaly , while ssa1 holds .
this has its counterpart in the monotonicity condition for ssa1 that remains unaltered , while the concavity for ssa2 transforms into a more strict requirement abandoning the collinearity .
new results are the geometric proof for concavity of minimal surfaces , the proof that ssa1 implies only monotonicity independent of the interval configuration , a new proof that ssa2 requires the integrated ncc using the raychaudhuri equation and finally the violation of ssa2 but not of ssa1 ( here only numerical ) for lorentz anomalous cfts . for time
dependent backgrounds we first provided a new simple strategy for understanding if and when ssa1 and ssa2 violation occurs with collinear intervals , that does not requires direct checking of monotonicity or concavity .
second we have reviewed , in a slightly different form , the result by wall that local ncc implies validity for ssa2 , while making manifest the connection with the energy condition by showing that the reason for the violation comes from the developing of null or time - like distances between the most inward and outward geodesics entering the ssa2 inequality .
furthermore we have provided an explicit example in both cases by using the vaidya metric . also some discussion on why violation of strong subadditivity
occurs has been provided .
following are two appendixes containing the results and the proofs ( to my knowledge new ) for what interval configuration , as a function of the slopes , gives the strongest bound on the entanglement entropy inequalities ssa1 and ssa2 , and explicit formulas ( and some derivation ) for the vaidya metric example .
an interesting point of view , that we did not discuss but is worth mentioning , is the result from @xcite where it was shown that virasoro conditions in bosonic string theory imply the ncc on the background geometry .
we would like to suggest that , perhaps , this is a hint that energy conditions may have some uv justification .
the hypothesis is that non respecting ncc ( or analogous conditions ) means that the metric we are using does not consistently arise as a background in theories that correctly quantize gravity , and one of the dual symptoms is not respecting strong subadditivity .
we would like to point out three possible hints for future research .
the first one is the problem of quantum bulk corrections to ee and the question if they do respect or not the boundary strong subadditivity inequality .
for example we can introduce the mutual information @xmath241 and rewrite the ssa2 inequality as , @xcite @xmath242 if the intervals @xmath9 and @xmath10 entering the mutual information are well disconnected , the classic holographic computation gives @xmath243 , as @xmath244 .
thus , for @xmath9 , @xmath10 and @xmath11 disconnected the inequality ( [ inmut ] ) just produces a classical @xmath245 result .
quantum corrections in the bulk clearly affect the above inequality ; how to compute them is still a discussed argument as the hrt ( or ryu - takayanagi ) surfaces are not string worldsheets as for the holographic description of wilson loops , but rather just geometrical surfaces .
thus @xmath246 corrections or higher genus computations are not the correct answer .
the question has so far received two different answers in @xcite and @xcite , with pro and con arguments for both .
the question is then if there are reasons to believe that quantum bulk corrections do respect or not strong subadditivity , and/or should they be constrained by energy conditions in doing so ? a second direction for future research , and also one of the original motivations for the present paper , is the question of what is the logical relationship between strong subadditivity and energy conditions .
we know that the integrated ncc implies ssa2 for static geometries with generic adjacent intervals , and that the two are equivalent for a ssa2-bound - maximizing configuration , and we also know that local ncc implies ssa2 in time dependent systems
. can we deform either the energy condition or the strong subadditivity , weakening or strengthening depending on the case , in order to make the correspondence one to one , in the widest possible range of cases ?
finally we would like to generalize as most as possible of the present paper to generic d - dimensional theories .
part of this work is straightforward , part quite complicated .
we hope to come back to these issues and more in the future .
i would like to thank aron c. wall and horacio casini for email correspondence at the early stage of this work , and diego trancanelli for suggestions and comments on the paper .
this work was founded by fapesp fellowship 2013/10460 - 9 .
in this appendix we want to study what are the configurations of one dimensional adjacent connected intervals @xmath247 such that the ee is maximally constrained by the inequalities of strong subadditivity ( [ ssa1 ] ) and ( [ ssa2 ] ) .
as explained in @xcite , in theories that respect causality and unitarity , the dependence on the space - like interval @xmath9 of @xmath248 can be only through the causal domain of dependence of @xmath9 that , in two dimensions ( without global issues ) , is determined exclusively by the two endpoints of the interval ; if the theory has lorentz invariance the dependence should be further restricted to the proper distance between them .
this is the only assumption we will make on @xmath249 , that does not come from strong subadditivity .
the requirement for @xmath247 to be space - like ( or light - like in some appropriate limit ) may , at first sight , appear excessive ; we can , for example , consider time - like @xmath9 and/or @xmath11 while still having space - like separated end points of @xmath24 and @xmath82 ( and consequently also @xmath203 ) , and an apparently meaningful ssa2 inequality ( [ ssa2 ] ) . with space - like end points for @xmath24 and @xmath82 and consider the ssa1 ( [ ssa1 ] ) .
] one way to see that this is not possible is from the explicit proof of ssa1 and ssa2 @xcite , where hilbert spaces associated to all the three @xmath250 and @xmath11 are involved .
another way is to show contradiction for ( [ ssa2 ] ) . and @xmath11 are time - like such that the end points of @xmath24 and @xmath82 are at barely space - like distances .
note that in this example the intersection between the causal domains of @xmath24 and @xmath82 determine a region whose most outwards end points in the space direction fix what we call @xmath251 ; this segment does not coincide with @xmath10 , as it should be by causality and would indeed happen if @xmath9 and @xmath11 were space - like
. violation of ssa2 occurs .
, scaledwidth=70.0% ] consider the picture [ figuraa ] where @xmath9 and @xmath11 are time - like such that @xmath24 and @xmath82 are barely space - like , that is with positive proper distance tending to zero . the intersection between their domains of dependence should determine the space - like segment @xmath10 , but it clearly does not . if one still tries to use ( [ ssa2 ] ) , written to respect causality as @xmath252 , ends up with an inequality where the proper lengths @xmath253 , @xmath254 and @xmath255 can be made arbitrary small while @xmath256 in general remains finite . as we know that @xmath249 is a monotonically increasing function , this clearly leads to a contradiction .
let us label the different possible configurations of intervals @xmath247 by the signs of their slopes : @xmath257 for positive and @xmath258 for negative .
we have eight cases that , assuming parity invariance for the space @xmath166-coordinate direction , are related as below @xmath259 as in figure [ figuracldcr ] we call @xmath260 and @xmath261 the space @xmath166 and time @xmath165 coordinate distances , with @xmath142 the slope defined by the counterclockwise angle measured from the @xmath166-axis ( as usual ) , @xmath262 , @xmath263 , @xmath264 . time and space distances of composed intervals , for instance @xmath24 , will be the sum of the corresponding ones for @xmath9 and @xmath10 ; we will call slope of any interval the ratio between the time and space total distances .
the case @xmath265 has been already considered in section [ cft ] , to fix @xmath66 to be monotonically increasing and concave , so these properties will be assumed .
further we will work in lorentzian signature so , for instance @xmath266 etc ... the general strategy is to increase ( resp .
decrease ) as much as possible the right - hand ( resp .
left - hand ) sides of ( [ ssa1 ] ) and ( [ ssa2 ] ) , by increasing ( resp . decreasing ) the proper lengths of the intervals is monotonically increasing .
] by varying @xmath65 while leaving @xmath160 generic ( they will be adjusted in some cases ) .
if a variation of some parameter @xmath142 acts in such a way to contemporary grow the smallest side of the inequality ( either ssa1 or ssa2 ) and reduce the largest , then this variation is applied as it makes the bound from that inequality on the function @xmath66 stronger .
instead it may happen that changing some @xmath142 s leads to either increasing or decreasing both sides of the inequality at the same time .
still , using the fact that @xmath66 is monotonically increasing and concave , if we have two intervals with proper lengths @xmath267 and @xmath268 depending on that @xmath142 , @xmath269 , on opposite sides of either ssa1 or ssa2 , there are cases where , without knowing the exact form of the function @xmath66 , we can still say if @xmath270 augments or diminishes : @xmath271 every other variation of @xmath267 and @xmath268 leads to an undetermined sign for @xmath270 .
let us start with the second inequality ( [ ssa2 ] ) and consider the possible configurations listed in the left side of ( [ ppp ] ) .
from any of the pictures of figure [ figuracldcr ] we can immediately understand that minimizing @xmath253 and at the same time maximizing @xmath272 by varying @xmath126 ( or equivalently with @xmath273 ) , is possible only if @xmath274 and @xmath275 ( increasing @xmath126 ) or if @xmath276 and @xmath277 ( decreasing @xmath126 ) .
if any of these conditions is fulfilled we obtain a stronger bound from ssa2 by changing @xmath126 until it reaches its limiting value @xmath278 or the value for which the condition itself is no longer valid ; if @xmath126 has reached a value different from @xmath278 , or if the above conditions were not fulfilled from the very beginning , the way to maximize @xmath279 is to compare derivatives , as on the right hand side of either ( [ der1 ] ) or ( [ der2 ] ) . in the present case @xmath280 , thus if it is strictly positive and @xmath281 we should increase @xmath126 either to @xmath282 or to the maximum value for which @xmath203 becomes flat . if @xmath283 ( either as an initial condition or by effect of the growth of @xmath126 ) , then @xmath284 and @xmath285 .
the condition to maximize @xmath279 is @xmath286 that simple algebra transforms into @xmath287 this means that if the slope of @xmath24 , @xmath288 , is higher or equal then the slope of @xmath203 , @xmath289 , then a continuous increment of @xmath126 will lead to a continuous maximization of @xmath279 , until @xmath126 reaches its limiting value of 1 .
thus we always end up with @xmath130 .
this rule that higher slope for an interval means higher rate of change for its proper length when compared with a second interval sharing the same end point , under a variation of the associated @xmath142-angle , is general and applies even when both slopes are negative , just considering absolute values . if @xmath290 the effect of varying @xmath126 would be undetermined , unless we knew the exact form of @xmath66 . after having sent @xmath291 ,
we focus on @xmath292 in order to maximize @xmath293 .
we can have either positive , null or negative @xmath294 ; if @xmath295 then @xmath296 and it can be either higher or lower ; if @xmath297 the sign of @xmath289 may be positive or negative ( in which case its absolute value is certainly lower than the one of @xmath294 ) .
in all the above cases but one , sending @xmath298 will result in stricter ssa2 inequality , either because of contemporary maximization of @xmath272 and minimization of @xmath299 , or because the rate of decrease for @xmath299 is higher then for @xmath272 ; the only exception being when @xmath300 but lower then @xmath289 .
thus we are led to two cases @xmath301 @xmath302 the final step is to send @xmath303 in the first case . as this is not possible by only acting on @xmath292 , as it may lead to a decrease of @xmath293 , we will contemporary change the value of @xmath304 in order to keep invariant @xmath299 under the change of value of @xmath292 .
that there is a solution for @xmath304 satisfying this condition is guaranteed by the fact that , changing @xmath304 , @xmath299 ranges between @xmath305 and infinity .
this operation keeps invariant everything in ( [ ssa2 ] ) but @xmath272 that a simple computation shows increases .
thus for every set of parameters belonging to the first case above , there is another set of values with @xmath303 and the new @xmath304 , such that ssa2 with these new parameters gives a stricter bound .
the discussion is parallel to the previous section .
we can have either a positive or negative value for @xmath288 and @xmath289 ; in all cases , but when @xmath289 is positive and higher of @xmath288 ( also positive ) , we can move @xmath306 in order to create a stronger bound from ssa2 . on the other side ,
as we always have @xmath295 and either positive but lower , or negative @xmath289 , an increase in the value @xmath307 will certainly lead to a stricter ssa2 inequality .
thus the two possibilities are @xmath308 @xmath309 working in a parallel way to the corresponding ppn case we can make the first case above to @xmath310 , by adjusting the value of @xmath311 .
if @xmath312 , then a continuous increment of @xmath126 will lead to a continuous maximization of @xmath279 , until @xmath126 reaches its limiting value of 1 .
the increase of @xmath293 by varying @xmath292 works similarly , only now we may have @xmath130 ; thus we end up with any of the following cases : @xmath313 @xmath314 @xmath315 it is now possible to reduce all the above to either a ppn or npp configuration by sending the undetermined @xmath142 ( @xmath292 , @xmath126 and either of those for respectively the three cases ) to minus its value , while modifying the corresponding @xmath166 coordinate in order to keep the proper length of either @xmath82 or @xmath24 invariant , as already done in the previous cases .
this operation keeps invariant everything in ( [ ssa2 ] ) but @xmath272 that a simple computation shows increases ( this is not true for a general set of positive @xmath65 , but can be proved in the above listed configurations ) .
thus for every set of values belonging to the three cases above , there is another set of values with @xmath316 and the new @xmath317 , such that ssa2 applied to these new parameters gives a more strict bound than the older case .
then respectively ppp passes to either ppn ( first case , @xmath130 ) or npp ( second case , @xmath133 ) or either ppn or npp , as we wish ( third case ) . from here
on , the procedure of modifying parameters to strengthen the ssa2 inequality bound proceeds as in the previous corresponding sections . here @xmath318 may be positive or negative . in the first case @xmath319 ( as well as @xmath320 ) so we can reduce @xmath253 at an higher rate than @xmath272 by sending @xmath321 and npn becomes ppn ( with now the positive @xmath289 higher then @xmath294 , thus @xmath292 remains undetermined ) . in the second case
@xmath288 is either positive or negative ; if it is positive again @xmath321 ( decreasing @xmath253 and increasing @xmath272 until it becomes eventually flat , and then decreasing it as well but at a lower rate ) so again ppn with undetermined @xmath292 .
if instead @xmath322 , its absolute value is lower then @xmath289 , then @xmath126 can not be changed and we consider @xmath294 . if it is positive @xmath323 and we obtain npp ; if it is negative we have @xmath288 and @xmath294 negative and their moduli lower than @xmath289 , so @xmath324 are undetermined : @xmath325 @xmath326 @xmath327 @xmath328 in the last case
we can apply the same procedure used in the previous section , that is sending ( for example ) @xmath329 while changing @xmath304 in order to keep fixed @xmath299 .
again it can be checked that this increases @xmath272 , and thus we end up with npp . from here on we work as in the first two sections .
we can finally say that making ssa2 inequality stricter leads to either ppn with @xmath330 or npp with @xmath310 , that is the a and c segments become light - like and on the same `` time side '' with respect to b. the slope of b remains general and positive ( as the negative values are covered by the parity transformation ( [ ppp ] ) ) . to make the ssa2 bound stronger , but the result always leads to an undetermined change for the inequality , and even the trick of modifying some of the @xmath166 coordinate as done before does not help ( as either there are non positive solutions or the new set of parameter leads to a weaker bound ) .
the analysis deals with all the single cases and it is quite boring , so we avoid presenting it in detail . ] here the work is easier as the @xmath10 interval only appears on one side of the inequality .
moreover the sign on @xmath126 and @xmath292 only affect one side as well .
this leads to the choice of ppp over any other configuration ( @xmath331 and @xmath332 are unaffected by @xmath333 but sign changes leading to the ppp configuration always decrease @xmath253 and @xmath299 , everything else being fixed ) , and @xmath334 ( it reduces both @xmath253 and @xmath299 while keeping invariant @xmath331 and @xmath332 ) .
a further attempt to shift the value of @xmath324 towards zero in order to decrease @xmath335 and @xmath336 fails , as the slopes of @xmath9 and @xmath11 are always lower ( @xmath10 is light - like ) then respectively the slopes of @xmath24 and @xmath82 ( and we do not fit the criteria of equation ( [ der2 ] ) , nor does it work the trick of changing the value of @xmath337 to keep @xmath253 fixed while increasing @xmath331 , as the solution for @xmath337 is in general not positive ) .
so the most strict bound on @xmath66 from ssa1 , with generic @xmath160 , comes from the configuration @xmath338 .
the main reference is @xcite .
i will often refer to formulas written there while trying to remain consistent here .
the metric of vaidya is @xmath175 which is a solution with negative cosmological constant and an energy momentum tensor with non - zero component @xmath176 we will pick @xmath177 to be a step function centerd at @xmath178 and @xmath179 a delta function .
if the delta function is positive ( @xmath180 ) the metric respects the ncc ( see for example section 3.3 of @xcite ) and it is given by ads inside for @xmath181 and btz outside for @xmath182 ; if instead the delta is negative ( @xmath183 ) the ncc is violated and we have btz inside for @xmath181 and ads outside for @xmath182 . using a scaling symmetry ( see eq . 3.18 of @xcite )
we can fix @xmath184 .
the physics is better understood in @xmath339 coordinates , where @xmath165 is related to @xmath340 by the definition @xmath341 with @xmath342 , having set @xmath343 for ads and @xmath344 for btz .
the solutions are @xmath345 the shell at @xmath178 moves towards the center at @xmath346 as the time @xmath165 increases , transforming , as it moves , ads into btz if ncc is respected , or the opposite when ncc is violated .
note that the coordinates @xmath94 and @xmath340 are continuous across the shell , while @xmath165 has a finite discontinuity .
the metric in @xmath339 coordinates is we want to solve the equation for a geodesic when ncc is violated ( see also appendix a of @xcite , while the easier case for ncc respected is treated in section 3.2.2 ) .
the case we will consider is when the geodesic s end points on the bulk s boundary are at the same time @xmath98 , and separated by a space distance @xmath348 .
as the metric is @xmath166 independent there is a conserved charge @xmath186 ; further the metric is also @xmath165 independent but when crossing the shell , thus we can use also a charge e , conserved independently in ads and btz . they can be computed to be @xmath349 where the overdot represents a derivative with respect to the affine parameter @xmath350 defined on the space - like geodesic as @xmath351 as the lagrangian along the `` time '' @xmath350 is @xmath352 we can write the equation @xmath353 that leads to the two equations ( the second obtained dividing by @xmath354 and rewriting @xmath355 ) @xmath356 there are two additional equations we need to derive relating the internal and external @xmath94 derivative on the shell ; they are obtained by locally extremizing the geodesic by varying the intersection point with the shell , @xmath357 ( @xmath358 ) given two generic nearby geodesic points , one in ads and one in btz @xcite .
the geodesic will experience a refraction with the difference between the two @xmath359 as @xmath360 the space - like geodesics with two end points on the boundary and crossing the shell ( twice ) are the ones we are interested in ; these geodesics are symmetric with respect to the axis that passes through their vertex and the middle between the two end points ( as we will fix the two end points to be at the same time @xmath165 ) .
thus also the two crossing points on the shell are at the same time @xmath165 ( different by the one at the boundary ) , so @xmath361 inside the shell and for btz @xmath362 .
this simplifies the computation . equating @xmath363
we have derived an equation for @xmath195 : @xmath364 using the first of ( [ eqn2 ] ) , with @xmath365 @xmath366 plugging ( [ a ] ) inside ( [ ep ] ) and using the second of ( [ eqn1 ] ) for btz to find an expression for @xmath367 depending only on @xmath368 , we arrive at @xmath369 a geodesic with nonzero energy will depend on two parameters , the radius of the two crossing points with the shell @xmath185 and the momenta @xmath186 , that we will soon see to coincide with the radius of the vertex of the geodesic . if the geodesic does nt cross the shell , or if we are concentrating on the arc after the crossing , for symmetry @xmath370 .
that geodesic will then depend only on @xmath186 in the first case , or on @xmath186 and an additional parameter @xmath371 in the second .
this @xmath371 will be then tuned to be a function of @xmath186 and @xmath185 as we want this geodesic arc , going from the shell inward , to be glued correctly to two symmetric pieces of a second geodesic coming from the boundary to the shell and having @xmath185 and the same @xmath186 as parameters .
an example of three different geodesics in different parameters range is shown in figure ( [ figura1 ] ) . above the shell represented by the plane @xmath178 is ads ( positive @xmath340 ) , below is btz ( negative @xmath340 ) .
the boundary of vaidya is at @xmath372 , its center at @xmath346 . at the shell
we see the refraction , where the geodesics bend at an angle ; the solution at negative @xmath340 is the one for btz , it depends on @xmath186 and a parameter @xmath371 that is adjusted so that this solution meets the one in ads at the radius @xmath185 for the same two values of the affine parameter , @xmath373 and @xmath374 .
more properties will be discussed later.,scaledwidth=60.0% ] let us start by solving the geodesic equation for btz in @xmath375 given by the first of ( [ eqn1 ] ) .
there are two types of solutions , either going from the boundary towards the center of btz and back , or starting from the center and moving towards the boundary and back , with the vertex that will be either at smaller or higher @xmath94 then @xmath185 ; further its value will be respectively @xmath186 for @xmath376 or one for @xmath377 in the first case and one for @xmath376 or @xmath186 for @xmath377 in the second ( also note that the first solution is ( 3.36 ) of @xcite with @xmath370 when @xmath378 . ) .
the values of @xmath373 and @xmath374 at which the geodesic reaches the shell at @xmath178 and @xmath380 will be given later ; the value of @xmath381 when the geodesic reaches its vertex is instead @xmath382 more complicated is the solution for the ads part of the geodesic , as the energy ( [ en ] ) is here nonzero .
the solution depends on the range of the parameters @xmath368 ; here we list the relevant formulas 1 .
@xmath190 : we solve the first equation of ( [ eqn1 ] ) for the piece of the geodesic on the right of the btz solution ( the choice @xmath383 for the sign will be explained shortly ): @xmath384 it goes from the shell in @xmath385 at @xmath386 to the boundary at @xmath387 . as @xmath388 should always be positive ( because the metric is @xmath166-independent having a non monotonically behaviour of @xmath389 from the vertex till the boundary , can only increase the length ) and from ( [ var2 ] ) @xmath390 , we have that @xmath391 as well , thus selecting the btz solution @xmath392 .
further the relation ( [ a ] ) between @xmath393 and @xmath391 tells us that also @xmath393 for the present range of parameters ; this explains also the choice of sign in ( [ ra1 ] ) as we conventionally consider an affine parameter @xmath350 increasing from left to right along the geodesic .
the value of @xmath394 is fixed by requiring @xmath395 : @xmath396 the second arc in ads , going from the boundary at @xmath397 to the shell at @xmath373 , can be easily obtained by symmetry . the value of @xmath373 is simply given by the relation @xmath398 and can be checked to produce again @xmath185 when inserted inside both @xmath392 and @xmath399 .
the values of @xmath400 and @xmath401 are obtained respectively from integrating the first of ( [ var ] ) ( with the integration constant conventionally fixed so that @xmath402 ) and from the definition of @xmath340 ( [ v ] ) , where @xmath403 again comes from integrating the second of ( [ var ] ) with integration constant such that @xmath404 ( for btz @xmath405 and we see the finite discontinuity of @xmath165 in passing through the shell ) .
we can also compute the boundary time @xmath98 and distance @xmath348 of the two ending points of the geodesic , as a function of @xmath406 : @xmath407 with appellf1 a kind of hypergeometric function coming from the integration of @xmath166 between the vertex and the shell and @xmath408 the distance in @xmath166 between the crossing with the shell and the boundary point @xmath409 and @xmath410 we will also need to reverse the formula ( [ tb ] ) to obtain the value of @xmath186 as function of @xmath185 and a given @xmath98 .
@xmath411 before going on let me comment that this last equation ( [ pxt ] ) does not work for any value of @xmath98 and @xmath185 , but there is just a restricted range for which the geodesic parameterized by @xmath412 correctly has boundary time equal to @xmath98 .
we will take this into account when plotting parameter regions of geodesics matching some fixed boundary time .
@xmath218 : equation and solution for @xmath399 are the same as the case above . now however we need to glue the btz solution @xmath413 ( positive @xmath414 means now @xmath393 and @xmath415 ) . working as
before we obtain @xmath416 @xmath374 is as before and @xmath373 can be computed using the new @xmath417 value inside @xmath381 . @xmath400 and @xmath401 are computed as in the previous case as the equations for @xmath98 , @xmath188 , @xmath408 and @xmath418 are the same as ( [ dxb ] ) , ( [ dxsb ] ) , ( [ tb ] ) and ( [ pxt ] ) , even in this parameter range .
@xmath226 that is @xmath419 : again a sign switch with both @xmath420 and @xmath415 .
thus the equation for @xmath421 picks up the negative sign , and accordingly changes the solution @xmath422 @xmath423 @xmath424 @xmath400 and @xmath401 are computed as in the previous cases .
now we have slightly different formulas for @xmath425 , @xmath408 @xmath426 @xmath427 ( @xmath117 is the usual step function ) while ( [ tb ] ) and ( [ pxt ] ) for @xmath98 and @xmath418 still apply if using the new quantity @xmath408 above . 4 . @xmath428
: not interesting as the geodesic never goes to the boundary .
the goal is to check weather two `` concentric '' geodesics linked to the end points of the intervals @xmath10 and @xmath203 respectively , all at the same fixed boundary time @xmath98 and choosing @xmath9 to be symmetric to @xmath11 , develop or not a time - like distance in the interior . as the generic problem is complicated , we will limit to check the distance between the vertices ( that , given the space - like boundary conditions are the most likely to be time - like separated ) .
as the two vertices are at the same @xmath166 , the problem will be in the @xmath429 plane .
the first step is to consider two generic points in this plane , we will call @xmath430 the coordinates of the point with the lowest value of @xmath94 and @xmath431 the other , @xmath432 ( if @xmath433 and @xmath434 the distance is obviously time - like ) .
the procedure will be to shoot out of @xmath430 two light - like geodesics inside the plane @xmath435 and towards the boundary ( increasing r ) bordering a space - like region . then we will check if @xmath431 is included inside this region or not .
we will start considering both @xmath430 and @xmath431 past the shell , inside btz and , for the moment , both belonging to the same parameter range , among the ones listed above ( this will be generalized later ) .
then the relevant metric is @xmath436 a light - like geodesic has @xmath437 , so @xmath438 and the space - like region is either @xmath439 if @xmath440 or @xmath441 if @xmath442
. in the @xmath443 coordinates we are using , with @xmath444 this means , as @xmath445 by assumption , the line element is space - like iff if instead both @xmath430 and @xmath431 belong to ads @xmath447 integrating between @xmath448 and @xmath449 we have the space - like - type - inequalities for btz ( the log can also be expressed as an arcth , as one wishes ) singling out either one connected region for @xmath450 or two disconnected ones for @xmath451 : obviously one has to check that with the parameters values considered , not only the two vertices both belong to btz , but also the light - like geodesic does nt cross the shell .
the crossing case will be shown soon .
for ads instead there is always a single connected space - like region @xmath453 the three cases are shown in figure [ figura2b ] . and @xmath442 and ads .
the picture for ads is similar to the first plot , just the curve is not given by a log but by 1/r .
@xmath430 is at the intersection of the two light - like geodesics with the space - like region in between ( or above and below as in the second graph ) coloured.,scaledwidth=100.0% ] two cases are missing , that is when @xmath454 and @xmath455 , thus crossing the btz horizon and having two different parameter - range geodesics , and when one vertex is in btz and the other in ads . to determine the light - like geodesics
the recipe is simply to integrate the null line element for the correct domain .
so if @xmath454 and @xmath455 we need to integrate @xmath456 and @xmath457 until @xmath458 and then from there to @xmath449 ; the integral around @xmath458 is divergent but the sum can be easily regularized ( integrating from @xmath448 until @xmath459 and from @xmath460 to @xmath449 and then sending @xmath461 ) obtaining the light - like distance between @xmath462 and what we may call @xmath463 ( the @xmath340 coordinate of the light - like geodesic at @xmath464 ) , to be : @xmath465.\ ] ] when instead we have a light - like geodesic crossing the shell , we have first to obtain the crossing point @xmath466 , and then integrate accordingly in the two different regions ( before and past @xmath466 ) .
the formula for @xmath466 is ( @xmath467 ) @xmath468\ ] ] so @xmath469 and ( here necessary @xmath440 otherwise this geodesic goes towards smaller @xmath340 and never crosses the shell starting from btz ; also the complementary case is not possible when , starting from a point in ads we have a light - like geodesic that crosses the shell , as they always goes for higher values of @xmath340 . )
@xmath470 and move in the @xmath440 region inside btz ; the lower space - like region when @xmath442 does not extend to @xmath440 , as there is not such a space - like geodesic crossing the horizon .
in the second we focus on the crossing of the shell starting from @xmath471 that happens at at @xmath472.,scaledwidth=80.0% ] let us now apply these inequalities to study in which parameter range the distance between the two geodesic vertices becomes time - like . as the boundary time has to be the same for both geodesics , the four parameters describing them are not independent .
in particular we can freely choose a couple @xmath473 plus @xmath474 and then use equation ( [ pxt ] ) to determine @xmath475 , so that @xmath476 . for simplicity we will restrict our analysis to situations
were the two geodesics belong to the same case , even if a more general computation is possible following the above discussion .
we can easily evaluate @xmath477 , first by calling `` f '' the geodesic with higher @xmath186 , let us fix it to be @xmath478 ( this will become a condition on the three parameters @xmath230 , if not respected just switch names ) and `` i '' the other . then , as the vertex @xmath94-value of a geodesic is always @xmath186 , with @xmath479 by definition : @xmath480
so we end up with ( [ btzc2 ] ) to become where we have defined @xmath483 the convention we made that @xmath484 can be shown to be equivalent to @xmath222 , and we also should impose @xmath223 . further the parameter space @xmath221 should be additionally constrained by @xmath225 , that we have seen is fulfilled by the definition of @xmath224 only on a restricted parameter range ; this will be implemented numerically in all the examples .
we can sweep this parameter space to look for what volume satisfies the inequalities ( [ in21 ] ) and ( [ in22 ] ) and what violates them .
the result is that they are always violated as shown in figure [ figura5 ] again @xmath484 is equivalent to @xmath222 , and thus the parameter space is @xmath228 plus the additional constrain @xmath487 that is @xmath229 that further restricts it , plus the usual @xmath225 . again
always violation , as shown in figure [ figura5b ] this is the most interesting situation as we will end up with both space - like and time - like vertices . here
the choice @xmath484 does not translate to an easy condition on the parameter space , so it is more convenient to avoid it and independently deal with the cases @xmath234 and @xmath235 .
the inequalities representing space - like distances between the vertices thus are : the parameter range for @xmath230 is then limited by the three constraints @xmath232 , @xmath233 and @xmath225 . inside this volume
we separately deal with the subspaces @xmath234 and @xmath235 ; now we can apply ( [ in11 ] ) and ( [ in12 ] ) on the relevant space and see if they are satisfied or not .
it turns out that when @xmath235 , @xmath236 is always respected , while @xmath237 is respected in some region and violated in its complementary , as is shown in figure [ figura7 ] . on the other side when @xmath234 @xmath238 is always respected , while @xmath239 is respected in some region and violated in its complementary ( with respect to the total space ) .
this is shown in figure [ figura8 ] .
now we have btz outside and ads inside the shell . skipping the derivation , that is easier then the one in the previous sections and can be found in @xcite , we limit ourselves in listing the relevant formulas for our purposes .
the boundary time is @xmath489 with @xmath490 @xmath491 the equivalent of ( [ pxt ] ) for the above @xmath98 is instead given by @xmath492 finally the total @xmath166 coordinate length covered by the geodesic at the boundary can be computed to be in the corresponding formula for @xmath493 of @xcite in front of the log . ]
@xmath494 with this we can repeat the analysis of the previous sections ; for geodesics crossing the shell and whose vertices are in ads , the difference between their @xmath340-coordinates becomes ( @xmath235 and @xmath495 in ads for the usual symmetry argument ) @xmath496 and the space - like condition is @xmath497 or more easily @xmath498 given @xmath499 by ( [ pxads ] ) ( restricting to the parameter space such that @xmath235 , and with the corresponding version of ( [ csl ] ) given by @xmath500 on the complementary @xmath234 ) , we can already visualize the parameter volume that respect the first and second of ( [ csl ] ) .
again we also implement the constrain on the parameter space such that the difference between @xmath501 and @xmath502 is smaller then some arbitrarily small number .
numerical results in random parameter domains show that vertices are always at space - like distance one from the other .
s. aoki , t. iritani , m. nozaki , t. numasawa , n. shiba and h. tasaki , `` on the definition of entanglement entropy in lattice gauge theories , '' jhep * 1506 * ( 2015 ) 187 [ arxiv:1502.04267 [ hep - th ] ] .
+ s. ghosh , r. m. soni and s. p. trivedi , `` on the entanglement entropy for gauge theories , '' arxiv:1501.02593 [ hep - th ] .
+ d. radicevic , `` notes on entanglement in abelian gauge theories , '' arxiv:1404.1391 [ hep - th ] .
v. balasubramanian , a. bernamonti , j. de boer , n. copland , b. craps , e. keski - vakkuri , b. muller and a. schafer _ et al .
_ , `` holographic thermalization , '' phys .
d * 84 * ( 2011 ) 026010 [ arxiv:1103.2683 [ hep - th ] ] .
p. v. buividovich and m. i. polikarpov , `` entanglement entropy in gauge theories and the holographic principle for electric strings , '' phys .
b * 670 * ( 2008 ) 141 [ arxiv:0806.3376 [ hep - th ] ] .
+ h. casini , m. huerta and j. a. rosabal , `` remarks on entanglement entropy for gauge fields , '' phys .
d * 89 * ( 2014 ) 8 , 085012 [ arxiv:1312.1183 [ hep - th ] ] .
+ w. donnelly , `` entanglement entropy and nonabelian gauge symmetry , '' class .
grav . * 31 * ( 2014 ) 21 , 214003 [ arxiv:1406.7304 [ hep - th ] ] .
+ c. eling , y. oz and s. theisen , `` entanglement and thermal entropy of gauge fields , '' jhep * 1311 * ( 2013 ) 019 [ arxiv:1308.4964 [ hep - th ] ] .
+ d. n. kabat , `` black hole entropy and entropy of entanglement , '' nucl .
b * 453 * ( 1995 ) 281 [ hep - th/9503016 ] .
b. czech , j. l. karczmarek , f. nogueira and m. van raamsdonk , `` the gravity dual of a density matrix , '' class .
* 29 * ( 2012 ) 155009 doi:10.1088/0264 - 9381/29/15/155009 [ arxiv:1204.1330 [ hep - th ] ] . | we study in detail the relationship between strong subadditivity for a boundary field theory and energy conditions for its bulk dual in 2 + 1 dimensions .
we provide a discussion of known facts and new results organized from the simplest case of a static system with collinear intervals to a time dependent one in a generic configuration , with particular focus on the holographic geometric description .
* strong subadditivity and holography * + andrea prudenziati 2em |
this work was supported in part by nsf - dmr-9705482 .
we also acknowledge partial support from nsf - dmr-96322526 ( the johns hopkins m.r.s.e.c . for nanostructured materials ) . to find the minimum energy we transform @xmath74 in eq .
( [ eq : tot ] ) into @xmath132 space , using eqs .
( [ eq : ee ] ) @xmath27 ( [ eq : orbital ] ) .
this leads to the following energy expressions in @xmath133 space : and @xmath138 is the total number of mn sites .
we obtain eqs .
( [ eq : nonzerok ] ) @xmath27 ( [ eq : del ] ) by minimizing the above expression with respect to all @xmath139 and @xmath140 . without the second neighbor elastic energy term , @xmath141 and @xmath142 minimizing eq .
( [ eq : aeqe ] ) become singular when any of @xmath143 , @xmath144 , and @xmath145 is zero . with nonzero @xmath41
this singularity has been uniquely solved for @xmath146 , while at @xmath111 , it is not . to find the energy term with @xmath111 , we take the @xmath147 limit .
that corresponds to the uniform strain energy , i.e. , the energy related to the change of the lattice parameters from the original cubic structure . here
the problem of the choice of the limiting process arises , because the calculation shows that the different directions of the limiting process of @xmath148 have given different energies and different uniform strains .
because the lower energy state is favored after all , the appropriate limiting process will be the one which gives the minimum uniform strain energy , and it determines the uniform strain also .
when @xmath149 , this appropriate limiting process has been found to satisfy the condition of @xmath143 , @xmath144 , and @xmath150 .
as far as @xmath143 , @xmath144 , and @xmath145 are nonzero , the limits are different only in the order of @xmath151 .
therefore , in the @xmath152 limit , any @xmath147 limit process satisfying the above condition gives the correct expression of the minimum energy term with @xmath111 .
it also gives a unique uniform strain .
r. von helmolt , j. wecker , b. holzapfel , l. schultz , and k. samwer , phys .
lett . * 71 * , 2331 ( 1993 ) ; s. jin , t. h. tiefel , m. mccormack , r. a. fastnacht , r. ramesh , and l. h. chen , science * 264 * , 413 ( 1994 ) ; k. liu , x. w. wu , k. h. ahn , t. sulchek , c. l. chien , and j. q. xiao , phys .
b * 54 * , 3007 ( 1996 ) . c. h. chen and s - w .
cheong , phys .
lett . * 76 * , 4042 ( 1996 ) .
y. murakami , h. kawada , h. kawata , m. tanaka , t. arima , y. moritomo , and y. tokura , phys .
lett . * 80 * , 1932 ( 1998 ) .
y. yamada , o. hino , s. nohdo , and r. kano , phys .
lett . * 77 * , 904 ( 1996 ) .
y. okimoto , t. katsufuji , t. ishikawa , t. arima and y. tokura , phys .
b * 55 * , 4206 ( 1997 )
. n. w. aschcroft , and n. d. mermin , _ solid state physics _ ( saunders college publishing , fort worth , tx , 1976 ) .
h. kawano , r. kajimoto , m. kubota , and h. yoshizawa , phys .
b * 53 * , 2202 ( 1996 ) ; * 53 * , 14709 ( 1996 ) . e. o. wollan and w. c. koehler , phys . rev . * 100 * , 545 ( 1955 ) .
j. kanamori , j. appl .
* 31 * , 14s ( 1960 ) .
a. j. millis , phys .
b * 53 * , 8434 ( 1996 ) .
s. j. l. billinge , r. g. difrancesco , g. h. kwei , j. j. neumeier , and j. d. thompson , phys .
* 77 * , 715 ( 1996 ) j. b. a. a. ellemans , b. van laar , k. r. van der veen , and b. o. loopstra , j. solid state chem . * 3 * , 238 ( 1971 ) . .
coordinates of site @xmath30 , orbital states , ionic displacements , and uniform strains for the minimum energy configuration of fig .
[ fig1](b ) , when @xmath103 , @xmath98 , @xmath99 and @xmath104 .
[ cols="^,^,^,^,^,^,^,^,^",options="header " , ] | a model of localized classical electrons coupled to lattice degrees of freedom and , via the coulomb interaction , to each other , has been studied to gain insight into the charge and orbital ordering observed in lightly doped manganese perovskites .
expressions are obtained for the minimum energy and ionic displacements caused by given hole and electron orbital configurations .
the expressions are analyzed for several hole configurations , including that experimentally observed by yamada _ et al .
_ in @xmath0 .
we find that , although the preferred charge and orbital ordering depend sensitively on parameters , there are ranges of the parameters in which the experimentally observed hole configuration has the lowest energy .
for these parameter values we also find that the energy differences between different hole configurations are on the order of the observed charge ordering transition temperature .
the effects of additional strains are also studied .
some results for @xmath1 are presented , although our model may not adequately describe this material because the high temperature phase is metallic . over the last few years
much attention has been focused on manganese perovskite - based oxides , most notably the pseudocubic materials @xmath2 .
( here @xmath3 is a rare earth element such as la , and @xmath4 is a divalent alkali metal element such as ca or sr . )
the initial motivation came from the observation that for some range of @xmath5 , and temperature , @xmath6 , resistance can be reduced by a factor of up to @xmath7 in the presence of a magnetic field.@xcite two other interesting physical phenomena occurring in this class of materials are charge ordering and orbital ordering.@xcite in this paper , we study the connection between the two .
the important electrons in @xmath2 are the mn @xmath8 electrons ; their concentration is @xmath9 . for many choices of @xmath3 , @xmath4 , and @xmath5 , especially at commensurate @xmath5 values , the @xmath8 charge distribution is not uniform and it indeed appears that a fraction @xmath5 of mn ions have no @xmath8 electron while @xmath9 have a localized @xmath8 electron . a periodic pattern of filled and empty sites is said to exhibit charge ordering .
there are two @xmath8 orbitals per mn ion .
a localized mn @xmath8 electron will be in one linear combination of these ; a periodic pattern of orbital occupancy is said to exhibit orbital ordering .
recently , murakami _ et al .
_ @xcite observed the charge ordering transition accompanying simultaneous orbital ordering in @xmath10 at 217 k ( well above the magnetic phase transition temperature 110 k ) .
it indicates that the interplay of the charge and orbital ordering to minimize the lattice energy could be the origin of the charge ordering . in this paper
we present an expression for the coupling between charge and orbital ordering , with different charge ordering patterns favoring different orbital orderings .
we also argue that the orbital ordering energy differences determine the observed charge ordering in lightly doped manganites .
localized charges induce local lattice distortions , which must be accommodated into the global crystal structure ; the energy cost of this accommodation is different for different charge ordering patterns . to model the charge and orbital ordering , we assume that the electrons are localized classical objects , so that each mn site is occupied by zero or one @xmath8 electron , and each @xmath8 electron is in a definite orbital state .
this assumption seems reasonable in the lightly doped materials such as @xmath0 , which are strongly insulating at all temperatures,@xcite but may not be reasonable for the @xmath1 composition,@xcite where the charge ordered state emerges at a low temperature from a metallic state .
we proceed by calculating the energies of different charge ordering patterns , emphasizing the 1/8 doping case .
it is practically impossible to consider all possible charge ordering configurations .
therefore , we consider the three configurations shown in fig .
[ fig1 ] , which are the only ones consistent with the following basic features of the hole - lattice implied by the experimental results by yamada _
et al._@xcite : invariance under translation by two lattice constants in the @xmath5 or @xmath11 direction , four in the @xmath12 direction , and an alternating pattern of occupied and empty planes along @xmath12 direction .
the configuration in fig .
[ fig1](b ) is the one proposed by yamada _
et al._@xcite to explain their experimental results for @xmath0 . for localized electrons
there are three energy terms : the coupling to the lattice , which will be discussed at length below , the coulomb interaction , and the magnetic interaction .
first , we argue that the coulomb energy can not explain the observed ordering pattern or transition temperature .
we take as reference the state with one @xmath8 electron per mn and denote by @xmath13 the charge of a hole on a mn site . from the classical coulomb energy @xmath14
one finds that the difference in energy between the configurations in fig .
[ fig1 ] is @xmath15 where @xmath16 is a site containing a hole and @xmath17 is the difference in charge between the two configurations .
we estimated the above infinite sum by repeated numerical calculations for larger and larger volumes of the unit cells around the origin .
we find that fig .
[ fig1](c ) has the lowest energy ; 12 mev/@xmath18 lower than fig .
[ fig1](b ) , and 27 mev/@xmath18 lower than fig .
[ fig1](a ) .
to estimate the magnitude of the coulomb energy differences , we need an estimate for the dielectric constant @xmath19 , which we obtain from the measured reflectivity for @xmath20,@xcite and the lyddane - sachs - teller relation@xcite @xmath21 . at frequencies greater than the greatest phonon frequency
the reflectivity is close to 0.1 , implying @xmath22 ; the reflectivity is near unity between @xmath23 ev , and @xmath24 ev , implying @xmath25 . because both @xmath0 and @xmath20 are insulating and have similar compositions , their static dielectric constants are expected to be similar . using @xmath26 , the energy difference between different configurations of holes is only around 2.4 mev , or 30 k per hole , which is small compared to the observed charge ordering temperature of 150 k @xmath27 200 k of these materials . the inconsistency with the experimentally observed hole configuration and the smallness of the energy difference scale indicate that the electrostatic energy is not the main origin of charge ordering for this material .
even though the magnetic and charge ordering transitions show a correlation in @xmath0,@xcite we do not think that the magnetic contribution to charge and orbital ordering is as important as the lattice contribution for three reasons .
first , in undoped @xmath28 , the orbital ordering and the structural phase transition occur at around 800 k and the magnetic ordering at around 140 k,@xcite suggesting that the magnetic effects are relatively weak .
second , in @xmath0 the mn spins are ferromagnetically ordered with moment close to the full mn moment at temperatures greater than the charge ordering temperature,@xcite and ferromagnetic order does not favor one charge configuration over another .
third , although in @xmath0 antiferromagnetic order appears at the charge ordering transition , the antiferromagnetic moment is very small ( less than 0.1 of the full mn moment),@xcite so the energy associated with this ordering must be much less than 140 k / site associated with magnetic ordering in @xmath28 .
therefore , we think that the canted antiferromagnetism occuring upon charge ordering in @xmath0 ( ref .
7 ) is not the cause but the effect of the charge and orbital ordering .
we now turn our attention to the lattice energy .
a classical model for the lattice distortions of the insulating perovskite manganites has been derived in ref .
10 , and shown to be consistent with experimental results on @xmath28 .
this model is adopted here with an additional term , an energy cost for shear strain .
we now briefly outline the model , which is explained in more detail in ref .
10 and the appendix .
the ionic displacements included are the vector displacement @xmath29 of the mn ion on site @xmath30 , and the @xmath31 directional scalar displacement @xmath32 ( @xmath33 , @xmath11 , and @xmath12 ) of the o ion which sits between the mn ion on site @xmath30 and the mn ion on site @xmath34 .
for convenience , @xmath29 and @xmath32 are defined to be dimensionless in the following way : the lattice constant of the ideal cubic perovskite is @xmath35 , the mn ion position in the ideal cubic perovskite is @xmath36 , the actual mn ion position is @xmath37 , and the actual o ion position is @xmath38 . the lattice energy is taken to be harmonic and depends only on the nearest neighbor mn - o distance and the first and second nearest neighbor mn - mn distances .
the spring constants corresponding to these displacements are @xmath39 , @xmath40 , and @xmath41 as shown in fig .
[ fig2 ] . because @xmath39 and @xmath40 involve bond stretching , while @xmath41 involves bond bending , @xmath42 is expected .
thus , @xmath43 , where @xmath44^{2 } \nonumber \\ & & + \left [ \left(\frac{\delta^a_{i+a - b}-\delta^b_{i+a - b}}{\sqrt{2}}\right)- \left(\frac{\delta^a_{i}-\delta^b_{i}}{\sqrt{2}}\right)\right]^{2}. \end{aligned}\ ] ] in the above equations @xmath45 denotes @xmath5 , @xmath11 , and @xmath12 , and @xmath46 represents @xmath47 , @xmath48 , and @xmath49 .
@xmath50 was not considered in ref .
10 .
the shear modulus produced by this term is important , because without it , a mn ion on site @xmath51 can have arbitrary large @xmath11 directional displacement relative to the mn ion on site @xmath30 at no cost of energy . for this reason ,
the model with @xmath52 has singularities , whose proper treatment requires @xmath53 in our model .
however , still we expect @xmath41 will be much smaller than @xmath39 or @xmath40 . therefore , in order to simplify the calculation , the @xmath54 limit has been taken after the expression of minimized energy and equilibrium ionic displacements have been obtained .
second , we consider the electronic degree of freedom . we parameterize the electron density by the variable @xmath55 . if an electron is present on site @xmath30 , @xmath56 ; if no electron is present , @xmath57 .
if there is an electron on site @xmath30 , the electron orbital state , which is a linear combination of the two @xmath8 orbitals , is parameterized by an angle @xmath58 as follows .
@xmath59 with @xmath60 .
the electron orbital state couples to the distortion of the surrounding oxygen octahedra through the jahn - teller distortion .
the coupling is given by @xmath61 \nonumber \\ & = & -\lambda\sum_{i , a}(1-h_{i})v^{a}_{i } \cos{2(\theta_{i}+\psi_{a } ) } , \end{aligned}\ ] ] where @xmath62 if a hole is present on site @xmath30 , it attracts the surrounding oxygens equally , giving rise to a breathing distortion energy given by @xmath63 the parameter @xmath64 represents the strength of the breathing distortion relative to the jahn - teller distortion .
finally , following kanamori,@xcite we include a phenomenological cubic anharmonicity term given by @xmath65 the sign has been chosen so that the electron orbital states of @xmath66 , @xmath67 , or @xmath68 , with @xmath69 , @xmath70 , and @xmath71 pointing toward nearest oxygen ions are favored when @xmath72 is positive .
the total energy , which is the sum of all the above energy terms , is given by @xmath73 we minimized @xmath74 about @xmath75 s and @xmath32 s for fixed hole and orbital configurations .
these are conveniently expressed in terms of the variables @xmath76 , @xmath77 , @xmath78 , and @xmath79 defined in the following way .
@xmath80 the details are shown in the appendix .
the minimized energy per mn ion may be written as @xmath81 where @xmath82 ( \beta h_{\vec{k}}-c^{a}_{\vec{k}})(\beta h_{-\vec{k } } - c^a_{-\vec{k } } ) , & \mbox{if $ k_a\neq 0 $ } \\ 0 , & \mbox{if $ k_a= 0 $ } \end{array } \right .
\label{eq : nonzerok } \\ { \cal e}_{\vec{k}=0}&=&-\frac{\lambda^2}{k_1 + 2k_2}[3(\beta h_0)^2 + \sum_{a}(c^a_0)^2 ] .
\label{eq : zerok}\end{aligned}\ ] ] the long wave length strain @xmath83 , and the @xmath84 components of the ionic displacements are given as @xmath85}{(k_1 + 2k_2)k_1 } \frac{1-e^{-ik_a}}{1-\cos{k_a } } ( \beta h_{\vec{k } } -c^a_{\vec{k } } ) , & \mbox{if $ k_a\neq0 $ } \\
0 , & \mbox{if $ k_a=0 $ } \end{array } \right .
\\ \delta^a_{\vec{k}\neq 0}&= & \left\ { \begin{array}{ll } -i\frac{\lambda\sin{k_a}}{(k_1 + 2k_2)(1-\cos{k_a } ) } ( \beta h_{\vec{k } } -c^a_{\vec{k } } ) , & \mbox{if $ k_a\neq 0 $ } \\ 0 , & \mbox{if $ k_a=0$. } \end{array }
\right .
\label{eq : del}\end{aligned}\ ] ] because @xmath55 s and @xmath86 s are bounded by @xmath87 , we can not treat @xmath78 s and @xmath88 s as independent variables to minimize @xmath74 .
therefore , we minimize @xmath74 over the orbital variable @xmath58 s at fixed hole configurations ; the ground state is then the hole configuration of the lowest energy . for @xmath0 ,
we consider the three hole configurations shown in fig .
[ fig1 ] , each of which is a bravais lattice , with a unit cell containing one mn site with a hole and seven mn sites without holes . the orbital configuration may be different in different unit cells of the lattice defined by the holes .
we consider the case where the orbital configuration is the same in each unit cell .
in addition to that , we also consider all possible two sublattice symmetry breakings .
therefore , we have seven ( if no symmetry breaking ) or fourteen ( if two - sublattice symmetry breaking ) orbital variable @xmath58 s .
@xmath89 in eq .
( [ eq : tot ] ) for each configuration is expressed in terms of those variables through eqs .
( [ eq : anharm ] ) , ( [ eq : hole ] ) , ( [ eq : orbital ] ) , ( [ eq : nonzerok ] ) , and ( [ eq : zerok ] ) , and is minimized about @xmath58 s . for this minimization , we use the _ findminimum _ routine in _ mathematica _ in the following way : for each set of parameters , and for each configuration , we check the local minimal values by using 50 @xmath27 200 random starting values of @xmath58 s . according to ref .
10 , @xmath90 ranges over 0.04 @xmath27 0.05 , and @xmath91 is between 0 and 1 .
@xmath92 ranges around 0.0002 , and @xmath93 ev.@xcite recently , a local breathing distortion of @xmath94 @xmath95 has been directly observed in @xmath96.@xcite the jahn - teller distortion is estimated around 0.15 @xmath95 from the mn - o distances of @xmath28.@xcite this implies that the breathing distortion and the jahn - teller distortion in these materials have similar order of magnitude , i.e. , @xmath97 .
we varied @xmath64 in the range of 0 @xmath27 10 , and @xmath92 in the range of 0 @xmath27 0.00035 , with @xmath98 , @xmath99 , and @xmath100 ev .
for each set of those parameters , the minimum energy per hole for each fixed hole configuration in fig .
[ fig1 ] has been found . by comparing them ,
we find the most favored hole configuration for each @xmath64 and @xmath92 , which is shown in fig .
[ fig3 ] as a plot in @xmath64-@xmath92 plane . at large @xmath64 ( @xmath101 ) ,
the configuration shown in fig .
[ fig1](c ) is the most favored , and that shown in fig .
[ fig1](a ) is the least favored .
this can be related to the fact that , in @xmath11-@xmath12 and @xmath12-@xmath5 directional planes , fig . [
fig1](c ) has the most even distribution of holes , and fig .
[ fig1](a ) has the least even distribution . for large @xmath64 ,
the contraction of oxygen octahedra toward holes is strong , and an uneven distribution of holes generates larger strains and elevates minimum energies . particularly , the square hole net squeezes the electron orbital at the center of the square along the direction perpendicular to the square plane . in the cubic hole configuration of fig .
[ fig1](a ) , the six squeezed electron orbitals point toward the cubic center , putting the electron orbital at the center at high energy , which is consistent with our result that fig .
[ fig1](a ) has far higher minimum energies than figs .
[ fig1](b ) and [ fig1](c ) in large @xmath64 limit . as @xmath64
is decreased into the range of 2 @xmath27 5 , the favored hole configuration becomes that of fig .
[ fig1](b ) , which is the experimentally observed hole configuration .
we expect that the difference of the energy per hole between the ground state hole configuration and the next lowest energy hole configuration corresponds approximately to the charge ordering temperature .
the calculation results indicate that , when @xmath64 is in the range of 2.0 @xmath27 2.5 or around @xmath102 and @xmath103 , the charge ordering temperature is around 100 @xmath27 200 k , which is consistent with experimental results . as @xmath64
is decreased further , the most favored hole configuration changes further and the temperature difference scale decreases .
figure [ fig3 ] also shows the tendency that the configuration of fig .
[ fig1](c ) becomes more favored as @xmath92 increases .
we think this occurs because the anharmonicity energy distorts the oxygen octahedra tetragonally , which can be more easily accommodated by the tetragonal hole configuration of fig .
[ fig1](c ) . in table
[ table1 ] , we have shown an example of the orbital states , ionic displacements , and uniform strains corresponding to the minimum energy configuration for fig .
[ fig1](b ) when @xmath103 , @xmath98 , @xmath99 and @xmath104 .
@xmath5 , @xmath11 , and @xmath12 directions are shown in fig .
[ fig1 ] .
the nearest mn - mn distance is unit .
@xmath105 is defined in such a way that @xmath106 s and @xmath107 s , where @xmath108 , @xmath109 , and @xmath110 are integers , represent the coordinates of the sites indexed by @xmath30 .
@xmath111 parts of the ionic displacements have been subtracted to find the non - uniform parts of the displacements .
the energy expressions in eqs .
( [ eq : nonzerok ] ) and ( [ eq : zerok ] ) are adequate for bulk materials .
when the material is grown on a substrate as a thin film , generally there is a strain generated by lattice mismatch between the film and the substrate materials . to see the effect of this strain , we add a term proportional to @xmath112 ( @xmath113 , @xmath11 , or @xmath12 ) to the energy , which corresponds to an @xmath114 directional strain . using a parameter @xmath115 , we replace @xmath116 in eq .
( [ eq : zerok ] ) by the following expression : @xmath117.\ ] ] we repeated similar calculations to find the favored hole configurations for different values of the applied strain , parameterized by @xmath115 .
the applied strain breaks cubic symmetry .
some of the hole configurations also break cubic symmetry . for these cases
the energy depends on the relative orientation of the strain and hole symmetry breakings .
we consider all possible orientations and find the lowest energy state .
we have varied @xmath115 between @xmath118 and @xmath119 , and @xmath64 between 0 and 7 , with @xmath103 , @xmath98 , and @xmath99 .
the results are shown as a phase diagram in @xmath64-@xmath115 plane in fig .
[ fig4 ] .
it shows that fig .
[ fig1](c ) configuration is favored more as @xmath120 increases .
this feature can be understood in the following way . for small @xmath115 s ,
the leading correction to the minimum energy for each hole configuration is @xmath121 , where @xmath122 represents @xmath123 .
therefore the configuration which has a larger @xmath124 will show greater change in energy for a given @xmath115 . since the hole configuration in fig .
[ fig1](c ) has tetragonal symmetry , which is compatible with the jahn - teller distortion , it has the largest @xmath125 . therefore , as @xmath120 increases , fig . [ fig1](c ) is more favored than figs . [ fig1](a ) and [ fig1](b ) . because the energy changes linearly with @xmath115 , the phase boundaries are straight lines for small @xmath115 , and have cusps at @xmath126 , as shown in fig .
[ fig4 ] .
typical variations of @xmath127 corresponding to changing @xmath120 from 0 to 0.4 are about 2 % .
the results indicate that the strain generated by substrates can change ordered hole configuration and ordering temperature .
our results indicate that the interaction of the electronic state and the lattice can be the origin of the charge ordering in this material , even though the details of the results are dependent on specific choice of @xmath39 , @xmath40 , @xmath128 , and @xmath72 .
a similar calculation has been done for @xmath129 , hole concentration 1/2 .
we choose the three hole configurations in fig .
[ fig5 ] to compare the minimum energies .
each configuration has an alternating hole distribution in different set of directions : @xmath5 , @xmath11 , and @xmath12 directions for fig .
[ fig5](a ) , @xmath5 and @xmath11 directions for fig .
[ fig5](b ) , and @xmath11 direction for fig .
[ fig5](c ) . figure [ fig5](b )
is the experimentally observed hole configuration.@xcite as we have done for @xmath130 , we consider both the case where hole and orbital state have the same unit cell , and the case where the orbital state is composed of the two hole sublattices
. calculations for @xmath103 , @xmath98 , and @xmath99 show that when @xmath64 is large , configuration in fig .
[ fig5](a ) is the most favored and fig .
[ fig5](c ) is the least favored . as @xmath64 is decreased , the favored configuration is changed between @xmath64=0.5 and 0.7 . after that fig .
[ fig5](c ) is the most favored and fig .
[ fig5](a ) is the least favored .
when @xmath64 is large , the holes prefer to distribute evenly because of the same reason as in @xmath130 case .
in contrast , when @xmath64 is small , electron sites prefer to have more neighboring electron sites to gain orbital energy .
our results are not consistent with the experimental results for @xmath1,@xcite which indicate that configuration fig .
[ fig5](b ) is the ground state .
this inconsistency may arise because our model involves only localized electrons , while for @xmath131 the charge ordering state arises from a metallic phase .
modifications of our model to include hole hopping are desirable . in summary
, we have shown that the lattice effect could play an important role in the charge ordering transition observed in perovskite manganites . |
the planck mass , @xmath5 , is a fundamental quantity of stellar astrophysics .
the chandrasekhar mass to order unity is @xmath6 , where @xmath7 is the proton mass .
the maximum mass of a star is approximately the chandrasekhar mass multiplied by the square of the ratio of radiation pressure to gas pressure ( eddington 1917 ) .
the minimum mass of a black hole is within order unity of the planck mass .
variation of newton s constant affects supernova cosmology via change in the planck mass over cosmic time and was first considered by gaztaaga ( 2002 ) . at that time
there were 42 sne available ; there are now 581 ( suzuki 2012 ) .
uzan ( 2003 ) and narimani , moss , & scott ( 2010 , 2012 ) advise that constraining the constancy of dimensional quantities is perilous .
preferred quantities are , for example , the gravitational fine structure constant " , @xmath8 .
the lunar laser ranging experiment initiated by nasa s apollo mission is an @xmath9 experiment , monitoring the specific potential energy of the earth - moon system @xmath10 , where t is the time of flight of earth launched photons .
measuring the luminosity distance of galaxies , @xmath11 , from type ia supernovae is an @xmath9 experiment , as @xmath12 , assuming a fixed fraction , f , of @xmath13 is turned into energy and stellar luminosities are calibrated by hydrogen burning stars .
specifically , to within a numerical constant of order unity , @xmath14 .
speculation about varying g began with dirac ( 1937 ) , who noted that the ratio of the electrostatic and gravitational forces between an electron and a proton was of the same order as the number of times an electron orbits a proton in the age of the universe .
he conjectured that @xmath9 might decay as the inverse of cosmic time .
this 20th century gravity problem ( which is sometimes tackled anthropically ) has been totally eclipsed in the last decade by the cosmological constant problem ( sol 2013 ) .
the contribution to the vacuum energy density of fluctuations in the gravitational field is larger than is observationally allowed by some 120 orders of magnitude . instead
, the vacuum energy density is of the same order of magnitude as the present mass density of the universe .
although ongoing type ia supernova observations indicate that the equation of state of einstein s general relativity is the best fit , this gross cosmological constant problem provides no comfort for constant g orthodoxy .
garcia - berro ( 2007 ) review astronomical measurements and constraints on the variability of fundamental constants generally .
garcia - berro ( 2006 ) fit a polynomial g(z ) = @xmath15 to the snia data , suggestive of a g larger in the past .
verbiest ( 2008 ) measure orbital period rates of pulsars and set a limit of @xmath16 . from white dwarf cooling , garcia - berro et al .
( 2011 ) derive an upper bound @xmath17 and corsico ( 2013 ) find a white dwarf pulsation limit of @xmath18 .
tomaschitz ( 2010 ) considers a gravitational constant scaling linearly with the hubble parameter , and fits the snia hubble diagram and agn source counts , concluding that further observational constraints are required .
furthermore , the luminosity of degenerate carbon core supernovae is proportional to the mass of carbon burned to @xmath4ni .
the precise mechanism which powers a type ia supernova explosion is a matter of lively debate , and we do not know yet whether a detonation or a deflagration occurs . what it is clear is the close correlation between the mass of nickel synthesised in the outburst and the luminosity .
this is discussed in quantitative experiments by gaztaaga ( 2002 ) .
the type snia standard candle is thought to result from a high degree of regulation of this quantity , such as would be provided by approach to a physical limit , the chandrasekhar mass .
however , one may conjecture that the fraction of @xmath13 that is turned into energy may also vary with z. like variation of g with z , this is also an issue for constraints on the equation of state of the universe arising from supernova measurements .
there are two constraints on varying g , that which has been established from the lunar distance since 1969 ( current epoch in table 1 ) and astrophysical constraints acting over cosmic time , such as the ages of the oldest stars .
according to the theories of g variation reviewed by faulkner ( 1976 ) , g may have been larger in the past and may be considered to follow a @xmath19 decline to the current epoch .
williams and dickey ( 2002 ) placed a 1@xmath20 limit of @xmath21 = 1.1 @xmath22 10@xmath23 per year in recent time . if g exceeded the present value by 3 @xmath24 @xmath21 @xmath24 + 7.3 @xmath22 10@xmath25 per year 13.7 gyrs ago , and we assume the supernova luminosity scales with the chandrasekhar mass , we obtain figure 1 .
this includes variation of the density term in the friedmann equation , i.e. @xmath26(t ) with two cases ( 1 ) @xmath26(t ) + @xmath27(t ) = 1 to retain flatness and ( 2 ) the dark energy density @xmath27 = 0.73 with @xmath28 .
we characterize type ( 2 ) models by the value of the curvature , @xmath29 , at z = 0.5 .
we adopt the wmap9 limits on curvature ( hinshaw 2013 ) , 0.0065 @xmath30 0.0012 . the current supernova data ( suzuki 2012 ) are shown in figure 1 . if we assume @xmath31cdm cosmology with w = 1 , current snia data constrain g stability to @xmath21 = ( 3,+7.3 ) @xmath22 10@xmath25 per year , this constraint is obtained if we adopt the standard model of cosmology as a prior .
we now ( 1 ) reexpress this as a constraint on @xmath32 , and ( 2 ) invert the argument to constrain @xmath33 , given laboratory limits on g variation .
\(1 ) to determine the upper limit on @xmath32 we calculate @xmath34 to compare the data with the prediction , marginalizing over h@xmath35 , and show this in figure 2 .
the contours of @xmath34 are oriented close to vertical , resulting in clear limits on g variation .
this constraint , our main result , 3
@xmath36 @xmath21 @xmath2 7.3 per year , may be expected to strengthen towards parity with the laboratory , in the era of dark energy experiments such as lsst , see weinberg ( 2013 ) .
an equivalent dimensionless limit is 0.5 @xmath37 1 , where 1/@xmath38 is the age of the universe .
@cllcll@ + @xmath21 & current epoch & & @xmath21 & cosmic time + @xmath39 & & & @xmath39 & + 2@xmath407 & lunar laser ranging & ( 1 ) & 0 @xmath404 & big bang nucleosynthesis & ( 2 ) + 40 @xmath4050 & binary pulsar & ( 3 ) & -1.42 + 2.48 - 2.27 & planck+wmap+bao & ( 4 ) + 230 & psr j04374715 & ( 7 ) & 0@xmath4016 & helioseismology & ( 5 ) + & & & -6@xmath4020 & neutron star mass & ( 6 ) + & & & 300,+730&this paper + [ cols= " < " , ] \(2 ) this constraint is obtained if we adopt the standard model of cosmology as a prior .
however , these snia data are conventionally used as a measurement of @xmath27 .
there is therefore a degeneracy between this and @xmath41 addressed by the same data .
we can quantify the degeneracy using the generalization of the friedmann equation as a polynomial by mould ( 2011 ) .
mould showed that if , such a polynomial is adopted to fit figure 1 , @xmath42 relationships ( degeneracies ) between the @xmath43 coefficients result from the limited available constraints ( snia , cmb ) .
if the snia data are used to constrain the equation of state of the universe with w @xmath44 1 , there is therefore a degeneracy between w and @xmath41 addressed by the same data . for z @xmath3 1 and zero curvature , @xmath45 from equations ( 3 ) & ( 11 ) of mould ( 2011 ) . for g stability to 2 parts in 10@xmath46 per year , @xmath47 = 0.0137 at z = 1 , which corresponds to @xmath48 0.03
the current experimental uncertainty in w ( rapetti 2013 ) is 0.07 .
both quantities therefore need to be constrained jointly . on the other hand ,
if one is prepared to adopt the lunar laser ranging results as a prior on dark energy experiments valid over all of cosmic time , the effect of variable g is constrained so that @xmath1w @xmath2 0.02 ( 95% confidence ) .
this is not a negligible contribution to the @xmath33 error budget , and it should not be ignored ( @xmath49 mortonson 2014 ) .
finally , the coupling of @xmath11 , f , and @xmath13 is direct .
analytically , 2@xmath50 = @xmath1f / f = 1.5@xmath1g / g = @xmath51 .
our limit on @xmath47 is thus degenerate with an equivalent limit on @xmath1f / f .
pragmatically , the key result here devolves from an assumed t@xmath52 variation of @xmath9 .
its basis is historical and traces back to dirac s ( 1937 ) large numbers hypothesis and the steady state universe , neither of which have any real traction today . other parameterizations are possible and even natural , such as 1+z scaling .
one form is @xmath9 = @xmath53 + @xmath54 .
in this case we obtain 0.02 @xmath55 0.04 .
if g @xmath3 1+z , the @xmath56 coefficient in equation ( 1 ) is promoted to @xmath57 ; that is , it becomes an anti - radiation pressure term .
assuming @xmath58 = 0 , the resulting degeneracies can be expressed ( mould 2011)@xmath59 where @xmath60 any g variation that scales as 1+z is traded off against @xmath27 , according to ( from table 2 ) @xmath61 , when snia and cmb anisotropy data measure cosmological parameters simultaneously . a universe with just conventional dark energy and radiation " like this has an age in units of 1/h@xmath35 obtained by integrating equation ( 1 ) with @xmath62 = 1 + z. @xmath63 .
@xmath64 @xmath65_{arsinh(b)}^\infty\ ] ] @xmath66 where @xmath67 and @xmath68 . for @xmath69 = 0.27 , x = 0.575 at z = 0 and
the age is 0.745 .
this is a second contradiction with the standard model of cosmology , as planck finds an age approximately one in these units ( ade et al 2013 , efstathiou 2013 ) . a further contradiction with the age of the globular clusters is mildly ameliorated by higher central temperatures of stars ( gm@xmath70 , where m , r are the stellar mass and radius ) during the epoch of reionization , when they were formed , and the extraordinary temperature sensitivity of the cn cycle of fusion , but for most of the low mass stars lifetime core temperatures are close to normal and ages are only mildly affected ( vandenberg 1977 ) .
how severe a constraint on g @xmath3 1+z is this ?
error analysis gives terms in @xmath71 .
the first and last of these are of order a few percent and the second is smaller .
this parameterization can therefore be rejected with 99% confidence .
g @xmath72 would be less unacceptable for large n , but is not a natural parameterization .
0.25 truein ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss@xmath43nw@xmath73f@xmath73 + vacuum010.662 + textures12/30.964 + `` curvature''21/32.294 + matter3010.40 + radiation4 + 1/3494.4 + + the @xmath74 coefficients have been evaluated at @xmath56 = 0.27 .
our conclusions from this work are as follows . 1 . the validity of the snia standard candle depends on the stability of g and the stability of f , the fraction of the chandrasekhar mass turned into energy .
we have considered the former in this paper and derived a constraint on the gravitational fine structure constant which can be compared with other combined astrophysical - cosmological constraints . but this is inextricably degenerate with possible evolution of f due to changes over cosmic time of snia progenitor astrophysics . with this caveat we set a snia cosmology constraint on g stability in the standard cosmology of ( 3,+7.3 ) @xmath22 10@xmath25 per year .
this limit is two orders of magnitude weaker than that from lunar laser ranging .
but that is a current epoch result and complements , but does not replace , a constraint that spans cosmic time .
the limit is also two orders of magnitude weaker than that arising from the great sensitivity to density of big bang nucleosynthesis .
the snia standard candle , however , has the distinct advantage of covering the last 10@xmath75 years of cosmic time , rather than the first twenty minutes .
our result is an update of gaztaaga ( 2002 ) , who found @xmath76 12 @xmath22 10@xmath77 /yr for @xmath27 = 0.8 , @xmath78 = 0.2 .
this is a 1@xmath20 limit , like ours , and directly comparable since @xmath79 1 .
the order of magnitude more supernovae now available have allowed us to relax their flat universe assumption , but has also relaxed their limit on @xmath41 . 5 .
caution would dictate that experiments to measure the equation of state of the universe carry the caveat that f and g stability is assumed .
for the latter , lunar laser ranging is available as a prior and limits @xmath80 to 0.02 , but the former has not been quantified and demands further understanding of the snia mechanism .
a parameterization g @xmath3 1+z is interesting on theoretical grounds related to the unity of forces .
however , with such an equation of state the expansion age of the universe is too short .
that parameterization can therefore be rejected with 99% confidence .
we are grateful to chris blake and michael murphy for helpful advice and to an anonymous referee for emphasizing the issue of the quantum of @xmath4ni .
caastro is the arc s centre of excellence for all - sky astrophysics , funded by grant ce11001020 .
ade , p. et al 2013 , a&a , in press , astro - ph 1303.5076 + copi , c. davis , a. & krauss , l. 2004 , phys .
92 , 171301 + corsico , a. , 2013 , jcap , 06 , 032 + dirac , p. 1937 , nature , 139 , 323 + eddington , a. s. 1917 , mnras , 77 , 596 + efstathiou , g. 2013 , astro - ph 1311.3461 + faulkner , d. 1976 , mnras , 176 , 621 + garcia - berro , e. 2006 , ijmpd , 15 , 1163 + garcia - berro , e. 2007 , a&arv , 14 , 113 + garcia - berro , e. 2011 , jcap , 05 , 21 + gaztaaga , e. 2002 , phys rev d , 65 , 023506 + guenther , b. , krauss , l. & demarque , p. 1998 , apj 498 , 871 + hinshaw , g. 2013 , apjs , 208 , 19 + kaspi , v. , taylor , j. & ryba , m. 1994 , apj 428 , 713 + li , y .- c 2013 , phys rev d , 88 , 084053 + mortonson , m. 2014 , arxiv , 1401.0046 + mould , j. 2011 , pasp , 123 , 1030 + muller , j. & biskupek , l. 2007 , class .
24 , 4533 + narimani , a. 2012 , ap&ss , 341 , 617 + narimani , a. 2010 , ijmpd , 19 , 2289 + rapetti , d. 2013 , mnras , 432 , 973 + riess , a. 2012 , apj , 752 , 76 + sol , j. 2013 , jphcs , 453 , 2015 + suzuki , n. , 2012 , apj , 746 , 85 + thorsett , s. 1996 , phys .
77 , 1432 + tomaschitz , r. 2010 , ap&ss , 325 , 259 + uzan , j .-
2003 , rvmp , 75 , 403 + vandenberg , d. 1977 , mnras , 181 , 695 + verbiest , j. 2008 , apj , 679 , 675 + weinberg , d. 2013 , phr 530 , 87 + williams , j. & dickey , j. 2002 , http://ilrs.gsfc.nasa.gov/docs/williams_lw13.pdf | astrophysical cosmology constrains the variation of newton s constant in a manner complementary to laboratory experiments , such as the celebrated lunar laser ranging campaign .
supernova cosmology is an example of the former and has attained campaign status , following planning by a dark energy task force in 2005 . in this paper we employ the full snia dataset to the end of 2013 to set a limit on g variation . in our approach
we adopt the standard candle delineation of the redshift distance relation .
we set an upper limit on its rate of change @xmath0 of 0.1 parts per billion per year over 9 gyrs .
by contrast lunar laser ranging tests variation of g over the last few decades .
conversely , one may adopt the laboratory result as a prior and constrain the effect of variable g in dark energy equation of state experiments to @xmath1w @xmath2 0.02 .
we also examine the parameterization g @xmath3 1+z .
its short expansion age conflicts with the measured values of the expansion rate and the density in a flat universe . in conclusion ,
supernova cosmology complements other experiments in limiting g variation .
an important caveat is that it rests on the assumption that the same mass of @xmath4ni is burned to create the standard candle regardless of redshift .
these two quantities , f and g , where f is the chandrasekhar mass fraction burned , are degenerate .
constraining f variation alone requires more understanding of the snia mechanism . *
constraining a possible variation of g with type ia supernovae * jeremy mould & syed a. uddin centre for astrophysics & supercomputing , swinburne university _ stars : ( supernovae):general gravity
( cosmology ) : distance scale stars : ( white dwarfs ) _ |
the study of doped antiferromagnets has become a central topic in the area of strongly correlated electrons systems , mainly after anderson s proposition@xcite about the probable existence of non - fermi liquid behavior once a mott insulating state is hole - doped . a valuable experiment that gives access to single - particle properties of many - electron systems
is the angle resolved photoemission spectroscopy ( arpes ) . even though in the last years there has been a rapid pace of developments in arpes experiments , there are still controversies in many strongly interacting cases , like the cuprates , where the difficulty to distinguish a coherent @xmath4-function peak from an incoherent part of the spectra does not allow a proper interpretation of the available data@xcite .
a useful microscopic model widely used to describe the low - energy physics involved in photoemission experiments of antiferromagnetic mott insulators@xcite is the @xmath0 model on the square lattice @xcite .
attention has also been concentrated on triangular mott insulators .
these systems are attractive due to the presence of both , strong correlation and geometrical magnetic frustration , leading to non - conventional behavior .
experimental realizations of such systems are @xmath5@xcite , the @xmath6 family of the organic charge transfer salts @xmath7 @xcite , and the silicon surfaces k / si ( 111):b and sic(0001)@xcite . in a recent publication , koretsune and
ogata@xcite have studied the effect of finite doping in the triangular heisenberg model .
they found that a resonating - valence - bond ( rvb ) ground state is favored when @xmath2 is positive , while a nagaoka s ferromagnetic ground state is stabilized for negative @xmath2 .
more recently , the role of the @xmath2 sign and frustration has been addressed in the superconductor @xmath8 where the @xmath9 atoms form the triangular lattice @xcite .
these remarkable issues lead us to investigate the spectral function of a hole in a triangular mott insulator , so as to analize the consequences of the geometrical frustration and the @xmath2 sign on the hole motion .
in particular , our aim is to study the existence of quasiparticle ( qp ) excitations and higher - energy features in the spectral functions . to this purpose , we focus on the hole dynamics in the triangular @xmath0 model , and we solve it using the lanczos exact diagonalization method and the self - consistent born approximation ( scba ) . in the non - frustrated case , it is known that the scba reproduces quite well the exact diagonalization results on small clusters@xcite and the angle - resolved photoemission spectroscopy ( arpes ) experiments@xcite . in particular , the scba predicts the existence of a well defined quasiparticle peak ( magnetic polaron ) at low energies for the whole brillouin zone ( bz ) .
furthermore , the scba is able to capture finite lifetime resonances above the qp peaks .
these resonances have been explained by the string picture since their energies scale as @xmath3 , as it happens in the ising limit . in our triangular frustrated case ,
the presence of the @xmath10 nel order generates two mechanisms for hole motion in the scba effective hamiltonian , one being tight - binding like ( or free ) and the other , magnon - assisted .
we will show that the @xmath2 sign tunes the interference between both mechanisms , producing pronounced differences in the low energy as well as in the higher energy region of the spectral functions .
for the strong coupling regime , we find an excellent agreement between the scba and the exact results on a 21 sites cluster . in the thermodynamic limit ,
the scba spectral functions present a rich structure .
notably , for @xmath11 a quasiparticle excitation is always defined for all @xmath12 of the brillouin zone , while for @xmath13 in the strong coupling regime the quasiparticle signal survives only around the goldstone magnetic wave vectors . by computing the quasiparticle energy scaling with @xmath14 it is possible to get some insight of the magnetic cloud surrounding the hole@xcite .
we argue that for @xmath15 the local environment around the hole enhances its af character , while for @xmath16 the ferromagnetic character is favored .
we find that the single - hole dynamics in the thermodynamic limit behaves in the opposite way as in very small clusters , like the three - sites problem@xcite .
furthermore , for small values of @xmath14 and @xmath17 , higher energy resonances above the quasiparticle peak appear showing a @xmath3 energy scaling . although the triangular @xmath18 model is completely different to the square lattice case , we continue identifying these resonances with string - like excitations . for @xmath19 , instead , we do not observe higher energy resonances that can be identified with strings .
this behavior observed at higher energies can be also traced back to the particular features of the magnetic environment around the hole mentioned above .
the paper is organized as follows . in section
ii we obtain the effective hamiltonian and the one hole properties within the scba . in section iii we present the results : ( iiia ) the comparison between exact and scba results in finite size clusters , ( iiib ) the quasiparticle features , and ( iiic ) the general description of the spectral functions . finally , in section iv we draw the concluding remarks .
the @xmath0 model is defined as @xmath20 where the first term represents the hopping between nearest neighbors ( @xmath21 ) of the triangular lattice , with the constraint of no double occupancy , @xmath22 ) , and the second term represents the antiferromagnetic ( af ) heisenberg part with @xmath23 .
the study of the single hole motion in an af requires a correct description of the magnetic low energy sector since the injected hole will couple strongly with these spin excitations . in the literature
there is firm numerical @xcite and analytical @xcite evidence that the triangular heisenberg antiferromagnet has a @xmath24 nel order ground state , and that spin wave theory reproduces quite well its low energy behavior@xcite .
so that , we assume an ordered ground state with a magnetic wave vector @xmath25 lying in the @xmath26 plane , and spin waves as the low energy excitations . in order to simplify the calculation
we perform a unitary transformation to local spin quantization axis primed operators in order to have a ferromagnetic ground state in the @xmath27 direction .
then , we use the spinless fermion representation for electrons @xmath28 and the holstein - primakov representation for spin operators ( to order @xmath29 ) @xmath30 , where @xmath31 creates a spinless hole at site @xmath32 , and @xmath33 creates a spin deviation .
we replace the above expresions in eq .
( [ tj ] ) and , after fourier and bogoliubov transformations , the effective hamiltonian results @xmath34\end{aligned}\ ] ] with @xmath35 and @xmath36 the bare hole and magnon dispersions , respectively .
@xmath37 is the bare hole - magnon vertex interaction with the geometric factors @xmath38 and @xmath39 ( @xmath40 s are the positive vectors to nearest neighbors ) , and @xmath41 and @xmath42 are the usual boguliubov coefficients . in the hamiltonian ( [ hsw ] ) constant terms related to
magnetic energy are omitted the free hopping hole term implies a finite probability of the hole to move without emission or absorption of magnons .
this is a direct consequence of the underlying _ non - collinear _ magnetic structure .
the hole - magnon interaction adds another mechanism for the charge carrier motion which is magnon - assisted .
the latter is responsible for the spin - polaron formation when a hole is injected in a non - frustrated antiferromagnet@xcite .
we will show the existence of a subtle interference between both processes that turns out to be dependent on the @xmath2 sign .
an important quantity that allows us to study the interplay between such processes is the retarded hole green function that is defined as @xmath43 , where @xmath44 is the undoped antiferromagnetic ground state with a @xmath10 nel order . in the scba @xmath44 is the spin wave af ground state , and after neglection of the crossing diagrams , the dyson s equation for the self energy at zero temperature results @xmath45 we have solved numerically this self - consistent equation for @xmath46 , and calculated the hole spectral function @xmath47 and the quasiparticle ( qp ) weight @xmath48 where the qp energy is given by @xmath49 .
we have computed the one hole spectral function @xmath50 using the lanczos exact diagonalization method ( ed ) and the scba on finite size clusters . to test the reliability of the scba predictions
we have compared both techniques on the @xmath51 sites cluster , which preserves the spatial symmetries of the infinite triangular lattice@xcite . in general , we have found a good agreement between ed and scba results for all momenta of the bz and for @xmath52 . in fig .
[ fig1 ] we show the spectral functions at three representative @xmath53 points of the bz ( see inset fig .
[ fig1 ] ) in the strong coupling regime , @xmath54 , and for both signs of the transfer integral @xmath2 . and @xmath55 for momenta @xmath56 and @xmath57 .
@xmath12 shown as filled circles in the inset ( the crosses represent the other momenta ) .
the upper ( lower ) panel displays the results for positive ( negative ) @xmath2 .
the solid and dashed lines are the exact and scba results , respectively . @xmath58 and
@xmath59 has been used.,scaledwidth=45.0% ] the exact structure of the spectral functions strongly depends on the momentum because the hole strongly scatter off antiferromagnetic fluctuations@xcite .
furthermore , the geometrical frustration is manifested in a marked dependence of the spectral function with the @xmath2 sign .
remarkably , the scba reproduces very well these exact results in the whole range of energies . in some cases from the scale of the figure
it can not be clearly distinguished a quasiparticle peak at low energies . by decreasing @xmath60
appropriately we have checked , however , the existence of a qp peak for all @xmath53 of the bz . in particular , for negative ( positive ) @xmath2 the @xmath61
( @xmath62 ) point becomes the momentum of the quasiparticle ground state .
the energy of the hole can be splitted in two contributions : @xmath63 , where @xmath64 is the energy of the localized hole , and @xmath65 is the energy gain due to delocalization of the hole@xcite . as the scba does not take into account the localized hole energy we have shifted the exact energy spectra by an amount @xmath66 for @xmath55 in order to make a proper comparison with the scba results .
the agreement between both techniques is similar to that found in the @xmath67 square lattice besides the presence of stronger antiferromagnetic fluctuations in our frustrated case .
it was shown that this agreement is a consequence of the small leading nonzero contribution to the self energy from the crossing diagrams@xcite .
the momentum and the @xmath2 sign dependence of the spectra aforementioned is strengthened in the thermodynamic limit .
in particular , we observe important differences in the quasiparticle weight @xmath68 through the bz under the change of @xmath2 sign .
using lattice sizes up to @xmath69 we have extrapolated the results to the thermodynamic limit . in fig.[fig2 ] we show the values of @xmath68 as a function of @xmath14 for representative points along high symmetry axes of the bz . .
upper ( lower ) panel is for @xmath16 ( @xmath15 ) .
the location of the @xmath70 is displayed in fig .
[ fig3 ] , scaledwidth=45.0% ] the most salient feature we have obtained is the vanishing of the quasiparticle weight at some momenta for positive @xmath2 . at the @xmath71 and @xmath72 points @xmath73 below @xmath74 and @xmath75 , respectively .
it is worth noticing the robustness and then the rapid decay to zero of @xmath76 as @xmath77 .
a similar behavior has the qp signal at @xmath78 , the ground state momentum , but it goes to zero only when @xmath79 . finally , the qp signal is very weak , but finite , at the magnetic wave vector @xmath80 for all the @xmath14 range studied ; for instance , @xmath81 when @xmath82 .
surprisingly , for @xmath83 , @xmath84 does not seem to approach one as it turns out in the square lattice case@xcite .
the non existence of quasi - particle excitations is a striking manifestation of the strong interference between the free and magnon - assisted hopping processes , tuned by the @xmath2 sign . for the strong coupling regime and a large region of the bz ,
the coherence of the qp is lost because of the quite different group velocities of the spin fluctuation and the tight binding hole motion , suggesting the likelihood of a spin - charge separation scenario@xcite for the triangular @xmath85 model . and positive @xmath2.,scaledwidth=25.0% ] on the other hand , for negative @xmath2
the quasiparticle weight is finite for all momenta and for @xmath86 .
this behavior is similar to the one found in the non - frustrated case @xcite .
it is interesting to note that now the qp ground state momentum is @xmath71 ( @xmath80 ) for @xmath87 ( @xmath88 ) .
the spectral weight of the qp becomes the most robust for the ground state momenta .
the interchange in the @xmath12 dependence of the qp weight , evidenced in a comparison between both panels of fig .
[ fig2 ] , could be thought as a remnant of the particle - hole symmetry that shifts the momenta by @xmath89 under @xmath2 sign reversal . in fig .
[ fig3 ] the shaded areas correspond to the regions where the qp weight is finite for @xmath90 and positive @xmath2 .
there is no qp signal for momenta outside the neighborhood of the magnetic goldstone modes ( @xmath91 ) .
the presence of qp signal in the non - equivalent vertices of the bz is an artifact of the used approximation : the spin wave magnetic ground state has a definite chirality ( @xmath92 ) , but our vertex interaction does not discriminate between @xmath93 , resulting the whole spectral function degenerate in both momenta.the shaded areas share the same symmetries with the triangular magnetic brillouin zone .
we have noted that as @xmath94 they shrink to the magnetic goldstones modes .
conversely , when @xmath95 there is a finite qp weight anywhere and therefore , the full bz becomes shaded .
for the particular value @xmath96 , the qp weight vanishes only around the @xmath71 point , as it is displayed by the solid line of fig .
this figure also shows the strong @xmath12 dependence of the qp weight for both @xmath2 signs .
again , a signature of the remnant particle - hole symmetry mentioned above is evidenced : the dashed line ( @xmath15 ) resembles the solid one ( @xmath16 ) shifted by the wave vector @xmath89 .
path ( see fig .
[ fig3 ] ) solid line is for positive @xmath2 and dashed line for negative @xmath2 . , scaledwidth=45.0% ] in a previous work the hole dynamics in a triangular antiferromagnet has already been studied using the scba @xcite , but the vanishing of the quasiparticle weight for positive @xmath2 has been overlooked .
our findings , however , are supported by a detailed and careful treatment of the parameters involved : the convergence factor @xmath97 , the extrapolation to the thermodynamic limit of the cluster sizes @xmath98 , and the number of frequencies @xmath99 @xcite .
points of the brilluoin zone for @xmath90 and negative @xmath2.,scaledwidth=45.0% ] by computing the quasiparticle energy scaling with @xmath14 it is possible to get some insight of the magnetic cloud surrounding the hole@xcite .
it is known that for a ferromagnetic polaron the qp energy scales as @xmath100 , while for an af polaron it follows a @xmath101 dependence . in our case
we have estimated a @xmath102 scaling , being @xmath103 ( @xmath104 ) for positive ( negative ) @xmath2 in the range @xmath105 .
a possible interpretation of such results is the following : the non - collinear short range magnetic correlations around the hole can be thought as a superposition of ferro and af correlations , and then for @xmath15 the local environment around the hole enhances its af character , while for @xmath16 the ferromagnetic character is favored .
it should be noticed that the opposite behavior holds in a simple toy model consisting of a three - sites @xmath85 model with one hole@xcite .
actually , we have checked that for cluster sizes smaller than @xmath55 , this behavior is reversed .
for instance , for @xmath106 , @xmath107 for positive @xmath2 while for negative @xmath2 it turns out @xmath108 . in this section
we discuss the main features of the spectral functions in the thermodynamic limit .
the arising structures can be traced back to the two hole motion processes present in the effective hamiltonian ( [ hsw ] ) .
the spectral functions at the @xmath78 , @xmath72 , @xmath80 and @xmath71 points are displayed in fig .
[ fig5 ] for @xmath90 and negative @xmath2 .
the spectra extend over a frequency range of order @xmath109 ; that is , the non - interacting electronic energy scale .
it is possible to discern two distinct regions : a low energy sector of magnetic origin ( magnon - assisted ) , and a high energy sector with a broad resonance related to the free hopping process .
this resonance , located at an energy near @xmath110 , has a finite lifetime of order @xmath111 and a dispersion of order @xmath2 .
furthermore , the low energy sector consists of a quasiparticle peak and higher energy resonances with a narrow dispersion of order @xmath112 as the momentun varies .
we will show below that the latter resonances can be identified with string excitations . at @xmath78 and @xmath72 both sectors
are well separated in energy , while at @xmath80 and @xmath71 they overlap . from @xmath78 to @xmath71
there is a spectral weight transfer from the high energy sector to the qp peak as it can be also noticed quantitatively in the lower panel of fig .
[ fig2 ] . in fig .
[ fig6 ] we show the magnetic resonances above the qp peak for @xmath113 ( at the ground state momentum @xmath71 ) .
we choose this particular parameter value in order to make them more visible , although they are clearly noticeable for @xmath114 . in the inset
it is shown the energies of the first three resonances _ versus _ @xmath3 .
the linear dependence of their energies allow us to associate these resonances with string excitations of a hole confined by an effective linear potential of order @xmath112@xcite .
as we have previously stated , there is a dynamical enhancement of the af environment around the hole when @xmath2 is negative , and this effect is the responsible for the strong confining potential seen by the hole , like in the square lattice case . point for @xmath113 and negative @xmath2 .
inset : energy scaling versus @xmath3 of the first three peaks labeled as i , ii and iii.,scaledwidth=45.0% ] for positive @xmath2 ( fig .
[ fig7 ] ) , the low energy sector consists of a qp peak , when it exists ( @xmath78 and @xmath80 ) , and there is no signature of string - like excitations .
this result is a manifestation of the ferromagnetic - like environment around the hole , pointed out before , for this @xmath2 sign .
again the high energy sector has a broad bandwith of order @xmath115 .
it is important to notice that at the ground state momentum @xmath78 the qp is very robust , being the only relevant feature of the spectral function .
otherwise , at @xmath72 and @xmath71 the spectral functions are almost structureless .
points of the brilluoin zone for @xmath90 and positive @xmath2.,scaledwidth=45.0% ]
we have studied the effect of geometrical frustration on a single hole moving in an antiferromagnetic background .
we chose the triangular @xmath0 model which at half filling presents a non - collinear magnetic ground state . using the spinless fermion representation and the self - consistent born approximation we derived an effective hamiltonian with two different hole motion mechanisms .
one of them arises due to spin - flip processes ( magnon - assisted ) and the other one has its origin from the ferromagnetic components of the non - collinear magnetic order .
the latter allows the free hole motion without emission or absorption of magnons and it is absent in collinear antiferromagnets .
the reliability of the analytical method ( scba ) is supported by the excellent agreement with the exact diagonalization method on small size clusters . unlike previous works@xcite
, we have performed a detailed description of the whole structures of the spectral function and we were able to trace them back to the two hole motion processes mentioned above .
in particular , we have done a careful analysis of quasiparticles , strings , and free hole resonances in the thermodynamic limit .
we have found new and reliable results of the low energy sector of the spectra .
our main result is the vanishing of the quasiparticle excitation for a _ positive _ integral transfer @xmath2 at some momenta and at finite @xmath112 .
also , we have uncovered other important change of the spectra under @xmath2 sign reversal : the presence of string - like excitations only for @xmath2 negative .
we suggest that these features are a consequence of different magnetic environment around the hole tuned by the @xmath2 sign .
based on the quasiparticle energy scaling with @xmath14 , we argue that for negative @xmath2 the local environment around the hole enhances its af character , while for the other sign the ferromagnetic character is favored . for small clusters ( @xmath116 ) , this behavior is reversed , in agreement with the three - sites problem worked out by koretsune and ogata.@xcite a deeper study of the magnetic environment using the hole wave function within the scba will be presented elsewhere @xcite .
the dynamics of a hole in the triangular lattice , for negative @xmath2 , exhibits the same features of the well known square lattice case@xcite : a well defined quasiparticle , string excitations , and antiferromagnetic correlations around the hole . on the contrary , for positive
@xmath2 the dynamics is more irregular : no well defined quasiparticle , no string excitations , and ferromagnetic correlations around the hole . for decades
researchers have tried to look for spin - liquid phases that upon hole doping , give rise to non - conventional excitations@xcite .
now we give firm evidence that such non - conventional excitations can be found in _ non - collinear _ spin - crystal phases like the one present in the triangular antiferromagnet .
we acknowledge helpful discussions with a. e. feiguin , a. greco , a. dobry , and d. poilblanc .
a. e. t. thanks the hospitality received at the universite paul sabatier in toulouse under the pics france - argentine 1490 .
we thank g. martins for pointing out an error in an earlier version of this manuscript .
this work was done under pict grant no .
n03 - 03833 ( anpcyt ) and was partially supported by fundacin antorchas .
99 p. w. anderson , science * 235 * , 1196 ( 1987 ) ; , 1839 ( 1990 ) .
a. damascelli , z. hussain , and z .- x .
shen , rev .
phys . * 75 * , 473 ( 2003 ) ; j. w. allen , solid state comm . * 123 * , 469 ( 2002 ) .
r. b. laughlin , , 1726 ( 1997 ) .
b. o. wells , z. -x .
shen , a. matsuura , d. m. king , m. a. kastner , m. greven , and r. j. birgeneau , , 964 ( 1995 ) .
e. dagotto , rev .
phys . * 66 * , 763 ( 1994 ) r. coldea , d. a. tennant , a. m. tsvelik , and z. tylczynski , , 1335 ( 2001 ) .
r. h. mckenzie , science * 278 * , 820 ( 1997 ) ; comments cond . mat . * 18 * , 309 ( 1998 ) .
l. i. johansson , f. owman , and p. mrtensson , surf .
sci . * 360 * , l478 ( 1996 ) ; h. h. weitering , x. shi , p. d. johnson , j. chen , n. j. dinardo , and k. kempa , , 1331 ( 1997 ) .
l. o. manuel , c. j. gazza , a. e. feiguin , and a. e. trumper , j. phys . :
matter * 15 * , 2435 ( 2003 ) .
t. koretsune and m. ogata , , 116401 ( 2002 ) .
m. ogata , cond - mat/0304405 ; b. kumar and s. shastry , cond - mat/0304210 ; q. h. wang , d. h. lee , and p. a. lee , cond - mat/0304377 ; g. baskaran , cond - mat/0303649 ; r. e. schaak , t. klimczuk , m. l. foo , and r. j. cava , cond - mat/0305450 . g. martinez and p. horsch , , 317 ( 1991 ) .
d. poilblanc , h. j. schulz , and t. ziman , , 6435 ( 1992 ) .
b. bernu , c. lhuillier , and l. pierre , , 2590 ( 1992 ) ; l. capriotti , a. e. trumper , and s. sorella , , 3899 ( 1999 ) . l. o. manuel , a. e. trumper , and h. a. ceccatto , , 8348 ( 1998 ) . a. e. trumper , l. capriotti , and s. sorella , , 11529 ( 2000 ) . c. l. kane , p. a. lee , and n. read , , 6880 ( 1989 ) .
w. apel , h. u. everts , and u. krner , eur .
j. b * 5 * , 317 ( 1998 ) .
z. liu and e. manousakis , , 2425 ( 1992 ) .
p. bran , d. poiblanc , and r. b. laughlin , nucl .
b * 473 * , 707 ( 1996 ) .
m. azzouz and th .
dombre , , 402 ( 1996 ) .
f. lema , ph.d .
thesis , instituto balseiro , bariloche , 1997 .
a. e. trumper _ et al .
_ , in preparation . | we investigate the single hole dynamics in the triangular @xmath0 model .
we study the structure of the hole spectral function , assuming the existence of a 120@xmath1 magnetic nel order . within the self - consistent born approximation ( scba )
there is a strong momentum and @xmath2 sign dependence of the spectra , related to the underlying magnetic structure and the particle - hole asymmetry of the model . for positive @xmath2 , and in the strong coupling regime ,
we find that the low energy quasiparticle excitations vanish outside the neighbourhood of the magnetic goldstone modes ; while for negative @xmath2 the quasiparticle excitations are always well defined . in the latter
, we also find resonances of magnetic origin whose energies scale as @xmath3 and can be identified with string excitations .
we argue that this complex structure of the spectra is due to the subtle interplay between magnon - assisted and free hopping mechanisms .
our predictions are supported by an excellent agreement between the scba and the exact results on finite size clusters .
we conclude that the conventional quasiparticle picture can be broken by the effect of geometrical magnetic frustration . |
hawking radiation refers to the theoretically predicted black hole emission of thermal radiation with a temperature of @xmath3 , where @xmath4 is the mass of the black hole and @xmath5 is the gravitational constant @xcite .
the detection of hawking radiation is extremely difficult because the hawking temperature of a black hole with solar mass @xmath6 [ kg ] is of the order of @xmath7 [ k ] , which is much smaller than the temperature of the cosmic microwave background @xmath8 [ k ] .
the unruh effect refers to the prediction that a uniformly accelerated observer sees the minkowski vacuum as a thermal bath with a temperature of @xmath9 , where @xmath1 is the acceleration of the observer @xcite .
the unruh effect is related to the hawking radiation via the principle of equivalence .
therefore , any experimental verification of the unruh effect will be useful as an analogy of the hawking radiation .
p. chen and t. tajima pointed out the possibility of testing the unruh effect using an extremely intense laser @xcite .
an electron accelerated by the laser will see the minkowski vacuum as a thermal bath ; then , radiation might be produced via the unruh effect , which is referred to as unruh radiation . using a laser with an intensity of the order of @xmath10 [ w/@xmath11 @xcite , the unruh temperature reaches @xmath12 [ k ] , which is much higher than room temperature . thus ,
unruh radiation is interesting as evidence of the unruh effect .
possible signatures of unruh radiation have been theoretically investigated by several authors .
the authors in refs . investigated the unruh radiation produced by scattering between an accelerated electron and the rindler particle of the thermal bath originating from the unruh effect .
however , the importance of quantum interference in predicting the unruh radiation has been pointed out @xcite ; we take this interference into account in our theoretical investigation of the signature of the unruh radiation @xcite .
the paper is organized as follows . in sec .
ii , we review our results in ref .. adopting a model consisting of a charged particle and an electromagnetic field , we calculate the energy flux while taking the quantum interference term into account . in sec .
iii , we show that the results in sec .
ii are consistent with the first - order quantum correction to the larmor radiation , which was previously obtained using a different approach in refs ..
we adopt a model consisting of a point particle ( with a mass @xmath2 and a charge @xmath13 ) and an electromagnetic field @xmath14 , which are coupled to each other .
the action of the model is given by @xmath15 with @xmath16 , @xmath17 , and @xmath18 , where @xmath19 is the field strength and @xmath20 is the metric of the minkowski spacetime , for which we follow the convention @xmath21 .
the equations of motion are derived from the action as follows : @xmath22 @xmath23 where we introduce the external force @xmath24 that accelerates the particle with acceleration @xmath1 using a uniform electromagnetic field @xmath25 .
the solution of eq .
( [ 032301 ] ) is given by @xmath26 where @xmath27 satisfies @xmath28 , and @xmath29 represents the retarded green s function that satisfies @xmath30 .
note that the background part of the electromagnetic field @xmath31 does not contribute to the radiation energy flux , but instead , simply uniformly accelerates the particle .
we introduce the inhomogeneous solution of the electromagnetic field @xmath32 as follows @xmath33 decomposing the trajectory of the particle @xmath34 into two parts @xmath35 and substituting ( [ 032303 ] ) into ( [ 032302 ] ) , we obtain the linearized equation of ( [ 032302 ] ) , @xmath36 in the following discussion , we omit the third - order time derivative term of the radiation reaction force @xmath37 in ( [ mddotz ] ) , because the contribution of this term to the perturbation @xmath38 is of the order of @xmath39 , so as to avoid the boiling vacuum that is due to the schwinger effect . therefore , @xmath0 is much smaller than unity .
] as mentioned in the paper by s. zhang @xcite
. moreover , this term corresponds to the effect caused by the short - distance dynamics , whose scale is of the order of @xmath40 , which is much smaller than the compton length @xmath41 . to obtain the energy flux of the unruh radiation ,
we first calculate the two - point correlation function , @xmath42 where @xmath43 represents the wightman function for vacuum fluctuations of the field and @xmath44 does not contribute to the energy flux ; therefore , we subtracted it from the two - point function .
the inhomogeneous solution is written as @xmath45 with @xmath46 , where @xmath47 satisfies the relation @xmath48 and the trajectory @xmath49 is in the r - region ( see the left panel of fig .
[ fig1 ] ) . using the expansion in ( [ 032401 ] )
, we obtain @xmath50 , where @xmath51 and @xmath52 are defined as @xmath53 and @xmath54 , respectively .
then , we obtain the formula for the inhomogeneous part @xmath32 @xmath55 where we define @xmath56 as @xmath57 and @xmath58 satisfies the relation @xmath59 ( see the left pane of fig .
[ fig1 ] ) . using ( [ 032402 ] ) , we obtain the formula ( see ref . for details ) @xmath60_{s } \nonumber \\ & & = \left ( \frac{e}{4 \pi } \right)^2 \frac{\dot{\bar{z}}^{\alpha } ( \tau_-^x)}{\rho_0 ( x ) } \frac{\dot{\bar{z}}^{\beta } ( \tau_-^x)}{\rho_0 ( x ) } + \left [ \frac{e}{4 \pi \rho_0 ( x ) } \frac{e}{4 \pi \rho_0 ( y ) } \frac{e^{\beta}_{(-)i } ( y ) e^{\alpha i}_{(+ ) } ( x)}{2 m } i_2 ( x , y ) \right .
\nonumber \\ & & \left .
+ \frac{e}{4 \pi \rho_0 ^ 3 ( x ) } \frac{e}{4 \pi \rho_0 ( y ) } e^{\beta}_{(-)i } ( y ) x^i \eta^{\alpha a } a \epsilon_{a}^ { \ a'}x_{a ' } \frac{i}{2 m } ( i_1 ( x , y ) - i_3 ( x , y ) ) \right ] \nonumber \\ & & + \left [ ( x , \alpha ) \leftrightarrow ( y , \beta ) \right],\end{aligned}\ ] ] where the functions @xmath61 , @xmath62 and @xmath63 are defined as @xmath64 for the f - region @xmath65 , where the dimensionless parameter @xmath66 is defined as @xmath67 . using the formulas of the energy - momentum tensor @xmath68 and the energy flux @xmath69 @xmath70 , we obtain the energy flux expressed as @xmath71 at a large distance ( i.e. , @xmath72 ) . @xmath73 and @xmath74 are the classical part ( the larmor radiation ) and the quantum part ( the unruh radiation ) , respectively , defined by @xmath75 with @xmath76,\end{aligned}\ ] ] where we defined @xmath77 and @xmath78 with @xmath79 . note that @xmath80 contains two terms that diverge in the coincidence limit @xmath81 .
this divergence derives from the point particle approximation @xcite .
the approximation can be removed by taking into account the finite - size effect of the particle . here , for simplicity , we omit the divergent terms .
this procedure does not change our conclusions , as long as the cutoff value is of the order of @xmath82 the right panel of fig .
[ fig1 ] plots @xmath83 and @xmath80 as functions of @xmath84 .
we see that @xmath80 is typically negative ; thus , the quantum part of the radiation flux is typically negative
. however , the total radiation flux of the classical part and the quantum part is positive ; therefore , the quantum radiation appears to suppress the total radiation .
this property is consistent with the predictions of the model based on the massless scalar field @xcite .
the quantum correction to the larmor radiation has been investigated using quantum field theory for a time dependent background@xcite .
the field theoretical approach to the quantum radiation from an accelerated charge is different from that described in the previous section .
however , the results are consistent with those presented in the previous section . in the previous section
, we adopted the unit @xmath85 ; here , we explicitly include @xmath86 .
according the results in refs .
, in the non - relativistic limit of the velocity of a charged particle , @xmath87 , and the energy of the quantum larmor radiation is expressed as the combination of the classical larmor radiation @xmath88 of the zeroth order of @xmath86 , and the quantum part @xmath89 of the first order of @xmath86 , @xmath90 where @xmath91 where dots indicate differentiation with respect to @xmath92 or @xmath93 .
we may use the relations @xmath94 , @xmath95 , and @xmath96 , for evaluating the radiation energy .
we roughly estimate the order of the radiation rate as @xmath97 note that the first - order correction to the larmor radiation ( [ 032503 ] ) is negative and that its amplitude is of the order of @xmath98 ; these two statements are consistent with the results of ( [ 032601 ] ) .
therefore , the quantum radiation obtained in the previous section is related to the quantum correction to the larmor radiation .
the quantum correction to the larmor radiation ( [ quantumlarmor ] ) is non - local in time , which might be useful for understanding the quantum radiation that is due to the unruh effect . in the r - region , @xmath47 is defined by the proper time of the particle s trajectory intersecting with the past light cone ; @xmath99 is similarly defined by the future light cone .
for an observer in the f - region , @xmath99 is the proper time of the hypothetical trajectory in the l - region intersecting with the past light cone .
( right ) functions @xmath83 ( red line ) and @xmath80 ( blue line ) with divergent terms omitted .
the sign of @xmath83 is positive ; the sign of @xmath80 is typically negative .
[ fig1],title="fig:",width=215 ] in the r - region , @xmath47 is defined by the proper time of the particle s trajectory intersecting with the past light cone ; @xmath99 is similarly defined by the future light cone . for an observer in the f - region , @xmath99 is the proper time of the hypothetical trajectory in the l - region intersecting with the past light cone .
( right ) functions @xmath83 ( red line ) and @xmath80 ( blue line ) with divergent terms omitted .
the sign of @xmath83 is positive ; the sign of @xmath80 is typically negative .
[ fig1],title="fig:",width=238 ]
we would like to thank j. yokoyama , t. suyama , and k. fukushima for helpful discussions .
we also thank prof .
p. chen for critical discussions that took place at the beginning of this study . | in this study , we investigate the signature of the unruh effect in quantum radiation from an accelerated charged particle interacting with vacuum fluctuations .
because a charged particle in uniformly accelerated motion exhibits thermal random motion around the classical trajectory because of the unruh effect , its quantum radiation might be termed unruh radiation .
we show that the energy flux of the quantum radiation is negative and that its amplitude is smaller than the classical larmor radiation by a factor of @xmath0 , where @xmath1 is the acceleration and @xmath2 is the mass of the particle .
the total radiation flux of the classical larmor radiation and the quantum radiation is positive ; therefore , the quantum radiation appears to suppress the total radiation .
interestingly , the results are consistent with the prediction for the quantum correction to classical larmor radiation , which were obtained using a different approach . |
in @xcite ( hereafter pi ) we have proposed three general models for the potential of the barred - spiral galaxy ngc 1300 , which reflect three different geometries of the system . in the extensive study presented in @xcite ( hereafter pii )
we have found several response models that reproduce at an acceptable degree morphological features of ngc 1300 .
varying the pattern speed in models with all three assumed geometries , we concluded that the main features of ngc 1300 were better reproduced in models clustered around two @xmath0 values , namely 16 and 22 km s@xmath1 kpc@xmath1 .
taking into account the index used for the comparison of the models ( pii ) and the empirical by eye assessment of the resulting morphologies , we sorted out the following four best " cases .
we remind that there was no single model reproducing all the features in an optimal way ) : 1 .
* model 1 * : a 2d ( case a in pi ) model , rotating with @xmath0=22 km s@xmath1 kpc@xmath1 .
the main success of this model is the reproduction of the bar morphology .
* model 2 * : a 2d ( case a in pi ) model , rotating with @xmath0=16 km s@xmath1 kpc@xmath1 .
the main success of this model is the reproduction of the spiral morphology .
* model 3 * : a thick disc model of cylindrical geometry ( case b in pi ) , rotating also with @xmath0=16 km s@xmath1 kpc@xmath1 .
the main success of this model is a nice reproduction of the bar s morphology .
its response at the spiral region succeeds in giving the density maxima of the model roughly at about the same area occupied by the spiral arms of the galaxy .
however , the response at the spiral region of the model forms a spurious ring structure . 4 .
* model 4 * : another thick disc model of cylindrical geometry ( case b in pi ) , rotating this time with @xmath0=22 km s@xmath1 kpc@xmath1 ( as model 1 ) , develops a spiral structure that is worth to be presented as a separate model , since the spiral arms are formed through a different dynamical mechanism than that in model 2 . to better compare with the orbital models .
isocontours of a smoothed k - band image are overplotted to emphasize the morphological details.,width=302 ] we will follow the detailed orbital analysis of these four models in the subsequent sections .
nevertheless , we want to underline , that the goal of our paper is to present the dynamical mechanisms behind the observed structures . besides the morphological similarities between response models and ngc 1300 we have chosen the four cases , because they represent different dynamical mechanisms leading to the formation of the bar and the spirals .
this gives us the opportunity to compare different possible dynamics that lead to the support of similar structures in the galactic disc . in that sense
the four models should be considered as typical cases for the description of different dynamics behind them .
a characteristic difference of the model potential from other _ barred - spiral _ models we find in the relevant literature @xcite is its asymmetry . as we can see in the near - infrared image of ngc 1300 ( fig .
3b in pi ) , the two spiral arms have different morphologies .
both arms emerge out of the ends of the bar as a continuation of an ansae structure . however , the ends of the ansae , i.e. the regions where the spiral arms begin , are not at equal distances from the center of the galaxy .
the bar itself is asymmetric with its lower end being closer to the center of the galaxy .
both spiral arms are strong at their beginning , close to the ends of the bar , in the near - infrared , even after correcting for star formation ( pi ) .
notwithstanding , they weaken as we move azimuthally away from the bar . in order to clearly observe the exact shape of the spiral arms we divide in fig .
[ relima ] the deprojected k - band image with the mean surface brightness at each radius . the arm at the left side of fig . [ relima ] could be considered as continuous , although its surface brightness varies azimuthally and has a clear minimum at a height corresponding roughly to the middle of the bar . however , the right arm in fig .
[ relima ] is conspicuously discontinuous and has a second strong part at the opposite end of the bar than the one it emerges from .
all these asymmetries are taken into account in the potential we use ( pi ) and thus in our orbital calculations .
we study a fully asymmetric dynamical behaviour .
we note that the spiral arms have in the k - band relative sharp features .
this supports a quasi - stable character , since transient features , due to different wave packages , are more likely to be smooth . in order to facilitate our calculations and the adaption of the potential models to our computer programs
, we have flipped the potential around the central column and then we have rotated it by @xmath3 counterclockwise .
this allows us to study models in which the sense of rotation of the galaxy is counterclockwise and the major axis of the bar is close to the y - axis of our cartesian coordinate system .
all figures in the present paper have this orientation . km s@xmath1 kpc@xmath1 .
we have one stable ( l@xmath4 ) and five unstable ( l@xmath5 , l@xmath6 , l@xmath7 , l@xmath8 , l@xmath9 ) lagrangian points indicated with arrows.,width=302 ] equations of motion are derived from the hamiltonian @xmath10 where @xmath11 are the coordinates in the cartesian frame of reference as defined above , rotating with angular velocity @xmath12 .
@xmath13 is the potential in cartesian coordinates and can be anyone of the three models presented in pi .
e@xmath14 is the numerical value of the jacobi constant and dots denote time derivatives .
as we have seen in pii the range of @xmath12 values for which a certain similarity between response models and galaxy could be traced was @xmath15 km s@xmath1 kpc@xmath1 .
beyond that range the resemblance of the response models with the morphology of the galaxy was visibly problematic .
so we did not follow the evolution of the responses in models with pattern speeds out of this range .
for @xmath16 km s@xmath1 kpc@xmath1 the isocontours of the effective potentials have a remarkable shape .
this is shown in fig .
[ effpot2d22 ] , for our model 1 ( @xmath17 km s@xmath1 kpc@xmath1 ) . besides the stable lagrangian point at the center of the galaxy ( not indicated on fig .
[ effpot2d22 ] ) and the usual equilibrium points at the sides of the bar ( l@xmath5 , l@xmath6 ) , there appears one more stable lagrangian point ( l@xmath4 ) , as a result of the complicated landscape " at the upper part of the bar . at this region
we have _ two _ unstable lagrangian points ( l@xmath7 and l@xmath18 ) .
l@xmath18 appears close to the middle of the upper semi - major axis of the bar , at a lower e@xmath14 value ( @xmath19 ) than that of l@xmath7 ( @xmath20 ) .
thus , between the unstable points l@xmath18 and l@xmath7 appears l@xmath4 , which is stable .
nevertheless , close to the lower end of the bar , we have only one unstable lagrangian point , l@xmath9 , as usual ( fig .
[ effpot2d22 ] ) . due to the asymmetry of the potential
, l@xmath9 has a larger e@xmath14 value ( @xmath21 ) than that of l@xmath7 .
we also note , that l@xmath5 and l@xmath6 in this case are unstable ( see pi ) .
however , as we will see below , the associated simple periodic orbits are stable in large intervals of the jacobi constant . to our knowledge
it is the first time that a systematic orbital analysis is done in a galactic effective potential with more than two equilibrium points roughly along the major axis of the bar . [
cols="^,^ " , ] the histogram with the jacobi constant statistics for model 1 is given in fig .
[ h2d22inout ] . with grey
is given the statistics of the particles at @xmath22 kpc , while the black bars that come on top of the grey ones refer to the total number of particles and appear at higher levels beyond a certain value of the jacobi constant .
we observe that we do not have any particles for e@xmath14 @xmath23 . by looking at fig .
[ char2d22 ] we realize that no particles are trapped around stable banana - like periodic orbits . only small segments of the characteristics of b4_1 and b5_1 exist at e@xmath14 @xmath24 .
especially close to the upper local maximum of the zvc , close to l@xmath5 , there are only unstable b4_1 periodic orbits .
the interval e@xmath14 @xmath24 includes only a small segment of stable periodic orbits belonging to the b5_1 banana - like family .
practically the characteristics at the right part of fig .
[ char2d22 ] belong to families that are not related with the model .
the grey histogram , that refers to the bar particles , shows that these particles have jacobi constants in the interval @xmath25
e@xmath14 @xmath26 , with a peak at @xmath27 e@xmath14 @xmath28 . what is the orbital behaviour at these jacobi constants ?
the range of the jacobi constants roughly covers the region between the last x1 orbits and the local maxima of the zvc close to the l@xmath5 and l@xmath6 points ( fig .
[ char2d22 ] ) . knowing the range of e@xmath14 s that contributes mostly to the orbital content of the bar , we investigate the surfaces of section at these jacobi constants .
the ( x , x ) surface of section for e@xmath14 = @xmath29 , close to the mode of the distribution , is given in fig .
[ sos170ic ] .
there are depicted about @xmath30 consequents with black points .
the only two conspicuous islands of stability belong to the asymmetric x4 retrograde orbits .
the one with negative velocity is like the dashed orbit in fig .
[ ox4 ] , while the other one , with positive velocity , is its symmetric with respect to the y=0 axis .
apart from these two islands of stability the rest of the surfaces of section is dominated by chaos .
we know however , that at this e@xmath14 we encounter most particles that contribute to the bar structure .
= @xmath31 in model 1.,width=302 ] the chaotic sea can be in principle generated by a single chaotic orbit , if we integrate it for a long time . however , starting with different initial conditions and by integrating for relatively short times we do not obtain necessarily similar orbital morphologies . integrating for finite time intervals a large number of initial conditions we simulate the situation of following the trajectories of individual particles on the galactic disc , and this allows us to understand whether they support a specific morphology within this time interval .
the red symbols overplotted on the surface of section indicate 587 initial conditions on the ( x , x ) surface of section in fig .
[ sos170ic ] , that have been integrated for time corresponding to 7 pattern rotations .
we will refer to all of the resulting trajectories also as orbits " , despite the fact that the chaotic orbit is essentially one .
the supported morphologies we found were quite distinct , and this allowed us to proceed to a crude by eye classification as follows : around 7% of the trajectories were related with the x4 retrograde family , being either regular quasi - periodic orbits trapped around them , or related sticky chaotic orbits . a typical orbit of this kind is given in fig . [ orbits170_no]a .
another 40% of the integrated trajectories were chaotic , staying in the inner bar region without ever exceeding a radius r=4 kpc .
they were essentially supporting the part of the bar without the ansae .
such an orbit is presented in fig .
[ orbits170_no]b .
only 1% of the chaotic orbits could be associated with x2-like orbits ( fig .
[ orbits170_no]c , cf . with fig . [ ox2 ] ) .
the remaining 52% of the integrated initial conditions belong to orbits in the chaotic sea , which share a common feature .
these particles , after spending part of the integration time in the central bar region as in fig .
[ orbits170_no]b , they visit the immediate neighborhood above and below the central part contributing to the formation and support of the ansae .
such orbits are presented in fig .
[ orb_ansae ] .
we note that l@xmath18 appears at e@xmath14 = @xmath19 , very close to the e@xmath14 value of the surface of section we study , indicating that the appearance of this type of orbits is associated with the presence of the stable and unstable manifolds of the unstable l@xmath18 equilibrium point .
this type of _ chaotic _ orbits give the strong ansae type morphology of the bar of the response model 1 .
the other morphological feature , that characterizes model 1 , are its open spirals ( fig . [ resp ] ) . despite the fact , that they do not match the ngc 1300 spirals , they are a strong feature of its grand - design morphology .
they consist of particles at @xmath32 kpc . as we can conclude from fig .
[ h2d22inout ] the mode , the value that occurs most frequently in the distribution of these particles , is found for @xmath33 e@xmath14 @xmath34 . in order to find
the orbital content of the model spirals we follow the same procedure .
the ( x , x ) surface of section for e@xmath14 = @xmath35 is presented in fig .
[ sos167 ] . being closer to the local maxima of the zvc ( fig . [ char2d22 ]
) there appears in the diagram also a region to the left of the central area of the surface of section . at the central region
the only conspicuous islands of stability are again those belonging to x4 , as in the previous case we studied at e@xmath14 = @xmath36 .
we have integrated again a large number of initial conditions for a time equal to 7 pattern rotations . in this case
the number of orbits that support a bar with ansae increased to 63% , while an additional 26% abandon the bar region and by going through the l@xmath37 location visit the area of the disc and the spirals .
typical orbits of the kind we describe can be observed in fig .
[ orb167 - 2d22 ] .
the mechanism that shapes the open spirals of model 1 is based on the presence of the family generated at the unstable lagrangian point l@xmath37 , which has e@xmath14 = @xmath20 . for a jacobi constant just greater than it , at e@xmath14 = @xmath38 , which is the jacobi constant of the surface of section we study ,
the unstable periodic orbit associated with the open spiral pattern already exist .
the chaotic sea is structured by the presence of the manifolds of this unstable periodic orbits .
for this particular response model , the mechanism that builds its spiral arms is the one valid also for the spirals of ngc 4314 @xcite . this mechanism , as in the case of ngc 4314 , creates spirals with a strong part , that reaches azimuthally angles up to @xmath39 starting from the ends of the bar .
details about how chaotic spirals can be formed this way can be found in @xcite , @xcite , @xcite , @xcite , @xcite ( see also articles by the same authors in @xcite and references therein ) .
our analysis shows however , that the spirals of ngc 1300 can not be related with this mechanism .
thus , we continue the investigation of our response models in order to find a better explanation of the dynamical origin of the ngc 1300 spirals .
we remind though , that the model 1 bar reproduces in a satisfactory way the bar of ngc 1300 .
in pii we have seen that by lowering the pattern speed of the system in the 2d case to @xmath40 km s@xmath1 kpc@xmath1 we obtained an excellent matching of the location of the spirals between model and galaxy .
this case is our model 2 of the present paper .
we summarize its morphology with the help of fig .
[ resp2d16 ] .
there are three main successes of model 2 in its comparison with the ngc 1300 spirals : ( i ) the abrupt break of the upper arm , just after it emerges from the upper end of the bar .
this is indicated in fig .
[ resp2d16]a with an arrow labeled a " .
( ii ) the spiral fragment at the lower right side of the bar . an arrow labeled b " points in fig .
[ resp2d16]a to this feature .
( iii ) the location and the pitch angle of the rather continuous left arm , indicated with the arrow labeled c " .
the overplotted on the model isophotes of the deprojected k - band image , show clearly the agreement of these features ( fig .
[ resp2d16]b ) .
this time the spiral structure is modeled at a very satisfactory level . however , the only feature of the bar that agrees with the ngc 1300 bar is its size .
the model bar is thicker than the one of the galaxy and has only a weak ansae character . particularly interesting
is the comparison of the effective potential isocontours ( fig .
[ effpot2d16 ] ) with the resulting response model .
now , for @xmath40 km s@xmath1 kpc@xmath1 , the lagrangian points have moved to larger distances from the center .
we have the usual two stable l@xmath41 and l@xmath42 equilibrium points ( at approximately @xmath43 and @xmath44 ) and the two unstable l@xmath7 and l@xmath9 at @xmath45 and @xmath46 respectively .
besides them there is a third unstable equilibrium point , l@xmath47 , at @xmath48 . as a consequence ,
there is one more equilibrium point , l@xmath49 , between l@xmath7 and l@xmath47 , which in this case is unstable .
all lagrangian points appear beyond the end of the barred - spiral morphology of the response model altogether .
contrarily to model 1 , all the lagrangian points in model 2 are within a relative narrow ring of width about 1.8 kpc
. however , the @xmath50 ratios we find deviate more among themselves . by considering the upper length of the bar and the l@xmath37 radius
we find this time @xmath50=1.44 , while we have @xmath50=1.68 if we consider the lower side of the bar and the l@xmath51 radius .
we underline the presence of an isocontour in the effective potential , which we have drawn with a thicker line and to which we point with a thick white arrow .
there is a striking similarity between the shape and the location of this isocontour in fig .
[ effpot2d16 ] and the border of the red region that surrounds the barred - spiral structure of the response model 2 in fig .
[ resp2d16 ] .
we discuss the role of this isophote for the dynamics of model 2 in the following paragraphs .
the characteristics of the main families of periodic orbits in an ( e@xmath14 , @xmath52 ) diagram , similar to the one presented in fig .
[ char2d22 ] for model 1 , is given in fig .
[ char2d16 ] .
the evolution of the x1 , x2 families is similar to that of model 1 , but the jacobi constant interval between the last stable x1 and the corotation region has increased significantly . between them
we have calculated segments of characteristics belonging to families with asymmetric morphologies we have classified as 4/1 , 7/1 and 8/1 .
we also note the presence of segments of several 3/1 type families close to the largest jacobi constants , where we still encounter x1 and x2 orbits . at all jacobi constants in fig .
[ char2d16 ] we have always orbits belonging to x4 and its bifurcations .
it is beyond the scope of the present paper to study the interconnections between the various families .
this time we have found periodic orbits extended roughly along the major axis of the system , which could contribute to the bar structure . in fig .
[ pobar2d16 ] we present some of them independently of their stability .
( a ) the longest stable x1 orbit at e@xmath14 = @xmath53 .
it has loops that now reach distances about 5 kpc from the center , ( b ) two 3/1 orbits at @xmath54 ( dark grey ) and @xmath55 ( light grey ) , that are close to symmetric with respect to the y - axis and together can be considered as bar supporting , ( c ) a 4/1 orbit from the tiny stable part of the characteristic of this family at e@xmath14 = @xmath56 and finally at ( d ) the 4/1 representative at e@xmath14 = @xmath57 .
km s@xmath1 kpc@xmath1 .
we have two stable ( l@xmath5 , l@xmath6 ) and four unstable ( l@xmath7 , l@xmath18 , l@xmath58 , l@xmath9 ) lagrangian points indicated with black arrows .
the white thick arrow points to an isocontour , that surrounds the bar - spiral morphology.,width=302 ] we also give the morphology of the two families above the 4/1 , at larger @xmath52 . in (
a ) a stable 8/1 ( e@xmath14 = @xmath59 ) and in ( b ) an unstable 7/1 ( e@xmath14 = @xmath60 ) ( fig .
[ pospir2d16 ] ) . in order to trace the orbits of the particles that successfully reproduce the spiral structure of ngc 1300 in model 2 , we follow the same procedure as in model 1 .
we isolate an area around the bar of the response model , which includes the spiral arms and we calculate the jacobi constants of the particles in this area .
the histogram with the jacobi constant statistics at this region is given in fig .
[ histoedw ] .
the histogram shows that the peak happens at the bin @xmath61 e@xmath14 @xmath62 , while practically the jacobi constants of all particles are @xmath63 e@xmath14 @xmath64 .
most particles on the spirals of model 2 have these e@xmath14 values . by looking at fig . [ char2d16 ]
we realize that at this region exist the 7/1 and 8/1 families , the 4/1 " family with double multiplicity , as well as a segment of the 4/1 characteristic . only for the 8/1 family
the hnon index is @xmath65 ( black segments in fig .
[ char2d16 ] .
this analysis shows that whatever the dynamical mechanism behind the appearance of the spiral structure is , it is related with particles on trajectories inside corotation .
the maxima of the zvc s are near l@xmath5 , l@xmath6 .
[ char2d16 ] on the ( x , x ) surface of section for e@xmath14 = @xmath66 there are no islands of stability discernible . for this jacobi constant
all families of periodic orbits we mentioned existing at the @xmath63 e@xmath14 @xmath64 interval are unstable . again in this case we integrated a large number of initial conditions on a grid for time equal to seven pattern rotations .
the inspection of these orbits helped us realize that the vast majority of the chaotic orbits during the integration time have loops at their apocentra almost on the thick isocontour of the effective potential drawn in fig .
[ effpot2d16 ] .
the red region surrounding the barred - spiral structure in fig .
[ resp2d16 ] is formed by the same orbits .
as these orbits perform their loops , they stay longer time at the region of the heavy isopotential in fig .
[ effpot2d16 ] and this is the way the local density maximum corresponding to the spiral arms are formed . in fig .
[ np2d16_1535 ] we give six typical spiral supporting orbits in model 2 with e@xmath14 = @xmath66 ( upper row ) and with e@xmath14 = @xmath67 ( lower row ) .
we plot them having as background the response of the model in order to clearly see the relation between orbits and response morphology .
the initial @xmath68 conditions are @xmath69 ( a ) , @xmath70 ( b ) , @xmath71 ( c ) and @xmath72 ( d ) , @xmath73 ( e ) and @xmath74 ( f ) .
we have practically the same dynamical behaviour at every jacobi constant in the interval @xmath63 e@xmath14 @xmath64 .
orbits of similar morphology are found by integrating initial conditions at all these surfaces of section , and contribute to the formation of the spiral pattern .
the only difference is in the mean radius of these chaotic orbits , which depends on the jacobi constant .
their existence over a @xmath75e@xmath14 range give the observed thickness of the model spirals .
the reason for which we have the formation of an open spiral and not a ring is that the orbits that participate in the formation of the structure are chaotic .
longer stability intervals along the characteristics of the involved families would trap a large number of quasi - periodic orbits around them leading to the formation of rings on the response models .
even the size of the islands of the stable 8/1 orbits are tiny , so the amount of trapped orbits is small . however , despite the fact that the stable periodic orbit influences a small part of the phase space and the fact that apart from the 8/1 family all other periodic orbits are unstable , we observe a similarity between the morphology of the existing periodic orbits ( see fig . [ pospir2d16 ] ) and the chaotic orbits that support the spiral structure ( fig .
[ np2d16_1535 ] ) .
this dynamical behaviour is similar to the one found by @xcite for the orbits that support the outer envelope of the bar in the potential of ngc 4314 .
model 2 gave an acceptable solution for the spirals of ngc 1300 .
however , out of the morphological features of the bar only its size matches the size of the ngc 1300 bar .
the combination of the two acceptable solutions , namely model 1 for the bar and model 2 for the spirals , in a single model with two pattern speeds is problematic .
a careful inspection of the ngc 1300 morphology shows that a bar rotating with a different pattern speed than the spirals will after a certain time visit the spiral region and the combined structure will change drastically in time . a way out of this dilemma
is offered by our next model we have chosen to present , i.e model 3 .
the two models we have presented up to now have a planar geometry . for models in the general class of thick disc models ( pi ) , we have realized in pii , that in the range of @xmath12 around 16 km s@xmath1 kpc@xmath1 @xmath76 the bar was becoming considerably thinner than the corresponding 2d models with the same pattern speed . in this section we describe the orbital behaviour in a thick disc model ( see pi ) , rotating with @xmath12=16 km s@xmath1 kpc@xmath1 . around this @xmath12 value
the @xmath77 index in pii improves in models , where we weight the jacobi constants we use in our models .
however , the thinning of the bar and the enhancement of the density at the ends of the bar that emphasizes its ansae character , is a basic feature of all thick disc models with relatively low patterm speeds . the best similarity with the ngc
1300 image is observed at @xmath12=16 km s@xmath1 kpc@xmath1 @xmath78 ( see fig . 3 in pii ) .
the @xmath50 ratios we find considering the l@xmath37 and l@xmath51 lagrangian points are this time 1.5 and 1.7 respectively .
the response of our model 3 is given in fig .
[ respo3dd16 ] .
the feature that appears considerably improved compared with model 2 is the bar .
the length of the bar remains the same , however the ends of the bar appear thinner and elongated at the region which corresponds to the ansae of the ngc 1300 bar ( see also the corresponding section in pii ) . on the other hand
less successful is the model in describing the spiral of the galaxy . at the same region , where we obtained the spiral arms in the 2d model 2 ,
now a ring is formed .
the density along this ring is not constant , but the resulting feature does not match the spiral of the galaxy as good as the model with the planar geometry ( cf . figs .
[ resp2d16 ] and [ respo3dd16 ] ) .
below we describe the differences in the orbital dynamics between the two models , that are reflected in the morphology of their responses . the evolution of the ( e@xmath14 , @xmath52 ) characteristics of the main families of periodic orbits for model 3 is given in fig .
[ char3dd16 ] .
in contrast with model 2 , the x1 characteristic is stable also at the raising part of the curve beyond e@xmath14 @xmath79 .
simultaneously , already for e@xmath14 @xmath80 , there exists a stable 4/1 family , symmetric , with orbits of rhomboidal morphology .
the families attributed to higher order resonances ( 7/1 , 8/1 ) still exist , and even at a larger jacobi constant interval than in model 2 .
this time we find a stable part at the 7/1 characteristic , as well as a small stable part of 8/1 .
the morphology of the orbits of these families is similar to that of the corresponding families of the planar model ( fig .
[ pospir2d16 ] ) .
finally , we mention the presence of several 3/1 families , also with stable parts , at @xmath81 e@xmath14 @xmath82 .
looking first for bar supporting orbits , we have found that they belong to x1 , 4/1 and even to 3/1 families . in fig .
[ bs3dd ] we present stable bar supporting periodic orbits with loops , that can be considered as a skeleton of a bar structure .
an arrow labeled a " points to the innermost orbit , which belongs to the x1 family and has e@xmath14 = @xmath83 .
the morphological evolution of x1 , also in this model , follows that of the propeller orbits described in @xcite . orbit
b " , belongs also to the x1 family , has e@xmath14 = @xmath84 , i.e. it is at the beginning of the raising part of the x1 characteristic .
the third orbit of the sample , named
c " , is from the 4/1 rhomboidal family at e@xmath14 = @xmath85 and has asymmetric loops , with the lower loop being larger than the upper one .
finally the outermost periodic orbit ,
d " , is an orbit of multiplicity 3 , the characteristic of which is the segment , like an extension , at the local maximum of the 4/1 characteristic .
this orbit has e@xmath14 = @xmath86 .
= 16 km s@xmath1 kpc@xmath1 .,width=302 ] model 3 differs from the other models studied up to now , in that there are always stable bar supporting orbits up to the 4/1 resonance region . besides the stable simple periodic orbits
the surfaces of section are characterized by the presence of islands of stable periodic orbits of higher multiplicity and sticky regions around them .
we give a typical example in fig .
[ sos3dd16 ] .
it is the ( x , x ) surface of section for e@xmath14 = @xmath87 .
the stable periodic orbit at the center of the island of stability belongs to the 4/1 family .
its shape is indicated in a box at the lower right corner of the figure .
another stable periodic orbit at this jacobi constant is a periodic orbit of multiplicity three .
the three islands of stability , to the left of the 4/1 island , belong to it and its shape is presented in a box at the upper left corner of fig .
[ sos3dd16 ] .
note the darker areas , that appear as bridges between the islands of the triple orbit and the island of stability of the 4/1 orbit . as we can observe from the morphology of these two periodic orbits both of them
are bar supporting .
bar supporting are also the sticky and other chaotic orbits that we integrated for time equal to seven pattern rotations .
we can say , that the bar of model 3 is built to a large extent in the classical way , i.e. with regular orbits trapped around stable periodic orbits .
this bar has many common features with the ngc 1300 bar .
we use model 4 to present another mechanism , that we found to support a spiral structure resembling that of ngc 1300 .
the mechanism involves orbits around the lagrangian points l@xmath5 and l@xmath6 .
the efficiency of this mechanism in supporting spirals was more evident in models with thick discs , especially in the range 20 km s@xmath1 kpc@xmath1 @xmath8823 km s@xmath1 kpc@xmath1 .
model 4 is a thick disc model ( see pi ) rotating with @xmath8922 km s@xmath1 kpc@xmath1 . unstable .
[ char3dd16 ] = @xmath87 .
the big island belongs to a stable orbit of the 4/1 family , while the three islands to a triple periodic orbit .
the morphology of the two orbits is given in boxes.,width=302 ] the overall response of the model and the relation of it with the k - band morphology of ngc 1300 is given in fig .
[ respo3dd22 ] .
the best reproduced feature of this model with respect to the rest of the models we have examined until now is the beginning of the left spiral , which emerges from the lower end of the bar .
there is clearly an increase of the local density of the model along the direction , where the left spiral of the galaxy extends , as we can see in fig .
[ respo3dd22]b .
we also can observe that the right spiral arm of the model breaks at about the same region as the spiral of the galaxy and also that there is an increase of the local density of the model close the lower end of the bar , at the right side , in general agreement with the fragment of spiral arm we find at the same region at the image of the galaxy .
a feature of the effective potential of this model worth pointing out is the characteristic asymmetry of the lagrangian points l@xmath37 and l@xmath51 with respect to the major axis of the bar .
the @xmath50 ratios we find for this model considering the l@xmath37 and l@xmath51 lagrangian points , are 1.15 and 1.16 respectively .
we have taken three sets of particles from three dense regions of the model spirals in order to find their jacobi constants .
this is given in fig .
[ sele3dd22 ] , where these regions are labeled with `` 1 '' ( upper part of the left arm ) , `` 2 '' ( lower part of the right arm ) and `` 3 '' ( spiral segment at the right side , close to the lower end of the bar ) .
the statistics of the jacobi constants are given in fig .
[ histo3dd22 ] .
green and red correspond to the the boxes labeled with `` 1 '' and `` 2 '' in fig .
[ sele3dd22 ] respectively , while the histogram of region `` 3 '' is given in yellow - grey .
the peak of the jacobi constants of the particles in all three regions is at @xmath90 e@xmath14 @xmath91 , while we observe , that the jacobi constants of all particles we consider , have @xmath92 e@xmath14 @xmath93 .
so we focus our investigation to spiral supporting orbits at these jacobi constants . as regards periodic orbits , we give in fig .
[ char3dd22 ] with heavy lines the ( e@xmath14 , @xmath52 ) characteristics of the families that are involved in the enhancement of the spiral arms in model 4 .
the left and right vertical lines indicate the extent of the e@xmath14 interval that we have to investigate as results from the statistics presented in fig .
[ histo3dd22 ] , i.e. @xmath92 e@xmath14 @xmath93 .
the two lines between them show the part of the interval , where we have the two highest bars of the histogram , i.e they are at e@xmath14 = @xmath85 and @xmath87 respectively .
the darkness of the grey shade reflects the number of particles at the bins of the histogram in fig .
[ histo3dd22 ] .
essentially the interval of the jacobi constants we study includes parts of the characteristics of three families with stable parts . l@xmath5 and l@xmath6 are stable in this model .
we study two characteristic surfaces of section for two close by values of e@xmath14 . in fig .
[ sos3dd ] we present them for e@xmath14 = @xmath94 in ( a ) and for e@xmath14 = @xmath87 in ( b ) . in ( a )
we are located still before the local maximum of the zvc at negative @xmath52 s , close to l@xmath6 .
the surface of section is practically chaotic everywhere .
however , we discern a darker region at the right side of the left part of the surface of section . in
( b ) the two regions , the central and the left one , have merged and a family of stable banana like orbits , originated at l@xmath6 appears ( invariant curves in the dense region at @xmath95 kpc ) .
it is surrounded by a sticky region , while another dark , sticky , region can be observed at the right part of the central region .
we followed the same procedure as in the previous models and we integrated for time equal to 7 pattern rotations a large number of initial conditions on both surfaces of section of fig .
[ sos3dd ] . in the subsequent figures we group the orbits of the particles on the spirals as follows : figure [ orb3dd_229a ] presents characteristic orbits at the sides of the model bar . with grey
are the trajectories at the right side .
we observe the loops at areas that correspond to the ends of the right side of the bar .
simultaneously these orbits leave rather empty the region between 5@xmath9610 kpc .
with black are given orbits at the left side of the bar this time .
most of the loops of these orbits are along the region of the left spiral arm of the model , while at their lower left side they perform in most cases a single loop that brings them closer to the bar .
more efficiently acts this mechanism at e@xmath14 = @xmath87 . for the same reasons we explained in the case of e@xmath14 = @xmath94 , the orbits at fig .
[ orb3dd_228r ] , which have e@xmath14 = @xmath87 , contribute to the formation of the spiral features . with black
we give orbits of various sizes surrounding l@xmath6 .
the loops of the large size orbits enhance the lower and upper part of the left spiral arm , while the small sized contribute to the formation of the central part . with grey
we give orbits surrounding l@xmath5 .
summarizing the response of model 4 we can say , that density maxima aligned along the location of the spiral arm features of ngc 1300 , are provided by the banana - like orbits .
it is remarkable also that the characteristic break of the right arm , just after it emerges from the bar , and the presence of the spiral segment close to the lower right end of the bar are also reproduced in an acceptable degree by model 4 ( fig .
[ respo3dd22 ] ) . however , the drawback of the mechanism is that one can not get rid of the additional features that are brought in the model from the same orbits that reinforce the spirals . as we can see in fig .
[ respo3dd22]b , there is a lot of additional structure that appears together with the spirals to the left of the spiral arm .
this structure is formed by the part of the orbits that is far from the bar .
it is a quite robust feature and can not be smoothed out even by increasing the dispersion in the velocities of the initial conditions of response models with @xmath97 km s@xmath1 kpc@xmath1 , always in the thick disc approximation .
we finally note that in model 4 we have the clear evolution of two spiral features that start with an overdense region almost on the major axis of the bar , but away from it .
the lower of the two starts in fig .
[ respo3dd22 ] at about @xmath98 and goes out of the frame at @xmath99 .
this feature , as well as its symmetric with respect to the center of the galaxy at the upper part of the model , are related with orbits associated with the outer resonances , beyond corotation @xcite .
however , it seems that they do not have any correspondence in the near - infrared morphology of ngc 1300 .
the class of models that is not represented in our analysis is the one that assumed a spherical component representing the major part of the bar ( without the ansae ) ( see pi ) .
this class of models did not improve the resemblance between model and galaxy .
however , in pii we have seen that the model of this class especially with @xmath100 km s@xmath1 kpc@xmath1 had a closer resemblance to the galaxy in its very central part .
the orbital analysis has shown that the inclusion of an extended spherical component introduces in the system circular , and nearly circular , x1 orbits at the major part of the characteristic of this family .
the characteristic remains close to the @xmath101 axis at long e@xmath14 intervals . in the particular model
the presence of the circular x1 orbits secures a round central region of radius @xmath102 0.5 kpc . for larger jacobi constants the orbits of the x1 family become asymmetric with respect to the x - axis .
their morphology becomes elliptical with one focus at the origin of the axes of our system and the other at negative y. this follows the tendency of the central region of the galaxy to become elongated towards negative y. the results of all models presented here do not alter qualitatively even if we chose different projection parameters on the sky for the disc of ngc 1300 .
we find similar dynamical mechanisms acting also for the projection parameters ( pa , ia)=106.6@xmath103 ) , which are those one finds by fitting an exponential disc to the regions outside the main bar in the k - band image , as in @xcite in agreement with @xcite .
a few examples have been presented in @xcite .
the dynamical mechanisms are similar , the @xmath0 numerical values however , differ .
following the framework established in pi , we performed extensive modeling of the stellar component of ngc 1300 in three classes of models reflecting different geometries .
as we have seen in pii , we obtained acceptable solutions for the bar or the spirals not only for more than one pattern speed , but for different geometries as well .
anyhow , the main parameter that determines the morphology of our response models is @xmath0 .
the orbital analysis of the present paper , shed light on the various `` promising '' models and provided feedback , that could be used for the determination of the optimum solution .
however , the most interesting result was not the determination of the optimum parameters for a `` best '' ngc 1300 model , but the dynamical mechanisms themselves , which are leading to the desired morphology
. even models to which our analysis was not pointing to as best matching the ngc 1300 morphology , have a special interest , since their parameters are equally realistic for a barred - spiral system in general as those best applying to the specific galaxy .
summarizing the four models we have chosen to present , we found that there have been two values of @xmath0 for which our models were giving the best matching morphologies to ngc 1300 in the three classes of models we considered , namely @xmath0=16 km s@xmath1 kpc@xmath1 and 22 km s@xmath1 kpc@xmath1 .
the new dynamical mechanisms that we found deserving special attention and extensive presentation , were the following : * in a 2d model rotating with @xmath0=22 km s@xmath1 kpc@xmath1 , a strong bar with ansae is supported practically _ only by chaotic orbits_. this bar has similar morphology with the bar of ngc 1300 .
decisive role in the support of such a structure by chaotic orbits plays the presence on the effective potential of multiple lagrangian points roughly along the major axis of the bar
. we can understand this by considering an @xmath104 curve , which is not monotonically decreasing , but has a wavy character .
then it is possible to get the same @xmath0 at two ( or more ) points along the @xmath104 curve .
this is the case of the bar in our model 1 .
it is an alternative mechanism for the support of the ansae morphology based only on chaotic orbits .
probably it can be applied to many bars showing this feature . * in a 2d model rotating with @xmath0=16 km s@xmath1 kpc@xmath1 , our model 2 , the morphology of the spiral arms of ngc 1300 can be reproduced by chaotic orbits at e@xmath14 s between the 4/1 resonance and corotation . in our case
the e@xmath14 values of the orbits that support the spirals are well inside corotation .
the spirals are supported by the bouncing " of the chaotic orbits on the walls " of the effective isopotential drawn with a heavy line in fig .
[ effpot2d16 ] . * in a thick disc model rotating with @xmath0=22 km s@xmath1 kpc@xmath1 , we found a mechanism supporting a spiral structure similar to that of ngc 1300 with orbits around the lagrangian points l@xmath5 and l@xmath6 ( our model 4 ) .
these orbits are either regular orbits trapped around stable periodic orbits at the region , or sticky chaotic orbits , or even chaotic trajectories from a pool of a chaotic sea .
all of them contribute to the support of the spiral arms .
the possibility that a set of banana - like orbits can support a spiral morphology has been indicated by @xcite and @xcite ( see also * ? ? ?
we do not include in our enumeration of the new mechanisms the formation of a typical ansae bar by regular orbits as in our model 2 . by `` chaotic orbits '' we actually mean trajectories , that occur after integrating for a given time initial conditions , which we select out of a common chaotic sea , i.e. essentially we examine fragments of a single chaotic orbit .
if at a certain e@xmath14 a percentage of these trajectories supports a given structure ( a bar or a spiral ) it practically means , that the orbit that fills the chaotic sea supports this structure during the same percentage of its , long , integration time .
our result , that surfaces of section without any discernible island of stability provide support to a specific morphological feature , means that the chaotic sea is structured in such a way , that allows this dynamical behaviour . despite the fact
, that there is no obvious structure in a chaotic sea , it is known that the manifolds associated with the unstable periodic orbits existing on the surface of section do structure the chaotic sea ( e.g. * ? ? ?
* ) . in our models
the spirals that we discuss as reproducing the spiral arms of ngc 1300 are _ inside corotation_.
however , the general relation between manifolds and resulting spiral structure is expected to be similar to the one described by @xcite , @xcite and @xcite for spirals related with the family of unstable periodic orbits originated at the unstable lagrangian points l@xmath37 and l@xmath51 _ at corotation_. we do not point to a specific family of unstable orbits , the existence of which determines that of the spirals in model 2 or the ansae type bar in model 1 .
it is the collective dynamical behaviour in chaotic seas within a specific e@xmath14 range .
the dynamical mechanisms that shape a structure are in general different . in model 2
it is the restriction of the chaotic orbits by a characteristic isocontour of the effective potential and the loops of the orbits as they `` bounce '' on it . on the other hand in model 1
it is the trapping of the orbits in the central area and the presence of the double l@xmath37 points that shape the morphology of the bar .
the structures that are supported by chaotic orbits are dictated to a large extent by the shape of the isocontours of the effective potential .
the fact that star clusters in the spiral arms of ngc 1300 are not well aligned as in other grand - design spiral galaxies ( e.g. ngc 2997 ) @xcite is in agreement with a chaotic character of the orbits of the stars that build them .
it has been proven quite difficult to point to a best ngc 1300 model .
the assumption of different pattern speeds for the bar and the spirals seems less probable , due to the relative locations of the two components .
this is not concluded from the apparent connection of the spiral arms and the bar , but mainly from the fact that we do not find two different , separated zones , where bar and spirals rotate without interacting .
a bar rotating faster than the spirals will reach the spiral arms at a certain fraction of the spiral period and then there will be an interference between the two components .
in such a case the observed morphology would be just a snapshot in a transient morphology of the barred - spiral system .
an indication against the transient spiral hypothesis is that the spiral arms exist in k - band and they are even sharp features as we mentioned in the introduction . if we have to deal with a more or less stationary morphology , the regions where bar and spirals rotate should be rather separated . as an example we mention the ngc 3359 model by @xcite , where in a two pattern speeds model the orbits of the spiral component change their topology in such a way as to allow the bar to rotate within a certain radius leaving the spirals outside it ( their fig .
20 ) . ,
while the particles of region `` 3 '' are represented by the yellow - grey bars.,width=264 ] , @xmath52 ) diagram for the main families in model 4 .
only the families drawn with heavy lines play a role in the formation of the spirals of the model .
the vertical lines indicate the extent of the interval of jacobi constants we study ( see text).,width=309 ] from the large number of response models we presented in pii it was evident that it was more difficult to find a response spiral matching the spiral arms of ngc 1300 , than to model the bar .
the reason is that at a certain range of @xmath0 values , in all three general classes of models we studied , the length of the bar is close to the length of the bar of the galaxy .
from the two mechanisms that we found supporting the galaxy s spiral , the one in model 2 ( chaotic orbits between 4/1 and corotation ) has more advantages than that of model 4 ( regular and chaotic orbits around l@xmath5 and l@xmath6 ) .
model 2 follows closer the pitch angle of the left arm of ngc 1300 , while the spiral fragment at the right of the lower end of the bar is more conspicuous in model 2 than in model 4 .
the chaotic spirals of model 2 with planar geometry can be combined with the bar of the thick disc model 3 in a model with a single pattern speed @xmath0=16 km s@xmath1 kpc@xmath1 .
on the other hand the left spiral of model 4 is combined with additional features ( fig .
[ respo3dd22 ] , which we can not ignore in the comparison with the k - band image of the galaxy .
there are regions of density maxima to the left of the left arm , that are formed due to the same orbits that give the density maxima along the response density maxima corresponding to the left arm of ngc 1300 .
these superfluous features could not be eliminated by increasing the velocity dispersion of the initial conditions . by combining model 1 with model 4 in a single model rotating with @xmath0=22 km s@xmath1 kpc@xmath1
, we have a 2d bar combined with a spiral in a thick disc beyond the bar .
thus , taking into account the advantages and disadvantages of all models we have found having a good relation with the ngc 1300 morphology we want to model , we conclude that model 3 for the bar and model 2 for the _ stellar _ spirals , both rotating with @xmath0=16 km s@xmath1 kpc@xmath1 , give the best representation of the corresponding structures of ngc 1300 .
we underline that the two preferable @xmath0 values ( 16 and 22 km s@xmath1 kpc@xmath1 ) are not chosen among equally good models with nearby pattern speeds , but it is a result of the extensive modeling presented in pii , that we have best responses for these two values .
= @xmath94 in ( a ) and for e@xmath14 = @xmath87 in ( b ) in model 4 . at this jacobi constant
we find most particles contributing to the spiral structure.,width=340 ] a comparison of our results with the modeling of ngc 1300 by other authors has been done in pii .
we want here to add that our model 1 gives lagrangian points ( fig .
[ effpot2d22 ] ) that are close to the corotation radius of the fast " model by @xcite , 104@xmath105 , ( l@xmath37 , l@xmath51 ) , as well as with the corotation radius of @xcite , @xmath10283 kpc , ( l@xmath4 , l@xmath5 , l@xmath6 ) .
as we explained there is no need for a two pattern speeds assumption for having multiple lagrangian points in a model .
below we enumerate our conclusions : 1 .
we could reproduce structures similar to those of the bar or the spiral of ngc 1300 by more than one model .
all models presented in the paper have some nice agreement with the morphology of the galaxy .
the comparison of these models with the desired barred - spiral morphology , as well as the comparison among themselves , indicates as best a configuration with a single pattern speed , where the whole barred - spiral structure extends inside corotation .
the bar apparently is composed by regular orbits ( model 3 ) , while the _ stellar _ spirals are composed by chaotic orbits ( model 2 ) .
we find an ansae - type bar morphology that can be constructed completely out of chaotic orbits ( model 1 ) .
this requires the presence of multiple unstable lagrangian points roughly along the major axis of the bar in the effective potential ( fig .
[ effpot2d22 ] ) .
this is an alternative way of building ansae - type bars .
chaotic orbits with jacobi constants between 4/1 and corotation can build a spiral .
the dynamical mechanism , described in our model 2 , is based on the loops these orbits perform at their apocentra , as they move in a region defined by the isocontour of the effective potential that separates the regions inside and outside corotation .
another dynamical mechanism that can support a spiral is based on banana - like orbits .
these orbits are regular as well as chaotic ( model 4 ) . 5 . the recently discussed spirals , that are associated with the stable and unstable manifolds of the unstable lagrangian points l@xmath37 and l@xmath51 , along the major axis of the bar , do not play any important role in our models
we find these spirals in some cases , as the one of model 1 .
they emerge from the ends of the bar and fade out after completing an azimuthal distance of about @xmath39 . however , they are never in agreement with the ngc 1300 spirals . 6 . the morphology of the bars and spirals that are supported by chaotic orbits in our models are determined to a large extent by the shape of the isocontours of the effective potential ( figs .
[ effpot2d22 ] and [ effpot2d16 ] ) .
we thank prof . g. contopoulos for fruitful discussions .
we also thank an anonymous referee for many constructive comments .
p.a.p thanks eso for a two - months stay in garching as visitor , where part of this work has been completed .
all image processing was done with the eso - midas system .
this work has been partially supported by the research committee of the academy of athens through the project 200/739 .
athanassoula e. , romero - gomez m. , masdemond j.j . ,
2009 , mnras 394,67 contopoulos , g. , 1980 , a&a , 81 , 198 contopoulos , g. , 2004 , order and chaos in dynamical astronomy " , 2004 , springer - verlag , new york contopoulos , g. , 2009 , in chaos in astronomy " , contopoulos g. , patsis p.a .
( eds ) , p. 3
, springer - verlag berlin , heidelberg contopoulos g. & patsis p.a .
2006 , in astrophysical disks " , fridman a.m. , marov m.y .
, kovalenko i.g .
( eds . ) , springer , dordrecht , the netherlands , 2006 , p.131 contopoulos g. , patsis p.a .
( eds ) 2009 , chaos in astronomy `` , springer - verlag berlin , heidelberg england m. , 1989 , apj 344 , 669 grosbl p. , dottori h. , 2008 , a&a 490,87 harsoula m. , kalapotharakos c. , 2009 , mnras 394 , 1605 hnon m. , 1965 , anap , 28 , 499 kalapotharakos c. , patsis p. a. , grosbl p. , 2010
, mnras 403,83 ( pi ) kalapotharakos c. , patsis p. a. , grosbl p. , 2010b
, _ submitted _ ( pii ) kaufmann d.e .
, contopoulos g. , 1996 , a&a , 309 , 381 kaufmann d.e . , patsis p. a. , 2005 , apj , 624 , 693 laurikainen e. , salo h. , 2002 , mnras , 337 , 1118 lindblad p.a.b . , kristen h. 1996 , a&a 313 , 733 grosbl p. , patsis p.a .
, pompei e. , 2004 , a&a , 423 , 849 patsis p.a . , 2006 , mnras 369l , 56 patsis p. a. , athanassoula e. , quillen a. c. , 1997 , apj , 483 , 731 patsis p. a. , kalapotharakos c. , 2010 , in ' ' tumbling , twisting , and winding galaxies : pattern speeds along the hubble sequence " , e. m. corsini and v. p. debattista ( eds . ) , mem.s.a.it , in press ( arxiv:1002.1259 ) patsis p. a. , kaufmann , d. e. , gottesman s.t . , boonyasait , v. , 2009 , mnras , 394 , 142 romero - gomez m. , masdemont j.j .
, athanassoula e. , garcia - gomez c , 2006 , a&a 453 , 39 voglis n. , stavropoulos i. , kalapotharakos c. , 2006 , mnras 372 , 901 voglis n. , tsoutsis p. , efthymiopoulos c. 2006 , mnras 373 , 280 tsoutsis p. , efthymiopoulos c. , voglis n. , 2008 , mnras 387 , 1264 tsoutsis p. , kalapotharakos c. , efthymiopoulos c. , contopoulos g. , 2009 , a&a 495 , 743 | we present the orbital analysis of four response models , that succeed in reproducing morphological features of ngc 1300 .
two of them assume a planar ( 2d ) geometry with @xmath0=22 and 16 km s@xmath1 kpc@xmath1 respectively .
the two others assume a cylindrical ( thick ) disc and rotate with the same pattern speeds as the 2d models .
these response models reproduce most successfully main morphological features of ngc 1300 among a large number of models , as became evident in a previous study .
our main result is the discovery of three new dynamical mechanisms that can support structures in a barred - spiral grand design system .
these mechanisms are presented in characteristic cases , where these dynamical phenomena take place .
they refer firstly to the support of a strong bar , of ansae type , almost solely by chaotic orbits , then to the support of spirals by chaotic orbits that for a certain number of pattern revolutions follow an n:1 ( n=7,8 ) morphology , and finally to the support of spiral arms by a combination of orbits trapped around l@xmath2 and sticky chaotic orbits with the same jacobi constant .
we have encountered these dynamical phenomena in a large fraction of the cases we studied as we varied the parameters of our general models , without forcing in some way their appearance .
this suggests that they could be responsible for the observed morphologies of many barred - spiral galaxies .
comparing our response models among themselves we find that the ngc 1300 morphology is best described by a thick disc model for the bar region and a 2d disc model for the spirals , with both components rotating with the same pattern speed @xmath0=16 km s@xmath1 kpc@xmath1 .
in such a case , the whole structure is included inside the corotation of the system .
the bar is supported mainly by regular orbits , while the spirals are supported by chaotic orbits .
[ firstpage ] galaxies : kinematics and dynamics galaxies : spiral galaxies : structure |
the presence of chiral interactions , motivated by experimental systems @xcite , can result in extremely rich phase transition phenomena in otherwise simple systems @xcite . in this respect ,
we study here a @xmath0 state clock spin - glass model in @xmath1 spatial dimensions , using renormalization - group theory .
our system has both competing ferromagnetic and antiferromagnetic interactions , as in the usually studied spin - glass models @xcite , and competing left - chiral and right - chiral interactions @xcite .
the global phase diagram is calculated in temperature , antiferromagnetic bond concentration @xmath2 , random chirality strength , and right - chirality concentration @xmath3 .
we find an extremely rich phase diagram , with a ferromagnetic phase , a multitude of different chiral phases , a chiral spin - glass phase , and a critical ( algebraically ) ordered phase @xcite .
the ferromagnetic and chiral phases accumulate at the disordered phase boundary and form a devil s staircases @xcite , where different ordered phases characteristically intercede at all scales of phase - diagram space .
in fact , a continuum of devil s staircases is found .
shallow and deep reentrances of the disordered phase , bordered by fragments of regular and temperature - inverted devil s staircases , are seen .
the extremely rich phase diagrams are presented as continuously and qualitatively changing videos .
the * @xmath4state clock spin glass * is composed of unit spins that are confined to a plane and that can only point along @xmath5 angularly equidistant directions , with hamiltonian @xmath6 where @xmath7 , @xmath8 , at each site @xmath9 the spin angle @xmath10 takes on the values @xmath11 with @xmath12 , and @xmath13 denotes that the sum runs over all nearest - neighbor pairs of sites . as a ferromagnetic - antiferromagnetic spin - glass system @xcite , the bond strengths @xmath14 , with quenched ( frozen ) ferromagnetic - antiferromagnetic randomness , are @xmath15 ( ferromagnetic ) with probability @xmath16 and @xmath17 ( antiferromagnetic ) with probability @xmath2 , with @xmath18 .
thus , the ferromagnetic and antiferromagnetic interactions locally compete in frustration centers .
recent studies on ferromagnetic - antiferromagnetic clock spin glasses are in refs .
@xcite . in the * @xmath4state chiral clock double spin glass * introduced here
, frustration also occurs via randomly frozen left or right chirality @xcite .
the hamiltonian in eq .
( 1 ) is generalized to random local chirality , @xmath19,\ ] ] where , for each pair of nearest - neighbor sites @xmath13 with @xmath20 ahead of @xmath9 in the coordinate direction , frozen ( quenched ) @xmath21 ( left chirality ) or @xmath22 ( right chirality ) , and the delta function @xmath23 for @xmath24 .
the overall concentrations of left and right chirality are respectively @xmath25 and @xmath3 , with @xmath26 .
the strength of the random chiral interaction is @xmath27 , with temperature divided out . with no loss of generality ,
we take @xmath28 . thus , the system is chiral for @xmath29 , chiral - symmetric for @xmath30 , chiral - symmetry - broken for @xmath31 .
the global phase diagram is in terms of temperature @xmath32 , antiferromagnetic bond concentration @xmath2 , random chirality strength @xmath33 , and chiral symmetry - breaking concentration @xmath3 .
we solve the chiral clock double spin - glass model with @xmath0 states by renormalization - group theory , in @xmath1 spatial dimensions , with length rescaling factor @xmath34 .
our solution is , simultaneously , the migdal - kadanoff approximation @xcite for the cubic lattice and the exact solution @xcite for the @xmath1 hierarchical lattice based on the leftmost graph of fig .
exact calculations on hierarchical lattices are also currently widely used on a variety of statistical mechanics problems.@xcite . .
the corresponding hierarchical lattice is obtained by the repeated self - imbedding of the leftmost graph .
the recursion relations are exact for this @xmath35 hierarchical lattice . ] under the renormalization - group transformation described below , the hamiltonian of eq .
( 2 ) maps onto the more general form @xmath36 where @xmath8 can take @xmath5 different values , so that for each pair @xmath37 of nearest - neighbor sites , there are q different interaction constants @xmath38 which are in general different at each locality ( quenched randomness ) . here , @xmath39 is the angle between consecutive clock - spin directions .
the largest element of @xmath40 at each locality @xmath37 is set to zero , by subtracting the same constant @xmath41 from all @xmath5 interaction constants , with no effect on the physics ; thus , the @xmath42 other interaction constants are negative .
the local renormalization - group transformation is achieved by the sequence , shown in fig . 1 , of bond movings @xmath43 and decimations @xmath44 where @xmath45 and @xmath41 are the subtractive constants mentioned above , and prime marks the interaction of the renormalized system .
the starting double - bimodal quenched probability distribution of the interactions , characterized by @xmath2 and @xmath3 as described above , is not conserved under rescaling .
the renormalized quenched probability distribution of the interactions is obtained by the convolution @xcite @xmath46 where @xmath47 as in eq .
( 4 ) and @xmath48 represents the bond moving and bond decimation given in eqs . and .
similar previous studies , on other spin - glass systems , are in refs .
for numerical practicality the bond moving and decimation of eqs .
( 5 ) and ( 6 ) are achieved by a sequential pairwise combination of interactions , each pairwise combination leading to an intermediate probability distribution resulting from a pairwise convolution as in eq .
we effect this procedure numerically , by generating @xmath49 interactions that embody the quenched probability distribution resulting from each pairwise combination .
each of the generated @xmath49 interactions is described by @xmath5 interaction constants , as explained above ( eq .
( 4 ) ) . at each pairwise convolution as in eq . ( 7 ) , @xmath49 randomly chosen pairs are matched by ( 5 ) or ( 6 ) , and a new set of @xmath49 interactions is produced .
the different thermodynamic phases of the system are identified by the different asymptotic renormalization - group flows of the quenched probability distribution @xmath50 .
for all renormalization - group flows , originating inside the phases and on the phase boundaries , eq .
( 7 ) is iterated until asymptotic behavior is reached .
thus , we are able to calculate the global phase diagram of the chiral clock double spin - glass model .
the global phase diagram of the @xmath0 state chiral clock double spin - glass model in @xmath1 spatial dimensions , in temperature @xmath32 , antiferromagnetic bond concentration @xmath2 , random chirality strength @xmath33 , and right - chirality concentration @xmath3 , is a four - dimensional object , so that only the cross - sections of the global phase diagram are exhibited . for the interactions @xmath51 in which @xmath52 is maximum and
therefore 0 ( and the other four interactions are negative ) is shown in this figure , with @xmath53 .
the projections of @xmath54 onto two of its four arguments are shown in each panel of this figure .
the other four @xmath55 have the same fixed distribution .
thus , chirality is broken locally , but not globally . ] at a given location @xmath37 , under consecutive renormalization - group transformations , are shown .
the @xmath56 angular value of each interaction @xmath57 is indicated in the figure panels .
bottom panel : inside the spin - glass phase . the corresponding lyapunov exponent is @xmath58 and the average interaction diverges as @xmath59 , where @xmath60 is the number of renormalization - group iterations and @xmath61 is the runaway exponent .
top panel : at the phase boundary between the spin - glass and disordered phases .
the corresponding lyapunov exponent is @xmath62 and the average non - zero interaction remains fixed at @xmath63 . as indicated by the lyapunov exponents , chaos is stronger inside the spin - glass phase than at its phase boundary . ] figs .
2 show the calculated sequence of phase diagrams for the ferromagnetic @xmath64 , on the left side of the figure , and antiferromagnetic @xmath65 , on the right side , systems with quenched random left- and right - chiral interactions .
the horizontal axis @xmath3 is the concentration of right - chiral interactions .
phase diagrams for different random chirality strengths @xmath33 are shown .
the system exhibits ferromagnetic ( f ) , a multitude of different chiral , and spin - glass ( s ) ordered phases .
the antiferromagnetic system also shows an algebraically ( a ) ordered ( critical ) phase , in which every point is a critical point with divergent correlation length @xcite . in all cases ,
the ferromagnetic and different chiral phases accumulate as different devil s staircases @xcite at their boundary with the disordered ( d ) phase . figs .
3 show the calculated sequence of phase diagrams for the left - chiral @xmath66 , on the upper side , and quenched random left- and right - chiral @xmath67 , on the lower side , system with in both cases quenched random ferromagnetic and antiferromagnetic interactions .
the horizontal axis is the random chirality strength @xmath33 .
the consecutive phase diagrams are for different concentrations of antiferromagnetic interactions @xmath2 .
the system exhibits ferromagnetic ( f ) , a multitude of different chiral , spin - glass ( s ) , and algebraically ordered ( a ) phases .
the ferromagnetic and different chiral phases accumulate as different devil s staircases @xcite at their boundary with the disordered ( d ) phase .
note shallow and deep reentrances of disorder @xcite at @xmath68 and @xmath69 , respectively , surrounded by regular and temperature - inverted devil s staircases . the full richness of the continuum of widely varying devil s staircase phase diagrams
can best be seen in video form , four of which are accessible at http:// web.mit.edu/physics/berker/temperature_delta_c0_scanp.avi , web.mit.edu/physics/berker/temperature_delta_c05_scanp.avi , web.mit.edu/physics/berker/temperature_c_p1_scandelta.avi , web.mit.edu/physics/berker/temperature_c_p0_scandelta.avi these videos effectively exhibit a very large number of calculated phase diagram cross - sections .
the renormalization - group mechanism for the algebraically ordered ( critical ) phase is that , all renormalization - group trajectories originating inside this phase flow to a completely stable fixed point ( sink ) that occurs at finite temperature ( finite coupling strength).@xcite in all other ordered phases , the trajectories flow to strong ( infinite ) coupling . in the ferromagnetic phase ,
the interaction @xmath70 becomes asymptotically dominant . in the chiral phases , in the renormalization - group trajectories , one of the chiral interactions from the right - hand side of eq .
( 4 ) , @xmath71 , becomes asymptotically dominant .
however , in each of the separate phases , it takes a characteristic number @xmath60 of renormalization - group transformations , namely a length scale of @xmath72 , to reach the dominance of one chiral interaction .
therefore , these chiral phases are distinct .
after the dominance of one chiral interaction , the renormalization - group trajectory follows the periodic sequence @xmath73 resulting from matching @xmath0 and @xmath34 .
by contrast , the asymptotic flow in the spin - glass phase is to a strong - coupling fixed distribution @xmath50 . since
, at each locality , the largest interaction in @xmath74 is set to zero and the four other interactions are thus made negative , by subtracting the same constant from all five interactions without affecting the physics , the quenched probability distribution @xmath50 , a function of five variables , is actually composed of five functions @xmath55 of four variables , each such function corresponding to one of the interactions being zero and the other four , arguments of the function , being negative .
4 shows one of the latter functions : the part of the fixed distribution , @xmath54 , for the interactions @xmath51 in which @xmath52 is maximum and therefore 0 ( and the other four interactions are negative ) is shown in this figure .
the projections of @xmath54 onto two of its four arguments are shown in each panel of this figure .
the other four @xmath55 have the same fixed distribution .
thus , chirality is broken locally , but not globally . another distinctive mechanism , that of chaos under scale change @xcite or , equivalently , under spatial translation @xcite , occurs within the spin - glass phase and differently at the spin - glass phase boundary @xcite , in systems with competing ferromagnetic and antiferromagnetic interactions @xcite and , more recently , with competing left- and right - chiral interactions @xcite . fig .
5 gives the asymptotic chaotic renormalization - group trajectories of the spin - glass phase and of the phase boundary between the spin - glass and disordered phases .
the five interactions @xmath75 at a given location @xmath37 , under consecutive renormalization - group transformations , are shown .
chaos is measured by the lyapunov exponent @xcite , which we here generalize , by the matrix form , to our multi - interaction case : @xmath76 where the function @xmath77 gives the largest eigenvalue of its matrix argument @xmath78 and the vector @xmath79 is @xmath80 with @xmath53 , at step @xmath81 of the renormalization - group trajectory .
the product in eq .
( 8) is to be taken within the asymptotic chaotic band , which is renormalization - group stable or unstable for the spin - glass phase or its boundary , respectively .
thus , we throw out the first 100 renormalization - group iterations to eliminate the transient points outside of , but leading to the chaotic band . subsequently , typically using 1,000 renormalization - group iterations in the product in eq .
( 8) assures the convergence of the lyapunov exponent value @xmath82 .
spin - glass chaos occurs for @xmath83 @xcite and the more positive @xmath82 , the stronger is chaos , as seen for example in the progressions in figs . 6 and 7 of ref .
@xcite . in the spin - glass phase of the currently studied system , the lyapunov exponent is @xmath58 and the average interaction diverges as @xmath59 , where @xmath60 is the number of renormalization - group iterations and @xmath61 is the runaway exponent . at the phase boundary between the spin - glass and disordered phases ,
the lyapunov exponent is @xmath62 and the average non - zero interaction remains fixed at @xmath84 . as indicated by the lyapunov exponents , chaos is stronger inside the spin - glass phase than at its phase boundary .
it is thus seen that chirality and chiral quenched randomness provides , in a simple model , remarkably rich phase transition phenomena .
these include a multitude of chiral phases , continuum of widely varying devil s staircases , shallow and deep reentrances of the disordered phase surrounded by regular and temperature - inverted devil s staircases , a critical phase , and a chiral spin - glass phase with chaotic rescaling behavior inside and differently at its boundary .
the widely varying continuum of devil s staircase phase diagrams are best seen in video form , four of which are accessible at http:// web.mit.edu/physics/berker/temperature_delta_c0_scanp.avi , web.mit.edu/physics/berker/temperature_delta_c05_scanp.avi , web.mit.edu/physics/berker/temperature_c_p1_scandelta.avi , web.mit.edu/physics/berker/temperature_c_p0_scandelta.avi s. ostlund , phys . rev .
* 24 * , 398 ( 1981 ) .
m. kardar and a. n. berker , phys .
lett . * 48 * , 1552 ( 1982 ) .
d. a. huse and m. e. fisher , phys .
lett . * 49 * , 793 ( 1982 ) .
d. a. huse and m. e. fisher , phys . rev . * 29 * , 239 ( 1984 ) .
r. g. caflisch , a. n. berker , and m. kardar , phys .
b * 31 * , 4527 ( 1985 ) .
t. alar and a. n. berker , phys .
e * 94 * , 032121 ( 2016 ) .
h. nishimori , _ statistical physics of spin glasses and information processing _ ( oxford university press , oxford , 2001 ) .
a. n. berker and l. p. kadanoff , j. phys .
a * 13 * , l259 ( 1980 ) .
a. n. berker and l. p. kadanoff , j. phys .
a * 13 * , 3786 ( 1980 ) .
p. bak and r. bruinsma , phys .
49 * , 249 ( 1982 ) .
a. fukuda , y. takanishi , t. isozaki , k. ishikawa , and h. takezoe , j. mat .
chem . * 4 * , 997 ( 1994 ) .
a. n. berker and s. ostlund , j. phys .
c * 12 * , 4961 ( 1979 ) .
r. b. griffiths and m. kaufman , phys .
b * 26 * , 5022r ( 1982 ) .
m. kaufman and r. b. griffiths , phys .
b * 30 * , 244 ( 1984 ) . s. r. mckay and a. n. berker , phys . rev .
b * 29 * , 1315 ( 1984 ) .
m. hinczewski and a. n. berker , phys .
e * 73 * , 066126 ( 2006 ) .
m. kaufman and h. t. diep , phys .
e * 84 * , 051106 ( 2011 ) .
m. kotorowicz and y. kozitsky , cond .
matter phys .
* 14 * , 13801 ( 2011 ) .
j. barre , j. stat .
phys . * 146 * , 359 ( 2012 ) . c. monthus and t. garel , j. stat .
mech . - theory and experiment , p05002 ( 2012 ) .
z. z. zhang , y. b. sheng , z. y. hu , and g. r. chen , chaos * 22 * , 043129 ( 2012 ) .
chang and r. shrock , phys .
a * 377 * , 671 ( 2013 ) .
xu , l .- s .
wang , and x .- m . kong , phys .
a * 87 * , 012312 ( 2013 ) .
s. hwang , d .- s .
lee , and b. kahng , phys .
e * 87 * , 022816 ( 2013 ) .
r. f. s. andrade and h. j. herrmann , phys .
e * 87 * , 042113 ( 2013 ) .
r. f. s. andrade and h. j. herrmann , phys .
e * 88 * , 042122 ( 2013 ) . c. monthus and t. garel , j. stat .
- theory and experiment , p06007 ( 2013 ) . o. melchert and a. k. hartmann , eur . phys .
j. b * 86 * , 323 ( 2013 ) .
fortin , j. phys .- condensed matter * 25 * , 296004 ( 2013 ) .
y. h. wu , x. li , z. z. zhang , and z. h. rong , chaos solitons fractals * 56 * , 91 ( 2013 ) .
p. n. timonin , low temp . phys . * 40 * , 36 ( 2014 ) .
b. derrida and g. giacomin , j. stat .
* 154 * , 286 ( 2014 ) . m. f. thorpe and r. b. stinchcombe , philos .
royal soc . a - math .
sciences * 372 * , 20120038 ( 2014 ) .
a. efrat and m. schwartz , physica * 414 * , 137 ( 2014 ) . c. monthus and t. garel , phys .
b * 89 * , 184408 ( 2014 ) . t. nogawa and t. hasegawa , phys . rev .
e * 89 * , 042803 ( 2014 ) .
m. l. lyra , f. a. b. f. de moura , i. n. de oliveira , and m. serva , phys .
e * 89 * , 052133 ( 2014 ) . v. singh and s. boettcher , phys .
e * 90 * , 012117 ( 2014 ) .
xu , x. zhang , z .- q .
liu , k. xiang - mu , and r. ting - qi , eur .
j. b * 87 * , 132 ( 2014 ) .
y. hirose , a. oguchi , and y. fukumoto , j. phys .
japan * 83 * , 074716 ( 2014 ) .
v. s. t. silva , r. f. s. andrade , and s. r. salinas , phys .
e * 90 * , 052112 ( 2014 ) .
y. hotta , phys .
e * 90 * , 052821 ( 2014 ) .
s. boettcher , s. falkner , and r. portugal , phys .
a * 91 * 052330 ( 2015 ) .
s. boettcher and c. t. brunson , eur .
lett . * 110 * , 26005 ( 2015 ) .
y. hirose , a. ogushi , and y. fukumoto , j. phys .
japan * 84 * , 104705 ( 2015 ) .
s. boettcher and l. shanshan , j. phys .
a * 48 * , 415001 ( 2015 ) .
a. nandy and a. chakrabarti , phys
. lett . * 379 * , 43 ( 2015 ) .
m. j. p. gingras and e. s. srensen , phys . rev .
b. * 46 * , 3441 ( 1992 ) .
g. migliorini and a. n. berker , phys .
b. * 57 * , 426 ( 1998 ) .
m. j. p. gingras and e. s. srensen , phys .
b. * 57 * , 10264 ( 1998 ) .
m. hinczewski and a.n .
berker , phys . rev .
b * 72 * , 144402 ( 2005 ) . c. n. kaplan and a. n. berker , phys . rev .
100 * , 027204 ( 2008 ) . c. gven , a. n. berker , m. hinczewski , and h. nishimori , phys .
e * 77 * , 061110 ( 2008 ) .
m. ohzeki , h. nishimori , and a. n. berker , phys .
e * 77 * , 061116 ( 2008 ) .
v. o. zelik and a. n. berker , phys .
e * 78 * , 031104 ( 2008 ) . g. glpinar and a. n. berker , phys .
e * 79 * , 021110 ( 2009 )
kaplan , m. hinczewski , and a.n .
berker , phys .
e * 79 * , 061120 ( 2009 ) .
e. ilker and a. n. berker , phys .
e * 89 * , 042139 ( 2014 ) .
m. demirta , a. tuncer , and a. n. berker , phys .
e * 92 * , 022136 ( 2015 ) .
e. cladis , phys .
lett . * 35 * , 48 ( 1975 ) .
f. hardouin , a. m. levelut , m. f. achard , and g. sigaud , j. chim . phys . * 80 * , 53 ( 1983 ) .
j. o. indekeu , a. n. berker , c. chiang , and c. w. garland , phys . rev .
a * 35 * , 1371 ( 1987 ) .
r. r. netz and a. n. berker , phys .
* 68 * , 333 ( 1992 ) .
s. kumari and s. singh , phase transitions * 88 * , 1225 ( 2015 ) .
h. saleur , nucl .
b * 360 * , 219 ( 1991 ) .
j. l. jacobsen , j. salas , and a. d. sokal , j. stat .
119 * , 1153 ( 2005 ) .
j. l. jacobsen and h. saleur , nucl .
b * 743 * , 207 ( 2006 ) .
y. ikhlef , mod .
* 25 * , 291 ( 2011 ) .
j. l. jacobsen and c. r. scullard , j. phys .
a * 45 * , 494003 ( 2012 ) .
j. l. jacobsen and j. salas , nucl .
b * 875 * , 678 ( 2013 ) . c. r. scullard and j. l. jacobsen , j. phys .
a * 49 * , 125003 ( 2016 ) .
r. bondesan , s. caracciolo , and a. sportiello , arxiv:1608.02916 [ cond-mat.stat-mech ] ( 2016 ) s. r. mckay , a. n. berker , and s. kirkpatrick , phys .
lett . * 48 * , 767 ( 1982 ) .
s. r. mckay , a. n. berker , and s. kirkpatrick , j. appl .
* 53 * , 7974 ( 1982 ) .
a. n. berker and s. r. mckay , j. stat . phys . *
36 * , 787 ( 1984 ) .
a. j. bray and m. a. moore , phys .
* 58 * , 57 ( 1987 ) .
e. j. hartford and s. r. mckay , j. appl .
70 , 6068 ( 1991 ) .
m. nifle and h. j. hilhorst , phys .
* 68 * , 2992 ( 1992 ) .
m. nifle and h. j. hilhorst , physica a * 194 * , 462 ( 1993 ) .
m. cieplak , m. s. li , and j. r. banavar , phys .
b * 47 * , 5022 ( 1993 ) .
f. krzakala , europhys . lett . * 66 * , 847 ( 2004 ) .
f. krzakala and j. p. bouchaud , europhys . lett . * 72 * , 472 ( 2005 ) .
m. sasaki , k. hukushima , h. yoshino , and h. takayama , phys .
lett . * 95 * , 267203 ( 2005 ) .
j. lukic , e. marinari , o. c. martin , and s. sabatini , j. stat .
mech . : theory exp .
l10001 ( 2006 ) .
p. le doussal , phys .
lett . * 96 * , 235702 ( 2006 ) .
t. rizzo and h. yoshino , phys .
b * 73 * , 064416 ( 2006 ) .
h. g. katzgraber and f. krzakala , phys .
lett . * 98 * , 017201 ( 2007 ) . h. yoshino and t. rizzo , phys .
b * 77 * , 104429 ( 2008 ) .
j. h. pixley and a. p.
young , phys rev b * 78 * , 014419 ( 2008 ) .
t. aspelmeier , phys .
lett . * 100 * , 117205 ( 2008 ) .
t. aspelmeier , j. phys .
a * 41 * , 205005 ( 2008 ) .
t. mora and l. zdeborova , j. stat .
phys . * 131 * , 1121 ( 2008 ) .
n. aral and a. n. berker , phys .
b * 79 * , 014434 ( 2009 )
. q. h. chen , phys .
b * 80 * , 144420 ( 2009 ) .
t. jrg and f. krzakala , j. stat .
mech . : theory exp .
l01001 ( 2012 ) .
w. de lima , g. camelo - neto , and s. coutinho , phys .
lett . a * 377 * , 2851 ( 2013 ) .
w. wang , j. machta , and h. g. katzgraber , phys .
b * 92 * , 094410 ( 2015 ) . v. martin - mayor and i. hen , scientific repts . * 5 * , 15324 ( 2015 ) . z. zhu , a. j. ochoa , s. schnabel , f. hamze , and h. g. katzgraber , phys . rev .
a * 93 * , 012317 ( 2016 ) .
w. wang , j. machta , and h. g. katzgraber , phys .
b * 93 * , 224414 ( 2016 ) .
j. marshall , v. martin - mayor , and i. hen , phys .
a * 94 * , 012320 ( 2016 ) .
l. a. fernandez , e. marinari , v. martin - mayor , g. parisi , and d. yllanes , arxiv:1605.03025 [ cond-mat.dis-nn ] ( 2016 ) .
p. collet and j .- p .
eckmann , _ iterated maps on the interval as dynamical systems _ ( birkhuser , boston , 1980 ) .
r. c. hilborn , _ chaos and nonlinear dynamics _ , 2nd ed .
( oxford university press , new york , 2003 ) . | the chiral clock spin - glass model with @xmath0 states , with both competing ferromagnetic - antiferromagnetic and left - right chiral frustrations , is studied in @xmath1 spatial dimensions by renormalization - group theory .
the global phase diagram is calculated in temperature , antiferromagnetic bond concentration @xmath2 , random chirality strength , and right - chirality concentration @xmath3 .
the system has a ferromagnetic phase , a multitude of different chiral phases , a chiral spin - glass phase , and a critical ( algebraically ) ordered phase .
the ferromagnetic and chiral phases accumulate at the disordered phase boundary and form a spectrum of devil s staircases , where different ordered phases characteristically intercede at all scales of phase - diagram space .
shallow and deep reentrances of the disordered phase , bordered by fragments of regular and temperature - inverted devil s staircases , are seen .
the extremely rich phase diagrams are presented as continuously and qualitatively changing videos . |
for a clock to work , according to , it is necessary to have a ) different energy states present , and b ) coherence between these states .
this might be easier to understand if we first look at the more familiar situation for space instead of time . in order to have a highly accurate definition of a point in space we need , according to , a well localized wavepacket .
this contains a large spread of momentum states . indeed if we had only _
one _ momentum component the would be completely spread out everywhere and in no definite position at all .
so localization in position requires a spread of momentum states .
but this is not enough .
in addition these momentum states must be _
coherent_. if one momentum state alone is spread out everywhere , it is only the phases between the different momentum states that allows them to cancel everywhere but in the vicinity of one point . and
so if these states are added _ incoherently _ the system is still not localized . and
so is it also with energy eigenstates and time .
a single energy eigenstate by itself does not tic , it must `` beat '' or interfere with something else .
typically , with atoms or molecules , we have a second energy eigenstate nearby and we observe the oscillating electromagnetic field resulting from the beating of the two levels . or
we can suppose a @xmath1-meson clock @xcite or neutrino clock using the oscillations in flavor ( strangeness , electron - muon number ) resulting from the interference of different mass eigenstates .
if the two atomic or molecular levels are not coherent , the emitted photons will be `` chaotic '' and the electromagnetic field has no well - defined phase suitable for a clock ; if the two kaon or neutrino states @xcite are incoherent there will be no oscillations .
we thus see that in addition to the slowing down of time which is possible by relativistic effects , there is a further aspect to time as physical process : coherence and incoherence .
to deal with such questions of coherence and incoherence in it is not enough to think about the wavefunction of the system in question . we need a quantity which can describe something like an average over different wavefunctions .
the tool for this in is the density matrix .
the is easy to understand if we suppose having in our minds the complete , big of the whole world , or at least that part of it that might interact with what we want to study .
what we want to study will in general only be a subsystem of what is described by the big and involves only a few of the total number of variables .
let us take the simple case where this subsystem is a two - state system , with basis states @xmath2 .
( one state would be _ too _ simple . )
then the big can be written @xmath3 where the @xmath4 s are all the variables outside the subsystem , with @xmath5 a big number like @xmath6 , and the kets @xmath7 are the wavefunctions for these variables .
the `` 1 '' or `` 2 '' indicate that these in general depend on the state of the subsystem . now
if we only look at the subsystem and not at the @xmath4 s , that is are only concerned with operators that act on @xmath2 , then all we need to know is the `` reduced '' density matrix of the two - state system .
this is a 2 x 2 matrix which can be expressed @xcite as follows @xmath8 where the interpretation of this matrix can be given as : @xmath9 where of course the `` amounts '' are probabilities and are normalized to one : ( @xmath10 ) .
the justification for calling the off - diagonal elements the degree of coherence between the two states may be seen by considering two extreme cases : let @xmath11 and @xmath12 be either the same or orthogonal . in the first case
@xmath13 can be factored out in eq [ [ psi ] ] and @xmath14 is just a product wavefunction ; @xmath15and @xmath16 are completely coherent . in the second case
no operator working just on the @xmath2 subspace can cause the two states to interfere , they are completely incoherent .
if we think of the @xmath4 s as describing an environment in which our subsystem is immersed , we can say that to the degree the subsystem is correlated with the environment , the subsystem is internally incoherent .
finally , this environment could be a `` measuring apparatus '' .
this is something that establishes a correlation usually a strong one between the state of the subsystem and the @xmath17 , which now represent the `` apparatus '' . in the limit of very strong correlation
we can know what @xmath13 is just by looking at the @xmath17 , which is of course how a measurement instrument should function . on the other hand ,
we thus see how the measurement `` collapses the '' of the subsystem : @xmath15 and @xmath18 become incoherent @xcite when the @xmath17 states are very different .
an easy and limiting case of incoherence to understand is the thermal state .
the density matrix operator @xmath19 for a thermal state is proportional to @xmath20 where @xmath21 is the hamiltonian operator , @xmath22 the temperature and @xmath23 the boltzmann constant .
looking at this expression you might be inclined to see only the usual boltzmann factor of statistical physics , which it is .
however , more important for our discussion is the fact that it is an _
operator_. furthermore this operator is a function of the hamiltonian operator only , @xmath22 and @xmath23 being just numerical parameters .
this means that there is a definite basis in which @xmath19 is diagonal , namely the basis of energy eigenstates @xmath24 : @xmath25 this diagonal property has the physical interpretation that different energy eigenstates are _
incoherent_. that is , consider the expectation value of some operator : @xmath26\sim c_e \delta_{e , e'}{\cal o}_{e , e'}\sim c_e { \cal o}_{e , e } \;,\ ] ] in words , the contributions to @xmath27 for each energy state are to be added separately and incoherently .
one obvious consequence of this incoherence between energy states is that nothing can be time dependent : as we well know the thermal state is _
stationary_. that is , if we consider the time dependence of any operator @xmath28 in the thermal state we get @xmath29= tr\bigl[\rho{\cal o}\bigr]= constant \;,\ ] ] since we can use the fact that @xmath30 commutes with @xmath19 and the permutation property of the trace to bring the @xmath30 factors together and cancel them .
is the thermal state the only state like this ?
obviously not , for eq [ [ tra ] ] we see that the important point is eq [ [ del ] ] , the diagonality , and not the precise value of the @xmath31 .
the above arguments go through for _ any _ @xmath32 and not just eq [ boltz ] . for this reason we wrote a general coefficient @xmath31 and
not the boltzmann factor .
so if it s not the thermal state , what is the characterizing feature of the [ [ del ] ] ?
the property which will be true of any state describe by eq[[del ] ] is that it is _
stationary_. since there is no coherence between energy states no operator can show a time dependence .
conversely , if we know on physical grounds that some system is stationary , we know that its energy states are incoherent .
an important case of this is the radiation coming from a thermal source .
the radiation itself , say neutrinos from the sun , need not follow a boltzmann factor .
but since the sun , ( to a very high degree of approximation ) is stationary , we know that different energy states are incoherent . understanding this helped to clear up a long - standing confusion on the treatment of neutrino oscillations @xcite .
limiting cases like perfect coherence for a pure state , or incoherence of energy states for stationary systems , are easy to understand . however , we are often confronted with the more difficult situation of partial and time - dependent coherence .
typically we imagine the problem of a subsystem in a pure , perfectly coherent , state at @xmath33 ; which as time goes on becomes more incoherent as it interacts with the external world .
although this is in principle a perfectly well defined problem : just go ahead , find the total at any time and average out the unobserved variables by calculating eq [ [ rhomat ] ] there is no simple way to do this perfectly rigorously for a many - variable system .
we must resort to some kinds of plausible approximations or guesses .
a simple and reasonable picture , from which however we can learn a lot , amounts to assuming that the outer world or environment has steady properties with time and there is a loss of coherence at a constant rate for our subsystem .
formally this is like the `` golden rule '' calculation used in decay problems or scattering theory .
for the two state system as described by eq [ [ rhomat ] ] we can give a quite complete description of the phenomenology of the problem on this basis and draw a number of interesting conclusions .
to set up the framework for this approach , we first parameterize the 2x2 matrix for @xmath19 in terms of the familiar pauli matrices @xmath34 : @xmath35 which is appropriately normalized to @xmath36 .
the two - state subsystem is governed by an internal hamiltonian which completely describes its behavior in the absence of interaction with the environment .
it is also a 2 x 2 matrix which we can parameterize in the same way : @xmath37 where we have left out a possible constant term which simple represents an overall energy shift and has no interesting effects .
these are just mathematical definitions , but because of the familiar identification of the pauli matrices with spin , they allow a simple intuitive picture .
the 3-vector @xmath38 is like a `` polarization '' .
according to eq [ [ pmat ] ] its vertical or `` z '' component represents the relative amounts of the two states `` 1 '' or `` 2 '' .
the `` transverse '' or x - y components , according to eq [ [ pmat ] ] represent the degree of coherence within the subsystem .
in generalizing to more than two- state systems , this remains true ; the diagonal elements represent the amount of the quality " in question and the off - diagonal elements the coherence between them .
note that this is necessarily a basis - dependent picture : we must specify which `` quality '' we are talking about .
naturally if the system is isolated and has no interactions with the outside world the situation is very simple .
the evolution is governed by the above hamiltonian and the equation of motion for @xmath19 given by the basic formula @xmath39\ ] ] using the algebra of the @xmath40 this translates into @xmath41 pictorially , the `` polarization '' * p * precesses around a `` pseudo - magnetic field '' * v*. note that * v * can be time dependent if we wish .
so far we havent done very much , this is just equivalent to the evolution of a spin in a magnetic field .
observe that with this equation the length of * p * can not change : @xmath42 that is , * p * rotates without shrinking .
this represents the fact that a pure state remains a pure state for an isolated system .
the non - trivial part comes in when we try to handle the interaction with the outside world . as mentioned ,
if we make the assumption of an outside world with constant properties , interacting repeatedly with our subsystem such that there is a steady loss of coherence , that is a constant rate of decrease of the off - diagonal elements , we can obtain a plausible generalization @xcite of eq [ [ pdot ] ] .
this has the form @xmath43 this involves one new parameter , @xmath44 .
this is the damping or decoherence rate and describes the rate of loss of coherence between the two basis states .
@xmath45 means the `` transverse '' or @xmath46 components of * p*. if we mentally turn off the first term in the equation , the @xmath44 term leads to an exponential decay of the overlap or coherence terms in @xmath19 , as in the `` golden rule '' , and this rate is independent of the time when we start .
if our subsystem is immersed in a large essentially infinite system it is difficult to imagine any other behavior .
we have made one inessential simplification in writing eq [ [ pdota ] ] . in order to concentrate on the interesting interaction between the decoherence and the internal evolution
we have left out a possible term @xmath47 in which there would be a direct relaxation between the two basis states .
for example in the problem of two states of a chiral molecule separated by a barrier @xcite we are assuming that the temperature is low enough that there is no direct jumping over the barrier .
this is the simple classical relaxation process which can easily be dealt with but is not essentially relevant to the quantum mechanical aspects on which we wish to focus .
similarly , it will be seen that at long times * p * in eq [ [ pdota ] ] tends to zero .
this may not be exactly correct if the internal hamiltonian is such that there is a constant energy difference between the two basis states , large compared to an ambient temperature .
then @xmath48 should end up being given by the boltzmann factor and not be zero .
so we implicitly make the assumption that any constant component @xmath49 is small compared to the temperature .
the most important aspect of the presence of the damping term is that now the length of @xmath38 can shrink : @xmath50 in general , and so the density matrix can change its `` purity '' .
that is , even without explicitly `` quality changing '' relaxations the length of p will inevitably decrease , if there is a @xmath51 component .
such a component however , will be produced , in general , by the `` rotation '' induced by @xmath52 .
although we shall explain below how to arrive at a microscopic understanding of @xmath44 , even the simple phenomenological eq [ [ pdota ] ] allows us to arrive at some interesting conclusions .
one of these is the `` zeno effect '' .
the zeno or watched pot effect , perhaps first proposed by alan turing @xcite , goes something like this . according to the `` copenhagen school '' s treatment of the `` measurement problem '' ( all these things are in `` ... '' because i m not sure they exist ) , a measurement leaves the state of the system in one particular eigenstate of the measurement .
for example after a position measurement , the object is in some definite position not in a superposition of positions .
so we or turing might say , let s keep rapidly measuring the object s position .
it will repeatedly be fixed to the spot and not be able to move at all ! sometimes this `` paradox '' is used to say there s something wrong with or something of the sort .
but actually it represents perfectly reasonable behavior and in fact is just a solution to eq [ [ pdota ] ] , for the case of large @xmath44 . to see this ,
consider the behavior of eq [ [ pdota ] ] in the case of a constant @xmath52 , say along the x - axis .
if we start @xmath38 along the z - axis , it will tend to rotate in a circle in z - y plane : @xmath48the probabilities or `` amounts '' we have defined as represented by @xmath48 will oscillate . now if we turn on @xmath44 , the oscillations will be damped and gradually die out . as we make @xmath44 larger and larger the oscillations will become overdamped and disappear completely .
finally in the limit of very strong damping one finds @xcite , @xcite @xmath53 in the limit of large @xmath44 , @xmath48 is practically frozen and will hardly budge from the spot ! according to this formula the time scale , initially set by the oscillation time @xmath54 , gets stretched by a factor @xmath55 , which can be enormous when it turns out @xmath44 is something like the interaction rate in a gas ( see formula for @xmath44 below ) and @xmath54 is the tunneling time between two states of a chiral molecule . in this way we gave a second answer to hund s paradox on the stability of the optical isomers @xcite ( the first was parity violation @xcite ) . on a practical level
this means that in relaxation processes at low where the classical barrier - hopping type of mechanisms are frozen out and quantum tunneling takes over , we have `` anti - intuitive '' relaxation , where a bigger @xmath44 means _ slower _
relaxation @xcite .
you might say the repeated `` measurements '' by the environment have fixed the state a la zeno- watched pot but in any case it s just a result of a simple solution to eq [ [ pdota ] ] and has nothing mysterious about it .
at this point we should warn of a misunderstanding which sometimes arises .
the arguments , or similar arguments leading to eq [ [ zeno ] ] or its equivalent , depend very much on our dealing with a system with two , or in any case a finite number of , levels .
they do _ not _ apply to a continuous density of levels , as for the decay of a quasi - stable state into the continuum .
this misunderstanding has occasionally led some people to erroneously suggest that the decay of a particle or atomic or nuclear level could be inhibited by observing if it has decayed or not .
this is obviously silly .
mathematically the difference between the two cases is that with two or finite number of levels @xmath48 has a `` flat - top '' near @xmath33 in the absence of damping .
that is , @xmath56 for small times ; while for a true decay problem we have @xmath57 , a linear behavior .
this leads to qualitatively different behavior with respect to `` stopping and restarting '' the system as we do when we turn on the damping to get the zeno - watched pot behavior .
another nice way of undertanding this ( suggested to me by michael berry ) is to consider the decay process as the tunneling through a barrier , as in the gamow picture of alpha decay .
now , when there is tunneling through a barrier connecting two _ discrete _ states , as in the chiral molecule problem @xcite , the origin of the zeno - watched pot effect may be understood as follows .
tunneling between two discrete states is a very delicate process and depends very much on the quasi - degeneracy of the energy between the two states . with @xmath58 the energy splitting
, the tunneling probability goes as @xmath59 , with @xmath60 the tunneling energy . since @xmath60 is typically very small ,
so must @xmath58 be small to have a non - neglible tunneling probability .
now if some external influence is causing the energy levels to move around constantly , even by small amounts , say by some shifting of the two potential wells in the two - state tunneling problem , the degeneracy is constantly being lifted and the tunneling is slowed down . depending on the ratio of @xmath61
it can be practically stopped .
this leads to the zeno - watched pot effect .
but the situation is entirely different if we have a quasi - bound state decaying into a _ continuum _ of states , as in the alpha decay situation .
even if the energy of the quasi - bound state is fluctuating , it will always find some continuum state with which it is degenerate .
hence there is no zeno - watched pot effect .
we now come to the microscopic understanding of @xmath44 , the damping or decoherence rate . using our general ideas , we can derive a nice formula for this , which furnishes considerable insight into the `` decoherence '' or measurement process " . from its definition in eq [ [ pdota ] ] and
looking at eq [ [ pmat ] ] we see that @xmath44 is a rate parameter , one which gives the rate at which the overlap between @xmath12 and @xmath11 is decreasing with time .
we take our subsystem to be interacting with the big environment in a constant , steady manner
this is implicit in our assumption of a constant @xmath44 with no further time dependence .
we can model this situation by looking at the subsystem as being bombarded by a constant flux from the environment .
this may be thought of as particles , or excitations like phonons or quasiparticles .
our basic idea is that if the two states 1 and 2 of the subsystem scatter this incoming flux in a way which is different for 1 and 2 , then the overlap @xmath62 decreases .
we recall the concept of the @xmath63 matrix of scattering theory .
this is an operator which turns the incoming wavefunction into the outgoing wavefunction : @xmath64 now the important point here is that states 1 and 2 may scatter the incoming differently , so we have two @xmath63 matrices , @xmath65 and @xmath66 @xcite . if these are different the incoming particle or excitation , which was _
un_correlated with the state of the subsystem before the scattering , will be correlated with it afterwards .
so every scattering decreases the overlap and if we work out what this means for @xmath44 we get the following formula : @xmath67 the expression is proportional to the incoming flux since it is a rate parameter and @xmath68 refers to the incoming state , with a possible average implied if we have many kinds of incoming states .
`` i m '' means imaginary part .
there is also a significance to the real part , which is an energy shift of the subsystem induced by the environment @xcite .
since usually we have @xmath69 for the @xmath63 matrix , the formula says that the decoherence rate is related to the `` unitarity deficit '' produced by the fact that the different components of the subsystem do nt interact the same way with the outside world .
eq [ [ d ] ] has two illuminating limits : @xmath70 if both states of the subsystem interact equally with the outer world , @xmath71 , there is no decoherence .
this is in accord with our feelings about `` measurement '' .
if the `` outside '' does not respond to the differences in the subsystem there is no `` measurement '' and no `` reduction '' of the state .
however , we do not need to use this language , the equation stands by itself .
note an important point : interaction with the environment is possible without decoherence .
the other interesting limit occurs if one state , say 1 , does nt interact so @xmath72 then @xmath73 this result follows from an application of the optical theorem of scattering theory @xmath74 and @xmath75 .
this corresponds to the naive expectation that the decoherence rate is the scattering rate .
this is just one particular limit however , and there can be subtleties such as effects of phases on @xmath44 @xcite .
there ia a classical conundrum which says , in apparent contradiction to common sense , that the entropy of a isolated system can not increase .
this is particularly simple to show in where the definition of entropy is @xmath76 $ ] . since for an isolated system the evolution of @xmath19
is governed by a hamiltonian , we get for the time dependence of the entropy , just as in eq .
( [ tra ] ) , @xmath77= tr\bigl[\rho\ , ln \rho\bigr]= constant$ ] .
this argument says that as long as the time dependence of @xmath19 is governed by a hamiltonian , the entropy is constant .
this constancy would also apply to the entropy @xmath78 $ ] of our two - state system where @xmath19 is given by eq .
( [ rhomata ] ) if it were isolated not coupled to the environment . or
as we learn from the argument , it could even be coupled to the environment ; but in such a way that the evolution of @xmath19 is given by a hamiltonian .
however we see from eq .
( [ d ] ) that the coupling to the environmnet is not governed by a single hamiltonian but rather by two hamiltonians , giving the two @xmath63 matrices .
if we had one hamiltonian we would have @xmath79 , in which case @xmath71 , and there is no decoherence .
hence there is a connection between @xmath44 and the entropy increase of the two state system .
in fact diagonalizing eq .
( [ rhomata ] ) and taking the trace , we find for the entropy @xmath80=ln { 2}-{1\over 2}\bigl((1+p)ln(1+p)+(1-p)ln(1-p)\bigr)$ ] . taking the time derivative , we find for the entropy change of the two state system @xmath81 where we used eq .
( [ shr ] ) .
the @xmath82 refers to the limit of small @xmath83 .
it seems intuitively clear that the rate of entropy increase and the decoherence rate should be closely related and eq .
( [ ent ] ) expresses this quantitatively . by appropriate generalization of the various quantities ,
this could be extended to systems larger than just two states .
note @xmath84 is necessarily positive .
furthermore , in thermal equilibrium where there is no coherence between states of different energy i.e @xmath85 , there is no entropy production .
in recent years our subject has moved from the theoretical- philosophical to the domain of the almost practical with the realization of quantum behavior for large , essentially macroscopic devices``mesoscopic systems '' .
this has been given additional impetus in view of the possible use of such devices for the implementation of the `` quantum computer '' .
furthermore the subject is interesting in connection with the idea to my mind wrong that there might be some limit where large objects do nt obey the rules of quantum mechanics .
decoherence is of course the main question for the observability of such effects and their possible practical use .
one of the devices that has been studied in detail is the rf squid , where by suitable adjustment of parameters it is possible to bring the system into the configuration of the familiar double potential well separated by a tunneling barrier .
the `` x '' coordinate stands for the flux in the superconducting ring , and the system obeys in the absence of decoherence a schroedinger equation in this variable .
the states localized in one of the potential wells represent the supercurrent flowing in a given sense around the ring , with a very large number ( microamps ) of electrons reversing direction when we go from one well to the other .
creation of quantum linear combinations of these states , which can occur by tunneling , would certainly be impressive evidence for the general applicabilty of quantum mechanics .
some beautiful experiments @xcite @xcite using microwave technique have seen evidence for such combinations in squid systems .
we have suggested @xcite a particularly simple way to both see the quantum linear combination of the two states and to measure the decoherence time of the system , hopefully directly sensitive only to the true quantum decoherence .
it involves the idea of `` adiabatic inversion '' .
this occurs when a slowly varying external field can cause a quantum system to reverse its state , as when a spin `` follows '' a rotating magnetic field and goes from up to down .
this phenomenon is also often refered to as a `` level crossing '' .
it is an intrinsically quantum mechanical phenomenon and , important for us is hindered when the decoherence time is short compared to the time in which the inversion takes place .
we propose to produce such an inversion in the squid by sweeping an external field and then observing the reversal of the direction of the flux in the squid .
note that the system need not be `` observed '' until the procedure is over our method is `` noninvasive '' .
when the sweep is faster than the decoherence time the inversion should take place , and when it is slower it should be inhibited .
we are witnessing the transition from quantum ( tunneling allowed ) to classical ( tunneling forbidden ) behavior as the decoherence is increased .
going from fast to slow sweeps , the sweep time where the inversion begins to become unsuccesful thus gives a determination of the decoherence time .
a possible difficulty here is that the sweep can not be too fast , otherwise the procedure becomes non - adiabatic .
however , our estimates indicate that a region of squid parameters and temperature should exist where fast - enough sweeps are possible without violating adiabaticity .
in order to study these points in more detail we have developed numerical simulations of such systems , both for the study of the adiabatic inversion ( the logical not ) as well as for a two - squid system operating as a cnot quantum logic gate @xcite . in fig 1
we show the results of a simulation @xcite for the one - squid adiabatic inversion .
the decoherence time @xmath86 was chosen to be about 39 000 units and simulated as a random flux noise .
the squid parameters were @xmath87 , giving a time unit of @xmath88 and so @xmath89 .
this decoherence time would correspond to about @xmath90 in the estimate @xmath91 , with @xmath92 @xcite .
the simulation included the first 8 quantum levels of the squid so that possible effects of non- adiabaticity are taken into account .
the vertical axis in fig 1 shows the probability for finding the flux in the ring reversed after the sweep .
we see that while the inversion is essentially always successful for sweep times less than decoherence time , it becomes progressively less so for longer sweeps .
hence we are seeing the transition from quantum towards classical behavior , and a measuremnt of when this takes place furnishes a determination of the decoherence time . the gradual fall - off seen after
the completion of the longest sweep is indicative of another difficulty , relaxation .
our final states will in general not be distributed according to thermal equilibrium , and the final detection should take place quickly on the time scale for relaxation .
although the above ideas are rather simple and phenomenological , they can be applied to a wide variety of interesting problems , where of course the various quantities like @xmath52 and @xmath93 must be adapted to each particular case .
these range from understanding the permanence of optical or `` chiral '' isomers ( hund s paradox ) @xcite , to the study of neutrino oscillations @xcite , to the mesoscopic devices @xcite , @xcite , and the design of quantum logic gates @xcite just discussed . here
, however , i would like to conclude in a more speculative vein , coming back to the question of the need for coherence between energy states to define time .
is the loss of coherence always a more or less accidental happening , depending on the particulars of the situation ? or is there something universal about it ?
there is of course one kind of interaction which universally and inevitably couples to energy states and which in principle `` measures '' or distinguishes them : gravity @xcite .
gravity couples to mass / energy and so states of different mass / energy interact differently .
indeed one can derive the gravitational redshift by considering the @xmath1 meson `` clock '' @xcite . for this clock
, we have two mass / energy eigenstates and interference effects between them tell time . by the same token
this must imply some degree of decoherence between the different mass / energy states due to gravitational interactions .
there ought to be an imaginary or dispersive part to the redshift . naturally because the coupling , newton s constant , is so small we expect the effects to be negligible under everyday conditions .
it is nevertheless amusing to see what the formulas look like .
we shall incidentally find that there is in fact an `` everyday '' application , involving the passage of oscillating neutrinos near galaxies .
qualitatively , we expect in regions of large and rapidly varying gravitational fields that different energy states become incoherent . since gravity couples universally , this can not be dismissed as an incidental effect .
it will affect all energy states , and thus time is slowed down equally for all processes .
if the effects become strong , time in some sense becomes ill - defined or stands still .
we try to model the due to gravity by calculating @xmath44 for the @xmath1 clock .
there are two interfering mass eigenstates and an environment interacting gravitationally with these two states .
if we take this environment to be a flux of particles , we can use our formula eq [ [ d ] ] .
the calculation of the s matrices is ( non relativistically ) the same as for coulomb scattering , where the `` charge '' is the mass .
the two components of the clock have different masses @xmath94 and so @xmath93 are different and thus there is a contribution to @xmath44 . in an impact parameter representation we have where @xmath98 , with g the newton constant and @xmath99 the energy of the incoming particle . as is usual in coulomb problems
there is a divergence at large distances , and so we introduce a large distance cutoff @xmath100 which must be interpreted according to the physical conditions . taking the two @xmath63 s and the imaginary part we finally get with @xmath102 , which we have taken to be small and expanded to lowest order .
@xmath103 is @xmath104 .
we can try to estimate this under thermal conditions , as in the early universe , where all dimensional quantities are related to the temperature @xmath22 .
we end up with @xmath105 as would have been expected , t must be at least on the order of the planck scale for this to be significant .
and then the rate is @xmath106 , which is still small , unless we are considering @xmath103 also on the planck scale .
the result eq [ [ difin ] ] is of order @xmath107 since decoherence , as explained , is a unitarity effect and so comes in to second order in the coupling .
although conceptually interesting this seems very remote from any potentially observable effects .
nevertheless there is , as promised , a present - day " application .
we now know there are neutrino oscillations .
so possibly if the time scale for the release of the neutrinos in the big bang were short compared to their oscillation time ( then ) , they would still be oscillating in step today . instead of the @xmath1 clock
we have a neutrino clock . a beautiful clock if we could ever read it a precise timepiece going back to the very first minutes of the big bang .
but there s a difficulty : in traveling to us the neutrinos will pass near various mass and gravitational field inhomogeneities , like galaxies , and these will tend to mix up the phases . because of the huge mass of a galaxy , @xmath108 will no longer be small .
one finds @xcite , with typical mass parameters for neutrinos and galaxies , that essentially any neutrino traveling through or near a galaxy will be decohered .
unfortunately , there will be no oscillations to read .
so much for the big bang clock , but at least there s a `` practical '' application of theory involving gravity . note that these ideas , based on the different interaction of different mass / energy states , do not contradict the classical equivalence principle according to which different masses move in the same way .
the interaction is so cleverly constructed that while different masses follow the same path classically , they get different quantum phases @xcite , and these are essentially what concern us here .
the interplay between the classical and quantum is subtle and interesting .
the last has certainly not been said on these questions . on the treatment of neutrino oscillations in a thermal environment , , phys .
d 36(1987)2273 and chapter 9 of g.g .
raffelt , _ stars as laboratories for fundamental physics _
chicago press , 1996 ) .
two level systems in media and ` turing s paradox ' , r.a .
harris and , phys .
b 116(1982)464 .
quantum damping and its paradoxes , , in _ quantum coherence _
, j. s. anandan ed .
world scientific , singapore ( 1990 ) .
see ref 4 , and for a more formal derivation , focused however on neutrino applications : `` non - abelian boltzmann equation for mixing and decoherence '' g. raffelt , g. sigl and , phys rev let .
70(1993)2363 .
the sophisticated reader will recognized that although a very simple and intuitive picture arises through the use of the s matrix with distinct scattering events , one could reach a similar result by using a hamiltonian with a continuous time evolution and depending on a quantum number with values `` 1 '' and `` 2 '' .
study of macroscopic coherence and decoherence in the squid by adiabatic inversion , paolo silvestrini and , physics letters * a280 * 17 - 22 ( 2001);www.arxiv.org , cond - mat/0004472 .
also adiabatic inversion in the squid , macroscopic coherence and decoherence , paolo silvestrini and , _ macroscopic quantum coherence and quantum computing _ , pg.271 ,
d. averin , b. ruggiero and p. silvestrini , kluwer academic / plenum , new york ( 2001 ) www.arxiv.org , cond - mat/0010129 .
design of adiabatic logic for a quantum cnot gate , valentina corato , paolo silvestrini , , and jacek wosiek , www.arxiv.org , cond - mat/0205514 , physics letters * a309 * 206 ( 2003 ) .
adiabatic evolution of a coupled- qubit hamiltonian , v.corato , p. silvestrini , , and j. wosiek cond - mat/0310386 ; physical review * b 68 * , 224508 ( 2003 ) .
valentina corato , paolo silvestrini , , and jacek wosiek , to be published .
we use a c program developed from the work of j. wosiek by summer students at the max - planck - institute , a.t .
goerlich and p. korcyl of the jagellonian university . | we discuss the notion of quantum mechanical coherence in its connection with time evolution and stationarity .
the transition from coherence to decoherence is examined in terms of an equation for the time dependence of the density matrix .
it is explained how the decoherence rate parameter arising in this description is related to the `` unitarity defect '' in interactions with the environment as well as the growth in entropy of the system .
applications to the `` zeno - watched pot effect '' and gravitational interactions are given .
finally , some recent results on applications to macroscopic coherence with the rf squid , where the transition from quantum to classical behavior could be studied experimentally , are shown .
= -0.35 cm = 0.3 cm @xmath0 presented at _ time and matter _ , venice , august 2002 2.0pc there s no doubt that the concept of time is absolutely fundamental , be it for philosophers , physicists or the man and woman on the street .
kant even snubbed those haughty mathematicians and put time before number , making the intuition of time the origin of arithmetic @xcite .
peculiar to physicists however , and probably not agreed upon or well understood by these other demographic groups , is that with the 20th century lessons of relativity and behind them , physicists in their overwhelming majority will agree that time should be viewed as a physical process .
there seems to be no avoiding even in principle the need to understand how the clock works as a physical object .
we are familiar with the slowing down of time in special relativity , perhaps most easily understood as a transformation of fields due to velocity .
the various fields change in such a way , according to relativity , that any physical process including some `` clock'' goes slower .
similarly , gravitational fields can affect the clock , as in the red shift .
while these are interesting enough , there are further and perhaps less familiar aspects of time as a physical process resulting from quantum mechanics
. this will be the subject of our discussion . |
standard discussions of the mott - hubbard transition are generally concerned with lattices of equivalent sites . at zero temperature , the metal - insulator transition develops when the on - site electron - electron repulsion @xmath4 reaches some critical value @xmath5 which , usually , also corresponds to the onset of characteristic magnetic correlations . in this paper
we are concerned with the less common case where the system is ionic , encompassing `` anions '' and `` cations '' .
earlier workers , including nagaosa and takimoto@xcite , and egami , ishihara and tachiki @xcite considered the simplest two - site ionic generalization of the hubbard model ( henceforth dubbed ionic hubbard model , ihm ) , which exhibits the transition from a band insulator to a mott insulator .
such a system is , in the absence of electron - electron repulsions , @xmath4 = 0 , a regular band insulator .
as @xmath4 increases above some critical value @xmath5 , a band insulator - mott - hubbard insulator transition is expected to take place . when @xmath4 is sufficiently large , the inequivalence between anion and cation should in fact become irrelevant , and the ground state of a large @xmath4 system of equivalent sites is mott - insulating , with antiferromagnetic correlations . however , owing to the residual inequivalence of the two ionic sites , it will also exhibit other properties , which have only partly been explored so far .
of special interest is the anomalous behaviour of the polarization of the ionic solid across the transition .
a very interesting quantity in this regard is the born effective charge @xmath6 associated with an infinitesimal `` dimerizing '' displacement of the ionic lattice , corresponding to a frozen @xmath7 optical phonon .
resta and sorella @xcite studied the ihm in @xmath2 using lanczos diagonalization and found , by the berry phase method , very striking indications of a polarization anomaly at @xmath0 .
ortiz , ordejn , martin and chiappe @xcite proposed a simple hartree - fock explanation to the anomaly , namely that at a first order magnetic transition between an af insulator and a band magnetic insulator , the polarization may also abruptly jump .
a number of important questions are apparently still open at this stage , in particular 1 . what is the physical origin of the polarization anomaly at the band - mott transition , and what is its connection with symmetry ? 2 . what is the nature precisely at @xmath8 , of the `` threshold '' state between the band and the mott insulator ? 3 .
how does the polarization anomaly depend on the dimensionality ?
in particular , will it survive in 2d and 3d , and if so , with what strength ?
4 . bearing in mind that most experiments measure just only @xmath9 , can we identify a simple experimentally accessible quantity which could signal the @xmath6 anomaly in magnitude and sign ? in this paper
we set out to discuss principally these questions .
we will do it by studying more closely the same simple ionic hubbard model @xcite , in particular by analysing the symmetry of its possible ground states , by calculating effective charges and excitation spectra , and generally by seeking to understand its properties by exact diagonalizations supplemented when necessary by the simple hartree - fock approximation .
firstly , we find that the mott insulator does , for appropriate boundary conditions , possess odd symmetry under inversion , contrary to the ionic band insulator , which has even parity .
secondly , hartree - fock reveals that one clue to the polarization anomaly lies in an effective reversal of the on - site energy offset ( accompanied with the vanishing of the associated band gap ) , taking place _ for the minority spin species only _ at @xmath5 .
thirdly , in agreement with hartree - fock , exact results suggest that at @xmath5 not only the spin gap , but also the true charge gap vanishes , indicating a _ metallic _ `` threshold '' state poised precisely at the brink between the band insulator and the mott insulator .
fourthly , a fresh study of two different 2d lattices , namely square and honeycomb , shows that a direct band insulator - mott hubbard insulator transition and the associated polarization anomaly may survive in higher dimensions too , particularly in the 2d honeycomb lattice . here , the anomaly is weaker than in d=1 ( a jump instead of a divergence )
. finally , we propose the piezoelectric effect in a non - centrosymmetric lattice , here exemplified precisely by the 2d honeycomb lattice , as the experimentally accessible quantity that will directly and strikingly change sign at the band insulator - mott insulator transition .
the polarization calculations are carried out using the berry phase technique , first introduced by king - smith and vanderbilt @xcite and by resta @xcite , and further extended to a general many - body case by ortiz and martin @xcite , and applied to the ihm by resta and sorella .
the rest of this paper is structured as follows . in section [ sec : model ]
we introduce the ionic hubbard model and describe briefly the classical limit , @xmath10 , useful for understanding later the full quantum case . in section [ sec : magnetism ] we first present a discussion of the magnetic behavior expected by the model , followed by a more detailed discussion of symmetry , and by a finite - size study of the level crossings connected with the transition at @xmath5
. section [ sec : hf ] contains the hartree - fock theory of the band- mott insulator transition , and of the polarization anomaly .
section [ sec : manybody ] presents the full many - body calculation of the polarization , done by means of lanczos diagonalization plus berry phase , here specialized to the 2d honeycomb case , and finally in section [ sec : discussion ] we give a discussion of the possible detectability of the transition via the piezoelectric effect , followed by our general conclusions .
we consider the hubbard hamiltonian in 1d ( linear chain ) and 2d bipartite lattices ( square and honeycomb lattices ) .
all lattices being bipartite , they are composed of @xmath11 and @xmath12 sublattices . to simulate ionicity , @xmath11 and @xmath12 are made inequivalent by onsite energies @xmath13 and @xmath14 respectively . because of the energy difference @xmath1 between the @xmath11 and the @xmath12 sublattices , the electron population of the @xmath11 sublattice is less than that of the @xmath12 sublattice for @xmath15 ( or viceversa for @xmath16 ) .
the sublattices are connected by electron hopping .
we assume a filling of one electron per site with equal numbers of spin up and of spin down electrons .
sites of the @xmath11 sublattice are denoted by @xmath17 , those of the @xmath12 sublattice by @xmath18 , where the vectors @xmath19 connect an @xmath11 site with its neighbouring @xmath12 sites , @xmath20 . in the linear chain , where the length of the unit cell is 2 ,
@xmath21 = 0,1 , @xmath22 = @xmath23 . in the square lattice ( square side of unit length ) , @xmath21 = 0,1,2,3 and @xmath24 = @xmath25(1,0 ) , @xmath26 = @xmath25(0,1 ) . in the honeycomb lattice
( unit length is the side of the hexagon ) , @xmath21 = 0,1,2 and @xmath27 = @xmath28 , @xmath29 = @xmath30 , @xmath31 = @xmath32 .
+ the hamiltonian is @xmath33\ ] ] where we have used standard notation , and u is the hubbard onsite electron - electron repulsive interaction .
electrons hop with matrix elements @xmath34 between neighbouring sites from the @xmath11 to the @xmath12 sublattice along the @xmath35 directions .
we denote the ground state energy of our @xmath36-electron system as @xmath37 .
when in need to distinguish between the number of spin up and spin down electrons , the ground state energy will be denoted by @xmath38 . unless otherwise specified @xmath39 .
we shall study the charge gap and the spin gap of the system defined respectively as : @xmath40 and @xmath41 .
+ it is instructive at the outset to consider what happens when all @xmath42 . as sketched in fig([class ] ) if we set the hopping @xmath43 in the hamiltonian ( [ hubbard ] ) the model is classical .
it has a first order transition at @xmath44 with a change of the macroscopic polarization per unit cell @xmath45 .
both the charge and spin gaps are finite and coincide in the region @xmath46 where they are given by @xmath47 . for @xmath48 the charge gap remains finite @xmath49 while the spin gap vanishes .
note that precisely at @xmath44 , where the transition takes place , both the charge and the spin gap vanish .
this kind of transition s expected to persist for @xmath50 , where it takes a standard band insulator , with a charge and a spin gap , over to an antiferromagnetic insulator , with a charge gap , and gapless spin excitations for large @xmath4 .
in this section we discuss the polarization properties of an electron system close to an antiferromagnetic transition .
as discussed by ortiz and martin antiferromagnetism appears to play a crucial role and remarkably affects the behaviour of the polarization.@xcite let us consider a finite electron hopping @xmath51 . for @xmath52 ,
one electron per site , the model is in fact described by a completely filled , spin independent band , separated by a finite gap ( @xmath53 , for large @xmath1 ) from the second , empty band . in the opposite limit of large @xmath4 the standard strong coupling analysis of the hubbard model leads to a charge gap @xmath54 , and to the well known heisenberg hamiltonian @xcite @xmath55 for spins : @xmath56 with an antiferromagnetic superexchange coupling @xmath57 .
the mapping to the heisenberg model implies that for large @xmath4 the model has gapless spin excitations . in one dimension
they have been named `` spinons''@xcite , and can be derived from the exact bethe ansatz solution of the 1d heisenberg model.@xcite in higher dimensions , where antiferromagnetic long range order is believed to exist at @xmath58 in 2d @xcite and 3d@xcite , the gapless modes are spin waves .
a band insulator - mott insulator transition should therefore occur at some finite coupling @xmath5 .
of course , @xmath5 will differ from its classical value @xmath59 .
there will be in general the possibility of intermediate metallic phases covering a range of @xmath4 values .
even if the insulator - insulator transition is direct , quantum fluctuations may drive its character , for example from first order to second order , or else may split it into more than one transition @xcite . quantum antiferromagnets with long range order for @xmath60 ( or quasi long range order in one dimension ) can be described at the critical level , by the well known non linear sigma model , with action @xmath61 where @xmath62 is a unit vector describing the local orientation of the antiferromagnetic order parameter , and the coupling constant @xmath63 depends on the microscopic parameters @xmath64 of the model .
this effective model is well known to have a second order transition predicting that a spin gap opens up continuously for @xmath65 , alias @xmath66 @xcite .
we will for simplicity assume in the following that the transition is unique and is always second order even though a hartree - fock calculation @xcite in the one dimensional model and our own in the 2d square lattice indicates the opposite .
in fact , the hartree - fock approximation may fail anyway to describe correctly the order of the transition , as it is not appropriate to describe the gapless spin - wave excitations of the model in the ordered phase .
contrary to the linear chain and the square lattice , the hartree - fock method however correctly predicts ( see sec .
[ sec : hf ] ) a second order phase transition in the honeycomb lattice , at least for the parameter values studied here . the magnetic mott insulator for @xmath67 has a charge gap , and is therefore fully described by the nonlinear sigma model , at least in more than one dimensions .
this model does not present any further phase transition but the one at the critical coupling @xmath68 , which is thus related to @xmath5 .
it is possible however that the charge transition to a band insulator might occur for @xmath4 values different from @xmath5 , as suggested by fabrizio _
et al _ @xcite .
we shall return to this point below . in a finite system with -say- @xmath69 sites
a level crossing may occur for some particular boundary conditions , if allowed by symmetry .
in particular , in the one dimensional model , it can be easily proved that there exists a finite value @xmath70 at which the ground state undergoes a change in the eigenvalue of the inversion symmetry operator @xmath71 .
inversion symmetry around the site @xmath72 is defined by the following relations : @xmath73 where @xmath74 is the electron vacuum state , by definition invariant under inversion symmetry .
the additional relation ( [ inversion ] ) is necessary to completely define the inversion transformation @xmath75 in the whole hilbert space .
inversion does not interchange the @xmath11 and the @xmath12 sublattice and clearly commutes with the hamiltonian ( [ hubbard ] ) for any @xmath4 , provided the boundary conditions are real , namely periodic or antiperiodic ( see sec . [
sec : manybody ] ) . since @xmath76 , the identity , the inversion has obviously two eigenvalues , @xmath23 .
we will show in the following that the ground state @xmath77 satisfies @xmath78 for @xmath52 , whereas for large @xmath4 there is a change of sign and @xmath79 , so that _ a level crossing must occur at an intermediate coupling _ @xmath80 .
in the non - interacting system the ground state is a direct product of a spin up and a spin down slater determinants , both possessing the same orbitals of the lowest band .
both slater determinants have a definite parity @xmath81 and the inversion eigenvalue of the global wavefunction is given by their product @xmath82 .
hence the band insulating state is even under inversion . in the large-@xmath4 mott insulator instead
we can use the mapping to the heisenberg hamiltonian ( [ heis ] ) , whose ground state in arbitrary dimensions can be generally written as @xcite : @xmath83 where @xmath84 is the spin lowering operator at the site @xmath85 , @xmath86 is the ferromagnetic state along the spin - up direction , and the wavefunction @xmath87 is real .
@xmath87 is also subject to the well known ( `` marshall sign '' ) condition @xcite , i.e. the sign of the wavefunction is determined by the number of spin flips in the @xmath12 sublattice ( @xmath85 odd ) : @xmath88 the action of the inversion symmetry @xmath71 on the spin lowering operators can be easily found by applying the definition given in eqs.([inversion],[inversion1 ] ) , namely @xmath89 , and thus @xmath71 maps an element of the basis @xmath90 to another one @xmath91 where @xmath92 and @xmath93 and can be determined using the canonical anticommutation rules to restore the order of the creation operators @xmath94 after the application of the inversion operator @xmath71 to the ferromagnetic state @xmath95 . since inversion symmetry does not change the marshall sign we arrive to the conclusion that the inversion eigenvalue of the heisenberg wavefunction coincides with the inversion eigenvalue of the ferromagnetic state @xmath95 which is simple to compute . doing that we find that for @xmath96 , the inversion eigenvalue can change sign depending on the boundary conditions ( b.c . ) :
@xmath97 this finally proves our initial statement ; in particular a level crossing from an even state to an odd one _ has _ to occur in a periodic ring with @xmath98 or in an antiperiodic one with @xmath99 .
on the other hand , there will not necessarily be a level crossing in a periodic ring with @xmath99 and an antiperiodic one with @xmath98 . in summary , we conclude that the demise of the band insulator occurs via a symmetry change , whose finite - size signature is a parity switch from even to odd in the appropriate boundary conditions . as will be discussed in sec.([sec : manybody ] ) the change of polarization in a many - body system can be obtained using a form of averaging over the boundary conditions .
thus the averaging necessarily include both periodic ( pbc ) and antiperiodic ( apbc ) boundary conditions .
now , in one or in the other , depending on l , a level crossing will necessarily occur at some finite @xmath70 .
> from the theory of polarization a strong variation of the polarization can be expected as a function of @xmath4 around @xmath70 even in presence of a perturbations such as a dimerization @xmath100 ( see sec.([sec : manybody ] ) ) .
therefore , within the hypothesis that there exists only a well defined mott transition at a critical value @xmath5 in the thermodynamic limit , we may expect that : @xmath101 that is , the @xmath4 value where the level crossing occurs for large size is just the critical value of the magnetic transition .
understanding charge and spin excitation gaps is a crucial point .
we have studied these quantities in the d=1 case as a function of @xmath102 , performing calculations on finite rings with pbc or apbc . by considering the sequence of closed shells with one electron per site
, there is no level crossing and a finite size scaling analysis can be safely applied to the charge and spin gaps ( see the end of sec.[sec : model ] for their definitions ) . for a general finite size system ,
the lowest order correction to any gap should be of the form @xmath103 .
we have used a three parameter fit @xmath104 including also a higher order @xmath105 term , to improve the accuracy . in fig .
( [ spinchargecl ] ) we show the finite size calculations of the gaps for the 6 and the 12 site ring , as well as the result extrapolated to the thermodynamic limit with the finite size scaling ( [ fit ] ) . in fig .
( [ spinchargeop ] ) we also present the results for the spin and charge gaps for open shell rings of 6 and 12 sites .
the overall behavior of the gaps for the largest open shell ring ( l=12 ) is in agreement with the infinite size extrapolation of the closed shell rings , supporting the validity of our finite size scaling . starting with the band insulator at small @xmath4 , and increasing @xmath4 , we find that both charge and spin gaps are very close , and decrease together until they appear to close at some @xmath5 . for @xmath60 ,
the charge gap turns sharply upwards , while the spin gap does nt . precisely at @xmath106 ,
our fit suggests finite size gap corrections of the form @xmath107 , which implies not only spin , but also charge gapless excitations .
this being the case , the system at @xmath5 is metallic .
the nature of this metal is unknown and deserves further investigation.@xcite > from our calculations it is hard to say whether charge and spin gaps will vanish at exactly the same @xmath5 , or at two slighly different values , as very recently proposed in @xcite .
nonetheless it is suggestive that large finite size charge gaps and small spin gaps become slightly inverted after extrapolation , which goes precisely in the direction of a charge gap closing at a slightly smaller @xmath4 than the spin gap .
we defer all discussion of the possible two - transition scenario to the work of fabrizio _ et al .
_ @xcite and we will not further dwell on it in this paper , where we consider for simplicity a single @xmath5 . in the previous analysis of a level crossing in the model ( [ hubbard ] ) we did not explicitly use the exact bethe ansatz solution of 1d systems .
in fact the result that the inversion symmetry @xmath71 for large @xmath4 has the same eigenvalue of the corresponding ferromagnetic state @xmath95 remains valid also in d=2 , an so does the evenness of the band insulator at small @xmath4 .
unfortunately , unlike d=1 , the d=2 inversion symmetry , transforming @xmath108 leaves the ferromagnetic state invariant on a bipartite lattice .
thus , a level crossing can not be argued based on identically the same symmetry argument . as it turns out , however , it is again possible to generalize the argument by using a more elaborate symmetry operator , which changes eigenvalue in going from the @xmath52 state to large @xmath4 state . in the square lattice
this symmetry operator is easily identified .
it is the mirror symmetry across the diagonal of the square lattice @xmath109 with @xmath110 odd ( see picture [ figsquare ] ) . in the honeycomb lattice
one can also find a symmetry operation with the same property .
however , it is much more involved and we will not discuss it in detail here .
we only mention that this symmetry is obtained by a @xmath111 rotation around a site followed by an additional gauge transformation @xmath112 with suitable angles @xmath113.@xcite based on this analysis , we can therefore conclude that , upon averaging over the boundary conditions , there will be , both in the square and in the honeycomb lattice , a level crossing when the system is in the mott state , but none in the band insulator state .
therefore , we should expect a polarization anomaly , and a metallic state at the transition , also in these two dimensional cases .
however , before moving on to do numerical work and check these expectations in these more difficult problems , it is wise to solve them in simple mean - field which , as the d=1 case demonstrated,@xcite is always very instructive .
we shall consider the hubbard model on the bipartite honeycomb lattice , defined in sec .
( [ sec : model ] ) , and on the simpler square lattice . in the latter case , in order to remove the nesting degeneracy of the non interacting 2d fermi surface
, we have also studied the effect of the next - nearest neighbor hopping . in the hartree -fock approximation the ground state of the hamiltonian
is approximated by a slater determinant .
we may further assume that this slater determinant @xmath114 is factored in spin space @xmath115 with this choice it is simple to obtain the hartree - fock hamiltonian by linearizing the interaction term : @xmath116 in this way we obtain : @xmath117 where @xmath118 where we have employed a fourier transform in the two sublattices and correspondingly we have defined the complex function : @xmath119 in the square lattice case the presence of the next - nearest hopping implies a further term : @xmath120 with @xmath121 to be added to @xmath122 in eq.([hfham ] ) . for a uniform solution in both sublattices
the average spin densities are given by : @xmath123 the parameters @xmath124 and @xmath125 , defining the hartree -fock hamiltonians are given therefore by : @xmath126 the constant is obtained after little algebra @xmath127 where @xmath128 and @xmath129 are the total number of spin up and spin down particles ( @xmath130 ) . at one electron per site
all the bonding bands @xmath131 by minimizing the total energy we obtain two self consistent equations for the variational parameters @xmath132 : @xmath133 these equations can be easily solved by inserting a trial initial value for @xmath134 and @xmath135 in the rhs and iterating the result until selfconsistency is reached . for small u only a self consistent solution with @xmath136
is possible , with a small charge transfer from the electron - rich site @xmath12 to the electron - poor site @xmath11 .
the plot of the quantity @xmath137 , is shown in fig .
( [ hfmag ] ) . for large @xmath4
a broken symmetry solution with @xmath138 is possible for @xmath4 sufficiently large .
a finite value of the staggered magnetization @xmath139 is stable , and is given by : @xmath140 for large enough @xmath4 , @xmath1 can be neglected and we approach asymptotically the standard antiferromagnetic solution , where @xmath141 it is clear therefore that after the magnetic transition , which occurs first in hf , there is a slightly larger critical @xmath4 when one of the two bands becomes gapless with @xmath142 . a plot of the hartree - fock band gaps @xmath143 and @xmath144 is shown in fig .
( [ hfgap ] ) . the self consistent equations ( [ selfhf ] ) for the 2d square lattice case are not modified by the presence of @xmath145 at least in the insulating hf phases where the bonding bands @xmath146 are full and the antibonding bands @xmath147 are empty for both spin up and down electrons .
however for @xmath148 we found a finite @xmath4 interval between the band insulator at small @xmath4 and the mott - hubbard insulator at large @xmath4 where the bonding and the antibonding bands do overlap in a finite energy range , leading to a fully metallic behavior ( see fig.[hfgaptp ] ) . here
the above analysys should be slightly modified to take into account the gain in energy obtained by occupying some states of the antibonding band .
we will not study the square lattice tt model further , since the general three - phase sequence : band insulator extended metal phase
mott - hubbard insulator is not our main concern .
and although the hartree - fock approximation is by no means exact , it certainly represents a good starting point in dimension larger than one , so that we expect the tt model to have indeed an extended metal phase between the two insulators . according to hartree - fock ,
the two insulator phases are however still adjacent on the two sides of @xmath5 in the square lattice with first neighbor hopping ( t=0 ) , and in the honeycomb lattice , with any hopping .
in the former , the nesting property of the non interacting fermi surface affects dramatically the hf solution , making it effectively one dimensional , and leading to an unphysical first order transition.@xcite moreover , even beyond hartree - fock , the first - neighbor square lattice model is not generic enough .
an arbitrarily small next neighbor hopping will take it away to a tt , where the two insulators are most likely no longer adjacent .
we conclude that , at least at the hf level , the simplest model which generically possesses a transition between a mott and a band insulator in 2d is the honeycomb lattice . here
i ) the hf transition to a magnetic phase is second order ; ii ) there is no nesting of the non interacting fermi surface , which consists of just two k - points ; and iii ) the hf solution is always insulating ( band or mott ) as a function of @xmath4 with the excep[tion of a point where there is semimetallic behavior . for the above reasons , rather well justified at the hf level , in 2d we shall restrict our study to the honeycomb lattice case . in the honeycomb hf solution , there is a small difference between the two critical values of @xmath4 , one where magnetism sets in , and the next where a band gap closes , fig .
( [ hfgap ] ) .
bearing in mind the discussion of sec .
[ [ sec : chspgap ] ] , we can not judge whether this difference and its sign is just an artifact of the hf calculation or a real one . because of the crudeness of hartree fock , most likely the second .
the evolution of the full hartree - fock band structure for increasing @xmath4 ( not shown ) is also instructive .
the bands display a full gap both in the band insulator , and in the magnetic insulator .
the presence of a spin gap in the latter is clearly an artifact due to breaking of spin rotation invariance characteristic of hartree - fock . at the upper critical @xmath4 ,
where the charge gap closes , there is linear band dispersion , a point - like fermi surface , and a dirac massless spectrum centered at the 2d zone boundary .
therefore we could expect a semimetallic state to exist at the treshold between the two different insulators .
we argue that close to the point where the effective gap @xmath125 changes sign , the polarization will change dramatically as a function of @xmath4 .
to verify our statement , we make use of the theory of the polarization , in the geometric phase formulation @xcite , @xcite , @xcite , @xcite . in the hartree - fock case , this amounts to calculate the contribution to the polarization of the spin - polarized band which becomes gapless at the critical point , when one effective gap parameter changes sign . in order to better illustrate that , we consider a model system consisting of spinless noninteracting fermions , described by the hamiltonian ( [ hfham])with a density of 1/2 fermion per site . with periodic boundary conditions
, we can calculate analytically the change of polarization as we vary continuously the fermion energy gap @xmath1 , or alternatively as we introduce a small dimerizing distortion @xmath100 in the hopping . in the linear chain , without any distortion , @xmath149 , the variation of the polarization as a function of the energy difference @xmath1 between anions and cations is described by a step function : @xmath150 ( see ref.@xcite ) .
when we vary @xmath1 without changing its sign , the polarization does not change , while when the sign changes , the polarization jumps by @xmath151 as anticipated . we introduce a dimerizing distortion ( `` optical phonon - like '' ) @xmath100 which amounts to a change of the hopping @xmath152 for electrons hopping to the right and the left respectively of any @xmath12 site .
we can evaluate analytically the `` born effective charge '' , which is the derivative of the polarization with respect to @xmath100 at @xmath153 0 : @xmath154 where @xmath155 , @xmath156 .
@xmath157 and @xmath158 are complete elliptic functions of the first and second kind . in this model , we have thus recovered the divergent effective charge anomaly discovered numerically in the 1d many - body calculation of resta and sorella @xcite .
the same analysys holds for two dimensional or higher d - dimensional lattices provided that at the transition point @xmath159 the bonding and the antibonding bands touch at a @xmath160 dimensional ( fermi- ) surface , ( nesting ) . that is the case in particular for the @xmath3 square lattice but not for the honeycomb lattice . in the honeycomb lattice , using the geometric phase technique for the mentioned spinless fermion model , it is straightforward to verify that without any distortion the polarization does not change as we vary @xmath1 . introducing a distortion along the @xmath161 direction : @xmath162 @xmath163 @xmath164 , @xmath165 @xmath166 , one can then verify that the component of the change in polarization which is orthogonal to @xmath161 vanishes , at least in first order in @xmath100 . along @xmath161 , it is difficult to derive an analytic expression for the change in polarization .
we can however extract the singular part of the effective charge , in the neighbourhood of the zero of @xmath1 @xmath167 the effective charge has therefore a finite symmetric jump when @xmath1 changes sign .
since this is only the singular part , there will be in addition a smooth background contribution shifting this jump with respect to zero .
( in d=1 , the singularity was infinite , so there the shift was irrelevant ) . in conclusion
, if the hartree - fock approximate picture were correct , we should expect such a polarization jump to emerge in the full many - body calculation for the 2d honeycomb lattice of the next section .
as it will turn out , the jump is indeed verified , see fig .
( [ z ] ) .
moreover , the quantitative size of the jump is remarkably close to _ twice _ the value we have just obtained analytically .
this result may be understood by assuming that each of the two bands with opposite spins , are forced to close together their single particle gaps , due to the spin rotation invariance of the model .
this also suggests that the jump in the effective charge in 2d , similarly to the jump in the polarization in 1d , maybe due to topological reasons and therefore rather general and weakly dependent on the parameters defining the model .
we now proceed to do proper polarization or more precisely , effective charge calculation for a 2d interacting case .
we restrict our study of the hamiltonian ( [ hubbard ] ) to the honeycomb lattice ( a sort of 2d hexagonal bn ) , which is slightly more interesting because it does not possess inversion symmetry , and can have , _
e.g. , _ a piezoelectric effect ( see sec.[sec : discussion ] ) .
the sites of the @xmath11 sublattice are given by @xmath168 , with @xmath169 , @xmath170 is the lattice constant and @xmath171 , @xmath172 .
the lattice is space periodic under translations @xmath173 and @xmath174 .
the reciprocal lattice vectors are @xmath175 and @xmath176 .
the sites of the @xmath12 sublattice are given by : @xmath177 , @xmath178 where the three vectors @xmath19 which connect neighouring @xmath11 and @xmath12 sites are @xmath179 , @xmath180 and @xmath181 .
obviously @xmath182 .
we can also consider , for the purpose of mimicking a uniaxial stress along the @xmath161 direction : @xmath183 and @xmath184 .
the hamiltonian depends parametrically on @xmath100 , @xmath185 . in order to calculate the difference in the polarization of the system as the parameter
is varied between its values 0 and @xmath100 , we consider families of hamiltonians @xmath186 obtained from ( [ hubbard ] ) by introducing a complex hopping and substituting @xmath187 @xmath188 , where the parameter k is given by @xmath189 , with @xmath190 , @xmath191 .
this is equivalent to imposing generalised boundary conditions on the original hamiltonian @xcite : if @xmath192 is an eigenfunction of @xmath193 , then generalized boundary conditions imply @xmath194 , @xmath191 .
periodic boundary conditions correspond to @xmath195 , antiperiodic to @xmath196 .
the polarization difference between two states which are characterised by the initial and final value of the distortion , 0 and @xmath100 is given as an integral over the generalised boundary conditions : @xmath197 where @xmath198 is the unit cell volume , @xmath199 and @xmath200 take alternatively the values 1 or 2 , @xmath201 is the many - body generalisation of the geometric phase : @xmath202 and @xmath203 is the ground state of @xmath186 , which satisfies periodic boundary conditions . in the numerical calculation
, we adopt a cell with 8 lattice sites ( @xmath204 ) and 8 electrons .
the number of spin - up and spin - down electrons is 4 , and the dimension of the hilbert space is only 4900 .
the system is really very small , but we expect that averaging over the boundary conditions will reduce the finite size effects . in the calculation
we are interested to study the behaviour of @xmath205 , the polarization difference between the undistorted ( @xmath206 ) and the distorted ( @xmath207 ) case , for different values of the onsite interaction @xmath4 .
we use a discretized form to calculate the geometric phase numerically @xmath208 the geometric phase however is only defined modulo 2@xmath209 and an uncertainty in the result for the difference in polarization arises .
there will be an ambiguity in the value of the polarization difference up to a quantum .
however the polarization difference should remain unambiguous for neighboring values of @xmath100 .
we can thus fix the 2@xmath209 uncertainty in @xmath210 by requiring that @xmath210 is continuous in @xmath100 for a given value of @xmath4 .
we may also require that it is continuous for neighboring values of @xmath4 for fixed @xmath100 , so long as we do not cross a point of degeneracy in the electronic spectrum .
+ convergence in the integration with respect to @xmath211 has not been easy and we had to use grids with 150 or 300 points to get reliable results . at the end
we found that in the undistorted case , similarly to the one - dimensional case , there is a critical point @xmath212 , where the geometric phase changes discontinuously by @xmath209 for one @xmath211 in the 2d brillouin zone .
as in one dimension this effect maybe related to the presence of a level crossing as a function of @xmath102 for a particular boundary condition .
introducing next the dimerizing distortion , we found that the effective charge , as in the one - dimensional case , is positive in the region below @xmath212 and negative above . at @xmath212
it is discontinuous , with a finite jump , instead of the one - dimensional divergence . a plot of the effective charge as a function of @xmath102 is shown in fig .
( [ z ] ) .
the transition between an ionic band insulator and a mott - hubbard insulator as exhibited by the prototype model ( eq.[hubbard ] ) is quite interesting .
we have extended existing studies in d= 1 , probing deeper into the transition .
we have also carried out newer investigations in d=2 , where same kind of transition is shown to take place , with some differences between the square and the honeycomb lattice . with reference to the list of questions presented in the introduction
, we can now formulate the following answers .
the physical origin of the polarization anomaly is connected with the symmetry switch between an even state , typical of the band insulator , to an odd state , typical of the mott insulator .
the symmetry switch , which we@xcite and others@xcite had noted earlier for finite size , is established here for arbitrary size .
moreover , in d=1 , even and odd refer to simple inversion symmetry , in d=2 the pertinent symmetry operation is different and depends on the lattice .
the threshold state between the ionic and the mott insulator , where the polarization abruptly changes sign , is one where the charge gap seems to close , and is therefore metallic .
this result underlines the profound difference between the two types of insulator , and invites further studies , which are now beginning to appear@xcite of this transition .
\3 . in d=2
the same model has again a band - to - mott insulator transition .
the polarization anomaly is generally weaker , to a degree which depends on the lattice . in one dimension
the effective charge diverges at the transition undergoing an abrupt sign change from @xmath213 to @xmath214 . in two dimensions this sign change persists .
we have shown that in the 2d honeycomb lattice the divergence turns to a jump , again with a change of sign . here
the magnitude of this jump is found to coincide almost exactly with @xmath215 , suggesting that this jump could be an experimentally detectable quantity dependent only on the bulk lattice constant and no other details of the actual material .
the total piezoelectric coefficient @xmath216 might be used to detect the polarization anomaly , because it is sensitive to the sign of @xmath6 .
it is conventionally divided into two contributions : the first , @xmath217 , is purely electronic , and is related to a uniform stress of the solid ( a simple scaling of all distances in the unit cell ) ; the second , more important contribution , is obtained by keeping the unit cell fixed and by changing the distance between atoms . in lattices without inversion symmetry , the piezoelectric coefficient @xmath216 is directly proportional , in magnitude _ and sign _ , to the effective charge @xmath6 , in the form ( we omit here for simplicity tensorial and vectorial indices ) @xcite @xmath218 where the constant @xmath219 represents the internal strain parameter .
an anomaly in @xmath6 will clearly reflect in the piezoelectric coefficient .
piezoelectric measurements could therefore permit the detection of a band - to - mott insulator transition , provided the system does not possess inversion .
the honeycomb lattice which we have studied in d=2 is the simplest one that does not possess inversion symmetry .
our study shows that the transition is not suppressed by piezoelectric strain .
we have not been able to identify , at present , a likely compound where this kind of transition could be experimentally detected .
it would seem possible , nevertheless , that suitable systems could be engineered , especially in the organic world .
we are thankful to m. fabrizio , s. de gironcoli , a. parola and r. resta for useful discussions .
n.g . acknowledges financial support through an eec fellowship under the hcm programme contract no .
work at sissa was co - sponsored by infm through pra htsc , and by the european union through contract erbchrxct940438 .
part of the calculations were performed during n.g.s visit to cineca , under the icarus2 project , contract no .
stefano cozzini is acknowledged for his help with the parallelization of the codes .
t. egami , s. ishihara and m. tachiki , science * 261 * , 1307 ( 1993 ) .
n. nagaosa and j.takimoto j.phys.soc.jpn.*55(8 ) * , 2735 ( 1986 ) ; ibidem 2745 .
r. resta and s. sorella , 4738 ( 1995 ) .
g. ortiz , p. ordejn , r.m .
martin and g. chiappe , 13515 ( 1996 ) .
king - smith and d. vanderbilt , , 1651 ( 1993 ) .
r. resta , ferroelectrics * 136 * , 51 ( 1992 ) ; ibidem rev .
66 * , 889 ( 1994 ) .
g. ortiz and r. m. martin , 14202 ( 1994 ) .
g. baskaran , z. zou , and p.w .
anderson , solid state commun .
, * 63 * , 973 ( 1987 ) .
see e.g. l. faddeev , l. takhtajan , phys .
lett . * a 85 * , 375 ( 1981 ) and references therein .
reger and a.p .
young , , 5978 ( 1988 ) .
k. j. runge , 7229 ( 1992 ) ; ibidem 12292 .
s. chakravarty , b. halperin and d. nelson , 1057 ( 1988 ) ; , 2344 ( 1989 ) .
e. lieb and d. mattis , j. math . phys . * 3 * , 749 ( 1962 ) .
a. parola , private communication .
d. vanderbilt and r.d .
king - smith , , 4442 ( 1993 ) .
c. gros , 6865 ( 1996 ) , ibidem z. phys .
b * 86 * , 359 ( 1992 ) .
a finite size gap minimum , which we believe is compatible with charge gap closing in the thermodynamic limit , was obtained in a 64-site dmrg calculation by k. schnhammer , o. gunnarson and r.m .
noack , , 2504 ( 1995 ) .
we are grateful to prof .
schnhammer for pointing out this reference to us .
m. fabrizio a. o. gogolin and a. a. nersesyan unpublished .
r. m. martin , 1607 ( 1972 ) . s. de gironcoli , s. baroni and r. resta , 2853 ( 1989 ) . | we study the quantum phase transition between a band ( `` ionic '' ) insulator and a mott - hubbard insulator , realized at a critical value @xmath0 in a bipartite hubbard model with two inequivalent sites , whose on - site energies differ by an offset @xmath1 .
the study is carried out both in @xmath2 and @xmath3 ( square and honeycomb lattices ) , using exact lanczos diagonalization , finite - size scaling , and berry s phase calculations of the polarization .
the born effective charge jump from positive infinity to negative infinity previously discovered in @xmath2 by resta and sorella is confirmed to be directly connected with the transition from the band insulator to the mott insulating state , in agreement with recent work of ortiz _
et al._. in addition , symmetry is analysed , and the transition is found to be associated with a reversal of inversion symmetry in the ground state , of magnetic origin .
we also study the @xmath2 excitation spectrum by lanczos diagonalization and finite - size scaling .
not only the spin gap closes at the transition , consistent with the magnetic nature of the mott state , but also the _ charge _ gap closes , so that the intermediate state between the two insulators appears to be metallic .
this finding , rationalized within unrestricted hartree - fock as due to a sign change of the effective on - site energy offset @xmath1 for the minority spin electrons , underlines the profound difference between the two insulators .
the band - to - mott insulator transition is also studied and found in the same model in @xmath3 . there
too we find an associated , although weaker , polarization anomaly , with some differences between square and honeycomb lattices .
the honeycomb lattice , which does not possess an inversion symmetry , is used to demonstrate the possibility of an inverted piezoelectric effect in this kind of ionic mott insulator .
-0.75 truecm |
conformal field theory ( cft ) has attracted considerable interest over the years in problems ranging from high - energy physics to statistical and condensed matter physics @xcite . in particular
, it provides a full understanding of the physical properties of the emerging quantum criticality of one - dimensional ( 1d ) quantum problems .
the low - energy relativistic spectrum of a 1d lattice model with a continuous symmetry is described in terms of representations of a certain current algebra @xcite .
this affine symmetry determines the operator content of the theory and all possible scaling dimensions of the operators are then fixed by the conformal invariance of the underlying wess - zumino - novikov - witten ( wznw ) model @xcite .
the cft approach allows the computation of correlation functions , the finite - size spectrum as well as the entanglement spectrum of the underlying lattice model @xcite .
the different critical phases are identified by the central charge @xmath6 of the wznw cft which fixes the low - temperature behavior of the specific heat and the scaling of the entanglement entropy @xcite . on top of the full description of the properties of the fixed point , the cft approach also gives access to the natural basis to investigate the effect of perturbations around it @xcite . as a result of a strongly relevant perturbation ,
the conformal symmetry might be lost and a mass gap is generated by the interaction .
a second possible scenario is a massless renormalization group ( rg ) flow where in the far infrared ( ir ) limit the conformal symmetry is restored with a smaller central charge @xcite .
a well - known example of the latter phenomenon is the massless rg flow between consecutive virasoro minimal models @xmath7 and @xmath8 perturbed by a negative @xmath9 relevant perturbation @xcite .
these rg flows might be studied by means of the power of integrability methods in case of integrable perturbations @xcite . in absence of integrability , numerical approaches , as the truncated conformal space approach @xcite or his improvement @xcite are efficient methods to fully determine the ir properties of a perturbed cft . in this paper
, we investigate the possible occurrence of a massless rg flow for the su(@xmath0)@xmath1 wznw cft with central charge @xmath10 perturbed by its adjoint primary field @xmath11 with scaling dimension @xmath12 .
the hamiltonian density of the resulting model is defined as follows : @xmath13 where @xmath14 are the chiral currents which satisfy the su(@xmath0)@xmath1 current algebra : @xmath15 with a similar definition for the right current and @xmath16 are the structure constants of the su(@xmath0 ) group . in eq .
( [ model ] ) , @xmath17 denotes the normal ordering of operator @xmath18 and a summation over repeated indices is assumed throughout the paper .
the adjoint primary field can be expressed in terms of the su(@xmath0)@xmath1 wznw field @xmath19 with scaling dimension @xmath20 @xcite : @xmath21 @xmath22 being the su(@xmath0 ) generators transforming in the fundamental representation normalized according to @xmath23 . the main physical motivation to study
the ir properties of the hamiltonian ( [ model ] ) stems from its application to 1d heisenberg spin chain models .
when @xmath24 , it is well - known that model ( [ model ] ) with @xmath25 accounts for the ir properties of spin-@xmath26 heisenberg chain @xcite .
a massless flow with emerging su(2)@xmath3 quantum criticality is obtained for @xmath27 and odd @xmath2 whereas a massive behavior occurs in other cases @xcite .
this result is directly related to the famous haldane s conjecture that integer heisenberg spin chain has a spectral gap while in the half - integer case a massless behavior in the su(2)@xmath3 universality class is stabilized @xcite . for general @xmath0 , it has been recently proposed that model ( [ model ] ) with @xmath5 governs the quantum phase transition between dimerized and 1d su(@xmath0 ) symmetry - protected topological phases @xcite . in this paper
, it will be shown that the su(@xmath0)@xmath1 perturbed cft ( [ model ] ) describes the low - energy limit of weakly coupled su(@xmath0 ) spin ladder and su(@xmath0 ) spin chain models with symmetric rank-@xmath2 tensor representation .
we will investigate the ir properties of model ( [ model ] ) and try to extend the known @xmath24 results to general @xmath0 .
we will see that the field theory ( [ model ] ) is massive for all @xmath0 and @xmath2 when @xmath28 .
in contrast , when @xmath29 , using a mapping onto @xmath2-leg su(@xmath0 ) spin ladder , a massless rg flow to su(@xmath0)@xmath3 cft is expected when @xmath0 and @xmath2 have no common divisor . in the simplest @xmath5 and @xmath4 case we perform a direct approach by means of gepner s parafermions ( gp ) @xcite . in this respect , we conclude that the su(3)@xmath30 cft perturbed by its adjoint primary field enjoys a massless rg flow down to the su(3)@xmath3 universality class . the rest of the paper is organized as follows . in sec .
ii , we present the mapping of model ( [ model ] ) onto @xmath2-leg su(@xmath0 ) spin ladder . using known results in su(@xmath0 ) spin chains ,
we deduce our main conclusion on the ir properties of the field theory ( [ model ] ) . in sec .
iii , a direct approach is developed in the simplest @xmath5 and @xmath4 case . finally , our concluding remarks are given in sec .
iv and the paper is supplied with two appendices which provide additional information .
in this section , we show that model ( [ model ] ) describes the low - energy physics of weakly coupled @xmath2-leg su(@xmath0 ) spin ladder with lattice hamiltonian : @xmath31 where @xmath32 denote the su(@xmath0 ) spin operators , which transform in the fundamental representation of su(@xmath0 ) ( represented by the young diagram @xmath33 ) , on the ith site and the @xmath34 leg of the spin ladder .
we assume that the ladder has open transverse boundary conditions , i.e. , the outer spins @xmath35 do not interact with @xmath36 ones ( see fig .
the spin ladder ( [ 2leg ] ) can , in principle , be manufactured in the context of ultracold alkaline - earth or ytterbium atoms where the existence of an high su(@xmath0 ) symmetry has been recently demonstrated experimentally @xcite . in absence of the interchain coupling ( @xmath37 ) ,
the model is @xmath2 decoupled sutherland models @xcite .
the latter is integrable and displays a quantum critical behavior in the su(@xmath0)@xmath3 universality class with central charge @xmath38 @xcite . in the weak - coupling limit @xmath39
, one can then perform a low - energy approach to deduce the physical properties of the spin ladder ( [ 2leg ] ) . in this respect
, we will extend the results of ref . to the general @xmath40 case .
-leg su(@xmath0 ) spin ladder . ] the low - energy properties of the sutherland model can be obtained by starting from the u(@xmath0 ) hubbard model at @xmath41 filling with large repulsive @xmath42 interaction @xcite . at low - energy below the charge gap , the su(@xmath0 ) operators in the continuum limit are described by @xcite : @xmath43 where @xmath44 ( @xmath45 being the lattice spacing ) and the fermi momentum is @xmath46 since the underlying hubbard model is @xmath41 filled ( one fermion per site ) . in eq .
( [ spinop ] ) , @xmath47 are the left and right su(@xmath0)@xmath3 currents , the @xmath48 and @xmath49 parts are related to the su(@xmath0)@xmath3 primary fields which transform respectively in the fundamental representation ( @xmath33 ) and the antisymmetric one ( @xmath50 ) with dimension @xmath51 of the su(@xmath0 ) group .
the su(@xmath0)@xmath3 currents can be expressed in terms of the underlying left ( right)-moving dirac fermions @xcite : @xmath52 where , here , there is no sum on @xmath53 and we have @xmath54 . the @xmath48 term of eq .
( [ spinop ] ) reads as follows in terms of the dirac fermions : @xmath55 where @xmath56 denotes an average over the charge degrees of freedom which are fully gapped in the large @xmath42 limit , and @xmath57 can be chosen real for a matter of convenience .
the field @xmath58 is the su(@xmath0)@xmath3 wznw primary field which transforms in the fundamental representation of su(@xmath0 ) . in the non - abelian bosonization approach @xcite ,
it is described ( see eq .
( [ 2kf ] ) ) by the identity : @xmath59 and has scaling dimension @xmath60 . with these basic facts at hands
, the continuum description of the decoupled su(@xmath0 ) spin ladder is given by the hamiltonian density : @xmath61 - \gamma j^a_{l r } j^a_{l l } , \label{contfreeham}\ ] ] where @xmath62 is the spin velocity and @xmath63 so that the perturbation is a marginal irrelevant current - current term .
we will discard this perturbation in the following which gives logarithmic corrections @xcite . in the weak - coupling limit
, one can use the low - energy description ( [ spinop ] ) to get the continuum limit of the @xmath2-leg su(@xmath0 ) spin ladder .
the leading contribution stems from the @xmath48 part of the spin operator ( [ spinop ] ) and we find : @xmath64 + j_{\perp } \lambda^2 \sum_{l=1}^{k-1 } \left({\rm tr } ( g_l t^a ) \ ; { \rm tr } ( g^{\dagger}_{l+1 } t^a ) + h.c.\right ) , \label{contham}\ ] ] where the perturbation is strongly relevant with scaling dimension @xmath65 .
the leading interaction in the weak - coupling regime takes thus the form of @xmath2 coupled su(@xmath0)@xmath3 wznw models . using the su(@xmath0 ) identity : @xmath66 it is then useful to rewrite the interacting part of model ( [ contham ] ) in the following form : @xmath67 with @xmath68 and @xmath69 .
the two strongly relevant perturbations in eq .
( [ vint ] ) are of very different nature .
indeed , one observes that @xmath70 is invariant under an su(@xmath0)@xmath71 @xmath72 su(@xmath0)@xmath73 symmetry : @xmath74 , @xmath75 being two independent su(@xmath0 ) matrices . in stark contrast , @xmath76 is only su(@xmath0 ) invariant : @xmath77 , with @xmath42 belonging to su(@xmath0 ) . in close parallel to the @xmath24 case @xcite , one way to separate the different degrees of freedom of the problem is to consider the following cft embedding , built from the product of @xmath2 su(@xmath0)@xmath3 cfts : @xmath78 where @xmath79 is a coset cft with central charge @xmath80 . in the @xmath5 case
, @xmath81 corresponds to the @xmath82 parafermionic cft @xcite . for general @xmath2
, @xmath83 is the sum of @xmath84 consecutive coset models cfts @xmath85 which enjoy an extended @xmath86 symmetry @xcite .
the su(@xmath0)@xmath1 cft is generated by su(@xmath0)@xmath1 chiral currents @xmath87 which are the sum of @xmath2 su(@xmath0)@xmath3 currents : @xmath88 .
since @xmath70 is su(@xmath0)@xmath71 @xmath72 su(@xmath0)@xmath73 invariant , one expects that it does not depend on the su(@xmath0)@xmath1 cft but only on the @xmath79 cft .
in fact , one can check using the following operator product expansion ( ope ) for the left current @xcite : @xmath89 that @xmath90 , i.e. , @xmath70 can not depend on the su(@xmath0)@xmath1 primary fields . in contrast , @xmath70 is a relevant primary field of the @xmath83 cft with scaling dimension @xmath91 and is expected to open a gap @xmath92 for these discrete degrees of freedom .
however , depending on the sign of the coupling constant @xmath93 , the relevant perturbation might give a massless rg flow as the one between consecutive minimal models @xcite .
when @xmath5 , it has been shown in ref . that @xmath70 is in fact related to a perturbation of the @xmath82 cft : @xmath94 where @xmath95 is the hamiltonian density of the @xmath96 parafermionic cft which is generated by the chiral currents @xmath97 with conformal weights @xmath98 @xcite
model ( [ paraintdeforfateev ] ) is known to be an integrable massive field theory for all sign of @xmath93 when @xmath0 is even @xcite .
in contrast , when @xmath0 is odd , it displays a massless rg flow from the @xmath96 fixed point to the minimal model series @xmath99 in the ir limit when @xmath100 , while for @xmath101 , it is again a massive field theory @xcite . in the following ,
we thus consider only the latter case , i.e. , @xmath102 , for general @xmath0 and @xmath2 , where a spectral gap @xmath92 is expected to be formed for the @xmath83 degrees of freedom . in the low - energy limit @xmath103 , the ir properties of the spin ladder model ( [ 2leg ] ) are then governed by the su(@xmath0)@xmath1 cft with a certain perturbation which stems from the @xmath76 contribution in eq .
( [ vint ] ) . to perform this low - energy limit
, it is convenient to consider the full euclidean action of model ( [ vint ] ) . in this respect
, we denote by @xmath104 the action of the su(@xmath0)@xmath3 wznw model : @xmath105 where @xmath106 is an su(@xmath0 ) field and @xmath107 is the famous wznw topological term .
the action of the @xmath2-leg su(@xmath0 ) spin ladder problem is then : @xmath108 for a ferromagnetic interchain coupling @xmath109 , we have @xmath110 and the second term in the right - hand side of eq .
( [ actionkleg ] ) opens a spectral gap @xmath92 as discussed above . in the strong - coupling regime ,
the configuration @xmath111 ( @xmath112 ) minimizes the term with coupling constant @xmath93 in eq .
( [ actionkleg ] ) .
one can then integrate out these degrees of freedom to get an effective action for the @xmath113 field when @xmath103 : @xmath114 with @xmath115 .
one observes that the effective action ( [ actiongeff ] ) describes an su(@xmath0)@xmath1 wznw model perturbed by the @xmath116 field with a positive coupling constant .
the latter perturbation corresponds to the trace of the adjoint field @xmath117 as it can be seen from the definition ( [ adjointfield ] ) and the identity ( [ sunident ] ) .
the low - energy properties of the @xmath2-leg su(@xmath0 ) spin ladder with a ferromagnetic interchain coupling @xmath102 are thus captured by the perturbed su(@xmath0)@xmath1 wznw model ( [ model ] ) or equivalently by the action ( [ actiongeff ] ) .
when @xmath28 , one can immediately argue from eq .
( [ actiongeff ] ) that in the ground state : @xmath118 , @xmath119 , i.e. , elements of the center of the su(@xmath0 ) group .
fluctuations around these minima give rise to massive degrees of freedom .
when @xmath29 , the situation is less clear and a massless rg flow is an intriguing possibility .
in this respect , the connection to @xmath2-leg su(@xmath0 ) spin ladder might give some interesting information . in the @xmath24 case , it is well established that an adiabatic continuity occurs between weak and strong coupling limits for all sign of @xmath120 @xcite .
the @xmath2-leg su(2 ) spin ladder with ferromagnetic interchain coupling @xmath109 is known to be gapless ( respectively fully gapped ) with one gapless bosonic mode , i.e. , @xmath121 when @xmath2 is odd ( respectively even ) @xcite .
the latter result is in full agreement with the massless rg flow of model ( [ actiongeff ] ) from su(2)@xmath1 to su(2)@xmath3 when @xmath2 is odd @xcite .
unfortunately , there are no known numerical results for su(@xmath0 ) spin ladder with @xmath122 and general @xmath2 except for a two - leg su(3 ) spin ladder where an adiabatic continuity has been shown numerically when @xmath102 @xcite .
assuming such continuity when @xmath109 for all @xmath0 , the nature of the ir properties of model ( [ actiongeff ] ) when @xmath122 can then be inferred from a lattice strong - coupling limit of model ( [ 2leg ] ) with @xmath123 . in the latter limit , the @xmath2-leg su(@xmath0 ) spin ladder ( [ 2leg ] )
is equivalent to a single 1d su(@xmath0 ) spin chain where the spin operator belongs to the @xmath2th symmetric representation of su(@xmath0 ) which is described by a young tableau with one line of @xmath2 boxes .
the phase diagram of 1d su(@xmath0 ) heisenberg spin chain in different representations is rather well understood @xcite .
if @xmath2 and @xmath0 have no common divisor , the su(@xmath0 ) heisenberg spin chain in the @xmath2th symmetric representation is known to display an su(@xmath0)@xmath3 quantum critical behavior @xcite .
when @xmath124 a spin gap phase is expected while if @xmath2 and @xmath0 have a common divisor different from @xmath125 the situation is less clear .
we will thus consider here only @xmath2-leg su(@xmath0 ) spin ladder when @xmath2 and @xmath0 have no common divisor .
this result , together with the identification ( [ actiongeff ] ) , leads us to the conjecture that the su(@xmath0)@xmath1 wznw model ( [ model ] ) perturbed by its adjoint primary field has a massless rg flow for @xmath27 to su(@xmath0)@xmath3 cft if @xmath2 and @xmath0 have no common divisor . it might be interesting to relate the proposed massless rg flow to symmetry protection of critical phases with an su(@xmath0 ) symmetry
recently , the massless rg flows of su(2)@xmath1 wznw models have been classified non - perturbatively thanks to a selection rule based on the global anomaly of the @xmath126 discrete symmetry of the center of the su(2 ) group @xcite . for perturbations preserving su(2 ) and @xmath126 symmetries ,
a massless rg flow between su(@xmath127)@xmath1 and su(@xmath127)@xmath128 can occur if only if @xmath2 and @xmath129 have the same parity due to anomaly matching mecanism @xcite .
the latter result stems from the fact that the @xmath126 orbifold of the su(2)@xmath1 wznw model is a consistent cft , i.e. , modular invariant , without global anomaly only if @xmath2 is even @xcite .
anomaly matching requires then a selection rule for the massless rg flow between su(2)@xmath1 wznw models @xcite . for perturbations invariant under su(2 ) and @xmath126 symmetries , a massless rg flow between su(2)@xmath1 and su(2)@xmath3 wznw models is only possible when @xmath2 is odd @xcite
. it might be interesting to extend this argument for a perturbation invariant under su(@xmath0 ) .
on top of the latter symmetry , the field theory ( [ model ] ) enjoys an @xmath96 symmetry , @xmath130 , which corresponds to the one - step translation symmetry of the underlying su(@xmath0 ) spin ladder . in this respect , in close parallel to the @xmath24 case , described in ref .
, we consider an @xmath96 orbifold of su(@xmath0)@xmath1 wznw models .
the spectrum of these models have been determined in ref . .
when @xmath0 is odd modular invariants exist for all @xmath2 while for even @xmath0 , @xmath2 should be even to define a consistent cft @xcite .
the latter result gives a selection rule for the massless rg flow between su(@xmath0)@xmath1 wznw theories with @xmath96 invariant perturbation as the field theory ( [ model ] ) .
in particular , we observe that the conjectured massless flow su(@xmath0)@xmath1
@xmath131 su(@xmath0)@xmath3 when @xmath2 and @xmath0 have no common divisor is compatible with the anomaly matching mechanism of the @xmath96 symmetry .
it is worth noting that the ir massless flow for @xmath5 might be explored perturbatively in the large-@xmath0 limit .
the scaling dimension of the su(@xmath0)@xmath30 perturbed adjoint field is indeed very close to two when @xmath132 : @xmath133 .
a perturbative rg approach is thus called for to find a non - trivial fixed point for model ( [ model ] ) with @xmath5 in the large-@xmath0 limit .
such analysis is similar in spirit to the one - loop rg approach of the massless flow between consecutive minimal models @xmath134 and @xmath135 induced by the @xmath9 perturbation when @xmath136 @xcite .
the analogy can be made more precise by considering the su(@xmath0)@xmath30 fusion rules of @xmath11 with itself which occurs at the one - loop level .
the latter can be derived by exploiting the level - rank duality and the fusion rules of su(2)@xmath137 @xcite : @xmath138 where @xmath139 corresponds to an su(@xmath0)@xmath30 primary field when @xmath140 which transforms in a representation of su(@xmath0 ) described by the following young tableau : @xmath141 the primary field @xmath139 has scaling dimension @xmath142 and is an irrelevant contribution when @xmath143 .
the situation is thus in striking parallel to the minimal model perturbed by the @xmath9 primary field which enjoys the fusion rule : @xmath144 , @xmath145 being an irrelevant field @xcite .
we then expect a non - trivial fixed point in the large-@xmath0 limit .
the details of the perturbative analysis will be presented elsewhere @xcite .
finally , in appendix a , we relate the field theory ( [ model ] ) with @xmath5 to a single su(@xmath0 ) spin chain problem where the spin operators belong to symmetric representation ( @xmath146 ) with dimension @xmath147 of the su(@xmath0 ) group .
it paves the way to the direct numerical investigation of the conjecture for finite @xmath0 in the simplest @xmath5 case .
in this section , we investigate the conjectured massless rg flow for model ( [ model ] ) by means of an direct approach in the simplest case , i.e. , @xmath5 and @xmath4 .
the approach is based on the gepner s parafermions @xcite which extends the @xmath148 parafermionic approach @xcite of the known massless rg flow for @xmath24 , which is reviewed in appendix b. in this respect , it is useful to consider the following conformal embedding : @xmath149 where the coset su(3)@xmath150 describes the so - called gp cft with central charge @xmath151 @xcite .
the su(3)@xmath30 primary field @xmath152 transforms in the left and right su(@xmath153 ) representations with highest weights @xmath154 and @xmath155 , @xmath156 and @xmath157 being weights respectively in the @xmath154 and @xmath155 representations . introducing two left and right bosonic fields @xmath158 , one can relate these primary fields to the one in the gp cft @xcite : @xmath159 where @xmath160 denotes the gp primary field with holomorphic dimension : @xmath161 @xmath162 being the sum of all positive roots of the lie algebra of su(3 ) : su(3 ) .
the su(3)@xmath150 gp primary fields have the following identification @xcite : @xmath163 where @xmath164 are dynkin labels and for notational clarity , we have omitted the weights in the right sector for the gp and su(3)@xmath30 primary fields .
the identification ( [ gpsu3ident ] ) leads to eight gp primary fields @xcite : @xmath165 .
the gp primary fields , which appear in the expression of the adjoint su(3)@xmath30 primary field @xmath166 , are together with their holomorphic dimensions : @xmath167 these fields are hermitean operators due to the identification ( [ gpsu3ident ] ) .
the fusion rules between these primary fields have been derived in ref .
: @xmath168 where the @xmath169 are three gp primary fields with holomorphic dimension @xmath170 which can be chosen as : @xmath171 and @xmath172 . with all these results
, one can derive some representations of the su(3)@xmath30 primary fields in terms of two bosons and the gp primary fields .
let us first consider the fundamental representation of su(3 ) with highest weight @xmath173 and weights @xmath156 : @xmath174 . using the decomposition ( [ gepnerprimary ] )
, we find the expression of the trace of the su(3)@xmath30 wznw primary @xmath19 with scaling dimension @xmath175 : @xmath176 where @xmath177 are the fundamental weights of su(3 ) with the property @xmath178 .
the bosonic field @xmath179 is a compactified field with the following redundancy according to eq .
( [ gepnerprimary ] ) : @xmath180 @xmath181 being integers and the identification ( [ identisu3 ] ) involves the root lattice q which is generated by the simple roots @xmath182 ( @xmath183 ) of su(3 ) .
one can repeat the analysis with the conjugate representation of su(3 ) with highest weight @xmath184 to find : @xmath185 we now consider the adjoint representation of su(3 ) with dimension eight .
the highest weight for the latter representation is @xmath186 with weights @xmath187 , the weight ( 0,0 ) being doubly degenerate . using the decomposition
( [ gepnerprimary ] ) , model ( [ model ] ) for @xmath5 and @xmath4 reads then as follows : @xmath188 where @xmath189 is the hamiltonian density of the su(3)@xmath150 cft and @xmath190 is the dual field of @xmath179 .
expression ( [ effbisgpsu3 ] ) can also be reproduced by considering the ope @xmath191 from eqs .
( [ tracewznwsu3 ] , [ traceconjwznwsu3 ] ) and the fusion rules ( [ gpsu3fusion ] ) .
the ir properties of model ( [ effbisgpsu3 ] ) are difficult to determine . in close parallel to the @xmath24 case ( see appendix b ) , we single out the perturbation with the @xmath192 field which takes the form of the so - called homogenous sine - gordon ( hsg ) model : @xmath193 the latter is an integrable massive field theory for all sign of @xmath194 and corresponds to the generalization of the integrable model of @xmath148 parafermions perturbed by their first thermal operator @xcite . when @xmath195 , i.e. @xmath196 , we have in our conventions @xmath197 ( @xmath198 ) . at lower energy than the mass gap of the hsg model , model ( [ effbisgpsu3 ] ) becomes equivalent to a double sine - gordon model with scaling dimension one : @xmath199 the latter model has a spectral gap and the bosonic @xmath179 field is pinned in the minima of the sine - gordon potential ( @xmath200 ) : @xmath201 @xmath202 being integers .
the vacuum expectations values of the bosons are thus described by the weight lattice @xmath203 which is generated by the fundamental weights . using the identification ( [ identisu3 ] ) on the bosons , we find that the ground state is only three - fold degenerate since the ratio p / q of the lattices is isomorphic to the center of su(3 ) : @xmath204 @xcite .
when @xmath205 , we have now @xmath206 . from the fusion rule ( [ gpsu3fusion ] ) @xmath198 , we expect that one enters a phase where @xmath207 and the disorder fields @xmath208 of gp should condense now .
using the fusion rules ( [ gpsu3fusion ] ) , we need to consider the second - order in perturbation theory to generate an effective low - energy model for the @xmath179 bosons after the integration of the massive gp degrees of freedom : @xmath209 where @xmath210 and @xmath211 are the luttinger parameters for the bosonic fields @xmath212 .
the scaling dimension of the perturbation is @xmath213 and thus irrelevant .
one expects the emergence of a quantum critical behavior with two gapless bosons , i.e. , with central charge @xmath214 .
however , we need to determine the actual values of the luttinger parameters which should be fixed since the symmetry of the initial model ( [ effbisgpsu3 ] ) is su(3 ) .
when @xmath215 , we have @xmath207 and @xmath216 in eq .
( [ tracewznwsu3 ] ) seems to be naively short ranged .
however , we need to considering higher - order in perturbation theory to deduce the low - energy expression of @xmath216 . in this respect , by performing the ope between ( [ tracewznwsu3 ] ) and the interacting hamiltonian ( [ effbisgpsu3 ] ) , we see that the contribution of the @xmath217 fields disappear and we get the ir representation of @xmath216 : @xmath218 since the three terms should have the same scaling dimension , we have necessarily @xmath219 and the scaling dimension of @xmath220 in the ir limit is : @xmath221 .
since we expect a massless flow to su(3)@xmath3 in the far ir , the su(3)@xmath30 wznw primary field g will transmute to the su(3)@xmath3 one with scaling dimension @xmath222 .
we have thus @xmath223 . the double sine - gordon model ( [ sgmodel2ndorder ] ) , which describes the ir physics of model ( [ model ] ) when @xmath27 , becomes marginal and identifies with the marginal irrelevant su(3)@xmath3 current - current perturbation @xcite . in summary , model ( [ model ] ) for @xmath5 and @xmath4 displays a massless rg flow to su(3)@xmath3 when @xmath27 with the ultraviolet - infrared transmutation : @xmath224 @xmath225 being the su(3)@xmath3 chiral currents and @xmath226 is the su(3)@xmath3 wznw primary field .
in this work , we have identified an ir massless rg flow for the su(@xmath0)@xmath1 wznw model perturbed by its relevant adjoint primary field . using a mapping onto @xmath2-leg su(@xmath0 ) spin ladder and assuming a weak - strong coupling continuity , we have shown that this model has critical properties in the su(@xmath0)@xmath3 universality class when @xmath0 and @xmath2 have no common divisor .
this result is the generalization to su(@xmath0 ) of the haldane s conjecture on spin-@xmath26 su(2 ) heisenberg chain whose physical properties are governed by the su(2)@xmath227 wznw model perturbed by its adjoint primary field .
the massless rg flow , presented in this paper , is consistent with the extension of the selection rule on wznw models based on the global anomaly of the @xmath126 symmetry @xcite .
futhermore , we have confirmed the existence of the massless rg flow for the special @xmath4 and @xmath5 case by means of a direct approach using gepner s parafermions .
the resulting non - trivial ir fixed point for @xmath5 can be investigated by a perturbative rg approach in the large-@xmath0 limit since the scaling dimension of the adjoint primary field is close to two when @xmath132 . as perspectives
, it will be interesting to have a direct approach of the ir properties of the perturbed su(@xmath0)@xmath1 wznw to complement the analysis of the massless rg flow identified in this paper . in this respect ,
a truncated conformal space approach in the simplest @xmath4 and @xmath5 case will be very useful as it has been done recently for su(2)@xmath1 perturbed cft @xcite
. a semiclassical approach of model ( [ model ] ) might also be very helpful to interpret the massless rg flow reported in this work as the result of some non - linear sigma model with a topological term as in the @xmath24 case @xcite .
finally , direct numerical calculations of @xmath2-leg su(@xmath0 ) spin ladder with @xmath102 and su(@xmath0 ) spin chain models with symmetric rank-@xmath2 tensor representation might also be fruitful to reveal the massless rg flow . we hope that future studies will shed light on theses questions and others results on su(@xmath0)@xmath1 perturbed cft will be obtained .
the author is grateful to r. santachiara , a. m. tsvelik , v. bois , s. capponi , t. quella , h. saleur , k. totsuka , and a. weichselbaum for useful discussions .
we relate here the field theory ( [ model ] ) to an su(@xmath0 ) spin chain problem where the spin operators belong in symmetric rank-@xmath2 tensor representation of su(@xmath0 ) .
such mapping might be useful to investigate numerically the massless rg flow proposed in this paper .
the starting point of the analysis is the existence of an integrable su(@xmath0 ) model with degrees of freedom in symmetric rank-@xmath2 tensor representation , introduced by andrei and johannesson ( aj ) @xcite .
the latter model is the su(@xmath0 ) generalization of bethe - ansatz integrable spin-@xmath26 heisenberg chain models which represent unstable su(2)@xmath227 quantum critical points @xcite .
it has been numerically proved that the aj model displays a quantum critical behavior in the general su(@xmath0)@xmath1 wznw universality class @xcite . since the latter cft has many relevant primary fieds ,
the critical point is expected to be very fragile . in this respect , let us introduce the following bilinear - biquadratic su(@xmath0 ) spin chain model to analyse the stability of the aj model in the simplest @xmath5 case : @xmath228 where @xmath229 denotes spin operators on site @xmath230 which transforms in the symmetric representation @xmath146 of the su(@xmath0 ) group . model ( [ ajmodel ] )
interpolates between the aj model for @xmath231 with su(@xmath0)@xmath30 quantum criticality and the pure heisenberg model in @xmath146 representation for @xmath232 . as already stressed in section ii.c , the latter displays an su(@xmath0)@xmath3 quantum critical behavior when @xmath0 is odd @xcite . model ( [ ajmodel ] ) might thus be a lattice description of a massless rg flow from su(@xmath0)@xmath30 to su(@xmath0)@xmath3 when @xmath0 is odd . in this respect
, it will be interesting to investigate this model numerically by means of the density - matrix renormalization group calculations in the simplest @xmath4 case to shed light on this intriguing possibility .
a field theory analysis of this problem can be obtained by considering at @xmath233 an 1d u(@xmath0 ) fermionic hubbard model with fermions with @xmath0 flavors and @xmath2 colors at filling factor @xmath41 @xcite .
for this special filling , the latter model enjoys an @xmath96 symmetry , associated to the one - step translation invariance , which might protect it from a mass - gap generation @xcite .
the low - energy approach of model ( [ ajmodel ] ) can then be derived when @xmath234 by exploiting the fact that the aj model has an su(@xmath0)@xmath30 critical behavior : @xmath235 where @xmath87 are su(@xmath0)@xmath30 chiral currents and @xmath236 is the leading relevant perturbation to be found which describes the departure from the su(@xmath0)@xmath30 fixed point .
it should be invariant under the symmetries of the lattice model ( [ ajmodel ] ) , in particular , the one - step translation symmetry which takes the form of an @xmath96 symmetry .
the latter gives strong constraints on the identification of @xmath236 . in this respect ,
let us recall the spectrum of the su(@xmath0)@xmath30 cft .
it has @xmath147 primary operators with highest - weights @xmath237 such that the dynkin labels satisfy the constraint : @xmath238 . introducing @xmath239 as a young tableau row lengths
, we see that the young tableau can not have more than two columns .
the scaling dimensions of the su(@xmath0)@xmath30 primary fields are given by @xcite : @xmath240 with @xmath241 is the number of boxes in the young tableau and @xmath242 .
the primary field which transforms according to the @xmath2th antisymmetric representation of su(@xmath0 ) is described by the highest - weight @xmath243 ( @xmath244 being in the @xmath2th position ) and has scaling dimension according to eq .
( [ scalingdimsun2 ] ) : @xmath245 .
these operators can not appear in model ( [ ajmodelcont ] ) since they are not invariant under the one - step translation symmetry .
the most relevant operator , which is translational invariant , turns out to be the primary field in the adjoint representation with highest weight ( 1 , 0 , , 0 , 1 ) and scaling dimension @xmath246 .
we thus expect that the continuum limit of model ( [ ajmodel ] ) in the vicinity of @xmath247 is described by the field theory ( [ model ] ) .
the massless flow for odd @xmath0 from su(@xmath0)@xmath30 to su(@xmath0)@xmath3 can thus be directly investigated numerically from the 1d lattice model ( [ ajmodel ] ) by means of density matrix renormalization group calculations for instance .
in this appendix , we review for completeness the ir properties of model ( [ model ] ) with @xmath24 by using a mapping onto @xmath148 parafermionic cft @xcite . in the @xmath24 case ,
model ( [ model ] ) reads as follows : @xmath248 where @xmath249 are chiral su(2)@xmath1 currents and @xmath250 is the spin-1 su(2)@xmath1 primary field with scaling dimension @xmath251 .
we then consider the conformal embedding @xcite : @xmath252 where the @xmath148 parafermionic cft with central charge @xmath253 describes the critical properties of two - dimensional @xmath148 generalization of the ising model @xcite .
the su(2)@xmath1 primaries ( @xmath254 ) are related to the z@xmath1 parafermionic ones ( @xmath255 ) by @xcite : @xmath256 where @xmath257 and @xmath236 is a bose field with chiral components @xmath258 . from the identification ( [ suparaprim ] ) , we get : @xmath259 where @xmath260 ( @xmath261 ) are @xmath148 primary fields with scaling dimensions @xmath262 @xmath263 which describe long - distance correlations of the @xmath148 ising lattice spins . in eq .
( [ paraboso ] ) , @xmath264 is the first thermal operator with scaling dimension @xmath265 and our notations are such that @xmath266 in the @xmath148 ordered phase of the underlying @xmath148 ising model .
model ( [ modeln2 ] ) can then be rewritten as follows in this new basis : @xmath267 @xmath268 being the dual bose field ( @xmath269 ) .
the perturbation is strongly relevant and since the @xmath148 degrees of freedom are discrete , one expects the opening of a mass gap in this sector .
the next step of the approach is therefore to single out the thermal perturbation on the @xmath148 parafermions : @xmath270 which is an integrable massive field theory for all sign of @xmath194 @xcite .
when @xmath28 , we have @xmath271 and the order spin fields condense @xmath272 , leading thus to a sine - gordon model for the @xmath236 field : @xmath273 which is a massive field theory for all @xmath2 .
the situation turns out to be very different when @xmath27 . in the latter case
, we have now @xmath274 and the parafermionic belongs to the disordered phase where @xmath275 .
the integration over the massive parafermionic degrees of freedom in eq .
( [ modeln2para ] ) leads to an low - energy effective hamiltonian for the bosonic field @xmath236 which depends on the parity of @xmath2 @xcite .
we first consider the @xmath2 odd case .
the @xmath276 operator in eq .
( [ modeln2para ] ) carries a @xmath277 charge under @xmath148 .
we need to consider higher - order in perturbation theory to cancel out the @xmath148 charge of @xmath276 .
when @xmath2 is odd , the @xmath2th order of perturbation is necessary to suppress the @xmath276 contribution .
the low - energy hamiltonian for the bose field @xmath236 takes then the form of a sine - gordon model at @xmath278 : @xmath279 with @xmath280 , @xmath281 is the luttinger parameter and @xmath62 is a velocity .
the scaling dimension of the perturbation is @xmath282 and thus naively irrelevant .
we also notice that the approach gives the same low - energy effective field theory than the one derived directly by schulz from the abelian bosonization of half - integer @xmath283 heisenberg spin chain @xcite .
the global continuous symmetry of model ( [ modeln2para ] ) is su(2 ) and therefore the luttinger parameter @xmath281 should be fixed to a value compatible with this non - abelian symmetry .
one way to identify @xmath281 is to determine the ir limit of @xmath284 in eq .
( [ paraboso ] ) . by fusing this operator with the hamiltonian ( [ modeln2para ] ) at the @xmath285th order in perturbation theory
, the @xmath286 contribution disappears and one obtains the low - energy limit description of @xmath284 : @xmath287 which corresponds to the su(2 ) spin - singlet dimerization operator if @xmath288 @xcite .
the su(2 ) symmetry of the problem fixes thus @xmath289 and @xmath19 has scaling dimension @xmath170 , i.e. , corresponds in the far ir limit to the spin-1/2 su(2)@xmath3 wznw primary .
a massless rg flow to su(2)@xmath3 cft is therefore expected in the odd @xmath2 case when @xmath27 as it should be . as in the @xmath2 odd case
, one has to consider higher orders in perturbation theory to derive an effective theory for the field @xmath236 since the @xmath276 operator in eq .
( [ modeln2para ] ) average to zero in the @xmath148 symmetric phase .
when @xmath2 is even , one needs now the @xmath290 th order of perturbation theory to cancel out the @xmath276 operator in eq .
( [ modeln2para ] ) .
the resulting low - energy hamiltonian then reads as follows : @xmath291 we recover the same low - energy approach than the one derived by schulz in his study of integer heisenberg spin @xmath292 chain @xcite . as seen in the previous case , the su(2 ) symmetry
is fixed by the value of the luttinger parameter : @xmath288 .
model ( [ hcneven ] ) becomes the @xmath293 sine - gordon model which is massive and enjoys an hidden su(2 ) symmetry @xcite . in stark contrast to the odd @xmath2 case , one can not suppress the contribution of the spin field @xmath286 in eq .
( [ paraboso ] ) by considering higher - order in perturbation theory with the hamiltonian ( [ modeln2para ] ) .
the ir limit of @xmath284 always gives a short - ranged contribution . in summary , from this @xmath148 parafermionic approach , we conclude that model ( [ modeln2 ] ) is fully gapped for all sign of @xmath194 when @xmath2 is even as it should
. 101 p. di francesco , p. mathieu , and d. snchal , _ conformal field theory _
( springer , berlin , 1997 ) .
g. mussardo , _ statistical field theory : an introduction to exactly solved models in statistical physics _
( oxford university press , uk , 2009 ) .
a. o. gogolin , a. a. nersesyan , and a. m. tsvelik , _ bosonization and strongly correlated systems _ ( cambridge university press , uk , 1998 ) .
i. affleck , phys .
lett . * 55 * , 1355 ( 1985 ) .
i. affleck , nucl . phys .
b * 265 * , 409 ( 1986 ) . v. g. knizhnik and a. b. zamolodchikov , nucl .
b * 247 * , 83 ( 1984 ) .
a. b. zamolodchikov and .
v. a. fateev , sov .
. phys . * 43 * , 657 ( 1986 ) .
j. l. cardy , nucl .
b * 270 * , 186 ( 1986 ) .
a. m. lauchli , arxiv : 1303.0741 .
h. w. j. blote , j. l. cardy , and m. p. nightingale , phys .
lett . * 56 * , 742 ( 1986 ) .
i. affleck , phys .
. lett . * 56 * , 746 ( 1986 ) .
p. calabrese and j. l. cardy , j. phys .
a * 42 * , 504005 ( 2009 ) .
a. b. zamolodchikov , adv .
pure math .
* 19 * , 641 ( 1989 ) .
a. b. zamolodchikov , jetp lett * 43 * , 730 ( 1986 ) .
a. b. zamolodchikov , sov .
* 46 * , 1090 ( 1987 ) .
a. a. w. ludwig and j. l. cardy , nucl .
b * 285 * , 687 ( 1987 ) .
b. zamolodchikov , nucl .
b * 358 * , 524 ( 1991 ) .
t. klassen and e. melzer , nucl .
b * 370 * , 511 ( 1992 ) . v. p. yurov and a. b. zamolodchikov , int . j. mod . phys .
a * 5 * , 3221 ( 1990 ) ; * 6 * , 4557 ( 1991 ) .
r. m. konik and y. adamov , phys .
. lett . * 98 * , 147205 ( 2007 ) ; * 102 * , 097203 ( 2009 ) .
i. affleck and f. d. m. haldane , phys .
b * 36 * , 5291 ( 1987 ) .
r. m. konik , t. plmai , g. takcs , and a. m. tsvelik , nucl .
b * 899 * , 547 ( 2015 ) .
f. d. m. haldane , phys .
lett * 50 * , 1153 ( 1983 ) ; phys . lett . *
93a * , 464 ( 1983 ) .
h. nonne , m. moliner , s. capponi , p. lecheminant , and k. totsuka , epl * 102 * , 37008 ( 2013 ) . t. morimoto , h. ueda , t. momoi , and a. furusaki , phys . rev .
b * 90 * , 235111 ( 2014 ) .
v. bois , s. capponi , p. lecheminant , m. moliner , and k. totsuka , phys . rev .
b * 91 * , 075121 ( 2015 ) .
d. gepner , nucl .
b * 290 * , 10 ( 1987 ) .
a. v. gorshkov , m. hermele , v. gurarie , c. xu , p. s. julienne , j. ye , p. zoller , e. demler , m. d. lukin , and a. m. rey , nat . phys . * 6 * , 289 ( 2010 ) .
m. a. cazalilla , a. f. ho , and m. ueda , new j. phys .
* 11 * , 103033 ( 2009 ) .
f. scazza , c. hofrichter , m. hofer , p. c. de groot , i. bloch , and s. folling , nat . phys .
* 10 * , 779 ( 2014 ) .
x. zhang , m. bishof , s. l. bromley , c. v. kraus , m. s. safronova , p. zoller , a. m. rey , and j. ye , science * 345 * , 1467 ( 2014 ) . m. a. cazalilla and a. m. rey , rep .
77 * , 124401 ( 2014 ) .
s. capponi , p. lecheminant , and k. totsuka , arxiv : 1509.xxxxx .
b. sutherland , phys .
b * 12 * , 3795 ( 1975 ) .
i. affleck , nucl .
b * 305 * , 582 ( 1988 ) .
p. lecheminant and a. m. tsvelik , phys .
b * 91 * , 174407 ( 2015 ) .
r. assaraf , p. azaria , m. caffarel , and p. lecheminant , phys .
b * 60 * , 2299 ( 1999 ) . c. itoi and m .- h .
kato , phys .
b * 55 * , 8295 ( 1997 ) .
s. r. manmana , k. r. a. hazzard , g. chen , a. e. feiguin , and a. m. rey , phys .
a * 84 * , 043601 ( 2011 ) .
e. witten , commun .
phys . * 92 * , 455 ( 1984 ) .
i. affleck , d. gepner , h. schulz , and t. ziman , j. phys . a : math .
* 22 * , 511 ( 1989 ) .
p. lecheminant , phys .
b * 648 * , 323 ( 2007 ) .
p. lecheminant and h. nonne , phys .
b * 85 * , 195121 ( 2012 ) . a. b. zamolodchikov and v. a. fateev , sov .
jetp * 62 * , 215 ( 1985 ) .
d. gepner and z qiu , nucl .
b * 285 * , 423 ( 1987 ) .
p. bouwknegt and k. schoutens , phys
* 223 * , 183 ( 1993 ) .
v. a. fateev , int .
phys . a * 6 * , 2109 ( 1991 ) .
v. a. fateev and a. zamolodchikov , phys .
b * 271 * , 91 ( 1991 ) .
t. giamarchi , _ quantum physics in one dimension _
( clarendon press , oxford , uk , 2004 ) .
s. capponi , p. lecheminant , a. m. tsvelik , and a. weichselbaum , in preparation . i. affleck and e. h. lieb , lett .
* 12 * , 77 ( 1986 ) .
n. read and s. sachdev , nucl .
b * 316 * , 609 ( 1989 ) .
s. rachel , r. thomale , m. fuehringer , p. schmitteckert , and m. greiter , phys .
b * 80 * , 180420(r ) ( 2009 ) .
h. nonne , p. lecheminant , s. capponi , g. roux , and e. boulat , phys .
b * 84 * , 125123 ( 2011 ) .
j. dufour , p. nataf , and f. mila , phys .
b * 91 * , 174427 ( 2015 ) .
s. c. furuya and m. oshikawa , arxiv : 1503.07292 .
d. gepner and e. witten , nucl .
b * 278 * , 493 ( 1986 ) .
g. felder , k. gawedzki , and a. kupiainen , commun .
* 117 * , 127 ( 1988 ) .
s. g. naculich , h. a. riggs , and h. schnitzer , phys .
b * 246 * , 417 ( 1990 ) ; nucl .
b * 357 * , 687 ( 1990 ) .
p. lecheminant and r. santachiara , in preparation .
d. cabra , p. pujol , and c. von reichenbach , phys .
b * 58 * , 65 ( 1998 ) .
d. gepner , phys .
b * 222 * , 207 ( 1989 ) .
e. ardonne and k. schoutens , annals of physics * 322 * , 201 ( 2007 ) . c. r. fernandez - pousa , m. v. gallas , t. j. hollowood , and j. l. miramontes , nucl .
b * 484 * , 609 ( 1997 ) .
o. castro - alvaredo , a. fring , c. korff , and j. l. miramontes , nucl .
b * 575 * , 535 ( 2000 ) . p. dorey and j. l. miramontes , nucl .
b * 697 * , 405 ( 2004 ) .
r. citro , e. orignac , n. andrei , c. itoi , and s. qin , j. phys . : condens .
matter * 12 * , 3041 ( 2000 ) .
m. beria , g. p. brandino , l. lepori , r. m. konik , and g. sierra , nucl .
b * 877 * , 457 ( 2013 ) .
n. andrei and h. johannesson , phys .
a * 104 * , 370 ( 1984 ) .
h. johannesson , nucl .
b * 270 * , 235 ( 1986 ) .
f. c. alcaraz and m. martins , j. phys .
a : math . gen . * 22 * , l865-l870 ( 1989 ) .
m. fuehringer , s. rachel , r. thomale , m. greiter , and p. schmitteckert , ann .
( berlin ) * 17 * , 922 ( 2008 ) .
j. schulz , phys .
b * 34 * , 6372 ( 1986 ) . i. affleck , nucl .
b * 265 * , 448 ( 1986 ) . | we investigate the infrared properties of su(@xmath0)@xmath1 conformal field theory perturbed by its adjoint primary field in 1 + 1 dimensions .
the latter field theory is shown to govern the low - energy properties of various su(@xmath0 ) spin chain problems .
in particular , using a mapping onto @xmath2-leg su(@xmath0 ) spin ladder , a massless renormalization group flow to su(@xmath0)@xmath3 criticality is predicted when @xmath0 and @xmath2 have no common divisor .
the latter result extends the well - known massless flow between su(2)@xmath1 and su(2)@xmath3 wess - zumino - novikov - witten theories when @xmath2 is odd in connection to the haldane s conjecture on su(2 ) heisenberg spin chains .
a direct approach is presented in the simplest @xmath4 and @xmath5 case to investigate the existence of this massless flow . |
a full understanding of the complex dynamical scenarios accompanying the glass transition requires a resolution of the statistics of free energy barriers at any given temperature .
such information is very hard to come by .
trying to measure such free energy barriers using the energy landscape of a typical glass former is a daunting procedure that has never been achieved . in this paper
we propose a method that provides us with a decent approximation of the statistics of the free energy barriers .
we do this by measuring quenched " rather than annealed " relaxation times .
the usual procedure for extracting a relaxation time from numerical simulations of any glassy system is the annealed " procedure .
one takes many realizations of a super - cooled liquid at a given temperature , measures a typical relaxation function for each realization , averages the relaxation functions over the ensemble and finally extracts a relaxation time @xmath1 from this average relaxation function by , say , determining when the average function reaches @xmath2 of its initial value @xcite .
the subscript @xmath3 designates the ` main ' slow relaxation mechanism in glassy dynamics , and is the only one that is typically resolved in numerical simulations .
this procedure is automatically performed in experiments where the macroscopic nature of the relaxing systems provides self - averaging resulting in ` smooth ' relaxation functions that hardly fluctuate from realization to realization . in this paper
we propose that significant simulational insight on glassy relaxation can be gained by adopting a quenched " procedure . in this procedure
a relaxation time @xmath4 is extracted from each and every realization by determining when the appropriate relaxation function reaches a value of @xmath2 , cf .
[ relaxation ] upper panel . finally an average is taken over the ensemble to provide @xmath5 .
a typical such quenched relaxation time for a lennard - jones binary glass is shown in the lower panel of fig .
[ relaxation ] .
here we measured the self part of the intermediate scattering function for each realization separately ( upper panel ) , i.e. the function @xmath6}\right\ } \ . \label{deffk}\ ] ] where the index @xmath7 runs over half the particles in the binary mixture ( those with longer interaction length ; we assume the mass of all the particles to be the same , @xmath8 ) .
one observes the usual dramatic slowing down , such that @xmath5 grows rapidly when @xmath9 decreases , first in a pre - arrhenius form ( see inset ) , then in an arrhenius form , linear in @xmath10 , and later in a faster , super - arrhenius form , which is referred to as ` fragile ' behavior in the glass community @xcite . in this paper
we will be interested predominantly in the transitions to and from the arrehenius regime , for which the quenched procedure is particularly illuminating .
the advantage of the quenched over the annealed procedure is that it allows a particularly transparent treatment of the statistics of free energy barriers .
having an ensemble of @xmath4 values for a given ensemble of super - cooled systems , we can define a typical free energy scale by writing for each @xmath4 value an equation @xmath11 where we choose units such that bolzmann s constant equals unity .
inverting eq .
( [ defe ] ) , every value of @xmath4 yields a value of @xmath12 .
the free energy has in principle a contribution from an energy barrier @xmath13 and an entropic contribution stemming from a degeneracy @xmath14 , @xmath15 of course , this definition makes sense when the dynamics involves escaping an energy barrier .
the prefactor @xmath16 is by definition the relaxation time inside the confined state at every temperature @xmath9 , and we assume that it is fixed for all the realizations in the ensemble : @xmath17 the dependence on @xmath18 stems simply from the estimating @xmath19 , the typical particle velocity , from equipartition as @xmath20 and @xmath16 is then a typical length - scale ( a cage magnitude ) over @xmath19 . for our purpose the constant @xmath21
is determined by requiring that the plot of @xmath22 as @xmath23 , see fig .
[ relaxation ] ( red cross ) and is @xmath24 . -20pt
our first observation is that in the arrhenius regime the distribution of free energies @xmath12 is always gaussian , and is always very sharply peaked . since @xmath12 in the arrhenius regime is temperature independent ,
we conclude that @xmath25 , and @xmath26 . as an example
we show again the binary lennard - jones system for which we have used 1000 realizations for each value of @xmath9 to generate the distributions of @xmath12 shown in fig .
[ gaussarr ] .
we conclude that in the arrhenius regime to a very good approximation @xmath27 with the peak value @xmath28 _ and _ the width @xmath29 _ unchanged _ throughout the arrhenius regime . in the present regime
we denote @xmath30 .
it is noteworthy that the distributions are also very sharply peaked ; to see this note that the quenched average can be now computed from the gaussian integral as @xmath31 \ . \label{quenched}\ ] ] alternatively , @xmath32 for the data shown in fig .
[ gaussarr ] @xmath33 and @xmath34 .
thus for the arrhenius range of temperatures @xmath35 the contribution of the width @xmath36 is of the order of @xmath37 in eq .
( [ quenched ] ) .
obviously , for larger systems we expect @xmath29 to be even smaller , tending to zero for @xmath38 .
it appears that all the relaxation events contributing to the arrhenius regime belong to a tight group of similar events with a very well defined energy barrier @xmath39 and a negligible dispersion .
in fact , we find that this conclusion is not particular either to the lennard - jones system or to @xmath40 . in fig .
[ examplesarrhenius ] we show similar results for other glass formers with a ) purely repulsive potential @xcite at @xmath40 , b ) the lennard - jones system @xcite at @xmath41 , c ) the kob - andersen model @xcite at @xmath40 and d ) the polydisperse model @xcite at @xmath40 .
the physical interpretation of these observations is discussed below , cf .
[ modelsec ] .
at high temperatures we observe ( see for example fig .
[ fullt - lennard ] ) that @xmath42 does not have a fixed dispersion ( it widens when @xmath9 increases ) and its peak moves to lower free energy values .
this observation is generic for all the models studied here . the physical interpretation of this observation is again discussed in sect .
[ modelsec ] .
obviously , one might expect that when @xmath9 reaches a value of @xmath43 something interesting should happen .
the physical relaxation mechanism whose energy barrier is @xmath39 should run out of steam , and other mechanisms , if they exist , should emerge .
this phenomenon is beautifully seen in the lennard - jones system , see fig .
[ fullt - lennard ] .
indeed , at @xmath44 the arrhenius process begins to disappear and we observe at first a gradual increase in both @xmath28 ( the peak position ) and its width @xmath29 . in the system
for which @xmath45 , we find that at @xmath46 the high temperature gaussian peak disappears in favor of a second gaussian peak which appears to replace it .
we will denote by @xmath47 the temperature at which both process coexist with equal probability .
the gaussian of the second process marches to the right with increasing both of @xmath28 and @xmath29 .
the transition region is depicted in fig .
[ fullt - lennard ] and is shown in more detail in fig .
[ transition ] . for the systems of the size studied in these simulations , by the time that the second relaxation mechanism takes over , the contribution of the variance @xmath29 to the quenched average relaxation time eq .
( [ quenched ] ) is no longer negligible .
thus for our example at @xmath48 @xmath49 and @xmath50 such that @xmath51 . at lower temperature the dispersion effect will contribute more and more to the relaxation time .
of course , also here we expect @xmath29 to decrease to zero when @xmath38 .
it is useful however to keep the @xmath29 dependence for the simulations at hand .
-15pt we should note that eq .
( [ simple ] ) remains valid throughout the temperature range .
we can rewrite it , denoting @xmath52 and introducing @xmath53 for the weighted sum of the two different mechanisms in the form @xmath54 \ .\ ] ] we test this formula for the data corresponding to the distributions found in fig .
[ fullt - lennard ] and demonstrate the perfect agreement in fig .
[ fig7 ] .
-25pt note that the relaxation time varies in this range of temperatures by close to 5 orders of magnitude . -20pt
in ref . @xcite it was shown that relaxation times of models of the type discussed above exhibit density scaling . for different densities
@xmath55 plots of @xmath56 vs. @xmath10 fall on different curves @xcite .
once plotted as a function of @xmath57 with an appropriate value of @xmath58 the data collapses onto one curve .
density scaling is easiest to justify when one uses a purely repulsive potential of the form @xmath59 where @xmath60 sets the energy scale of the interparticle potential and @xmath61 is measured in dimensionless units .
then the only relevant length scale @xmath62 in the system is determined by the density as @xmath63 where @xmath55 is the density in dimensionless units . obviously , when we plot @xmath64 vs. @xmath10 ( see upper panel of fig .
[ fig8 ] ) we present a dimensionless number in terms of a dimensional quantity which is inappropriate . to correct this
, we need to non - dimensionalize @xmath10 by multiplying it by an energy scale .
the only energy available is @xmath65 , and we write @xmath66 thus by presenting @xmath64 vs. @xmath67 we should find data collapse as discussed in ref .
it is well known that the argument remains approximately valid also when the potential is not a pure power law but with an exponent @xmath58 that needs to be found by collapsing the data @xcite . in the second panel of fig .
[ fig8 ] we show the data collapse when plotted properly , with the density varying in the range @xmath68 $ ] .
additionally , in these systems the inter - particle potential possesses only one energy scale , which is nothing but the depth of the inter - particle potential , denoted as @xmath60 .
this energy scale must determine the typical barrier height that needs to be surmounted in the arrhenius regime .
accordingly , we expect @xmath39 to be proportional to @xmath60 .
this expectation is fully supported by the data shown in fig .
[ linear ] where for two models ( binary lennard - jones in 2 and 3 dimensions ) we have changed @xmath60 and measured @xmath39 . -7pt our additional contribution to the discussion of density scaling is in pointing out that it is obeyed separately by the two contributions to eq .
( [ simple ] ) . in the upper row of fig .
[ fig9 ] we show that @xmath28 exhibits a very nice data collapse when data for different densities are re - plotted as required . in the lower row
we show the same for @xmath29 .
we will use these scaling functions below in providing a model for the observed glassy dynamics . -1.8
an obvious worry about our method of analysis is that the dispersion @xmath29 depends on the system size , and therefore the temperature @xmath47 at which we observe the transition to fragility may depend on the system size .
this is of course correct .
nevertheless we discover that the value of @xmath47 converges to a value @xmath69 in the thermodynamic limit .
the data is shown for the binary lennard - jones model in fig .
[ lintstarf ] . from the data in the lower panel
one can conclude that for this model @xmath70
-15pt we did not perform an exhaustive analysis of the exponents of this law of convergence for the different models .
the purpose of this section is to establish one of the main claims of this paper , i.e. that with the transition to the super - arrhenius behavior one observes a change from a local to a cooperative relaxation mechanism . at temperatures below @xmath47
one expects a cooperative relaxation process to play an important role ; the proposition of this paper is that this is mainly due to the disappearance of the free volume that was available in the arrhenius regime ( see quantitative model below ) , requiring now a cooperative motion of many particles to ` collect ' enough free volume to allow a relaxation step .
although such a picture is current among some practitioners in the glass community , to the best of our knowledge it had never been shown explicitly .
we use in this section scaling arguments to establish this picture convincingly . adding the system size dependence to our typical free energy , we write now @xmath71.\ ] ] we expect @xmath72 to be determined by the energy landscape and not to depend strongly on the system size .
energy barriers are mostly sensitive to local arrangements of particles which are not system size dependent . on the other hand
@xmath73 $ ] is the entropic contribution to the free energy barrier .
our data showed that in the arrhenius regime the degeneracy factor @xmath14 was of the order of unity ; here we will show that our data strongly supports a cooperative process in the super - arrhenius regime , mainly due to the system size dependence of the degeneracy factor .
the flip side of having a cooperative mechanism is that there should exist a typical length @xmath74 which measures the degree of cooperativity and is increasing when the temperature decreases .
consider then a system of @xmath75 particles in the super - arrhenius regime , associated with a typical scale @xmath74 , and contained in a cubic box of size @xmath76 at some temperature @xmath9 .
as long as the system is small , i.e. @xmath77 , the whole system needs to cooperate in order to relax .
accordingly we expect the degeneracy factor to grow extensively like @xmath78 . when the system is large enough , i.e. when @xmath79 we expect to find @xmath80 , not changing with the system size .
in other words , @xmath81 where @xmath82 is typical inter particle distance . accordingly
the free energy should assume the form of a scaling function of @xmath83 as @xmath84,\ ] ] where @xmath85 is a temperature independent proportionality factor .
the scaling function @xmath86 has the following asymptotic dependence @xmath87
-15pt considering the small system limit , we conclude that there the free energy barrier should exhibit a logarithmic dependence on the system size .
this expectation is supported by the data shown in fig [ fvsn ] . in the upper panel
we show the system size dependence of the free energy value where the distribution function @xmath88 has a second peak , for different temperatures . with increasing system size
these values tend to saturate to an asymptotic value .
note that the dependence on the system size seems to be decreasing with increasing temperature , indicating the existence of a cooperative length scale .
we can not go to high temperatures in this study because at higher temperature the extraction of the second peak position become difficult as its height goes down drastically , merging with the arrhenius peak .
we focus on data where the position of the second peak can be estimated accurately . to show the logarithmic dependence on the system size we collapsed the data together by rescaling the free energy axis by @xmath89 with @xmath90 and the x axis by @xmath91
. for simplicity we fitted here a guess function for @xmath74 in the form @xmath92 @xcite . in the lower panel of fig [ fvsn ] one can see the convincing logarithmic dependence of the free energy on the system size for different temperatures .
the data shown above indicates two changes in the relaxation mechanism between the pre - arrhenius and the arrhenius regime and then between the arrhenius and the fragile regime .
we propose that the pre - arrhenius regime is a fluid regime , in which a well - defined cage is only beginning to form around every particle .
it is relatively easy to break this ill - formed cage , and there are many ways to do it . accordingly the free energy needed is made of a low value of @xmath13 and a finite value of @xmath14 such that when @xmath9 is increasing the gaussian peak that we observe marches to the left and becomes broader . in the arrhenius regime
a well defined cage has formed around each particle , and this cage is a more or less regular cluster of particles aggregating around each center particle .
we expect a roughly constant energy @xmath13 to suffice to break this kind of cage and the width of the distribution of @xmath12 should be relatively narrow as discussed below . towards the end of the arrhenius regime a second and maybe third layer of next - nearest neighbors begins to form around a center particle , and therefore it becomes more and more difficult for the single - particle relaxation to take place . the energy barrier increases but because of the fluidity of the second and third layer also the width of the distribution starts to increase . eventually a new , cooperative mechanism must be favored for dynamical relaxation .
note that the apparent mean energy of activation @xmath93 _ increases _ when the temperature decreases below the arrhenius regime .
this must mean that the single particle relaxation mechanism that is operative at the arrhenius regime is no longer available at lower temperatures , since if it were available it would have been selected .
in the temperature range where both mechanisms coexists we indeed find the associated @xmath28 and @xmath29 to be similar .
we thus offer a model of the observed phenomenology on the the following basis : \(i ) in the arrhenius regime the relaxation process is an elementary event like a @xmath94 process which is characterized by a relatively sharp free energy cost @xmath95 , @xmath96 where @xmath97 is the number of available inequivalent @xmath0 processes .
this energy cost @xmath39 is of the order of breaking through a local cage and the process involves only a few particles . due to the amorphous nature of the super - cooled liquid @xmath39 is a stochastic variable but the dispersion around @xmath39 is relatively small .
we denoted the variance in the distribution of @xmath39 as @xmath98 and noted that in this regime @xmath99 .
\(iii ) the precise reason for the blocking of this elementary relaxation channel when temperature is lowered is not known .
we will propose , however , that this blocking is due to the disappearance of available free volume per particle @xmath100 @xcite .
in other words , denoting by @xmath101 the minimal average volume that every particle occupies , there must be additional free volume @xmath102 for the @xmath94 process to occur . of course , the average volume per particle , which is the inverse of the density , @xmath103 , is determined by the equation of state which varies from one material to the other .
we will define @xmath104 as the temperature where the free volume disappears , and then @xmath105 . for the purposes of the present model we estimate the average volume per particle from the first order taylor expansion : @xmath106 where the partial derivative is computed at @xmath107 .
accordingly , since @xmath105 , @xmath108 the probability of having a local free volume @xmath109 is estimated from maximum entropy considerations as @xmath110 consequently the probability that @xmath111 , which is the probability that a @xmath0 process could occur , is @xmath112 finally , we write @xmath113 , where @xmath3 is the isobaric coefficient of thermal expansion at @xmath107 , @xmath114 with this we can write @xmath115 } \ , \label{probt1}\ ] ] where the dimensionless constant @xmath116 .
note that we can immediately estimate the probability of seeing the cooperative mechanism ( which we will refer to as the @xmath117 process ) as simply @xmath118}$ ] .
therefore we can estimate @xmath47 where the two processes are equiprobable from @xmath119 } \approx 1/2 \ , \ ] ] and therefore @xmath120 .
\(iv ) in the fragile regime we propose that eq .
( [ f1 ] ) retains its form , but now with a temperature dependent @xmath121 and a sizeable degeneracy @xmath122 .
our picture is that of a relaxation process that involves more and more cooperative motions of a larger and larger number of particles .
this obviously results in an increasing energy barrier @xmath121 which is indeed higher than @xmath39 but selected because the @xmath94 process is no longer available or it requires cooperation from the surroundings to be realized .
also an increase in the number of particles involved must increase @xmath123 simply due to the number of ways that these particles can be chosen .
thus we write in the fragile regime we now consider the emergence of the @xmath117 process as a consequence of the disappearing of the @xmath94 process as is presented for example in fig .
[ fullt - lennard ] .
we write @xmath125 df \ .\end{aligned}\ ] ] here @xmath126 and @xmath127 are the weights of the two relaxation channels respectively . in the arrhenius regime @xmath126
is composed of three factors ; remembering that @xmath128 , we write in that regime @xmath129}\times \frac{1}{\sqrt{2\pi\sigma^2}}e^{\frac{-(f - e_a ) ^2}{2\sigma^2}}\times e^{-f / t } \ . \label{w1ar}\ ] ] the first factor is the probability to have a @xmath94 process , eq . [ probt1 ] .
the second factor is the probability to select a free energy barrier of magnitude @xmath12 .
the third factor is the probability to overcome that particular barrier . in the larger temperature regime that includes the pre- and post - arrhenius regime we generalize eq.([w1ar ] ) to read @xmath129}\times \frac{1}{\sqrt{2\pi\sigma_1 ^ 2(t)}}e^{\frac{-[f-\langle f_1\rangle(t ) ]
^2}{2\sigma_1 ^ 2(t)}}\times e^{-f / t } \ . \label{w1}\ ] ] where @xmath130 .
this fit is supported by the scaling function shown in fig .
[ fig9 ] in the left lower panel , and @xmath131 is the temperature of the minimum of scaling function . in fig .
[ model1 ] we present the prediction of eq.([model ] ) as a function of temperature .
we chose the various parameters to agree with the functions shown in fig .
[ fig9 ] with @xmath133 and @xmath134 .
the model indeed reproduces very well the main observations discussed above .
we see the pre - arrhenius behavior and its transition to arrhenius behavior where expected .
we also observe the rapid decline of the arrehnius gaussian with the concurrent increase of the fragile gaussian which is marching to the right with increasing dispersion .
we should stress that this simple model is not meant to represent the full complexity of the observed processes , especially at low temperatures where the cooperative dynamics reigns supreme ; the point of this simple model is to stress the fundamental observation of this paper , i.e. that a compact and well defined relaxation mechanism appears to operate in the arrhenius regime , whereas this simple mechanism is blocked once temperature reduces , giving rise to another mechanism which is presumably cooperative and much more involved .
for all that we know one may encounter additional relaxation processes at still lower temperatures which are currently not available to numerical simulations . for a more systematic length scale which was derived as a nontrivial typical system size that separates when the lowest eigenvalue of the hessian matrix is determined by elastic modes and when by plastic modes see s. karmakar , e. lerner , i. procaccia , physica a , * 391 * , 1001 ( 2012 ) | generic glass formers exhibit at least two characteristic changes in their relaxation behavior , first to an arrhenius - type relaxation at some characteristic temperature , and then at a lower characteristic temperature to a super - arrhenius ( fragile ) behavior .
we address these transitions by studying the statistics of free energy barriers for different systems at different temperatures and space dimensions .
we present a clear evidence for changes in the dynamical behavior at the transition to arrhenius and then to a super - arrhenius behavior .
a simple model is presented , based on the idea of competition between single - particle and cooperative dynamics .
we argue that arrhenius behavior can take place as long as there is enough free volume for the completion of a simple @xmath0 relaxation process .
once free volume is absent one needs a cooperative mechanism to ` collect ' enough free volume .
we show that this model captures all the qualitative behavior observed in simulations throughout the considered temperature range . |
from recent cosmological observations , inflation appears to be the most successful cosmological model giving the large scale structure of the universe that existed in the very early universe .
current observations of the temperature anisotropy of the cosmic microwave background ( cmb ) suggest that an inflationary universe is consistent with the present universe .
although it is not always necessary , a typical slow - roll condition is an important ingredient for the inflationary scenario .
considering slow - roll inflation , there are two possible dynamics scenarios for velocity damping : the original ( supercooled ) inflation scenario in which the friction is induced by the hubble parameter , and the so - called warm inflation scenario in which a strong diffusion produces significant friction for the inflaton motion@xcite . in the strongly dissipative warm inflation ( sdwi )
scenario , dissipative effects dominate the friction term . if the dissipation is not so strong as to dominate the friction term while the radiation production occurs concurrently during the inflationary period@xcite , this scenario is termed the weakly dissipative warm inflation ( wdwi ) scenario .
the conventional criteria for discriminating warm inflation from supercooled inflation is not @xmath0 in the friction term but @xmath1 for the temperature during inflation , where @xmath2 denote the damping rate , hubble parameter and temperature . in this paper
we consider the significant thermal effect of the radiation created during warm inflation , for both sdwi ( @xmath0 ) and wdwi ( @xmath3 ) scenarios .
two different sources are considered for the radiation : one is supplied continuously from the inflaton field , and the other is supplied temporarily from the phase transition and the decay in the remote sector .
originally , the spectrum of the cosmological perturbations created during inflation has been expected to be scale - invariant and gaussian .
however , recent observations may suggest that small anomalies are significant in the spectrum , such as a small departure from exact scale - invariance and a certain non - gaussian character @xcite , both of which can help reveal the underlying theory based on the expected form of the inflationary action .
in particular , observation of a tiny shift in the spectrum index @xmath4 is an obvious example@xcite of scale invariance violation that can be used to determine the strict form of the inflaton potential for the inflation scenario .
observation of non - gaussian character in the spectrum may lead to a significant bound for the mechanism of generating curvature perturbations . determining how these anomalies arise and how they are related to the inflation dynamics and also to the generating mechanism of the cosmological perturbations
will depend on the cosmological model and the character of the underlying theory . in order to explain how these anomalies are explained , there are already many models of inflation and the generating mechanism of the curvature perturbations , in which the spectrum is generated ( 1 ) during inflation @xcite , ( 2 ) at the end of inflation @xcite , ( 3 ) after inflation by preheating@xcite and reheating@xcite or ( 4 ) late after inflation by curvatons @xcite or by inhomogeneous phase transition@xcite
the idea of inhomogeneous phase transition is very useful and unifies inhomogeneous scenarios , as it explains density perturbations from a spatially inhomogeneous transition accompanied by a gap in the scaling relation ; namely @xmath5 with @xmath6 , where @xmath7 denotes the cosmic - scale factor .
for example , @xmath8 and @xmath9 reproduces inhomogeneous reheating@xcite and creation of perturbations at the end of inflation@xcite , respectively . in this paper
we consider ( 1 ) and ( 2 ) , namely the creation of the curvature perturbations during inflation and at the end of inflation .
however , contrary to the conventional model , the inhomogeneities at the end of inflation are not induced by the fluctuations of the end - boundary of the inflaton field ( @xmath10 ) but by the inhomogeneous phase transition in the remote sector based on spatial inhomogeneities of the critical temperature ( @xmath11 ) .
recently , the origin of the cosmological perturbations in warm inflation has been revealed in ref.@xcite . in this paper , we follow the discussions in ref.@xcite . in warm inflation ,
the effective potential of the inflaton fields @xmath12 depends on temperature @xmath13 , which can be expressed as @xmath14 .
the damping rate @xmath15 of the inflaton fields may also depend on @xmath13 and @xmath12 .
therefore , the `` trajectory '' of warm inflation is generally given by inflaton fields @xmath12 and @xmath13 .
there may be multiple trajectories for inflation , and the correct one can not be chosen independent of the initial condition that is generally given by @xmath12 and @xmath13 .
the end - boundary of inflation may also be determined by @xmath16 and @xmath13 .
the situation is similar to multi - field models of inflation . to compare the situations for multi - field inflation and warm inflation , first consider two - field inflation with a fast - rolling inflaton @xmath17 that reaches its minimum during inflation .
we assume that the slow - rolling inflaton @xmath16 couples to the trigger field in the hybrid - type potential and this field determines the end of inflation .
in this case , the fluctuation of the e - foldings number @xmath18 is expected to be given by @xmath19 and @xmath20 , which naturally leads to@xcite @xmath21 however , for multi - field inflation the usual definition of the curvature perturbations on spatially flat hypersurfaces is given by using @xmath22(density ) , @xmath23(pressure ) and @xmath24(momentum perturbation ) . following the standard definition of the curvature perturbations ,
the curvature perturbations are found to be given by @xmath25 where @xmath26 on the steep potential is assumed . in this case , the evolution during inflation ( @xmath27 ) is crucial .
namely , at a `` bend '' in the inflation trajectory , where the fast - rolling field @xmath17 reaches its minimum , the correction caused by the multi - field evolution is given by @xmath28 which is consistent with the naive expectation given in eq.([bai - ve ] ) .
see also ref.@xcite for an alternative approach .
a similar situation appears in strongly dissipating warm inflation , in which the curvature perturbation is given by formula based on @xmath19 and @xmath20 . ]
@xmath29 again , the evolution during inflation is crucial , since from the time - integral of a decaying component we find @xcite , @xmath30 one might think that the result is trivial when the number of e - foldings is determined exclusively by the inflaton .
however , just as in multi - field inflation where the significant evolution of the curvature perturbation is important , in warm inflation the evolution of the curvature perturbation is very important . ) . for simplicity ,
the correction from the perturbations of the radiation is not considered in this paper , although it is important in warm inflation . ] from the langevin equation , the root - mean square fluctuation amplitude of the inflaton field @xmath19 after the freeze out is obtained to be@xcite @xmath31 where @xmath32 denotes the ratio between @xmath13 and @xmath33 , defined by @xmath34 , which is assumed to be @xmath35 during warm inflation .
the strength of the thermal damping is measured by the rate @xmath36 given by @xmath37 which can be used to rewrite the field equation as @xmath38 for sdwi , the curvature perturbations created during warm inflation are thus determined by the equation @xmath39 for wdwi , it is given by @xmath40 here we used @xmath41 , which is explained in the next section .
an interesting property is expected from the above equations .
for remote inflation , one may naturally expect that many fields are in thermal equilibrium and stay in their false vacuum states .
if there is a cascading symmetry breaking in the remote sector as @xmath13 evolves during inflation , it raises a gap in the hubble parameter .
this may cause a transition from wdwi to sdwi , or simply a gap in the parameters @xmath36 and @xmath32 . assuming instant decay in the remote sector
, @xmath13 may have a gap ( sudden increase ) at the transition .
this may happen if the energy released by the transition ( @xmath42 ) raises the effective temperature @xmath13 for @xmath43 .
otherwise , one may assume that the change of the temperature @xmath13 is negligible because of the slow - decay , and @xmath44 is a constant at the transition .
even if both @xmath13 and @xmath44 are constant at the transition , a gap in the curvature perturbation may appear from the gap in the hubble parameter .
we investigate these interesting possibilities in this paper .
more significantly , the hubble parameter may decrease with a large ratio @xmath45 if the major part of the remote sector decays .
then , the increase of @xmath36 and @xmath32 caused by the shift of the hubble parameter leads to the amplification ( or suppression ) of the scalar perturbation . for example
, we consider the simplest case in which the radiation released from the remote sector is not significant due to the slow - decay , but the shift of the hubble parameter is not negligible . for wdwi ( @xmath46 ) ,
the curvature perturbation is given by @xmath47 for sdwi ( @xmath48 ) it is given by @xmath49 these equations show that the curvature perturbation is suppressed at the gap .
note that a transition from wdwi to sdwi is also possible at the gap .
contrary to scalar perturbations , tensor perturbations do not couple strongly to the thermal background .
gravitational waves are thus generated by quantum fluctuations as in standard ( supercooled ) inflation@xcite .
we thus find @xmath50 which decreases at the gap . here
the `` gap '' is not always assumed to be large .
it may appear as a few percent of @xmath13 and @xmath33 to create a mild scale - dependence .
for the second example , we consider a gap characterized by @xmath51 with the explicit form of the damping rate given by @xmath52 .
if a small part of the false - vacuum energy is released by the phase transition caused by a field in the multi - field remote sector , the gap in the hubble parameter is negligible but the gap in the temperature can be significant . since the damping rate depends explicitly on the temperature , the gap is accompanied by an increase in the damping rate . using @xmath53 , there is a change in the curvature perturbations @xmath54 which may lead to a significant amplification of the curvature perturbations at the corresponding cosmological scale .
the amplification , if it exists in the observable scale , may appear as a temporal peak in the spectrum .
the tail of the peak is determined by the inflation attractor .
finally , we note that the gap in the temperature caused by the phase transition in the remote sector may * appear even if the energy density during inflation is not dominated by the remote sector . * in this sense , the gap in the temperature is a common and very important phenomenon in warm inflation , which requires further study .
details of these topics are discussed in sec.2 .
in conventional hybrid inflation , the typical form of the hybrid - type potential is given by @xmath55 the vacuum energy during inflation is @xmath56 , which is defined by the false vacuum @xmath57 and @xmath58 . here
the critical point is defined by @xmath59 because of the condition @xmath58 , the motion of the inflaton field must obey `` roll - in '' motion @xmath60 .
another possibility for realizing hybrid inflation is to consider `` roll - out '' motion of the inflaton field considering a non - trivial form of the couplings .
the scenario that realizes `` roll - out '' motion is called the inverted scenario of hybrid inflation@xcite .
the conditions related to the roll - in and roll - out motion are usually used to restrict inflaton potential .
therefore , removing the conditions for hybrid inflation related to `` roll - in '' and `` roll - out '' is useful .
a natural construction of the hybrid - type potential in the string theory is to consider brane distance as the inflaton .
the trigger field may be a tachyon that appears in the collision of a brane anti - brane pair@xcite , or a field localized on the target wall@xcite .
a thermal effect can cause an attractive force between branes , which can be used to construct thermal brane inflation@xcite , and to construct models of brane inflation based on defect inflation scenario@xcite and brane inflation with roll - out motion @xcite .
although brane inflation is not discussed explicitly in this paper , application of the remote inflation scenario to brane inflation is an attractive idea .
contrary to conventional hybrid inflation , either normal , inverted or brane - motivated , remote inflation does not have direct couplings between inflaton and trigger fields .
namely , remote inflation is a kind of hybrid - like inflation but it does not have the so - called hybrid - type potential .
the trigger fields @xmath61 are placed in a remote sector and stays in their false vacuum states during inflation , but they do not have direct couplings with the inflaton field @xmath16 . these trigger fields can stay in their false vacuum states because of the thermal effect from the background radiation . the background radiation is negligible in supercooled inflation , but it may have a significant effect in warm ( @xmath62 ) inflation .
a typical situation is shown in figure 1 , in which either roll - in or roll - out potential for inflaton are considered . ( 200,300)(100,170 ) [ fig : mass - dom ] in order to show how slow - roll inflation is realized in the warm inflation scenario , we first consider a homogeneous inflaton field interacting with thermal radiation@xcite .
the thermodynamic potential is given by @xmath63 where @xmath64 is the effective number of thermal particles .
the evolution equation for the inflaton field is given by @xmath65 where @xmath44 is the damping term and @xmath33 is the expansion rate of the universe . here
the subscript denotes the derivative with respect to the field .
the strength of the thermal damping is measured by the rate @xmath36 given by @xmath66 , which can be used to rewrite the field equation as @xmath38 the strongly dissipating phase of warm inflaton is defined by @xmath67 .
the production of entropy during inflation is caused by the dissipation of the inflaton motion , where the entropy @xmath68 is defined by the thermodynamic equation ; @xmath69 where the subscript denotes the derivative with respect to the temperature . for our calculation
, we need explicit forms of the energy density @xmath22 and the pressure @xmath70 , which are given by @xmath71 where @xmath72 is the contribution from the kinetic term of the inflaton field , and the friedman equation follows @xmath73 here we assumed that inflaton has the standard kinetic term .
the slow - roll approximation is very useful in estimating the order of magnitude of the physical quantities : @xmath74 where the second equation is obtained from the evolution equation for the radiation energy density .
the slow - roll trajectory of inflation for @xmath16 and @xmath13 is determined by eqs.([slou ] ) and ( [ rel - warm ] ) .
the gauge - invariant combinations for the curvature perturbation are defined by @xmath75 where @xmath76 is the momentum perturbation satisfying @xmath77 here @xmath78 is the perturbation of the comoving density , satisfying the evolution equation @xmath79 where @xmath80 is related to the shear perturbation and assumed to be finite . from the above definitions ,
@xmath81 hypersurfaces are identified with uniform density hypersurfaces ( @xmath82 ) at large scales due to the relation @xmath83 , where these two definitions of the curvature perturbation are identified ( @xmath84 ) .
we considered linear scalar perturbations of a friedman - robertson - walker(frw ) background : @xmath85dx^i dx^j.\ ] ] using the above definitions based on the energy - momentum , we find that the comoving curvature perturbation created during warm inflaton on spatially flat hypersurfaces @xmath86 is expressed as @xmath87 the result is obviously much smaller than the result expected from the perturbation of the number of e - foldings based on @xmath19 and @xmath20 , which is given by @xmath88 as we discussed in sec.1 , a similar situation appears for multi - field inflation , and the solution to this problem is given by the evolution of the curvature perturbations .
namely , for multi - field inflation a significant evolution at the `` bend '' in the trajectory leads to @xmath89 which recovers the expected form of the curvature perturbations based on @xmath19 and @xmath20 . for warm inflation
, a significant evolution appears from the time - integral of a decaying component in the evolution equation .
again , it leads to the result consistent with the expectation based on @xmath19 and @xmath20 . in this paper
, we thus consider that the curvature perturbations created during warm inflation are given by @xmath90 which is used without mentioning the origin .
an important source of density perturbations in warm inflation is thermal fluctuations , which causes significant difference from supercooled inflation .
namely , a comoving mode of thermal fluctuations during warm inflation is created by thermal noise , which can lead to significant field perturbations @xmath91 .
the behavior of a scalar field interacting with radiation can be studied by using the langevin equation @xmath92 where @xmath93 denotes a source of stochastic noise . from the langevin equation , the root - mean square fluctuation amplitude of the inflaton field @xmath19 after the freeze is obtained .
it is given by a simplified form @xmath94 for the typical ( @xmath95 ) warm inflation , and @xmath96 for weakly dissipating ( @xmath97 ) warm inflation .
using the relation given in eq,([rel - warm ] ) , the scalar curvature perturbation is expressed as @xmath98 cosmological perturbations created during remote inflation are calculated using these basic equations of warm inflation .
remote inflation requires symmetry restoration in the remote sector , which is caused by the radiation dissipated from the inflaton . to illustrate some typical features of finite temperature effects in the remote sector , here we consider a real scalar field and a potential : @xmath99 where @xmath100 is tuned so that the cosmological constant vanishes at the true minimum .
the phenomenon of high - temperature symmetry restoration can be understood by the finite - temperature effective potential given by @xcite @xmath101,\ ] ] where @xmath102 is the one - loop potential for zero - temperature with the classical field @xmath103 : @xmath104 where @xmath105 is a renormalization mass scale . at high temperatures
, @xmath106 can be expanded near the symmetric point ( @xmath107 ) as @xmath108 which suggests that the temperature - corrected effective mass at @xmath107 changes sign at the critical temperature given by @xmath109 more generally , one may introduce a background thermal bath which couples to @xmath110 .
assuming that the couplings of @xmath110 to the fields in the background thermal bath are more significant than the self - coupling , we obtain a typical form of the potential with a thermal correction term , which is given by @xmath111 where @xmath112 denotes the effective coupling of @xmath110 to the fields in the thermal bath . in this case , the critical temperature is given by @xmath113 in this paper , we assume that the couplings to the background fields are significant and @xmath114 is always given by @xmath115 .
we define the critical temperature for the fields in the remote sector by @xmath116 . since the false - vacuum energy in the remote sector determines the energy scale of inflation , @xmath100 in eq.([ini - the ] )
must be determined explicitly .
the simplest form of the remote sector would be given by a single - field quartic potential ; @xmath117 where the false - vacuum state is stabilized for @xmath118 . for the single - field remote sector model ,
we find @xmath119 since the false - vacuum energy @xmath100 in the remote sector must dominate the energy density during warm inflation , we have a condition @xmath120 , which leads to a simple equation @xmath121 it is also interesting to consider if a phase transition in the remote sector occurs during a warm inflationary phase characterized by @xmath62 .
the condition for the phase transition is given by @xmath122 , which leads to @xmath123 then , a condition @xmath124 is obtained . for @xmath125
, the phase transition may occur in supercooled inflation .
if the radiation is not supplied concurrently from the inflaton , the model becomes thermal inflation@xcite .
more generally , we may consider an effective symmetry in the remote sector , which forbids ( or reduces ) the quartic interaction . in this case
the potential is given by @xmath126 the condition for the false - vacuum domination @xmath120 leads to @xmath127 again , from @xmath128 , we obtain an obvious condition for the phase transition ; @xmath129 in standard inflation , the end of inflation is determined by @xmath130 , where the slow - roll condition is violated at @xmath131 .
the end - boundary of remote inflation is determined by either @xmath130 or @xmath114 .
if @xmath16 reaches @xmath130 ahead of @xmath13 , the situation is precisely the same as the conventional warm inflation . however , if @xmath13 reaches @xmath114 ahead of @xmath16 , @xmath114 in the remote sector determines the end of inflation .
also , if there is an inhomogeneous phase transition@xcite in the remote sector , it causes an inhomogeneous end of warm inflation .
contrary to the usual scenario of generating curvature perturbations at the end of inflation@xcite , here a model is considered in which the inhomogeneities are not caused by the inflaton - field perturbations at the end ( @xmath132 ) , but by the spatial inhomogeneities of the critical temperature ( @xmath133 ) in the remote sector . here
the sources of the inhomogeneities are light fields ( moduli ) in the model , which determines the value of @xmath112 and @xmath134 . from the definition of the critical temperature
, we find @xmath135 which does not vanish if @xmath134 or @xmath112 are determined by the vacuum expectation values of light fields , which obtain cosmological fluctuations at the horizon exit .
the effect can be used to add significant non - gaussianity to the spectrum , but to find the explicit form of the curvature perturbations we must solve @xmath136 at the end of warm inflation .
since the calculation is highly model - dependent and requires numerical study , we do not discuss details of this issue in this paper .
however , from a heuristic argument we can estimate the magnitude of the correction by @xmath137 \sim \frac{\phi}{\dot{\phi}}\left[\frac{\delta t_c}{t_c}\right ] \sim \frac{\phi}{\delta \phi}\left[\frac{\delta \phi}{\dot{\phi}}\right ] \left[\frac{\delta t_c}{t_c}\right],\ ] ] which can be large .
here we assumed that @xmath13 is given by a polynomial of @xmath16 , assuming that @xmath138 and @xmath139 with the inflaton potential given by @xmath140 .
the creation of the curvature perturbations caused by the inhomogeneous phase transition in the remote sector is thus important for the calculation of the curvature perturbations and its anomalies in the spectrum .
the essence of remote inflation is discussed in the previous section considering the typical situation with single - field remote sector . at the horizon exit ,
the single - field model discussed above creates the same spectrum as the conventional warm inflation .
however , for the creation of the cosmological perturbations at the inhomogeneous end of inflation , the inhomogeneous phase transition in the remote sector may leave a significant signature in the spectrum . therefore , the creation of the perturbations at the end of inflation is important in distinguishing the model . as we will show in this subsection , the significance of remote inflation is more apparent in its multi - field extension , where the gaps caused by the phase transitions in the remote sector may leave significant signatures in the spectrum . in general , generation of gaps and peaks in the spectrum
is difficult in conventional inflation .
we thus believe that observations of these signatures are important clues to remote inflation .
more precisely , in standard inflation , gap generation is caused by non - trivial interactions that become significant at a critical point @xmath141 .
otherwise , the gap in the spectrum is explained by the gap in the inflaton potential .
these mechanisms are highly model - dependent and require many additional sources for the critical point .
therefore , we believe that our idea , in which the gap is caused by the critical temperature , is more attractive than the usual scenarios . the effect from the phase transition in the multi - field remote sector can be categorized in the following ways .
1 . @xmath142 and @xmath143 , where the major component in the remote sector decays and supplies significant radiation .
2 . @xmath144 and @xmath145 , where the major component in the remote sector decays but due to the slow - decay it does not raise the temperature .
3 . @xmath142 and @xmath146 , where a minor component in the remote sector decays and quickly thermalizes during inflation .
note that the vacuum - energy domination by the remote sector is not necessary for the situation .
( i.e. , the situation is based on a small phase transition in a remote sector , which can appear in conventional warm inflation . )
the conditions for the situations and the expected corrections to the curvature perturbations are discussed in the followings , using an explicit form of @xmath44 : @xmath147 . in this case , the component in the remote sector that decays at the phase transition has dominated the false - vacuum energy until the phase transition .
we consider the situation in which the peak of the temperature ( @xmath148 ) caused by the phase transition exceeds the temperature before the phase transition , but the decay is not so fast as to dominate the energy density of the universe .
denoting the cosmological quantities before and after the phase transition by the subscripts @xmath149 and @xmath150 , the condition is given by @xmath151 . because of the gap in the hubble parameter , the ratios @xmath34 and @xmath152 increase after the gap .
moreover , the supplement of the radiation from the remote sector enhances @xmath153 and @xmath53 just after the gap .
here we consider the damping rate @xmath44 that is defined by @xmath154 where @xmath155 is the constant determined by the number of fields and the channels for the decay@xcite . for simplicity , we introduce rates defined by @xmath156 we thus find @xmath157 which lead to the ratio of the curvature perturbations @xmath158 which show that in wdwi the gap always causes suppression of the curvature perturbation , while in sdwi amplification / suppression is determined by the model .
this typical characteristic feature can be used to distinguish sdwi from wdwi .
if the enhancement of @xmath44 is significant , the situation with @xmath159 is possible . the instant transition from wdwi to sdwi leads to @xmath160 again , the gap may cause either suppression or amplification . in this case
, the component in the remote sector that decays at the phase transition has dominated the false - vacuum energy , but the creation of the radiation from the phase transition is very slow and thus the temperature is not raised significantly after the phase transition .
therefore , we find @xmath161 and @xmath162 in this situation .
we thus find @xmath163 which lead to the curvature perturbations @xmath164 which show that in both cases the gap causes suppression of the curvature perturbation .
if the enhancement of @xmath44 is significant , the situation with @xmath159 is possible .
the transition from wdwi to sdwi leads to @xmath165 in this case , the component in the remote sector that decays at the phase transition has not dominated the false - vacuum energy . the same situation can be considered in standard warm inflation . here
we assume for simplicity that the creation of the radiation from the phase transition is rapid and the radiation is quickly raised after the phase transition .
we consider the situation in which the peak of the temperature ( @xmath148 ) caused by the phase transition exceeds the temperature before the phase transition . because there is no significant gap in the hubble parameter
, the ratio @xmath166 is given by @xmath167 and can be neglected .
the creation of the radiation from the remote sector enhances @xmath153 and @xmath53 just after the gap , which leads to @xmath168 .
we thus find @xmath169 which lead to the curvature perturbations @xmath170 which show that for both wdwi and sdwi the gap always causes significant amplification of the curvature perturbation .
if the enhancement of @xmath44 is significant , the situation with @xmath159 is possible .
the instant transition from wdwi to sdwi leads to @xmath171 from the evolution equation for the radiation energy density , the radiation concurrently created from the inflaton is given by @xmath172 assuming that @xmath36 is given by the standard form @xmath173 , the above equation gives @xmath174 which show that the temporal increase in the radiation energy density caused by the decay in the remote sector leads to ( 1 ) enhancement of @xmath175 in wdwi or ( 2 ) suppression of @xmath175 in sdwi . as a result ,
the tail of the gap ( the typical time - length of the diffusion ) in wdwi is typically longer than in sdwi . considering the attractor of the trajectory @xmath176 from the equations of warm inflation given in this section , finding two independent solutions is possible , although it depends on the details of the inflaton potential .
the transition between different trajectories is usually disregarded because it is quite unlikely in conventional warm inflation . however , in remote inflation the transition from a low - temperature trajectory to a high - temperature trajectory may occur after the phase transition in the remote sector .
the phase transition in the remote sector may also cause significant change in the hubble parameter that may also lead to the change of the trajectory . besides the complexities of the decaying process in the remote sector , production of cosmological defects during inflation may leave significant signatures in the spectrum@xcite .
an interesting possibility that may arise in the context of an inflationary scenario@xcite is that monopoles are formed during inflation but are not completely inflated away .
then , at the confinement phase transition after inflation , strings connecting monopoles ( or necklaces ) are formed , whose length is initially much longer than the hubble radius .
such strings may leave significant cosmological signatures@xcite , such as dark matter and primordial black holes . besides strings and monopoles and their hybrids@xcite , cosmological domain walls
may also appear when a discrete symmetry is broken at the phase transition .
the creation of cosmological domain walls may cause serious cosmological problem .
cosmological domain walls can decay safely before domination if they are made unstable by a bias between the quasi - degenerated vacua , which can be induced by an effective interaction term that breaks the discrete symmetry .
note that for supergravity , domain walls caused by the r - symmetry are naturally safe because the supergravity interaction causes the required bias @xcite .
however , the instability caused by the explicit breaking term may not work effectively to erase these domain walls before domination , because the typical distance between domain walls created during inflation is much larger than the hubble radius .
namely , the distance between domain walls may still be larger than the hubble radius at the time when they must decay for successful cosmology .
this condition puts a trivial bound on the cosmological time of the domain - wall creation during inflation . on the other hand ,
if such domain walls are generated sufficiently late during the inflationary epoch and decay safely before domination , they can be used to create significant amplification of the small - scale density perturbations .
these topological defects ( monopoles , strings , domain walls and others@xcite ) may put bounds on the phase transition , but at the same time they can be used to create dark matter candidates , primordial black holes and small - scale density perturbations .
in this paper we proposed a new scenario of hybrid - like inflation .
in remote inflation , radiation raised continuously by a dissipating inflaton field causes symmetry restoration in a remote sector , and the false - vacuum energy of the remote sector dominates the energy density during inflation .
the situation is similar to hybrid inflation , but the hybrid - type potential is not required in this scenario .
remote inflation is terminated when the temperature reaches the critical temperature , or when the slow - roll condition is violated .
of course , the scenario is identical to thermal inflation if the radiation is not supplied by the inflaton . without introducing additional couplings , the inflaton field may roll - out like an inverted - hybrid model or quintessential inflation .
significant signatures of remote inflation can be observed in the spectrum . for a single - field remote sector ,
the difference is caused by the inhomogeneous phase transition in the remote sector .
for multi - field extension of the remote sector , successive phase transition is possible and it may cause significant difference in the spectrum .
remote inflation can predict strong amplification or suppression of small - scale perturbations , which may appear without introducing multiple inflation . note that the phase transition in the remote sector is possible during standard warm inflaton , irrespective of the domination by the false - vacuum energy in the remote sector . the inflaton field may have a run - away potential , which can be identified with quintessential potential . in this sense ,
the idea of remote inflation also provides a hybrid - type extension of quintessential inflation . for definiteness , we consider a concrete model of single field inflation with quartic potential and calculate the curvature perturbations and the spectral index of the model .
note first that the spectrum of remote inflation is basically the same as the warm inflation with the hybrid - type potential .
the potential of the model is given by @xmath177 where @xmath100 is the false - vacuum energy of the remote sector .
here we consider a flat potential with @xmath178 .
we consider the damping rate @xmath44 ; @xmath179 for strongly dissipating phase , the slow - roll velocity is given by @xmath180 which leads to the temperature during warm inflation @xmath181 here @xmath182 is the stefan - boltzmann constant .
a useful expression is @xmath183 from these equations , the curvature perturbation is obtained as @xmath184 obviously , a tiny @xmath185 is required for warm inflation .
note that remote inflation works without the hybrid - type potential and may solve problems in supercooled inflation , however it can not be free of the conditions related to warm inflation . in ref.@xcite ,
the spectral index of the curvature perturbations in warm inflation has been studied for general form of the dissipative coefficient .
the straightforward calculation of the spectral index using the useful expression in ref.@xcite gives @xmath186\sim\frac{3}{7}\frac{\eta}{r},\ ] ] where @xmath187 is given by @xmath188 the running of the spectral index for this model gives @xmath189@xcite .
is allowed for the running of the spectral index @xmath190@xcite . ] a more interesting and realistic application of remote inflation will be discussed in ref.@xcite .
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 ] ] .
a. hosoya and m. a. sakagami , `` time development of higgs field at finite temperature , '' phys .
d * 29 * , 2228 ( 1984 ) .
n. bartolo , e. komatsu , s. matarrese and a. riotto , `` non - gaussianity from inflation : theory and observations , '' phys .
rept . * 402 * , 103 ( 2004 ) [ arxiv : astro - ph/0406398 ] .
s. yadav and b. d. wandelt , `` evidence of primordial non - gaussianity @xmath191 in the wilkinson microwave anisotropy probe 3-year data at 2.8@xmath110 , '' phys .
lett . * 100 * , 181301 ( 2008 ) [ arxiv:0712.1148 ] .
a. r. liddle and d. h. lyth , `` cosmological inflation and large - scale structure , '' _ cambridge , uk : univ .
( 2000 ) 400 p_. t. matsuda , `` delta - n formalism for the evolution of the curvature perturbations in generalized multi - field inflation , '' arxiv:0906.2525 [ hep - th ] ; t. matsuda , `` evolution of curvature perturbation in generalized gravity theories , '' arxiv:0906.0643 [ hep - th ] ; t. matsuda , `` modulated inflation , '' phys .
b * 665 * , 338 ( 2008 ) [ arxiv:0801.2648 [ hep - ph ] ] ; t. matsuda , `` modulated inflation from kinetic term , '' jcap * 0805 * , 022 ( 2008 ) [ arxiv:0804.3268 [ hep - th ] ] ; t. matsuda , `` running spectral index from shooting - star moduli , '' jhep * 0802 * , 099 ( 2008 ) [ arxiv:0802.3573 [ hep - th ] ] ; t. matsuda , `` successful d - term inflation with moduli , '' phys .
b * 423 * , 35 ( 1998 ) [ arxiv : hep - ph/9705448 ] .
l. kofman , `` probing string theory with modulated cosmological fluctuations , '' arxiv : astro - ph/0303614 .
d. wands , k. a. malik , d. h. lyth and a. r. liddle , `` a new approach to the evolution of cosmological perturbations on largescales , '' phys .
d * 62 * , 043527 ( 2000 ) [ arxiv : astro - ph/0003278 ] ; d. h. lyth and d. wands , `` conserved cosmological perturbations , '' phys . rev .
d * 68 * , 103515 ( 2003 ) [ arxiv : astro - ph/0306498 ] ; c. gordon , d. wands , b. a. bassett and r. maartens , `` adiabatic and entropy perturbations from inflation , '' phys .
d * 63 * , 023506 ( 2001 ) [ arxiv : astro - ph/0009131 ] .
t. matsuda , `` curvatons and inhomogeneous scenarios with deviation from slow - roll , '' jcap * 0812 * ( 2008 ) 001 [ arxiv:0811.1318 [ hep - ph ] ] .
a. c. vincent and j. m. cline , `` curvature spectra and nongaussianities in the roulette inflation model , '' jhep * 0810 * , 093 ( 2008 ) [ arxiv:0809.2982 [ astro - ph ] ] .
j. r. bond , l. kofman , s. prokushkin and p. m. vaudrevange , `` roulette inflation with kaehler moduli and their axions , '' phys .
d * 75 * , 123511 ( 2007 ) [ arxiv : hep - th/0612197 ] .
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 ] .
d. h. lyth and a. riotto , `` generating the curvature perturbation at the end of inflation in string theory , '' phys .
rev . lett .
* 97 * , 121301 ( 2006 ) [ arxiv : astro - ph/0607326 ] ; 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 , `` non - standard kinetic term as a natural source of non - gaussianity , '' jhep * 0810 * , 089 ( 2008 ) arxiv:0810.3291 [ hep - ph ] . t. matsuda , `` brane inflation without slow - roll , '' jhep * 0703 * , 096 ( 2007 ) [ arxiv : astro - ph/0610402 ] .
g. dvali , a. gruzinov and m. zaldarriaga , `` a new mechanism for generating density perturbations from inflation , '' phys .
d * 69 * , 023505 ( 2004 ) [ arxiv : astro - ph/0303591 ] .
g. dvali , a. gruzinov and m. zaldarriaga , `` cosmological perturbations from inhomogeneous reheating , freezeout , and mass domination , '' phys .
d * 69 * , 083505 ( 2004 ) [ arxiv : astro - ph/0305548 ] .
k. enqvist , a. mazumdar and m. postma , `` challenges in generating density perturbations from a fluctuating inflaton coupling , '' phys . rev .
d * 67 * , 121303 ( 2003 ) [ arxiv : astro - ph/0304187 ] ; f. vernizzi , `` cosmological perturbations from varying masses and couplings , '' phys .
d * 69 * , 083526 ( 2004 ) [ arxiv : astro - ph/0311167 ] .
k. enqvist , a. jokinen , a. mazumdar , t. multamaki and a. vaihkonen , `` non - gaussianity from preheating , '' phys .
* 94 * , 161301 ( 2005 ) [ arxiv : astro - ph/0411394 ] ; k. enqvist , a. jokinen , a. mazumdar , t. multamaki and a. vaihkonen , `` non - gaussianity from instant and tachyonic preheating , '' jcap * 0503 * , 010 ( 2005 ) [ arxiv : hep - ph/0501076 ] ; a. jokinen and a. mazumdar , `` very large primordial non - gaussianity from multi - field : application to massless preheating , '' jcap * 0604 * , 003 ( 2006 ) [ arxiv : astro - ph/0512368 ] .
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 ] .
k. dimopoulos , g. lazarides , d. lyth and r. ruiz de austri , `` curvaton dynamics , '' phys .
d * 68 * , 123515 ( 2003 ) [ arxiv : hep - ph/0308015 ] ; q. g. huang , `` curvaton with polynomial potential , '' arxiv:0808.1793 [ hep - th ]
. t. matsuda , `` curvaton paradigm can accommodate multiple low inflation scales , '' class .
* 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 .
b * 659 * , 783 ( 2008 ) [ arxiv:0712.2103 [ hep - ph ] ] .
t. matsuda , `` cosmological perturbations from an inhomogeneous phase transition , '' arxiv:0902.4283 [ hep - ph ] .
t. matsuda , `` evolution of the curvature perturbations during warm inflation , '' arxiv:0905.0308 [ astro-ph.co ] .
a. n. taylor and a. berera , `` perturbation spectra in the warm inflationary scenario , '' phys .
d * 62 * , 083517 ( 2000 ) [ arxiv : astro - ph/0006077 ] .
d. h. lyth and e. d. stewart , `` more varieties of hybrid inflation , '' phys .
d * 54 * , 7186 ( 1996 ) [ arxiv : hep - ph/9606412 ] .
s. clesse and j. rocher , `` avoiding the blue spectrum and the fine - tuning of initial conditions in hybrid inflation , '' arxiv:0809.4355 [ hep - ph ] .
g. r. dvali and s. h. h. tye , `` brane inflation , '' phys .
b * 450 * , 72 ( 1999 ) [ arxiv : hep - ph/9812483 ] .
t. matsuda , `` non - tachyonic brane inflation , '' phys .
d * 67 * , 083519 ( 2003 ) [ arxiv : hep - ph/0302035 ] ; t. matsuda , `` f - term , d - term and hybrid brane inflation , '' jcap * 0311 * , 003 ( 2003 ) [ arxiv : hep - ph/0302078 ] ; t. matsuda , `` kaluza - klein modes in hybrid inflation , ''
d * 66 * , 107301 ( 2002 ) [ arxiv : hep - ph/0209214 ] .
g. r. dvali , `` infrared hierarchy , thermal brane inflation and superstrings as superheavy dark matter , '' phys .
b * 459 * , 489 ( 1999 ) [ arxiv : hep - ph/9905204 ] .
t. matsuda , `` topological hybrid inflation in brane world , '' jcap * 0306 * , 007 ( 2003 ) [ arxiv : hep - ph/0302204 ] ; t. matsuda , `` q ball inflation , '' phys .
d * 68 * , 127302 ( 2003 ) [ arxiv : hep - ph/0309339 ] .
t. matsuda , `` thermal hybrid inflation in brane world , '' phys .
d * 68 * , 047702 ( 2003 ) [ arxiv : hep - ph/0302253 ] ; t. matsuda , `` affleck - dine baryogenesis after thermal brane inflation , '' phys .
d * 65 * , 103501 ( 2002 ) [ arxiv : hep - ph/0202209 ] .
l. m. h. hall , i. g. moss and a. berera , `` scalar perturbation spectra from warm inflation , '' phys .
d * 69 * , 083525 ( 2004 ) [ arxiv : astro - ph/0305015 ] .
e. w. kolb and m. s. turner , `` the early universe , '' front .
* 69 * ( 1990 ) 1 .
m. bastero - gil and a. berera , `` warm inflation model building , '' arxiv:0902.0521 [ hep - ph ] .
r. jeannerot , j. rocher and m. sakellariadou , `` how generic is cosmic string formation in susy guts , '' phys .
d * 68 * , 103514 ( 2003 ) [ arxiv : hep - ph/0308134 ] .
a. vilenkin and e.p.s .
shellard , _ cosmic strings and other cosmological defects , ( cambridge university press , cambridge , 2000)_. t. matsuda , `` dark matter production from cosmic necklaces , '' jcap * 0604 * , 005 ( 2006 ) [ arxiv : hep - ph/0509064 ] ; t. matsuda , `` primordial black holes from monopoles connected by strings , '' astropart .
phys . *
30 * , 333 ( 2009 ) [ arxiv : hep - ph/0509061 ] ; t. matsuda , `` primordial black holes from cosmic necklaces , '' jhep * 0604 * , 017 ( 2006 ) [ arxiv : hep - ph/0509062 ] . t. matsuda , `` brane necklaces and brane coils , '' jhep * 0505 * , 015 ( 2005 ) [ arxiv : hep - ph/0412290 ] . t. matsuda , `` weak scale inflation and unstable domain walls , '' phys . lett .
b * 486 * , 300 ( 2000 ) [ arxiv : hep - ph/0002194 ] ; t. matsuda , `` on the cosmological domain wall problem in supersymmetric models , '' phys .
b * 436 * , 264 ( 1998 ) [ arxiv : hep - ph/9804409 ] .
m. eto , t. fujimori , t. nagashima , m. nitta , k. ohashi and n. sakai , `` dynamics of domain wall networks , '' phys .
d * 76 * , 125025 ( 2007 ) [ arxiv:0707.3267 [ hep - th ] ] ; m. eto , y. isozumi , m. nitta , k. ohashi and n. sakai , `` solitons in the higgs phase : the moduli matrix approach , '' j. phys .
a * 39 * , r315 ( 2006 ) [ arxiv : hep - th/0602170 ] .
y. zhang , `` warm inflation with a general form of the dissipative coefficient , '' jcap * 0903 * , 023 ( 2009 ) [ arxiv:0903.0685 [ hep - ph ] ] .
t. matsuda , `` remote inflation as hybrid - like sneutrino / mssm inflation , '' arxiv:0905.4328 [ hep - ph ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , `` five - year wilkinson microwave anisotropy probe ( wmap5 ) observations : cosmological interpretation , '' astrophys . j. suppl . * 180 * , 330 ( 2009 ) [ arxiv:0803.0547 [ astro - ph ] ] . | a new scenario of hybrid - like inflation is considered without using hybrid - type potential .
radiation raised continuously by a dissipating inflaton field keeps symmetry restoration in a remote sector , and the false - vacuum energy of the remote sector dominates the energy density during inflation .
remote inflation is terminated when the temperature reaches the critical temperature , or when the slow - roll condition is violated . without introducing a complex form of couplings , inflaton field
may either roll - in ( like a standard hybrid inflation ) or roll - out ( like an inverted - hybrid model or quintessential inflation ) on arbitrary inflaton potential .
significant signatures of remote inflation can be observed in the spectrum caused by ( 1 ) the inhomogeneous phase transition in the remote sector , or ( 2 ) a successive phase transition in the remote sector .
remote inflation can predict strong amplification or suppression of small - scale perturbations without introducing multiple inflation . since the inflaton may have a run - away potential ,
it is also possible to identify the inflaton with quintessence , without introducing additional mechanisms .
even if the false - vacuum energy is not dominated by the remote sector , the phase transition in the remote sector is possible during warm inflation , which may cause significant amplification / suppression of the curvature perturbations .
sit - hep / tm-55 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 . |
one hundred years after its discovery by j. j. @xcite , the triatomic hydrogen ion h@xmath1 is still driving research at the forefront of astrophysics and molecular spectroscopy . since the publication of the classic papers on reaction networks in interstellar clouds by @xcite and @xcite ,
h@xmath1 has been recognized as a cornerstone of astrochemical models .
detections of h@xmath1 absorption lines in the dense @xcite and diffuse @xcite interstellar medium were milestones for the discipline of astrochemistry and have profoundly impacted our understanding of molecule formation in interstellar space .
today , h@xmath1 observations are routinely used to trace the cosmic ray ionization rate in the diffuse interstellar medium @xcite .
similar to h@xmath0 , h@xmath1 exists in two different nuclear spin modifications . if the three proton spins align to @xmath2 , ortho - h@xmath1 is formed ; if the nuclear spins combine to @xmath3 , para - h@xmath1 is formed .
in both h@xmath0 and h@xmath1 the lowest rotational states are of para symmetry , the energy difference to the next higher ( ortho ) state is 170.5k in h@xmath0 @xcite and 32.9k in h@xmath1 @xcite . for all practical purposes ,
the two different nuclear spin modifications can be considered different species and efficient interconversion can only be accomplished by exchange of protons in chemical reactions .
recent evaluations of h@xmath1 and h@xmath0 observations @xcite in diffuse sightlines , found that the h@xmath1 nuclear spin excitation temperature @xmath4h@xmath1 ) is consistently lower than the excitation temperature @xmath5 of h@xmath0 , despite the fact that these two species undergo frequent collisions . while @xmath5 is typically @xmath670k , @xmath7(h@xmath1 ) tends to be around @xmath630k , implying an enrichment in the para - h@xmath1 population .
@xcite have shown that @xmath5 can be taken as the kinetic temperature of the cloud , indicating that there has to be a mechanism that shifts @xmath4h@xmath1 ) away from thermodynamic equilibrium .
two different scenarios can be envisioned : ( i ) collisions between h@xmath1 and h@xmath0 result in a non - thermal nuclear spin distribution for h@xmath1 ; ( ii ) collisions between h@xmath1 and h@xmath0 are too infrequent or too inefficient in the diffuse interstellar medium to bring h@xmath1 into thermal equilibrium , therefore the formation and destruction processes dominate the observed nuclear spin fractions . in the present work we address point ( i ) by presenting measurements of the para - h@xmath1 fraction in collisions with thermal h@xmath0 samples at interstellar temperatures .
the h@xmath1 / h@xmath0 collision system has been the subject of previous experimental @xcite and theoretical @xcite studies .
this work represents the first experimental study of h@xmath1/h@xmath0 nuclear spin exchange at interstellar temperatures @xmath8k . in the following
, we will abbreviate para - h@xmath1 by @xmath9-h@xmath1 and ortho - h@xmath1 by @xmath10-h@xmath1 . in the same way , para - h@xmath0 and ortho - h@xmath0
will be denoted by @xmath9-h@xmath0 and @xmath10-h@xmath0 , respectively .
normal - h@xmath0 , with a ( @xmath10-h@xmath0):(@xmath9-h@xmath0 ) ratio of 3:1 , will be abbreviated by @xmath11-h@xmath0 .
it is convenient to define the following parameters to quantify the fractions of molecules in the para configuration : @xmath12 where the terms on the right - hand sides of the equations represent the number densities of the respective species .
the measurements were performed using a chemical probing scheme in a temperature - variable 22-pole radiofrequency ( rf ) ion trap @xcite . in this spectroscopy approach
the popualtions of the lowest rotational states of h@xmath1 are probed by a laser - induced reaction ( lir ) that leads to the formation of arh@xmath13 , which can be detected with very high sensitivity .
the lir technique has been developed for spectroscopy of cold ions @xcite .
the specifics for the spectroscopy of h@xmath1 have been described in detail elsewhere @xcite ; here we will give a brief overview .
the h@xmath1 ions are produced in a ion source by electron impact ionization of h@xmath0 , followed by the exothermic h@xmath14 + h@xmath0 @xmath15 h@xmath1 + h formation reaction .
the ions are extracted and guided into the 22-pole trap .
the cylindrically symmetric rf trap consists of 22 stainless steel rods ( diameter 1 mm , length 40 mm ) that are planted alternatingly into two copper side plates . by applying rf fields ( @xmath617.7mhz ) of opposing sign to the side plates ,
the ions are stored radially .
the ions enter and leave the ion trap through small circular electrodes ( diameter 8 mm ) inside the side plates , which can be switched to load or unload the ion trap .
the 22-pole geometry is chosen to minimize rf heating , and it has been shown that stored ions can be cooled down to their lowest rotational states in collisions with an inert buffer gas @xcite .
the ion trap is mounted on a 10k cold head and the trap temperature can be varied using a small heating element .
the temperature is measured by a calibrated silicon diode . during ion storage helium and argon gas
are continuously bled into the trap through dedicated gas lines .
helium serves as buffer gas to sympathetically cool the h@xmath1 ions .
collisional cooling with a neutral inert buffer gas allows for the cooling of translational as well as rovibrational degrees of freedom .
this cooling method is applicable to almost all stable molecular ions , and with helium as a buffer gas temperatures down to a few degrees kelvin can be reached .
an overview of this cooling technique can be found in @xcite .
the argon gas is used to facilitate the action spectroscopy scheme . in this spectroscopic approach
, an endothermic chemical reaction is used that is triggered by the absorption of a photon .
the reaction product is detected by a mass spectrometer with high sensitivity . for the present experiment ,
the endothermic proton - hop reaction @xmath16 was utilized . for ground state h@xmath1
, the reaction can not proceed .
it becomes energetically possible when the h@xmath1 excitation overcomes the 0.55ev threshold , which is the case , e.g. , for any h@xmath1 ion with more than one vibrational quantum [ @xmath17+@xmath18@xmath19 1 , where @xmath17 and @xmath20 denote the quantum numbers in the symmetric stretch and asymmetric bending modes of h@xmath1 , respectively .
( see * ? ? ? * for details on h@xmath1 nomenclature . ) a tunable diode laser ( 1.33 - 1.40@xmath21 m ) with an output power of 10 - 15mw is coupled in along the trap axis .
it is suited to excite h@xmath1 from the three lowest - lying rotational states to the ( @xmath17@xmath22)=(03@xmath23 ) vibrational overtone ( see fig.[fig1 ] ) . with a photon energy of @xmath60.9ev ,
the laser - excited h@xmath1 ions are well above the threshold for reaction([eq : arh ] ) and thus the formation of arh@xmath13 is assumed to proceed rapidly .
the spectra are taken by slowly scanning the laser over one of the three transitions , while repeatedly loading the ion trap and unloading it
after variable storage times and counting the arh@xmath13 reaction product . to probe the low - temperature equilibrium population of the lowest h@xmath1 states , we exposed the stored ions to h@xmath0 with a thermal nuclear spin fraction .
we prepared gas samples with variable @xmath24 using a para hydrogen converter , consisting of a cryogenic container that is filled with a ferromagnetic catalyst .
routinely , we prepared ca.600 mbar of pure @xmath9-h@xmath0 ( purity @xmath1999% ) in a 1liter lecture bottle that is lined with teflon to reduce back - conversion . in a second step ,
we diluted the @xmath9-h@xmath0 gas with @xmath11-h@xmath0 until we reached the desired @xmath24 value .
the mixtures were tested before and after each measurement by raman spectroscopy . for each measurement
the temperature of the ion trap was chosen to corresponded to the excitation temperature of the measured @xmath24 value .
typically , back - conversion during several days of measurement resulted in an increase of @xmath5 by 3k .
furthermore , the low raman signal strength caused by the relatively low pressure of the h@xmath0 mixtures
accounts for the uncertainty in @xmath24 of @xmath25 . due to the ionization process and the exothermic formation reaction
, approximately 2/3 of the @xmath6500 captured h@xmath1 ions are highly excited during the initial capture .
they immediately react with ar to form arh@xmath13 .
however , these arh@xmath13 ions react with h@xmath0 and undergo the exothermic back - conversion of reaction([eq : arh ] ) . using short trap loading and unloading cycles , we carefully monitored and adjusted the h@xmath0 pressure in the trap until the back conversion occurred with a time constant of @xmath610ms , thus assuring that the h@xmath0 number densities are comparable for each data point . within the first 50ms of the 500ms storage time ,
all excited h@xmath1 ions are buffer gas cooled and the initial arh@xmath13 count has dropped below the background level of the detection system .
lacking the option to directly use a pressure gauge inside the trap volume , we use readings at a remote gauge to estimate the helium number density inside the trap to be on the order of @xmath6@xmath26@xmath27 .
the argon density is 2 - 3 orders of magnitude lower than the helium density .
it is adjusted such that it saturates the arh@xmath13 signal , i.e. , that every excited h@xmath1 ion finds an argon atom to react before the excitation decays .
the hydrogen density is lower still and difficult to infer precisely .
the best estimate comes from the 10ms time constant of the arh@xmath13+h@xmath0@xmath15ar+h@xmath1 back - reaction .
however , this time constant is likely to reflect not only the reaction rate , but also the time that it takes the excited ions to cool below the reaction threshold , as those h@xmath1 ions that do not cool before the next collision with argon are likely to react again to form arh@xmath13 .
assuming langevin rate coefficients for both the arh@xmath13/h@xmath0 and h@xmath1/h@xmath0 collision systems , we infer a lower limit of 60 h@xmath1/h@xmath0 collisions during the 500ms storage time , while the actual number may be quite a bit higher . to assure that the number of h@xmath0 collisions is sufficient to bring the nuclear spin of h@xmath1 into steady state , we performed a test with doubled storage time , which did not alter the results .
for the last 50ms of storage , the diode laser is activated .
the measured arh@xmath13 signal results from the competition between laser excitation initiating reaction ( [ eq : arh ] ) and the presence of h@xmath0 that causes the reverse reaction .
the saturated arh@xmath13 signal following long exposure to the laser field ( long compared to the @xmath610ms decay time constant of the back - reaction ) is therefore directly proportional to the arh@xmath13 lifetime .
consequently , it would be impractical to increase the number of h@xmath0 collisions by increasing the h@xmath0 number density inside the ion trap , as that would lead to a reduction of the arh@xmath13 signal strength .
due to the time - consuming preparation and verification of the @xmath9-h@xmath0 samples together with careful adjustment of the relative gas densities and the slow scanning routine , each experimental @xmath28 value corresponds to more than one week of laboratory time .
the he and h@xmath0 number densities in the present experiment are lower by at least three orders of magnitude compared to experiments in plasmas and discharges and thus 3-body collisions and disturbances of the h@xmath29 collision complex are unlikely .
we expect collisions with helium to occur on a @xmath610@xmath21s time scale , while the lifetime of the h@xmath29 complex should be on the order of nanoseconds or shorter @xcite .
furthermore , as the ions are produced in an external ion source and transferred into the ion trap , there are no free electrons interacting with the ions during storage .
the laser is activated for the last 50ms of storage to achieve sufficient signal strength . during these 50ms ,
it is conceivable that arh@xmath13 formation , followed by back - conversion to h@xmath1 via @xmath30 will change the ortho / para ratio of the stored ions . here
, we will estimate the size of this effect .
the formation rate of arh@xmath13 during the laser - on time is given by @xcite @xmath31 where @xmath11(h@xmath1 ) is the h@xmath1 number density inside the trap , @xmath32 the fraction of h@xmath1 ions in a rovibrational state that can be excited by a given laser frequency @xmath33 , @xmath34 the einstein coefficient for that transition and @xmath35 the spectral energy density .
the integral runs over the interaction volume which is defined by the spatial overlap between the laser beam and the ion cloud inside the trap .
the parameter of interest is @xmath32 , which under constant measurement conditions is proportional to @xmath36 . at a given frequency , the arh@xmath13 yield as a function of the laser - on time @xmath37 and
the arh@xmath13 lifetime @xmath38 is @xcite @xmath39 for @xmath40 this leads to an asymptotic arh@xmath13 signal given by @xmath41 . from the measured @xmath42 and @xmath38
, we determine the arh@xmath13 formation rate @xmath43 .
for the strongest line
the @xmath44(1,0 ) @xmath10-h@xmath1 line we have measured a signal strength of @xmath45 , which together with @xmath46ms corresponds to @xmath47s@xmath48 . during the 50ms laser - on time , this leads to a total of @xmath680 arh@xmath13 ions , a considerable fraction of the @xmath6500 stored ions . to derive how many of these arh@xmath13 ions will end up as @xmath10-h@xmath1 and @xmath9-h@xmath1 , respectively
, we use the nuclear spin selection rules for reaction ( [ eq : back ] ) . the analogous case of spin statistics for the proton transfer between o@xmath0 and h@xmath0
is described by @xcite , and has recently been studied experimentally by @xcite . for this exothermic proton exchange reaction
, we assume that the nuclear spin selection rules hold .
the fraction of arh@xmath13 ions that will populate is ( @xmath49 , while ( @xmath50 ) of the arh@xmath13 ions will form @xmath9-h@xmath1 . for any given transition
, this effect will result in the depletion of the nuclear spin manifold that is being probed .
for the analysis we assume that the laser interaction will not affect the relative strength of the two para lines ( and thus the rotational temperature ) , as the frequent helium collisions will keep the rotational states within the para manifold in equilibrium .
the depletion effect should be most pronounced for the @xmath44(1,0 ) ortho line , firstly because @xmath44(1,0 ) is the strongest line , and secondly because for high @xmath24 values , most of the arh@xmath13 ions will be converted to @xmath9-h@xmath1 .
the possible influence of the laser interaction therefore leads to increased uncertainty limits towards small values of @xmath28 .
for the data point with the highest @xmath24 value this corresponds to an uncertainty of @xmath611% in @xmath28 . the effect is included in the total uncertainty for each data point .
this procedure may over - estimate the effect , because during the 50ms laser - on time , further h@xmath1/h@xmath0 collisions will shift the populations back towards equilibrium .
furthermore , a strong laser depletion would deform the lineshapes of the measured transitions and we see no indication of that in our data .
the chemical probing spectra for the different values of @xmath24 are shown in figure[fig2 ] . plotted are normalized arh@xmath13 count rates for the @xmath51 , @xmath52 , and @xmath53 transitions , respectively .
the linewidths are dominated by the doppler width , which is on the order of @xmath54mhz at 70k .
the lines were fitted with gaussian profiles and translational temperatures were extracted from the width of the strongest line @xmath51 . from the relative arh@xmath13 yield of the @xmath52 and
the @xmath53 transitions , which are both within the para manifold , the rotational temperature @xmath55 was derived , using the einstein coefficients of @xcite .
assuming that the rotational populations are in equilibrium among the para- and ortho states , we calculated the fractional populations @xmath56 and @xmath57 for the lowest para ( 1,1 ) and ortho ( 1,0 ) levels , respectively , within the ortho and para manifolds for the given value of @xmath55 . then we used the measured arh@xmath13 yields @xmath58 and @xmath59 for the @xmath52 and @xmath51 transitions , normalized to the einstein @xmath34 coefficients to derive the relative para and ortho populations using @xmath60 from these populations we calculated @xmath28 for the stored h@xmath1 ions using the analog of equation ( [ eq : p3 ] ) .
all temperatures as well as the @xmath24 and @xmath28 values are given in table1 .
[ cols="^,^,^,^,^,^,^ " , ] in figure[fig3 ] the measured equilibrium values are compared to astronomical observations .
the observational h@xmath1 data are taken from @xcite , the h@xmath0 data from @xcite and @xcite ; they represent diffuse molecular clouds with average kinetic temperatures between 5070k and h@xmath0 column densities on the order of 10@xmath61@xmath62 .
the experimental results are close to the thermal curve , while the interstellar observations exhibit distinctly larger values of @xmath28 .
also plotted is the outcome of a chemical kinetics model @xcite , based on rate coefficients calculated using the micro - canonical model of @xcite .
as detailed in the original publication , the model requires three parameters as input , namely the branching fraction @xmath63 between the proton - hop and proton exchange channels , the fraction @xmath64 of the identity reaction and the temperature @xmath7 .
here we adopt @xmath65 , as observed recently in a liquid - nitrogen cooled discharge @xcite .
the identity branching fraction has a minor impact on the equilibrium conditions , here we used @xmath66 .
the model results agree very well with the experimental values .
the present measurements show that the nuclear spin equilibrium of h@xmath1 in collisions with h@xmath0 is close to the expected thermal values .
the discrepancy with the observations in diffuse sightlines remains , consequently , it appears that in these interstellar environments , h@xmath1 / h@xmath0 collisions are too infrequent to bring the two species into equilibrium .
@xcite have shown that the nascent @xmath28 value originating from the h@xmath14 + h@xmath0 @xmath67 h@xmath1 + h formation reaction is likely to follow a linear trend ( plotted in fig.[fig3 ] ) with @xmath68 .
the present results nurture the assumption that h@xmath1 might be formed along the nascent line and has not reached the thermal curve yet in the observed environments .
the dominant destruction process of h@xmath1 in the diffuse medium is the dissociative recombination ( dr ) with free electrons .
the number of reactive collisions @xmath69 with h@xmath0 that an h@xmath1 ion experiences in its lifetime can be estimated by @xmath70 where @xmath71 and @xmath72 denote the rate coefficients for reactive h@xmath0 collisions and dr , respectively
. adopting values of k@xmath73@xmath74s@xmath48 @xcite and @xmath75@xmath74s@xmath48 @xcite together with a n(h@xmath0)/n(e@xmath76 ) ratio of @xmath77 , we expect @xmath640 collisions per lifetime . while that number may be sufficient for thermalization , a smaller rate coefficient for reactive collisions , as suggested by deuteration experiments described by @xcite , would result in @xmath67 collisions , which may not be enough for thermalization .
however , one should bear in mind that more recent measurements by @xcite with a similar setup resulted in larger rate coefficients , more in line with the current assumptions .
another source of uncertainty for the astrophysical @xmath28 value is the nuclear - spin dependence of the dr rate coefficient .
recent calculations by @xcite suggest a large difference of almost an order of magnitude between @xmath9-h@xmath1 and @xmath10-h@xmath1 at low temperatures .
incidentally , the calculated rate coefficient for @xmath9-h@xmath1 is higher , which aggravates the problem since this would lead to a depletion of @xmath9-h@xmath1 .
presently , the theoretical calculations are supported by afterglow experiments @xcite , while measurements in storage rings see only a slight enhancement in the rate coefficient of @xmath9-h@xmath1 @xcite , however , the sampled temperatures in the storage rings may be too high @xcite .
as the present measurements indicate that the irregular nuclear spin populations of h@xmath1 do not have their origin in the h@xmath1 / h@xmath0 collision process itself , strengthened efforts to understand the h@xmath1 formation and destruction processes , as well as an accurate determination of the absolute rate coefficient of the thermalizing collisions with h@xmath0 , are called for .
we acknowledge support from the max planck society . k.n.c . and b.j.m .
have been supported by the national science foundation ( phy 08 - 55633 ) .
k.n.c . was supported by a nasa earth and space science fellowship .
s.g . and s.s .
have been supported by the deutsche forschungsgemeinschaft ( dfg ) in the framework of the collaborative research grant sfb 956 .
s.g . acknowledges financial support from the bonn - cologne graduate school of physics and astronomy ( bcgs ) .
we thank kisam park for providing the code to determine the nuclear - spin - dependent rate coefficients . | recent observations of h@xmath0 and h@xmath1 in diffuse interstellar sightlines revealed a difference in the nuclear spin excitation temperatures of the two species .
this discrepancy comes as a surprise , as h@xmath1 and h@xmath0 should undergo frequent thermalizing collisions in molecular clouds .
non - thermal behavior of the fundamental h@xmath1 / h@xmath0 collision system at low temperatures was considered as a possible cause for the observed irregular populations . here , we present measurements of the steady - state ortho / para ratio of h@xmath1 in collisions with h@xmath0 molecules in a temperature - variable radiofrequency ion trap between 45 - 100k .
the experimental results are close to the expected thermal outcome and they agree very well with a previous micro - canonical model .
we briefly discuss the implications of the experimental results for the chemistry of the diffuse interstellar medium . |
the notion of heat kernel has a long history .
the oldest and the best - known heat kernel is the gauss weierstrass function @xmath0 where @xmath1 and @xmath2 , which is a fundamental solution of the heat equation @xmath3 where @xmath4 is the laplace operator in @xmath5 . a more general parabolic equation @xmath6 , where @xmath7 is a uniformly elliptic operator with measurable coefficients @xmath8 , has also a positive fundamental solution @xmath9 , and the latter admits the gaussian bounds@xmath10 where the sign @xmath11 means that both @xmath12 and @xmath13 are true , but the positive constants @xmath14 and @xmath15 may be different for upper and lower bounds . the estimate ( [ gaussian ] ) was proved by aronson @xcite using the parabolic harnack inequality of moser @xcite .
the next chapter in the history of heat kernels was opened in differential geometry .
consider the heat equation ( [ heat ] ) on a riemannian manifold @xmath16 , where @xmath4 is now the laplace
beltrami operator on @xmath16 .
the heat kernel @xmath9 is defined as the minimal positive fundamental solution of ( [ heat ] ) , which always exists and is a smooth nonnegative function of @xmath17 ; cf .
the question of estimating the heat kernel on riemannian manifolds was addressed by many authors ( see , e.g. , @xcite ) .
apart from obvious analytic and geometric motivation , a strong interest to heat kernel estimates persists in stochastic analysis because the heat kernel coincides with the transition density of brownian motion on @xmath16 generated by the laplace beltrami operator .
one of the most powerful estimates of heat kernels was proved by li and yau @xcite : if @xmath16 is a complete riemannian manifold of nonnegative ricci curvature , then@xmath18 where @xmath19 is the geodesic distance on @xmath16 , and @xmath20 is the riemannian volume of the geodesic ball @xmath21 .
similar estimates were obtained by gushchin with coauthors @xcite for certain unbounded domains in @xmath5 with the neumann boundary condition .
an interesting question is what minimal geometric assumptions imply ( [ liyau ] ) .
the upper bound in ( [ liyau ] ) is know to be equivalent to a certain _ faber krahn_-type inequality ( see section [ secef ] ) .
the geometric background of the lower bound in ( [ liyau ] ) is more complicated and is closely related to the harnack inequalities .
in fact , the full estimate ( [ liyau ] ) is equivalent , on one hand , to the parabolic harnack inequality of moser ( see @xcite ) , and , on the other hand , to the conjunction of the volume doubling property and the poincar inequality ( see @xcite ) . for a more detailed account of heat kernel bounds on manifolds
we refer the reader to the books and surveys @xcite .
new dimensions in the history of heat kernels were literally discovered in analysis on fractals .
fractals are typically subsets of @xmath5 with certain self - similarity properties , like the sierpinski gasket ( @xmath22 ) or the sierpinski carpet ( @xmath23 ) .
one makes a fractal into a metric measure space by choosing appropriately a metric @xmath24 ( e.g. , the extrinsic metric from the ambient @xmath5 ) and a measure @xmath25 ( usually the hausdorff measure ) .
the next crucial step is introduction of a strongly local regular _ dirichlet form _ on a fractal , that is , an analogy of the dirichlet integral @xmath26 on manifolds , which is equivalent to construction of brownian motion on the fractal in question ; cf . @xcite .
this step is highly nontrivial and its implementation depends on a particular class of fractals . on @xmath27
brownian motion was constructed by goldstein @xcite and kusuoka @xcite , on @xmath23 , by barlow and bass @xcite .
kigami @xcite introduced a class of _ post - critically finite _ ( p.c.f . )
fractals , containing @xmath27 , and constructed the dirichlet form on such a fractal as a scaled limit of the discrete dirichlet forms on the graph approximations .
a strongly local regular dirichlet form canonically leads to the notion of the heat semigroup and the heat kernel , where the latter can be defined either as the integral kernel of the heat semigroup or as the transition density of brownian motion .
surprisingly enough , the dirichlet forms on many families of fractals admit continuous heat kernels that satisfy the _ sub - gaussian _ estimates@xmath28 where @xmath29 and @xmath30 are two parameters that come from the geometric properties of the underlying fractal . estimate ( [ subgauss ] ) was proved by barlow and perkins @xcite on @xmath27 , by kumagai @xcite on nested fractals , by fitzsimmons , hambly and kumagai @xcite on affine nested fractals and by barlow and bass on @xmath23 @xcite and on generalized sierpinski carpets @xcite ( see also @xcite ) .
in fact , @xmath31 is the hausdorff dimension of the space , while @xmath32 is a new quantity that is called the _ walk dimension _ and that can be characterized either in terms of the exit time of brownian motion from balls or as the critical exponent of a family of besov function spaces on the fractal ( cf .
@xcite).=-1 the purpose of this paper is to find convenient equivalent conditions for sub - gaussian estimates of the heat kernels on abstract metric measure spaces .
let @xmath33 be a locally compact separable metric space , @xmath25 be a radon measure on @xmath16 with full support and @xmath34 be a strongly local regular dirichlet form on @xmath16 ( see section [ secbasic ] for the details ) .
we are interested in the conditions that ensure the existence of the heat kernel @xmath35 as a measurable or continuous function of @xmath36 , and the estimates of the following type@xmath37 where @xmath38 and @xmath39 , @xmath40 are some nonnegative increasing functions on @xmath41 .
for example , ( [ liyau ] ) has the form ( [ rfi ] ) with @xmath42 and @xmath43 , while ( [ subgauss ] ) has the form ( [ rfi ] ) with @xmath44 and @xmath45 [ assuming that means that the ratio of both sides is bounded between two positive constants . ]
@xmath46 , which , in fact , follows from ( [ subgauss ] ) ] . to describe the results of the paper ,
let us introduce some hypotheses .
first , we assume that the metric space @xmath33 is unbounded and that all metric balls are precompact is a complete metric spaces .
the following partial converse is also true : if @xmath47 is complete and the volume doubling property ( [ vd ] ) holds , then all balls are precompact . however , since we do not always assume ( [ vd ] ) , we make an independent assumption of precompactness of the balls . ]
( although these assumptions are needed only for a part of the results ) . next ,
define the following conditions : @xmath48 the volume doubling property ( [ vd ] ) : there is a constant @xmath15 such that @xmath49 for all @xmath50 and @xmath51 ; @xmath48 the elliptic harnack inequality ( [ h ] ) : there is a constant @xmath15 such that , for any nonnegative harmonic function @xmath52 in any ball @xmath53 , @xmath54 where @xmath55 and @xmath56 are the essential supremum and infimum , respectively , ( see section [ secharnack ] for more details ) ; @xmath48 the estimate of the mean exit time ( [ ef ] ) , @xmath57 where @xmath58 is the first exist time from ball @xmath59 of the associated diffusion process , started at the center @xmath60 , and @xmath61 is a given function with a certain regularity ( see section [ secef ] for more details ) .
a typical example is @xmath62 for some constant @xmath30 . the conditions @xmath63 are known to be true on p.c.f
. fractals ( see @xcite ) as well as on generalized sierpinski carpets ( see @xcite ) so that our results apply to such fractals .
another situation where @xmath64 are satisfied is the setting of _ resistance forms _ introduced by kigami @xcite .
a resistance form is a specific dirichlet form that corresponds to a strongly recurrent brownian motion .
kigami showed that , in the setting of resistance forms on self - similar sets , ( [ vd ] ) alone implies ( [ h ] ) and ( [ ef ] ) with @xmath65 , for a suitable choice of a distance function .
examples with more general functions @xmath61 appear in @xcite and @xcite .
let us emphasize in this connection that our results do not depend on the recurrence or transience hypotheses and apply to both cases , which partly explains the complexity of the proofs .
a transient case occurs , for example , for some generalized sierpinski carpets .
another point worth mentioning is that we do not assume specific properties of the metric @xmath24 such as being geodesic ; the latter is quite a common assumption in the fractal literature .
this level of generality enables applications to resistance forms where the distance function is usually the resistance metric that is not geodesic . our first main result , which is stated in theorem [ tmain ] and which , in fact , is a combination of theorems [ tg=>fk ] , [ tdue ] , [ tholder ] , [ tnle ] , says the following : if the hypotheses @xmath66 are satisfied , then the heat kernel @xmath67 exists , is hlder continuous in @xmath36 and satisfies the following upper estimate : @xmath68 where @xmath69 and @xmath70 and the _ near - diagonal _ lower estimate @xmath71 where @xmath72 is a small enough constant .
furthermore , assuming that ( [ vd ] ) holds a priori , we have the equivalence ] where @xmath73 stands for the _ parabolic _ harnack inequality for caloric functions .
hence , we see that the `` difference '' between @xmath74 and ( [ h ] ) is the condition ( [ ef ] ) , that in particular provides a necessary space / time scaling for @xmath74 . ]
@xmath75 ( theorem [ tconv ] ) .
for example , if @xmath62 for some @xmath30 , then @xmath44 and @xmath76 .
hence , ( [ ue ] ) and ( [ nle ] ) become as follows:@xmath77 and@xmath78 it is desirable to have a lower bound of @xmath9 for all @xmath36 that would match the upper bound ( [ subu ] ) .
however , such a lower bound fails in general .
the reason for that is the lack of _ chaining properties _ of the distance function , where by chaining properties we loosely mean a possibility to connect any two points @xmath79 by a chain of balls of controllable radii so that the number of balls in this chain is also under control .
more precisely , this property can be stated in terms of the modified distance @xmath80 where @xmath81 is a parameter . the exact definition of @xmath82 is given in section [ secde ] , where it is also shown that @xmath83 where @xmath84 is the smallest number of balls in a chain of balls of radii @xmath85 connecting @xmath60 and @xmath86 .
as @xmath87 goes to @xmath88 , @xmath80 increases and can go to @xmath89 or even become equal to @xmath89 . if the distance function @xmath24 is geodesic then @xmath90 , which corresponds to the best possible chaining property .
in general , the rate of growth of @xmath80 as @xmath91 can be regarded as a quantitative description of the chaining properties of @xmath24 . for this part of our work
, we assume that@xmath92 which allows to define a function @xmath93 from the identity@xmath94 our second main result states the following : if ( [ ad ] ) and @xmath95 are satisfied , then@xmath96 where @xmath97 ( theorem [ ttwo ] ) . for example , the above hypotheses and , hence , the estimates ( [ twoe ] ) and ( [ twohk ] ) hold on connected p.c.f .
fractals endowed with resistance distance , where one has @xmath46 and @xmath98 for some constant @xmath31 . the estimate ( [ twohk ] ) on p.c.f .
fractals was first proved by hambly and kumagai @xcite .
in fact , we use the argument from @xcite to verify our hypotheses ( see remark [ exhk ] ) .
note that the dependence on @xmath17 in the estimates ( [ twoe ] ) and ( [ twohk ] ) in very implicit and is hidden in @xmath99 .
one can loosely interpret the use of this function in ( [ twoe ] ) and ( [ twohk ] ) as follows . in order to find a most probable path for brownian motion to go from @xmath60 to @xmath86 in time @xmath100 ,
one determines the optimal size @xmath97 of balls and then the optimal chain of balls of radii @xmath85 connecting @xmath60 and @xmath86 , and this chain provides an optimal route between @xmath60 and @xmath86 .
this phenomenon was discovered by hambly and kumagai in the setting of p.c.f .
fractals , where they used instead of balls the construction cells of the fractal .
as it follows from our results , this phenomenon is generic and independent of self - similar structures . if the distance function satisfies the _ chain condition _
@xmath101 , which is stronger than ( [ ad ] ) , then one can replace in ( [ twoe ] ) @xmath82 by @xmath24 and obtain ( [ rfi ] ) ( corollary [ cortwo ] ) .
in fact , in this case we have the equivalence @xmath102 ( corollary [ corunbound ] ) . in the setting of random walks on infinite graphs , the equivalence ( [ 3 + 2 ] )
was proved by the authors in @xcite .
of course , in this case all the conditions have to be adjusted to the discrete setting . for the sake of applications ( cf . , e.g , @xcite ) , it is desirable to replace the probabilistic condition ( [ ef ] ) in all the above results by an analytic condition , namely , by a certain estimate of the capacity between two concentric balls
this type of result requires different techniques and will be treated elsewhere . in section [ sechh ]
we revise the basic properties of the heat semigroups and heat kernels and prove the criterion for the existence of the heat kernel in terms of local ultracontractivity of the heat semigroup ( theorem [ tptom ] ) . in section [ secaux ]
we prove two preparatory results : @xmath103 where ( [ fk ] ) stands for a certain _ faber krahn inequality _ , which provides a lower bound for the bottom eigenvalue in any bounded open set @xmath104 via its measure ( theorem [ tg=>fk ] ) .
in turn , ( [ fk ] ) implies the local ultracontractivity of the heat semigroup , which by theorem [ tptom ] ensures the existence of the heat kernel .
@xmath105 implies the following estimate of the tail of the exit time from balls:@xmath106 ( theorem [ tef ] ) . in section [ secdue ]
we prove the upper estimate of the heat kernel , more precisely , the implication@xmath107 ( theorem [ tdue ] ) . the main difficulty lies already in the proof of the diagonal upper bound @xmath108 using ( [ fk ] ) , we obtain first some diagonal upper bound for the dirichlet heat kernels in balls , and then use kigami s iteration argument and ( [ psi2 ] ) to pass to ( [ due ] ) .
the latter argument is borrowed from @xcite .
the full upper estimate ( [ ue ] ) follows from ( [ due ] ) and ( [ psi2 ] ) . in section [ seclow ]
we prove the lower bounds of the heat kernel .
the diagonal lower bound @xmath109 follows directly from ( [ psi2 ] ) ( lemma [ lemdle ] ) . to obtain the near diagonal lower estimate ( [ nle ] ) ,
one estimates from above the difference @xmath110 where @xmath86 is close to @xmath60 , which requires the following two ingredients : the oscillation inequalities that are consequences of the elliptic harnack inequality ( [ h ] ) ( lemma [ lemosc ] and proposition [ posc ] ) ; the upper estimate of the time derivative @xmath111 ( corollary [ cdtpt ] ) .
combining them with ( [ ue ] ) , one obtains an upper bound for ( [ di ] ) , which together with ( [ dle ] ) yields ( [ nle ] ) ( theorem [ tnle ] ) .
the same method gives also the hlder continuity of the heat kernel ( theorem [ tholder ] ) . in section [ sectwo ]
we prove two - sided estimates ( [ twoe ] ) and ( [ twohk ] ) ( theorem [ ttwo ] ) .
for the upper bound , we basically repeat the proof of ( [ ue ] ) by tracing the use of the distance function @xmath24 and replacing it by @xmath82 .
the lower bound for large @xmath112 is obtained from ( [ nle ] ) by a standard chaining argument using the semigroup property of the heat kernel and the chaining property of the distance function . in section [ secconv ]
we prove the converse theorem [ tconv ] , which essentially consists of the equivalence ( [ eq ] ) .
we use the letters @xmath113 etc . to denote positive constant
whose value is unimportant and can change at each occurrence .
note that the value of such constants in the conclusions depend on the values of the constants in the hypotheses ( and , perhaps , on some other explicit parameters ) . in this sense , all our results are quantitative .
the relation @xmath114 means that @xmath115 for some positive constant @xmath15 and for a specified range of the arguments of functions @xmath116 and @xmath117 .
the relation @xmath118 means that both inequalities @xmath119 and @xmath120 hold but possibly with different values of constants @xmath121 that are involved in the expressions @xmath116 and/or @xmath117 .
[ sechh ] throughout the paper , we assume that @xmath47 is a locally compact separable metric space , and @xmath25 is a radon measure on @xmath16 with full support . as usual , denote by @xmath122 where @xmath123 $ ] the lebesgue function space with respect measure @xmath25 , and by @xmath124 the norm in @xmath125 . the inner product in @xmath126 is denoted by @xmath127 .
all functions on @xmath16 are supposed to be real valued .
denote by @xmath128 the space of all continuous functions on @xmath16 with compact supports , equipped with the @xmath129-norm .
let @xmath130 be dirichlet form in @xmath126 .
this means that @xmath131 is a dense subspace of @xmath126 , and @xmath132 is a bilinear , nonnegative definite , closed is called closed if @xmath131 is a hilbert space with respect to the following inner product : @xmath133 ] form defined for functions @xmath134 , which satisfies , in addition , the markovian property .
, then also the function @xmath135 belongs to @xmath131 and @xmath136 . ] the dirichlet form @xmath34 is called _ regular _ if @xmath137 is dense both in @xmath131 and in @xmath138 .
the dirichlet form is called _ strongly local _ if @xmath139 for all functions @xmath134 such that @xmath117 has a compact support and @xmath140 in a neighborhood of @xmath141 . in this paper , we assume by default that @xmath34 is a regular , strongly local dirichlet form . a general theory of dirichlet forms can be found in @xcite .
let @xmath142 be the generator of @xmath143 ; that is , @xmath142 is a self - adjoint nonnegative definite operator in @xmath126 with the domain @xmath144 that is a dense subset of @xmath131 and such that , for all @xmath145 and @xmath146@xmath147 the associated _ heat semigroup _
@xmath148 is a family of bounded self - adjoint operators in @xmath149 .
the markovian properties allow the extension of @xmath150 to a bounded operator in @xmath122 , with the norm @xmath151 , for any @xmath152 $ ] .
denote by @xmath153 the class of all borel functions on @xmath16 , by @xmath154 the class of bounded borel functions , by @xmath155 the class of nonnegative borel functions and by @xmath156 the class of borel functions that belong to @xmath122 .
by @xcite , theorem 7.2.1 , for any local dirichlet form , there exists a diffusion process @xmath157 with the initial point @xmath60 outside some properly exceptional set is called properly exceptional if it is borel , @xmath158 and@xmath159 for all @xmath160 ( see @xcite , page 134 and theorem 4.1.1 on page 137 ) . ]
@xmath161 , which is associated with the heat semigroup @xmath162 as follows : for any @xmath163 , @xmath164,@xmath165 consider the family of operators @xmath166 defined by@xmath167 for all functions @xmath168 ( if @xmath169 has a finite lifetime , then @xmath116 is to be extended by @xmath88 at the cemetery ) .
the function @xmath170 is a bounded borel function on @xmath171 .
[ remdoweneedthisextension ? ] it is convenient to extend it to all @xmath50 by setting@xmath172 so that @xmath173 can be considered as an operator in @xmath174 . obviously , @xmath175
if @xmath176 and @xmath177 . moreover , the family @xmath178 satisfies the semigroup identity @xmath179 indeed , if @xmath180 , then we have by the markov property , for any @xmath168,@xmath181 where we have used that @xmath182 with @xmath183-probability @xmath184 . if @xmath185 , then we have again@xmath186 because the both sides are @xmath88 . by considering increasing sequences of bounded functions ,
one extends the definition of @xmath187 to all @xmath188 so that the defining identities ( [ ptex ] ) and ( [ ptex0 ] ) remain valid also for @xmath189 [ allowing value @xmath190 for @xmath170 ] . for a signed function @xmath192 ,
define @xmath187 by@xmath193 provided at least one of the functions @xmath194 , @xmath195 is finite .
obviously , identities ( [ ptex ] ) , ( [ ptex0 ] ) are satisfied for such functions as well .
if follows from the comparison of ( [ ee ] ) and ( [ ptex ] ) that , for all @xmath196 , @xmath197 it particular , @xmath187 is finite almost everywhere .
the set of the above assumptions will be referred to as the _ basic hypotheses _ , and they are assumed by default in all parts of this paper . sometimes we need also the following property . the dirichlet form @xmath130 is called _ conservative _ ( or _ stochastically complete _ ) if @xmath198 for all @xmath1 .
let @xmath16 be a connected riemannian manifold , @xmath24 be the geodesic distance on @xmath16 , @xmath25 be the riemannian volume .
define the sobolev space@xmath199 where @xmath200 is the riemannian gradient of @xmath116 understood in the weak sense .
for all @xmath201 , one defines the energy form@xmath202 let @xmath131 be the closure of @xmath203 in @xmath204 .
then @xmath130 is a regular strongly local dirichlet form in @xmath126 .
[ remexamplesofdirichletformonfractals ] [ defhk ] the _ heat kernel _ ( or the _ transition density _ ) of the transition semigroup @xmath205 is a function @xmath9 defined for all @xmath1 and @xmath206 , where @xmath207 is a properly exceptional set containing @xmath208 , and such that the following properties are satisfied : for any @xmath1 , the function @xmath9 is measurable jointly in @xmath36 ; for all @xmath189 , @xmath1 and @xmath209,@xmath210 for all @xmath1 and @xmath211 , @xmath212 for all @xmath213 and @xmath211,@xmath214 the set @xmath215 is called the domain of the heat kernel .
let us extend @xmath9 to all @xmath79 by setting @xmath216 if @xmath60 or @xmath86 is outside @xmath215 .
then ( [ sym ] ) and ( [ semi ] ) hold for all @xmath79 , and the domain of integration in ( [ pt = pt ] ) and ( [ semi ] ) can be extended to @xmath16 . the existence of the heat kernel allows us to extend the definition of @xmath187 to all measurable functions @xmath116 by choosing a borel measurable version of @xmath116 and noticing that the integral ( [ pt = pt ] ) does not change if function @xmath116 is changed on a set of measure @xmath88 .
it follows from ( [ ee ] ) and ( [ pt = pt ] ) that , for any @xmath217 , @xmath218 for all @xmath1 and @xmath25-a.a .
@xmath50 . a measurable function @xmath9 that satisfies ( [ ept ] )
is called the heat kernel of the semigroup @xmath150 .
it is well known that the heat kernel of @xmath150 satisfies ( [ sym ] ) and ( [ semi ] ) although for _ almost _ all @xmath79 ( see @xcite , section 3.3 ) .
hence , the relation between the heat kernels of @xmath173 and @xmath150 is as follows : the former is defined as a pointwise function of @xmath36 , while the latter is defined almost everywhere , and the former is a pointwise realization of the latter , where the defining identities ( [ pt = pt ] ) , ( [ sym ] ) , ( [ ept ] ) must be satisfied pointwise . in this paper
the heat kernel is understood exclusively in the sense of definition [ defhk ] .
the existence of the heat kernel is not obvious at all and will be given a special treatment .
those who are interested in the settings where the pointwise existence of the heat kernel is known otherwise , can skip the rest of this section and go to section [ secaux ] .
[ lemuniq]let @xmath219 be the heat kernel of @xmath173 .
the function @xmath220 belongs to @xmath221 for all @xmath1 and @xmath50 .
for all @xmath1 , @xmath79 , we have @xmath222 and @xmath223 consequently , @xmath224 . if @xmath225 is another heat kernel , then @xmath226 in the common part of their domains .
\(a ) set @xmath227 and observe that , by ( [ sym ] ) and ( [ semi]),@xmath228 for all @xmath1 and @xmath211 . since @xmath229 is a borel function , we obtain that @xmath220 is borel .
the latter is true also if @xmath230 since in this case @xmath231 . setting in ( [ t2 ] ) @xmath232
, we obtain@xmath233 whence @xmath234 .
\(b ) by ( [ ptex ] ) , ( [ ptex0 ] ) we have @xmath235 for all @xmath1 , @xmath50 provided @xmath236 . setting @xmath237 _ { -}$ ] , we obtain@xmath238 _ { -}\,d\mu=-\int _ { m } [ p_{t } ( x,\cdot ) ] _ { -}^{2}\,d\mu,\ ] ] whence it follows that @xmath239 _
{ -}=0 $ ] a.e . ,
that is , @xmath240 a.e . on @xmath16 .
it follows from ( [ t2 ] ) that , for all @xmath79,@xmath241 inequality ( [ int1 ] ) is trivial if @xmath230 , and if @xmath209 then it follows from@xmath242
\(c ) let @xmath215 be the intersection of the domains of @xmath219 and @xmath225 . for all @xmath189 and @xmath1 , @xmath209 ,
we have @xmath243 applying this identity to function @xmath244 where @xmath245 , and using ( [ t2 ] ) , we obtain@xmath246 similarly , we have@xmath247 whence @xmath248 . following @xcite , page 67 , a sequence @xmath249 of subsets of @xmath16
will be called a _ regular nest _ if : each @xmath250 is closed ; @xmath251 for all @xmath252 ; @xmath253 as @xmath254 ( see @xcite for the definition of capacity ) ; measure @xmath255 has full support in @xmath250 ( in the induced topology of @xmath250 ) . [ defn ] a set @xmath256 is called _ truly exceptional _ if : @xmath207 is properly exceptional ; @xmath257 ; there is a regular nest @xmath258 in @xmath16 such that @xmath259 and that the function @xmath260 is continuous for all @xmath261 , @xmath1 , and @xmath262 . the conditions under which a truly exceptional set exists , will be discussed later on .
let us mention some important consequences of the existence of such a set .
[ lempt < fi]let @xmath207 be a truly exceptional set . if , for some @xmath263 , , and for an upper semicontinuous function @xmath264 $ ] , the inequality@xmath265 holds for @xmath25-a.a .
@xmath50 , then it is true for all @xmath266 .
similarly , if @xmath267 is a lower semicontinuous function and@xmath268 holds for @xmath25-a.a .
@xmath50 , then it is true for all @xmath266 .
this proof is essentially the same as in @xcite , theorem 2.1.2(ii ) .
assume that @xmath269 for some @xmath270 .
by definition [ defn ] , @xmath271 belongs to one of the sets @xmath250 .
since @xmath272 is continuous and , hence , @xmath273 is lower semicontinuous , the condition @xmath274 implies that @xmath275 for all @xmath60 in some open neighborhood @xmath276 of @xmath271 in @xmath250 . since measure
@xmath25 has full support in @xmath250 , we have @xmath277 so that @xmath278 in a set of positive measure , that contradicts the hypothesis .
the second claim follows from the first one with @xmath279 .
denote by @xmath280 the @xmath25-essential supremum of a function @xmath116 on a set @xmath281 , and by @xmath282the @xmath25-essential infimum .
[ cesup = sup1]let @xmath207 be a truly exceptional set .
then , for any @xmath263 , @xmath1 , and an open set @xmath283,@xmath284 function@xmath285 is upper semicontinuous . since @xmath286 for @xmath25-a.a .
@xmath50 , we conclude by lemma [ lempt < fi ] that this inequality is true for all @xmath287 , whence@xmath288 the opposite inequality follows trivially from the definition of the essential supremum .
the second identity in ( [ esi ] ) follows from the first one by changing @xmath116 to @xmath289 .
note that if @xmath9 is the heat kernel with domain @xmath290 , then we have by ( [ semi ] ) that , for all @xmath291 , @xmath292,@xmath293 where @xmath294 .
hence , if @xmath207 is truly exceptional , then the claims of lemma [ lempt < fi ] and corollary [ cesup = sup1 ] apply to function @xmath9 in place of @xmath295 , with any fixed @xmath245 .
[ lembgk ] let @xmath296 be the heat kernel with the domain @xmath290 such that @xmath207 is a truly exceptional set .
let @xmath297 $ ] be an upper semicontinuous function and @xmath298 be a lower semicontinuous function .
if , for some fixed @xmath1 , the following inequality : @xmath299 holds for @xmath300-almost all @xmath211 , then ( [ pta ] ) holds for all @xmath211 .
this lemma is a generalization of @xcite , lemma 2.2 , and the proof follows the argument in @xcite .
proof of lemma [ lembgk ] consider the set@xmath301 if @xmath302 then applying lemma [ lempt < fi ] to the function @xmath303 , we obtain that ( [ pta ] ) holds for all @xmath209 .
now fix @xmath209 .
since by fubini s theorem @xmath304 , ( [ pta ] ) holds for @xmath25-a.a .
@xmath305 . applying lemma [ lempt < fi ] to the function @xmath220
, we conclude that ( [ pta ] ) holds for all @xmath245 .
[ cesup = sup]under the hypotheses of lemma [ lembgk ] , if @xmath306 are two open subsets of @xmath16 then@xmath307 and@xmath308 this follows from lemma [ lembgk ] with functions @xmath309 and@xmath310 in conclusion of this section , let us state a result that ensures the existence of the heat kernel outside a truly exceptional set .
[ tbbck ] assume that there is a positive left - continuous function @xmath311 such that for all @xmath312 and @xmath1,@xmath313 then the transition semigroup @xmath173 possesses the heat kernel @xmath9 with domain @xmath290 for some truly exceptional set @xmath207 , and @xmath314 for all @xmath211 and @xmath1 .
if the semigroup @xmath162 satisfies ( [ ultra ] ) , then it is called _ ultracontractive _ ( cf .
it was proved in @xcite that the ultracontractivity implies the existence of a function @xmath67 that satisfies all the requirements of definition [ defhk ] except for the joint measurability in @xmath36 .
let us prove the latter so that @xmath9 is indeed the heat kernel in our strict sense .
given that @xmath9 satisfies conditions ( 2)(4 ) of definition [ defhk ] , we see that the statement of lemma [ lemuniq ] remains true because the proof of that lemma does not use the joint measurability .
in particular , for any @xmath1 , @xmath209 , the function @xmath220 is in @xmath126 .
also , the mapping @xmath315 is weakly measurable as a mapping from @xmath215 to @xmath126 because for any @xmath316 , the function @xmath317 is measurable . since @xmath318 is separable , by pettis s measurability theorem ( see @xcite , chapter v , section 4 ) the mapping @xmath319 is strongly measurable in @xmath320 .
it follows that the function@xmath321 is jointly measurable in @xmath211 as the composition of two strongly measurable mappings @xmath322 and a continuous mapping @xmath323 .
any open subset @xmath324 of @xmath16 can be considered as a metric measure space @xmath325 .
let us identify @xmath326 as a subspace in @xmath126 by extending functions outside @xmath324 by @xmath88 .
define @xmath327 as the closure of @xmath328 in @xmath131 so that @xmath327 is a subspace of both @xmath329 and @xmath330 .
then @xmath331 is a regular strongly local dirichlet form in @xmath330 , which is called the restriction of @xmath34 to @xmath324 .
let @xmath332 be the generator of the form @xmath333 and @xmath334 , @xmath335 , be the _ restricted _ heat semigroup .
define the _ first exit time _ from @xmath324 by@xmath336 the diffusion process associated with the restricted dirichlet form , can be canonically obtained from @xmath337 by killing the latter outside @xmath324 , that is , by restricting the life time of @xmath169 by @xmath338 ( see @xcite ) .
it follows that the transition operator @xmath339 of the killed diffusion is given by@xmath340 for all @xmath341 .
then @xmath342 is defined for @xmath116 from other function classes in the same way as @xmath173 .
also , extend @xmath343 to all @xmath344 by setting it to be @xmath88 if @xmath345 .
we say that the semigroup @xmath150 is _ locally ultracontractive _ if the restricted heat semigroup @xmath346 is ultracontractive for any metric ball @xmath347 of @xmath33 .
[ tptom]let the semigroup @xmath150 be locally ultracontractive .
then the following is true .
there exists a properly exceptional set @xmath256 such that , for any open subset @xmath104 , the semigroup @xmath348 possesses the heat kernel @xmath349 with the domain @xmath350 .
if @xmath351 are open subsets of @xmath16 , then @xmath352 for all @xmath1 , @xmath353 . if @xmath354 is an increasing sequence of open subsets of @xmath16 and @xmath355 , then @xmath356 as @xmath357 for all @xmath1 , @xmath358
. set @xmath290 .
let @xmath359 $ ] be an upper semi - continuous function such that , for some open set @xmath360 and for some @xmath1,@xmath361 for almost all @xmath362 .
then ( [ ptfi ] ) holds for all @xmath358 . for simplicity of notation ,
set @xmath363 to be @xmath88 for all @xmath36 outside @xmath324 ( which , however , does not mean the extension of the domain of @xmath364 ) .
proof of theorem [ tptom ] ( a ) since the metric space @xmath33 is separable , there is a countable family of balls that form a base .
let @xmath365 be the family of all finite unions of such balls so that @xmath366 is countable and any open set @xmath104 can be represented as an increasing union of sets of @xmath366 .
since any set @xmath367 is contained in a metric ball , the semigroup @xmath368 is dominated by @xmath346 and , hence , is ultracontractive . by theorem [ tbbck ] , there is a truly exceptional set @xmath369 such that the @xmath370 has the heat kernel @xmath371 in the domain @xmath372 . since the family @xmath366 is countable , the set@xmath373 is properly exceptional .
let us first show that if @xmath374 , @xmath375 are the sets from @xmath366 and @xmath376 , then @xmath377 it follows from ( [ ptom ] ) that , for any @xmath378,@xmath379 that is,@xmath380 setting here @xmath381 where @xmath382 , we obtain@xmath383 setting in ( [ om12 ] ) @xmath384 , we obtain@xmath385 combining the above two lines gives ( [ pt12 ] ) .
let @xmath324 be any open subset of @xmath16 and @xmath386 be an increasing sequence of sets from @xmath366 such that @xmath387 .
let us set@xmath388 this limit exists ( finite or infinite ) by the monotonicity of the sequence @xmath389 .
it follows from ( [ ptom ] ) that , for any @xmath341,@xmath390 by the monotone convergence theorem , we obtain@xmath391 for all @xmath1 and @xmath392 . comparing the above two lines , we obtain@xmath393 the symmetry of @xmath363 is obvious from ( [ ptomlim ] ) , and the semigroup property of @xmath364 follows from that of @xmath394 by the monotone convergence theorem . note that @xmath395 does not depend on the choice of @xmath396 by the uniqueness of the heat kernel ( lemma [ lemuniq ] ) .
\(b ) for two arbitrary open sets @xmath397 let @xmath398 and @xmath399 be increasing sequences of sets from @xmath366 that exhaust @xmath400 and @xmath401 , respectively . set @xmath402 so that @xmath403 and @xmath401 is the increasing union of sets @xmath404 ( see figure [ pic6 ] ) .
then @xmath405 and , hence , @xmath406 , which implies as @xmath254 that @xmath407 . .
] \(c ) let @xmath408 be an increasing sequence of open sets whose union is @xmath324 .
let @xmath409 be an increasing sequence of sets from @xmath366 that exhausts @xmath410 .
as in the previous argument , we can replace @xmath411 by @xmath412 so that @xmath413 .
rename @xmath414 back to @xmath411 , and assume in the sequel that @xmath415 .
similarly , replace @xmath416 by @xmath417 and assume in the sequel that @xmath418 .
arguing by induction , we redefine the double sequence @xmath419 in the way that it is monotone increasing not only in @xmath420 but also in @xmath421
. then we claim that @xmath422 indeed , if @xmath423 then @xmath424 for some @xmath421 and , hence , @xmath425 for some @xmath420 , which implies @xmath426 for @xmath427 .
finally , we have @xmath428 and@xmath429 whence it follows that@xmath430 \(d ) let @xmath431 be subset of @xmath324 .
then the semigroup @xmath368 is ultracontractive and possesses the heat kernel @xmath371 with the domain @xmath432 where @xmath433 is a truly exceptional set as in part ( a ) .
note that @xmath434 .
since @xmath435 in @xmath436 , we obtain by hypothesis that @xmath437 for almost all @xmath438 . by lemma [ lembgk ] , we conclude that this inequality is true for all @xmath439 .
exhausting @xmath324 be a sequence of subsets @xmath431 and using ( [ ptomlim ] ) , we obtain ( [ ptfi ] ) .
a priori we assume here only the basic hypotheses . all necessary additional assumptions are explicitly stated .
the main result of this section is theorem [ tg=>fk ] .
given an open set @xmath104 , define the green operator @xmath440 first for all @xmath341 by@xmath441 for all @xmath180 , where we admit infinite values of the integral . if @xmath116 @xmath442 and @xmath443 then @xmath444 is also defined by @xmath445 we have , for any open @xmath104 and all @xmath446,@xmath447 for any @xmath448 . in particular,@xmath449 indeed , integrating ( [ ptom ] ) in @xmath100 , we obtain@xmath450 obviously , ( [ gtau ] ) follows from ( [ gom ] ) for @xmath451 .
denote by @xmath452 the bottom of the spectrum of @xmath332 in @xmath453 , that is,@xmath454 for any open set @xmath104 , we will consider the _ mean exit time _
@xmath455 from @xmath324 as a function of @xmath456 .
also , set@xmath457 [ lg1 - 1]if @xmath458 then @xmath440 is a bounded operator on @xmath459 and it uniquely extends to each of the spaces @xmath460 , @xmath461 , @xmath462 , with the following norm estimates:@xmath463 \label{g1 - 1 } \vert g^{\omega}\vert_{l^{1}\rightarrow l^{1}}&\leq&\widetilde { e } ( \omega ) , \\[-1pt ] \label{g2 - 2 } \vert g^{\omega}\vert_{l^{2}\rightarrow l^{2}}&\leq&\widetilde { e } ( \omega ) .\end{aligned}\ ] ] moreover , @xmath464 and @xmath440 is the inverse in @xmath330 to the operator @xmath332 .
it follows from ( [ gtau ] ) that@xmath465 which implies that for any @xmath466 , @xmath467 hence , @xmath440 can be considered as a bounded operator in @xmath468 with the norm estimate ( [ gi - i ] ) .
estimate ( [ g1 - 1 ] ) follows from ( [ gi - i ] ) by duality . indeed , for any two functions @xmath469 , we have@xmath470 which follows from ( [ gdef ] ) and the symmetry of @xmath471 .
by linearity , ( [ gfh ] ) extends to all @xmath472 .
then , for any @xmath473 , we have@xmath474 & = & \sup_{h\in\mathcal{b}_{b } ( \omega ) \setminus \ { 0\ } } \frac{\int_{\omega}fg^{\omega}h\,d\mu}{\vert h\vert _ { \infty } } \\[-1pt ] & \leq&\sup_{h\in\mathcal{b}_{b } ( \omega ) \setminus \ { 0\ } } \frac{\vert g^{\omega}h\vert_{\infty}\vert f\vert _ { 1}}{\vert h\vert_{\infty } } \\[-1pt ] & \leq&\widetilde{e } ( \omega ) \vert f\vert_{1},\end{aligned}\ ] ] whence it follows that @xmath440 uniquely extends to a bounded operator in @xmath475 with the norm estimate ( [ g1 - 1 ] ) . the estimate ( [ g2 - 2 ] ) [ as well as a similar estimate for @xmath476 for any @xmath477 follows from ( [ gi - i ] ) and ( [ g1 - 1 ] ) by the riesz
thorin interpolation theorem . to prove ( [ lamin < ] ) , let us consider the following `` cut - down '' version of the green operator:@xmath478 where @xmath479 .
the same argument as above shows that @xmath480 can be considered as an operator in @xmath481 with the same norm bound@xmath482 on the other hand , using the spectral resolution @xmath483 of the generator @xmath484 , we obtain , for any @xmath473,@xmath485\\[-8pt ] & = & \int_{0}^{\infty}\varphi_{t } ( \lambda ) \,de_{\lambda}f \nonumber\\ & = & \varphi_{t } ( \mathcal{l}^{\omega } ) f,\nonumber\end{aligned}\ ] ] where@xmath486 since @xmath487 is a bounded function on @xmath488 , the operator @xmath489 is a bounded operator in @xmath481 . by the spectral mapping theorem , we obtain@xmath490 on the other hand , since @xmath491 is decreasing in @xmath492,@xmath493 whence@xmath494 by letting @xmath495 and observing that @xmath496 , we obtain@xmath497 which in particular means that @xmath498 .
consequently , the operator @xmath332 has a bounded inverse . passing in ( [ fit ] ) to the limit as @xmath495
, we obtain @xmath499 .
let @xmath324 be an open subset of @xmath16 .
we say that a function @xmath500 is _ harmonic _ in @xmath324 if@xmath501 [ lemg - g]let @xmath324 be an open subset of @xmath16 such that @xmath502 , and let @xmath276 be an open subset of @xmath324 . for any @xmath503 ,
the function @xmath444 is harmonic in @xmath276.[remdoweneeda ? ] for any @xmath504 , the function @xmath505 is harmonic in @xmath276 .
if @xmath506 then @xmath507 is defined as the extension of @xmath508 to @xmath324 by setting it to be equal to @xmath88 in @xmath509 .
proof of lemma [ lemg - g ] ( a ) set @xmath510 . to prove that @xmath52 is harmonic in @xmath276 , we need to show that @xmath511 , for any @xmath512 .
since by lemma [ lg1 - 1 ] @xmath513 , we have @xmath514 .
therefore , by the definition of @xmath515,@xmath516 \(b ) set @xmath517 .
any function @xmath512 can be considered as an element of @xmath518 by setting it to be @xmath88 in @xmath509 .
then both @xmath52 and @xmath519 are in @xmath327 whence@xmath520 denote by @xmath521 the open metric ball of radius @xmath51 centered at a point @xmath50 , and set@xmath522 that @xmath25 has full support implies @xmath523 whenever @xmath51 .
whenever we use the function @xmath524 , we always assume that @xmath525 for example , this condition is automatically satisfied if all balls are precompact . however , we do not assume precompactness of all balls unless otherwise explicitly stated .
we say that the _ elliptic harnack inequality _ ( [ condh ] ) holds on @xmath16 , if there exist constants @xmath526 and @xmath527 such that , for any ball @xmath59 in @xmath16 and for any function @xmath500 that is nonnegative and harmonic @xmath528 , @xmath529 we say that the _ volume doubling _ property ( [ vd ] ) holds if there exists a constant @xmath15 such that , for all @xmath50 and @xmath51[condvd ] @xmath530 it is known that ( [ vd ] ) implies that , for all @xmath79 and @xmath531 , @xmath532 for some @xmath29 ( see @xcite ) .
assume that @xmath533 hold .
let @xmath534 be an open subset of @xmath16 such that @xmath535 and let @xmath536 be a ball contained in @xmath324 .
[ lgom ] for any nonnegative function @xmath537 , [ remdoweneeda?copy1]@xmath538 [ lemgb - gb]for and any nonnegative function @xmath539,@xmath540 \(a ) since identity ( [ gom < e ] ) survives monotone increasing limits of sequences of functions @xmath541 , it suffices to prove ( [ gom < e ] ) for any nonnegative function @xmath542 .
then , by lemma [ lemg - g ] , the function @xmath543 is harmonic in @xmath544 . since @xmath545
, we can use the harnack inequality ( [ condh ] ) in ball @xmath347 , which yields@xmath546 \(b ) assume first that @xmath547 . by lemma [ lemg - g ] ,
the function @xmath548 is harmonic in @xmath59 . since @xmath549 , applying for this function argument ( [ fi1 ] ) , we obtain ( [ gbb ] ) .
an arbitrary nonnegative function @xmath550 can be represented as a sum in @xmath551@xmath552 where @xmath553 . applying ( [ gbb ] ) to each @xmath554 and summing up , we obtain ( [ gbb ] ) for @xmath541 .
a classical theorem of faber and krahn says that for any bounded open set @xmath555 , @xmath556 where @xmath347 is a ball in @xmath5 of the same volume as @xmath534 . if the radius of @xmath347 is @xmath557 , then
@xmath558 where @xmath14 is a positive constant depending only on @xmath420 , which implies that @xmath559 cf .
we refer to lower estimates of @xmath452 via a function of @xmath560 as _ faber
krahn inequalities_. a more general type of a faber
krahn inequality holds on a complete @xmath420-dimensional riemannian manifold @xmath16 of nonnegative ricci curvature : for any bounded open set @xmath561 and for any ball @xmath347 of radius @xmath557 containing @xmath534,[condfk0]@xmath562 where @xmath563 and @xmath564 ( see @xcite ) .
note that ( [ fkrn ] ) follows from ( [ fkrel ] ) ( apart from the sharp value of the constant @xmath14 ) because in @xmath5 we have @xmath565 .
it was proved in @xcite that , on any complete riemannian manifold , @xmath566 where ( _ ue _ ) is here the upper bound of the heat kernel in the li
yau estimate ( [ liyau ] ) . in section [ secdue ]
we will derive a general upper bound ( [ uee ] ) from a set of hypotheses containing a suitable version of ( [ fkrel ] ) . in this section
, we will deduce a faber
krahn inequality from the main hypotheses .
we fix from now on a function @xmath567 that is a continuous increasing bijection of @xmath568 onto itself , such that , for all @xmath569,@xmath570 for some constants @xmath571 , @xmath526 . in the sequel
we will use the inverse function @xmath572 .
it follows from ( [ fb ] ) that@xmath573 for all @xmath574 .
we say that the faber
krahn inequality ( [ fk ] ) holds if , for any ball @xmath347 in @xmath16 of radius @xmath557 and any open set @xmath575,[condfk ] @xmath576 with some positive constants @xmath577 .
we say that the mean exit time estimate ( [ ef ] ) holds if , for all @xmath180 and @xmath51,[condef ] @xmath578 with some constant @xmath526 .
we denote by @xmath579 and @xmath580 the upper and lower bounds of @xmath581 in ( [ eff ] ) , respectively .
[ tg=>fk]the hypotheses @xmath582 imply ( [ fk ] ) .
we have by ( [ lamin < ] ) and ( [ gtau ] ) @xmath583 it will be convenient to rename @xmath584 to @xmath585 and let the original ball @xmath347 be @xmath586 and @xmath587 .
fix a point @xmath344 so that @xmath588 , consider a numerical sequence @xmath589 , @xmath590 where @xmath591 is the parameter from ( [ condh ] ) , and the balls @xmath592 .
we have@xmath593 where @xmath420 is to be chosen ( see figure [ pic3 ] ) , whence@xmath594 . ] setting @xmath595 and using @xmath596 , we obtain , by lemma [ lemgb - gb],@xmath597 also , by ( [ gom1]),@xmath598 hence , collecting together the previous lines , we obtain@xmath599 choose any @xmath600 so that @xmath601 . using ( [ fb ] ) ,
( [ va ] ) and the monotonicity of @xmath602 , we obtain@xmath603 now choose @xmath420 from the condition @xmath604 and set @xmath605 .
we obtain then@xmath606\\[-8pt ] & \leq&cf ( r ) \biggl ( \frac{v ( r ) } { v ( r ) } \biggr ) ^{\nu}+cf ( r ) .\nonumber\end{aligned}\ ] ] using again ( [ va ] ) , ( [ fb ] ) and @xmath607 , we obtain @xmath608 which implies that the second term in ( [ esupg2 ] ) can be absorbed by the first one , thus giving @xmath609 since the point @xmath344 was arbitrary , covering @xmath324 by a countable family of balls like @xmath610 , we obtain @xmath611 which together with ( [ lag ] ) finishes the proof .
our main result in this section is theorem [ tef ] saying that the condition ( [ eff ] ) implies a certain upper bound for the tail @xmath612 of the exit time from balls .
the results of this type go back to barlow @xcite , theorem 3.11 . here
we give a self - contained proof in the present setting , which is based on the ideas of @xcite .
an alternative analytic approach can be found in @xcite .
for any open set @xmath104 , set@xmath613 [ lttail]for any open @xmath104 such that @xmath614 , we have , for all @xmath1 and @xmath615,@xmath616 denote @xmath617 , and observe that@xmath618 where @xmath619 is the time shift of trajectories . using the markov property , we obtain , for any @xmath448,@xmath620 whence @xmath621 ( see figure [ pic7 ] ) , . ] and ( [ px < ] ) follows . [ le<1-e]assume that the condition ( [ eff ] ) is satisfied
. then there are constants @xmath622 such that , for all @xmath180 , @xmath623 , and @xmath624,@xmath625 denoting @xmath626 and using lemma [ lttail ] , we have , for any , @xmath627 the condition ( [ eff ] ) implies that@xmath628 whence@xmath629 using these two estimates of @xmath630 , we obtain@xmath631 setting @xmath632 and choosing @xmath633 , we obtain@xmath634 if also @xmath635 , then we obtain ( [ exe ] ) .
clearly , the former condition will be satisfied provided@xmath636 which finishes the proof .
[ pf < exp]assume that the condition ( [ eff ] ) is satisfied .
then there exists constant @xmath637 such that , for all precompact balls @xmath638 with @xmath180 and for all @xmath639,@xmath640 where @xmath572 .
rename the center @xmath60 of the ball to @xmath641 so that the letter @xmath60 will be used to denote a variable point .
fix some @xmath531 to be specified later , and set @xmath642 $ ] .
set also @xmath643 , @xmath644 and@xmath645 where @xmath646 .
note that all @xmath647 are bounded by @xmath184 .
choose @xmath648 where @xmath85 is the constant from lemma [ le<1-e ] , and let @xmath649 be a point in @xmath650 for which @xmath651 fix @xmath652 , observe that @xmath653 and consider the following function in @xmath654:@xmath655 where @xmath656 ( see figure [ pic4 ] ) . since the ball @xmath657 is precompact , we have @xmath658 ( while for noncompact balls the exit point could have been at the cemetery ) . and @xmath659
] let us show that , for all @xmath660,@xmath661 indeed , we have by the strong markov property@xmath662 which proves ( [ u < uu ] ) .
it follows from ( [ u < uu ] ) that@xmath663 whence@xmath664 by lemma [ le<1-e ] , if @xmath665 then @xmath666 .
therefore , under hypothesis ( [ lar ] ) , we have@xmath667 whence it follows by iteration that@xmath668 where in the last inequality we have used that @xmath669 and @xmath670 .
condition ( [ lar ] ) can be satisfied by choosing @xmath671 this value of @xmath557 is legitimate only if @xmath672 , that is , if@xmath673 if ( [ r > ] ) is not fulfilled , then ( [ ex0 ] ) is trivially satisfied by choosing the constant @xmath15 large enough . assuming that ( [ r > ] ) is satisfied and defining @xmath557 by ( [ n= ] ) we obtain from ( [ n ] ) that@xmath674 whence ( [ ex0 ] ) follows .
[ tef]assume that ( [ ef ] ) holds .
then , for any precompact ball @xmath638 with @xmath180 and for any @xmath1 , @xmath675 where @xmath637 is the constant from lemma [ pf < exp ] and @xmath676 changing the variable @xmath557 in ( [ fidef ] ) , we obtain the following equivalent definitions of @xmath677 : @xmath678 where @xmath572 . proof of theorem [ tef ] denoting @xmath626 and using lem - ma [ pf < exp ] , we obtain that , for any @xmath639,@xmath679 taking the supremum in @xmath492 and using ( [ fidef2 ] ) , we obtain ( [ psi ] ) .
[ remfi ] it is clear from ( [ fidef ] ) that function @xmath680 is increasing in @xmath584 and decreasing in @xmath100 . also , we have , for any constants @xmath681,@xmath682 in particular , it follows that @xmath683 where@xmath684 hence , ( [ psi ] ) can be written also in the form@xmath685 clearly , @xmath686 .
let us show that @xmath687 for all @xmath688 .
since@xmath689 we see from ( [ fidef1 ] ) that @xmath690 .
it follows from ( [ fb ] ) and @xmath30 that @xmath691 given @xmath688 , choose @xmath557 so big that @xmath692 [ such @xmath557 exists by the second condition in ( [ frr ] ) ] .
then@xmath693 in order to prove that @xmath694 , it suffices to show that@xmath695 indeed , if @xmath557 is sufficiently small , then by the first condition in ( [ frr ] ) , @xmath696 whence @xmath697 .
another useful property of function @xmath40 is the inequality[remdoweuseit?]@xmath698 indeed , we have for any @xmath51 @xmath699 whence ( [ fi2s ] ) follows by taking @xmath129 in @xmath557 .
if @xmath61 is differentiable then the supremum in ( [ fidef1 ] ) is attained at the value of @xmath557 that solves the equation@xmath700 for example , @xmath62 then we obtain @xmath701 whence @xmath702 and@xmath703 consequently , ( [ psi ] ) becomes@xmath704 [ exf120 ] consider the following example of function @xmath705:@xmath706 where @xmath707 .
it is easy to see that ( [ fb ] ) is satisfied with @xmath708 and @xmath709 .
similarly to the previous example , one obtains that@xmath710 so that ( [ psi ] ) becomes@xmath711 [ lemtifi]the function @xmath712 satisfies the following inequality:@xmath713 for all @xmath714
. by ( [ fidef ] ) , we have , for any @xmath51,@xmath715 we claim that there exists @xmath51 such that@xmath716 indeed , the function @xmath717 is continuous on @xmath718 , tends to @xmath88 as @xmath719 and to @xmath89 as @xmath720 so that @xmath717 takes all positive values , whence the claim follows . with the value of @xmath557 as in ( [ rt ] ) , we have @xmath721 if @xmath722 then using the left - hand side inequality of ( [ fb ] ) , we obtain@xmath723 which together with ( [ rt ] ) yields @xmath724 substituting into ( [ lati ] ) , we obtain@xmath725 similarly , it @xmath726 then using the right - hand side inequality in ( [ fb ] ) we obtain@xmath727 whence it follows that @xmath728 [ cef]under the hypotheses of theorem [ tef ] , we have , for any @xmath180 , @xmath623 , @xmath1 , @xmath729 indeed , if @xmath730 , then ( [ ptau ] ) follows from theorem [ tef ] , lemma [ lemtifi ] and ( [ fb ] ) .
if @xmath731 , then ( [ ptau ] ) is trivial .
the following result will be used in the proof of theorem [ tdue ] below .
[ pdirhk ] let @xmath276 be an open subset of @xmath16 , and assume that , for any nonempty open set @xmath732,@xmath733 for some positive constants @xmath734 .
then the semigroup @xmath735 is ultracontractive with the following estimate:@xmath736 for any @xmath737 .
the next theorem provides pointwise upper bounds for the heat kernel .
[ tdue][tue]if the conditions @xmath738 are satisfied and all balls are precompact then the heat kernel exists with the domain @xmath739 for some properly exceptional set @xmath207 , and satisfies the upper bound[condue ] @xmath740 for all @xmath1 and @xmath741 , where @xmath572 and @xmath677 is defined by ( [ fidef ] ) .
as it follows from theorem [ tg=>fk ] , the hypotheses @xmath742 here can be replaced by @xmath743 . also , using ( [ fifi ] ) , one can write ( [ uee ] ) in the form@xmath744 as it was stated in the . a version of theorem [ tdue ] was proved by kigami @xcite under additional assumptions that the heat kernel is a priori continuous and ultracontractive , and using instead of ( [ fk ] ) a local nash inequality . in the case
@xmath62 , another version of theorem [ tdue ] was proved in @xcite , where the upper bound ( [ uee ] ) was understood for _ almost all _ @xmath36 .
the proof below uses a combination of techniques from @xcite and @xcite .
if function @xmath61 is given by ( [ f120 ] ) as in example [ exf120 ] , then @xmath40 is given by ( [ fi120 ] ) and ( [ uee ] ) becomes@xmath745 where @xmath746 .
proof of theorem [ tdue ] the hypothesis ( [ fk ] ) can be stated as follows : for any ball @xmath536 where @xmath50 and @xmath51 , and for any nonempty open set @xmath747 , we have @xmath748 where@xmath749 and @xmath750 are positive constants .
hence , ( [ fk ] ) implies by proposition [ pdirhk ] that@xmath751 in particular , the semigroup @xmath752 is ultracontractive and @xmath162 is locally ultracontractive . by theorem [ tptom ] ,
there exists a properly exceptional set @xmath256 ( containing @xmath208 ) such that , for any open subset @xmath104 , the semigroup @xmath348 possesses the heat kernel @xmath363 with the domain @xmath753 .
fix this set @xmath207 for what follows . by theorem [ tbbck ] , ( [ ab ] ) and ( [ bultra ] )
imply the following estimate : @xmath754 for any ball @xmath347 of radius @xmath557 , and for all @xmath1 , @xmath755 .
our next step is to prove the on - diagonal estimate @xmath756 for all @xmath160 and @xmath1 . to understand the difficulties ,
let us first consider a particular case when the volume function satisfies the following estimate : @xmath757 for all @xmath50 and @xmath623 , where @xmath758 is the exponent in ( [ fk ] ) [ e.g. , ( [ vf ] ) holds , if @xmath759 , @xmath760 and @xmath761 . in this case ,
the value @xmath762 in ( [ fk ] ) cancels our , and we obtain @xmath763 hence , by proposition [ pdirhk ] , the semigroup @xmath764 is ultracontractive , and by theorem [ tbbck ] we obtain the estimate@xmath765 for all @xmath160 and @xmath1 .
observing that@xmath766 we see that ( [ t1nu ] ) is equivalent to ( [ due ] ) .
although this argument works only under restriction ( [ vf ] ) , it has an advantage that it can be localized as follows . assuming that ( [ vf ] ) is satisfied for all @xmath767 with some fixed constant @xmath768 , ( [ fkc ] ) is satisfied for all open sets @xmath324 with a bounded value of @xmath769 , and ( [ eff ] ) is satisfied for all balls with a bounded value of the radius , one can prove in the same way that ( [ t1nu ] ) is true for @xmath770 for some @xmath771 .
the proof below does not allow such a localization in the general case.=-1 in the general case , without the hypotheses ( [ vf ] ) , the heat semigroup @xmath162 is not necessarily ultracontractive , which requires other tools for obtaining ( [ due ] ) . in the case of riemannian manifolds , one can obtain ( [ due ] ) from ( [ fk ] ) using a certain mean value inequality ( see @xcite ) but this method heavily relies on a specific property of the distance function that , which is not available in our generality .
we will use kigami s iteration argument that allows us to obtain ( [ due ] ) from ( [ ptb ] ) using , in addition , the hypothesis ( [ eff ] ) .
this argument is presented in an abstract form in @xcite , lemma 5.6 , that says the following .
assume that the following two conditions are satisfied : for any ball @xmath536,@xmath772 where function @xmath773 satisfies certain conditions should be monotone decreasing in @xmath100 and should satisfy the following doubling condition : if @xmath774 and @xmath775 , then @xmath776 for some constant @xmath777 .
this is obviously satisfied for the function @xmath778 given by ( [ psidef ] ) . ] ; for all @xmath180 , @xmath1 , and @xmath779,@xmath780 where @xmath81 is a sufficiently small where @xmath777 is the constant from the conditions for @xmath778 . ]
constant and @xmath541 is a positive increasing function on @xmath781 such that@xmath782 then the heat kernel on @xmath16 satisfies the estimate@xmath783 obviously , ( [ ptb ] ) implies ( [ psit ] ) with the function @xmath784 by corollary [ cef ] , ( [ eff ] ) implies ( [ ptau ] ) , which means that ( [ pxtau ] ) is satisfied provided @xmath785 for a sufficiently large constant @xmath15 ; hence , the function @xmath786 can be chosen as follows : @xmath787 which clearly satisfies ( [ intfi ] ) [ indeed , by ( [ rb ] ) , we have @xmath788 for all @xmath789 , whence ( [ intfi ] ) follows ] . by ( [ ptxx ] )
, we obtain@xmath790 where we have also used ( [ rb ] ) and ( [ va ] ) . by theorem [ tptom](d ) , we can replace here @xmath791 by @xmath792 outside @xmath207 , whence ( [ due ] ) follows .
now we prove the full upper estimate ( [ uee ] ) .
fix two disjoint open subsets @xmath793 of @xmath16 and use the following inequality proved in @xcite , lemma 2.1 : for all functions @xmath794,@xmath795\\[-9pt ] & & { } + \bigl(\mathbb { e}_{\cdot}\bigl(\mathbf{1}_{\{\tau_{v}\leq t/2\}}\mathcal{p}_{t-\tau_{v}}g(x_{\tau _ { v}})\bigr),f\bigr)\nonumber\end{aligned}\ ] ] ( see figure [ pic5 ] ) . ) . ]
assume in addition that @xmath796 and @xmath797 .
then , under the condition @xmath798 , we have@xmath799 almost surely , where@xmath800 here we have used that @xmath801 almost surely , which is due to the fact that @xmath337 is a diffusion and the set @xmath207 is properly exceptional .
it follows that@xmath802 estimating similarly the second term in ( [ ptfgb ] ) , we obtain from ( [ ptfgb])@xmath803 & & \qquad\leq s\vert f\vert _ { 1}\int_{u}\mathbb{p}_{x}\biggl ( \tau_{u}\leq\frac{t}{2}\biggr ) g ( x ) \,d\mu ( x)\\[-2pt ] & & \qquad\quad { } + s\vert g\vert_{1}\int_{v}\mathbb{p}_{y}\biggl ( \tau _ { v}\leq \frac{t}{2}\biggr ) f ( y ) \,d\mu ( y ) .\end{aligned}\ ] ] by @xcite , lemma 3.4 , we conclude that , for @xmath804-a.a .
@xmath805 and @xmath806 , @xmath807 a slightly different inequality ( i.e. , also enough for our purposes ) was proved in @xcite . for the case of heat kernels on riemannian manifolds , ( [ p < pb2 ] )
was proved in @xcite , lemma 3.3 . by theorem [ tef ] , we have@xmath808 for all @xmath160 and @xmath809 .
let @xmath810 then , for any @xmath811 , we obtain by ( [ psi1])@xmath812 using a similar estimate for @xmath813 , we obtain from ( [ p < pb2 ] ) that , for @xmath25-a.a .
@xmath814 and @xmath813,@xmath815 since the right - hand side here is a constant in @xmath36 , we conclude by theorem [ tptom](d ) that ( [ ptro ] ) holds for all @xmath816 and @xmath817 .
now fix two distinct points @xmath741 , set @xmath818 and observe that the balls @xmath819 and @xmath820 are disjoint .
since @xmath814 and @xmath813 , we conclude that ( [ ptro ] ) is satisfied for these points @xmath36 with the above value of @xmath584 .
let us estimate the quantity @xmath821 defined by ( [ s ] ) .
using the semigroup property and ( [ due ] ) , we obtain , for all @xmath822,@xmath823 observe that by ( [ va ] ) , for all @xmath824,@xmath825 applying this for @xmath826 and @xmath827 so that @xmath828 and @xmath829 , and substituting to the above estimate of @xmath830 we obtain@xmath831 using that @xmath832 $ ] as well as ( [ rb ] ) and ( [ va ] ) , we obtain @xmath833 which together with ( [ ptro ] ) yields @xmath834 on the other hand , we have by ( [ fidef2 ] ) @xmath835 where we have chosen @xmath836 . using the elementary estimate@xmath837 and its consequence@xmath838 we obtain@xmath839 whence@xmath840 choosing @xmath841 small enough and substituting this estimate to ( [ pts ] ) , we obtain ( [ uee ] ) .
[ rembdd]it is desirable to have a localized version of theorem [ tdue ] when the hypotheses are assumed for balls of bounded radii and the conclusions are proved for a bounded range of time . as was already mentioned in the proof , kigami s
argument requires the ultracontractivity of @xmath842 for _ all _ balls , and ( [ eff ] ) should also be satisfied for all balls , because , loosely speaking , one deals with the estimate of @xmath843 for an exhausting sequence of balls @xmath844 ( see @xcite or @xcite ) . as we will see in section [ secrvd ] , the hypotheses @xmath845 for all balls imply that the space @xmath33 is unbounded .
note that all other arguments used in this paper do admit localization .
the harnack inequality ( [ condh ] ) is a standing assumption in this subsection .
the main result is proposition [ posc ] that is heavily based on the oscillation inequality of lemma [ lemosc ] .
the latter is considered as a standard consequence of ( [ condh ] ) , but we still provide a full proof to emphasize the use of the precompactness of balls .
[ lemh]let @xmath347 be a precompact ball of radius @xmath557 in @xmath16 . if @xmath846 is harmonic in @xmath347 , and if @xmath847 in @xmath347 for some real constant @xmath841 , then@xmath848 where @xmath15 and @xmath849 are the same constants as in ( [ condh ] ) .
let @xmath850 be a cutoff function of @xmath347 in @xmath16 , that is , @xmath851 and @xmath852 in an open neighborhood of @xmath853 .
the function @xmath854 belongs to @xmath131 and is equal to @xmath855 in @xmath347 .
let us show that @xmath519 is harmonic in @xmath347 .
indeed , for any @xmath856 , we have@xmath857 because @xmath858 by the harmonicity of @xmath52 in @xmath347 , and @xmath859 by the strong locality as @xmath852 in a neighborhood of @xmath860 . applying ( [ condh ] ) to @xmath519
, we obtain ( [ h - a ] ) . for any function @xmath116 on any set @xmath861 , define the essential oscillation of @xmath116 in @xmath821 by@xmath862 the following statement
is well known for functions in @xmath5 ( see , e.g. , @xcite and @xcite , lemma 2.3.2 ) .
[ lemosc]there exists @xmath863 such that , for any precompact ball @xmath864 , for any nonnegative harmonic function @xmath52 in @xmath59 , and any @xmath865,@xmath866 where @xmath867 is a positive constant that depends on the constants in ( [ condh ] ) .
fix @xmath50 , and write for simplicity @xmath868
. consider first the case when @xmath869 [ where @xmath849 is a parameter from ( [ condh ] ) ] and set@xmath870 and@xmath871 clearly , @xmath872 . by lemma [ lemh ] , we have@xmath873 that is,@xmath874 similarly , applying lemma [ lemh ] to function @xmath875 , we obtain@xmath876 whence@xmath877 adding up the two inequalities yields@xmath878 whence@xmath879 that is,@xmath880 where @xmath881 . for an arbitrary @xmath882 ,
find a nonnegative integer @xmath421 such that @xmath883 iterating ( [ der ] ) , we obtain@xmath884 note that the constant @xmath15 in ( [ esiu ] ) can be assumed to be big enough , say @xmath885 .
then @xmath886 and ( [ eosc ] ) follows from the previous line with @xmath887 .
[ posc]let @xmath324 be an open subset of @xmath16 such that @xmath502 . fix a function @xmath888 and set @xmath510 .
then , for any precompact ball @xmath889 and all @xmath890$],@xmath891 where @xmath867 is the same constant as in lemma [ lemosc ] and@xmath892 write for simplicity @xmath893 . let us first prove that if @xmath236 , then @xmath894 by lemma [ lg1 - 1 ] , we have for the function @xmath895 that@xmath896 the function @xmath897 is harmonic in @xmath898 by lemma [ lemg - g ] and nonnegative by theorem [ tptom](b ) . by lemma [ lemosc ] , we obtain@xmath899 since @xmath900 , ( [ osc+ ] ) follows by adding up the two previous lines .
for a signed function @xmath116 , write @xmath901 and set@xmath902 then @xmath903 and @xmath904 are nonnegative and @xmath905 , whence it follows that@xmath906 applying ( [ osc+ ] ) separately to @xmath903 and @xmath907 and adding up the inequalities , we obtain ( [ oscu ] ) . in this section
, we assume only the basic hypotheses .
if @xmath217 , then , for any @xmath1 , the function @xmath908 is in @xmath909 and satisfies the heat equation@xmath910 where @xmath911 is the strong derivative in @xmath912 of the mapping @xmath913 ; cf .
@xcite and @xcite , section 4.3 .
the argument in the next lemma is well known in the context of the semigroup theory ( see @xcite ) , and we reproduce it here for the sake of completeness . for any @xmath914 and all @xmath915 , we have@xmath916 where @xmath917 .
let @xmath918 be spectral resolution in @xmath126 of the operator @xmath142 .
then we have@xmath919 whence@xmath920 since @xmath921 we obtain@xmath922 which was to be proved
. in the rest of this section , assume that @xmath9 is the heat kernel with domain @xmath215 .
[ cdtptl2]for any @xmath1 and @xmath245 , the function @xmath923 is strongly differentiable in @xmath318 and , for all @xmath292,@xmath924 setting @xmath925 for some @xmath81 and using ( [ ptpt ] ) , we obtain that the function@xmath926 satisfies ( [ dudt ] ) , that is,@xmath927 renaming @xmath928 by @xmath100 and @xmath929 by @xmath930 , we finish the proof .
set@xmath931 where the strong derivative @xmath932 is taken with respect to the first variable @xmath60 .
hence , for any @xmath245 and @xmath1 , @xmath933 is an @xmath481-function of @xmath60 .
[ lemqtpt]for all @xmath292 and @xmath245 , we have@xmath934 indeed , we have by ( [ ptpt])@xmath935 since @xmath936 is a bounded operator in @xmath481 , it commutes with the operator @xmath932 of strong derivation . applying the latter to the both sides of the above identity , we obtain ( [ p ] ) .
[ cdtpt]for all @xmath1 , @xmath245 and @xmath25-a.a .
@xmath209 , @xmath937 by lemma [ lemqtpt ] , we have , for all @xmath245 and @xmath25-a.a .
@xmath209,@xmath938 whence by corollary [ cdtptl2 ] , @xmath939 for any @xmath940 . choosing @xmath941 , we finish the proof of ( [ dtpt ] ) .
it follows easily from the identity@xmath942 that the function @xmath943 is differentiable for all fixed @xmath211 and@xmath944 arguing as in the previous proof , we obtain@xmath945 for all @xmath1 and @xmath211 .
however , for applications we need estimate ( [ dtpt ] ) for the strong derivative @xmath946 rather than for the partial derivative @xmath947 .
[ lemgu]if @xmath324 is an open subset of @xmath16 and if @xmath948 , then , for all @xmath1 and @xmath824 , the function @xmath949 satisfies in @xmath950 the equation@xmath951 by lemma [ lg1 - 1 ] , the green operator @xmath440 is a bounded operator in @xmath330 , and @xmath440 is the inverse operator to @xmath332 .
since the function @xmath952 satisfies the equation @xmath953 , applying @xmath440 proves the claim . in this subsection
we use the hypotheses @xmath582 .
as it follows from theorem [ tg=>fk ] and proposition [ pdirhk ] , under these hypotheses the heat semigroup @xmath764 is locally ultracontractive .
hence , by theorem [ tptom ] , for any open set @xmath104 , the heat kernel @xmath364 exists with the domain @xmath350 where @xmath954 is a fixed properly exceptional set ; cf .
the beginning of the proof of theorem [ tdue ] .
[ lemoscpb]let the hypotheses @xmath582 be satisfied , and let @xmath324 be an open subset of @xmath16 .
fix @xmath1 , @xmath955 and set @xmath956 where @xmath32 is the exponent from ( [ fb ] ) , and @xmath867 is the constant from lemma [ lemosc ] .
fix also a point @xmath957 and assume that the ball @xmath958 is precompact , and its closure is contained in @xmath324 .
then@xmath959 where @xmath960 and @xmath15 depends on the constants in @xmath579 and ( [ fb ] ) . by construction in the proof of theorem [ tptom ] ,
the heat kernel @xmath364 is obtained as a monotone increasing limit of @xmath394 as @xmath254 where @xmath961 is an exhaustion of @xmath324 by sets @xmath962 that are finite union of balls from a countable base and the convergence is pointwise in @xmath350 .
suppose for a moment that we have proved ( [ oscpb ] ) for @xmath962 instead of @xmath324 , that is,@xmath963\\[-8pt ] & & \qquad \leq\inf_{y\in b ( x,\rho ) \setminus \mathcal{n}}p_{t}^{u_{n } } ( x , y ) + c\biggl ( \frac{\rho}{\mathcal { r } ( t ) } \biggr ) ^{\theta}\sup_{y\in b ( x , r ) \setminus\mathcal{n}}p_{t/2}^{u_{n } } ( y , y)\nonumber\end{aligned}\ ] ] [ note that if @xmath420 is large enough , then @xmath964 . replacing on the right - hand side of ( [ ptun ] ) @xmath394 by a larger value @xmath364 and letting @xmath254 on the left - hand side , we obtain ( [ oscpb ] ) . to prove ( [ ptun ] ) , rename for simplicity @xmath962 into @xmath276 and recall that , by construction in the proof of theorem [ tptom ] , the domain of @xmath371 is @xmath432 where @xmath433 is a truly exceptional set in @xmath276 , that is contained in @xmath207 .
it follows from corollary [ cesup = sup1 ] , that , for any @xmath965 @xmath966 and a similar identity for @xmath967 and @xmath56 .
hence , it suffices to prove that@xmath968 where@xmath969 set@xmath970 where @xmath932 is the strong derivative in @xmath971 with respect to the variable @xmath86 . applying corollary [ cdtpt ] to the heat kernel @xmath371
, we obtain , for @xmath25-a.a .
@xmath972,@xmath973 by lemma [ lemgu ] , we have @xmath974 . since for all @xmath975@xmath976 and @xmath882
, we obtain by proposition [ posc ] and @xmath977 that@xmath978 since @xmath979 , we have by ( [ fb])@xmath980 whence it follows that@xmath981 note that this inequality is true for any @xmath557 such that @xmath982 and @xmath983 . choosing @xmath984 , we obtain ( [ oscpb ] ) .
[ tholder]let the hypotheses @xmath985 be satisfied .
then , for any open set @xmath986 , the heat kernel @xmath363 is hlder continuous in @xmath60 and @xmath86 in @xmath350 .
fix @xmath392 , @xmath1 , and choose @xmath987 so small that @xmath988 where @xmath989 is defined by ( [ r= ] ) .
using theorem [ tdue ] , ( [ vd ] ) , and ( [ fb ] ) , we obtain that , for any @xmath990,@xmath991 where we have used that @xmath992 .
therefore , by lemma [ lemoscpb],@xmath993 in particular , if @xmath994 is close enough to @xmath60 , then we have@xmath995 which means that @xmath996 is hlder continuous in @xmath350 . [ corholder]let the hypotheses @xmath997 be satisfied , and let @xmath638 be a precompact ball , such that @xmath160 . then for all @xmath998 and @xmath100 , such that @xmath999 and for all @xmath1000 , the following estimate holds : @xmath1001 set @xmath1002 . then the condition @xmath1003 from the previous proof is satisfied because @xmath1004 by ( [ r= ] ) and ( [ tror ] )
. hence , ( [ pt - pt ] ) follows from ( [ oscptom ] ) .
[ lemdle]assume that @xmath1005 are satisfied
. then there exists @xmath81 such that , for all precompact balls @xmath638 with @xmath160 and for all @xmath1006,@xmath1007 choose @xmath557 from the condition @xmath1008 which implies @xmath1009 and , hence , @xmath1010 .
hence , it suffices to prove that@xmath1011 setting @xmath536 , we have by ( [ ptom])@xmath1012 by corollary [ cef ] , we obtain@xmath1013 whence it follows that , for small enough @xmath1014,@xmath1015 therefore,@xmath1016 where we have used that @xmath1017 . finally , using ( [ rb ] ) and
( [ va ] ) we obtain ( [ pb > ] ) .
[ tnle]assume that the hypotheses @xmath997 are satisfied , and all metric balls in @xmath47 are precompact
. then there exist @xmath1018 such that[condlle]@xmath1019 for all @xmath623 , @xmath1006 and @xmath1020 , provided @xmath1021 ( see figure [ pic9 ] ) .
consequently , the following . ]
inequality:[condnle ] @xmath1022 holds for all @xmath1 and @xmath741 satisfying ( [ dd ] ) .
obviously , ( [ nle ] ) follows from ( [ lle ] ) by letting @xmath1023 , so that it suffices to prove ( [ lle ] ) .
let @xmath998 and @xmath100 be such that@xmath1024 where @xmath1025 is the constant from lemma [ lemdle ] , and @xmath1026 is to be defined below .
then the hypotheses of lemma [ lemdle ] and corollary [ corholder ] are satisfied .
writing for simplicity @xmath626 , we obtain by ( [ lle ] ) and ( [ pt - pt ] ) that , for any @xmath1027,@xmath1028 choosing @xmath1029 sufficiently small , we obtain ( [ lle ] ) .
combining theorems [ tue ] , [ tholder ] and [ tnle ] , we obtain the main result : [ tmain]if the hypotheses @xmath1030 are satisfied and all metric balls are precompact , then the heat kernel exists , is hlder continuous in @xmath36 , and satisfies ( [ uee ] ) and ( [ nle ] ) . under the hypotheses of theorem [ tmain ] , assume that the volume function @xmath524 satisfies the uniform estimate@xmath1031 with some @xmath29 , and function @xmath705 be as follows : @xmath1032 where @xmath1033.[remwhysuchexampleexists ? ] then @xmath1034 and the heat kernel satisfies the estimate@xmath1035 it follows that @xmath1036 for any @xmath32 from the interval @xmath1037 .
let us verify that the following upper bound fails:@xmath1038 where @xmath746 .
indeed , by ( [ nle ] ) we have@xmath1022 provided @xmath1039 . assuming that @xmath1040 and setting @xmath1041 we obtain with a prescribed distance @xmath1042 can be guaranteed , provided the space @xmath1043 is connected . ] @xmath1044 while it follows from ( [ ptupbad ] ) that@xmath1045 since @xmath1046 , the exponent of @xmath100 under the exponential is negative so that the right - hand side of ( [ pupab ] ) becomes as @xmath1047 much smaller than that of ( [ plowab ] ) , which is a contradiction . another way to see a contradiction is to observe that ( [ ptupbad ] ) implies ( [ eff ] ) with function @xmath1048 ( cf .
@xcite ) , which is incompatible with ( [ eff ] ) with function ( [ f12 ] ) [ although this argument requires the conservativeness of @xmath1049 .
the conclusion is that in general ( [ pupbad ] ) does not imply ( [ ptupbad ] ) . for comparison ,
let us note that if @xmath1050 and the underlying space is a riemannian manifold , then ( [ pupbad ] ) does imply ( [ ptupbad ] ) ; cf .
we say that a sequence @xmath1051 of points in @xmath1052 is an @xmath85-_chain _ between points @xmath79 if@xmath1053 one can view an @xmath85-chain as a sequence of _ chained _ balls @xmath1054 of radii @xmath85 , that connect @xmath60 and @xmath86 ; that is , the center of @xmath1055 is @xmath60 , the center of @xmath1056 is @xmath86 , and the center of @xmath1057 is contained in @xmath1058 for any @xmath1059 -chain connecting @xmath60 and @xmath86 . ] ( see figure [ pic8 ] ) .
for any @xmath81 and all @xmath79 , define@xmath1060 where the infimum is taken over all @xmath85-chains @xmath1061 between @xmath36 with arbitrary @xmath1062 .
it is obvious that @xmath80 is a decreasing left - continuous function of @xmath85 and @xmath1063 furthermore , @xmath1064 it is clear that @xmath82 is an extended metric in the sense that @xmath82 satisfies all the axioms of a metric except for finiteness .
if an @xmath85-chain exists for any two points @xmath36 , then @xmath1065 , and hence @xmath82 is a true metric .
[ lemn]if @xmath1066 for some @xmath79 and @xmath81 , then there exists an @xmath1067-chain @xmath1051 between @xmath36 such that @xmath1068 here stands for the least integer upper bound of the argument .
it follows from ( [ dedef ] ) by the triangle inequality that always @xmath1069 hence , denoting by @xmath84 the minimal value of @xmath1062 for which there exists an @xmath85 chain @xmath1070 between @xmath60 and @xmath86 , we obtain@xmath1071 the number @xmath84 can be also viewed as the minimal number in a sequence of chained balls of radii @xmath85 connecting @xmath60 and @xmath86 . proof of lemma [ lemn ] if @xmath1072 , then also @xmath1073 , and hence @xmath1074 is an @xmath85-chain with @xmath1075 .
assume @xmath1076 , and let @xmath1077 be a @xmath85-chain between @xmath36 , such that@xmath1078 which exists by hypothesis .
set @xmath1079 . then ( [ xixi+1 ] ) implies@xmath1080 whence@xmath1081 if @xmath1082 , then @xmath1083 and @xmath1084 , whence it follows that@xmath1085 hence , there is an index @xmath1086 such that both @xmath1087 and @xmath1088 are smaller than @xmath1089 .
this implies that @xmath1090 so that by removing the point @xmath1091 from the chain we still have an @xmath85-chain . continuing this way , we finally obtain an @xmath85-chain satisfying ( [ n < ] ) .
if @xmath1092 , then it follows from ( [ de > ] ) and ( [ fb ] ) that@xmath1093 in this section , we make an additional assumption that , for all @xmath1094,@xmath1095 in particular , ( [ fe ] ) implies the finiteness of @xmath1096 for all @xmath81 .
define the function @xmath1097 as follows:@xmath1098 if @xmath232 , then @xmath1099 . if @xmath1100 , then it follows from ( [ fe ] ) and ( [ fei ] ) that @xmath1101 .
[ lemfe]if ( [ fe ] ) is satisfied , then the function @xmath1102 satisfies the identity@xmath1103 for all distinct @xmath79 and @xmath1 . since the function @xmath1104 is continuous and @xmath80 is left - continuous in @xmath85 , we have @xmath1105 assume from the contrary that @xmath1106 and note that , for any @xmath1107 , we have by definition of @xmath85 that@xmath1108 on the other hand , @xmath1109 and @xmath1110 hence,@xmath1111 which contradicts ( [ eps ] ) .
[ ttwo]let all metric balls be precompact .
let the hypotheses @xmath1112 and ( [ fe ] ) be satisfied , and let @xmath93 be the function from ( [ epsdef ] ) .
then , for all @xmath741 and @xmath1113 , we have @xmath1114 where @xmath1115 and @xmath1116 for the upper bound in ( [ two ] ) while @xmath1117 is a small enough positive constant for the lower bound . the proof of theorem [ ttwo ] is preceded by a lemma .
[ lemdee]for all distinct @xmath79 and @xmath1 , we have@xmath1118 where @xmath1119 .
let us first show that , for all @xmath81 and some @xmath1120 , @xmath1121 by ( [ fidef1 ] ) , we have , for all @xmath51,@xmath1122 choosing @xmath1123 we obtain the right - hand side inequality in ( [ fif ] )
. by ( [ fidef1 ] ) , the left - hand side inequality in ( [ fif ] ) is equivalent to@xmath1124 that is , to@xmath1125 if @xmath1126 , then this is trivially satisfied provided @xmath1127 .
if @xmath1128 , then by ( [ fb ] ) we have@xmath1129 which proves the previous inequality and , hence , ( [ fif ] ) . putting in ( [ fif ] ) @xmath1130 and using @xmath1131 , which is true by lemma [ lemfe ] , we obtain@xmath1132 whence ( [ df ] ) follows . proof of theorem [ ttwo ] if @xmath232 , then @xmath1133 , and ( [ two ] ) follows from theorems [ tue ] and [ tnle ] .
assume in the sequel that @xmath1092 .
let us first prove the lower bound in ( [ two ] ) , that is,@xmath1134 by theorem [ tnle ] , we have @xmath1022 provided @xmath1135 for some @xmath72 .
set @xmath1136 where @xmath1137 will be chosen later . consider first the case @xmath1138 . by ( [ df ] ) ,
we have@xmath1139 applying ( [ tfi ] ) with @xmath1140 , we obtain@xmath1141 whence by ( [ rb])@xmath1142 provided @xmath1117 is small enough .
since @xmath1143 , we see that the condition ( [ detr ] ) is satisfied and , hence , ( [ twol ] ) follows from ( [ nle ] ) .
assume now that @xmath1144 .
by lemma [ lemn ] , there is an @xmath85-chain @xmath1145 connecting @xmath60 and @xmath86 and such that@xmath1146 by ( [ semi ] ) , we have@xmath1147\\[-8pt ] & \geq&\int_{b_{1}}\cdots\int_{b_{n-1}}p_{{t / n } } ( z_{0},z_{1})p_{{t / n}}(z_{1},z_{2})\cdots\nonumber\\ & & \hspace*{54.8pt}{}\times p_{{t / n}}(z_{n-1},z_{n})\,dz_{1}\cdots dz_{n-1},\nonumber\end{aligned}\ ] ] where @xmath1148 , @xmath1149 , @xmath1150 .
we will estimate @xmath1151 from below by means of ( [ nle ] ) .
for that , we need to verify the condition@xmath1152 by ( [ te ] ) , we have @xmath1153 it follows from ( [ tze ] ) and ( [ ne ] ) that@xmath1154 whence by ( [ rb ] ) @xmath1155 clearly , if @xmath1117 is small enough , then@xmath1156 since in ( [ chain ] ) @xmath1157 and @xmath1158 , it follows from ( [ ee ] ) that@xmath1159 hence , by ( [ nle ] ) and ( [ va]),@xmath1160 therefore , ( [ chain ] ) implies@xmath1161\\[-8pt ] & \geq&\frac{\exp ( -cn ) } { v ( x,\mathcal{r } ( t ) ) } \nonumber\\ & \geq&\frac{\exp ( -c({d_{\varepsilon } ( x , y ) } ) /{\varepsilon } ) } { v ( x,\mathcal{r } ( t ) ) } .\nonumber\end{aligned}\ ] ] by lemma [ lemdee ] , we have@xmath1162 substituting into ( [ 4 ] ) , we obtain ( [ twol ] ) .
to prove the upper bound in ( [ two ] ) , we basically repeat the proof of theorem [ tue ] with @xmath24 being replaced by @xmath82 for an appropriate @xmath85 . fix some @xmath81 and
denote by @xmath1163 the ball in the metric @xmath82 .
it follows from ( [ eps > ] ) that @xmath1164 which allows to modify lemma [ pf < exp ] as follows : for all @xmath1165 and @xmath623@xmath1166 provided the values of parameters @xmath557 and @xmath492 satisfy the conditions@xmath1167 indeed , ( [ rr ] ) is analogous to estimate ( [ n ] ) from the proof of lemma [ pf < exp ] for @xmath24-balls , which was proved using @xmath1168 . to repeat the proof for the metric @xmath82 , we need the precompactness of @xmath82-balls , that follows from ( [ de > ] ) , and the condition ( [ eff ] ) for @xmath82-balls of radii @xmath1169 , that follows from ( [ b = b ] ) , provided @xmath1170 . consequently , the statement of theorem [ tef ] is modified as follows : for all @xmath180 and @xmath714 , @xmath1171 where @xmath1172 indeed , arguing as in ( [ 3a ] ) and using ( [ rr ] ) we obtain that , under the assumptions of ( [ rl]),@xmath1173 setting here @xmath1174 yields@xmath1175 finally , minimizing the right - hand side of ( [ g ] ) in @xmath1176 , we obtain ( [ pg ] ) .
let us show that if @xmath1177 then@xmath1178 we have@xmath1179 whereas@xmath1180 it follows that@xmath1181 which proves ( [ p ] ) .
hence , we can rewrite ( [ pg ] ) in the form @xmath1182 provided the relation ( [ te2 ] ) between @xmath1183 is satisfied . as in the last part of the proof of theorem [ tue ] , we apply ( [ pg1 ] ) with @xmath1184 for fixed @xmath1020 .
note that in ( [ vrr ] ) @xmath1185 can be replaced by a larger value @xmath1186 .
the rest of the argument goes through unchanged , and we obtain @xmath1187 provided @xmath1188 by ( [ te ] ) , condition ( [ 18 ] ) can be satisfied by setting @xmath1189 .
[ ctwo]under the hypotheses of theorem [ ttwo ] , we have @xmath1190 where @xmath1115 and @xmath1117 is a large enough constant for the upper bound and a small enough positive constant for the lower bound
. the lower bound in ( [ twod ] ) follows from ( [ 4 ] ) , while the upper bound follows from ( [ ptupe ] ) and @xmath1191 provided @xmath1117 is chosen large enough to ensure @xmath1192 where @xmath14 is the constant from ( [ ptupe ] ) .
estimate ( [ twon ] ) follows then from ( [ nede ] ) .
[ exhk]a good example to illustrate theorem [ ttwo ] and corollary [ ctwo ] is the class of post critically finite ( p.c.f . )
fractals , where the heat kernel estimate ( [ twon ] ) was proved by hambly and kumagai @xcite . without going into the details of @xcite , let us mention that @xmath19 is the resistance metric on such a fractal @xmath16 , and @xmath25 is the hausdorff measure of @xmath16 of dimension @xmath1193 .
one has in this setting @xmath1194 , in particular , ( [ vd ] ) is satisfied .
hambly and kumagai proved that ( [ eff ] ) is satisfied with @xmath1195 where @xmath1196 ; cf .
@xcite , theorem 3.8 . condition ( [ fe ] ) follows from the estimate@xmath1197 proved in @xcite , lemma 3.3 , with @xmath1198 , as ( [ ne < ] ) implies that@xmath1199 the harnack inequality ( [ condh ] ) on p.c.f .
fractals was proved by kigami @xcite .
hence , corollary [ ctwo ] applies and gives on unbounded p.c.f .
fractals estimate ( [ twon ] ) . the same estimate was proved in @xcite , theorem 1.1 , for bounded p.c.f .
fractals using a different method . note that ( [ vd ] ) implies ( [ ne < ] ) with @xmath1200 [ where @xmath31 comes from ( [ va ] ) ] , provided all balls in @xmath16 are connected . indeed , ( [ vd ] ) implies by the classical ball covering argument that any ball of radius @xmath557 can be covered by at most @xmath1201 balls of radii @xmath1202 .
consequently , any point @xmath1203 can be connected to @xmath60 by a chain of @xmath85-balls containing at most @xmath1201 balls .
taking @xmath1204 we obtain ( [ ne < ] ) with @xmath1200 .
therefore , hypothesis ( [ fe ] ) is satisfied automatically for @xmath62 with @xmath1205 .
estimate ( [ twon ] ) means that the diffusion process goes from @xmath60 to @xmath86 in time @xmath100 in the following way .
the process first `` computes '' the value ) is satisfied with @xmath1206 , we obtain from ( [ te ] ) @xmath1207 whence@xmath1208 and@xmath1209\end{aligned}\ ] ] ] of @xmath85 as a function of @xmath17 , then `` detects '' a shortest chain of @xmath85-balls connecting @xmath60 and @xmath86 and finally goes along that chain ( see figure [ pic10 ] ) .
-balls for two distinct values of @xmath1210 provide different routes for the diffusion from @xmath60 to @xmath86 for two distinct values of time @xmath100 . ] this phenomenon was first observed by hambly and kumagai on p.c.f .
fractals , but it seems to be generic . hence , to obtain matching upper and lower bounds , one needs , in addition to the usual hypotheses , also the information encoded in the function @xmath84 , namely , the graph distance between @xmath60 and @xmath86 on any @xmath85-net approximation of @xmath16 .
the statement of theorem [ ttwo ] can be simplified if the metric space @xmath33 possesses an additional property as follows .
we say that a metric space @xmath33 satisfies the _ chain condition _ if there exists a constant @xmath1211 such that , for any positive integer @xmath420 and for all @xmath79 , there is a sequence @xmath1212 of points in @xmath16 such that @xmath1213 , @xmath1214 and@xmath1215 for example , any geodesic metric satisfies the chain condition .
[ lemde]if @xmath33 satisfies the chain condition , then @xmath101 for any @xmath81 .
indeed , fix @xmath81 and two distinct points @xmath79 , and choose @xmath420 so big that @xmath1216 .
let @xmath1217 be a sequence from the chain condition .
then it is also an @xmath85-chain , whence@xmath1218 which was to be proved .
[ cortwo]let the metric space @xmath33 satisfy the chain condition , and let all metric balls be precompact . if the hypotheses @xmath1219 are satisfied , then , for all @xmath741 and @xmath1 , @xmath1220 where @xmath40 is defined by ( [ fidef1 ] ) .
since by lemma [ lemde ] @xmath101 , condition ( [ fe ] ) is obviously satisfied . since @xmath1221 , we can replace in ( [ two ] ) @xmath82 by @xmath24 , which together with ( [ fifi ] ) yields ( [ twosided ] ) .
[ remnle ] @xmath1222obviously , estimate ( [ twosided ] ) ( should it be true ) implies ( [ uee ] ) .
we claim that ( [ twosided ] ) implies also ( [ nle ] ) ; moreover , the parameter @xmath1029 in ( [ nle ] ) can be chosen to be arbitrarily large , say @xmath1223 .
indeed , we need to show that if@xmath1224 where @xmath1029 is a ( large ) given constant , then@xmath1225 which amounts to@xmath1226 where we have renamed @xmath1227 to @xmath1029 .
indeed , by ( [ fidef1 ] ) we have@xmath1228 so that ( [ txi ] ) is equivalent to@xmath1229 if @xmath1230 , then by ( [ rb])@xmath1231 if the ratio @xmath1232 is large enough then , using @xmath1233 , we obtain that @xmath1234 whence ( [ t / xi ] ) follows .
if @xmath1232 is bounded by a constant , say @xmath1235 ( which includes also the case @xmath1236 ) , then by ( [ rb])@xmath1237 whence ( [ txi ] ) follows again .
we assume only basic hypotheses in this subsection .
moreover , we use neither the locality of @xmath1238 nor the existence of the process @xmath337 .
we state and prove some basic properties of the dirichlet problem in the abstract setting , that will be used in the proof of theorem [ tconv ] .
fix an open set @xmath104 such that @xmath1239 , and consider the following weak dirichlet problem in @xmath324 : given a function @xmath1240 , find a function @xmath1241 such that@xmath1242 where the second condition is a weak boundary condition and means that @xmath1243 .
[ lemdir1 ] for any @xmath1240 , problem ( [ d ] ) has a unique solution @xmath52 .
if @xmath52 solves ( [ d ] ) and @xmath1244 is another function such that @xmath1245 @xmath1246 then @xmath1247 .
moreover , the identity @xmath1248 holds if and only if @xmath1249 .
\(a ) the condition @xmath1250 implies that@xmath1251 for all @xmath1252 .
hence , @xmath327 is a hilbert space also with respect to the inner product @xmath1253 .
the harmonicity of @xmath52 in ( [ d ] ) means that @xmath1254 equivalently , this means for the function @xmath1255 that@xmath1256 since @xmath1257 , the functional @xmath1258 is a bounded linear functional in @xmath327 , and equation ( [ vfi ] ) has a unique solution @xmath1259 by the riesz representation theorem .
then @xmath1260 is a unique solution of ( [ d ] ) .
\(b ) setting @xmath1261 and noticing that @xmath1262 , we obtain using ( [ ufi])@xmath1263 hence , @xmath1247 , and the equality is attained when @xmath1264 , that is , when @xmath1265 . in what follows ,
denote by @xmath584 the resolvent operator of ( [ d ] ) , that is , @xmath1266 .
obviously , @xmath584 is a linear operator in @xmath131 . since by lemma [ lemdir1 ] @xmath1267
, we see that the norm of the operator @xmath584 in @xmath131 is bounded by @xmath184 .
[ lemdir2 ] if @xmath119 , then @xmath1268 .
in particular , if @xmath236 , then @xmath1269 . if @xmath1270 , then also @xmath1271 .
if @xmath1272 is an increasing sequence from @xmath131 and @xmath1273 as @xmath254 , then @xmath1274 a.e . in @xmath324 as @xmath254
\(a ) the function @xmath1275 is harmonic in @xmath347 and satisfies the boundary condition @xmath1276 @xmath1277 . by @xcite , lemma 4.4 , the latter condition implies @xmath1278 . substituting @xmath1279 into ( [ ufi ] )
, we obtain @xmath1280 . on the other hand , by @xcite ,
lemma 4.3 , @xmath1281 , whence it follows that @xmath1282 and , hence , @xmath1283 .
consequently , @xmath1276 and @xmath1284 .
\(b ) set @xmath1266 and @xmath1285 so that @xmath1286 and @xmath1287 .
setting @xmath1288 and @xmath1289 , we see that @xmath1290 , @xmath1291 and @xmath1292 . by @xcite ,
lemma 4.4 , we conclude that @xmath1293 . on the other hand , we have @xmath1294 whence @xmath1295 .
it follows that @xmath1296 . by lemma [ lemdir1 ]
we conclude that @xmath1297 . since the opposite inequality is true by the definition of a dirichlet form , we see that @xmath1298 .
it follows from lemma [ lemdir1 ] that @xmath1299 , which implies @xmath1300 .
\(c ) since @xmath584 is a bounded operator in @xmath131 , we see that @xmath1301 as @xmath1302 .
it follows that also @xmath1303 .
then there is a subsequence of @xmath1304 that converges to @xmath1305 almost everywhere in @xmath324 . finally , since the sequence @xmath1306 is monotone increasing , the entire sequence @xmath1306 also converges to @xmath1305 almost everywhere in @xmath324 .
the next lemma states useful consequences of the main hypotheses and motivates the statement of theorem [ tconv ] below .
it is also used in the proof of corollary [ corue ] .
[ lemconv]let all metric balls be precompact
. then the following implications are true : ( [ condh ] ) implies that the metric space @xmath33 is connected ; ( [ eff ] ) implies that the dirichlet form @xmath130 is conservative ; ( [ eff ] ) implies that @xmath1307 .
\(a ) assume that @xmath33 is disconnected , and let @xmath324 be a nonempty open subset of @xmath16 such that @xmath1308 is also nonempty and open .
there is a big enough ball @xmath1309 such that the intersections of @xmath1310 both with @xmath324 and @xmath1308 are nonempty , where @xmath849 is the parameter from ( [ condh ] ) .
since @xmath1311 is a compact set , there is a cutoff function @xmath52 of @xmath1311 in @xmath324 ; that is , @xmath1312 and @xmath1313 in a neighborhood of @xmath1311 .
obviously , @xmath1314 in @xmath1308 .
we claim that @xmath52 is harmonic in @xmath347 .
indeed , for every function @xmath1315 , we have @xmath1316 and@xmath1317 since @xmath1318 and , hence , @xmath1313 in a neighborhood of @xmath1319 , we obtain by the strong locality of @xmath1320 that @xmath1321 .
since @xmath1322 and @xmath1323 in @xmath1308 , it follows that @xmath1324 .
hence @xmath511 and @xmath52 is a nonnegative harmonic function in @xmath347 .
however , the function @xmath52 does not satisfy ( [ condh ] ) because @xmath52 takes in @xmath1310 the values @xmath184 and @xmath88 .
\(b ) by corollary [ cef ] , ( [ eff ] ) implies that@xmath1325 for any @xmath180 , @xmath623 , @xmath1 .
using this estimate and ( [ ptom ] ) , we obtain@xmath1326 as @xmath1023 , we see that @xmath1327 , which proves the stochastic completeness .
\(c ) if @xmath1328 then @xmath1329 so that the exit time from @xmath1330 is @xmath89 and @xmath579 fails . in the next statement , we use weaker versions of ( [ uee ] ) and ( [ nle ] ) that will be denoted by @xmath1331 and @xmath1332 .
namely , in each of these conditions we assume that the heat kernel exists as a measurable integral kernel of the heat semigroup @xmath162 and satisfies the estimates ( [ uee ] ) and ( [ nle ] ) for all @xmath1 and for _ almost _ all @xmath79 .
note that unlike the conditions ( [ uee ] ) and ( [ nle ] ) , their weak versions do not use the diffusion process @xmath1333 .
[ tconv]assume that all metric balls are precompact and@xmath1307
. then the following sets of conditions are equivalent : @xmath1334 ; @xmath1335 , and the heat kernel is hlder continuous outside a properly exceptional set ; @xmath1336 .
note that , by lemma [ lemconv ] , ( i ) implies that @xmath1307 .
however , neither of conditions ( ii ) or ( iii ) implies that @xmath16 is unbounded because ( ii ) is satisfied on any compact riemannian manifold .
proof of theorem [ tconv ] the implication @xmath1337 is contained in theorem [ tmain ] , and the implication @xmath1338 is trivial . in what follows
we prove the implication @xmath1339 . assuming ( iii ) ,
let us first show that @xmath16 is connected .
indeed , let @xmath16 split into a disjoint union of two nonempty open sets @xmath1340 and @xmath401 . by the continuity of the paths of @xmath1333
, we have @xmath1341 for all @xmath1 and @xmath1342 , @xmath1343 , whereas by ( [ nle ] ) we have @xmath1344 whenever @xmath1345 .
this contradictions proves the connectedness of @xmath16 . by @xcite ,
corollary 5.3 , ( [ vd ] ) , the connectedness , and the unboundedness of @xmath16 imply the _ reverse volume doubling _ ( [ rvd ] ) ; that is , the following inequality holds:[condrvd ] @xmath1346 which holds for all @xmath50 , @xmath569 , with some positive constants @xmath1347 . by @xcite ,
theorem 2.2 and section 6.4 ( see also @xcite ) , @xmath1348 imply @xmath579 . ) is essential for @xmath1349 ( see @xcite , theorem 2.2 ) .
in fact , it was shown in @xcite and @xcite that @xmath1350 imply also @xmath580 provided the dirichlet form is conservative . in our setting the conservativeness of the dirichlet form can also be proved but a direct proof of @xmath1351 is shorter . ]
let us now prove @xmath1351 , that is , @xmath1352 for all @xmath160 and @xmath623 , where @xmath207 is a properly exceptional set .
it suffices to prove that there is a constant @xmath1353 such that , for any ball @xmath1354,@xmath1355 indeed , the function@xmath1356 is quasi - continuous by @xcite , theorem 4.2.3.[remcheckthisinfot ] by @xcite , proposition 6.1 , if @xmath1357 for almost all @xmath344 , where @xmath841 is a constant and @xmath324 is an open set , then @xmath1358 for all @xmath1359 where @xmath207 is a properly exceptional set .
hence , ( [ pbb ] ) implies that@xmath1360 for some properly exceptional set @xmath1361 . taking the union of such sets @xmath1362 where @xmath347 varies over a countable family @xmath821 of all balls with rational radii and whose centers form a dense subset of @xmath16
, we obtain a properly exceptional set @xmath207 such that ( [ pfn ] ) holds for any ball @xmath1363 . approximating any ball @xmath347 from inside by balls of the family @xmath821
, we obtain ( [ pfn ] ) for all balls , which implies ( [ pf ] ) .
now let us prove ( [ pbb ] ) . by the comparison principle of @xcite , proposition 4.7
( see also @xcite , lemma 4.18 ) , we have , for any nonnegative function @xmath1364 , @xmath1365}{\mathop{\operatorname{esup}}}_{y\in b\setminus({1/2})b}p_{s}f ( y)\ ] ] for almost all @xmath1366 .
let @xmath1367 be a small positive constant to be specified below , and set @xmath1368 .
it follows from @xmath1332 and ( [ va ] ) that@xmath1369 provided @xmath1006 .
the initial value of @xmath85 is given by the condition @xmath1332 but we are going to further reduce this value of @xmath85 in the course of the proof .
assume that @xmath100 varies in the following interval:@xmath1370 the left - hand side inequality in ( [ tr ] ) implies by ( [ fb ] ) that @xmath1371 chose @xmath1367 from the identity@xmath1372 so that ( [ rt ] ) implies@xmath1373 integrating ( [ infp ] ) over @xmath1374 and using ( [ vd ] ) and ( [ zeta ] ) , we obtain@xmath1375\\[-8pt ] & \geq&c\zeta^{\alpha } \nonumber\\ & = & c^{\prime}\varepsilon^{\alpha/\beta}\nonumber\end{aligned}\ ] ] for almost all @xmath1376
. on the other hand , for almost all @xmath1377 , we have by @xmath1378 and lemma [ lemtifi ] @xmath1379 where we have used that @xmath1380 and @xmath1381 . using ( [ va ] ) and ( [ fb ] ) we obtain@xmath1382 finally , it follows from ( [ tr ] ) and @xmath1383 that @xmath1384 whence@xmath1385 for almost all @xmath1377 . combining ( [ eqs-2 ] ) , ( [ eqs-4 ] ) and ( [ eqs-3 ] ) , we obtain , for almost all @xmath1386,@xmath1387}{\mathop{\operatorname{esup}}}_{b\setminus k}p_{s}f \\ & \geq&c^{\prime}\varepsilon^{\alpha/\beta}-c\biggl ( \frac { 1}{\varepsilon } \biggr ) ^{\alpha/\beta}\exp\biggl ( -c\biggl ( \frac{1}{\varepsilon } \biggr ) ^{{1}/({\beta^{\prime}-1})}\biggr ) \\ & \geq&\frac{1}{2}c{^{\prime}}\varepsilon^{\alpha/\beta},\end{aligned}\ ] ] provided @xmath85 is chosen small enough .
the path @xmath1388 is a continuous path in @xmath1389 and , hence , can be integrated in @xmath100 .
it follows from the previous inequality that@xmath1390 which finishes the proof of @xmath1351 .
we are left to prove that @xmath1391 . by @xcite , theorem 3.1 ( see also @xcite and @xcite , theorem 5.3 ) , @xmath1392 imply the _ parabolic _ harnack inequality for bounded caloric function and , hence , the harnack inequality ( [ condh ] ) for bounded harmonic functions ( note that this result uses the precompactness of the balls ) .
we still have to obtain ( [ condh ] ) for all nonnegative harmonic functions .
note that by @xcite , theorem 2.1 , @xmath1393 in particular , for any ball @xmath347 , we have @xmath1394 . given a function @xmath500 that is nonnegative and harmonic in a ball @xmath1309 , set @xmath1395 for any @xmath262 , and denote by @xmath1396 the solution of the dirichlet problem@xmath1397 cf .
section [ secharm ] . since @xmath1398
, we have also @xmath1399 . since the sequence @xmath1400 increases and @xmath1401 ( cf .
@xcite , theorem 1.4.2 ) , it follows by lemma [ lemdir2 ] that @xmath1402 almost everywhere in @xmath347 .
each function @xmath1396 is bounded and , hence , satisfies the harnack inequality in @xmath347 , that is,@xmath1403 replacing in the right - hand side @xmath1396 by a larger function @xmath52 and passing to the limit in the left - hand side as @xmath254 , we obtain the same inequality for @xmath52 , which was to be proved .
[ corue]assume that all metric balls are precompact , @xmath1404 , and the dirichlet form @xmath1320 is conservative .
then the following sets of conditions are equivalent : @xmath1405 ; @xmath1406 . in the view of theorem [ tconv ]
, it suffices to prove that @xmath1407 . by lemma [ lemconv ] ,
( [ condh ] ) implies the connectedness of @xmath16 . by @xcite , corollary 5.3 ,
@xmath1408 provided @xmath16 is connected and unbounded , which is the case now .
by @xcite , theorem 2.2 , the conservativeness and @xmath1409 imply ( [ eff ] ) .
many equivalent conditions for @xmath1410 were proved in @xcite under the standing assumptions @xmath1411 and the conservativeness of @xmath1412 .
of course , each of these conditions can replace @xmath1378 in the statement of corollary [ corue ] .
[ corunbound]assume that all metric balls are precompact , @xmath1307 , and @xmath33 satisfies the chain condition .
then the following two sets of conditions are equivalent : @xmath1030 ; the heat kernel exists and satisfies the two - sided estimate ( [ twosided ] ) . the implication @xmath1413 is contained in corollary [ cortwo ] .
let us prove the implication @xmath1414 .
estimate ( [ twosided ] ) implies ( [ uee ] ) as well as ( [ nle ] ) with any value of @xmath1029 , in particular , @xmath1223 ; cf .
remark [ remnle ] .
by @xcite , lemma 4.1 , ( [ nle ] ) with @xmath1223 implies ( [ vd ] ) .
finally , by theorem [ tconv ] , we obtain @xmath1415 .
we briefly list the lettered conditions used in this paper with references to the appropriate places in the main body .
@xmath1416 ( section [ secharnack ] ) ; @xmath1417 ( section [ secharnack ] ) ; @xmath1418 ( section [ secef ] ) ; @xmath1419 ( section [ secef ] ) ; @xmath1420 ( section [ secdue ] ) ; @xmath1421 provided @xmath1422 ( section [ seclow ] ) ; @xmath1423 ( section [ secrvd ] ) .
the authors thank martin barlow , jiaxin hu , jun kigami , takashi kumagai and alexander teplyaev for valuable conversations on the topics of this paper .
the authors are grateful to the unnamed referees for careful reading of the manuscript and for the useful remarks . | we prove equivalent conditions for two - sided sub - gaussian estimates of heat kernels on metric measure spaces . and . |
the nearby virgo cluster is a largely extended and complex structure of over 1300 galaxies and represents the major feature of the local supercluster .
based on the pioneering work by de vaucouleurs ( 1961 ) , the photographic las campanas survey by binggeli et al .
( 1985 , hereafter the virgo cluster catalog or vcc ) and a complementary collection of galaxy redshifts ( binggeli et al . 1993 ) , a number of subclumps and gravitationally unbound clouds were identified in virgo from imaging and recessional velocity data .
centrally located are two separate subclumps each dominated by a giant elliptical galaxy , i.e. m87 and m86 ( see binggeli 1999 ) . on a large scale
there is another structure along the north - south axis of the virgo cluster , defined by the northern m87/m86 subclumps ( called `` cluster a '' in binggeli et al .
1987 ) and the southern galaxy concentration around m49 , called here the `` m49 subclump '' (= `` cluster b '' ) .
these subclusters are bound to the south and west by three galaxy clouds named w , w ( de vaucouleurs 1961 ) , and m ( ftaclas et al .
1984 ) .
the knowledge of a precise distance to the virgo cluster or more generally a good understanding of its three dimensional structure plays an important role in many research areas of extragalactic astronomy .
however , despite the large effort to resolve these issues over the past decades , the spatial extension of the virgo cluster remained highly uncertain .
the reasons are twofold .
firstly , only a relative small number of cluster member galaxies had accurate distance measurements .
secondly , the typical target objects were spiral galaxies which reside , according to the morphology density relation ( dressler 1980 ) , in the outskirts of galaxy clusters .
so it should come as no surprise that using accurate distance indicators like cepheids and supernovae of type ia revealed a significant distance spread among the virgo spirals ranging from 15mpc ( eg .
graham et al .
1999 ; saha et al .
2001 ) to 25mpc ( eg .
saha et al .
these heavily disputed results simply reflect the line - of - sight depth of the large spiral halo of the virgo cluster and demonstated why accurate distance measurements of only a few spirals are of limited use .
progress towards a precise mapping of the 3d - structure of the virgo cluster region was made only when the numbers of spiral galaxy distances were increased ( e.g. yasuda et al .
1997 ) . alternatively , fundamental plane distance measurements ( gavazzi et al .
1999 ) and surface brightness fluctuations ( neilsen & tsvetanov 2000 ; tonry et al .
2001 ) were employed to locate the more centrally concentrated early - type giant elliptical galaxies .
similar to the giant brethren , the less luminous and more elusive dwarf elliptical ( de ) galaxies are well confined to the highest galaxy densities in a cluster ( the morphology density relation for dwarfs , binggeli et al .
but it is their occurrence in large numbers in cluster cores ( binggeli et al .
1985 ; ferguson & binggeli 1994 ; jerjen & dressler 1997 ) that makes them unique and even more valuable than giant ellipticals .
dwarf ellipticals are the only galaxy type that flag the gravitational center(s ) in a galaxy cluster _ and _ are available in statistically sufficient numbers . despite the great potential des offer to examine the densest regions of galaxy clusters only little
work has been done with des to date due to the lack of an accurate and practical distance indicator .
correlations between global parameters of des such as the effective surface brightness luminosity relation ( binggeli & cameron 1991 ) or the shape parameter luminosity relation ( jerjen & binggeli 1997 ; binggeli & jerjen 1998 ) have considerable scatter and thus are not reliable to measure individual distances . instead , the resolution power of the hst has to be used to resolve the stellar populations of des to establish distances by means of the tip magnitude of the red giant branch ( trgb , e.g. karachentsev et al .
while this approach works well for nearby des it becomes exceedingly difficult ( crowding effects ) and expensive ( long integration times ) to obtain good s / n stellar photometry at the required limiting magnitude with larger distances .
this explains why the trgb method was applied to only one de ( vcc1104 or ic3388 ) in the virgo cluster ( harris et al .
1998 ) to date .
considering the limitation of the trgb method , another distance indicator has emerged as a substitute for measuring accurate distances to des beyond 10mpc from the local group .
this is the surface brightness fluctuation method based on the discrete sampling of an _ unresolved _ stellar population in a galaxy with a ccd detector and the resulting poisson fluctuations in the number of stars within a resolution element ( tonry & schneider 1988 ) . the method has been extensively tested in the sculptor group ( jerjen et al .
1998 ) , centaurus a group ( jerjen et al .
2000 ) , m81 group , the canes venatici cloud and the near field ( jerjen et al .
first results from dwarf ellipticals in the more distant fornax cluster ( jerjen 2003 ) and centaurus cluster ( mieske et al . 2003 ) were reported recently .
the high accuracy of the method ( @xmath19% ) opens up the possibility to measure precise distances to cluster des in a simple and efficient way .
the goal of this paper is threefold .
first , we will measure the surface brightness fluctuation distances of a sample of 16 dwarf elliptical galaxies in the direction of the virgo cluster .
we introduce the dwarf galaxies with their basic properties in 2 . in
3 we describe the observations and data reduction . in
4 we will carry out the sbf analysis and calculate the fluctuation signals .
we determine the sbf distances of the sample galaxies in 5 .
second , we will investigate the distance distribution of the dwarfs in the context of the 3d - structure of the cluster in 6 .
lastly , we will estimate rough metallicities and ages for our early - type galaxies in 7 using the fluctuation magnitudes and b - r broadband colour .
our conclusions are given in 8 .
to follow - up the question on the 3d - distribution of the de population in the virgo cluster and to search for the gravitational centre(s ) of the cluster , we have studied a sample of 16 early - type dwarfs from the virgo cluster catalog ( vcc ; binggeli et al .
1985 ) . as it is a main requirement for the successful application of the sbf method , galaxies were primarily selected on their morphological appearance , i.e. type `` de(,n ) '' or `` ds0(,n ) '' and on their apparent size , i.e. an isophotal radius @xmath20 and small ellipticity . within these constraints , dwarfs were chosen in a way to get a good coverage in velocity space ( @xmath21kms@xmath22@xmath231850kms@xmath22 ) and in the celestial distribution ( @xmath24 ; @xmath25 ) .
a few dwarfs like vcc0810 simply came into our sample because they were close neighbours of our target galaxy .
finally , we added the nucleated de galaxy vcc1104 ( ic3388 ) to the sample .
vcc1104 is the only virgo de for which an independent distance measurement is available ( harris et al .
1998 ) .
[ fig1 ] shows the galaxy distribution in the central region of the virgo cluster with our sample galaxies highlighted as filled circles . in projection , all except two dwarfs are located in clustera ( m87/m86 subclump ) .
vcc0929 ( ngc4415 ) and vcc0856 ( ic3328 ) are both clusterb objects . in table
[ tbl-1 ] we list the fundamental parameters of our dwarfs . the vcc numbers ( col .
[ 1 ] ) , other names ( col . [ 2 ] ) , and the morphological types ( col . [ 3 ] ) were taken from binggeli et al .
the j2000 coordinates are give in cols.[4 - 5 ] and the cluster region where the galaxy is found in col .
[ 6 ] . the total @xmath0 magnitudes ( col .
[ 7 ] ) are taken from barazza et al .
( 2003 ) and the heliocentric velocities ( col . [ 8 ] ) are from binggeli et al .
( 1985 , 1993 ) and conselice et al .
( 2001 ) .
a full account of the photometric properties of the galaxies based on new ccd data , including a comprehensive discussion of the characteristics of the surface brightness profiles , colour gradients , and structure parameters , is given elsewhere ( barazza et al .
we note that the des have total absolute @xmath0 luminosities in the range @xmath26 and thus are at the bright end of the type - specific luminosity function for des ( sandage et al .
1985 ; jerjen & tammann 1997 ) .
as such these systems are typical cluster des with no counterparts known in low density environments i.e. galaxy groups or in the field as satellite galaxies .
only four of our des are fainter than the m31 companion ngc 205 ( @xmath27 ) , a de that is among the brightest early - type dwarfs found outside of clusters .
deep ccd images of the virgo des have been acquired in the @xmath0 and @xmath1 passbands in service mode using the two first units of the very large telescope at eso paranal observatory over a period of two semesters : at ut1+fors1 during an observing run on july 1014 , 1999 and at ut2+fors2 during dark time periods in march may , 2000 .
the detectors of the fors ( focal reducer / low dispersion spectrograph ) instruments are thinned and anti - reflection coated tektronix ( fors1 ) and site ( fors2 ) ccds with @xmath28 pixels . by default ,
service observations were taken in standard resolution mode , with a high gain and a pixel scale of @xmath29pixel@xmath22 that yields a field of view of @xmath30 .
the ccds were read out in the four - port mode , i.e. four amplifiers read out one quarter of the ccd each .
target galaxies were positioned close to the center of one of the four quadrants leaving the other three `` empty '' and thus useful as night flats .
three exposures of 400600sec durations with slightly different pointings were secured in each filter for each galaxy .
the bessell @xmath1 broadband filter was used at ut1+fors1 while bessell @xmath0 and @xmath31 filters were employed at ut2+fors2 .
table [ tbl-2 ] summarises the log of the observations .
template bias frames were subtracted from all ccd images .
twilight sky exposures were median - averaged after level normalisation to obtain twilight masterflats .
flatfielding with the masterflats did produce no satisfactory results in all cases and reliable nightflats , which might have worked better , were unfortunately impossible to construct from the science frames . to remove the background gradients noticeable in some of the images , above all in the @xmath1 ( or @xmath32 ) frames , we fitted instead third - order polynomials to the sky background ( see barazza et al .
2003 for more details ) .
the systematic uncertainty in the residual images due to this sky flattening problem was measured less than 0.1% over the full ccd area in all passbands , with the random errors being much smaller .
all ccd data were acquired on nights with photometric sky quality . in service mode ,
a series of landolt ( 1992 ) photometric standard stars are observed in various passbands under different airmasses every night .
fields which contained saturated standard stars when using short exposure times ( @xmath33sec ) were reobserved in the low gain mode .
the analysis of fields taken in both gain modes confirmed the inverse gain ratios for the four ccd quadrants given in the fors1 + 2 user manual issue 2.1 of eso .
finally , we scaled the low gain frames to match the high gain images . in that way , photometric data over a larger magnitude range became available . from each series of standard stars observations we computed photometric zero points , extinction coefficients and color terms using the iraf package .
for the observing period 65 the extinction coefficients and color terms could be compared with the values reported by the vlt quality control and trending services ( hanuschik & silva 2002 ) finding good agreement : @xmath34 the photometric zero points varied significantly over the observing period 65 .
therefore , we used the photometric zero points we determined for each observing night from our own set of standard stars .
formal photometric accuracies in the different passbands are 0.03mag ( bessell @xmath0 ) , 0.02mag ( bessell @xmath1 ) , and 0.02mag ( @xmath35 ) .
qc information for the observing period 63 was not available .
all images of a galaxy were registered by matching the positions of @xmath36 reference stars on each ccd frame using _ starfind _ , _ xyxymatch _ and _ imalign_. the sky - subtracted images taken in the same passband were cleaned from cosmic rays with _
crreject _ and co - added with _
imcombine_. finally , the resulting master images were flux calibrated .
the @xmath1-band galaxy images were prepared for the sbf analysis following the recipes of jerjen et al .
( 2000 ; 2001 ) . this includes * using daophot ii ( stetson 1987 ) to clean the images from foreground stars , globular clusters , and background galaxies brighter than magnitude @xmath37mag ( @xmath38500 adus in @xmath1 and 600 adus in @xmath32 ) . * modeling the 2-dimensional galaxy surface brightness distribution . *
subtracting the best - fitting galaxy model from the master image .
* noise normalization of the residual image .
between three and seven square subimages ( hereafter sbf fields ) were then defined on a fluctuation image within the 25.5 @xmath1magarcsec@xmath39 isophotal limit thereby avoiding areas with previously identified disturbing sources or non - radial irregular features such as dust or spiral features ( jerjen et al .
2000 ; jerjen et al .
2001 ; barazza et al .
the size of the sbf fields was chosen between @xmath40 and @xmath41 pixels depending on the apparent size of the galaxy .
assuming an average seeing of @xmath42arcsec ( see table[tbl-2 ] ) , a field carries the sbf signal from @xmath43 to @xmath44 independent points . in total
we defined 80 sbf fields in our 16 sample galaxies .
the sbf fields were fourier transformed and the azimuthally averaged power spectra calculated . from isolated bright stars on the master image we determined the point spread function ( psf ) profile .
we then fitted a linear combination of the flux normalized and exposure time weighted psf power spectrum and a constant at the observed galaxy power spectrum @xmath45 , demanding a least - squares minimisation .
data points at low spatial frequencies ( @xmath46 ) were omitted as they are likely affected by imperfect galaxy model subtraction .
[ psgallery1 ] and [ psgallery2 ] show the power spectrum of each sbf field with the best fitting analytic function indicated as solid line
. tables [ tbl-3 ] and [ tbl-4 ] summarize the quantities measured in the sbf analysis : col . 1 galaxy name and sbf field number , col
. 2 pixel size of the sbf field , col .
3 magnitude @xmath47 of a star yielding 1 adu per second on the ccd , col .
4 mean galaxy surface brightness within the sbf field in adu , col .
5 sky brightness in adu , col .
6 exposure time normalized amplitude @xmath48 of the best least squares fit at wave number @xmath49 with fitting error in brackets , col
. 7 the scale - free white noise component @xmath50 in the power spectrum , indicating the ratio of sky to mean galaxy surface brightness within the sbf field .
the contribution from distant background galaxies that are fainter than the cutoff magnitude @xmath51@xmath1mag to the fluctuation signal @xmath48 was estimated with the modified jensen et al .
( 1998 ) equation as discussed in jerjen et al .
( 2001 ) : @xmath52 where @xmath53 is the sbf field pixel size in arcsec .
the measured signal - to - noise s / n@xmath54 ( col .
8) in a sbf field is between 2.3 ( f1 for vcc0856 ) and 13.0 ( f3 from vcc1087 ) . the relative contribution of @xmath55 to the signal @xmath48 ( col .
9 ) is of the order of three percent .
a rich system of globular clusters ( gcs ) in the halo of a target galaxy is another source of unwanted surface brightness fluctuations .
but as we pointed out elsewhere ( jerjen 2003 ) the expected number of gcs in bright cluster dwarf ellipticals is quite low and only a minor issue . the gc frequency ( @xmath56)luminosity relation for de , ns studied in the fornax and virgo clusters ( miller et al .
1998 ) predicts 1040 gcs for galaxies in the luminosity range covered by our dwarfs .
all gcs would be brighter than our cutoff luminosity and excised during the image cleaning process .
therefore , we applied no further correction to the measured sbf power .
in the last section we fourier - analysed the selected sbf fields and measured the fluctuation signals .
these signals were converted into a stellar fluctuation magnitude @xmath2 using the formula @xmath57 .
furthermore , the @xmath58 colour for each field was measured from the cleaned @xmath0 and @xmath1 master images .
both quantities were corrected for foreground extinction using the iras / dirbe maps of dust ir emission ( schlegel et al .
the results are listed in the tables [ tbl-5 ] and [ tbl-6 ] .
the power spectrum fitting error is between 2 and 11% .
other sources of minor errors are the psf normalization ( @xmath592% ) , the shape variation of the stellar psf over the ccd area ( 12% ) and the uncertainty in the photometric calibration ( @xmath60mag in @xmath0 , 0.02mag in @xmath1 and @xmath32 ) . if we further adopt a 16% error for the foreground extinction ( schlegel et al .
1998 ) , the formal combined error for a single @xmath61 measurement is between 0.05 and 0.14mag ( col . 3 ) .
the error associated with the local colour ( col .
4 ) has been obtained through the usual error propagation formula from the uncertainties in the sky level determination , the photometry zero points , and galactic extinction .
the distance modulus of a program dwarf was then determined by least - squares fitting the set of @xmath62 $ ] data to the calibration equations @xmath63 ^ 2 - 1.39\\ & & \quad \mbox{for}\,\,0.80<(b - r)_0<1.35 . \nonumber\end{aligned}\ ] ] these linear and parabolic equations were described and discussed at length in jerjen et al .
( 2001 ) and jerjen ( 2003 ) .
they were deduced from the theoretical @xmath64 relation for mainly old , metal - poor stellar populations as predicted by worthey s ( 1994 ) stellar population synthesis models using the evolutionary isochrones from the padova library ( bertelli et al .
1994 ) and an empirical zero point ( jerjen et al .
fig.[mbarcalib ] illustrates the calibration diagram showing model @xmath58 and @xmath65 values as a function of age , metallicity , and stellar mixture of the underlying stellar population .
the solid lines mark the two best - fitting analytical expressions ( eqs.1 and 2 ) . to obtain the sbf distance of a galaxy with integrated colour redder than @xmath66
( e.g. vcc0929 , vcc1010 , and vcc1355 ) is straightforward as the colour fluctuation magnitude relation is unambiguously represented by eq . 1 .
however , the task is more difficult for galaxies in the colour regime where both calibration equations are valid ( @xmath67 ) .
there , the decision which of the equations to use has to be made based on the observed trend between colour and fluctuation magnitude for the analysed sbf fields . in some cases there is also independent distance information available that renders one of the two solutions more likely than the other .
the requirement , or desideratum to cover a colour range as large as possible has to be borne in mind during the sbf field selection , but sometimes it is impossible to achieve due to the lack of a colour gradient in the dwarf ( barazza et al .
2003 ) . for galaxies like vcc0815 , vcc0846 , vcc0856 , vcc0928 , vcc0940 , and
vcc1422 we found sufficient colour range in the sbf fields .
the results are illustrated in fig .
[ cmdiagrams ] where we show that the colour fluctuation magnitude diagrams of these dwarfs , with the exception of vcc0928 , are best fitted with the linear calibration equation . judged from the colour
fluctuation magnitude relation over the observed colour range , vcc0928 might fit into either the linear or parabolic branch , but the latter is strongly preferred by the observed velocity of @xmath68 km @xmath69 ( cf .
table 1 ) ; the linear branch solution would put that galaxy at too large a distance to be compatible with a negative velocity ( cf .
also 6 ) . for vcc1104
the colour range covered by the fields is very small , leaving the calibration ambiguous .
however , there is a strong constraint put by the trgb distance of @xmath70mag ( @xmath9mpc ) from the hst observations of harris et al .
( 1998 ) . as the tip magnitude of the red giant branch has been shown to be an accurate distance indicator for old and metal - poor stellar populations ( e.g. da costa & armandroff 1990 ; gratton et al .
1997 ; salaris & cassisi 1998 ) , this result is useful to identify the correct sbf distance .
the correlation of @xmath2 with the @xmath58 colour over the range of galactocentric radii probed by the four subframes in vcc1104 is plotted in fig .
[ cmdiagrams ] .
we found a good agreement with the sbf distance modulus @xmath71mag ( @xmath7mpc ) based on the linear branch while the distance modulus of @xmath72mag from the parabolic branch appears too small and less compatible with the trgb result given the quoted uncertainties .
therefore , we will adopt a true distance modulus of @xmath71mag for vcc1104 in the following .
the most interesting situation is found for the dwarfs vcc0009 and vcc0490 . as shown by their colour fluctuation magnitude diagram in fig .
[ cmdiagrams ] , the @xmath73 data points of each galaxy form two separate groups and the least - squares fitting moves them right onto the two calibration branches .
this is the first time that galaxies have been found with data on both branches .
although unusual , this finding is well explained with an age difference between the underlying local stellar populations ( see 7 ) .
it provides first empirical support for the relative offset of the two theoretical branches .
we further note that vcc0490 has two sbf fields ( f5 and f6 ) with similar @xmath58 colours but with fluctuation magnitudes that differ by more than 0.4mag .
this apparent discrepancy can only be understood by overlaying the data on the calibration diagram .
the case of vcc0490 nicely demonstrates the importance to examine many sbf fields in a galaxy to estimate its distance correctly .
there are four galaxies left , viz .
vcc0810 , vcc1036 , vcc1087 , and vcc1261 , where the colour range is too small and where no independent distance information as in the case of vcc1104 is available to break the colour branch ambiguity .
their data can be technically fitted either by the linear or the parabolic equation and thus two distances can be inferred for them .
we have to leave it to future investigations to find further support for one of the two distances because of lack of independent evidence that could help to resolve the ambiguity .
we show in fig .
[ cmdiagramstwo ] the best fits of the dwarfs colour fluctuation magnitude diagrams to both calibration equations and quote the two distances in table[distmods ] .
we list in table [ distmods ] the derived sbf distances for all program galaxies .
the mean 1@xmath5 uncertainty in @xmath74 , neglecting the calibration ambiguity in those four cases , is 0.17 mag , corresponding to 1.4 mpc or 8% of the distance .
the dominant source of this uncertainty is the error in the @xmath58 colour through its strong amplification by the steep colour - fluctuation magnitude relation .
this can be compensated only by having a large number of fields for the sbf analysis ; the more fields the smaller the resulting error .
the relatively large galaxy sample with sbf distances available allows us to check the quality of our distance calibration . for that purpose we followed neilsen & tsvetanov ( 2000 ) and tested our data for a possible correlation between the colour and distance modulus . fig .
[ colordistmod ] shows the distance modulus for each galaxy versus the mean color of that galaxy . in the four cases where we have two different distance
moduli the data points are interconnected with a vertical line .
at a first glance , there seems to be a slight trend in the data for redder galaxies to have shorter distances .
but looking more closely , this impression comes mainly from the relatively short distance of the reddest dwarf vcc0929 .
do we have reasons to think that vcc0929 is indeed at the near side of the virgo cluster ?
vcc0929 is a clusterb object and in projection close to m49 ( see fig.[fig1 ] ) . as m49
has a reported distance of @xmath75mag ( neilsen & tsvetanov 2000 ; tonry et al .
2001 ) coinciding with the mean distance of the m87 subcluster , i.e. the more nearby of the two major clumps of clustera ( see following section ) , the short sbf distance of @xmath76mag for vcc0929 is indeed fully consistent with its location in the cluster . aside from vcc0929
, there is certainly no significant colour trend discernible in fig .
[ colordistmod ] .
the individual sbf distances of our des range from 14.9 to 21.3 mpc ( table [ distmods ] , not counting the ambiguous cases ) . clearly , all of these galaxies are members of the virgo cluster , as was expected from their redshift and morphological appearance . the mean distance modulus , giving each solution of the four ambiguous cases half weight , is @xmath77 = 31.23 @xmath78 0.06 , or @xmath79 = 17.6 @xmath78 1 mpc .
this value is in good accord with the mean virgo cluster distance derived for giant early - type galaxies from the sbf ( tonry et al .
2001 : @xmath74 = 31.15 , or 17.0 mpc ) and other methods ( neilson & tsvetanov 2000 , kelson et al .
2000 ) , as well as for virgo spiral galaxies ( e.g. graham et al . 1999 ; but see saha et al . 1997 and
ekholm et al .
1999 for dissident views ) .
however , as the distance distribution of des is broad and highly non - symmetric ( see below ) , a discussion of the cluster mean distance is not very meaningful anymore ; the derived mean value is obviously highly sensitive to the inclusion or exclusion of single objects of the sample .
a coarse measure of the distance spread is provided by the standard deviation of de distance moduli : @xmath80 = 0.245 mag .
this has of course to be compared with the mean distance error for a single dwarf which , as quoted above , is @xmath81 = 0.169 .
the true 1@xmath5 distance dispersion of the dwarf ellipticals in the cluster is then @xmath82 or 1.45 mpc .
this corresponds to a total 1@xmath5 or 2@xmath5 cluster _ depth _ of the virgo cluster of ca .
3 or 6 mpc , respectively , in accord with the large distance range of 14.9 to 21.3 mpc found for the dwarfs ( cf .
table [ distmods ] ) . how does this compare with the tangential cluster dimension ? according to binggeli ( 1999 ) ,
the angular width of the virgo cluster is @xmath83 ( as can also be seen from fig . 1 ) .
assuming spherical symmetry for the cluster and a mean distance of 17mpc , we would then expect a ( @xmath38 2@xmath5 ) front - to - back depth of only 2.4mpc ( @xmath84mag ) .
hence this is evidence that the distribution of virgo cluster des is significantly elongated along the line of sight .
there have been previous reports of a _ very _ extended ( @xmath85 mpc ) , cigar - like distribution of virgo des ( young & currie 1995 ) .
while such a huge depth in the de population as claimed by young & currie has been suspected to be in large part a finger - of - god effect due to unaccounted - for observational errors ( binggeli & jerjen 1998 ) , it is definitively ruled out by our data . on the other hand ,
the slightly prolate distribution of des found here is in excellent agreement with neilsen & tsvetanov ( 2000 ) and west & blakeslee ( 2000 ) who used the sbf method to determine the distances to giant ellipticals in virgo .
these authors found a similar cluster depth of @xmath86mpc around m87 .
moreover , arnaboldi et al .
( 2000 ) inferred from the luminosity function of intracluster planetary nebulae that the front end of the virgo cluster is 14%18% ( or @xmath87=0.28 $ ] to 0.36 ) closer to us than the virgo core region around m87 .
the filamentary distribution of virgo _ spirals _ has of course been known for a long time .
yasuda et al .
( 1997 ) have argued that the distribution of spiral galaxies in the virgo cluster is best described as an elongated structure along the line - of - sight with a depth of @xmath88mpc from m87 .
however , the evidence that such an elongation , if only a milder one , holds for the early type galaxies in the core region as well is unexpected and new , and it is nicely confirmed by our sbf distances of dwarf ellipticals .
but there is more to the distance distribution than just the dispersion .
consider fig .
[ distmodhisto ] , upper panel , where the distance distribution of our 16 dwarf ellipticals is shown ( black histogram , dwarfs with two possible distances being weighted 0.5 at each value ) . the binned distribution is clearly _ bimodal _ with a broad major galaxy concentration at @xmath89mag ( @xmath11mpc ) , a sharp drop at @xmath90mag ( @xmath91mpc ) and a narrow secondary peak at @xmath12mag ( @xmath13mpc ) .
these strutural features are further emphasised when we add the sbf results for 24 giant ellipticals and ngc4486b from neilsen & tsevtanov ( 2000 ) and tonry et al .
( 2001 ) that are found in our surveyed area ( dashed histogram ) .
the 11 ellipticals studied in both references are weighted half at each distance . to test the bimodality in the distance distribution against the conservative assumption that the entire galaxy sample was drawn from a single gaussian distribution we used the adaptive kernel method ( vio et al .
1994 ) employing a gaussian kernel .
thereby , distances for giant elliptical galaxies with two independent measurements ( neilsen & tsevtanov 2000 ; tonry et al .
2001 ) were averaged while distances for dwarf ellipticals with two possible distances were weighted half at each distance .
the lower panel in fig .
[ distmodhisto ] shows the adaptive kernel distribution of the 41 early - type virgo galaxies as a solid line .
this line was compared with the result from 1000 simulations of 41 randomly distributed objects having the same statistical properties as our sample galaxies , i.e. a mean distance modulus of @xmath92 and a standard deviation of @xmath930.4 . the mean adaptive kernel distribution and the @xmath94 confidence level lines for these simulations are shown as dashed and dotted lines .
the two galaxy concentrations at @xmath95mag and @xmath12mag were found to be significant at the 99% ( 2.5@xmath5 ) and 89% ( @xmath14 ) level , respectively .
it should also be noted that any asymmetric feature in the true distance distribution tends to be washed out by the errors in the individual distances , i.e. the two peaks will in reality be much narrrower .
what could this bimodality mean ?
it has previously been noted ( binggeli et al .
1993 , binggeli 1999 ) that the virgo cluster core ( cluster a ) shows a pronounced double structure in the projected spatial distribution and in the kinematics of the galaxies .
the main , most massive and galaxy - rich clump of cluster a , and of the whole cluster is clearly centered on the supergiant e galaxy m87 ( ngc4486 ) .
a secondary , less massive and less rich clump , lying about 1@xmath96 nw to m87 , is centered on the giant galaxy m86 ( ngc4406 ) .
members of this secondary clump are found to have systematically smaller radial velocities ; in particular , all negative velocities tend to be clustered in the m86 region ( binggeli et al .
1993 , binggeli 1999 ) .
this has been interpreted in terms of an infall , or subcluster merging scenario .
given the observed kinematic pattern , the m86 subcluster seems to be falling into the main m87 subcluster _ from the backside_. in consequence , we might expect to see a bimodality in the distance distribution of virgo cluster galaxies : a somewhat larger , broader population peaking at a slightly lower - than - average distance , and a smaller , possibly narrower population peaking at a slightly higher - than - average distance , with a separation of 1 to a few mpc .
not only is this exactly what we see in fig .
[ distmodhisto ] , very intriguingly the distance modulus of the primary , higher and closer peak , @xmath97 = 31.00 , roughly coincides with the mean sbf distance modulus for m87 measured by tonry et al .
( 2001 ) and neilsen & tsevtanov ( 2000 ) : @xmath98mag .
likewise , the position of the secondary peak , @xmath99 = 31.33 , is perfectly matched by these authors measurements for the sbf distance of m86 : @xmath100mag .
the bimodal distance distribution of e and de galaxies in the core of the virgo cluster is therefore naturally explained by the subcluster merging scenario proposed before ( binggeli et al.1993 , binggeli 1999 ) .
the m87 and m86 subclusters seem to be separated by @xmath101 0.3 , or @xmath102 2.5 mpc . with individual distances and radial velocities available for a fairly large sample of es and des
, we can now also check whether the galaxies show the expected infall pattern . for this purpose
we have plotted in fig .
[ veldistplot ] distance versus heliocentric velocity for all galaxies condisered . as the radial velocity dispersion of early - type galaxies in the virgo cluster is @xmath103kms@xmath22 ( binggeli et al .
1993 ) , we should not expect to see any streaming motions in the velocity range @xmath104kms@xmath22 .
however , there is a clear infall signature traced out by the galaxies with negative velocities . as mentioned before , most of these are clustered around m86 ( ngc4406 ) , which has a negative velocity itself .
they are on average lying on somewhat larger distances than the cluster mean .
ngc4419 is an exception , lying in the north of the cluster and being classified as sa in the vcc .
this galaxy likely belongs to a population of surrounding late - type galaxies that have recently fallen into the cluster ( see the discussion in vcc ) .
whether vcc0810 , for which two possible distances have been derived ( cf . table [ distmods ] ) , is approaching the cluster from behind for the first time or has already fallen through it while still maintaining such a large relative ( negative ) velocity , is difficult to judge without detailed modeling .
the long distance of vcc0810 would in any case appear more likely now , as it is perfectly in accord , like the distances and velocities of m86 , vcc00928 , vcc0815 , and vcc0846 , with the virgocentric infall model of kraan - korteweg ( 1986 ) .
the metallicities of de galaxies are basically unknown .
rakos et al .
( 2003 ) estimated a [ fe / h ] range from @xmath105 to solar with a mean value around @xmath106 .
the primary problem in obtaining accurate information about the stellar populations in des is because their surface brightnesses are very low , making spectral observations difficult regardless of distance . in the light of those technical obstacles ,
the much easier to measure field - to - field variation of the sbf magnitude and @xmath3 colours across a galaxy may serve as a valuable diagnostic tool to get rough metallicities and ages of the unresolved stellar populations . given , however , that the age - metallicity degeneracy steadily increases for populations younger than 8 gyrs ( i.e. @xmath107 $ ] data points of models move down along the linear branch in fig .
[ metalmagplot ] ) , it is essential to base our estimates on age constraints for early - type galaxies as provided by studies using more accurate methods .
stellar absorption - line indices were measured by caldwell et al .
( 2003 ) to disentangle the effects of age and metallicity in a sample of virgo early - type galaxies .
the strong balmer lines found in these dwarfs were interpreted within the lick index system as primarily being caused by young age , rather than by low metallicity .
derived galaxy ages ranged from 116 gyrs.this interpretation stands in contrast to maraston & thomas ( 2000 , 2001 ) , who showed that a mix of an old metal - rich and an old metal - poor component can produce strong balmer and metal lines without invoking a young population .
a full investigation of this issue is beyond the scope of this paper but we want to recall instead another result from age - sensitive narrow - band photometry in the modified strmgren filtersystem by rakos et al .
these authors derived a mean age of @xmath108gyrs for a sample of 27 fornax dwarf ellipticals and thus provide observational support for a lower age limit of @xmath109 gyrs for luminous early - type dwarf galaxies in nearby galaxy clusters .
if the main age of the stellar populations in our des is 8 gyrs or older , as we assume in the following , @xmath1-band fluctuations and @xmath58 colours depend mostly on metallicity as both quantities closely track the temperature of the rgb whose colour is governed by metallicity .
this relation allows a direct comparison of the colour range covered by the sbf measurements for an individual dwarf with the corresponding metallicity range .
more quantitatively , we compared each set of observed @xmath107 $ ] data points with the predictions of a grid of synthetic stellar populations ( fig.[mbarcalib ] ) : single - burst populations covering the \{age=8 , 12 , 17 gyr } @xmath110 \{[fe / h]=@xmath111 , @xmath112 , ... , @xmath113 , @xmath16 , @xmath114 , @xmath115 } parameter space ( with [ fe / h]@xmath116 in the case of 17gyrs due to model limitations )
. taking into account that local dwarf ellipticals show quite a diversity of evolutionary histories with multiple epochs of starformation ( e.g. da costa 1998 ; grebel 1998 ) we also added a set of composite models where the previously defined single - burst populations were mixed at the 10 , 20 and 30% levels ( in mass ) with a second , intermediate - age generation of 5gyrs old stars with solar metallicity .
the @xmath58 and @xmath65 values for each model were computed with worthey s on - line model interpolation engineworthey / dial / dial_a_pad.html ] with a standard salpeter imf , using the evolutionary isochrones from the padova library ( bertelli et al .
1994 ) and the empirical zero point from jerjen et al .
( 2001 ) .
the model values are graphically presented in fig .
[ mbarcalib ] . single - burst , old , and metal - poor stellar populations have fainter and bluer @xmath1-band fluctuations .
their data define the parabolic branch while more metal - rich and/or composite single - burst populations with a minor contribution of intermediate - age stars make up the linear branch .
the plot symbols along the linear branch in fig .
[ mbarcalib ] covering the @xmath117gyrs age range indicate that this part of the colour fluctuation luminosity relation , although degenerated to some level , is primarily a metallicity sequence . by comparing the location of the sbf data for each de in figs .
6 and 7 with fig .
[ mbarcalib ] we get rough estimates of the metallicities . table [ metallist ] summarises the results and a brief discussion of each de is given here : * vcc0009 * : appears to be among the purest breed of our sample .
the locations of the sbf data on both calibration branches suggest star formation episods at 17 and 812 gyrs .
the metallicity of this major population is [ fe / h]@xmath118 .
the data is further consistent with a pollution at the @xmath119% level with a more metal - rich population of intermediate age stars ( 5 gyrs old ) .
* vcc0490 * : similar to vcc0009 , this dwarf has sbf data points on the parabolic branch and thus appears to have a genuine old stellar population ( 17 gyrs ) with a metallicity of [ fe / h]@xmath118 and a second population ( 812 gyrs ) with similar metallicity .
traces of spiral arms found in the @xmath1-band image of this galaxy ( jerjen et al . 2001 ; barazza et al .
2002 ) may be evidence for an even younger stellar population .
we note that the central part of the dwarf with the spiral structure was not used in the sbf analysis ( see 4 ) . * vcc0810 * : depending on whether the data is on the parabolic or linear branch ( see fig .
[ cmdiagramstwo ] ) we find the following solutions for the stellar population . _ parabolic branch _ :
main population has an age of 17 gyrs with @xmath120[fe / h]@xmath121 and a 20% contribution of a 5 gyrs old population with solar metallicity . _
linear branch _ : 812 gyr old population with metallicity @xmath113 .
* vcc0815 * : is together with vcc0928 the bluest de in our sample ( b r=1.15 ) .
because the sbf data is located on the linear branch it is also the most metal - poor dwarf with a metallicity of [ fe / h]@xmath122 .
the main stellar age is between 8 and 12 gyrs .
the presence of a second stellar population remains undetermined .
* vcc0846 * : the metallicity of the stellar population ( 812 gyrs ) is in the range of @xmath123 [ fe / h ] @xmath121 .
the presence of a second stellar population remains undetermined .
* vcc0856 * : the metallicity of the main stellar population aged between 8 and 12 gyrs is the range of @xmath123[fe / h ] @xmath121 .
caldwell et al .
( 2003 ) derived for vcc0856 a much younger age of @xmath124gyrs and a 0.6dex higher metallicity ( @xmath125 to @xmath126 ) .
the unique spiral structure observed in this early - type dwarf ( jerjen et al .
2000 ) could be optical evidence for recent starformation and thus a population of younger stars in this galaxy .
we note that the central part of the dwarf with the spiral structure was not used in the sbf analysis ( see 4 ) . * vcc0928 * : the location of the sbf data on the parabolic branch implies a major population of 90% of 17 gyrs
old stars ( metallicity range of @xmath127[fe / h]@xmath128 ) with 10% of 5 gyr old , more metal - rich stars . * vcc0929 * : is the most metal - rich de ( [ fe / h]@xmath129 ) in our sample as estimated from @xmath130 .
caldwell et al .
( 2003 ) derive a similar metallicity of [ fe / h]@xmath131 and an age of the stellar population of @xmath132gyrs .
this is in good agreement with the lower age limit ( 8gyrs ) we assumed for our metallicity estimates . *
vcc0940 * : the metallicity of the main stellar population is [ fe / h]@xmath133 with an assumed stellar age of 812 gyrs .
* vcc1010 * : metallicity of the 812 gyrs old stellar population is [ fe / h]@xmath134
. a younger population may be present as indicated by the spiral arms ( barazza et al .
we note that the area with the spiral structure was not used in the sbf analysis ( see 4 ) .
caldwell et al .
( 2003 ) derived a slightly younger age of @xmath135gyrs but with a similar metallicity @xmath136 .
* vcc1036 * : depending on whether the data is on the parabolic or linear branch we find the following solutions for the stellar population
. _ parabolic branch _
: age 17 gyrs with @xmath137\approx 1.0 $ ] , or _
linear branch _ : [ fe / h]@xmath138 and a typical stellar age of 812 gyrs . the independent measurements by caldwell et al . ( 2003 ) give an age of @xmath139gyrs and a metallicity @xmath140 , favouring our latter results and thus supporting a distance modulus of @xmath141 for this dwarf . * vcc1087 * : _ parabolic branch _
: age 17 gyrs with [ fe / h]@xmath142 , or _
linear branch _ [ fe / h]@xmath138 and stellar age of 812 gyrs . * vcc1104 * : the sbf data suggests a metallicity of @xmath143 , assuming an age of 812 gyrs .
* vcc1261 * : similar to vcc0810 we estimate , depending on which calibration branch is used , _ parabolic branch _ : age 17 gyrs with @xmath120[fe / h]@xmath144 and 20% of a 5 gyrs old population with solar metallicity , or _ linear branch _ : a slightly more metal - rich ( [ fe / h]@xmath118 ) , 812 gyrs old population . * vcc1355 * : sbf data is located at the intersection of the two branches .
independent of the stellar age , but assuming it to be older than 8 gyrs , we derive a metallicity of @xmath145 . * vcc1422 * : metallicity of the stellar population is the range of @xmath123[fe / h ] @xmath121 with an assumed stellar age between 812 gyrs .
caldwell et al .
( 2003 ) derive a metallicity of [ fe / h]@xmath146 and an age of the stellar population of @xmath147gyrs .
the presence of a younger population may be indicated by the observed small central edge - on disk ( barazza et al .
but we note that this part of the dwarf was not used in the sbf analysis ( see 4 ) .
the broadband colour based metallicities for our virgo des range from [ fe / h]@xmath148 to @xmath16 .
the mean of 0.95 is 0.2 dex more metal - rich than the observed mean for a sample of eight luminous des in the fornax cluster ( jerjen 2003 ) .
the latter sample contains none of the metal - rich des as vcc0929 and vcc1010 and is centred at a slightly fainter absolute magnitude .
we calculated the absolute @xmath149 magnitudes for the virgo des ( table [ metallist ] ) from their @xmath0 magnitudes ( barazza et al .
2003 ) , adopting a mean @xmath150 colour ( bothun et al .
1989 ) , and employing the newly derived distance moduli .
[ metalmagplot ] shows the location of the virgo des in the metallicity-@xmath151 plane .
the virgo dwarfs share a common region with the mentioned fornax des in the interval @xmath152 .
the points scatter nicely around the metallicity - luminosity relation as set out by the low - luminous des and dsphs in the local group ( da costa 1998 ) and 10 fornax des ( held & mould 1994 ) supporting qualitiatively our results . also included in the figure , for comparison , are the metallicites from caldwell et al .
( 2003 ) for the four common dwarfs vcc0856 , vcc0929 , vcc1010 , and vcc1422 .
it is not surprising that their data preferentially occupy the region above the metallicity - luminosity relation since higher order balmer lines were used as age indicators .
this leads to younger ages ( @xmath153gyrs ) and higher metallicites as pointed out by rakos et al .
( 2001 ) .
clearly , our metallicity and age results from the sbf and @xmath58 measurements are valid under the assumption that the studied sbf fields in each galaxy represent the dominant stellar populations and that a lower age limit of 8 gyrs is correct for those systems . within these constraints
, they offer a valuable alternative to obtain rough metallicities and ages of the unresolved stellar populations of early - type dwarf galaxies .
we have presented the first sbf analysis for 16 dwarf elliptical galaxies in the virgo cluster .
our main results and conclusions are : 1 . following a similar analysis by jerjen ( 2003 ) for fornax cluster des
, we have shown that the sbf method , by employing a good imager at a 8m - class ground - based telescope such as fors@vlt , is an efficient tool to determine accurate distances to dwarf ellipticals as far out as @xmath154 mpc .
the semi - empirical calibration of the @xmath1-band fluctuation magnitudes is in most cases straightforward to achieve by chosing many fields over the face of the galaxy with differing @xmath58 colour .
the mean uncertainty for an individual galaxy in @xmath155 is , to be conservative , @xmath38 0.2 mag , or roughly 10% in distance .
2 . the individual sbf distances of the 16 des range from 14.9 to 12.3 mpc , proving virgo cluster membership for all of them . however , the distance spread is considerable , being roughly twice as large as would be expected if the cluster were spherical symmetric .
so even the early - type component of the virgo cluster is slightly elongated along the line - of - sight , confirming similar findings based on the distribution of giant es and intracluster pne .
the combined giant+dwarf virgo early - type sbf distances show a significantly bimodal distribution .
there is a broad primary clump of galaxies at @xmath156 16 mpc and a narrow secondary clump around 18.5 mpc .
these features can be associated with the two major virgo core subclusters centered on m87 and m86 .
the latter , smaller subcluster seems to be falling into the dominating m87 clump from the backside , separated by ca .
a trace of the expected infall pattern is indeed seen in a plot of sbf distance versus heliocentric radial velocity .
the large depth and bimodality of the distance distribution of virgo cluster early types renders a clear - cut definition of the ` virgo cluster mean distance ' difficult .
as the m87 subclump , judged from x - ray images , is undoubtedly the most massive structure of the virgo complex , it is most natural and physically meaningful to identify it with the virgo cluster proper . in this case
the mean distance of the ` virgo cluster ' is @xmath38 16 mpc , rather than @xmath38 17 mpc .
the surface brightness fluctuations of a galaxy also serve as a constraint on its stellar contents . in particular
, the model calibration of the observed colour - fluctuation magnitude relation can be used to derive metallicities for the virgo des that agree with narrow - band and spectroscopic measurements .
the method works if the dwarfs are at least as old as 8 gyr , for which there is independent evidence .
the sbf metallicies of the present virgo des range from [ fe / h ] @xmath15 to @xmath16 .
6 . given the great potential of the sbf method to determine the distances of virgo des in an efficient way ( low costs in terms of telescope time ) and with 10% accuracy , it should prove extremely rewarding for our understanding of the 3d structure of that most nearby cluster of galaxies , if this work could be extended and a massive observational campaign be started to get a good fraction of the roughly 800 more dwarf ellipticals in the virgo cluster left .
we like to thank the eso staff of the vlt ut1+ut2 for providing excellent quality images in the service observing mode and the anonymous referee for carefully reading the manuscript .
the authors are grateful to the _ swiss national science foundation _ for the financial support of this research project .
da costa , g. s. 1998 , stellar astrophysics for the local group : viii canary islands winter school of astrophysics , eds .
a. aparicio , a. herrero , and f. sanchez .
cambridge ; new york : cambridge university press , p. 351
cllccclr 0009 & ic3019 & de1,n & 12h09m22.3s & + 13@xmath157 & a & 14.04 & 1804 + 0490 & ic0783 & ds0(3),n & 12h21m38.8s & + 15@xmath158 & a & 13.97 & 1293 + 0810 & 13o42 & de0,n & 12h25m33.6s & + 13@xmath159 & a & 16.68 & @xmath160340 + 0815 & 13o47 & de2,n & 12h25m37.1s & + 13@xmath161 & a & 15.95 & @xmath160700 + 0846 & 13o48 & de1,n & 12h25m50.4s & + 13@xmath162 & a & 16.19 & @xmath160730 + 0856 & ic3328 & de1,n & 12h25m57.9s & + 10@xmath163 & b & 14.19 & 972 + 0928 & 12o42 & de6,n & 12h26m39.6s & + 12@xmath164 & a & 16.13 & @xmath68 + 0929 & ngc4415 & d : e1,n & 12h26m40.9s & + 08@xmath165 & b & 13.69 & 910 + 0940 & ic3349 & de1,n & 12h26m47.1s & + 12@xmath166 & a & 14.81 & 1563 + 1010 & ngc4431 & ds0(5),n & 12h27m26.6s & + 12@xmath167 & a & 13.86 & 913 + 1036 & ngc4436 & de6/ds0(6),n & 12h27m41.6s & + 12@xmath168 & a & 13.89 & 1163 + 1087 & ic3381 & de1,n & 12h28m15.1s & + 11@xmath169 & a & 14.15 & 645 + 1104 & ic3388 & de5,n & 12h28m27.9s & + 12@xmath170 & a & 15.49 & 1704 + 1261 & ngc4482 & d : e5.n & 12h30m10.4s & + 10@xmath171 & a & 13.59 & 1850 + 1355 & ic3442 & de2,n & 12h31m20.0s & + 14@xmath172 & a & 14.52 & 1332 + 1422 & ic3468 & e1,n : & 12h32m14.2s & + 10@xmath173 & a & 13.80 & 1372 + lcccccccccc 0009 & @xmath0 & 2000 - 03 - 28 & 1.28 & @xmath174 & 0.60 & @xmath1 & 1999 - 07 - 14 & 1.55 & @xmath174 & 0.65 + 0490 & @xmath0 & 2000 - 03 - 28 & 1.34 & @xmath174 & 0.70 & @xmath1 & 1999 - 07 - 10 & 1.41 & @xmath174 & 0.45 + 0810 & @xmath0 & 2000 - 05 - 01 & 1.41 & @xmath175 & 0.90 & @xmath1 & 1999 - 07 - 13 & 1.57 & @xmath175 & 0.60 + 0815 & @xmath0 & 2000 - 05 - 01 & 1.41 & @xmath175 & 0.90 & @xmath1 & 1999 - 07 - 13 & 1.57 & @xmath175 & 0.60 + 0846 & @xmath0 & 2000 - 05 - 01 & 1.41 & @xmath175 & 0.90 & @xmath1 & 1999 - 07 - 13 & 1.57 & @xmath175 & 0.60 + 0856 & @xmath0 & 2000 - 03 - 28 & 1.29 & @xmath174 & 0.85 & @xmath1 & 1999 - 07 - 13 & 1.39 & @xmath174 & 0.55 + 0928 & @xmath0 & 2000 - 04 - 01 & 1.36 & @xmath176 & 0.80 & @xmath31 & 2000 - 04 - 05 & 1.30 & @xmath176 & 0.60 + 0929 & @xmath0 & 2000 - 04 - 05 & 1.19 & @xmath174 & 0.70 & @xmath1 & 1999 - 07 - 14 & 1.47 & @xmath174 & 0.60 + 0940 & @xmath0 & 2000 - 04 - 01 & 1.36 & @xmath176 & 0.80 & @xmath31 & 2000 - 04 - 05 & 1.30 & @xmath176 & 0.45 + 1010 & @xmath0 & 2000 - 04 - 01 & 1.29 & @xmath174 & 0.70 & @xmath31 & 2000 - 04 - 01 & 1.25 & @xmath174 & 0.65 + 1036 & @xmath0 & 2000 - 04 - 01 & 1.29 & @xmath174 & 0.65 & @xmath31 & 2000 - 04 - 01 & 1.25 & @xmath174 & 0.65 + 1087 & @xmath0 & 2000 - 04 - 01 & 1.27 & @xmath174 & 0.65 & @xmath31 & 2000 - 04 - 01 & 1.24 & @xmath174 & 0.60 + 1104 & @xmath0 & 2000 - 04 - 05 & 1.71 & @xmath176 & 0.55 & @xmath31 & 2000 - 04 - 05 & 1.51 & @xmath175 & 0.45 + 1261 & @xmath0 & 2000 - 04 - 01 & 1.43 & @xmath174 & 0.85 & @xmath31 & 2000 - 04 - 05 & 1.23 & @xmath174 & 0.50 + 1355 & @xmath0 & 2000 - 03 - 30 & 1.39 & @xmath176 & 0.90 & @xmath31 & 2000 - 04 - 05 & 1.40 & @xmath176 & 0.45 + 1422 & @xmath0 & 2000 - 04 - 01 & 1.56 & @xmath174 & 0.75 & @xmath31 & 2000 - 04 - 05 & 1.21 & @xmath174 & 0.50 + lcccccccc vcc0009 f1 & 100 & 27.11 & 1992.9 & 10326.7 & 0.070(0.003 ) & 0.005 & 10.9 & 0.02 + f2 & 100 & & 1393.1 & & 0.065(0.003 ) & 0.007 & 7.9 & 0.02 + f3 & 100 & & 480.2 & & 0.082(0.004 ) & 0.017 & 4.3 & 0.02 + f4 & 100 & & 516.7 & & 0.082(0.006 ) & 0.017 & 4.5 & 0.02 + f5 & 100 & & 681.3 & & 0.067(0.005 ) & 0.012 & 4.7 & 0.02 + f6 & 100 & & 1412.5 & & 0.069(0.002 ) & 0.006 & 8.4 & 0.02 + f7 & 100 & & 556.8 & & 0.090(0.005 ) & 0.015 & 5.3 & 0.02 + vcc0490 f1 & 100 & 27.12 & 952.1 & 26184.5 & 0.078(0.003 ) & 0.017 & 4.1 & 0.02 + f2 & 80 & & 404.8 & & 0.107(0.007 ) & 0.032 & 3.1 & 0.01 + f3 & 100 & & 486.4 & & 0.098(0.003 ) & 0.029 & 3.2 & 0.02 + f4 & 100 & & 932.6 & & 0.076(0.003 ) & 0.017 & 4.1 & 0.02 + f5 & 100 & & 766.3 & & 0.074(0.003 ) & 0.020 & 3.4 & 0.02 + f6 & 50 & & 3568.1 & & 0.051(0.003 ) & 0.005 & 7.8 & 0.03 + f7 & 100 & & 498.5 & & 0.082(0.004 ) & 0.029 & 2.6 & 0.02 + vcc0810 f1 & 90 & 27.10 & 842.3 & 20052.6 & 0.078(0.005 ) & 0.013 & 5.1 & 0.02 + f2 & 60 & & 870.2 & & 0.081(0.006 ) & 0.014 & 5.0 & 0.02 + f3 & 60 & & 1243.2 & & 0.079(0.006 ) & 0.007 & 9.0 & 0.02 + f4 & 60 & & 1153.2 & & 0.082(0.006 ) & 0.008 & 8.3 & 0.02 + vcc0815 f1 & 60 & 27.10 & 942.0 & 20052.6 & 0.077(0.003 ) & 0.008 & 7.7 & 0.02 + f2 & 60 & & 449.8 & & 0.089(0.006 ) & 0.016 & 4.9 & 0.02 + f3 & 60 & & 653.5 & & 0.108(0.006 ) & 0.013 & 7.1 & 0.01 + f4 & 60 & & 1277.9 & & 0.085(0.003 ) & 0.007 & 10.0 & 0.02 + f5 & 60 & & 669.6 & & 0.091(0.003 ) & 0.012 & 6.6 & 0.02 + f6 & 80 & & 1001.0 & & 0.076(0.003 ) & 0.009 & 6.9 & 0.02 + vcc0846 f1 & 70 & 27.10 & 1236.3 & 20052.6 & 0.084(0.006 ) & 0.008 & 8.3 & 0.02 + f2 & 70 & & 1107.0 & & 0.075(0.004 ) & 0.009 & 6.8 & 0.02 + f3 & 70 & & 1246.2 & & 0.059(0.003 ) & 0.008 & 6.1 & 0.03 + f4 & 70 & & 325.2 & & 0.104(0.007 ) & 0.025 & 3.8 & 0.01 + f5 & 80 & & 981.9 & & 0.074(0.008 ) & 0.011 & 5.6 & 0.02 + vcc0856 f1 & 100 & 27.12 & 176.1 & 10619.7 & 0.120(0.007 ) & 0.049 & 2.3 & 0.01 + f2 & 100 & & 357.5 & & 0.084(0.004 ) & 0.028 & 2.8 & 0.02 + f3 & 70 & & 583.1 & & 0.074(0.002 ) & 0.015 & 4.4 & 0.02 + vcc0928 f1 & 60 & 27.35 & 503.9 & 10080.5 & 0.073(0.005 ) & 0.014 & 4.3 & 0.03 + f2 & 60 & & 1986.7 & & 0.062(0.006 ) & 0.006 & 7.3 & 0.04 + f3 & 60 & & 891.6 & & 0.063(0.004 ) & 0.009 & 5.3 & 0.04 + f4 & 60 & & 2172.5 & & 0.059(0.004 ) & 0.004 & 9.1 & 0.04 + f5 & 60 & & 2071.7 & & 0.050(0.004 ) & 0.006 & 5.7 & 0.05 + vcc0929 f1 & 120 & 27.12 & 2830.0 & 10408.2 & 0.046(0.002 ) & 0.005 & 7.2 & 0.03 + f2 & 120 & & 3650.1 & & 0.040(0.002 ) & 0.004 & 7.0 & 0.04
+ f3 & 120 & & 1553.6 & & 0.046(0.003 ) & 0.008 & 4.7 & 0.03 + f4 & 120 & & 1766.1 & & 0.048(0.003 ) & 0.007 & 5.2 & 0.03 + f5 & 120 & & 1212.6 & & 0.050(0.003 ) & 0.009 & 4.4 & 0.03 + f6 & 120 & & 1137.7 & & 0.051(0.002 ) & 0.010 & 4.4 & 0.03 + f7 & 120 & & 696.5 & & 0.052(0.003 ) & 0.014 & 3.2 & 0.03 + lcccccccc vcc0940 f1 & 90 & 27.35 & 2663.8 & 10080.5 & 0.058(0.002 ) & 0.003 & 10.2 & 0.04 + f2 & 100 & & 1535.7 & & 0.056(0.002 ) & 0.005 & 6.9 & 0.04 + f3 & 110 & & 661.8 & & 0.068(0.003 ) & 0.012 & 4.6 & 0.04 + f4 & 60 & & 342.3 & & 0.116(0.007 ) & 0.016 & 6.1 & 0.02 + vcc1010 f1 & 80 & 27.32 & 1303.8 & 8355.3 & 0.054(0.002 ) & 0.005 & 7.4 & 0.04 + f2 & 80 & & 1210.8 & & 0.052(0.004 ) & 0.005 & 6.8 & 0.04 + f3 & 80 & & 876.3 & & 0.060(0.004 ) & 0.008 & 5.9 & 0.04 + f4 & 80 & & 660.3 & & 0.075(0.004 ) & 0.010 & 6.0 & 0.03 + f5 & 80 & & 812.0 & & 0.064(0.007 ) & 0.009 & 5.6 & 0.04 + f6 & 80 & & 947.3 & & 0.059(0.004 ) & 0.007 & 6.2 & 0.04 + vcc1036 f1 & 60 & 27.32 & 1102.7 & 8355.3 & 0.078(0.004 ) & 0.006 & 9.6 & 0.03 + f2 & 60 & & 1734.3 & & 0.078(0.003 ) & 0.005 & 11.0 & 0.03 + f3 & 60 & & 1340.8 & & 0.074(0.003 ) & 0.005 & 9.7 & 0.03 + f4 & 60 & & 2122.0 & & 0.073(0.004 ) & 0.004 & 11.7 & 0.03 + vcc1087 f1 & 70 & 27.31 & 3198.6 & 7903.0 & 0.054(0.002 ) & 0.002 & 11.1 & 0.04 + f2 & 60 & & 3237.6 & & 0.056(0.002 ) & 0.003 & 11.4 & 0.04 + f3 & 60 & & 2998.2 & & 0.068(0.002 ) & 0.003 & 13.0 & 0.03 + f4 & 60 & & 2514.5 & & 0.058(0.002 ) & 0.003 & 10.4 & 0.04 + vcc1104 f1 & 90 & 27.33 & 2478.4 & 14213.5 & 0.108(0.004 ) & 0.006 & 12.0 & 0.02 + f2 & 90 & & 1933.2 & & 0.117(0.002 ) & 0.007 & 12.1 & 0.02 + f3 & 90 & & 1516.8 & & 0.102(0.003 ) & 0.009 & 8.9 & 0.02 + f4 & 90 & & 1913.6 & & 0.118(0.004 ) & 0.009 & 9.9 & 0.02 + vcc1261 f1 & 110 & 27.37 & 5418.7 & 7873.2 & 0.058(0.002 ) & 0.002 & 12.6 & 0.04 + f2 & 100 & & 4104.7 & & 0.054(0.001 ) & 0.002 & 10.9 & 0.05 + f3 & 110 & & 1945.4 & & 0.065(0.003 ) & 0.005 & 8.8 & 0.04 + f4 & 110 & & 983.1 & & 0.069(0.003 ) & 0.006 & 7.7 & 0.04 + vcc1355 f1 & 90 & 27.34 & 2377.9 & 10534.8 & 0.078(0.003 ) & 0.003 & 13.2 & 0.03 + f2 & 80 & & 2640.1 & & 0.075(0.003 ) & 0.003 & 12.7 & 0.03 + f3 & 90 & & 2360.9 & & 0.071(0.002 ) & 0.004 & 11.6 & 0.03 + f4 & 90 & & 1834.7 & & 0.076(0.003 ) & 0.004 & 11.1 & 0.03 + vcc1422 f1 & 120 & 27.38 & 3695.1 & 7571.6 & 0.076(0.002 ) & 0.003 & 13.0 & 0.03 + f2 & 120 & & 2209.8 & & 0.076(0.003 ) & 0.004 & 10.8 & 0.03 + f3 & 120 & & 1018.1 & & 0.092(0.003 ) & 0.007 & 9.4 & 0.03 + f4 & 120 & & 439.2 & & 0.109(0.002 ) & 0.014 & 6.6 & 0.02 + f5 & 120 & & 3987.8 & & 0.066(0.002 ) & 0.003 & 12.3 & 0.04 + f6 & 120 & & 876.2 & & 0.089(0.004 ) & 0.008 & 8.5 & 0.03 + cccc vcc0009 f1 & @xmath177 & @xmath178 & @xmath179 + f2 & & @xmath180 & @xmath181 + f3 & & @xmath182 & @xmath183 + f4 & & @xmath184 & @xmath185 + f5 & & @xmath186 & @xmath187 + f6 & & @xmath188 & @xmath189 + f7 & & @xmath190 & @xmath191 + vcc0490 f1 & @xmath192 & @xmath193 & @xmath194 + f2 & & @xmath195 & @xmath196 + f3 & & @xmath197 & @xmath198 + f4 & & @xmath199 & @xmath200 + f5 & & @xmath201 & @xmath202 + f6 & & @xmath203 & @xmath204 + f7 & & @xmath205 & @xmath206 + vcc0810 f1 & @xmath207 & @xmath208 & @xmath209 + f2 & & @xmath210 & @xmath211 + f3 & & @xmath212 & @xmath213 + f4 & & @xmath214 & @xmath215 + vcc0815 f1 & @xmath207 & @xmath216 & @xmath217 + f2 & & @xmath218 & @xmath219 + f3 & & @xmath220 & @xmath221 + f4 & & @xmath222 & @xmath179 + f5 & & @xmath223 & @xmath224 + f6 & & @xmath225 & @xmath179 + vcc0846 f1 & @xmath207 & @xmath184 & @xmath226 + f2 & & @xmath227 & @xmath226 + f3 & & @xmath228 & @xmath229 + f4 & & @xmath230 & @xmath231 + f5 & & @xmath232 & @xmath229 + vcc0856 f1 & @xmath233 & @xmath234 & @xmath196 + f2 & & @xmath235 & @xmath236 + f3 & & @xmath237 & @xmath213 + vcc0928 f1 & @xmath238 & @xmath239 & @xmath240 + f2 & & @xmath241 & @xmath221 + f3 & & @xmath242 & @xmath243 + f4 & & @xmath244 & @xmath217 + f5 & & @xmath245 & @xmath226 + vcc0929 f1 & @xmath192 & @xmath246 & @xmath247 + f2 & & @xmath248 & @xmath249 + f3 & & @xmath250 & @xmath251 + f4 & & @xmath252 & @xmath247 + f5 & & @xmath253 & @xmath251 + f6 & & @xmath254 & @xmath247 + f7 & & @xmath255 & @xmath256 + cccc vcc0940 f1 & @xmath207 & @xmath257 & @xmath258 + f2 & & @xmath259 & @xmath260 + f3 & & @xmath261 & @xmath262 + f4 & & @xmath263 & @xmath264 + vcc1010 f1 & @xmath207 & @xmath265 & @xmath266 + f2 & & @xmath267 & @xmath266 + f3 & & @xmath268 & @xmath269 + f4 & & @xmath270 & @xmath271 + f5 & & @xmath272 & @xmath273 + f6 & & @xmath274 & @xmath275 + vcc1036 f1 & @xmath233 & @xmath276 & @xmath277 + f2 & & @xmath278 & @xmath279 + f3 & & @xmath280 & @xmath260 + f4 & & @xmath281 & @xmath282 + vcc1087 f1 & @xmath233 & @xmath283 & @xmath251 + f2 & & @xmath284 & @xmath285 + f3 & & @xmath286 & @xmath287 + f4 & & @xmath288 & @xmath287 + vcc1104 f1 & @xmath192 & @xmath289 & @xmath226 + f2 & & @xmath290 & @xmath217 + f3 & & @xmath291 & @xmath292 + f4 & & @xmath293 & @xmath294 + vcc1261 f1 & @xmath207 & @xmath295 & @xmath292 + f2 & & @xmath296 & @xmath297 + f3 & & @xmath298 & @xmath297 + f4 & & @xmath299 & @xmath292 + vcc1355 f1 & @xmath238 & @xmath300 & @xmath258 + f2 & & @xmath301 & @xmath275 + f3 & & @xmath302 & @xmath275 + f4 & & @xmath303 & @xmath275 + vcc1422 f1 & @xmath238 & @xmath304 & @xmath204 + f2 & & @xmath305 & @xmath294 + f3 & & @xmath306 & @xmath204 + f4 & & @xmath307 & @xmath231 + f5 & & @xmath308 & @xmath294 + f6 & & @xmath309 & @xmath310 + llcc 0009 & ic3019 & 31.00@xmath78 0.10 & 15.85@xmath78 0.80 + 0490 & ic0783 & 31.31@xmath78 0.16 & 18.28@xmath78 1.46 + 0810 & 13o42 & 30.76@xmath78 0.13 & 14.19@xmath78 0.92 + & & 31.16@xmath78 0.22 & 17.06@xmath78 1.88 + 0815 & 13o47 & 31.48@xmath78 0.17 & 19.77@xmath78 1.68 + 0846 & 13o48 & 31.30@xmath78 0.19 & 18.20@xmath78 1.73 + 0856 & ic3328 & [email protected] & 18.03@xmath78 2.25 + 0928 & 12o42 & 31.42@xmath78 0.14 & 19.23@xmath78 1.34 + 0929 & ngc4415 & 30.86@xmath78 0.14 & 14.86@xmath78 1.04 + 0940 & ic3349 & 31.36@xmath78 0.20 & 18.71@xmath78 1.87 + 1010 & ngc4431 & 31.05@xmath78 0.15 & 16.22@xmath78 1.22 + 1036 & ngc4436 & 30.94@xmath78 0.12 & 15.42@xmath78 0.93 + & & 31.13@xmath78 0.17 & 16.83@xmath78 1.43 + 1087 & ic3381 & 31.27@xmath78 0.14 & 17.95@xmath78 1.26 + & & 31.39@xmath78 0.18 & 18.97@xmath78 1.71 + 1104 & ic3388 & 31.15@xmath78 0.19 & 16.98@xmath78 1.61 + 1261 & ngc4482 & 31.34@xmath78 0.16 & 18.54@xmath78 1.48
+ & & 31.74@xmath78 0.21 & 22.28@xmath78 2.34 + 1355 & ic3442 & 30.92@xmath78 0.17 & 15.28@xmath78 1.30 + 1422 & ic3468 & 31.64@xmath78 0.18 & 21.28@xmath78 1.92 + ccc vcc0009 & @xmath311 & @xmath113 + vcc0490 & @xmath312 & @xmath113 + vcc0810 & @xmath313 & @xmath143 + vcc0815 & @xmath314 & @xmath315 + vcc0846 & @xmath316 & @xmath317 + vcc0856 & @xmath318 & @xmath317 + vcc0928 & @xmath319 & @xmath143 + vcc0929 & @xmath320 & @xmath16 + vcc0940 & @xmath321 & @xmath322 + vcc1010 & @xmath323 & @xmath16 + vcc1036 & @xmath324 & @xmath325 + vcc1087 & @xmath326 & @xmath325 + vcc1104 & @xmath327 & @xmath143 + vcc1261 & @xmath328 & @xmath113 + vcc1355 & @xmath329 & @xmath145 + vcc1422 & @xmath330 & @xmath317 + | we have employed fors1 and 2 at the very large telescope at eso to acquire deep @xmath0 and @xmath1-band ccd images of 16 dwarf elliptical galaxies in the direction of the virgo cluster . for each dwarf
we measure the apparent @xmath1-band surface brightness fluctuation ( sbf ) magnitude @xmath2 and the @xmath3 colour in a number of fields at different galactocentric distances . from the field - to - field variation of the two quantities
we determine the sbf distance by means of the @xmath4 relation .
the derived distances of the dwarfs are ranging from 14.9mpc to 21.3mpc , with a mean 1@xmath5 uncertainty of 1.4 mpc or 8% of the distance , confirming that there is considerable depth in the distance distribution of early - type cluster members . for vcc1104 ( ic3388 )
our sbf distance modulus of @xmath6 ( @xmath7mpc ) is in good agreement with the harris et al .
( 1998 ) result of @xmath8mag ( @xmath9mpc ) based on hst observations and the tip magnitude of the red giant branch .
combining our results with existing distances for giant virgo ellipticals we identify two major galaxy concentrations in the distance distribution : a broad primary clump around @xmath10mag ( @xmath11mpc ) and a narrow secondary clump around @xmath12mag ( @xmath13mpc ) .
an adaptive kernel analysis finds the two concentrations to be significant at the 99% ( 2.5@xmath5 ) and 89% ( @xmath14 ) levels .
while the near - side clump of virgo early - type galaxies can be associated to the subcluster centered on m87 , the second clump is believed to be mainly due to the backside infalling group of galaxies around m86 . +
the ages and metallicities of the de stellar populations are estimated by combining the observed @xmath3 colours with worthey s stellar population synthesis models .
it appears that the virgo des cover a wider range in metallicity , from [ fe / h]@xmath15 ( vcc0815 ) to @xmath16 ( ngc4415 ) , than fornax cluster des .
the derived metallicities place the virgo des on the extension of the metallicity luminosity relation defined by the low - luminousity local group des .
the data further suggest an age range from genuinly old ( @xmath17gyrs ) stellar systems like ic3019 and ic0783 to intermediate - age ( @xmath18gyrs ) dwarfs like ngc4431 and ic3468 . |
jdw is supported by a sandia laboratories / purdue university excellence in science and engineering fellowship .
mjm acknowledges support from the miller family foundation .
the work by maryland ( eh and sds ) is supported by microsoft q. gac acknowledges support from the nsf dmr-0907172 .
50 c. gerl , s. schmult , h .-
tranitz , c. mitzkus , and w. wegscheider , appl .
. lett . * 86 * , 252105 ( 2005 ) .
m. j. manfra , l. n. pfeiffer , k. w. west , r. de picciotto , and k. w. baldwin , appl .
. lett . * 86 * , 162106 ( 2005 ) .
m. j. manfra , e. h. hwang , s. das sarma , l. n. pfeiffer , k. w. west , and a. m. sergent , phys .
lett . * 99 * , 236402 ( 2007 ) .
a. kumar , n. samkharadze , m. j. manfra , g. a. csthy , l. n. pfeiffer , k. w. west , arxiv:1007.1504v1 ( 2010 ) . b. grbi , r. leturcq , t. ihn , k. ensslin , d. reuter , and a. d. wieck , phys . rev .
* 99 * , 176803 ( 2007 ) .
s. p. koduvayur , y. lyanda - geller , s. khlebnikov , g. csthy , m. j. manfra , l. n. pfeiffer , k. w. west , and l. p. rokhinson , phys .
106 * , 016804 ( 2011 ) .
m. j. manfra , r. de picciotto , z. jiang , s. h. simon , l. n. pfeiffer , k. w. west , and a. m. sergent , phys .
lett . * 98 * , 206804 ( 2007 ) .
e. f. schubert , _ doping in iii - v semiconductors _ ( cambridge university press , cambridge , 1993 ) .
j. j. heremans , m. b. santos , k. hirakawa , and m. shayegan , j. appl .
* 76 * , 1980 ( 1994 ) .
nextnano3 simulator 1999 - 2008 walter schottky institute .
e. h. hwang and s. das sarma , phys .
b * 77 * , 235437 ( 2008 ) .
v. umansky , r. de picciotto , and m. heiblum , appl .
. lett . * 71 * , 683 ( 1997 ) .
k. lee , m. s. shur , t. j. drummond , and h. morko , j. appl . phys . * 54 * , 6432 ( 1983 ) . c. jiang , d. c. tsui , and g. weimann , appl .
. lett . * 53 * , 1533 ( 1988 ) .
j. a. seamons , d. r. tibbetts , j. l. reno , and m. p. lilly , appl .
lett . * 90 * , 052103 ( 2007 ) . e. h. hwang and s. das sarma , phys
b * 78 * , 075430 ( 2008 )
. t. m. lu , z. f. li , d. c. tsui , m. j. manfra , l. n. pfeiffer , and k. w. west , appl .
. lett . * 92 * , 012109 ( 2008 ) .
s. das sarma and e. h. hwang , phys .
* 83 * , 164 ( 1999 ) .
s. das sarma and e. h. hwang , phys .
b * 61 * , 7838 ( 2000 ) . | we report on a systematic study of the density dependence of mobility in a low - density carbon - doped ( 100 ) gaas two - dimensional hole system ( 2dhs ) . at t=50mk , a mobility of 2.6 @xmath0 10@xmath1 @xmath2/vs at a density p=6.2
@xmath0 10@xmath3 @xmath4 was measured .
this is the highest mobility reported for a 2dhs to date . using a back - gated sample geometry ,
the density dependence of mobility was studied from 2.8 @xmath0 10@xmath3 @xmath4 to 1 @xmath0 10@xmath5 @xmath4 .
the mobility vs. density can not be fit to a power law dependence of the form @xmath6 @xmath7 p@xmath8 using a single exponent @xmath9 .
our data indicate a continuous evolution of the power law with @xmath9 ranging from @xmath7 0.7 at high density and increasing to @xmath7 1.7 at the lowest densities measured .
calculations specific to our structure indicate a crossover of the dominant scattering mechanism from uniform background impurity scattering at high density to remote ionized impurity scattering at low densities .
this is the first observation of a carrier density - induced transition from background impurity dominated to remote dopant dominated transport in a single sample .
the two - dimensional hole system ( 2dhs ) offers an attractive platform for the study of strong carrier interactions parameterized by @xmath10 : the ratio of the coulomb energy to the fermi energy .
@xmath11 , where @xmath12 is the hole density and @xmath13 is the effective mass .
recent developments in the growth of carbon - doped ( 100 ) gaas heterostructures by molecular beam epitaxy ( mbe ) have resulted in 2dhss of unprecedented quality @xcite@xmath14 @xcite .
such structures have been utilized in the study the metal - to - insulator transition ( mit ) @xcite , fractional quantum hall physics in the 2nd landau level ( ll)@xcite , spin - orbit coupling in aharonov bohm rings @xcite and charge density wave formation in partially filled lls@xcite@xmath14@xcite .
these initial experiments and the prospect of studying strong correlations in the presence of tunable spin - orbit coupling provide strong motivation to understand the scattering processes presently limiting mobility in the highest quality samples . here
we present mobility vs. density data on an unprecedently high mobilty 2dhs .
one of the most exciting avenues for future research is the investigation of ultra - low density 2dhss at very large @xmath10 .
thus , our data and calculations will inform the design of new hole heterostructures of ever increasing quality .
carbon doping@xcite of 2dhss offers advantages over the more commonly used acceptor dopants beryllium and silicon .
carbon diffuses and surface segregates much less at typical mbe growth temperatures ( t@xmath7630 @xmath15c ) than beryllium@xcite .
additionally carbon can be incorporated as an acceptor on multiple crystallographic orientations , including on the high - symmetry ( 100 ) face of gaas .
silicon can also act as an acceptor to produce high quality 2dhss but so far high mobility ( @xmath16 @xmath17/vs ) silicon - doped 2dhss have only been realized on ( 311)a face@xcite .
the ( 311)a face has a well known mobility anisotropy between the [ @xmath18 and [ @xmath19 directions @xcite whereas carbon - doped structures on the ( 100 ) face have a significantly lower anisotropy between the [ @xmath20 and [ @xmath21 directions @xcite . furthermore , the high symmetry of the ( 100 ) orientation dramatically alters the nature of spin - orbit interactions in 2dhss as compared to quantization along the ( 311)a direction .
indeed , further experimental work is needed to fully exploit the potential benefits of carbon - doped ( 100 ) 2dhss .
our sample consists of a 20 nm al@xmath22ga@xmath23as / gaas / al@xmath22ga@xmath23as quantum well asymmetrically modulation doped with carbon at a density of 1 @xmath0 10@xmath24 @xmath4 above the quantum well at a setback of 80 nm .
fig.[qw ] shows a sketch of the device along with the numerically calculated @xcite band structure and heavy hole ground state wavefunction ( normalized to unity ) . for simplicity in simulation
the superlattice and buffer regions are truncated . in order to modulate the density in the quantum
well , we utilized a back - gate geometry .
the sample was first thinned to approximately 150 @xmath6 m and then cleaved into a hall bar approximately 2 mm @xmath0 9 mm .
ohmic contacts consist of in / zn dots positioned approximately 1 mm apart along the length of the hall bar and annealed at t = 430 @xmath15c .
the hall bar was subsequently fixed to a gold backgate evaporated on an undoped gaas substrate .
the carrier density was measured from minima in the longitudinal magnetoresistance , and the conductivity was obtained from four - terminal zero field measurements using standard lock - in techniques . as shown in fig .
[ vg ] , the 2dhs density depended linearly on voltage over the range measured .
modeling the structure as a parallel plate capacitor with one plate being the 2dhs and the other being the backgate we estimate the gate to be situated 175 @xmath6 m from the well .
the peak mobility @xmath6 at low temperature ( t = 50 mk ) was measured to be 2.6 @xmath0 10@xmath1 @xmath2/vs at a density of 6.2 @xmath0 10@xmath3 @xmath4 in an as - grown sample .
fig .
[ vg ] also shows the leakage current as a function of the gate voltage .
the linear dependence of the leakage current on the gate voltage and its small magnitude ( @xmath25 1.5 na as compared to an excitation current of 50 na ) suggest that the observed leakage represents parasitic current passing through the measurement circuit and not hard breakdown in the gaas .
in addition , the linear dependence of the density on the gate voltage also suggests that sharp breakdown did not occur .
fig .
[ mr ] shows a representative trace of the magnetoresistance at t = 50 mk .
the deep minima in the fractional quantum hall states around filling factor @xmath26 illustrate the high quality of the processed sample .
we note that this sample has also been studied at ultra - low temperatures ( t @xmath27 10 mk ) in which the first evidence of a fully formed fractional quantum hall state at @xmath28 in the 2nd landau level in a 2dhs was observed @xcite . in order to examine the scattering mechanisms limiting mobility in our system , we measured the dependence of the mobility on the 2d hole density modulated by the backgate as shown in fig .
[ n_v_mu ] .
as can be clearly seen on this log - log plot the data points do not fall on a straight line as would be expected for a single dominant scattering mechanism .
the mobility vs. density can not be fit to a power law dependence of the form @xmath6
@xmath7 p@xmath8 using a single exponent @xmath9 .
our data indicate a continuous evolution of the power law with @xmath9 ranging from @xmath7 0.7 at high density and increasing to @xmath71.7 at the lowest densities measured .
thus the data indicate the presence of multiple dominant scattering mechanisms over the range of density tested .
indeed at the lowest densities measured , the mobility decreases rapidly indicating that the system will eventually approach a finite density mit@xcite .
we emphasize , however , that @xmath29 , the product of the fermi wavevector and the mean free path , remains larger than 50 over the entire range of density tested .
it can be seen at high density that the mobility follows a power law behavior @xmath30 which is indicative of uniformly distributed charged background impurity ( bi ) scattering @xcite@xmath14 @xcite in 2d carrier systems . in @xmath31 .
( b ) comparison of experimental mobility data and theoretical results .
solid line represents rpa - boltzmann calculation and dashed line represents rpa - boltzmann calculation with the hubbard approximation in both plots.,width=288 ] however , the power law continuously shifts towards higher @xmath9 of approximately 1.7 at low density .
typically , an exponent @xmath32 is taken as an indication of the dominance of remote ionized ( ri ) impurity scattering@xcite@xmath33@xcite originating in the remote doping layer in 2d systems .
we emphasize that it is unusual that such a crossover behavior can be seen in a single sample while remaining in the high mobility ( or equivalently large @xmath29 ) regime .
it is interesting to note that our setback distance is 80 nm , a distance at which remote ionized impurity scattering usually makes a relatively minor contribution to the total scattering in samples with density @xmath34 @xmath4 .
nevertheless , it is clear that for the material parameters of our structure it dominates scattering at lower density .
a similar transition to remote ionized impurity scattering in low density 2d electron samples was report by jiang et al .
@xcite , but in these samples @xmath29 was substantially lower and the samples were approaching a conduction threshold .
unusual transport behavior was also recently reported in the 2dhs in an undoped electron - hole bilayer @xcite and subsequently explained by carrier inhomogeneities resulting from strong carrier - carrier interactions and non - linear screening @xcite .
it could be argued that t = 300 mk data is only marginally outside the range where phonons could be playing some role , but the fact that our t = 50 mk data displays the same functional dependence seems to rule out any critical scattering contributions from phonons .
interface roughness ( ir ) scattering should also be considered , but significant ir scattering should also manifest itself in a significant mobility anisotropy @xcite which was not observed in our sample . finally , scattering between the light and heavy hole bands@xcite as well as changes in the effective mass @xcite have also been reported .
however , both of these effects should decrease as the density is lowered and the fermi level moves towards the top of the valence band and away from light hole band . to explain our experimental observation we calculated the coulomb scattering rate due to background charged impurity scattering and remote charged impurity scattering using a boltzmann transport method@xcite@xmath14 @xcite .
screening was taken into account using the random phase approximation ( rpa ) .
the calculation was performed both with and without the inclusion of correlation effects via the hubbard approximation .
the analysis assumed a 3d background impurity concentration of @xmath35 @xmath36 , dopant setback from the center of the quantum well @xmath37 nm , hole effective mass @xmath38 where @xmath39 is the free electron mass , quantum well width @xmath40 nm , and a remote ionized impurity concentration @xmath41 @xmath4 .
fig .
[ theory](a ) shows a crossover of the mobility exponent changing from @xmath42 to @xmath43 which has the same qualitative behavior as seen in the data in fig .
[ n_v_mu ] .
the results of the calculations are compared with the experimental data in fig .
[ theory](b ) .
qualitatively , as the hole density is lowered , screening of the remote dopants by the 2d hole gas becomes less effective , and the dominant scattering mechanism transitions from being dominated by uniform background impurities to being dominated by remote dopants .
such a transition point in the density is governed entirely by @xmath44 , @xmath45 , and @xmath46 , and thus it is not surprising that this is the first time ( to our knowledge ) that such a transition has been observed in a single sample within the high mobility regime . to understand the transition in the transport mechanism observed in our data , it is important to realize@xcite that the impurity scattering strength in the transport theory carries the form - factor @xmath47 at low temperatures where @xmath48 is the typical momentum transfer for resistive scattering by impurities , and @xmath45 is the typical separation of the impurities from the 2d carrier layer . since @xmath49 , it is clear that lowering the carrier density would lead to stronger scattering by remote impurities which is exponentially suppressed at higher values of @xmath50 . for a screened coulomb potential with two impurity contributions
( remote and background charged impurity ) we can derive the approximate qualitative formula for the mobility @xmath51 where @xmath52 is a density independent constant and @xmath53 is the thomas fermi screening wave vector with the effective bohr radius @xmath54 . in the high - density limit ,
@xmath55 the mobility is proportional to the square of the sum of two wave vectors , i.e. , @xmath56 .
however , as @xmath57 ( or , density ) decreases the mobility behaves approximately @xmath58 .
thus , as long as strong localization does not set in , which is the usual situation for lower mobility samples@xcite , lowering carrier density should always lead to a continuous increase of the exponent @xmath9 as scattering from the remote dopant impurities becomes important .
this is exactly the theoretical behavior predicted in the theory ( fig .
[ n_v_mu ] ) , and experimentally observed in our extremely high - mobility hole sample . in conclusion , we measured the density dependence of mobility in a very high quality 2dhs .
the 50 mk mobility was found to be 2.6 @xmath0 10@xmath59 @xmath17/vs at a density of 6.2 @xmath0 10@xmath3 @xmath4 in a pristine sample .
the mobility appears to be limited by background charged impurity scattering at high density but in the low density regime is a stronger function of the density indicating an increasingly important scattering contribution from remote impurities . from this data
, we can surmise that increased 2dhs mobility can be realized at low density by significantly increasing the spacer thickness .
such experiments are currently underway .
our work also demonstrates that in samples of sufficiently high quality , where the 2d mit transition is pushed down to very low carrier densities , the theoretically predicted continuous transition from background impurity scattering limited transport to remote dopant scattering limited transport can be quantitatively verified by decreasing the carrier density in a single sample . |
in relativistic mean field theories with scalar and vector potentials the dirac hamiltonian for a fermion ( nucleon ) of mass @xmath3 is @xmath4 where @xmath5 and @xmath6 are the usual dirac matrices .
we have introduced the sum " and the difference " potentials defined by @xmath7 and @xmath8 . when @xmath9 or @xmath10 , due to the matrix structure of @xmath11 , the dirac hamiltonian is invariant under a su(2 ) symmetry @xcite . this is a general feature which does not depend on the particular form of @xmath0 and @xmath1 . in this paper
we will consider only the case @xmath12 ( @xmath2 ) and a harmonic oscillator potential with angular frequency @xmath13 for the @xmath14 potential , _
i.e. _ , @xmath15 . a more generalized form of the relativistic harmonic oscillator where the case @xmath10 is also discussed has been presented recently @xcite . following the details of that paper
it is easy to prove that , for the case @xmath9 , the lower component of the dirac spinor satisfies the differential equation @xmath16f_\kappa(r)=0 .
\label{eq : d2ordfofr2}\ ] ] as shown in @xcite , @xmath17 is not the non - relativistic angular momentum @xmath18 but rather the pseudo - orbital angular momentum of the lower component of the dirac spinor .
they are related by @xmath19 , where the quantum number @xmath20 determines whether spins are parallel or antiparallel .
pseudospin symmetry is exact when doublets with @xmath21 are degenerate , where @xmath22 is the pseudospin quantum number .
the eigenenergies are given by @xmath23 where @xmath24 is the radial quantum number related to the nodes of the lower component @xmath25 . from eq .
( [ eq : enrohpapertilde ] ) one concludes that the real solutions must have positive binding energy @xmath26 .
the second order equation ( [ eq : d2ordfofr2 ] ) , which only depends on @xmath17 , and also the eigenenergy equation ( [ eq : enrohpapertilde ] ) , show that there is no pseudospin orbit coupling .
therefore the states with same @xmath27 , but with @xmath28 and @xmath29 are degenerate : they can be seen as pseudospin doublets as discussed in @xcite .
thus , when @xmath14 is a harmonic oscillator potential and @xmath9 , pseudospin symmetry is exact and there are only positive - energy bound states , _
i.e. _ , @xmath14 acts as a binding potential .
this is an interesting result in view of the fact that the pseudospin symmetry obtained in the limit @xmath30 can not be realized for nuclear vector and scalar mean - fields which go to zero as @xmath31 .
in fact , in that case , @xmath32 acts as a binding negative central potential well and therefore
no bound states may exist when @xmath9 @xcite .
the spectrum of single particle states for the case @xmath9 , @xmath33 and @xmath34 is shown in fig .
[ fig : spcposignl ] ( a ) using the quantum numbers of the upper component , that can be related to the non - relativistic quantum numbers . in fig .
[ fig : spcposignl ] ( b ) we classify the same energy levels by the quantum numbers of the lower components @xmath35 .
the comparison between these two figures exhibits the pseudospin symmetry and its quantum numbers .
for example , the doublets @xmath36 $ ] and @xmath37 $ ] , which have the same pseudo angular momentum @xmath17 and the same @xmath24 , are , in the new notation , @xmath38 $ ] and @xmath39 $ ] , respectively .
therefore , the harmonic oscillator with @xmath9 and @xmath40 provides an example of exact pseudospin symmetry .
[ fig : spcposignl ] ( a ) also shows the @xmath41 degeneracy , which means that not only the states with same @xmath42 are degenerate ( pseudospin partners ) but also , for example , @xmath43 or @xmath44 have the same energy .
the non - relativistic limit of the eigenvalue equation in eq.([eq : enrohpapertilde ] ) is reached when @xmath45 , which , in turn , means that @xmath46 , giving @xmath47 this equation , valid for @xmath9 , shows that the energy is of second order in the ratio @xmath48 , meaning that the energy is zero up to first order in @xmath48 .
we can interpret this fact by saying that , up to this order , there is no non - relativistic limit for @xmath9 and therefore the theory is intrinsically relativistic and so is the pseudospin symmetry .
when @xmath49 ( @xmath50 ) a pseudospin orbit term shows up in the equation for the lower component given by @xmath51 where the prime denote the derivative with respect to @xmath52 and @xmath20 is the quantum number referred to before .
so when @xmath53 the denominator , @xmath54 , can become very small near the pole and the contribution of this term can be quite significant , as shown in @xcite .
this manifests the non - perturbative breaking of this symmetry when @xmath55 , as it is the case for the nucleus @xcite .
however , we may overcome this problem if we keep @xmath56 but a tensor potential @xmath57 is introduced in the dirac equation . in this case
a new term appears in the spin - orbit term , @xmath58 , as shown in @xcite .
because of the product @xmath59 , the pole structure in the denominator discussed before is absent . with @xmath33 and @xmath60 as a function of @xmath61.,width=234 ] with @xmath33 , @xmath62 and @xmath60.,width=234 ] if we choose a linear form for @xmath63 , we have an additional quadratic central potential with a frequency @xmath64 in eq .
( [ eq : d2ordfofr2 ] ) , so that we still have a global harmonic oscillator central potential .
moreover , we generate a pseudospin orbit term @xmath65 .
thus , with this new term , the @xmath17 degeneracy ( pseudospin symmetry ) can be broken perturbatively in the sense that can be made so small as the @xmath64 frequency .
we show in fig .
[ fig : oh_04 ] how this symmetry is broken for the @xmath66 states , for positive @xmath64 , and present the eigenenergies in the spectroscopic notation @xmath67 , where states with @xmath68 are denoted by @xmath69 respectively .
interestingly , from fig . [
fig : oh_05 ] we see that with the breaking tensor potential there is a value of @xmath64 for which @xmath70 ( @xmath71 ) , a state with @xmath72 rather than @xmath73 , becomes the ground state . finally , the relativistic nature of the pseudospin symmetry with harmonic oscillator potentials shown here suggests that this symmetry may be more important in ultra - relativistic systems .
thus , this symmetry may play a role in high density matter , such as the matter inside compact stars .
r. l. and m. m. thanks the nice atmosphere during the iwara ( olinda / pe ) where this work has been presented .
r. l. and m. m. acknowledge support from cnpq .
de c. also thanks support cnpq and fapesp .
p.a . and m.f .
acknowledges the financial support from fct ( pocti ) , portugal .
0 j. n. ginocchio , _ phys .
lett . _ * 78 * , 436 ( 1997 ) .
p. alberto , m. fiolhais , m. malheiro , a. delfino and m. chiapparini , _ phys .
lett . _ * 86 * , 5015 ( 2001 ) ; p. alberto , m. fiolhais , m. malheiro , a. delfino and m. chiapparini , _ phys .
_ * c65 * , 034307 ( 2002 ) .
r. lisboa , m. malheiro and p. alberto , _ phys .
_ * c67 * , 054305 ( 2003 ) .
r. lisboa , m. malheiro , a. s. de castro , p. alberto and m. fiolhais , _ phys .
rev . _ * c69 * , 024319 ( 2004 ) .
s. marcos , l. n. savushkin , m. lpez - quelle and p. ring , _ phys .
rev . _ * c62 * , 054309 ( 2000 ) . | we show that relativistic mean fields theories with scalar , @xmath0 , and vector , @xmath1 , quadratic radial potentials can generate a harmonic oscillator with exact pseudospin symmetry _ and positive energy bound states _ when @xmath2 .
the eigenenergies are quite different from those of the non - relativistic harmonic oscillator .
we also discuss a mechanism for perturbatively breaking this symmetry by introducing a tensor potential .
our results shed light into the intrinsic relativistic nature of the pseudospin symmetry , which might be important in high density systems such as neutron stars . -0.5
cm recently , there has been much interest in understanding the nuclear pseudospin symmetry in terms of relativistic dynamics @xcite .
however , in the presence of nuclear scalar and vector mean - field potentials this symmetry is broken non - perturbatively as it has been discussed in recent articles @xcite . in this work
we will show that for quadratic radial scalar , @xmath0 , and vector , @xmath1 , potentials with @xmath2 is possible to break the pseudospin symmetry by introducing a tensor potential that preserves the form of the harmonic oscillator central potential but generates a pseudospin orbit term . |
there has been considerable evidence of high energy particle production during the fracture of certain kinds of crystals@xcite . in particular ,
fracture induced nuclear transmutations and the production of neutrons have been clearly observed@xcite . the production of neutrons appears greatly enhanced if the solids being fractured are piezoelectric@xcite materials .
our purpose is to describe theoretically the manner in which the mechanical pressure in a piezoelectric stressed solid about to fracture can organize the energy so that neutrons can be produced .
the nuclear physics involves a standard weak interaction wherein collective radiation plus an electron can be captured by a proton to produce a neutron plus a neutrino @xmath0 the required collective radiation energy may be produced by the mechanical elastic energy storage via the piezoelectric effect . by the _
definition _ of a _ piezoelectric material _ , the conversions of energy of the form @xmath1 are allowed . in terms of the electric field
@xmath2 and the crystal strain tensor @xmath3 , the precise definition of the piezoelectric tensor @xmath4 is discussed in sec.[pi ] . the final result may be expressed as the effective interaction hamiltonian @xmath5 wherein the tensor coefficients @xmath6 describe piezoelectricity as shown in fig .
some implications of the conversion from mechanical energy into electromagnetic energy are quite striking . for example , a piezoelectric ignition system can be constructed wherein a sharp mechanical impulse to a piezoelectric material can induce a sharp voltage spike across the sample with the resulting spark igniting a fire in a surrounding gas . more dramatically , the rocks crushed in earthquakes contain piezoelectric quartz .
the mechanical impulse causing micro - cracks in the rocks can thereby produce impulse earthquake lightning flashes . in sec.[fs
] we review the stresses and strains which accompany micro - cracks in rocks that are being fractured .
elasticity theories of such micro - cracks are well known@xcite .
the central result is as follows .
if @xmath7 denotes the elastic stress required to break the chemical bonds on an area of a micro - crack and @xmath8 denotes the surface tension of the free face of a crack , then the fracture stress @xmath9 required to create a crack of half length @xmath10 is given by @xmath11 for brittle fracture . in sec.[np ] , the manner in which the conversion of mechanical to electrical energy takes place is explored .
it is shown that copious electromagnetic energy is emitted in the radio frequency microwave regime .
the radiation accelerates the electrons allowing for nuclear transmutations in forms following from eq.([intro1 ] ) . in the concluding sec.[conc ] , the number of neutrons produced by rock fractures will be estimated .
the energy per unit volume @xmath12 of a piezoelectric material obeys @xmath13 wherein @xmath14 is the temperature , @xmath15 is the entropy , @xmath16 is the stress tensor , @xmath3 is the strain tensor , @xmath17 is the electric dipole moment per unit volume and @xmath2 is the electric field . the adiabatic piezoelectric tensor may be defined as @xmath18 to quadratic order , the mechanical electric field interaction energy @xmath19 follows from eq.([pi2 ] ) ; it is @xmath20 leading to the quantum operator in the effective hamiltonian and the feynman diagram of eq.([intro3 ] ) .
the adiabatic electric susceptibility of the material at constant strain is defined @xmath21 while the same susceptibility at constant stress is given by @xmath22 the elastic response tensor , @xmath23 determines the difference between between the two susceptibilities in eq.([pi4 ] ) and ( [ pi5 ] ) ; i.e. the thermodynamic identity is that @xmath24 for a complex frequency @xmath25 with @xmath26 , there are dynamical electric susceptibilities @xmath27 and @xmath28 . the dynamical version of eq.([pi7 ] ) is easily obtained .
phonon modes described by the dynamical phonon propagator @xmath29 affect the dynamical susceptibilities via @xmath30 eq.([pi7 ] ) is the zero frequency limit @xmath31 of eq.([pi8 ] ) .
the dynamical dielectric response tensor @xmath32 appears in the polarization part of the photon propagator@xcite .
the feynman diagrams contributing to the polarization part of the photon propagator in a piezoelectric system are shown in fig .
these are equivalent to eq.([pi8 ] ) and explain why mechanical acoustic frequencies appear in the electrical response of piezoelectric materials .
shown in fig . [ fig3 ] is a crystal under stress @xmath33 inducing a micro - crack of width @xmath34 and length @xmath35 .
the energy @xmath36 required to create a micro - crack of half width @xmath37 and length @xmath38 is given by@xcite @xmath39 b^2 , \label{fs1}\ ] ] wherein @xmath40 is the surface tension of the micro - crack interface , @xmath41 is the material young s modulus and @xmath42 is the poisson ratio .
the maximum of the elastic micro - crack energy per unit length ( @xmath43 ) represents the energy barrier to micro - crack creation . in detail , @xmath44 , \nonumber \\ u=2\gamma_s a = \frac{4\gamma_s ^2 } { \pi } \left[\frac{\cal e}{(1-\nu^2)\sigma_f ^2 } \right ] .
\label{fs2}\end{aligned}\ ] ] the stress level @xmath45 which nucleates a micro - crack is thereby the well known result@xcite @xmath46 the tensile strength @xmath45 of the material is then given by eq.([intro4 ] ) wherein the broken chemical bond strength @xmath47 is determined by young s modulus @xmath41 and the poisson ratio @xmath48 . employing the values of material constants for fused quartz
, we can estimate at least the powers of ten that would apply to piezoelectric rocks such as granite rocks .
the values are @xmath49 in satisfactory agreement with the elastic theory as reviewed in sec.[ts ] .
some comments are in order : ( i ) for quartz , the value of @xmath50 .
( ii ) for the brittle fracture of quartz , the macroscopic fracture surface experimentally exhibits micro - cracks with a length @xmath51 . (
iii ) as is usual in fractures @xmath52
, i.e. @xmath53 for the problem at hand .
( iv ) the velocity of sound @xmath54 compared with the velocity of light @xmath55 obeys @xmath56 .
the ratio of phonon frequencies to photon frequencies in cavities of similar length scales thereby obey @xmath57 the importance of the above eq.([ne2 ] ) is that the phonon modes enter into the dynamic dielectric response function @xmath58 in virtue of eq.([pi8 ] ) .
the neutron production rate at the fracture stress @xmath45 is here considered due to energetic electrons scattering off protons which are naturally present in ( say ) granite as water or organic molecules .
the feynman diagram in the fermi theory limit of the standard model is shown in fig .
[ fig4 ] described in eq.([intro1 ] ) . to begin to analyze the production of neutrons via the reaction eq.([intro1 ] ) , one must calculate the mean energy of electrons in condensed matter when accelerated by an electric field @xmath59 the electron energy is estimated to be @xmath60 if @xmath61 represents the the mean squared electric field strength in a bandwidth @xmath62 , then eq.([np1 ] ) , implies @xmath63 if @xmath64 denotes the dominant frequency in the spectrum of electric field fluctuations , then eq.([np3 ] ) is more simply written @xmath65 so that the ratio of the energy to the rest energy of the electron is @xmath66 the value of @xmath67 is critical for measuring whether or not there is sufficient radiation energy to allow for the reaction in eq.([intro1 ] ) . to estimate the electric field ,
one notes that the stress at fracture @xmath9 is in large part due to the electric field strength @xmath68 in virtue of eq.([ne1 ] ) .
since @xmath69 one finds @xmath70 the frequency of a sound mode localized on a micro - crack of width @xmath34 for a reasonable sound velocity in rock is in the microwave range @xmath71 one should then observe electromagnetic microwave emission when the sound mode is turned into an electromagnetic mode via the piezoelectric effect . in virtue of eqs.([np5
] ) , ( [ meaf3 ] ) and ( [ meaf4 ] ) one finds @xmath72 .
the threshold value of @xmath73 for eq.([intro1 ] ) to be possible without radiation is @xmath74 so that the energy renormalized by radiation is above threshold by a wide margin , i.e. @xmath75 . the electron energies on the surface of a micro - crack in a stressed environment with an external stress @xmath45 obey @xmath76 the transition rate per unit time for eq.([intro1 ] ) by the usual standard has been computed@xcite as @xmath77 the transition rate per unit time per unit area of micro - crack surfaces may be found from @xmath78 wherein @xmath79 is the number of protons per unit micro - crack area in the first few layers of the quartz granite .
typical values are @xmath80 if the fracture takes place with hydraulic fracture processes , then the neutron production rate will be about a factor of ten higher due to the higher water concentration on the micro - crack surface areas .
it is in the nature of piezoelectric matter that strong mechanical disturbances give rise to strong electromagnetic responses .
this is true for piezoelectric rocks such as granite which contain large amounts of quartz . for large scale piezoelectric rock fracturing ,
as takes place in earthquakes , electromagnetic responses in many frequencies , from radio frequaecy to gamma ray frequancy , are to be expected .
some have attributed earthquake lights and/or lightning@xcite to the phenomena discussed in this work .
we have employed the standard model of weak interactions along with the known theory of piezoelectric materials to explain the experimental evidence that fracturing piezoelectric rocks produces neutrons .
we have also explained why such fracturing processes produce microwave radiation .
the elastic energy micro - crack production ultimately yields the macroscopic fracture whose acoustic vibrations are converted into electromagnetic oscillations .
the electromagnetic microwaves accelerate the condensed matter electrons which then scatter from protons to produce neutrons and neutrinos .
this work also may have implications for a better understanding of radiative processes associated with earthquakes@xcite . | a theoretical explanation is provided for the experimental evidence that fracturing piezoelectric rocks produces neutrons .
the elastic energy micro - crack production ultimately yields the macroscopic fracture .
the mechanical energy is converted by the piezoelectric effect into electric field energy .
the electric field energy decays via radio frequency ( microwave ) electric field oscillations .
the radio frequency electric fields accelerate the condensed matter electrons which then collide with protons producing neutrons and neutrinos . |
the nucleon - nucleon interaction proceeds via the meson exchange , and therefore the existence of many - particle @xmath3 forces is a natural consequence of nonlinearity of the strong interaction theory .
indeed , only two - particle forces , fig.1a , are possible in the lowest order of linear theory containing only quadratic terms of the meson field @xmath4 in the meson lagrangian density .
the situation is different in nonlinear theories . in the theory with @xmath5 terms
the meson may turn into two ones thus giving rise to three - particle forces , fig.1b , in addition to the pairing ones . in the same way the lowest - order four - particle forces , fig.1c , appear in the theory with @xmath6 terms .
no higher - power terms exist in renormalizable theories , but the branching of meson is possible in higher orders thus giving rise to higher @xmath3 forces , fig.1d .
for this reason the empirical information on many - particle forces is of fundamental importance for the physics of strong interactions . at present the only such information is provided by the calculations for the few - nucleon systems . according to these calculations something
should be added to the two - particle @xmath3 forces to get the agreement between the calculated and observed quantities . in practice
only three - particle forces are added ( see ref.@xcite and the references therein ) since the problem is difficult even in this case .
indeed , the forces are characterized by a number of adjustable parameters describing the strength , range , the spin - isospin structure etc . , and
the observed quantities are expressed through these parameters in all orders of the perturbation theory . in such conditions
it is hardly possible to distinguish , for instance , between the genuine three - particle interaction and the combination of the latter with the four - particle one . as shown in the present paper such possibility may be provided by the experimental data on nuclear single - particle states .
it is based on the facts that ( a ) the contribution of the two - particle @xmath3 forces to the nuclear single - particle potential is fixed by experiment and ( b ) no correlation effects contribute to the single - particle energies @xcite . for these reasons
the many - particle contribution may be detected by comparing the observed energies of single - particle states with the calculations including the two - particle @xmath3 forces only .
we are using the untraditional approach by m.baranger @xcite .
it is based on the spectral representation for the single - particle propagator @xcite @xmath7 where @xmath8 is the ground - state wave function of nucleus @xmath9 which is chosen to be even - even the ground state thus being nondegenerate , @xmath10 the sums over @xmath11 and @xmath12 thus running over complete sets of states of @xmath13 and @xmath14 nuclei .
the short - time behaviour of a particle or hole which is suddenly created in the ground state @xmath8 is described by the following relations @xmath15 where @xmath16 , @xmath17 .
the single - particle states are treated as eigenstates of the single - particle hamiltonian @xmath18 , eq.(4 ) , @xmath19 thus being the doorway states for the single - nucleon transfer reactions@xcite . as a result of the correlation effects
these states are distributed over actual states of both the @xmath13 and @xmath14 nuclei .
however the properties of the distribution process permit the determination of the single - particle energies from the experimental data . introducing the spectroscopic factors @xmath20 multiplying eqs .
( 3)(5 ) by @xmath21 and integrating over @xmath22 and @xmath23 we get @xmath24 @xmath25 @xmath26 @xmath27 the determination of single - particle energies by using the eqs .
( 3a)(5a ) and ( 8) will be discussed in subsect.3.2 .
the ref.@xcite derivation of the explicit form of the single - particle hamiltonian @xmath28 should be specified by taking into account the meson - exchange nature of the @xmath3 interaction .
first , the contemporary meson - exchange forces such as the obe @xcite , paris @xcite and bonn @xcite have no hard repulsive core , and therefore the problem of elimination of short - range correlations ( see the discussion in ref.@xcite ) does not really exist .
so there are no actual reasons preventing from a direct use of bare @xmath3 forces for nuclear structure calculations ( i.e. the preliminary calculation of the brueckner @xmath29-matrix @xcite or some different way of the hard core elimination is not necessary ) .
second , the commutator technique of ref.@xcite does not apply in this case because of the retardation . for this reason the field - theoretical approach @xcite
should be used instead . in this approach the single - particle green function @xmath30 obeys the dyson equation @xmath31 where @xmath32 is the kinetic energy and the mass operator @xmath33 consists of the energy - independent part @xmath34 , which is just the nuclear single - particle potential , and the energy - dependent one @xmath35 which is responsible for the correlation effects : @xmath36 as shown in refs.@xcite the quantity @xmath35 vanishes in the @xmath37 limit @xmath38 ( the dots in the rhs are the higher - power terms of @xmath39 ) and therefore @xmath40 the decomposition ( 11 ) thus being unambiguous .
let us show that the single - particle hamiltonian of eq.(4 ) is @xmath41 whereas the quantity @xmath42 of eq.(5 ) is defined by eq.(12 ) .
first let us note that according to the time - energy heisenberg relation the infinite @xmath43 value is equivalent to the infinitely short time interval , the eqs.(13 ) and ( 14 ) thus meaning that the hamiltonian @xmath28 is indeed responsible for the short - time behaviour of the particle ( hole ) .
now let us use the spectral representation @xcite @xmath44 and the identity @xmath45 in the @xmath46 limit .
putting eq.(16 ) into eq.(15 ) we get the following asymptotic expansion @xmath47 the quantities @xmath48 and @xmath49 being just the lhs of eqs .
( 3),(4 ) and ( 5 ) respectively . as seen from eqs.(11),(12 ) and
( 14 ) the dyson equation ( 10 ) may be written in the form @xmath50 putting eq.(17 ) into eq.(18 ) we again get eqs.(3)(5 ) , but now the quantities @xmath28 and @xmath51 are determined . in terms of bare @xmath3 forces ( for a moment
let us take into account the pairing forces only ) the single - particle potential is defined by the first - order diagrams of fig.2 provided the nucleons interact without the retardation @xcite .
this is however not the case for the meson - exchange forces including both the momentum and the energy transfer . as a result of the latter the exchange diagrams of fig.2b have the @xmath52 asymptotics thus contributing to the quantity @xmath35 rather than @xmath34 @xcite .
let us illustrate this for the bonn @xmath53 potential @xcite .
it is the sum of the terms @xmath54 in the four - momentum space , the form of the meson - nucleon vertices being specified by the lorentz symmetry of the mesons .
the latter is however disregarded because it is irrelevant for the energy dependence . with this remark
the expression for the fig.2b diagram becomes @xmath55 considering the case of the monopole formfactor , @xmath56 , and using the identity @xmath57 we get @xmath58\right . \nonumber\\ & & \hspace{-1cm}-\left .
\left(1-(\lambda^2-\mu^2)\frac\partial{\partial\lambda^2 ( q)}\right)\frac1{2\omega_\lambda(q)}\left[\sum_j \frac{\psi_j(x)\psi^+_j(x')}{\varepsilon - e_j+\omega_\lambda(q ) } + \sum_k\frac{\psi_k(x)\psi^+_k(x')}{\varepsilon - e_k -\omega_\lambda(q)}\right]\right\ } .\end{aligned}\ ] ] the sign is also irrelevant because the bonn @xmath53 potential is the sum of terms with different signs . in the @xmath46 limit
we get @xmath59dq\ .\ ] ] the integral is convergent because the integrand has the @xmath60 asymptotics .
in addition to fig.2b the quantity @xmath35 includes an infinite sum of higher - order feynman diagrams describing all kinds of the correlation effects ( pauli , particle particle , particle hole , ground state etc . ) .
thus the only contribution to the nuclear single - particle potential is provided by the hartree diagrams of fig.2a and the hartree - like ones of fig.3 resulting from the many - particle @xmath3 forces , the single - particle energies thus being free of the correlation effects .
the latter is the feature of the ref.@xcite approach , thus permitting the model - independent studies of nuclear structure .
the fig.2a contribution to the single - particle potential is the convolution of the two - particle @xmath3 forces with the nucleon density distribution in nucleus .
the latter is determined by the combined analysis of electron - nucleus and 1 gev proton - nucleus elastic scattering data @xcite .
the two - particle forces are determined from the deuteron properties and the elastic @xmath3 scattering phase shifts at the energies below the pion production threshold . in this way the `` pairing '' contribution to nuclear single - particle potential is fixed by the experimental data thus being independent of any nuclear model . we have chosen the bonn @xmath53 potential @xcite for the two - particle forces , the motivation being discussed in sect.4 .
this choice enables us to check the status of nuclear relativity @xcite by calculating the scalar and vector fields in nuclear interior . using the parameters of table 5 from ref.@xcite and the equilibrium nuclear matter density value @xmath61 @xmath62 we get @xcite @xmath63 these values are close to those provided by the dirac phenomenology @xcite . in such conditions it is reasonable to treat the single - particle wave functions @xmath64 as dirac bispinors obeying the dirac equation .
the single - particle hamiltonian resulting from the bonn @xmath53 potential is @xcite @xmath65 @xmath66 is the free nucleon mass , @xmath67 and @xmath68 are dirac matrices .
the scalar field @xmath69 consists of the isoscalar and isovector parts resulting from the exchange by @xmath70 and @xmath71 mesons .
in addition to the isoscalar and isovector parts resulting from the exchange by @xmath72 and @xmath73 mesons the vector field @xmath74 includes the coulomb potential . the small isovector quantity @xmath75 results from the tensor part of the @xmath73 meson - nucleon coupling .
the details of the calculations are described in ref.@xcite .
they also may be found in the extensive literature on the walecka model @xcite , see for instance ref.@xcite .
the dirac equation @xmath76 is equivalent to schrdinger - like one for a particle with the effective mass @xmath77 in the central @xmath78 and spin - orbit potentials @xcite @xmath79 the terms arising from the quantity @xmath75 are omitted here , but actually they are taken into account in the calculations .
the quantities in eqs.(24)(28 ) should be supplied by the subscript `` pair '' since they describe the contribution from the two - particle forces .
the contribution from the many - particle forces is looked for as the following expansion in terms of the nucleon density distribution @xmath80 @xmath81\ ] ] the @xmath82 term results from three - particle forces , fig.3a , the @xmath83 one is of four - particle origin , fig.3b , etc .
( the linear density terms enter the two - particle contribution , eq.(27 ) ) .
the potential @xmath84 is assumed to be equally distributed between the scalar and vector fields , @xmath85 the only motivation for this assumption is the aim to have as little free parameters as possible . as mentioned in sect.2
the single - particle energies may be determined from experiment by using the sum rules ( 3a)(5a ) .
the most suitable situation is that for the cases when the absolute value of the single - particle energy @xmath86 exceeds the width @xmath87 of the distribution region . in these cases all states ,
over which the single - particle one is distributed , belong to the same nucleus , and therefore the sum rules ( 3a)(5a ) are saturated by only one term in the lhs , the first for the hole states and the second for the particle ones .
such situation occurs for the peaks in the cross sections of quasielastic knockout reactions @xmath88 and @xmath89 @xcite leading to the hole states with @xmath90 indeed , according to eq.(8 ) the width @xmath87 depends on the single - particle wave function @xmath91 rather than the energy @xmath92 , thus being roughly the same for all single - particle states .
so it is reasonable to identify @xmath70 with the largest observed value @xmath93 mev .
the latter is the width of the peaks corresponding to the @xmath94 hole states . for these reasons
the average energies of the above peaks obeying the eq.(31 ) condition may be identified with the single - particle energies within the experimental accuracy of @xmath95 mev . in this way we demonstrated that the ref.@xcite approach permits the model - independent determination of the single - particle energies .
it is worth mentioning that the experimental data for the @xmath96 energies are independent of those for the two - particle contribution to the @xmath96 potential .
we used the facts that the cross section of the quasielastic knockout reaction leading to the fixed nuclear state is proportional to the spectroscopic factor of this state , and the absolute values of the @xmath97-factors are not necessary when all states , over which the single - particle one is distributed , belong to the same nucleus ( in this case the relative values are sufficient ) .
this is however not the case for weakly bound single - particle states with @xmath98 .
such states are distributed over actual ones of both @xmath13 and @xmath14 nuclei , and therefore the @xmath97-factors from both the pickup and stripping reactions are necessary .
but the @xmath97-factors are determined with a rather low accuracy because of both experimental and theoretical ambiguities , and therefore it is unclear how to pin together the @xmath97-factors from the two reactions . for these reasons
the energies of weakly bound single - particle states are yet unknown .
one should also bear in mind that the low - lying states of @xmath99 nuclei forming the fermi - surface of the closed - shell nucleus @xmath9 are landau
migdal quasiparticles @xcite rather than the single - particle states of nucleon .
indeed , the correlation term @xmath35 is included in the quasiparticle energies in contrast to the single - particle ones , see ref.@xcite for details .
the observed energies of neutron and proton single - particle states in @xmath1zr are plotted in fig.4 together with the results of the calculations .
as seen from the figure the lowest `` pair '' states @xmath94 are significantly underbound whereas the higher states , especially the @xmath100 ones , are overbound .
the same compression of `` pair '' single - particle states occurs in @xmath2pb , fig.5 , where this effect is even more pronounced , and also in @xmath0ca , fig.6 .
this means that the potential well , including two - particle forces only , is somewhat too wide but insufficiently deep .
the isoscalar potentials in @xmath1zr , `` pair '' , actual and the many - particle contribution , are plotted in fig.7 . as seen from the figure , the many - particle contribution ( i.e. the difference between the actual and `` pair '' wells ) consists of the repulsive and attractive parts , the radius of the latter being less than that of the former .
so the expansion ( 29 ) has to contain at least two terms of different sign obeying the above condition . the most simple possibility is provided by the sum of the three - particle repulsion , @xmath101 , and the four - particle attraction , @xmath102 .
of course we can not guarantee the absence of contributions from higher many - particle forces .
it only should be mentioned that the above possibility corresponds to the least number of free parameters .
taking into account the possible contribution of many - particle forces to the isovector nuclear potential the quantity @xmath84 is chosen as @xmath103\rho^-(r)\bigg\}\ , \nonumber\\ \rho(r ) & = & \rho_n(r)+\rho_p(r)\ , \quad \rho^-(r)\ = \
\rho_n(r)-\rho_p(r)\ .
\end{aligned}\ ] ] @xmath104 and @xmath105 are neutron and proton density distributions in nucleus . as seen from fig.4 the description of single - particle energies
is improved by including the many - particle contribution .
the results , which are labelled as `` tot '' in fig.4 , are obtained using the woods saxon - like density distributions of ref.@xcite which are folded with the nucleon electromagnetic form factor , see ref.@xcite for details
. however the nuclear potential is expressed through the point densities since the finite size of nucleon is taken into account in the @xmath3 forces .
for this reason we used the point densities @xmath106 which are obtained from those of ref.@xcite by usual deconvolution procedure
. it should be mentioned that the electron and proton elastic scattering data underlying these densities are sensitive to nucleon density distributions in the surface region of nucleus , whereas the single - particle energies are sensitive to those in nuclear interior .
therefore the observed single - particle energies may be used to specify the nucleon density distributions .
we tried many different forms of @xmath80 .
the most appropriate one is found to be @xmath107 where @xmath108 is the fourth order hermite function .
the best fit parameters @xmath109 and @xmath110 of neutron and proton density distributions are shown in table 1 .
the best fit strength parameters are found to be the same for all nuclei .
they are @xmath111 the calculations with the specified densities are labelled as `` tot1 '' in figs .
neutron and proton density parameters [ cols=">,^,^,^,^",options="header " , ] slightly better agreement for @xmath2pb is provided by the following strength parameters @xmath112 with the following form of the nucleon density distributions @xmath113 the corresponding parameters @xmath109 and @xmath110 are those of the fourth row in table 1 , the results are labelled as `` tot2 '' in fig.5 .
the largest discrepancy between the observed single - particle energies and the `` tot1 '' results for @xmath1zr and @xmath0ca as well as the `` tot2 '' ones for @xmath2pb is less than 3 mev , the average discrepancy is 1.65 mev .
the specified nucleon density distributions are plotted in figs.8 and 9 .
as seen from the figures the neutron densities have a pronounced dip in the center of nucleus which increases with increasing mass number .
we also calculated the 1 gev proton elastic scattering cross sections within the glauber
sitenko theory @xcite using both the specified densities and those of ref.@xcite .
the results , see fig.10 , clearly show that the agreement with experiment is equally good for both the ref.@xcite and the specified densities .
the possible reasons for the difference between the eq.(34 ) and the eq.(35 ) strength parameters are as follows .
( i ) expression ( 32 ) corresponds to the zero - range forces @xmath114 whereas the actual forces may be of finite range .
( ii ) the above forces should have been folded with the two - particle density and the three - particle one rather than the products of single - particle densities .
( iii ) as mentioned above , the contributions from higher many - particle forces may be present .
the difference is however small for the charge - symmetric nuclear matter with the equilibrium density @xmath61 @xmath62 . indeed , as follows from eqs .
( 32),(34 ) and ( 35 ) the contributions of three(four)-particle forces to the isoscalar single - particle potentials are @xmath115 the first values in the rhs correspond to the eq.(34 ) parameters whereas those in parentheses refer to the eq.(35 ) set .
they are not small compared to the `` pair '' value @xmath116 mev , see eqs .
( 24 ) and ( 27 ) .
but they nearly compensate each other thus giving rise to the conclusion that the isoscalar part of the nuclear single - particle potential is mainly of the two - particle origin .
this conclusion is supported by the fact that according to eq.(30 ) @xmath117 thus being considerably less than the pair values @xmath118 mev , @xmath119 mev , see eq.(24 ) .
the situation is different for the isovector part of the potential . with @xmath120 eq.(32 )
gives @xmath121 for both eq.(34 ) and eq.(35 ) parameters whereas the `` pair '' value provided by the bonn @xmath53 potential is @xmath122mev , the isovector nuclear potential thus being mainly of many - particle origin .
the reason is due to the fact that the two - particle contribution arises from the exchange by isovector mesons @xmath73 and @xmath71 which are weakly coupled to nucleon @xcite .
in this way we demonstrated that the many - particle @xmath3 interaction includes at least the three - particle repulsion and the four - particle attraction . this result is restricted because no information is obtained for the many - particle forces which do not contribute to the nuclear single - particle potential .
but it provides a very instructive example of the situation where the experimental data on complex nuclei are more appropriate for the fundamental problem than those on few - nucleon systems .
indeed , the latter ones are described by solving the complicated many - particle quantum mechanical problem including the interaction in all orders of the perturbation theory . on the contrary
the single - particle states are solutions of the one - particle problem which is much more simple .
in addition the nuclear single - particle potential is expressed through the @xmath3 interaction in first order of the perturbation theory , the result thus being visual ( see fig.7 ) .
they are essentially based on the choice of the bonn potential for the two - particle @xmath3 forces .
the reason for this choice is a very high level of confidence : ( a ) the physics underlying the bonn potential is absolutely clear , ( b ) it contains only one adjustable parameter , ( c ) the two - nucleon data are described with @xmath123datum=1.9 @xcite ( according to this reference the bonn @xmath53 version we used in the present work is completely equivalent to the full one ) . of course our results
may be changed with the progress of the knowledge about the @xmath3 interaction .
we insist however that the acceptable new - fashioned potential must be of higher level of confidence than the bonn one .
this means that ( a ) the underlying physics should be as clear as that for the bonn potential , ( b ) the number of free parameters may be also only one but the @xmath124/datum value should be considerably less or ( c ) the @xmath124/datum value may be the same but no free parameters should be present . \2 . in the ref.@xcite experiments the energy of the knocked - out nucleon is only about 100 mev
. this may be insufficient to neglect the final - state inelastic interactions leading to additional excitation of the final nucleus . as a result of such excitations the average energies of the peaks
may be shifted from the single - particle energy values because the reaction mechanism is not a pure quasielastic knockout in this case .
to get rid of this ambiguity the additional quasielastic knockout @xmath125 or @xmath126 experiments are desired in which the energy of the knocked - out nucleon would be about 1 gev .
we hope that our work will stimulate such experiments .
: : two - particle ( a ) , three - particle ( b ) , four - particle ( c ) and higher ( d ) nucleon - nucleon interaction forces .
the wavy lines are for mesons , the full ones are for nucleons .
: : first - order hartree ( a ) and exchange ( b ) diagrams .
: : contributions of three - particle ( a ) and four - particle ( b ) forces to nuclear single - particle potential .
: : single - particle energies of neutron and proton states in @xmath1zr .
the energy scale is shown from the left of each figure .
the labels are @xmath127 for the observed energies , `` pair '' for the calculations taking into account the two - particle forces only , `` tot '' for those including the many - particle forces and using the density distributions from ref.@xcite , and `` tot 1 '' for the case when both the many - particle forces and specified densities are included .
: : the same for @xmath2pb .
the label `` tot 2 '' is for the case when the individual strength parameters for @xmath2pb are used .
: : the same for @xmath0ca .
: : isoscalar potential in @xmath1zr . the full , dashed and dot - dashed lines are for actual , `` pair '' and many - particle part of the potential respectively .
: : neutron density distributions in @xmath0ca , @xmath1zr and @xmath2pb . fig.9 .
: : the same for protons .
: : elastic scattering of 1 gev protons on @xmath0ca , @xmath1zr and @xmath2pb .
the calculations with the specified densities are shown by full line , those with the woods saxon - like ones from ref.@xcite are plotted by dashed line , the dots are for the experimental data . | as follows from the energies of single - particle states in @xmath0ca , @xmath1zr and @xmath2pb nuclei the contribution of many - particle @xmath3 forces to the nuclear single - particle potential is at least the sum of repulsive and attractive parts resulting from three - particle and four - particle forces respectively .
in addition the specified nucleon density distributions in the above nuclei are determined from both the 1 gev proton - nucleus elastic scattering and the single - particle energies .
-1 cm 0.25 cm .2 cm |
to obtain information about unresolved stellar populations in galaxies , one can look at colours , ranging from the uv to the infrared . such a method is popular especially when studying faint objects .
since colours are strongly affected by dust extinction , one prefers to use spectra when studying brighter objects .
these have the additional advantage that abundances of various elements can be studied , from the strength of absorption lines ( e.g. rose 1985 ; snchez - blazquez et al . 2003 ; carretero et al .
although the uv and the near ir are promising wavelength regions to study , the large majority of spectral studies is done in the optical , mainly because of our better understanding of the region . to obtain information ,
one generally compares model predictions with measurements of strong absorption line strengths .
this method , which is insensitive to the effects of dust extinction ( e.g. , macarthur 2005 ) very often uses the lick / ids system of indices , which comprises definitions of 25 absorption line strengths in the optical spectral range , originally defined on a low resolution stellar library ( fwhm@xmath18 - 11 ) ( gorgas et al . 1993 , worthey et al . 1994 , worthey & ottaviani 1997 ) .
there is an extensive list of publications in which this method has been applied , mainly for early - type galaxies ( see for example the compilation provided in trager et al .
1998 ) .
predictions for the line - strength indices of the integrated light of stellar clusters and galaxies are obtained with stellar population synthesis models .
these models calculate galaxy indices by adding the contributions of all possible stars , in proportions prescribed by stellar evolution models .
the required indices of all such stars are obtained from observed stellar libraries , but since the spectra in these libraries were often noisy , and did not contain all types of stars present in galaxies , people used the so - called fitting functions , which relate measured line - strength indices to the atmospheric parameters ( @xmath2 , @xmath3 , ) of library stars .
the most widely used fitting functions are those computed on the basis of the lick / ids stellar library ( burstein et al .
1984 ; gorgas et al .
1993 ; worthey et al .
1994 ) there are , however , alternative fitting functions in the optical range ( e.g. , buzzoni 1995 ; gorgas et al . 1999 ; tantalo , chiosi & piovan 2007 ; schiavon 2007 ) and in other spectral ranges ( e.g. , cenarro et al .
2002 ; mrmol - queralt et al .
2008 ) .
the most important parameters that can be obtained using stellar population synthesis are age ( as a proxy for the star formation history ) and metallicity ( the average metal content ) .
there are several methods to perform stellar population analysis based on these indices . by far
the most popular method is to build key diagnostic model grids by plotting an age - sensitive indicator ( e.g. h@xmath4 ) versus a metallicity - sensitive indicator ( e.g. , mg@xmath5 , @xmath6 ) , to estimate the age and metallicity ( e.g. trager et al .
2000 ; kuntschner et al .
there are other methods that , e.g. , employ as many lick indices as possible and simultaneous fit them in a @xmath7 sense ( e.g. , vazdekis et al .
1997 ; proctor et al . 2004 ) .
an alternative method is the use of principal component analysis ( e.g. covino et al . 1995 ; wild et al . 2009 ) . in the last decade
the appearance of a generation of stellar population models that predict full spectral energy distributions ( seds ) at moderately high spectral resolution has provided new means for improving the stellar population analysis ( e.g. , vazdekis 1999 , hereafter v99 ; bruzual & charlot 2003 ; le borgne et al .
these models employ newly developed extensive empirical stellar spectral libraries with flux - calibrated spectral response and good atmospheric parameter coverage . among the most popular stellar libraries
are those of jones ( 1999 ) , cat ( cenarro et al . 2001a ) , elodie ( prugniel & soubiran 2001 ) , stelib ( le borgne et al .
2003 ) , indo - us ( valdes et al .
2004 ) , and miles ( snchez - blzquez et al . 2006d , hereafter paper i ) .
models that employ such libraries are , e.g. , vazdekis ( 1999 ) , vazdekis et al .
( 2003 ) ( hereafter v03 ) , bruzual & charlot ( 2003 ) , le borgne et al .
( 2004 ) . alternatively
, theoretical stellar libraries at high spectral resolutions have also been developed for this purpose ( e.g. , murphy & meiksin 2004 ; zwitter et al .
2004 ; rodrguez - merino et al . 2005 ; munari et al . 2005 ; coelho et al . 2005 ; martins & coelho 2007 ) .
models that use such libraries are , e.g. , schiavon et al .
( 2000 ) , gonzlez - delgado et al .
( 2005 ) , coelho et al .
( 2007 ) .
there are important limitations inherent to the method that prevent us from easily disentangling relevant stellar population parameters .
the most important one is the well known age / metallicity degeneracy , which not only affects colours but also absorption line - strength indices ( e.g. worthey 1994 ) .
this degeneracy is partly due to the isochrones and partly due to the fact that line - strength indices change with both age and metallicity .
the effects of this degeneracy are stronger when low resolution indices are used , as the metallicity lines appear blended .
alternative indices that were thought to work at higher spectral resolutions , such as those of rose ( 1985 , 1994 ) , have been proposed to alleviate the problem .
new indices with greater abilities to lift the age / metallicity degeneracy have been proposed with the aid of the new full - sed models ( e.g. , vazdekis & arimoto 1999 ; bruzual & charlot 2003 ; cervantes & vazdekis 2009 ) .
these models allow us to analyze the whole information contained in the observed spectrum at once .
in fact , there is a growing body of full spectrum fitting algorithms that are being developed for constraining and recovering in part the star formation history ( e.g. , panter et al . 2003 ; cid fernandes et al .
2005 ; ocvirk et al . 2006ab ; koleva et al .
furthermore the use of these ssp seds , which have sufficiently high spectral resolution , has been shown to significantly improve the analysis of galaxy kinematics both in the optical ( e.g. , sarzi et al . 2006 ) and in the near - ir ( e.g. , falcn - barroso et al .
2003 ) , for both absorption and emission lines . here
we present single - age , single - metallicity stellar population ( ssp ) seds based on the empirical stellar spectral library miles that we presented in paper i ) and cenarro et al .
( 2007 , hereafter paper ii ) .
these models represent an extension of the v99 seds to the full optical spectral range .
these miles - models are meant to provide better predictions in the optical range for intermediate- and old - stellar populations . in paper
i we provided all relevant details for miles , which has been obtained at the 2.5 m issac newton telescope ( int ) at the observatorio del roque de los muchachos , la palma .
the library is composed of 985 stars covering the spectral range @xmath8 3540 - 7410 at 2.3 ( fwhm ) .
miles stars were specifically selected for population synthesis modeling . in paper
ii we present a homogenized compilation of the stellar parameters ( @xmath2 , @xmath3 , ) for the stars of this library .
in fact the parameter coverage of miles constitutes a significant improvement over previous stellar libraries and allows the models to safely extend the predictions to intermediate - aged stellar populations and to lower and higher metallicities .
the models that we present here are based on an empirical library , i.e. observed stellar spectra , and therefore the synthesized seds are imprinted with the chemical composition of the solar neighbourhood , which is the result of the star formation history experienced by our galaxy . as the stellar isochrones ( girardi et al .
i.e. the other main ingredient feeding the population synthesis code are solar - scaled , our models are self - consistent , and scaled - solar for solar metallicity . in the low metallicity regime , however , our models combine scaled - solar isochrones with stellar spectra that do not show this abundance ratio pattern ( e.g. , edvardsson et al . 1993 ; schiavon 2007 ) . in a second paper
we will present self - consistent models , both scaled - solar and @xmath9-enhanced , for a range in metallicities .
for this purpose we have used miles , together with theoretical model atmospheres , which are coupled to the appropriate stellar isochrones .
we note , however , that the use of our ( base ) models , which are not truly scaled - solar for low metallicities , does not affect in a significant manner the metallicity and age estimates obtained from galaxy spectra in this metallicity regime ( e.g. , michielsen et al .
2008 ) . for the high metallicity regime , where massive galaxies reside and show enhanced ratios , the observed line - strengths for various popular indices ( e.g. , mg@xmath5 , fe5270 ) are significantly different from the scaled - solar predictions .
it has been shown that by changing the relative element - abundances , and making them @xmath9-enhanced , both the stellar models ( e.g. , salaris , groenewegen & weiss 2000 ) and the stellar atmospheres ( e.g. tripicco & bell 1995 ; korn , maraston & thomas 2005 ; cohelo et al .
2007 ) are affected .
however it has also been shown that base models , such as the ones we are presenting here , can be used to obtain a good proxy for the abundance ratio , if appropriate indices are employed for the analysis ( e.g. yamada et al .
in fact we confirm the results of , e.g. , snchez - blzquez et al .
( 2006b ) , de la rosa et al .
( 2007 ) and michielsen et al .
( 2008 ) , who obtain a linear relation between the proxy for , obtained with scaled - solar models , and the abundance ratio estimated with the aid of models that specifically take into account non - solar element mixtures ( e.g. , tantalo et al .
1998 ; thomas et al . 2003 ; lee & worthey 2005 ; graves & schiavon 2008 ) . in section [ sec : models ] we describe the main model ingredients , the details of the stellar spectral library miles and the steps that have been followed for its implementation in the stellar population models . in section
[ sec : miles_ssp_seds ] we present the new single - age , single - metallicity spectral energy distributions , ssp seds , and provide a quantitative analysis for assessing their quality . in section
[ sec : system ] we show line - strengths indices as measured on the newly synthesized ssp spectra and propose a new reference system , as an alternative to the lick / ids system , to work with these indices . in section [ sec : colours ] we provide a discussion about the colours measured on the new ssp seds . in section [ sec : applications ] we apply these models to a set of representative stellar clusters and galaxies to illustrate the capability of the new models . in section
[ sec : web ] we present a webpage to facilitate the use and handling of our models . finally in section
[ sec : conclusions ] we provide a summary .
the ssp seds presented here represent an extension to the full optical spectral range of the v99 models , as updated in v03 .
we briefly summarize here the main ingredients and the relevant aspects of this code .
we use the solar - scaled theoretical isochrones of girardi et al .
a wide range of ages and metallicities are covered , including the latest stages of stellar evolution .
a synthetic prescription is used to include the thermally pulsing agb regime to the point of complete envelope ejection .
the isochrones are computed for six metallicities @xmath100.0004 , 0.001 , 0.004 , 0.008 , 0.019 and 0.03 , where 0.019 represents the solar value .
in addition we include an updated version of the models published in ( girardi et al .
1996 ) for @xmath100.0001 .
the latter calculations are now compatible with girardi et al .
the range of initial stellar masses extends from 0.15 to 7@xmath11 .
the input physics of these models has been updated with respect to bertelli et al .
( 1994 ) with an improved version of the equation of state , the opacities of alexander & ferguson ( 1994 ) ( which results in a rgb that is slightly hotter than in bertelli et al . ) and a milder convective overshoot scheme . a helium fraction was adopted according to the relation : @xmath12 . for the tp - agb phase girardi
( 2000 ) adopt a simple synthetic prescription that , for example , does not take into account the third dredge - up .
an improved treatment for this difficult stellar evolutionary phase has been recently included by marigo et al .
the effects of such improvements become relevant for the near - ir spectral range ( see e.g. , marigo et al .
2008 ; maraston 2005 ) .
the theoretical parameters of the isochrones @xmath2 , @xmath3 , are transformed to the observational plane by means of empirical relations between colours and stellar parameters ( temperature , metallicity and gravity ) , instead of using theoretical stellar atmospheres .
we mostly use the metallicity - dependent empirical relations of alonso , arribas & martnez - roger ( 1996 , 1999 ; respectively , for dwarfs and giants ) .
each of these libraries are composed of @xmath13 500 stars and the obtained temperature scales are based on the ir - flux method , i.e. only marginally dependent on model atmospheres .
we use the empirical compilation of lejeune , cuisinier & buser ( 1997 , 1998 ) ( and references therein ) for the coolest dwarfs ( @xmath14 and giants ( @xmath15 ) for solar metallicity . for these low temperatures we use a semi - empirical approach to other metallicities on the basis of these relations and the model atmospheres of bessell et al .
( 1989 , 1991 ) and the library of fluks et al .
( 1994 ) . the empirical compilation of lejeune et al .
( 1997,1998 ) was also used for stars with temperatures above @xmath138000k .
we apply the metal - dependent bolometric corrections given by alonso , arribas & martnez - roger ( 1995 , 1999 ; respectively , for dwarfs and giants ) . for the sun
we adopt @xmath16 and a bolometric magnitude of 4.70 .
we have computed our predictions for several imfs : the two power - law imfs described in vazdekis et al .
( 1996 ) ( i.e unimodal and bimodal ) , which are characterized by its slope @xmath17 as a free parameter , and the multi - part power - law imfs of kroupa ( 2001 ) ( i.e. universal and revised ) . in our notation the salpeter ( 1955 ) imf corresponds to a unimodal imf with @xmath18 .
further details of the imf definitions are given in appendix a of v03 .
we set the lower and upper mass - cutoff of the imf to 0.1 and 100m@xmath19 , respectively .
to compute the ssp seds in the full optical spectral range we use the miles library , for which the main characteristics are given in paper i. the stars of the library were selected to optimize the stellar parameter coverage required for population synthesis modelling .
this is , in fact , one of the major advantages of this library . another important feature of the library is that the stellar spectra were carefully flux - calibrated . to reach this goal ,
all the stars were observed through a wide ( 6@xmath20 ) slit to avoid selective flux losses due to the differential refraction effect , in addition to the higher resolution setups used to achieve the blue and red parts of the stellar spectra .
telluric absorptions that were present in the redder part of the spectra were properly removed .
we refer to paper i for the technical details on how these aspects were tackled . to prepare this library for its implementation in the models we identified those stars
whose spectra might not be properly representative of a given set of atmospheric parameters .
we used the simbad database for identifying the spectroscopic binaries in the miles sample .
a number of these stars were removed from the original sample as they were found non essential as miles contains sufficient stars with similar atmospheric parameters .
we also checked for those stars with high signal of variability ( @xmath21 mag ) , according to the combined general catalogue of variable stars of kholopov et al .
( 1998 ) ( the electronically - readable version provided at cds ) . some stars with signs of variability
were removed , when no such high variability was expected according to their spectral types and atmospheric parameters and , at the same time , we were able to identify alternative stars in the library with similar parameters and no such sign of variability .
various stars were discarded because emission lines were detected in their spectra .
for various technical reasons we also discarded from the original miles list an additional number of stars : those with very low signal - to - noise in the blue part of the spectrum or for which the atmospheric parameters were lacking or had problems in the continuum .
each stellar spectrum of this selected subsample of stars was compared to a synthetic spectrum of similar atmospheric parameters computed with the algorithm described in v03 ( appendix b ) ( see also [ sec : synthesis ] ) employing the miles stellar library . to perform this test we first excluded from the list the target star for which we wished to synthesize a similar spectrum .
we then compared the synthesized stellar spectrum with the observed one .
this comparison was found to be very useful for identifying stellar spectra with possible problems and , eventually , discard them . in case of doubt , particularly for those regions of parameter space with poorer coverage , we compared the observed spectrum of our star with the spectra of those stars , with the closest atmospheric parameters , which were selected by the algorithm to synthesize our target star .
as an example of this method , fig .
[ fig : interpolator ] shows the results for stars of varying temperature , gravity and metallicity .
according to these careful inspections and tests we marked 135 stars of the miles database . while 60 of these stars were removed from the original sample , we kept 75 stars , which were found to be useful for improving the coverage of certain regions of parameter space .
for the latter , however , we decreased the weight with which they contribute when we synthesize a stellar spectrum for a given set of atmospheric parameters . in practise
this is performed by artificially decreasing their signal - to - noise , according to the prescription adopted in [ sec : synthesis ] . for the interested reader we provide explanatory notes for these stars in an updated table of the miles sample that can be found in * http://miles.iac.es*. figure[fig : miles - parameters ] shows the parameter coverage of the selected subsample of miles stars for dwarfs and giants ( separated at @xmath22 ) compared with the original miles sample .
the 75 stars with decreased weights are also indicated .
the parameters of miles plotted here are the result of an extensive compilation from the literature , homogenized by taking as a reference the stars in common with soubiran , katz & cayrel ( 1998 ) , which have very well determined atmospheric parameters .
we refer the reader to paper ii for an extensive description of the method and the adopted parameters . for comparison
we show the parameter coverage of some popular stellar libraries ( stelib : leborgne et al .
2003 ; lick / ids : gorgas et al .
1993 , worthey et al . 1994 ; indo - us : valdes et al 2004 ; and elodie 3.1 : prugniel et al .
for elodie 3.1 we use the parameters determined with the tgmet software ( see prugniel & soubiran 2004 for details ) to maximize the number of plotted stars .
we adopt the mean value for those stars with repeated observations .
[ fig : miles - parameters ] shows that , at solar metallicity , all types of stars are well represented in all libraries . at lower metallicities , however , miles shows a much better coverage of the parameter space for both , dwarfs and giants , than other libraries . particularly relevant
is the presence of dwarfs with temperatures above @xmath23 and metallicities .
with these stars we are in position to overcome a major limitation of current population synthesis models based on empirical stellar libraries providing with predictions for metal - poor stellar populations in the age range 0.1 5gyr . furthermore , there are sufficient stars to compute models for metallicities as low as , particularly for very old stellar populations where the temperature of the turnoff is lower .
another advantage with respect to the other libraries is the coverage of metal - rich dwarf and giant stars that allow us to safely compute ssp seds for .
also for even higher metallicities miles shows a good stellar parameter coverage and , therefore , it is possible to compute seds for .
these predictions will be presented and we will make them available in our website in a future paper , where we feed the models with the teramo isochrones ( pietrinferni et al .
2004 ) , that reach these high metallicity values . these supersolar metallicity predictions are particularly relevant for studying massive galaxies and it is the first time they can be safely computed , without extrapolating the behaviour of the spectral characteristics of the stars at higher metallicities as it is done , e.g. , in models based on the lick / ids empirical fitting functions ( e.g. worthey 1994 ; thomas et al .
finally , it is worth noting that these new model predictions also extend the range of ssp metallicities and ages of our previous model seds based on the jones ( 1999 ) ( v99 ) and cenarro et al .
( 2001a ) ( v03 ) stellar libraries .
a quantitative analysis of the quality of the models based on the atmospheric parameter coverage of the employed libraries is provided in section [ sec : quality ] .
we use the method described in v99 and v03 for computing the ssp seds on the basis of the selected subsample of miles stars . in short ,
we integrate the spectra of the stars along the isochrone taking into account their number per mass bin according to the adopted imf . for this purpose ,
each requested stellar spectrum is normalized to the corresponding flux in the v band , following the prescriptions adopted in our code , which are based on the photometric libraries described in
[ sec : ingredients ] .
we refer the interested reader to the papers previously mentioned for a full description of the method .
the ssp sed , @xmath24})$ ] , is calculated as follows : @xmath25})&=&\int_{m_{\rm l}}^{m_{\rm t } } s_{\lambda}(m , t,{\rm [ m / h]})n(m , t)\times \nonumber \\ & & { f_v}(m , t,{\rm [ m / h]})dm , \label{eq : ssp}\end{aligned}\ ] ] where @xmath26})$ ] is the empirical spectrum , normalized in the v band , corresponding to a star of mass @xmath27 and metallicity [ m / h ] , which is alive at the age assumed for the stellar population @xmath28 .
@xmath29 is the number of this type of star , which depends on the adopted imf . @xmath30 and @xmath31 are the stars with the smallest and largest stellar masses , respectively , which are alive in the ssp .
the upper mass limit depends on the age of the stellar population . finally , @xmath32})$ ] is its flux in the v band , which comes from transforming the theoretical parameters of the isochrones .
it is worth noting that the latter is performed on the basis of empirical photometric stellar libraries , as described in section [ sec : ingredients ] , rather than relying on theoretical stellar atmospheres , as it is usually done in other stellar population synthesis codes .
this avoid errors coming from theoretical uncertainties as the incorrect treatment of convection , turbulence , non - lte effects , incorrect or imcomplete line list , etc ( see worthey & lee 2006 ) .
we obtain the spectrum for each requested star with a given atmospheric parameters interporlating the spectra of adjacent stars using the algorithm described in v03 .
the code identifies the miles stars whose parameters are enclosed within a given box around the point in parameter space ( @xmath33 , @xmath34 , [ m / h]@xmath35 ) . when needed , the box is enlarged in the appropriate directions until suitable stars are found ( for example , less and more metal - rich ) .
this is done by dividing the original box in 8 cubes , all with one corner at that point .
this reduces the errors in case of gaps and asymmetries in the distribution of stars around the point .
the larger the density of stars around the requested point , the smaller the box is .
the sizes of the smallest boxes are determined by the typical uncertainties in the determination of the parameters ( cenarro et al .
2001b ) . in each of the boxes ,
the stars are combined taking into account their parameters and the signal - to - noise of their spectra .
finally , the combined spectra in the different boxes are used to obtain a spectra with the required atmospheric parameters by weighting for the appropiate quantity . for a full description of the algorithm
we refer the reader to v03 ( appendix b ) .
it is worth noting that , within this scheme , a stellar spectrum is computed according to the requested atmospheric parameters , irrespective of the evolutionary stage .
we are aware that this approach is not fully appropriate for the tp - agb phase , where o - rich , c - rich and stars in the superwind phase are present .
in fact , in mrmol - queralt et al .
( 2008 ) we have found non negligible differences in the co bandhead at 2.3@xmath36 between 19 agb stars , all from the miles library , and rgb stars of similar parameters .
this , in principle , could introduce an error in our predictions as very few c - rich , o - rich or superwind phase tp - agb stars can be found in miles .
however , the contribution of such stars to the total flux budget in the v - band is just a few percent , mainly for stellar populations of intermediate - ages ( 0.1 1.5gyr ; see e.g. , bruzual 2007 ) .
percival & salaris ( 2009 ) have recently shown that systematic uncertainties associated with the three fundamental stellar atmospheric parameters might have a non negligible impact on the resulting ssp line - strength indices .
the interested reader is refered to that paper for details and for a suite of tests showing the effects of varying these parameters .
in particular , a relatively small offset in the effective temperature of 50 - 100 k , which is of the order of the systematic errors in the conversion from temperature to colours used here ( alonso et al . 1996 ) may change the age of a 14gyr stellar population by 2 - 3 gyr and alliviate the so - called zero - point problem for which the ages of the globular clusters ( gcs ) are older than the most recent estimations of the age of the universe .
furthermore , they show that , in many cases , there is a mistmatch in scales between the underlying models ( the isochrones ) and the adopted stellar library in the stellar population models .
we showed already in paper ii that the alonso et al .
( 1996 , 1999 ) photometric library did not show any offset with the miles atmospheric parameters . therefore , there is not a missmatch in scales between the models and the stellar library in the models presented here .
however , inspired by this work we decided to compare our stellar parameters with those from several recently published works , e.g. , gonzlez - hernndez & bonifacio ( 2009 ) , casagrande et al .
( 2006 ) , ramrez & melndez ( 2005 ) , to see if our parameters should be corrected .
we show here , for ilustrative purposes , the comparison of our parameters with those using the new temperature scale of gonzlez - hernndez & bonifacio ( 2009 ) , whom derived teff using the infrared flux method ( blackwell et al .
1990 ) , as it is the work with a larger number of stars in common with miles ( 215 ) and because it is the one showing larger differences . in fig.[fig :
gonzalezbonifacio ] we plot the difference in temperature for the dwarf and giant stars in common between gonzlez - hernndez & bonifacio ( 2009 ) and miles as a function of temperature and metallicity .
we performed linear fits weighted with the errors obtaining the following relations : @xmath37 and metallicity @xmath38 } } \label{eq : hbfeh}\ ] ] obtaining an rms of 138k in the two cases .
the mean temperature offset of 59k and 54k for dwarfs and giants , respectively .
this is in good agreement with the offsets obtained by gonzlez - hernndez & bonifacio ( 2009 ) when comparing their temperature scale with that of alonso et al .
( 1996 ) and alonso et al .
finally we did not find any significant offset in metallicity or gravity among these samples .
we computed an alternative ssp model sed library by transforming the miles temperatures to match this scale .
these models can be used to better assess the uncertainties involved in the method .
we comment on these models in sections [ sec : colours ] and [ sec : gonzalezbonifaciossps ] .
however , we do not find any strong reason for adopting the gonzlez - hernndez & bonifacio ( 2009 ) since the accuracy at high metallicities is worse than in the alonso et al .
( 1996,1999 ) ( the temperature scale of gonzlez - hernndez & bonifacio ( 2009 ) has been optimized for low metallicities while the errors in the temperature determination for metal rich giant stars are very large ) .
furthermore , as we already mentioned above , the temperature scale of the isochrones and stellar libraries in our models do not show any offset .
lc + spectral range & @xmath8 3540.5 - 7409.6 + spectral resolution & fwhm @xmath39 , @xmath40 km s@xmath41 + linear dispersion & 0.9 / pix ( 51.5 km s@xmath41 ) + continuum shape & flux - scaled + telluric absorption residuals & fully cleaned + units & f@xmath42/l@xmath43@xmath41m@xmath44 , l@xmath45 + + imf type & unimodal , bimodal , kroupa universal , kroupa revised + imf slope ( for unimodal and bimodal ) & 0.3 3.3 + stellar mass range & 0.1 100m@xmath43 + metallicity & @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath50 , @xmath51 , @xmath52 + age ( ) & @xmath53gyr ( only for imf slopes @xmath54 1.8 ) + age ( ) & @xmath55gyr + age ( ) & @xmath56gyr + table [ tab : sedproperties ] summarizes the spectral properties of the newly synthesized ssp seds .
the nominal resolution of the models is [email protected] , which is almost constant along the spectral range . for applications requiring to know the resolution with much greater accuracy , we refer the reader to fig . 4 of paper i. this resolution is poorer than that of the models we presented in v99 ( fwhm=1.8 ) , but this is compensated by the large spectral range @xmath8 3540.5 - 7409.6 , with flux - calibrated response and no telluric residuals . for computing the seds we have adopted a total initial mass of 1m@xmath43 .
table [ tab : sedproperties ] also summarizes the ssp parameters for which our predictions can be safely used .
we discuss this issue in section [ sec : quality ] . in fig .
[ fig : ssp_spectra ] we show the new spectral library of ssp models for different ages and metallicities .
all the spectra are plotted at the nominal resolution of the models ( fwhm=2.3 ) .
the ssp age - sequence shows weaker 4000 break and stronger balmer line strengths as the age decreases . the largest balmer line strengths ,
however , are reached for ages of @xmath130.4 gyr . on the other hand
the metallic absorption lines become more prominent with increasing age .
it can be seen that the continuum of the models with larger ages is heavily lowered by the strengthening of the metallic features ( line blanketing ) .
this is the reason why the lick / ids indices were defined using psedocontinua instead of real continua ( worthey et al .
1994 ) . in the ssp
sed metallicity - sequence it can be seen that balmer line - strengths decrease , and metallicity - sensitive features get stronger for increasing metallicity .
these two ssp sed sequences clearly illustrate the effects of the age / metallicity degeneracy . in fig .
[ fig : ssp_spectra_imf ] we illustrate the effects of varying the imf slope ( @xmath17 ) and the imf shape .
the left panel shows a sequence of ssp spectra of solar metallicity and 10gyr , where the slope is varied for the unimodal imf .
for old stellar populations this imf type shows the largest variations as the fraction of low mass stars vary dramatically when varying the imf slope .
we see that the reddest part of the spectrum shows the largest sensitivity to this parameter .
these variations do no evolve linearly as a function of @xmath17 .
in fact the ssp seds corresponding to the flatter imfs look rather similar . on the other hand ,
the effects caused by varying @xmath17 quickly become stronger for the steepest imfs , notably redward of @xmath135500 , including the na line at @xmath135800 and the tio molecular band around @xmath136200 and @xmath136800 .
the right panel shows that the effects of varying the imf slope are much less significant for the bimodal imf .
this is because for this imf type the contribution from stars with masses lower than 0.6m@xmath43 is decreased . since this imf more closely resembles the kroupa imfs , it is probably more realistic
. finally , we note that the ssp seds for the two kroupa ( 2001 ) imfs and the bimodal imf with slope 1.3 ( as well as the salpeter imf ) do not show significant differences . in this section
we try to provide some quantitative way of estimating the reliability and the quality of the synthesized ssp seds , and to estimate the size of the errors due to an incomplete coverage of the stellar parameter space of the input library .
we perform this evaluation as a function of ssp age , metallicity and imf . to do this , we have taken advantage of our algorithm for synthesizing a representative stellar spectrum for a given set of atmospheric parameters ( see section [ sec : synthesis ] ) to compute for each ssp a parameter , @xmath57 , as follows @xmath58}}_i^j}{\sigma_{\rm [ m / h]_{\rm m}}}\right)^2}}}}\right ] { n_i { f_v}_i } } { \sum_{i=1}^{n_{m_{\rm t}}}{n_i { f_v}_i } } \label{eq : q}\ ] ] we integrate the term enclosed within the square - brackets along the isochrone , from the smallest stellar mass , @xmath30 , to the most massive star , @xmath31 , which is alive at the age of the stellar population @xmath28 , weigthing by the flux in the v band , @xmath59 , and the number of stars , @xmath60 , according to the adopted imf .
@xmath61 is the number of stars that have been found within cube @xmath62 in parameter space ( where @xmath63 , see section [ sec : synthesis ] ) . for a star with given @xmath64 , @xmath3 and [ m / h ] , the size of these cubes
is given by @xmath65,@xmath66 and @xmath67}}_i^j$ ] .
@xmath68 , @xmath69 and @xmath70}_{\rm m}}$ ] is the minimum uncertainty in the determination of @xmath64 , @xmath3 and , respectively .
following v03 we adopt @xmath71 , @xmath72 and @xmath73}}_{\rm m } = 0.09 $ ] .
the size of the smallest cube is obtained by multiplying the minimum uncertainty by the factor @xmath74 .
therefore , according to the square - bracketed term , for each star along the isochrone , the larger the total number of stars found in these 8 cubes and the smaller the total volume , the larger the quality of the synthesized stellar spectrum is .
we also compute , for each ssp , the minimum acceptable value for this parameter , @xmath75 .
our algorithm ensures that at least we find stars in 3 ( out of 8) cubes to synthesize a stellar spectrum ( see appendix b of v03 for details ) .
furthermore the maximum enlargement that is allowed for each cube is assumed to be 1/10 of the total range covered by the stellar parameters in our library
. therefore the sizes for the largest cubes are @xmath76 , @xmath77 and @xmath73}}_{\rm m } = 0.41 $ ] .
note that for @xmath78 we also assume the condition that @xmath79 , which corresponds to the adopted @xmath78 limiting values when transforming to the @xmath2 scale . therefore to compute @xmath75 we will assume for each stellar spectrum along the isochrone the following three conditions that must be satisfied : _
i ) _ we find stars in three cubes , _ ii ) _ there is a single star for each of these three cubes , which implies that @xmath80=3 , and _
iii ) _ it is necessary to open the size of these cubes , generically , @xmath81 , @xmath820.34 , i.e. the star is found to be located within 1@xmath83 of @xmath84 , considering the two directions at either side of the requested parametric point . by substituting these values in eq.[eq : q ] we obtain @xmath85}}_m}{\sigma_{\rm [ m / h]_{\rm m}}}\right)^2 } } } } { \sum_{i=1}^{n_{m_{\rm t}}}{n_i { f_v}_i } } \label{eq : q_m}\ ] ] then a quality measure of the ssp sed , due to the atmospheric parameter coverage of the stellar library feeding the models , is given by the normalized parameter @xmath86 although the absolute value of @xmath87 depends on the assumptions adopted for computing @xmath75 , the @xmath87 parameter allows us to compare the quality of a given ssp sed compared to the seds computed for other ssp parameters .
in addition , @xmath87 allows us to compare the quality of ssp seds synthesized on the basis of different stellar spectral libraries , given the fact that the stellar spectra of these libraries are of sufficiently high quality .
furthermore these prescriptions are the same that we use for computing a stellar spectrum of given atmospheric parameters , which have been extensively tested in v03 . figure [ fig : quality_miles ] shows the value of @xmath87 as a function of the ssp age ( in gyr ) for different metallicities ( different line types ) and adopting a unimodal imf , with its slope increasing from the left to the right panel .
ssp seds with @xmath87 values above 1 can be considered of sufficient quality , and therefore they can be safely used .
the panels show that , as expected , the higher quality is achieved for solar metallicity . for the salpeter imf , i.e. the second panel , @xmath87 reaches a value of @xmath135 for stellar populations in the range 1 10gyr .
outside this age range the quality starts to decrease as low - mass and hotter ms stars are less numerous in the miles database .
a similar behaviour is found in fig .
[ fig : quality_miles ] for the ssps of metallicities , @xmath50 and @xmath49 , all with @xmath87 values in the range 3 4 .
despite the fact that for all these metallicities the value of @xmath87 sharply decreases toward younger stellar populations , the computed ssp seds can be considered safe . for lower metallicities
we obtain significantly lower @xmath87 values , which gradually decrease toward younger ages . for old stellar populations we obtain @xmath87 values around 2 and around 1.5 for and , respectively .
interestingly , according to our quantitative analysis , the ssp seds synthesized for and ages above 10gyr are barely acceptable for imfs with slopes @xmath88 and @xmath18 .
these seds are particularly useful for globular cluster studies .
note the sharp increase of @xmath87 observed for ages around 1gyr .
this should be mostly attributed to the fact that for stars of temperatures larger than 9000k we do not take into account the metallicity of the available stars for computing a stellar spectrum of given atmospheric parameters ( see v03 for details ) .
therefore , the quality of the seds synthesized for the younger stellar populations is lower than stated by @xmath87 .
[ fig : quality_miles ] also shows that @xmath87 decreases slightly with increasing imf slope .
in fact , the largest values are obtained for the flatter imf ( @xmath890.3 ) , as a result of the smaller contribution of low mass stars , which are less abundant in miles .
this explains why @xmath87 drops significantly for the steepest imf .
it is worth noting that the ssp seds that are synthesized for the lowest metallicity can not be considered safe for imfs steeper than salpeter .
but for ssp seds computed on the basis of the atmospheric parameters corresponding to the lick ( left panel ) and stelib ( right panel ) stellar spectral libraries ( see the text for details ) .
to perform this test we selected the miles spectra for the stars in common with miles , whereas for the rest of the stars we synthesized their spectra with our interpolating algorithm , using the miles spectral database ( see the text for details).,width=307 ] for comparison we show in fig .
[ fig : quality_lick_stelib ] the @xmath87 values obtained for two of the most popular libraries that are used in stellar population synthesis , i.e. lick and stelib . the first library is mainly used to predict lick / ids line indices via the empirical fitting functions of worthey et al .
( 1994 ) ( e.g. , worthey 1994 ; vazdekis et al .
1996 ; thomas et al .
stelib is the library feeding the models of bruzual & charlot ( 2003 ) . for this purpose we used miles to assemble a stellar spectral library resembling lick , and another one resembling stelib . for each of these libraries we selected the miles spectra for the stars in common with miles with listed stellar parameters ( i.e. 213 for lick and 117 for stelib ) . for the remaining stars , which are not available in miles , we synthesized their spectra with our interpolating algorithm of section [ sec : synthesis ] , using miles .
the two panels of fig .
[ fig : quality_lick_stelib ] show that the obtained @xmath87 values are significantly lower than the ones obtained for miles for all metallicities and ages .
the left panel shows that the ssp predictions based on the lick library are not reliable for metallicities lower than @xmath901.3 .
the right panel shows that this limit occurs at higher metallicity ( @xmath900.7 ) for any prediction based on stelib .
overall the stelib ssp models are of much lower quality than the predictions based on the lick library , judging from the parameter coverage of these libraries .
note that there is a large dispersion in the @xmath87 values obtained for old stellar populations .
this feature is less prominent in fig .
[ fig : quality_miles ] , due to the fact that the distribution of stellar parameters is far more homogeneous for miles .
table [ tab : sedproperties ] summarizes the ranges in age , metallicity and imf slopes for which our miles seds can be considered safe .
we do , however , strongly recommend to inspect fig .
[ fig : quality_miles ] for a more accurate description of the quality of the ssp seds .
although not shown , we find that the @xmath87 values obtained for the two kroupa imf shapes and the bimodal imf with slope 1.3 are virtually identical to those shown in the second panel of fig . [
fig : quality_miles ] . for the bimodal imf
the @xmath87 values also decrease with increasing imf slope , but the effect is much lower than is the case for the unimodal imf . for an extensive comparison of our newly synthesized ssp seds with the models of other authors we refer the reader to the recently published paper of koleva et al .
( 2008 ) and cid - fernandes & gonzlez ( 2010 ) .
koleva et al . compared three ssp spectral libraries : galaxev ( bruzual & charlot 2003 ) , pegase - hr ( le borgne et al . 2004 ) and ours . whereas the two former codes use the padova 1994 stellar evolution models ( bertelli et al .
1994 ) , ours employ the padova 2000 models ( girardi et al .
2000 ) , whose red giant branch are about 50200k hotter .
each of these codes use different prescriptions for transforming the theoretical stellar parameters to the observational plane .
the main difference is that our models mostly rely on empirical relations , rather than on model atmospheres .
however , the key difference between these models is the stellar libraries that they use , i.e. , stelib , elodie ( version 3.1 ) and miles for galaxev , pegase - hr and ours , respectively .
koleva et al . (
2008 ) do this comparison by means of the full spectrum fitting approach , using two different algorithms : nbursts ( chilingarian et al .
2007 ) , which performs a parametric non - linear fit , and steckmap ( ocvirk et al .
2006ab ) , which employs a non - parametric formalism .
koleva et al .
( 2008 ) find that the quality of the models is very dependent on the atmospheric parameter coverage of the stellar spectral libraries .
the two fitting methods show consistent results when our seds are compared to the pegase - hr seds , obtaining residuals of the order of 1 per cent . as expected , the largest discrepancies are found for low metallicities ( ) , as shown in their fig . 2 .
however in fig .
[ fig : miles - parameters ] we show that miles provides a significantly better coverage than elodie 3.1 for this low - metallicity regime both , for dwarfs and giants .
furthermore the last two panels of fig . 2 of koleva et al .
show that the discrepancy between these two model sets are larger for ssps with ages smaller than @xmath135gyr , which can be explained by the lack of hot metal - poor dwarfs in the elodie library .
the fits performed by these authors reveals that the worst results are obtained when either pegase - hr or our models are compared with the seds of bruzual & charlot ( 2003 ) .
koleva et al .
conclude that this is due to the poor stellar atmospheric coverage of stelib , particularly for the non solar metallicity regime , as is also shown in our fig .
[ fig : miles - parameters ] . this has been recently confirmed by cid - fernandes & gonzlez ( 2009 ) by comparing the model spectra to milky way globular clusters .
when comparing our miles models with the models that we presented in v99 we show in fig .
[ fig : v99vsmiles_0.0 ] a representative ssp spectrum of solar metallicity and 10gyr , with the corresponding v99 model over plotted .
the latter is an updated version of the original v99 models , which has been computed with the same code used to synthesize the miles ssp seds . the v99 ssp spectrum , with a resolution of 1.8(fwhm ) ,
was smoothed to match the resolution of the miles model , i.e. , 2.3(fwhm ) .
the residuals are plotted on a similar scale .
the residuals show a systematic trend in the two narrow spectral ranges covered by the v99 models .
this must be attributed to errors in the flux - calibration quality of the latter , as was reported in v99 .
the same pattern is seen in the residuals of fig .
[ fig : v99vsmiles_1.7 ] , which shows a similar comparison at metallicity .
contrary to fig .
[ fig : v99vsmiles_0.0 ] , the residuals of fig .
[ fig : v99vsmiles_1.7 ] show much larger variations at shorter wavelength scales .
this translates into larger index strength variations in this low metallicity regime , showing that the jones ( 1999 ) library feeding the v99 models lacks important stars .
this comparison illustrates how relevant it is to synthesize ssp seds with stellar libraries with good atmospheric parameters coverage , such as miles .
1.8 ) than miles , was smoothed to match the resolution of the miles model ( [email protected] ) .
the residuals are plotted in the lower panel with the same scale.,width=316 ] but for metallicity .,width=316 ]
a major application of the ssp seds is to produce key line - strength indices that can be compared to the observed values . in practise
this has been the most popular approach for studying in detail the stellar content of galaxies .
so far the most widely used indices are those of the lick / ids system ( e.g. , gonzlez 1993 ; trager et al .
1998 , 2000 ; vazdekis et al . 1997 ; jrgensen 1999 ; snchez - blzquez et al .
2003 , 2006a , b , c , 2009 ) .
the standard lick / ids system of indices was defined on the basis of a stellar spectral library ( gorgas et al . 1993 ; worthey et al .
1994 ) , which contains about 430 stars in the spectral range @xmath84000 - 6200 .
the lick / ids system has been very popular because the stellar library contains stars with a fair range of stellar parameters and , therefore , is appropriate to build stellar population models .
the most popular application of this library has been the use of empirical fitting functions , which relate the index strengths , measured at the lick / ids resolution , to the stellar atmosphere parameters . because the value of the lick / ids indices depend on the broadening of the lines , authors with spectral data who wish to use the population models based on the lick / ids fitting functions ( e.g. , worthey 1994 ; vazdekis et al .
1996 ; thomas et al .
2003 ) , or compare them with previously published data , need to transform their measurements into the lick / ids system .
as the resolution of the lick / ids library ( fwhm@xmath138 - 11 ) is much lower than what is available with modern spectrographs , the science spectra are usually broadened to the lower lick / ids resolution .
this is performed by convolving with a gaussian function , whose width varies as a function of wavelength , as the resolution of the lick / ids system , apart from being low , also suffers from an ill - defined wavelength dependence ( see worthey & ottaviani 1997 ) .
this effect is particularly significant for systems with small intrinsic velocity dispersions , such as globular clusters and dwarf galaxies , for which part of the valuable information contained in the higher resolution galaxy spectra is lost .
after that , it is still necessary to correct the measurements for the line broadening due to the velocity dispersion of the stars in the integrated spectra of galaxies .
the correction should be calculated by finding a good stellar template of the galaxy by combining an appropriate set of stellar or synthetic spectra which matches the observed spectrum .
this is because this correction is highly sensitive to the strength of the index . despite of this
, many studies still use single polynomial obtained with a single star as a template , that may have very different index - values than those on the observed spectra ( see kelson et al .
2006 for a discussion of the systematic effects associated to this procedure ) . due to the fact that the shape of the continuum in the lick / ids stars is not properly calibrated , the use of the lick system requires a conversion of the observational data to the instrumental response curve of lick / ids dataset ( see the analysis by worthey & ottaviani 1997 ) that will affect , predominantly , the broad indices .
this is usually done by observing a number of lick stars with the same instrumental configuration as for the science objects .
then , by comparing with the tabulated lick index measurements , the authors find empirical correction factors for each index .
this experiment shows that even for the narrow indices , differences exist between the lick / ids stars and other libraries ( see worthey & ottavianni 1997 ; paper i ) . the origin of these differences is not clear , but they need to be corrected .
furthermore , the spectra of this library have a low effective signal - to - noise ratio due to the significant flat - field noise ( dalle ore et al .
1991 ; worthey et al . 1994 ; trager et al .
this translates into larger random errors in the indices , much larger than in present - day galaxy data .
it is worth noting that the accuracy of the measurements based on the lick system is often limited by these transformations , rather than by the quality of the galaxy data . with the advent of new stellar libraries , such as miles , with better resolution , better signal - to - noise , larger number of stars and a very complete coverage of the atmospheric parameter space , it is about time to revisit the standard spectrophotometric system at which the indices are measured .
it is not our objective to tell the authors at which resolution the indices need to be measured .
in fact , we recommend to use the flexibility provided by the model ssp seds and compare the measurements performed on each object with models with the same total broadening .
however , this might not be a valid option in many cases , e.g. , when a group of measurements is to be represented together in the same index - index diagram .
it is also convenient to agree on a standard resolution at which to measure line - strengths , so that measurements can be compared easily with previous studies , with colleagues , and to avoid systematic errors that might affect the conclusions .
one of the reasons why the lick / ids system is still popular nowadays is because most previous studies have calculated their indices in this system .
although it is a good standard procedure to compare new measurements with previous works , the fact that those are transformed into the lick / ids systems perpetuates the problem . for this reason , we propose in this section a new line index system , hereafter lis , with three new spectral resolutions at which to measure the lick indices .
note that this new system should not be restricted to the lick set of indices in a flux calibrated system .
in fact , lis can be used for any index in the literature ( e.g. , for the rose ( 1994 ) indices ) , including newly defined indices ( e.g. , cervantes & vazdekis 2009 ) .
we provide conversions to transform the data from the lick / ids system to lis as well as tables with index measurements in the new system for popular samples of milky way globular clusters , nearby elliptical galaxies and bulges . for this purpose
we use index tables already published by other authors , and transform them into the new system . to calculate the conversion , we have broadened the miles stars in common with the lick / ids library ( 218 stars in total ) to the new standard resolutions defined for both , globular clusters and galaxies , and compare the indices measured in both datasets .
third - order polynomials were fitted in all cases .
these transformations are given in appendix [ ap : transformations ] . in a forthcoming paper
, we will present empirical fitting functions for the lick indices at these new standard resolutions proposed here .
a preliminary version of these fitting functions can be found in martn - hernndez et al .
( 2007 ) , which were computed at the resolution 2.3 ( fwhm ) .
we concentrate here on the lick / ids suite of indices because it has been the most widely used in the literature .
the definition of this new system using the miles stellar library has two main practical advantages : * the resolution of the miles stars is constant as a function of wavelength . as a result
, the user will not have to degrade the spectra with a wavelength dependent function .
* additional offsets will not have to be applied to the data , as long as the measurements are flux calibrated ( in a relative sense ) .
therefore the system is universal , and the index measurements straightforward to perform .
it is very difficult to choose an appropriate resolution at which to define a new system , as it depends on the intrinsic broadening of the objects and the instrumental resolution .
the lick / ids indices have been measured in a variety of objects , from globular clusters to giant elliptical galaxies , covering a very wide range in velocity dispersion , and in data with a variety of instrumental resolutions .
the appropriate resolution also depends on which indices are going to be measured ; while some lick / ids indices do not change their sensitivity to age and metallicity with broadening , others ( e.g. ca4227 , fe4531 ) do . for these reasons
, we consider it appropriate to define three different standard resolutions .
the first one has been specifically chosen for stellar clusters and dwarf galaxy data , while the other two are more appropriate for working with massive galaxy spectra . when choosing a standard resolution at which to compare indices of gcs , we were guided by the resolution of existing high - quality gc data .
schiavon et al .
( 2005 ) presented a spectral library of 40 galactic gcs with a resolution of @xmath91(fwhm ) .
this represents an important comparison database for extragalactic gc systems , and a logical upper limit for our choice of resolution .
however , the majority of extragalactic spectral data for gcs have lower resolution . data from the gemini / gmos collaboration ( e.g. , norris et al . 2008 ; pierce et al .
2006ab ) exhibits a typical fwhm of @xmath135 . very similar to that are the vlt / fors spectra presented by puzia et al .
keck / lris data from the sagesbrodie/sages/[email protected] ] group has been generally taken at slightly higher resolutions @xmath132.84 ( e.g. strader et al . 2005 ; beasley et al . 2006 ; cenarro et al . 2007 ;
chomiuk et al .
2008 ; beasley et al .
2009 ) . since it is considerably more straightforward to degrade higher resolution data to lower resolution than it is the attempt to correct lower resolution indices to higher resolution , we suggest using a resolution ( fwhm ) of 5 in gc index comparisons , a choice which encompasses all of the above data .
this choice reflects a factor of 2 improvement in resolution over the lick / ids system .
authors willing to compare their data with the models presented here will only have to broad their spectra to a resolution of 5 ( fwhm ) ( this corresponds to @xmath93 at 5000 ) and measure the indices in there .
if they wish to compare with previous published data in the lick / ids system , they can use the conversion provided in table [ ap : transformations ] . for galaxies we have decided to define two different standard resolutions .
the first one , discussed in this subsection , is fixed at resolution fwhm=8.4(this corresponds to @xmath94 at 5000 ) .
the reasons for this choice are twofold : @xmath95 it matches , roughly , the resolution of the lick / ids system in the h@xmath4 mg triplet region . the indices in this regions ( e.g. , h@xmath4 , mg@xmath5 , fe5270 ) are the most widely used in the literature and , therefore , the measurements in the new system will not differ considerably from previous measurements .
@xmath95 200@xmath96 is roughly the mean velocity dispersion for the early - type galaxies in the sloan digital sky survey ( bernardi et al .
2003a , b , c ) .
this system is particularly appropriate for studies of dwarf and intermediate - mass galaxies , where the total broadening of the spectra ( @xmath97 ) is not higher than @xmath98 at 5000 . in this case , the author will only have to broad their spectra to the proposed total @xmath83 using a gaussian broadening function and measure the indices .
if , on the contrary , the total broadening of the object is higher than this value , the author has the option of using the system lis-14.0 , or to correct his / her measurements to a total broadening of 8.4(fwhm ) using polynomials ( see , e.g. , kuntschner 2000 ; snchez - blzquez et al .
2006a ) .
in a similar way as for lis-5 , previous data in the lick system can be converted into this new system using the polynomials given in table [ conversion.gal ] .
the third system is defined at a resolution of fwhm=14 ( this corresponds to @xmath99357kms@xmath41 at 5000 ) .
lis-14.0 has been designed especially for the study of massive galaxies . in the spectra obtained for these galaxies the broadening
is dominated by the velocity dispersion of the stars .
the total broadening is higher than the standard lick / ids resolution for which , in order to compare with the models , usually one had to perform a correction due to velocity dispersion of the galaxy .
as mentioned above , this correction is one of the largest sources of systematic errors in the measurement of the indices ( e.g. kelson et al .
2006 ) . for studies that only analyse the stronger lick indices in a sample of galaxies with velocity dispersions @xmath100 200 kms@xmath41 , we suggest to broaden all the spectra by an amount such that the total resolution ( @xmath101 ) matches fwhm=14 , before measuring the indices , where @xmath102 represents the instrumental resolution and @xmath103 the velocity dispersion of the galaxy .
this will help us to avoid to have to perform any further correction to the indices due to broadening .
fig [ fig : lis_resolutions ] shows the synthetic spectra of an ssp with 10gyr and solar metallicity broadened to the resolution of the different lis- systems .
the polynomials required to transform the indices on the lick / ids system to this resolution are provided in table [ conversion.gal3 ] .
most spectrographs keep the resolution constant in units of wavelength , rather than velocity .
therefore , the broadening of the spectra , if done in linear wavelength scale , should be always performed using gaussians of a fixed number of pixels . if the broadening of the data has to be made with a gaussian with a constant width in kms@xmath41 .
then the broadening in the spectra should be performed on a spectrum with a logarithmic wavelength scale .
it is common practise in studies of stellar populations to observe stars in common with the lick / ids library .
this is done to derive offsets to transform the data onto the spectrophotometric system of the lick library .
we stress here that this is necessary , _ only _ because the lick / ids spectra suffer from a number of calibration problems .
however , these offsets will not correct properly any uncertainty in the calibration of the science spectra , as the position of the line - index on the detector changes for extragalactic objects due to their redshift .
lrrrrrrrrrrrrr + name & h@xmath104 & h@xmath105 & cn@xmath106 & cn@xmath107 & ca4227 & g4300 & h@xmath108 & h@xmath109 & fe4383 & ca4455 & fe4531 & c4668 & h@xmath4 + ngc5927 & @xmath110 & 0.1134 & @xmath111 & @xmath112 & 0.8685 & 4.9847 & @xmath113 & @xmath114 & 2.9922 & 0.6743 & 2.6763 & 3.0431 & 1.7502 + & @xmath115 & 0.0269 & @xmath116 & @xmath117 & 0.0176 & 0.0282 & @xmath118 & @xmath119 & 0.0493 & 0.0247 & 0.0424 & 0.0629 & 0.0335 + ngc6218 & @xmath120 & 2.8266 & @xmath121 & @xmath122 & 0.2969 & 3.3555 & @xmath123 & @xmath124 & 0.1026 & 0.0319 & 1.3028 & 0.1446 & 2.8334 + & @xmath125 & 0.0123 & @xmath126 & @xmath126 & 0.0098 & 0.0251 & @xmath127 & @xmath128 & 0.0358 & 0.0127 & 0.0344 & 0.0535 & 0.0278 + ngc6284 & @xmath129 & 2.6239 & @xmath130 & @xmath131 & 0.3717 & 3.8327 & @xmath132 & @xmath133 & 0.8237 & 0.1649 & 1.8133 & 0.5586 & 2.5439 + & @xmath134 & 0.0187 & @xmath126 & @xmath135 & 0.0128 & 0.0248 & @xmath136 & @xmath137 & 0.0442 & 0.0161 & 0.0363 & 0.0575 & 0.0272 + ngc6356 & @xmath138 & 0.6216 & @xmath139 & @xmath140 & 0.6710 & 5.4774 & @xmath141 & @xmath142 & 2.3657 & 0.4230 & 2.6000 & 1.7702 & 1.7484 + & @xmath125 & 0.0117 & @xmath143 & @xmath126 & 0.0100 & 0.0142 & @xmath144 & @xmath145 & 0.0272 & 0.0122 & 0.0253 & 0.0440 & 0.0189 + ngc6388 & @xmath146 & 1.0513 & @xmath147 & @xmath148 & 0.5320 & 4.8704 & @xmath149 & @xmath150 & 2.4337 & 0.3673 & 2.5714 & 1.7768 & 2.1259 + & @xmath151 & 0.0062 & @xmath152 & @xmath153 & 0.0049 & 0.0092 & @xmath154 & @xmath155 & 0.0158 & 0.0066 & 0.0120 & 0.0196 & 0.0098 + ngc6441 & @xmath156 & 1.1013 & @xmath157 & @xmath158 & 0.6307 & 4.8525 & @xmath159 & @xmath160 & 2.7413 & 0.3912 & 2.6228 & 1.7832 & 2.0552 + & @xmath161 & 0.0176 & @xmath126 & @xmath116 & 0.0150 & 0.0237 & @xmath162 & @xmath163 & 0.0333 & 0.0157 & 0.0320 & 0.0501 & 0.0240 + ngc6528 & @xmath164 & 0.3512 & @xmath165 & @xmath166 & 1.0559 & 5.6030 & @xmath167 & @xmath168 & 4.6496 & 0.7368 & 2.9746 & 4.4146 & 1.8889 + & @xmath169 & 0.0294 & @xmath170 & @xmath171 & 0.0194 & 0.0303 & @xmath172 & @xmath173 & 0.0509 & 0.0276 & 0.0447 & 0.0796 & 0.0351 + ngc6553 & @xmath174 & 0.8167 & @xmath175 & @xmath176 & 1.2451 & 5.7732 & @xmath177 & @xmath178 & 3.9143 & 0.6909 & 3.3030 & 3.7248 & 2.0026 + & @xmath179 & 0.0432 & @xmath180 & @xmath181 & 0.0328 & 0.0476 & @xmath182 & @xmath183 & 0.0777 & 0.0460 & 0.0672 & 0.1140 & 0.0521 + ngc6624 & @xmath184 & 0.6427 & @xmath185 & @xmath186 & 0.6539 & 5.3411 & @xmath187 & @xmath188 & 2.4164 & 0.4004 & 2.5801 & 1.8939 & 1.7645 + & @xmath189 & 0.0136 & @xmath190 & @xmath143 & 0.0092 & 0.0144 & @xmath191 & @xmath128 & 0.0252 & 0.0128 & 0.0218 & 0.0366 & 0.0199 + ngc6626 & @xmath192 & 2.4215 & @xmath193 & @xmath194 & 0.3655 & 3.7828 & @xmath195 & @xmath196 & 0.7371 & 0.0931 & 1.6467 & 0.6247 & 2.3904 + & @xmath197 & 0.0178 & @xmath143 & @xmath116 & 0.0113 & 0.0260 & @xmath198 & @xmath199 & 0.0381 & 0.0124 & 0.0351 & 0.0571 & 0.0278 + ngc6637 & @xmath200 & 0.3993 & @xmath201 & @xmath202 & 0.5761 & 5.5794 & @xmath203 & @xmath204 & 2.1029 & 0.3539 & 2.4863 & 1.7632 & 1.7367 + & @xmath205 & 0.0107 & @xmath153 & @xmath190 & 0.0071 & 0.0105 & @xmath206 & @xmath207 & 0.0214 & 0.0096 & 0.0200 & 0.0303 & 0.0141 + ngc6626 & @xmath192 & 2.4215 & @xmath193 & @xmath194 & 0.3655 & 3.7828 & @xmath195 & @xmath196 & 0.7371 & 0.0931 & 1.6467 & 0.6247 & 2.3904 + & @xmath197 & 0.0178 & @xmath143 & @xmath116 & 0.0113 & 0.0260 & @xmath198 & @xmath199 & 0.0381 & 0.0124 & 0.0351 & 0.0571 & 0.0278 + ngc6637 & @xmath200 & 0.3993 & @xmath201 & @xmath202 & 0.5761 & 5.5794 & @xmath203 & @xmath204 & 2.1029 & 0.3539 & 2.4863 & 1.7632 & 1.7367 + & @xmath205 & 0.0107 & @xmath153 & @xmath190 & 0.0071 & 0.0105 & @xmath206 & @xmath207 & 0.0214 & 0.0096 & 0.0200 & 0.0303 & 0.0141 + ngc6981 & @xmath208 & 1.7715 & @xmath209 & @xmath210 & 0.3862 & 3.5881 & @xmath211 & @xmath212 & 0.1925 & 0.0895 & 1.3977 & 0.3620 & 2.5091 + & @xmath213 & 0.0127 & @xmath143 & @xmath126 & 0.0091 & 0.0208 & @xmath214 & @xmath215 & 0.0360 & 0.0126 & 0.0333 & 0.0475 & 0.0255 + name & fe5015 & mg@xmath106 & mg@xmath107 & mgb & fe5270 & fe5335 & fe5406 & fe5709 & fe5782 & na5895 & tio@xmath106 & tio@xmath107 + ngc5927 & 4.9167 & 0.0756 & 0.1999 & 3.5827 & 2.2866 & 2.0628 & 1.3108 & @xmath216 & 0.8942 & 4.4516 & 0.0377&@xmath217 + & 0.0715 & 0.0007 & 0.0009 & 0.0358 & 0.0442 & 0.0524 & 0.0381 & @xmath218 & 0.0317 & 0.0396 & 0.0008 & @xmath135 + ngc6218 & 2.8312 & 0.0232 & 0.0582 & 1.1352 & 0.8644 & 1.0777 & 0.4432 & @xmath219 & 0.2400 & 1.1520 & 0.0101 & @xmath220 + & 0.0691 & 0.0006 & 0.0008 & 0.0343 & 0.0311 & 0.0490 & 0.0289 & @xmath221 & 0.0275 & 0.0387 & 0.0009 & @xmath222 + ngc6284 & 3.2827 & 0.0375 & 0.0854 & 1.5008 & 0.9413 & 1.2086 & 0.6797 & @xmath223 & 0.3272 & 2.2529 & 0.0083 & @xmath224 + & 0.0752 & 0.0007 & 0.0009 & 0.0323 & 0.0364 & 0.0454 & 0.0328 & @xmath225 & 0.0227 & 0.0420 & 0.0008 & @xmath135 + ngc6356 & 4.1852 & 0.0647 & 0.1602 & 2.8145 & 1.7640 & 1.8961 & 1.0925 & @xmath226 & 0.5489 & 3.1437 & 0.0224 & @xmath227 + & 0.0446 & 0.0005 & 0.0006 & 0.0214 & 0.0231 & 0.0326 & 0.0210 & @xmath228 & 0.0191 & 0.0262 & 0.0005 & @xmath143 + ngc6388 & 4.2186 & 0.0502 & 0.1310 & 2.1997 & 1.9315 & 1.9042 & 1.1532 & @xmath229 & 0.6698 & 3.6860 & 0.0219 & @xmath230 + & 0.0244 & 0.0002 & 0.0003 & 0.0107 & 0.0158 & 0.0146 & 0.0125 & @xmath231 & 0.0096 & 0.0143 & 0.0003 & @xmath152 + ngc6441 & 4.3328 & 0.0640 & 0.1586 & 2.7556 & 2.0015 & 1.9567 & 1.1619 & @xmath232 & 0.8016 & 3.9727 & 0.0128 & @xmath233 + & 0.0591 & 0.0006 & 0.0007 & 0.0269 & 0.0379 & 0.0432 & 0.0253 & @xmath234 & 0.0239 & 0.0265 & 0.0006 & @xmath143 + ngc6528 & 5.2490 & 0.1033 & 0.2382 & 3.7516 & 2.4408 & 2.5976 & 1.7286 & @xmath235 & 0.8249 & 5.1220 & 0.0575 & @xmath236 + & 0.0807 & 0.0009 & 0.0009 & 0.0395 & 0.0457 & 0.0479 & 0.0476 & @xmath237 & 0.0364 & 0.0402 & 0.0010 & @xmath116 + ngc6553 & 5.7870 & 0.0897 & 0.2324 & 3.9036 & 2.7011 & 2.5835 & 1.3858 & @xmath238 & 1.1831 & 3.7845 & 0.0523 & @xmath239 + & 0.0971 & 0.0011 & 0.0013 & 0.0514 & 0.0702 & 0.0753 & 0.0521 & @xmath240 & 0.0454 & 0.0563 & 0.0010 & @xmath116 + ngc6624 & 4.3141 & 0.0628 & 0.1554 & 2.7574 & 1.8628 & 1.8745 & 1.1159 & @xmath241 & 0.6466 & 2.5670 & 0.0337 & @xmath242 + & 0.0422 & 0.0004 & 0.0005 & 0.0166 & 0.0243 & 0.0342 & 0.0256 & @xmath243 & 0.0214 & 0.0219 & 0.0005 & @xmath190 + ngc6626 & 3.3106 & 0.0364 & 0.0811 & 1.4305 & 1.1532 & 1.1603 & 0.7001 & @xmath244 & 0.4725 & 1.9530 & 0.0187 & @xmath245 + & 0.0669 & 0.0005 & 0.0006 & 0.0299 & 0.0322 & 0.0365 & 0.0272 & @xmath246 & 0.0269 & 0.0322 & 0.0007 & @xmath143 + ngc6637 & 4.0857 & 0.0501 & 0.1388 & 2.5752 & 1.6781 & 1.6068 & 0.9602 & @xmath247 & 0.4894 & 2.4638 & 0.0263 & @xmath248 + & 0.0365 & 0.0004 & 0.0005 & 0.0164 & 0.0212 & 0.0237 & 0.0170 & @xmath249 & 0.0144 & 0.0225 & 0.0004 & @xmath190 + ngc6626 & 3.3106 & 0.0364 & 0.0811 & 1.4305 & 1.1532 & 1.1603 & 0.7001 & @xmath244 & 0.4725 & 1.9530 & 0.0187 & @xmath245 + & 0.0669 & 0.0005 & 0.0006 & 0.0299 & 0.0322 & 0.0365 & 0.0272 & @xmath246 & 0.0269 & 0.0322 & 0.0007 & @xmath143 + ngc6637 & 4.0857 & 0.0501 & 0.1388 & 2.5752 & 1.6781 & 1.6068 & 0.9602 & @xmath247 & 0.4894 & 2.4638 & 0.0263 & @xmath248 + & 0.0365 & 0.0004 & 0.0005 & 0.0164 & 0.0212 & 0.0237 & 0.0170 & @xmath249 & 0.0144 & 0.0225 & 0.0004 & @xmath190 + ngc6981 & 2.7705 & 0.0265 & 0.0532 & 1.1555 & 0.9544 & 0.7882 & 0.3914 & @xmath250 & 0.1239 & 1.2737 & 0.0070 & @xmath251 + & 0.0619 & 0.0006 & 0.0007 & 0.0304 & 0.0323 & 0.0394 & 0.0218 & @xmath252 & 0.0151 & 0.0367 & 0.0006 & @xmath135 + one of the reasons many authors keep using the lick / ids system is to compare with previous studies , which are usually presented in this system .
we provide here with a set of indices , obtained from several references already transformed to the new systems .
the data are available at the cds and in the electronic version of the paper .
we have selected several samples of early - type galaxies and globular clusters from the literature and transformed the line - strength indices to the new systems defined here .
this is intended to provide a benchmark of indices properly transformed into lis and measured on high quality data that other authors can use to compare with .
the selected samples are the following : * puzia et al .
this sample comprises 12 galactic globular clusters .
lick indices were measured from long - slit spectra obtained with the boller & chivens spectrograph mounted on the 1.5 m telescope in la silla . the wavelength coverage of the spectra ranges from 3400 to 7300 with a final resolution of @xmath136.7 .
all the spectra were observed with a slit with of 3@xmath2530 .
we transformed the lick indices into the new system using the transformations provided in this paper ( table [ conversion . ] ) .
* schiavon et al .
the sample consists of 40 galactic globular clusters , obtained with the blanco 4-m telescope at the cerro tololo observatory .
the spectra cover the range @xmath133350 - 6430 with 3.1 ( fhwm ) resolution .
the signal - to - noise ratio of the flux calibrated spectra ranges from 50 to 240@xmath41 at 4000 and from 125 to 500@xmath41 at 5000 .
the sample has been carefully selected to contain globular clusters with a range of ages and metallicities .
the indices provided in table [ table.index ] have been measured directly in the spectra degraded to the lis-5.0 resolution .
* trager et al . ( 2000 ) .
this sample consist of 40 galaxies drawn from that of gonzlez ( 1993 ) intended to cover , relatively uniformly , the full range of velocity dispersion , line - strength and color displayed by local elliptical galaxies . the line - strength indices ( i.e. h@xmath4 , mgb , fe5270 , fe5335 ) were measured from long - slit data obtained with the ccd cassegrain spectrograph on the 3 m shane telescope of lick observatory .
the spectral features were measured within r@xmath254/8 aperture .
the indices were transformed into the new system applying the transformations from tables [ conversion.gal ] and [ conversion.gal3 ] .
* snchez - blzquez et al .
( 2006a ) .
this sample consists of 98 early - type galaxies including ellipticals and lenticulars , spanning a wide range in central velocity dispersion ( @xmath83 ) .
the sample contains galaxies in different environments : isolated galaxies , galaxies in groups , and in several clusters .
the central spectra were extracted within an aperture 0.6 kpc ( assuming h@xmath255=70 kms@xmath41 ) .
the spectra were corrected from the presence of nebular emission using gandalf ( sarzi et al .
2006 ) , instead of using the [ oiii]@xmath256-based correcion published in the original paper .
we use the models provided in this paper as templates .
the spectral range covers from @xmath133500/3700 ( depending on the run ) to @xmath135200/5700 , with spectral resolution ranging from 3.6 to 6.56 ( fwhm ) ( see original paper for more details ) .
the indices provided in table [ table.index ] have been measured in the emission - free spectra previously degraded at the different lis resolutions . for those galaxies with a @xmath83 larger than the lis resolution
an additive correction was made using the kinematic templates ( the combination of ssp - models that better reproduced the galaxy spectra obtained in the measurement of @xmath83 using ppxf ( cappellari & emsellem 2004 ) ) , following the prescriptions in kelson et al .
the correction was obtained as the difference in the indices measured in these templates degraded at the lis - resolution and at the @xmath257 , where @xmath103 is the velocity dispersion of the galaxy and @xmath102 the instrumental resolution .
* yamada et al . ( 2006 ) .
the sample consists of 14 galaxies , morphologically classified as e in the virgo cluster .
the galaxies where selected with three criteria : ( 1 ) they had to be classified as elliptical galaxy in the rc3 catalogue ( de vaucouleurs et al .
1991 ) ; ( 2 ) they had to be free from strong emission lines ; ( 3 ) they have to cover a wide range of luminosity evenly distributed along the color - magnitude of the virgo cluster . these extremely high quality spectra , which have been extracted within an aperture of r@xmath254/10 , have signal - to - noise ratio well above @xmath1100@xmath41 for all the galaxies , and about @xmath13500@xmath41 for some of them . the spectra cover a wavelength range from 3800 to 5800 , and the resolution ranges from 2.0 to 3.1 , depending on the observational runs , which have been performed at the william herschel telescope ( wht ; 4.2 m ) and subaru ( 8.2 m ) telescopes . for full details of the sample
we refer the reader to the original paper .
three galaxies have a velocity dispersion larger than the resolution of the lis-8.4 and the indices in this system were corrected in the same way as in the sb06b sample * kuntschner et al .
this sample consists of 48 e / s0 galaxies in the sauron survey of early - type galaxies ( de zeeuw et al .
2002 ) at the william herschel 4.2 m telescope in la palma .
we use here the line strength measurements ( i.e. h@xmath4 , fe5015 , mg@xmath5 ) averaged within r@xmath254/8 aperture .
the spectral coverage of the observed sauron integral - field observations ranges from 4800 to 5380 at resolution @xmath134.2 ( fwhm ) .
the indices of table [ table.index ] were obtained applying the transformations of table [ conversion.gal ] and [ conversion.gal3 ] to the indices transformed into the lick systems provided in the original paper .
+ table [ table.index ] shows a portion of the tables to illustrate its format and content .
the table contains the lick / ids indices measured at the different lis resolutions with the index definitions provided in trager et al .
the full tables are available in the electronic version of the paper .
[ fig : lis5 ] shows the commonly used index - index diagnostic diagram h@xmath4 vs. [ mgfe ] in the new lis-5.0 system for the puzia et al .
( 2002 ) and schiavon et al .
( 2005 ) globular cluster samples .
we use the indices in table [ table.index ] . note that we do not plot the model indices for ssps younger than 10gyr for , according to our q@xmath258 parameter analysis . in fig .
[ fig : lis8.4 ] and fig .
[ fig : lis14 ] we show the h@xmath4 vs. [ mgfe ] and h@xmath4 vs. [ mgfe50 ] diagrams for the four galaxy samples both in lis-8.4 and lis-14.0 .
note that for all the galaxies of kuntschner et al .
( 2006 ) and for 36 galaxies in the snchez - blzquez et al .
( 2006a ) sample it was not possible to plot the [ mgfe ] index , as either one or the two iron indices ( fe5270 , fe5335 ) required to calculate this index were outside the covered spectral range .
note that we used gandalf ( sarzi et al .
2006 ) to remove nebular emission from the galaxy spectra of snchez - blzquez et al .
( 2006a ) and yamada et al .
we refer the reader to all these original works for a detailed discussion of the obtained results .
figure [ models - age ] shows the predicted time evolution of our solar metallicity and salpeter imf models for a selection of line - strength indices measured at four different spectral resolutions : the nominal resolution of the models , i.e. fwhm = 2.3 , and the resolutions of lis-5.0 , lis-8.4 , and lis-14.0 ( i.e. , fwhm@xmath05 , 8.4 and 14 , respectively ) .
it can be seen that the sensitivity of some indices to variations in resolution is barely noticeable .
this is the case for the wide molecular index definitions , i.e. , cn , mg and tio . for some lines , however , broadening
the spectra means loosing part of the information contained in them ( e.g. ca4227 ) .
the most critical situation when analyzing data by means of diagnostic diagrams based on these indices happens when the relative variation of the index with age depends on resolution .
authors should keep this in mind in order to choose the right system and diagnostics to work with the data .
our recommendation is to use the most straightforward approach possible , which is to work with the model seds smoothed to match the total resolution of the data .
this approach allows the user to take full advantage of the whole information contained in the data ( see v99 for an extensive discussion ) .
interestingly we note that the time evolution for most lick indices shows a minimum for intermediate - age stellar populations in the range 0.11.0gyr , as already mentioned in sec .
[ sec : sedsbehaviour ] .
this is particularly evident for the cn indices .
[ models - age - balmer ] is similar to fig .
[ models - age ] but here several balmer index definitions are shown .
we plot the standard lick h@xmath4 index and the four higher - order balmer indices of worthey & ottaviani ( 1997 ) .
we also include two of the age indicators of vazdekis & arimoto ( 1999 ) , the recently defined h@xmath259 index of cervantes & vazdekis ( 2009 ) and the d4000 break ( e.g. , gorgas et al .
we see that for all the balmer index definitions the dependence on resolution is marginal or virtually nil ( e.g. the 4000break ) .
the two h@xmath260 of vazdekis & arimoto ( 1999 ) are specifically defined to be insensitive to resolution .
for this reason we only plot , for h@xmath261 , the values corresponding to the resolution of lis-5.0 and for h@xmath262 the ones for lis-8.4 as these two resolutions fall within the stability range of these indices , respectively .
we did not include the third index defined in vazdekis & arimoto ( 1999 ) , h@xmath263 , as the resolution of lis-14.0 is larger than is allowed for this index .
interestingly , being almost negligible , we see that , whereas the sensitivity to resolution for the new h@xmath264 index is slightly larger than that for the standard h@xmath4 lick index for old stellar populations , the situation is reverse for ages smaller than @xmath131gyr .
the higher - order balmer line index definitions of worthey & ottaviani ( 1997 ) also show larger resolution sensitivity for younger stellar populations .
note that all balmer indices , except the d4000 break , peak at @xmath130.3gyr and , therefore , the fits obtained on the basis of these indices can have two solutions .
kauffmann et al .
( 2003 ) take advantage of the behaviour of these indices to define diagnostics for constraining the star formation history . figure [ models - z ] and fig .
[ models - z - balmer ] shows the variation of the line - strength indices as a function of metallicity predicted by our new models for stellar populations of 10gyr and salpeter imf . as in the previous figures the predictions
are shown for three different resolutions ; fwhm= 5.0 , 8.4 and 14 .
we see a resolution - dependent differential behaviour as function of metallicity for several indices , such as for example ca4227 , fe5335 , fe5782 .
this shows that a lot of care has to be taken to match the resolution of data and models when comparing both as , otherwise , artificial trends can be derived ( see kelson et al .
[ models - z - balmer ] shows the somewhat opposite dependence on the resolution of two h@xmath4 index definitions : h@xmath264 shows larger resolution dependence than h@xmath4 for higher metallicities .
note however the insensitivity to metallicity achieved by this h@xmath264 definition for , which is not seen for the commonly used h@xmath4 lick index .
such insensitivity is also shown by the two h@xmath265 indices of vazdekis & arimoto ( 1999 ) plotted here .
we are not describing here in detail the index bahaviour as function of metallicity and age , but the metallicity insensitivity of the two tio molecular indices for is particularly noteworthy . in fig .
[ fig : hbeta_transf ] we focus specifically on the behaviour of the most popular age indicator , the lick h@xmath4 index , as a function of age ( plotted in linear scale , for ages @xmath11gyr ) and metallicity . in this figure
we compare the index computed on the basis of the empirical fitting functions of worthey et al .
( 1994 ) , using the vazdekis et al .
( 1996 ) code , as updated in this work and transformed to the lis-8.4 system following table [ conversion.gal ] , with the h@xmath4 index measured on the ssp seds , once smoothed to the lis-8.4 resolution .
we see that the models based on miles provide larger values for solar and supersolar metallicities and for than those based on the lick / ids fitting functions .
this implies that when using miles models the obtained ages are older by about @xmath132gyr for ssps around @xmath1310gyr .
another interesting effect at metallicities is that , for ages above @xmath1312gyr , h@xmath4 starts increasing instead of decreasing as expected .
this effect , not seen at higher metallicities , is attributed to the contributions of the horizontal branch ( hb ) stars ( e.g. maraston & thomas 2000 ) .
therefore the h@xmath4 index , rather than indicating the age of the stellar population due to their sensitivity to the @xmath2 of the turnoff stars , become sensitive to the hb morphology .
this result is in agreement with maraston & thomas ( 2000 ) .
we see , however , that this sudden increase in the strength of the h@xmath4 index is less visible for the new predictions for the three lowest metallicities plotted in this figure .
in fact , the effect virtually disappears for .
this result would have important implications for the interpretation of integrated spectra of composite stellar systems , for which the balmer lines , such as the h@xmath4 index , are the primary age diagnostics .
in fact , maraston & thomas ( 2000 ) have suggested that a an spread in metallicity of an old stellar populations can account for most of the dispersion of h@xmath4 values observed in the ellipticals family due to the higher h@xmath4 index of the metal poor stars without the need of a young population as had been previously suggested in other studies ( e.g. gonzlez 1993 ) . as we obtain with miles higher h@xmath4 strengths , the required fraction for the metal - poor component should be smaller than is suggested by maraston & thomas ( 2000 ) .
it is very important to note that in fig .
[ fig : hbeta_transf ] we are comparing the same synthesis code , with identical hb prescriptions .
the different behaviour at low metallicities between the two sets of predictions should be fully attributed , therefore , to the limitations of the lick stellar library , as it becomes evident when comparing the @xmath87 values obtained for these two sets of predictions in this low metallicity regime ( see the second panel of fig .
[ fig : quality_miles ] and first panel of fig .
[ fig : quality_lick_stelib ] ) parameter , we only plotted the h@xmath4 values for ages larger than 10gyr for the miles models of .
however we plotted the h@xmath4 values obtained for the lick / ids fitting functions for , despite the fact that their corresponding @xmath87 values are lower than 1 ( see the first panel of fig .
[ fig : quality_lick_stelib ] ) ] .
one of the main advantages of miles is that the stellar spectra were carefully flux - calibrated .
in fact we obtained for each star of the library a wide - slit spectrum to avoid selective flux losses due to the differential refraction effect ( see paper i for details ) .
this allows us to measure accurate colours from the synthesized ssp seds .
this is performed by convolving the sed with the corresponding filter responses . to obtain the johnson @xmath266 broad band colour we use the filters of buser & kurucz ( 1978 ) , with the zero points set with vega . to obtain the zero point for the @xmath267 and @xmath268 filters , we used two different vega seds from hayes ( 1985 ) and from bohlin & gilliland ( 2004 ) , the latter based on high signal to noise space telescope imaging spectroph ( stis ) observations and extrapolated to higher wavelengths using kurucz model atmospheres .
after normalising the synthetic spectra to the sed of vega we measured the @xmath266 colours by applying the selected filter transmission curves .
the results are shown in fig .
[ fig : colours ] .
we also plot , for comparison , the photometric version of this colour following vazdekis et al .
( 1996 ) which is computed with the alonso et al .
( 1996 , 1999 ) calibrations .
the figure shows a good agreement for nearly all ages and metallicities , within the typical photometric uncertainties derived from the zero point of the @xmath267 and @xmath268 filters , as shown by the colours differences using two different normalizations of the vega spectrum .
we obtain rather similar time evolution colour curves for all the metallicities , with an offset of @xmath130.02mag . for the ssps with the offset can be as large as @xmath130.05mag ( miles seds colours slightly bluer ) , which is well below the photometric errors of the colours .
note , however , that for this metallicity the miles models are barely acceptable , as shown by the q parameter ( see fig . [
fig : quality_miles ] ) .
this is because of the stellar parameter coverage of the miles library and the fact that colour transformation are not accurate at such low metallicity . a rather similar offset ( @xmath130.015mag . )
has been quoted in paper i ( see their fig .
8) by comparing the @xmath266 colours derived from miles stellar spectra to the values tabulated in the lausanne photometric database ( mermilliod et al .
1997 ) for the same stars .
as discussed in paper i this offset sets an upper limit to the systematic uncertainties of our photometry . for the sake of clarity
we do not show in fig .
[ fig : colours ] the colours obtained from the model seds computed with the temperature scale of gonzlez - hernndez & bonifacio ( 2009 ) .
these models provide redder @xmath266 colour by @xmath130.02mag , i.e. similar to the zero point uncertainties .
this is because for a given temperature the stellar population model selects cooler spectra as the gonzlez - hernndez & bonifacio ( 2009 ) temperatures are hotter than miles by 51k .
the results presented here show that the colours derived from our models , either those predicted via photometric libraries or the ones measured from our miles ssp seds , are consistent .
such agreement shows the reliability of the temperatures adopted in paper ii , which are consistent with the temperature scales of alonso et al .
( 1996,1999 ) , i.e. the photometric libraries that feed our models . we also provide , in electronic form and on the web ( see section [ sec : web ] ) , colours measured directly in the miles spectra using combinations of various filter systems that are commonly used in the literature ( e.g. , sdss , hst ) .
colour derived from the ssp seds for different ages and metallicities ( as indicated within the panel ) and the kroupa universal imf with the zero point set with two different vega spectra in black ( very thin ) and red ( thicker ) ( see the text for details ) .
the thickest green line represents the photometric colour computed by vazdekis et al .
( 1996 ) , as updated in this work , which is based on extensive empirical photometric stellar libraries . ]
we compare , in this section , our new model seds with a set of representative stellar clusters of varying age and metallicity .
we also illustrate the use and capabilities of these models for the study of galaxies with varying age , metallicity , abundance ratio and mass , by giving a few examples for some well - studied galaxies . ) .
the obtained residuals are plotted at the same scale .
note that no continuum removal was applied when performing the fit . ]
galactic globular clusters allow us to uniquely test our evolutionary synthesis models , since independent age and metallicity estimates are available through deep cmd analyses .
these studies have concluded that , in most cases , globular clusters can be considered as formed in a single burst , which is characterized by a single age and a single metallicity , i.e. an ssp .
note , however , that there are clusters for which two stellar components have been identified ( e.g. , meylan 2003 ) and that stochastic effects ( e.g. , cervio et al . 2002 ) may affect significantly their integrated spectra .
we have selected three representative galactic globular clusters , with varying metallicity , from the sample of schiavon et al .
( 2005 ) , which is composed of 40 clusters .
these authors obtained integrated spectra of high signal - to - noise ( s / n@xmath13250 ) covering the range @xmath833506430 with 3.1(fhwm ) resolution .
[ fig : gcs ] shows the spectra and the best ssp sed fits overplotted .
we used , for this purpose , our ssp sed library with kroupa universal imf .
the ssp seds were conveniently smoothed to match the instrumental resolution of the observations ( i.e. @xmath269 ) .
we normalized the stellar cluster spectra and the model seds at 5000 , but we did not modify the spectral shape of the data ( and the models ) to perform the fits .
the best fit is understood as the one that minimizes the residuals , which are also plotted in the figure at the same scale .
the obtained ages and metallicities fully agree with the cmd - derived estimates of the harris ( 1996 ) catalog . for 47tuc
we obtained the same fit as in v99 .
for ngc5904 we derived a younger age , in agreement with the very deep cmd study of marin - franch et al .
( 2009 ) who have estimated for this cluster a relative age of 0.83 , i.e. the cluster is located within their young gc branch .
we also found a younger age for this cluster , by means of a line - strength analysis that included the newly defined h@xmath264 age indicator of cervantes & vazdekis ( 2009 ) .
the higher h@xmath4 value obtained for this cluster is attributed in part to the presence of an increased fraction of blue stragglers ( see cenarro et al .
2008 ) .
the residuals of fig .
[ fig : gcs ] do not show any significant colour difference in the covered spectral range , but some very low frequency variations that can be fully attributed to differences in the flux calibration .
in fact , note that these low frequency residuals are qualitatively identical among the three clusters .
however the residuals show differences in various molecular bands , mainly in the blue part of the spectra for all the clusters .
in particular , cn features at @xmath133860 and @xmath134150 , are much stronger in the clusters than in the models . we also see a residual in the g band at @xmath134300 , being this absorption weaker in the data . the cn - strong absorption , and
its anticorrelation with ch , is a well known feature in the galactic globular cluster family , which has been shown in both their individual stars ( e.g. , norris & freeman 1979 ) and its integrated spectra ( e.g. rose & tripicco 1986 ) .
a similar result is obtained in v99 for 47tuc , but using a different input stellar spectral library and also a different spectrum for the cluster .
such cn - strong absorption feature has also been reported for extragalactic globular clusters , like ngc1407 ( cenarro et al .
2007b ) . their lick
/ ids c@xmath1074668 vs. cn@xmath107 diagram shows the same anticorrelation with respect to the scaled - solar abundance ratios .
overall , we find that the residuals blueward of @xmath134300 are significantly larger than for the redder part of the spectra for all the clusters
. we do not discard possible effects due to the fact that the milky way globular clusters show oxygen - enhanced abundance ratios for low metallicities .
in fact cassisi et al .
( 2004 ) have shown that such effects become relevant for the spectral ranges covered by the @xmath267 , and mostly @xmath270 , broadband filters .
therefore an appropriate modelling for these clusters requires working with such @xmath9-enhanced stellar evolutionary isochrones as well as the use of stellar spectra with similar abundance pattern .
in addition we also require to use specific stellar spectra with cn - strong absorption features and cn - enhanced isochrones . in fig .
[ fig : m67 ] we show our sed fit to the integrated spectrum of the standard open cluster m67 ( schiavon , caldwell & rose 2004 ) .
unlike for the globular clusters of fig .
[ fig : gcs ] these authors obtained the integrated spectrum for m67 by co - adding individual spectra of cluster members , weighted according to their luminosities and relative numbers .
the obtained age and metallicity is in good agreement with these authors ( see also schiavon 2007 ) and with our isochrone fitting results ( vazdekis et al . 1996 ) within the uncertainties .
the residuals only show a low frequency pattern , which is related to differences in the flux calibration . however the higher frequency residuals are negligible .
it is particularly interesting that , unlike in fig . [ fig : gcs ] , we do not see the residuals blueward of 4300 .
this shows that all relevant spectral types are present in the miles library .
therefore the residuals obtained for galactic globular clusters can be attributed to the non - solar abundance ratios present in these clusters .
however we propose to use our base models in the same way that we describe in section [ sec : alpha ] to deal with such abundance patterns .
we illustrate in fig .
[ fig : gccenarro ] the potential use of these models for the study of extragalactic globular cluster systems .
we plot a keck / lris spectrum of a globular cluster of the massive elliptical galaxy ngc1407 ( cenarro et al .
2007b ) , with mean s / n@xmath1340 .
the model seds were properly smoothed to match the resolution of the observed spectrum ( @xmath271 ) .
we obtain an age of 17.78 gyr and a metallicity of @xmath272 in agreement , within the errors , with the values derived from the detailed line - strength analysis performed by these authors .
the line - strengths derived from our base models have already been compared to the line - strengths of the integrated spectra of galactic globular clusters ( e.g. , mendel et al .
2007 ; cenarro et al . 2008 ; cervantes & vazdekis 2009 ) .
these comparisons confirm the well known model zero - point problem affecting the age estimates obtained from the balmer lines , i.e. the derived ages are larger than the cmd - derived ages .
note that the same problem appears when using other models .
for an extensive description and discussion of this problem we refer the reader to the above papers and to , e.g. , gibson et al .
( 1999 ) , vazdekis et al .
( 2001a ) , and schiavon et al ( 2002 ) . in fig .
[ fig : gonzalezbonifaciossps ] we plot several line - strength diagnostic diagrams to obtain the ages and metallicities of the galactic globular cluster sample of schiavon et al .
we use , for this purpose , our reference models and the alternative set of models described in section [ sec : gonzalezbonifacio ] , which adopt the recently published temperature scale of gonzlez - hernndez & bonifacio ( 2009 ) .
we see that the data fall outside the model grid towards ages larger than the oldest predicted ssp ( see also fig .
[ fig : lis8.4 ] ) . this is older than the ages derived from cmds .
if we use the model grid that adopt the gonzlez - hernndez & bonifacio ( 2009 ) temperature scale we obtain , however , lower h@xmath264 values , i.e. , we alleviate in part the model zero - point problem obtaining slighly younger ages for the globular clusters than with the reference models , using the miles temperatures .
this is because when synthesizing a ssp sed the code assigns a cooler spectrum ( i.e. with lower balmer index values ) to a requested star of given parameters , as the gonzlez - hernndez & bonifacio ( 2009 ) temperatures are @xmath1351k hotter than in miles .
note also that these models provide slightly larger values for the metallicity indices .
however , as pointed out in section [ sec : gonzalezbonifacio ] , we do not find a strong physical reason to use the gonzlez - hernndez & bonifacio ( 2009 ) temperature scale , particularly for metal - rich objects .
note that fig .
[ fig : gonzalezbonifaciossps ] also shows that a significantly larger temperature offset ( @xmath13150k ) , i.e. , much larger than the current uncertainties in the temperature scale , is required to be able to overcome the model zero - point .
therefore other aspects , such as those discussed in vazdekis et al .
( 2001a ) and schiavon et al .
( 2002 ) should be considered .
it is remarkable the large age sensitivity of the h@xmath264 index , which allows us to obtain nearly orthogonal model grids , particulalry for @xmath273 .
we see that our age estimates for the most metal - rich globular clusters do not depend on the metallicity indicator in use .
however the obtained metallicities might differ among the plotted diagnostic diagrams , indicating a departure from the scaled - solar abundance pattern .
in fact , based on the h@xmath264 vs. cn2 diagram we obtain for many globular clusters metallicities that are significantly larger than the ones inferred from the h@xmath264 vs. [ mgfe ] plot .
therefore these clusters show cn2 values that are larger than expected according to their total metallicities , as traced by the [ mgfe ] index ( e.g. thomas et al . 2003 ) .
note that characteristic cn - strong pattern emerged in the residuals obtained from the full - spectrum fits shown in fig .
[ fig : gcs ] . finally , note that these index - index diagrams show that there are a group of clusters with apparently younger ages .
cenarro et al .
( 2008 ) found a clear correspondence between the h@xmath4 values and the specific frequency of blue stragglers ( bss ) , such that the higher the bss ratio the younger the apparent spectroscopic age .
we refer the reader to that paper for further details .
in this section we show the use of our models for studying galaxy spectra . with this aim
we have chosen four galaxies ( out of 14 ) of the virgo early - type galaxy sample of yamada et al .
( 2006 ) , with spectra of extraordinarily high s / n ( s / n @xmath1100@xmath41 at the h@xmath265 feature ) , extracted within an aperture of r@xmath254/10 .
our subsample of galaxies ( ngc4387 , ngc4473 , ngc4489 and ngc4621 ) cover a range in age , metallicity , abundance ratio and mass , as these galaxies were selected alongside the colour - magnitue relation of the virgo cluster .
the resolution of the subaru spectrum for ngc4489 is [email protected] , whereas for the other three galaxies , observed at the william herschel telescope ( vazdekis et al .
2001b ) , the resolution is [email protected] .
such high quality spectra were obtained to be able to measure the h@xmath260 age indicator of vazdekis & arimoto ( 1999 ) , which is virtually insensitive to metallicity .
mean luminosity weighted ages and metallicities were obtained by plotting h@xmath260 versus the total metallicity indicator [ mgfe ] ( gonzalez 1993 ; vazdekis et al .
2001b ; thomas et al .
the derived values are tabulated in table 5 of yamada et al .
we did not attempt here to fit the spectra of these galaxies , but in fig .
[ fig : galaxies ] we overplot for each galaxy the ssp sed with the closest age and metallicity , as listed in their table 5 .
the model seds were smoothed to match the total resolution of the galaxy spectra ( i.e. taking into account the instrumental resolution and galaxy velocity dispersion ) , which are indicated in the panels .
the residuals of the comparison for the low - velocity dispersion ( @xmath83 ) , old galaxy , ngc4387 , are very small . for the other galaxy with low @xmath83 but young age , ngc 4489 , the residuals in the blue part are as small as those obtained for ngc4387 .
however , we see a pattern in the residuals at longer wavelengths . if the stars of this galaxy would have formed in two bursts then a characteristic residual pattern would be seen , namely a surplus of light in the blue and a deficit in the red , which is not the case .
the good flux - calibration of the models presented here allows this type of analysis . unlike the other galaxies , the spectrum of ngc4489 was taken in one of the subaru observing runs , for which the correction of the continuum shape to a flux scale is less accurate than is achieved for the observing runs at the wht
the same applies to the feature seen in the residuals at @xmath134225 . for the most massive galaxies , ngc4473 and ngc4621 we see in fig .
[ fig : galaxies ] positive residuals for the fe features , e.g. , fe4383 , fe5270 , fe5335 , and negative residuals for the mg feature at @xmath135170 .
this shows the well known enhancement , which is characteristic of massive galaxies ( e.g. , peletier 1989 ; worthey , faber & gonzlez 1992 ; gonzlez 1993 ) .
we show here how our base models can be safely used for studying galaxies with non - solar element mixtures .
a good example of such use is shown in the line - strength analysis of yamada et al .
a proxy for the abundance ratio is obtained with the aid of the base model grids and an age indicator that is almost insensitive to metallicity .
if a galaxy spectrum is enhanced in mg / fe , we obtain a higher metallicity when this age indicator is plotted vs. a mg - dominated index than when the index is fe - dominated .
yamada et al .
use the h@xmath260 index of vazdekis & arimoto ( 1999 ) , which requires spectra with extremely high s / n , and the metallicity indicators mg@xmath5 and fe3 ( @xmath274 defined in kuntschner , 2000 ) .
irrespective of the metallicity indicator in use , such optimized age indicators provide orthogonal model grids , and the lines of constant age are essentially horizontal , which means that a given measurement for the age indicator corresponds to a unique age ( see cervantes & vazdekis 2009 ) .
the metallicity difference obtained in this way , [ z@xmath275/z@xmath276 , is a good proxy for the abundance ratio , as was shown in , e.g. , snchez - blzquez et al .
( 2006b ) , de la rosa et al .
( 2007 ) and michielsen et al .
( 2008 ) . in fig .
[ fig : heuristic ] we plot the values obtained by yamada et al .
( 2006 , table 5 ) for this proxy vs. the estimates derived with the aid of the models of thomas et al .
( 2003 ) , which specifically account for non scaled - solar element partitions . to apply these models we use the lick / ids version of the h@xmath4 , mg@xmath5 and fe3 indices of yamada et al .
( 2006 ) . as an example , table 5 of yamada et al . lists [ z@xmath275/z@[email protected] for ngc4473 and ngc4621 .
following fig .
[ fig : heuristic ] this translates to @xmath277 on the abundance ratio scale of thomas et al .
note that this [ mg / fe ] abundance scale is also a considerable approximation because _ i ) _ the use of scaled - solar isochrones , _ ii ) _ it is based on three synthetic spectra to model the ssps and _ iii ) _ a fixed element ratio pattern is adopted .
note also that in fig .
[ fig : heuristic ] we combine the results from two different model predictions , and that our method requires extrapolating to higher metallicities to obtain @xmath278 ( see the corresponding plots in yamada et al .
this explains in part the larger dispersion obtained for the higher abundance ratios .
it is worth noting that the newly defined h@xmath264 index of cervantes & vazdekis ( 2009 ) provides similar age disentangling power as the h@xmath260 index , but with the advantage that h@xmath264 requires spectra of similar qualities as usually employed for measuring the standard h@xmath4 index ( see cervantes & vazdekis 2009 ) .
h@xmath4 can also be used for this purpose , but the resulting model grids are significantly less orthogonal , increasing the error bars of the [ z@xmath275/z@xmath276 determinations .
finally , to obtain the total metallicity , these optimized age indicators can be plotted versus the combined indices [ mgfe ] ( gonzlez 1993 ) or [ mgfe] ( thomas et al .
2003 ) , which have been shown to be virtually insensitive to the abundance ratio . in the second paper of this series we will present an updated version of fig .
[ fig : heuristic ] , where we use both the scaled - solar and the @xmath9-enhanced version of our miles models , in order to be fully consistent .
in addition , these models reach higher metallicities , avoiding in this way the extrapolation that might be required to obtain [ z@xmath275/z@xmath276 .
however , we have shown here that our base models , which combine scaled - solar isochrones with an empirical stellar library , can be used for studying galaxies with @xmath9-enhanced element partitions via the line - strength analyses .
figure [ fig : hbeta_jones_miles ] shows the very popular diagnostic diagram h@xmath4 vs. [ mgfe ] , which is obtained from our miles ssp seds once smoothed to match the lis-14.0 resolution . for comparison we show the equivalent model grid from the v99 models , as updated in this work , which use the jones ( 1999 ) stellar spectral library .
[ fig : hbeta_jones_miles ] shows that the v99 model grid is significantly more orthogonal than the grid obtained with the miles ssp seds , mainly due to the lower sensitivity of h@xmath4 index to metallicity variations . the miles ssp grid looks , in fact , more similar to the equivalent standard grid obtained on the basis of the empirical fitting functions of worthey et al .
( 1994 ) on the lick / ids system .
we plot the radial line - strength profile for the virgo galaxy ngc4387 , which has been obtained from low resolution keck / lris spectra by snchez - blzquez et al .
we deliberately do not plot the observational errors , as we are only interested in illustrating the net effect on the trend .
note that we also plotted in fig .
[ fig : galaxies ] the higher resolution spectrum obtained at the william herschel telescope for the central aperture r@xmath279 , i.e. , 1.6arcsec ( yamada et al .
according to the v99 model grid , the observed gradient can be attributed to a decreasing metallicity with increasing galactocentric distance . using the miles ssp grid ,
the gradient can be explained by a combination of increasing age and decreasing metallicity .
we have investigated the possible cause for the differences between the two model grids .
as we use here an updated version for the v99 models , which is now based on the same isochrones and transformations to the observational plane that we employ for the miles models , these two ingredients can not be the the ones we are looking for .
we identified three other possible sources : _
i ) _ intrinsic differences in the spectra of miles and jones ( 1999 ) stellar libraries , mostly around the h@xmath4 feature ( such differences have been quoted before by worthey & ottaviani ( 1997 ) ; _ ii ) _ differences in the stellar parameters adopted here ( i.e. paper ii ) and in v99 and _ iii ) _ differences in the algorithms employed for computing / assigning a stellar spectrum to a given set of atmospheric parameters . in order to understand this issue , we synthetized seds using the 245 stars , both in the jones and miles libraries , with the following characteristics : ( a ) models using both stellar libraries . miles and jones.(b ) models using miles spectra assigning them the two set of stellar parameters , those used in v99 and in miles ( c ) models using miles stellar spectra and with both algorthims , the one used in v99 and the new algortihm described in v03 . for the sake of brevity we do not show here this detailed study but we summarize the main results .
we found that these model differences are mostly originated by the stellar spectrum assignation scheme employed in v99 , which works at a significantly more coarse stellar parameter resolution .
the other aspects described above also have a significant impact , but to a lesser degree .
furthermore the h@xmath4 index measurements on the ssp seds synthesized with the same algorithm but exchanging the stellar spectral libraries are consistent with the h@xmath4 index measurements obtained on a star by star basis .
unfortunately , this problem might have some impact on certain studies that make use of the h@xmath4 index derived from v99 models .
given the improvement on the stellar spectrum assignation scheme used for the miles ssps , we will provide an updated version of the v99 predictions in our website using the new algorithms .
vs. [ mgfe ] diagnostic diagram obtained from our miles base model seds ( solid line ) and v99 ( dotted line ) , which uses the jones ( 1999 ) stellar library .
the indices were measured on the ssp spectra once smoothed to match the lis-14.0 resolution .
we plot the radial line - strength profile for the virgo galaxy ngc4387 , which has been obtained from low resolution keck / lris spectra by snchez - blzquez et al .
the size of the filled circles decrease with increasing galactocentric distance : 1.25 , 2.5 , 5 , 10 , 20arcsec . ]
the exploitation of stellar population synthesis models has depended in part on the development of tools by users .
some authors offer software to either help to interpret or aid the user to handle the provided predictions .
a few examples include lector ( v99 ) , galaxev ( bruzual & charlot 2003 ) , pgase ( fioc & rocca - volmerange 1997,1999 ) , or ez ages ( graves & schiavon 2008 ) . in an effort to continue and extend these initiatives , we have created a new website ( http://miles.iac.es ) to provide support for using the spectra of both the miles stellar library and the model predictions described in this paper , but also in v99 at 1.8 resolution .
the webpage also handles the near - ir stellar spectral library cat and models that we published in cenarro et al .
( 2001a ) , cenarro et al .
( 2001b ) , cenarro et al .
( 2002 ) and v03 .
line - strength predictions based on empirical fitting functions ( including the ones of the lick / ids system ) are also included .
furthermore the webpage includes photometric quantities such as fluxes , colours , mass - to - light ratios and surface brightness fluctuations , which have been published in vazdekis et al .
( 1996 ) and blakeslee , vazdekis & ajhar ( 2001 ) .
it is very important to note that all these predictions will be continuously updated with the latest developments and with the feedback from users .
these comply with the standards of the virtual observatory .
this webpage is intended for users not only to retrieve both the stellar spectra and the synthetic ssp sed predictions in this project , but also to provide webtools to facilitate the handling and transformation of the spectra . beyond the basic tools for the exploration and visualization of the whole dataset , at a first stage
, we also provide advanced webtools to : * generate synthetic stellar spectra for a given set of stellar atmosphere parameters .
these spectra are computed on the basis of our stellar libraries according to the mathematical algorithm described in appendix b of v03 ( see also section [ sec : synthesis ] ) . *
transform the stellar spectra or the ssp spectra to the instrumental setup of the data observed by the user .
this tool will allow the user to input the details of the instrumental configuration of the observed data ( spectral resolution and sampling ) and generate a set of spectra matching that setup .
the model / library spectra will be convolved with the observational psf ( typically gaussian ) to match the spectral resolution of the observed data . during this step ,
particular attention will be paid as to whether the user requires the fwhm to be constant as a function of wavelength ( in which case convolution will be performed in linear space ) , or constant in velocity ( in which case convolution will be performed in logarithmic space ) . finally , the data will be rebinned ( i.e. spectrally sampled , and if desired , in log scale ) .
* measure a variety of line - strength indices and colors on the spectra .
we make use of lector ( v99 ) and indexf ( cardiel 2007 ) for this purpose .
this tool allows to input newly defined indices by the user . *
convert lick / ids line - strength indices to the new line index system proposed here .
we use for this purpose the transformations described in appendix a. * estimate the best luminosity weighted age and metallicity , and alpha - enhancement for a given set of line indices using rmodel ( cardiel et al .
this program determines stellar population parameters ( e.g. age , metallicity , imf type and slope ) , using as input line - strength indices by interpolating ssp model predictions .
both linear and bivariate fits are computed to perform the interpolation . *
provide spectra for a given star formation history .
these tools are intended to be the first of many more sophisticated tools , that we will incorporate in the website over time , to facilitate the comparison between the observed data and our stellar library and models , and in this way make the analysis easier .
here we present stellar population model seds based on the empirical stellar spectral library miles . in paper
i we have presented and described this library , whereas in paper ii we have published the stellar parameter determinations . here
we present synthetic seds for single - age , single - metallicity stellar populations ( ssps ) covering the full optical spectral range , 3540.57409.6 , at moderately high resolution , [email protected] , which is virtually constant as a function of wavelength ( see fig . 4 of paper i ) .
the ssp seds have a reliable , accurate flux calibration , which is one of the major advantages of the miles library .
furthermore the redder part of the ssp spectra do not show telluric residuals , as these were carefully cleaned from the stellar spectra as shown in paper i. we consider the seds presented here as our base models , as they combine an empirical library , which is imprinted with the chemical composition of the solar neighbourhood , with scaled - solar stellar isochrones ( girardi et al .
2000 ) . therefore , whereas our models are self - consistent , and scaled - solar for solar metallicity , this is not the case for the low metallicity regime , as the observed stars there do not show this abundance pattern ( see schiavon 2007 ) . in a second paper
we will present self - consistent models , both scaled - solar and @xmath9-enhanced , for a range in metallicities . for this purpose
we have used miles , together with theoretical model atmospheres , which are coupled to the appropriate stellar isochrones .
an interesting feature of our base models is that they rely , as much as possible , on empirical ingredients .
apart from using miles stellar spectra , we also use transformations of the theoretical parameters of the isochrones to observational , measurable , quantities that are based on extensive empirical , photometric libraries , with marginal dependence on model atmospheres .
the implementation of miles in the population synthesis code has been done very carefully : each star of the library has been compared with other miles stars with similar atmospheric parameters , using the same interpolation algorithm used in the synthesis code . as a result of these tests and other considerations ,
we have removed or decreased the weight with which a given star contributes to the synthesis of a stellar spectrum for 135 ( out of 985 ) stars of miles . the unprecedented stellar parameter coverage of the miles library has allowed us to safely extend our optical ssp seds from intermediate to very old age regimes ( 0.0618gyr ) and the metallicity coverage of the ssps from supersolar ( ) to .
the very low metallicity ssps should be of particular interest for globular cluster studies .
in addition , we have computed ssp seds for a suite of imf shapes and slopes : unimodal and bimodal imf shapes as defined in vazdekis et al .
( 1996 ) with slopes varying in the range 0.3 - 3.3 ( salpeter : unimodal with slope 1.3 ) , and the segmented imfs of kroupa ( 2001 ) ( universal and revised cases ) .
in addition to our reference models , which adopt the miles temperatures published in paper ii , we provide an alternative set of models that adopt the recently published temperature scale of gonzlez - hernndez & bonifacio ( 2009 ) ( i.e. , about @xmath1351k hotter ) .
the most relevant feature of these models is that the well known model zero - point problem affecting the age estimates obtained from the balmer lines , i.e. the obtained ages are larger than the cmd - derived ages , is partially alleviated . to constrain the range of ages , metallicities and imfs where the use of these seds are safe
, we provide a quantitative analysis .
this analysis , which basically takes into account the parameter coverage of the stellar library feeding the models , clearly shows that our ssp seds can be safely used in the age range 0.0618gyr , and for all the metallicities , except for , for which only stellar populations with ages above @xmath1310gyr can be considered safe .
this applies to the standard imf shapes .
however our quantitative analysis has shown that the quality of the ssp seds decreases as we increase the imf slope , especially for the unimodal imf .
for example , for slopes above 1.8 , the synthesized ssp seds for are no longer safe for a unimodal imf . according to this analysis
our predictions are of significantly higher quality than any previous model predictions in the literature for all the ages and metallicities that have been computed in this work .
for example , any predictions based on the standard fitting functions of the lick / ids system ( worthey et al .
1994 ) , e.g. line - strength indices , with which most stellar population studies have been performed so far , should not be considered safe for ( for standard imf shapes ) .
this limit is set at higher metallicity , , for predictions based on stelib ( e.g. , bruzual & charlot 2003 ) .
the comparison of our new models , based on miles , with the ssp seds published in v99 , reveals systematic differences due to limitations in the flux calibration of the jones ( 1999 ) stellar library that feeds the latter models . as for the line - strengths the most important difference between these two models is found for the h@xmath4 index , which shows a lower metallicity sensitivity in the v99 models .
we show an example where this effect has a non - negligible impact on our interpretation of galaxy age / metallicity gradients .
the ssp seds increase our ability to perform the stellar populations analysis , which can be done in a very flexible manner .
observed spectra can be studied by fitting the full spectrum or by focussing on selected line indices , either standard or newly defined .
an example of the previous method can be found in koleva et al .
( 2008 ) , where various stellar populations models were compared following this approach .
our favourite approach is to analyze the spectrum of a galaxy at the resolution imposed by its velocity dispersion ( and its instrumental resolution ) , which require us to smooth the ssp seds to match the observed total resolution .
therefore , the line index analysis to obtain relevant stellar population parameters is performed in the system of the galaxy , either through full spectrum fitting or line index approaches .
as many authors prefer to plot the line - strength measurements of their whole galaxy sample in a single index - index diagnostic diagram , or to compare their data to tabulated indices in the literature , we propose a new line index system ( lis ) , which makes this method straightforward to apply .
with lis we avoid the well known uncertainties inherent to the widely employed lick / ids system . unlike the latter , lis has two main advantages , i.e. , a constant resolution as function of wavelength and a universal flux - calibrated spectral response .
data can be analyzed in this system at three different resolutions : 5 , 8.4 and 14 ( fwhm ) , i.e. @xmath99127 , 214 and 357@xmath280 at 5000 , respectively .
these resolutions are appropriate for studying globular cluster , low and intermediate - mass galaxies , and massive galaxies , respectively .
line - strength measurements given at a higher lis resolution can be smoothed to match a lower lis resolution .
furthermore we provide polynomials to transform current lick / ids line - index measurements in the literature to the new system .
we provide lis line - index tables for various popular samples of galactic globular clusters and galaxies .
we also show various popular index - index diagnostic diagrams for these samples in the lis system . as an application
, we have fitted a number of representative stellar clusters of varying ages and metallicities with our models , obtaining good agreement with cmd determinations . unlike for the open cluster m67 ,
our sed fits for a sample of galactic globular clusters show non negligible residuals blueward of 4300 , which mostly reflect the characteristic cn - strong features of these clusters among other deviations from the scaled - solar pattern .
we also applied our models to representative galaxies with high quality spectra , for which independent studies are available , obtaining good results .
we show that our base models can be used for studying line - strength indices of galaxies with @xmath9-enhanced element partitions .
examples of such a use can be found in , e.g. , vazdekis et al .
( 2001b ) , kuntschner et al .
( 2002 ) , carretero et al .
( 2004 ) , yamada et al .
the method consists in plotting a highly sensitive age indicator , such as those of vazdekis & arimoto ( 1999 ) , or the recently defined h@xmath281 index of cervantes & vazdekis ( 2009 ) versus mg ( e.g. , mg@xmath5 ) and versus fe ( e.g. , fe4383 , @xmath6 ) , obtaining a virtually orthogonal model grid where the estimated age does not depend on the metallicity indicator in use . unlike the age , for a enhanced galaxy
the obtained metallicities differ when plotted against , for example , mg@xmath5 and @xmath6 indices .
this metallicity difference , [ z@xmath282/z@xmath283 , can be used as a good proxy for the abundance ratio determined with the aid of stellar population models that specifically take into account non - solar element partitions .
although this proxy yields larger values , there is a linear relation between these two ways of estimating the abundance ratios ( e.g. , snchez - blzquez et al .
2006b ; de la rosa et al .
2007 ; michielsen et al . 2008 ) . in the second paper of this series
we will present both @xmath9-enhanced and scaled - solar self consistent ssp seds , all based on miles , but modified with the aid of model atmospheres . by applying these models to galaxies with varying abundance ratios , for which high s
/ n spectra are available , we confirm the linear relation mentioned above .
there is one advantage in using this proxy : the abundance ratios determined in this way do not depend on the specific details of modeling of @xmath9-enhanced stellar populations such as those that are implicit to the computations of the atmosphere , or the adopted element mixtures .
our fits to the stellar cluster and galaxy spectra of high quality have confirmed the high precision of the flux calibration of our ssp seds .
furthermore the colours that we obtained from our ssp seds are consistent with the ones that we derived via the employed photometric stellar libraries , within typical zero point uncertaintes .
such agreement shows the reliability of the temperatures adopted for the miles stars ( paper ii ) , which are consistent with the temperature scales of alonso et al .
( 1996,1999 ) .
such accuracy opens new applications for these models .
for example , the model seds can be used as templates for codes aiming at determining spectrophotometric redshifts , particularly for those based on narrow - band filters ( e.g. alhambra , pau ) .
finally we present a webpage from which not only these models and stellar libraries can be downloaded , but where we also provide a suite of user - friendly webtools to facilitate the handling and transformation of the spectra .
for example , once the user enters the details of the instrumental setup employed in the observations ( e.g. , psf , sampling ) , the users can obtain their favourite line strength indices and diagnostic diagrams , ready to plot their observational measurements .
apart of the model seds shown here , the webpage also provides predictions for a suite of observable quantities .
we are indebted to the padova group for making available their isochrone calculations .
we are grateful to p. coelho , j. l. cervantes , i. g. de la rosa , m. koleva , e. ricciardelli and j. gonzlez - hernndez for very useful discussions .
we also would like to thank j. a. perez prieto for helping us in the construction of the web page for the models .
we thank the referee for pointing out relevant aspects that were useful for improving the original manuscript .
the miles library was observed at the int on the island of la palma , operated by the isaac newton group at the observatorio del roque de los muchachos of the instituto de astrofsica de canarias .
this research has made an extensive use of the simbad data base ( operated at cds , strasbourg , france ) , the nasa s astrophysics data system article service , and the _ hipparcos _ input catalogue .
psb acknowledges the support of a marie curie intra - european fellowship within the 6th european community framework programme .
jfb acknowledges support from the ramn y cajal program financed by the spanish ministry of science and innovation .
ramn y cajal _ fellow of the spanish ministry of science and innovation .
this work has been supported by the programa nacional de astronoma y astrofsica of the spanish ministry of science and innovation under grants _ aya2007 - 67752-c03 - 01 , aya2007 - 67752-c03 - 02 and aya2006 - 15698-c02 - 02_. 99 alexander d. r. , ferguson j. w. , 1994 , apj , 437 , 879 alonso a. , arribas s. , martnez - roger c. , 1995 , a&a , 297 , 197 alonso a. , arribas s. , martnez - roger c. , 1996 , a&a , 313 , 873 alonso a. , arribas s. , martnez - roger c. , 1999 , a&as , 140 , 261 beasley m. a. , cenarro j. , strader j. , brodie j. , , 2009 , aj , 137 , 5146 beasley m. a. , strader j. , brodie j. , cenarro j. , geha m. , 2006 , aj , 131 , 814 bernardi m. , et al .
2003a , aj , 125 , 1817 bernardi m. , et al .
2003b , aj , 125 , 1849 bernardi m. , et al . 2003c , aj , 125 , 1866 bertelli g. , bressan a. , chiosi c. , fagotto f. , nasi e. , 1994 , a&as , 106 , 275 bessell m.s .
, brett j.m . ,
wood p.r . , scholz m. , 1989 , a&as , 77 , 1 bessell m.s . , brett j.m . , scholz m. , wood p.r . , 1991 , a&as , 89 , 335 blakeslee j. p. , vazdekis a. , ajhar e. a. , 2001 , mnras , 320 , 193 blackwell , d. e. , petford , a. d , arribas , s. , haddock , d. j. , selby , m. j. , 1990 , a&a , 232 , 396 bohlin r. c. , gilliland r. l. , 2004 , aj , 127 , 3508 bruzual g. , charlot s. , 2003 , mnras , 344 , 1000 bruzual g. , 2007 , in vazdekis a. & peletier r. eds .
iau symposium 241 , stellar populations as building blocks of galaxies .
cambridge univ . press ,
p. 125 .
burstein d. , faber s. m. , gaskell c. m. , krumm n. , 1984 , apj , 287 , 586 buser r. , kurucz r. l. 1978 , a&a , 70 , 555 buzzoni , a. 1995 , apjs , 98 , 69 cappellari m. , emsellem e. 2004 , pasp , 116 , 138 cardiel n. , 2007 , highlights of spanish astrophysics iv , proceedings of the 7th scientific meeting of the spanish astronomical society , f. figueras et al .
eds . , springer , cd - rom .
cardiel n. , gorgas j. , snchez - blzquez p. , cenarro a. j. , pedraz s. , bruzual g. , klement j. 2003 , a&a , 409 , 511 carretero , c. , vazdekis , a. , beckman , j. , snchez - blzquez , p. , gorgas , j. 2004 , apj , 609 , l45 casagrande , l. , portinari , l. , flynn , c. 2006 , mnras , 373 , 13 cassisi s. , salaris m. , castelli f. , pietrinferni a. , 2004 , apj , 616 , 498 cenarro a. j. , cervantes j. l. , beasley m. a , marin - franch a. , vazdekis a. , 2008 , apj , 689 , l29 cenarro a. j. , et al
. , 2007 , mnras , 374,664 ( paper ii ) cenarro a. j. , beasley m. a. , strader j. , brodie j. , forbes d. , 2007b , aj , 134 , 391 cenarro a. j. , gorgas j. , cardiel n. , vazdekis a. , peletier r. f. , 2002 , mnras , 329 , 863 cenarro a. j. , cardiel n. , gorgas j. , peletier r. f. , vazdekis a. , prada f. , 2001 , mnras , 326 , 959 cenarro a.j . , gorgas j. , cardiel n. , pedraz s. , peletier r.f . ,
vazdekis a. , 2001b , mnras , 326 , 981 cervantes , j. l. , vazdekis a. , 2009 , mnras , 392 , 691 cervio m. , valls - gabaud d. , luridiana v. , mas - hesse j. m. , 2002 , a&a , 381 , 51 chilingarian i. v. , prugniel , p. , silchenko o. k. , koleva m. , 2007 , in vazdekis a. & peletier r. eds .
iau symposium 241 , stellar populations as building blocks of galaxies .
cambridge univ . press ,
p. 175 .
chomiuk l. , strader j. , brodie j. , 2008 , aj , 136 , 234 cid fernandes , r. , mateus , a. , sodr , l. , stasiska , g. , gomes , j. m. , 2005 , mnras , 358 , 363 cid fernandes , r. , gonzlez - delgado , r. , 2010 , mnras , 403 , 780 coelho p. , bruzual g. , charlot s. , weiss a. , barbuy b. , ferguson j. , 2007 , mnras , 382 , 498 coelho p. , barbuy b. , melndez j. , schiavon r. p. , castilho b. v. , 2005 , a&a , 443 , 735 covino , s. , galletti , s. , pasinetti , l. e. , 1995 , a&a , 443 , 735 dalle ore c. , faber s. m. , gonzalez j. j. , stoughton r. , burstein d. , 1991 , apj 375 , 427 de la rosa , i. g. , de carvalho , r. r. , vazdekis , a. , barbuy , b. , 2007 , aj , 133 , 330 de vaucouleurs , g. ; de vaucouleurs , a. ; corwin , h. g. , jr .
; buta , r. j. ; paturel , g. ; fouqu , p. 1991 , _ third reference catalogue of bright galaxies _ , springer , new york .
de zeeuw p. t. , et al .
, 2002 , mnras , 329 , 513 edvardsson , b. , andersen , j. , gustafsson , b. , lambert , d. l. , nissen , p. e. , tomkin , j. , 1993 , a&as , 102 , 603 falcn - barroso j. , balcells m. , peletier r. , vazdekis a. , 2003 , a&a , 405 , 455 fioc m. , rocca - volmerange b. , 1997 , a&a , 326 , 950 fioc m. , rocca - volmerange b. , 1999 , preprint ( astro - ph/9912179 ) fluks m.a . , plez b. , th p.s . , de winter d. , westerlund b.e .
, steenman h.c . , 1994 ,
a&as , 105 , 311 gibson b. k. , madgwick d. s. , jones l. a. , da costa g. s. , norris j. e. , 1999 , aj , 118 , 1268 girardi l. , bressan a. , bertelli g. , chiosi c. , 2000 , a&as , 141 , 371 girardi l. , bressan a. , chiosi c. , bertelli g. , nasi e. , 1996 , a&as , 117 , 113 gonzlez j. j. , 1993 , ph.d . thesis , univ .
santa cruz gonzlez - delgado , r. cervio , m. , martins , l. p. , leitherer , c. , hauschildt , p. h. , 2005 , mnras , 357 , 945 gonzlez - hernndez , j. , bonifacio , 2009 , a&a , 497 , 497 gorgas j. , cardiel n. , pedraz s. , gonzlez j. j. 1999 , a&as , 139 , 29 gorgas j. , faber s. m. , burstein d. , gonzlez j. j. , courteau s. , prosser c. , 1993 , apjs , 86 , 153 graves , g. j. , schiavon , r. p. , 2008 , apjs , 177 , 446 harris w. , 1996 , aj , 112 , 1487 hayes , d. s. , 1985 , in d. s. hayes , l. e. pasinetti , & a. g. d. philip , eds .
iau symposium 111 , calibration of fundamental stellar quantities .
dordrecht , reidel , p. 225 .
jones l. a. , 1999 , ph.d .
thesis , univ . of north carolina , chapel hill j orgensen i. , 1999 , mnras , 306 , 607 kauffmann g. et al .
, 2003 , mnras , 341 , 33 kholopov p.n .
et al . , 1998 , _ the combined general catalogue of variable stars 4th edition .
; moscow : sternberg astronomical institute ; http://vizier.u-strasbg.fr/ii/214a kelson d. , illingworth g. , franx m. , van dokkum p. g. , 2006 , apj , 653 , 159 koleva , m. , prugniel , ph . , ocvirk , p. , le borgne , d. , soubiran , c. , 2008 , mnras , 385 , 1998 korn a. j. , maraston c. , thomas d. , 2005 , a&a , 438 , 685 kroupa p. , 2001 , mnras , 322 , 231 kuntschner h. 2000 , mnras , 315 , 184 kuntschner h. , smith , r. j. , colless , m. , davies , r. l. ; kaldare , r. , vazdekis , a. 2002 , mnras , 337 , 172 kuntschner h. et al .
2006 , mnras , 369 , 497 le borgne , j .- f . , rocca - volmerange , b. , prugniel , p. , lancon , a. , fioc , m. , soubiran , c. , 2004 , a&a , 425 , 881 le borgne , j .- f . , bruzual , g. , pell , r. , lancon , a. , rocca - volmerange , b. , sanahuja , b. , schaerer , d. , soubiran , c. , vlchez - gmez , r. , 2003 , a&a , 402 , 433 lee h .- c . , worthey g. , 2005 , apjs , 160 , 176 lejeune t. , cuisinier , f. , buser r. , 1997 , a&as , 125 , 229 lejeune t. , cuisinier , f. , buser r. , 1998 , a&as , 130 , 65 macarthur l. a. , 2005 , apj , 623 , 795 maraston c. , 2005 , mnras , 362 , 799 maraston c. , thomas d. , 2000 , apj , 541 , 126 marigo , p. , girardi , l. , bressan , a. , groenewegen , m. a. t. ; silva , l. ; granato , g. l. 2008 , a&a , 482 , 883 marin - franch a. et al .
, 2009 , apj , 694 , 1498 mrmol - queralt e. , cardiel n. , cenarro a. j. , vazdekis a. , gorgas j. , pedraz s. , peletier r. f. , snchez - blzquez p. , 2008
, a&a , 489 , 885 martn - hernndez j. m. , mrmol - queralt e. , gorgas j. , cardiel n. , snchez - blzquez p. , cenarro a. j. , peletier r. f. , vazdekis a. , falcn - barroso j. , 2007 , in vazdekis a. & peletier r. eds .
iau symposium 241 , stellar populations as building blocks of galaxies .
cambridge univ .
press , p. 99 .
martins , l. p. , coelho , p. , 2007 ,
mnras , 381 , 1329 mendel j. t. , proctor r. n. , forbes d. a. , 2007 , mnras , 379 , 1618 mermilliod j .- c . , mermilliod m. , hauck b. , 1997 , a&as , 124 , 349 meylan g. , 2003 , in `` new horizons in globular cluster astronomy '' , asp conference proceedings , eds .
g. piotto , g. meylan , s. g. djorgovski , m. riello .
296 , p.17 michielsen d. , et al .
2008 , mnras , 385 , 1374 munari u. , sordo r. , castelli f. , zwitter t. , 2005 , a&a , 442 , 1127 murphy , t. , meiksin , a. , 2004 , mnras , 351 , 1430 norris j. , freeman k. c. , 1979 , apj , 230 , l19 norris m. a. et al . , 2008 , mnras , 385 , 40 ocvirk , p. , pichon , c. , lancon , a. , thibaut , e. , 2006a , mnras , 365 , 74 ocvirk , p. , pichon , c. , lancon , a. , thibaut , e. , 2006b , mnras , 365 , 46 panter , b. , heavens , a. f. , jimenez , r. , 2003 , mnras , 343 , 1145 peletier r. j. , 1989 , ph.d . thesis , univ .
groningen percival , s. m. , salaris , m. , 2009 , apj , 703 , 1123 pierce m. , et al . , 2006a , mnras , 366 , 1253 pierce m. , et al . ,
2006b , mnras , 368 , 325 pietrinferni , a. , cassisi , s. , salaris , m. , castelli , f. 2004 , apj , 612 , 168 proctor , r. , forbes , d. , beasley , m. a. , 2004 , mnras , 355 , 1327 prugniel ph .
, soubiran c. , 2004 , preprint ( astro - ph/0409214 ) prugniel ph .
, soubiran c. , 2001 , a&a , 369 , 1048 prugniel , p. , soubiran , c. , koleva , m. , & le borgne , d. 2007 , preprint ( astro - ph/0703658 ) puzia t. , et al . , 2004 , a&a , 415 , 123 puzia t. h. , saglia r. p. , kissler - patig m. , maraston c. , greggio l. , renzini a. , ortolani s. , 2002 , a&a , 395 , 45 ramrez , i. , melndez , j. , 2005 , apj , 626 , 446 rodrguez - merino , l. h. , chavez , m. , bertone , e. , buzzoni , a. , 2005 , apj , 626 , 411 rose j. a. , 1994 , aj , 107 , 206 rose j. a. , 1985 , aj , 90 , 1927 rose j. a. , tripicco m. j. , 1986 , aj , 92 , 610 salaris m. , groenewegen m. , weiss a. , 2000 , a&a , 355 , 299 salpeter e. e. , 1955 , apj , 121 , 161 snchez - blzquez p. , gorgas j. , cardiel n. , cenarro j. , gonzlez j. j. , 2003 , apj , 590 , l91 snchez - blzquez p. , forbes d. a. strader j. , brodie j. , proctor r. , 2007 , mnras , 377 , 759 snchez - blzquez p. , et al .
, 2006d , mnras , 371 , 703 ( paper i ) snchez - blzquez p. , gorgas , j. , cardiel , n. , gonzlez , j. j. , 2006a , a&a , 457 , 787 snchez - blzquez p.,gorgas , j. , cardiel , n. , gonzlez , j. j. , 2006b , a&a , 457 , 809 snchez - blzquez p.,gorgas , j. , cardiel , 2006c , a&a , 457 , 823 snchez - blzquez p. et al . , 2009 , a&a , 499 , 47 sarzi m. , et al . , 2006 , mnras , 366 , 1151 schiavon r. p. , 2007 ,
apjs , 171 , 146 schiavon r. p. , rose j. a. , courteau s. , macarthur l. , 2005 , apjs , 160 , 163 schiavon r. p. , caldwell n. , rose j. a. , 2004 , aj , 127 , 1513 schiavon r. p. , faber s. m. , castilho b. v. , rose j. a. , 2002 , apj , 580 , 850 schiavon r. p. , barbuy , b. , bruzual , g. 2000 , apj , 532 , 453 soubiran c. , katz d. , cayrel r. , 1998 , a&as , 133 , 221 strader j. , brodie j. , cenarro , a. j. , beasley m. a. , forbes d. a. , 2005 , aj , 130 , 1315 tantalo r. , chiosi c. , bressan a. , 1998 , a&a , 333 , 419 tantalo r. , chiosi c. , piovan l. , 2007 , a&a , 462 , 481 thomas , d. , maraston , c. , bender , r. , 2003 , mnras , 339 , 897 trager s. c. , faber s. m. , worthey g. , gonzlez j. j. , 2000 , aj , 119 , 1645 trager s. c. , worthey g. , faber s. m. , burstein d. , gonzlez j. j. , 1998 , apjs , 116 , 1 tripicco m. j. , bell r. a. , 1995 , aj , 110 , 3035 valdes , f. , gupta , r. , rose , j. a. , singh , h. p. , bell , d. j. , 2004 , apjs , 152 , 251 vazdekis a. , cenarro a. j. , gorgas j. , cardiel n. , peletier r. f. , 2003 , mnras , 340 , 1317 ( v03 ) vazdekis a. , salaris m. , arimoto n. , rose j. a. , 2001a , apj , 549 , 274 vazdekis a. , kuntschner h. , davies r. l. , arimoto n. , nakamura o. , peletier r. , 2001b , apj , 551 , l127 vazdekis a. , 1999 , apj , 513 , 224 ( v99 ) vazdekis a. , arimoto n. , 1999 , apj , 525 , 144 vazdekis a. , peletier r. f. , beckman j. e. , casuso e. 1997 , apjs , 111 , 203 vazdekis a. , casuso e. , peletier r. f. , beckman j. e. 1996 , apjs , 106 , 307 wild v. , walcher c. j. , johansson , p. h. , tresse l. , charlot s. , pollo a. , le fevre o. , de ravel l. , 2009 , mnras , 395 , 144 worthey g. , lee h .- c . , 2006 , preprint ( astro - ph/0604590 ) worthey g. , ottaviani d. l. , 1997 , apjs , 111 , 377 worthey g. , 1994 , apjs , 95 , 107 worthey , g. , faber , s. m. , gonzalez , j. j. , burstein , d. , 1994 , apjs , 94 , 687 worthey , g. , faber , s. m. , gonzalez , j. j. , 1992 , apj , 398 , 69 yamada y. , arimoto n. , vazdekis a. , peletier r. , 2006 , apj , 637 , 200 zwitter , t. , castelli , f. , munari , u. , 2004 , a&a , 417 , 1055 [ lastpage ]
in this appendix we provide the required transformations that make it possible to convert lick / ids index measurements to the new system proposed here . for this purpose
we have compared the index measurements of the 218 miles stars in common with the lick library .
the indices were measured once the miles stellar spectra were broadened to 5.0 , 8.4 and 14.0 to obtain the conversions corresponding to lis-5.0 , lis-8.4 and lis-14.0 , respectively . to calculate these conversions we fitted 3@xmath284 order polynomials in all cases .
these polynomials are defined as follows : where i@xmath286 and i@xmath287 represent the index values in the lis ( at either of the three resolutions ) and lick / ids system , respectively .
figure [ plots ] shows the results of the comparison between the indices of the lick / ids library and of miles at fwhm=5resolution .
similarly , fig .
[ lick.gal ] and fig .
[ lick.gal3 ] show the comparison of the indices of the lick / ids stars and of the miles stars broadened to fwhm=8.4 and fwhm=14 , respectively .
the coefficients of these polinomia , which convert the lick / ids indices to the lis-5.0 , lis-8.4 and lis-14.0 , are given in table [ conversion .
] , table [ conversion.gal ] and table [ conversion.gal3 ] , respectively . | we present synthetic spectral energy distributions ( seds ) for single - age , single - metallicity stellar populations ( ssps ) covering the full optical spectral range at moderately high resolution ( [email protected] ) .
these seds constitute our base models , as they combine scaled - solar isochrones with a empirical stellar spectral library ( miles ) , which follows the chemical evolution pattern of the solar neighbourhood .
the models rely as much as possible on empirical ingredients , not just on the stellar spectra , but also on extensive photometric libraries , which are used to determine the transformations from the theoretical parameters of the isochrones to observational quantities .
the unprecedent stellar parameter coverage of the miles stellar library allowed us to safely extend our optical ssp seds predictions from intermediate- to very - old age regimes , and the metallicity coverage of the ssps from super - solar to .
ssps with such low metallicities are particularly useful for globular cluster studies .
we have computed ssp seds for a suite of imf shapes and slopes .
we provide a quantitative analysis of the dependence of the synthesized ssp seds on the ( in)complete coverage of the stellar parameter space in the input library that not only shows that our models are of higher quality than other works , but also in which range of ssp parameters our models are reliable .
the ssp seds are a useful tool to perform the stellar populations analysis in a very flexible manner . observed spectra can be studied by means of full spectrum fitting or by using line indices . for the latter
we propose a new line index system ( named lis ) to avoid the intrinsic uncertainties associated with the popular lick / ids system and provide more appropriate , uniform , spectral resolution .
apart from constant resolution as function of wavelength the system is also based on flux - calibrated spectra .
data can be analyzed at three different resolutions : 5 , 8.4 and 14 ( fwhm ) , which are appropriate for studying globular cluster , low and intermediate - mass galaxies , and massive galaxies , respectively .
furthermore we provide polynomials to transform current lick / ids line index measurements to the new system .
we provide line - index tables in the new system for various popular samples of galactic globular clusters and galaxies .
we apply the models to various stellar clusters and galaxies with high - quality spectra , for which independent studies are available , obtaining excellent results .
finally we designed a web page from which not only these models and stellar libraries can be downloaded but which also provides a suite of on - line tools to facilitate the handling and transformation of the spectra .
[ firstpage ] galaxies : abundances galaxies : elliptical and lenticular , cd galaxies : stellar content globular clusters : general |
quite often there are problems within quantum chromodynamics ( qcd ) involving a single heavy quark with momentum @xmath1 where @xmath2 is the on - shell heavy - quark mass and @xmath3 . in situations
when the characteristic residual momentum is small ( @xmath4 ) and light quarks and gluons have small momenta ( @xmath5 ) it is possible to use a simpler effective field theory heavy quark effective theory ( hqet ) ( see , e.g. , the review @xcite and the textbooks @xcite ) . its lagrangian is a series in @xmath6 .
qcd operators are also given by series in @xmath6 in terms of hqet operators .
here we shall consider @xmath7 renormalized heavy
light qcd quark currents @xmath8 where @xmath9 is a dirac matrix , and the index 0 means bare quantities .
they can be expressed via operators in hqet @xmath10 where @xmath11 @xmath12 is the bare hqet heavy - quark field satisfying @xmath13 , and @xmath14 are dimension-4 hqet operators with appropriate quantum numbers .
we shall not discuss @xmath6 corrections in this paper ; our main subject is the matching coefficients @xmath15 .
the coefficients @xmath15 have been calculated at one - loop order in the pioneering paper @xcite . at two loops ,
they were calculated in ref .
@xcite , and corrected in ref .
, the term with @xmath16 in the square bracket , @xmath17 , should be @xmath18 .
] all - order results in the large-@xmath19 limit were obtained in @xcite ( see also @xcite ) .
asymptotics of perturbative series were investigated in a model - independent way in ref .
@xcite . in the present paper
we calculate the matching coefficients @xmath20 up to three loops
. these coefficients are useful for obtaining matrix elements of qcd currents ( such as @xmath21 ) from results of lattice hqet simulations ( see , e.g. , refs .
@xcite ) or hqet sum rules ( see , e.g. , refs .
there are eight dirac structures giving non - vanishing quark currents in four dimensions : @xmath22}\,,\quad \gamma_\bot^{[\alpha } \gamma_\bot^{\beta ] } \rlap / v\,,\quad \gamma_\bot^{[\alpha } \gamma_\bot^{\beta } \gamma_\bot^{\gamma]}\,,\quad \gamma_\bot^{[\alpha } \gamma_\bot^{\beta } \gamma_\bot^{\gamma ] } \rlap / v\ , , \label{match : gamma}\end{aligned}\ ] ] where @xmath23 , and square brackets mean antisymmetrization .
the last four of them can be obtained from the first four by multiplying by the t hooft veltman @xmath24 .
we are concerned with flavour non - singlet currents only , therefore , we may also use the anticommuting @xmath25 ( there is no anomaly ) .
the currents renormalized at a scale @xmath26 with different prescriptions for @xmath27 are related by @xcite @xmath28 } q\right)_\mu & { } = { } & z_t(\mu ) \left(\bar{q } \gamma_5^{\mbox{\scriptsize hv } } \gamma^{[\alpha } \gamma^{\beta ] } q\right)_\mu\ , , \label{match : larin}\end{aligned}\ ] ] where the finite renormalization constants @xmath29 can be reconstructed from the differences of the anomalous dimensions of the currents .
multiplying @xmath9 by @xmath25 does not change the anomalous dimension . in the case of @xmath30}$ ] , multiplying it by @xmath24 just permutes its components , and also does not change the anomalous dimension , therefore , @xmath31 @xmath32 are known up to three loops @xcite .
the anomalous dimension of the hqet current ( [ intro : jtilde ] ) does not depend on the dirac structure @xmath9 .
therefore , there are no factors similar to @xmath33 in hqet .
multiplying @xmath9 by @xmath25 does not change the matching coefficient .
therefore , the matching coefficients for the currents in the second row of ( [ match : gamma ] ) are not independent : they can be obtained from those for the first row . in the @xmath34 rest frame ,
where @xmath35 , we have @xmath36 in particular , two matching coefficients are equal : @xmath37}}(\mu)\ , .
\label{match : ct}\ ] ] in the following we shall consider only the matching coefficients for the first four dirac structures in ( [ match : gamma ] ) .
in order to find the matching coefficients @xmath15 , we equate on - shell matrix elements of the left- and right - hand side of eq .
( [ intro : match ] ) .
they are obtained by considering transitions of the heavy quark with momentum @xmath38 ( [ intro : p ] ) to the light quark with momentum @xmath39 : @xmath40 the on - shell matrix elements are ) , yielding a finite @xmath15 . ]
@xmath41 where @xmath42 and @xmath43 are the bare vertex functions , @xmath44 and @xmath45 are the on - shell wave - function renormalization constants of the heavy and the light quark in qcd , @xmath46 is the on - shell wave - function renormalization constant of the hqet quark field @xmath47 , and @xmath48 differs from @xmath45 because there are no @xmath49 loops in hqet .
the difference between @xmath50 and @xmath51 is of order @xmath52 , and can be neglected .
it is most convenient to use @xmath53 , then the @xmath54 term is absent .
the qcd vertex has two dirac structures : @xmath55 this leads to @xmath56 the scalar vertex function @xmath57 can be obtained by multiplying the diagrams by a projector and taking the trace . for the first two dirac structures in ( [
match : gamma ] ) , the projector @xmath58 can be used ; for the next two structures , @xmath59 .
the hqet vertex has just one dirac structure .
therefore , @xmath60 here @xmath2 is the on - shell mass of the heavy quark ( because the external heavy quark with @xmath61 should be on its mass shell ) .
therefore , mass - counterterm vertices have to be taken into account on all @xmath49 lines .
if all flavours except @xmath49 are massless , all loop corrections to @xmath62 , @xmath46 , and @xmath48 contain no scale and hence vanish : @xmath63 , @xmath64 , @xmath65 .
if there is another massive flavour ( charm in @xmath66-quark hqet ) , this is no longer so .
the vertex @xmath62 and @xmath46 have been calculated up to three loops in ref .
the massless - quark on - shell wave - function renormalization constant in the presence of a massive flavour has been calculated up to three loops in ref .
@xcite ( an explicit expression can be found in appendix a of ref .
the three - loop renormalization of the hqet current @xmath67 has been calculated in ref .
@xcite .
if all flavours except @xmath49 are massless , @xmath68 , @xmath44 , and @xmath45 contain a single scale @xmath2 .
the on - shell heavy - quark wave - function renormalization constant @xmath44 has been calculated at three loops in @xcite ( and confirmed in ref .
@xcite ) ; @xmath45 has been found in ref .
the vertex @xmath68 is the subject of the present paper .
if there is another flavour with a non - zero mass @xmath69 , there are two scales , and calculations become more difficult .
the renormalization constant @xmath44 has been calculated in this case , up to three loops , in ref .
@xcite ( the master integrals appearing in this case are discussed in ref .
the vertex @xmath68 and @xmath45 with two masses are considered in sect .
[ s : calc ] and [ s : zq ] .
the three - loop renormalization of the qcd currents with all dirac structures @xmath9 has been obtained in ref .
@xcite .
the qcd quantities @xmath57 , @xmath44 , and @xmath45 in the numerator of ( [ match : main ] ) are calculated in terms of @xmath70 , the bare coupling of the @xmath71-flavour qcd .
if there is another massive flavour , say charm , they also contain its bare mass @xmath72 ; we re - express it via the on - shell mass @xmath73 .
these quantities do not involve @xmath26 .
the @xmath7 renormalization constant @xmath74 is expressed in terms of @xmath75 .
the hqet quantities @xmath62 , @xmath46 , and @xmath48 in the denominator of ( [ match : main ] ) are calculated via @xmath76 and @xmath77 , the bare coupling and the bare @xmath78-quark mass in the @xmath79-flavour qcd of heavy flavours with mass @xmath2 and the number @xmath80 of massive flavours with mass @xmath73 , so that @xmath81 and @xmath82 . in reality ,
@xmath83 and @xmath84 .
if the @xmath78 quark is considered massless , we can also include it in the number of light flavours @xmath85 , and set @xmath86 . ]
( we re - express @xmath77 via the on - shell mass @xmath73 , which is the same in both theories ) .
the @xmath7 renormalization constant @xmath87 is expressed in terms of @xmath88 . to combine all these quantities in ( [ match : main ] ) , we re -
express them via the coupling @xmath88 using the decoupling relation @xcite ( an explicit expression for @xmath75 via @xmath88 , including the necessary terms with positive powers of @xmath89 , is given in eq .
( 12 ) of ref .
@xcite ) .
there exists an exact relation @xcite between the matching coefficients @xmath90 and @xmath91 .
namely , the renormalized vector and scalar currents are related by @xmath92 where @xmath93 is the @xmath7 mass of the heavy quark @xmath49 . taking the on - shell matrix element of this equality between the heavy quark with @xmath94 and the light quark with @xmath95 and re - expressing both qcd matrix elements via the matrix element of the hqet current with @xmath96
, we obtain @xmath97 the ratio @xmath98 has been calculated at three loops in refs . @xcite .
comparing @xmath99 with the analytical result @xcite for @xmath98 provides a strong check of our calculations .
for @xmath100 , this ratio has been calculated in ref .
the calculation of @xmath57 at @xmath101 is a single - scale problem .
the calculation is almost completely automated and similar to refs .
the feynman diagrams were generated with ` qgraf ` @xcite and classified into various topologies with the help of ` q2e ` and ` exp ` @xcite .
some sample diagrams are shown in fig . [
fig : nm_diagrams ] ( the first four diagrams ) .
scalar feynman integrals were reduced to master integrals using integration - by - parts identities @xcite .
this was done in two independent ways : using the ` form ` @xcite package ` shell3 ` @xcite and the ` c++ ` program ` crusher ` @xcite implementing the laporta algorithm @xcite .
the master integrals for the case of a massless charm quark are known from ref .
@xcite ( see also comments in ref .
@xcite ) . as an independent check the automatic setup described in ref .
@xcite has been used and the bare vertex functions have been checked numerically with the help of ` fiesta ` @xcite . the results for the bare vertex functions with all four dirac structures @xmath9 can be found on the web page @xcite .
+ + next we study the influence of a non - zero @xmath78-quark mass , @xmath69 , on the @xmath66-quark matching coefficients @xmath20 .
then @xmath78-quark loops exist both in the full qcd and in the @xmath66-quark hqet .
the full - theory quantities in the numerator of ( [ match : main ] ) depend on the ratio of the on - shell quark masses , @xmath102 some sample diagrams contributing to @xmath57 are shown in fig .
[ fig : nm_diagrams ] ( all diagrams except the first four ones depend on @xmath103 ) .
the hqet quantities in the denominator of ( [ match : main ] ) contain a single scale @xmath73 . from the technical point of view
the calculation is similar to refs .
we have used ` crusher ` @xcite for the reduction .
the master integrals are known from refs .
@xcite , see also ref .
most of the needed terms of their @xmath89 expansions are known analytically , in terms of harmonic polylogarithms @xcite ( hpls ) of @xmath103 ( the status of these expansions is summarized in tables 14 of ref .
@xcite ) . from the requirement of cancellation of @xmath104 poles in the matching coefficients ( sect .
[ s : coefs ] ) we were able to find exact analytical expressions ( in terms of hpls ) for the @xmath105 terms of the master integrals 5.2 and 5.2a ( fig . 8 in @xcite ) ; formerly , they were known only as truncated series in @xmath103 . the corresponding entry in table 3 of ref .
@xcite needs updating .
we do not present these long expressions here , they can be found at @xcite .
we have checked the @xmath73 dependent results by taking the limit @xmath106 and reproducing the @xmath85 part of the results given in sect .
[ s : coefs ] .
another check is taking the limit @xmath107 .
if we set @xmath108 and re - express the renormalized matching coefficients @xmath20 via @xmath109 , the results of sect .
[ s : coefs ] with the substitution @xmath110 are reproduced .
the last ingredient of eq .
( [ match : main ] ) which we had to calculate is the on - shell wave - function renormalization constant @xmath45 of a massless quark .
the result for can be extracted from ref .
@xcite , however , we have performed an independent calculation .
the @xmath73-dependent part is a new result .
the calculation is similar to that of ref .
@xcite , where @xmath44 has been calculated .
the integrals contributing to @xmath45 reduce to tadpoles when the mass of the incoming particle is set to zero .
there is only one new type of diagram which is shown in fig .
[ fig : zq ] , all others can be reduced to known results . for completeness
we give here the result for an arbitrary gauge group .
it reads @xmath111 where @xmath112 has the dimensionality @xmath113 , and the single - scale contributions are @xcite @xmath114 + \mathcal{o}(\varepsilon^4)\ , , \nonumber\\ z_a & { } = { } & \frac{1}{192 } \biggl\ { 1 + \frac{10}{3 } \varepsilon + \frac{227}{9 } \varepsilon^2 - \left(16 \zeta_3 + \frac{1879}{54}\right ) \varepsilon^3 \nonumber\\ & & { } - \xi \left [ 1 - 3 \varepsilon + \frac{35}{3 } \varepsilon^2 + \left(8 \zeta_3 - \frac{407}{9}\right ) \varepsilon^3 \right ] \biggr\ } + \mathcal{o}(\varepsilon^4)\ , , \nonumber\\ z_l & { } = { } & - \frac{\varepsilon}{72 } \left[1 - \frac{5}{6 } \varepsilon + \frac{337}{36 } \varepsilon^2\right ] + \mathcal{o}(\varepsilon^4)\ , , \nonumber\\ z_h & { } = { } & - \frac{\varepsilon}{36 } \left[1 - \frac{5}{6 } \varepsilon + \frac{151}{36 } \varepsilon^2\right ] + \mathcal{o}(\varepsilon^4)\,,\end{aligned}\ ] ] where @xmath115 , @xmath116 is the bare gauge - fixing parameter . , the free gluon propagator is @xmath117 . ] the two - scale contribution ( fig .
[ fig : zq ] ) is given by @xcite @xmath118
in this section , we present the results for the matching coefficients of the different heavy
light currents for the colour group su(3 ) ( results for a general colour group can be found at @xcite ) .
for this purpose we decompose the coefficients as follows @xmath119 where @xmath96 , @xmath120 , @xmath121 , @xmath122 , and @xmath82 is the number of active flavours in hqet .
furthermore , we use the abbreviation @xmath123 ( @xmath2 is the on - shell @xmath66-quark mass ) .
we first present the results for @xmath124 .
the individual contributions read @xmath125 @xmath126 @xmath127 @xmath128 where @xmath129 .
the two - loop results , as well as the coefficients @xmath130 , are known from ref .
@xcite ; all remaining three - loop results are new .
it is instructive to re - write the matching coefficients for @xmath101 and @xmath131 via the leading @xmath132-function coefficient in hqet , @xmath133 .
we obtain @xmath134 at two loops , and the following results at three loops : @xmath135 a method to estimate higher loop contributions called naive nonabelianization has been formulated in ref .
it is based on the fact that each polynomial in @xmath71 can be re - written as a polynomial in @xmath19 .
usually it is relatively simple to calculate the term with the highest power of @xmath71 .
this means that we know the coefficient of the leading power of @xmath19 . neglecting subleading powers of @xmath19 we obtain an estimate of the full result .
the two - loop corrections to the matching coefficients ( [ nna2 ] ) were among the examples confirming naive nonabelianization @xcite . at three loops
we see that this prescription reproduces the correct signs and roughly the correct magnitude of the full results . in all cases
we observe a compensation between the @xmath136 and @xmath137 terms . in the case of @xmath138 and @xmath139
this compensation is almost complete , and naive nonabelianization works surprisingly well .
since the expressions for the charm - mass dependence are quite involved and not completely expressed in terms of hpls of @xmath103 ( the @xmath140 terms of the master integrals 5.2 , 5.2a , 5.3 , 5.3a are known analytically only as truncated series in @xmath103 , see table 3 in ref .
@xcite ) , we refrain from listing them here
. instead , we present an expansion of our results to the second order in @xmath103 .
the results in terms of the master integrals and expansions to higher orders can be obtained from the web page @xcite .
our results for the @xmath73-dependent coefficients read @xmath141 x^2 + \mathcal{o}(x^3)\ , , \nonumber\end{aligned}\ ] ] @xmath142 x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\rlap{\scriptsize/}v}^{hm}(x ) & { } = { } & c_{\rlap{\scriptsize/}v}^{hl } + \left ( -\frac{817}{1350 } + \frac{1}{18 } \pi^2 -\frac{1}{15 } \ln x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\rlap{\scriptsize/}v}^{lm}(x ) & { } = { } & 2 c_{\rlap{\scriptsize/}v}^{ll } + \left(\frac{7}{108 } \pi^2 -\frac{1}{18 } \pi^2 \ln 2 -\frac{1}{36 } \pi^2 \ln
x \right ) x \nonumber\\ & & { } + \left ( -\frac{1}{2 } -\frac{1}{18 } \pi^2 -\frac{2}{9 } \ln x + \frac{1}{3 } \ln^2 x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\rlap{\scriptsize/}v}^{mm}(x ) & { } = { } & c_{\rlap{\scriptsize/}v}^{ll } + \left ( \frac{1}{135 } \pi^2 -\frac{1}{36 } \pi^2 \ln x \right ) x \nonumber\\ & & { } + \left ( -\frac{3}{4 } -\frac{1}{18 } \pi^2 -\frac{2}{9 } \ln x + \frac{1}{3 } \ln^2 x \right ) x^2 + \mathcal{o}(x^3)\ , , \\
c_{\gamma_\bot}^{m}(x ) & { } = { } & c_{\gamma_\bot}^{l } -\frac{1}{24 } \pi^2\ , x + \left ( \frac{3}{2 } + \ln x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot}^{gm}(x ) & { } = { } & c_{\gamma_\bot}^{gl } + \biggl [ -\frac{130327}{11664 } \pi^2 + \frac{14089}{972 } \pi^2 \ln 2 -\frac{143}{1944 } \pi^3 + \frac{11}{144 } \pi^2 { l}\nonumber\\ & & \hphantom{c_{\gamma_\bot}^{gl}+\biggl[\biggr . } + \frac{409}{216 } \pi^2 \ln x \biggr ] x \nonumber\end{aligned}\ ] ] @xmath143 x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot}^{hm}(x ) & { } = { } & c_{\gamma_\bot}^{hl } + \left ( \frac{349}{4050 } -\frac{1}{54 } \pi^2 -\frac{1}{15 } \ln x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot}^{lm}(x ) & { } = { } & 2 c_{\gamma_\bot}^{ll } + \left ( \frac{77}{324 } \pi^2 -\frac{11}{54 } \pi^2 \ln 2 -\frac{11}{108 } \pi^2 \ln x \right ) x \nonumber\\ & & { } + \left ( -\frac{19}{54 } + \frac{1}{54 } \pi^2 + \frac{2}{27 } \ln x -\frac{1}{9 } \ln^2 x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot}^{mm}(x ) & { } = { } & c_{\gamma_\bot}^{ll } + \left ( \frac{11}{405 } \pi^2 -\frac{11}{108 } \pi^2 \ln x \right ) x \nonumber\\ & & { } + \left ( -\frac{29}{108 } + \frac{1}{54 } \pi^2 + \frac{2}{27 } \ln x -\frac{1}{9 } \ln^2 x \right ) x^2 + \mathcal{o}(x^3)\ , , \\
c_{\gamma_\bot\rlap{\scriptsize/}v}^{m}(x ) & { } = { } & c_{\gamma_\bot\rlap{\scriptsize/}v}^{l } -\frac{5}{24 } \pi^2\ , x + \left ( \frac{7}{6 } -\frac{1}{3 } \ln x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot\rlap{\scriptsize/}v}^{gm}(x ) & { } = { } & c_{\gamma_\bot\rlap{\scriptsize/}v}^{gl } + \biggl ( -\frac{59221}{3888 } \pi^2 + \frac{6295}{324 } \pi^2 \ln 2 -\frac{65}{648 } \pi^3 + \frac{25}{144 } \pi^2 { l}\nonumber\\ & & \hphantom{c_{\gamma_\bot\rlap{\scriptsize/}v}^{gl}+\biggl(\biggr . }
+ \frac{541}{216 } \pi^2 \ln x \biggr ) x \nonumber\\ & & { } + \biggl [ \frac{35653}{5184 } + \frac{631}{432 } \pi^2 -\frac{7}{108 } \pi^2 \ln 2 + \frac{313}{48 } \zeta_3 -\frac{5}{6 } \pi^2 \zeta_3 -\frac{19}{2160 } \pi^4 \nonumber\\ & & \hphantom{{}+\biggl[\biggr . }
-\frac{5}{6 } \zeta_5 -\frac{35}{36 } { l}\nonumber\\ & & \hphantom{{}+\biggl[\biggr . }
+ \left ( \frac{575}{432 } + \frac{9}{4 } \pi^2 -\frac{1}{8 } \zeta_3 -\frac{43}{540 } \pi^4 + \frac{5}{18 } { l}\right ) \ln x \nonumber\\ & & \hphantom{{}+\biggl[\biggr . }
+ \frac{1}{2 } \ln^2 x \biggr ] x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot\rlap{\scriptsize/}v}^{hm}(x ) & { } = { } & c_{\gamma_\bot\rlap{\scriptsize/}v}^{hl } + \left ( \frac{49}{4050 } -\frac{1}{54 } \pi^2 -\frac{1}{15 } \ln x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot\rlap{\scriptsize/}v}^{lm}(x ) & { } = { } & 2 c_{\gamma_\bot\rlap{\scriptsize/}v}^{ll } + \left ( \frac{35}{108 } \pi^2 -\frac{5}{18 } \pi^2 \ln 2 -\frac{5}{36 } \pi^2 \ln x \right ) x \nonumber\\ & & { } + \left ( -\frac{23}{54 } + \frac{1}{54 } \pi^2 + \frac{2}{27 } \ln x -\frac{1}{9 } \ln^2 x \right ) x^2 + \mathcal{o}(x^3)\ , , \nonumber\\ c_{\gamma_\bot\rlap{\scriptsize/}v}^{mm}(x ) & { } = { } & c_{\gamma_\bot\rlap{\scriptsize/}v}^{ll } + \left ( \frac{1}{27 } \pi^2 -\frac{5}{36 } \pi^2 \ln x \right ) x \nonumber\\ & & { } + \left ( -\frac{37}{108 } + \frac{1}{54 } \pi^2 + \frac{2}{27 } \ln x -\frac{1}{9 } \ln^2 x \right ) x^2 + \mathcal{o}(x^3)\,.\end{aligned}\ ] ]
we are now in the position to apply our results to the matrix elements between a @xmath0 or @xmath144 meson with momentum @xmath145 and the vacuum .
they are defined through @xmath146 where @xmath147 is the @xmath144 polarization vector .
the corresponding hqet matrix elements ( at @xmath148 ) in the @xmath34 rest frame are @xmath149 where the single - meson states are normalized by the non - relativistic condition @xmath150 we also remind the reader that @xmath151 , so that there are only two currents .
these two matrix elements are characterized by a single hadronic parameter @xmath152 due to the heavy - quark spin symmetry . from eq .
( [ match : div ] ) we have @xcite @xmath153 here @xmath154 where @xmath155 is the residual energy of the ground - state @xmath0 meson in the limit @xmath148 . neglecting @xmath6 corrections , we see that this equation coincides with ( [ match : mm ] ) .
we have checked that our results agree with the known formulas for @xmath98 at @xmath101 @xcite and with @xmath73 corrections taken into account @xcite . using the expressions of section [ s : coefs ] we find the following results for the ratios of decay constants .
again , we present an expansion to the second order in @xmath103 for the charm mass dependent terms . for the numerical evaluations we use an expansion to the eighth order in @xmath103 and the values @xmath156 and @xmath157 .
@xmath158 \left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^2 \nonumber\\ & & { } + \bigl [ - 129.211 - 0.198 n_h + 14.294 n_l - 0.006 n_h^2 - 0.005 n_h
n_l \nonumber\\ & & { } - 0.331 n_l^2 + \bigl ( 14.294 - 17.816 x + 11.697 x \ln x \nonumber\\ & & \qquad { } - 0.273 x^2 -6.082 x^2 \ln x + 4.889 x^2 \ln^2 x \bigr ) n_m \nonumber\\ & & { } + \bigl ( -0.005 - 0.040 x^2 \bigr ) n_h
n_m \nonumber\\ & & { } + \bigl ( -0.661 + 0.692 x -0.731 x \ln x \nonumber\\ & & \qquad { } + 0.879 x^2
+ 0.296 x^2 \ln x -0.444 x^2 \ln^2 x \bigr ) n_l n_m \nonumber\\ & & { } + \bigl ( - 0.331 + 0.195 x - 0.731 x \ln x \nonumber\\ & & \qquad { } + 1.213 x^2 + 0.296 x^2 \ln x - 0.444 x^2 \ln^2 x \bigr ) n_m^2 \bigr ] \left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^3 \nonumber\\ & & { } + \mathcal{o}\left(\alpha_s^4,\frac{\lambda_{\mbox{\scriptsize qcd}}}{m_b}\right ) \nonumber\\ & { } = { } & 1 - \frac{2}{3 } \frac{\alpha_s^{(4)}(m_b)}{\pi } - ( 6.370 + 0.189 ) \left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^2 \nonumber\\ & & { } - ( 77.549 + 6.575 ) \left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^3 + \mathcal{o}\left(\alpha_s^4,\frac{\lambda_{\mbox{\scriptsize qcd}}}{m_b}\right ) \nonumber\\ & { } = { } & 1 - 0.046 - ( 0.030 + 0.001 ) - ( 0.025 + 0.002 )
\nonumber\\ & { } = { } & 0.899 - 0.003 = 0.896 + \mathcal{o}\left(\alpha_s^4,\lambda_{\mbox{\scriptsize qcd}}/m_b\right)\,.\end{aligned}\ ] ] in the second line from the bottom the corrections from tree level , first , second and third order in @xmath159 are given separately . also the contributions stemming from the finite charm mass are separated ( the second number in the parentheses ) . in the first part of the last line ,
the @xmath73 correction is also separated .
power corrections @xmath160 are discussed in refs .
@xcite and amount to several per cent . for the second ratio we obtain for @xmath161 @xmath162 @xmath163 \left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^2 \nonumber\\ & & { } + \bigl [ - 70.923 - 0.666 n_h + 9.175 n_l - 0.026 n_h^2 - 0.016 n_h
n_l \nonumber\\ & & { } - 0.222 n_l^2 + \bigl(9.175 - 7.859 x + 6.031 x \ln x \nonumber\\ & & \qquad { } + 4.555 x^2 + 3.256 x^2 \ln x - 0.278 x^2 \ln^2 x \bigr ) n_m \nonumber\\ & & { } + \bigl ( - 0.016 - 0.074 x^2 \bigr ) n_h n_m \nonumber\\ & & { } + \bigl ( - 0.444 + 0.346 x - 0.366 x \ln x - 0.074 x^2 \bigr ) n_l n_m \nonumber\\ & & { } + \bigl ( - 0.222 + 0.097 x -0.366 x \ln x - 0.074 x^2 \bigr ) n_m^2 \bigr ]
\left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^3 \nonumber\\ & & { } + \mathcal{o}\left(\alpha_s^4,\frac{\lambda_{\mbox{\scriptsize
qcd}}}{m_b}\right ) \nonumber\\ & { } = { } & 1 - ( 3.367 + 0.142 ) \left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^2 \nonumber\\ & & { } - ( 38.530 + 3.973 ) \left(\frac{\alpha_s^{(4)}(m_b)}{\pi}\right)^3 + \mathcal{o}\left(\alpha_s^4,\frac{\lambda_{\mbox{\scriptsize qcd}}}{m_b}\right ) \nonumber\\ & { } = { } & 1 - ( 0.016 + 0.001 ) - ( 0.013 + 0.001 ) \nonumber\\ & { } = { } & 0.971 - 0.002 = 0.969 + \mathcal{o}\left(\alpha_s^4,\lambda_{\mbox{\scriptsize qcd}}/m_b\right)\,.\end{aligned}\ ] ] the coefficients of these perturbative series ( at @xmath101 and @xmath161 ) can be re - written via @xmath164 : @xmath165 again , naive nonabelianization @xcite works quite well for these ratios predicting the correct sign and order of magnitude .
asymptotics of the perturbative coefficients for the matching coefficients at a large number of loops @xmath166 have been investigated in ref .
@xcite in a model - independent way .
the results contain three unknown normalization constants @xmath167 .
let us in the following assume that the number of loops @xmath168 is much larger than one and compare our results with ref .
the asymptotics of the perturbative coefficients for @xmath169 contain @xmath170 and @xmath171 ( see ( 5.6 ) in ref .
@xcite ) ; in the case of @xmath172 it contains only @xmath170 ( see ( 5.9 ) in ref .
@xcite ) has been adopted in ref .
@xcite ; @xmath173 is the renormalization - group invariant mass ( the exact definition used here is given in the unnumbered formula after ( 3.8 ) in ref .
] : @xmath174 \frac{n_2}{n_0 } \right\ } \nonumber\\ & & { } \times\left(\frac{m}{\hat{m}}\right)^{(n+1)}_{l=-5/3}\ , .
\label{asympt}\end{aligned}\ ] ] the coefficient of @xmath175 is about @xmath176 at @xmath177 , and it seems reasonable to neglect this contribution . neglecting also @xmath178 corrections , we obtain @xcite @xmath179 our exact result @xmath180 agrees with this prediction reasonably well , thus confirming the simple relation ( 5.14 ) in ref . @xcite .
however , @xmath178 corrections are large and tend to break this agreement .
it is natural to expect that @xmath181 ( and higher ) corrections are also substantial at @xmath177 .
we have calculated the n@xmath182lo corrections to the matching coefficients of heavy
light currents in hqet .
our result takes into account effects due to the mass of the charm quark . strictly speaking
, our results should be used together with the n@xmath182lo @xmath132-function @xcite and anomalous dimensions of both the qcd currents and the hqet one .
the four - loop anomalous dimensions are known for some of the qcd currents ( for @xmath183 , @xmath121 the anomalous dimension is exactly zero ; for @xmath96 it is just the anomalous dimension of the @xmath7 mass @xcite with a minus sign ) .
however , the four - loop anomalous dimension of the hqet current is not known ( this anomalous dimension does not depend on the dirac structure @xmath9 ) .
( without dynamic @xmath78 quarks ) , then , in addition to hqet evolution without @xmath78 quark ( @xmath184 ) and with it ( @xmath185 ) , one has to know the decoupling relation at @xmath186 .
it is known with three - loop ( n@xmath182lo ) accuracy @xcite . ]
the effect of this unknown anomalous dimension cancels in ratios of @xmath0-meson decay constants , @xmath169 and @xmath187 , discussed in sect .
[ s : meson ] .
all the previous experience shows that contributions from anomalous dimensions are numerically much smaller than from matching coefficients . matching coefficients have renormalon singularities at the borel parameter @xmath188 , this is the position closest to the origin out of all possible ones
; this means that the factorial growth of their perturbative coefficients is fastest among all possible variants . on the other hand
, it is generally believed that anomalous dimensions have no renormalon singularities , and their perturbative series have finite convergence radii .
only @xmath21 has been measured experimentally @xcite .
our results can be used for predicting the @xmath144 decay constants .
we find that the perturbative series for @xmath169 and @xmath189 converge very slowly at best .
the effects due to the charm - quark mass are small and of the order of @xmath190 .
the matching coefficients @xmath20 can be used for extraction of @xmath0 ( and @xmath144 ) decay constants from lattice hqet simulations ( see ref .
@xcite for recent reviews ) , or from hqet sum rules .
we are grateful to t. van ritbergen for the package ` shell3 ` @xcite .
we thank rainer sommer for communication . this work was supported by the dfg through sfb / tr9 and the graduiertenkolleg `` hochenergiephysik und teilchenastrophysik '' , by nserc and the alberta ingenuity fund .
the feynman diagrams were drawn with the help of ` axodraw ` @xcite and ` jaxodraw ` @xcite
. m. della morte _ et al_. , jhep * 02 * ( 2008 ) 078 [ arxiv:0710.2201 [ hep - lat ] ] ; + m. della morte _
et al_. , proceedings of science ( lattice 2008 ) 226 [ arxiv:0810.3166 [ hep - lat ] ] ; + b. blossier _
et al_. , arxiv:0911.1568 [ hep - lat ] .
larin , j.a.m .
vermaseren , phys .
* b 259 * ( 1991 ) 345 ; + s.a .
larin , in _ quarks-92 _ , ed
. d.yu .
grigoriev , v.a .
matveev , v.a .
rubakov , p.g .
tinyakov , world scientific ( 2003 ) p. 201
[ arxiv : hep - ph/9302240 ] .
s. laporta , e. remiddi , phys .
* b 379 * ( 1996 ) 283 [ arxiv : hep - ph/9602417 ] ; + s. laporta , int .
j. mod .
phys . * a 15 * ( 2000 ) 5087 [ arxiv : hep - ph/0102033 ] .
p. marquard , j.h .
piclum , d. seidel , m. steinhauser , phys .
* b 678 * ( 2009 ) 269 [ arxiv:0904.0920 [ hep - ph ] ] .
e. remiddi , j.a.m .
vermaseren , int .
j. mod .
phys . * a 15 * ( 2000 ) 725 [ arxiv : hep - ph/9905237 ] .
d. matre , comput .
commun .
* 174 * ( 2006 ) 222 [ arxiv : hep - ph/0507152 ] ; arxiv : hep - ph/0703052 . m. neubert , phys .
* d 46 * ( 1992 ) 1076 .
t. van ritbergen , j.a.m .
vermaseren , s.a .
larin , phys .
* b 400 * ( 1997 ) 379 [ arxiv : hep - ph/9701390 ] .
m. czakon , nucl .
* b 710 * ( 2005 ) 485 [ arxiv : hep - ph/0411261 ] .
chetyrkin , phys .
* b 404 * ( 1997 ) 161 [ arxiv : hep - ph/9703278 ] .
vermaseren , s.a .
larin , t. van ritbergen , phys .
* b 405 * ( 1997 ) 327 [ arxiv : hep - ph/9703284 ] .
k. ikado _ et al . _
( belle collaboration ) , phys .
* 97 * ( 2006 ) 251802 [ arxiv : hep - ex/0604018 ] ; + b. aubert _ et al .
_ ( babar collaboration ) , phys .
* d 77 * ( 2008 ) 011107 [ arxiv:0708.2260 [ hep - ex ] ] .
vermaseren , comput .
* 83 * ( 1994 ) 45 .
d. binosi , l. theussl , comput .
* 161 * ( 2004 ) 76 [ arxiv : hep - ph/0309015 ] . | we consider the currents formed by a heavy and a light quark within quantum chromodynamics and compute the matching to heavy quark effective theory to three - loop accuracy . as an application
we obtain the third - order perturbative corrections to ratios of @xmath0-meson decay constants .
ttp09 - 41 + sfb / cpp-09 - 110 + alberta thy 16 - 09 , , , , , |
the investigation of heavy quarks production in high energy hadron collisions provides a method for studying the internal structure of hadrons .
some problems are the same in hadroproduction and photo-/electroproduction processes .
so , the review of the situation of the heavy quark hadroproduction can be useful for the interpretation of hera data . in this talk we present a short review of heavy quark hadroproduction .
the theoretical predictions are usually obtained in the nlo parton model @xcite .
the assumptions which are used for simplifications of the computations are considered in sect .
2 . in the case of one - particle distributions
even lo ( @xmath0 ) parton model with collinear approximation is enough for the data description , nlo contributions ( @xmath1 ) only change the normalizations . on the other hand ,
in the case of two - particle distributions , see sect .
3 , the collinear approximation has failed , and it is necessary to account for the transverse momenta of the incident partons . the possibility to include the transverse momenta of the incident partons in the framework of semihard theory @xcite , where the virtualities and polarizations of the gluons are taken into account , is considered in sect .
4 . in ref .
@xcite we presented the results for main and simplest subprocess , @xmath2 for hadroproduction , and @xmath3 for photo- and electroproduction .
the conventional nlo parton model expression for the heavy quark hadroproduction cross sections has the factorization form @xcite : @xmath4 where @xmath5 and @xmath6 are the structure functions of partons @xmath7 and @xmath8 in the colliding hadrons @xmath9 and @xmath10 , @xmath11 is the factorization scale ( i.e. virtualities of incident partons ) and @xmath12 is the cross section of the subprocess which is calculated in perturbative qcd . the last cross section can be written as a sum of lo and nlo contributions , @xmath13 , where @xmath14 is the renormalization scale , and @xmath15 as well as @xmath16 depend practically only on one variable @xmath17 , @xmath18 .
the expression ( 1 ) corresponds to the process shown schematically in fig . 1 with @xmath19 the main contribution to the cross section at small @xmath20
is known to come from gluon - gluon fusion , @xmath21 .
heavy quark production in parton model .
the principal uncertainties of any numerical qcd calculation of heavy flavour production are connected with the unknown values of the parameters : both scales , @xmath11 and @xmath14 , and the exact value of heavy quark mass , @xmath22 .
the values of both scales should be of the order of hardness of the considered process , however nobody can say what is better to use for scales , @xmath22 , @xmath23 or @xmath24 .
the phenomenological parton densities are sometimes ( at very small @xmath20 ) in contradiction @xcite with the general properties of perturbative qcd .
however it is just the region that dominates in the heavy quark production at high energies = 1.4gev , at lhc , @xmath25 = 14 tev , the product @xmath26 of two gluons ( both @xmath27 and @xmath28 are the integral variable ) is equal to @xmath29 . ] .
another problem of parton model is the collinear approximation .
the transverse momenta of the incident partons , @xmath30 and @xmath31 are assumed to be zero , and their virtualities are accounted for only via structure functions ; the cross sections @xmath15 and @xmath16 are assumed to be independent on these virtualities . the nlo parton model calculations of the total cross sections of @xmath32 and @xmath33 production , as functions of the beam energy , for @xmath34 and @xmath35 collisions can be found in @xcite .
these results depend strongly ( on the level of several times ) on the numerical values of quark masses as well as on the both scales , @xmath11 and @xmath14 .
some experimental data are in contradiction with each other , however generally they are in agreement with nlo parton model predictions
. + fig .
@xmath36-distributions for @xmath37 at @xmath25 = 1.8 tev , for different values of rapidity .
the nlo contributions to one - particle distributions lead only to renormalization of lo results , practically without correction of the shapes of a distributions @xcite .
it means that instead of more complicate calculation of @xmath36 , or rapidity distributions , in nlo , it is enough to calculate them in lo , and multiply after by k - factor @xmath38 which can be taken , say , from the results for total production cross sections .
the comparison of lo + nlo calculations with lo multiplied by k - factor is presented in fig .
2 taken from ref . @xcite .
the values of k - factors and their energy and scale dependences for several sets of structure functions were calculated in refs .
@xcite .
the experimental data for @xmath39-distributions of d - mesons produced in @xmath40 interactions @xcite are in agreement with the parton model distributions for bare quarks , as one can see in fig .
3 taken from @xcite .
it means that the fragmentation processes are not important here , or they are compensated by , say , recombination processes .
the shape of @xmath39-distributions does not depend practically on the mass of @xmath41-quark .
3 . experimental @xmath39 distributions for @xmath42 mesons , compared to the nlo parton model prediction for charm quarks .
the data on one - particle @xmath36-distributions , including the hadronic colliders data for the case of beauty production , also can be described by the nlo parton model , see @xcite .
the azimuthal angle @xmath43 is defined as an opening angle between two produced heavy quarks , projected onto a plane perpendicular to the beam . in the lo parton model
this angle between them is exactly @xmath44 . in the case of nlo parton model
a distribution over @xmath43 angle appears @xcite .
the investigation of such distributions is very important . in one - partricle distributions ,
the sum of lo and nlo contributions of the parton model practically coinsides with the lo contribution multiplied by @xmath45-factor .
so we can not control the magnitudes of lo and nlo contributions separately . in the case of azimuthal correlations
all difference from the trivial @xmath46 distribution comes from nlo contribution .
the experimental data on azimuthal correlations are claimed ( see @xcite and refs .
therein ) to be in disagreement with the nlo predictions , for the cases of charm pair hadro- and photoproduction at fixed target energies . the level of disagreements can be seen in fig .
4 ( solid histograms ) taken from ref . @xcite .
these data can be described @xcite , assuming the comparatively large intrinsic transverse momenta of incoming partons ( @xmath47 kick ) . for each event , in the longitudinal centre - of - mass frame of the heavy quark pair , the @xmath48 system is boosted to rest
. then a second transverse boost is performed , which gives the pair a transverse momentum equal to @xmath49 ; @xmath50 and @xmath51 are the transverse momenta of the incoming partons , which are chosen randomly , with their moduli distributed according to @xmath52 fig .
azimuthal correlation for charm production in @xmath40 collisions : nlo parton model and @xmath47 kick calculations versus the wa75 and wa92 data .
the dashed and dotted histograms in fig . 4 correspond to the nlo parton model prediction , supplemented with the @xmath47 kick , with @xmath53 gev@xmath54 and @xmath55 gev@xmath54 , respectively .
we see that with @xmath56 gev@xmath54 it is possible @xcite to describe the data .
however the large intrinsic transverse momentum significantly changes one - particle @xmath36-distributions of heavy flavour hadrons , which were in good agreement with the data .
the solid curves in fig . 5 taken from @xcite represent the nlo parton model predictions for charm quarks @xmath36-distributions which are in agreement with the data .
the effect of the @xmath47 kick results in a hardening of the @xmath57 spectrum . on the other hand , by combining the @xmath47 kick with @xmath58 gev@xmath54 and the peterson fragmentation @xcite , the theoretical predictions slightly undershoot the data ( dot - dashed curves ) . fig .
5 . charm @xmath57 distribution measured by wa92 and e769 , compared to the nlo parton model predictions , with and without the non - perturbative effects .
the @xmath47 kick can only very weakly change the @xmath39-distributions of produced @xmath41-quarks , fig . 3 , and after accounting the fragmentation these distributions can become too soft .
let us consider why the conventional nlo parton model with collinear approximation works reasonably for one - particle diustributions , and , at the same time , it is in disagreement with the data on azimuthal correlations .
the contribution of the processes of fig .
1 , which governs the heavy quark production can be written as a convolution of initial transverse momenta distributions , @xmath59 and @xmath60 , with squared modulo of perturbative qcd matrix element , @xmath61 : @xmath62 now there are two possibilities : + i ) the essential values of initial transverse momenta are much smaller than the transverse momenta of produced heavy quarks , @xmath63 , and + ii ) all transverse momenta are of the same order , @xmath64 . in lo for the first case ( i ) we can wait that one - dimentional @xmath36 distributions should be more broad than the distributions on the transverse momenta of the quark pair , because in lo @xmath65 .
nlo gives here a correction , numerically not very large , see fig .
2 . in this case one can replace both the initial distributions @xmath66 by @xmath67-functions , @xmath68 .
this reduces the expression ( 5 ) to the very simplified one : @xmath69 in total agreement with weizsaecker - williams approximation in qed . in the second case ( ii )
the distributions on the transverse momenta of the quark pair should be the same , or even more broad than the one - particle @xmath36-distributions , and namely this situation is realised @xcite . in this case
we can not wait a priory that the weizsaecker - williams approximation will give good results , however it works quite reasonably in the case of one - particle distributions .
in the case of distribution on the transverse momentum of the heavy - quark pair , @xmath70 we measure ( only approximately in nlo ) the distribution over the sum of transverse momenta of incident gluons . in this case
it seems to be senseless to replace the distributions @xmath59 and @xmath60 by @xmath67-functions , and to expect a reasonable agreement with the data .
the same can be said about the azimuthal correlations .
the @xmath71 kick @xcite effectively accounts for the transverse momenta of incident partons .
it uses the expression which can be written symbollically as @xmath72 and the main difference from the general qcd expression eq .
( 5 ) is that due to absence of @xmath30 in the matrix element the values of @xmath73 in eq .
( 4 ) should be different for different processes and kinematical regions .
the reason is that in eq .
( 5 ) the values @xmath66 decrease at large @xmath74 as a weak power ( see next sect . ) , i.e. comparatively slowly , and more important is the @xmath74 dependence of the matrix element . in the last one
the corrections of the order of @xmath75 , where @xmath76 is the qcd scale , are small enough when @xmath77 and they start to suppress a matrix element value when @xmath78 .
let us consider another approach , when the transverse momenta of incident gluons in the small-@xmath20 region appear from the diffusion of transverse momenta in the gluon evolution - factorization formulae can be found , for example , in refs . @xcite . ] .
this diffusion is described by the function @xmath79 determined @xcite as @xmath80}{d q^2 } \;,\ ] ] where @xmath81 is usual gluon structure function . in principle
the function @xmath79 which determine the probability to find gluon with fixed value of transverse momentum , @xmath82 , depends on three variables , @xmath20 , @xmath82 and gluon virtuality @xmath83 .
however at small @xmath20 in lla @xmath84 , and it leads to comparatively weak dependence of @xmath79 on @xmath85 ( strongly different from exponential dependence in eq .
( 4 ) ) due to weak @xmath83-dependences of phenomenological structure functions .
the exact expression for gluon @xmath82-distributions can be obtained , as a solution of the nonlinear evolution equation .
the calculations @xcite result in difference from our @xmath79 function only about 10 - 15% .
the matrix element @xmath86 accounting for the gluon virtualities and polarizations is much more complicate than the parton model one .
that is why we consider only lo contribution of the subprocess @xmath87 .
the differential cross section of heavy quarks hadroproduction has the form @xmath88 @xmath89 here @xmath90 and @xmath91 are the quarks rapidities in the hadron - hadron c.m.s .
( 9 ) enables to calculate straightforwardly all distributions concerning heavy flavour one - particle , or pair production .
however there exists a problem coming from infrared region .
gluon structure function in eq .
( 8) is not determined at small virtualities , so the function @xmath92 is unknown at the small values of @xmath93 and @xmath94 . to solve this problem
we will use the direct consequence of eq .
( 8) @xcite @xmath95 and rewrite @xcite the integrals in the eq .
( 9 ) as the sum of four contributions .
the first one is determined by the product of two gluon distributions , @xmath96 and @xmath97 , and it is the same as the conventional lo parton model expression .
next three terms contain the corrections to the parton model .
if the initial energy is not high enough , the first term dominates . in the case of very high energy
the first term can be considered as a small corrections , and our results are differ from the conventional ones . in the cases when the collinear approximation is available ,
our results only slightly differ from the parton model .
6 . cross section of beauty production in cdf .
it is illustrated in fig .
6 taken from @xcite , where we compare our calculations of @xmath10-quark @xmath36-distributions with the experimental results of cdf collaboration and with two variants of parton model calculations .
the distributions over the azimuthal angle @xmath43 can be found in @xcite .
the essential values of @xmath98 and @xmath99 in our calculations increase with increase the value of @xmath100 of detected @xmath10-quark . in the language of @xmath47
kick it means that the values of @xmath73 will be also increased .
the experimental results on total cross sections for charm and beauty production are in agreement with the conventional parton model predictions , using reasonable values of qcd scales and quark masses .
the data on @xmath39 and @xmath36 distributions are also in reasonable agreement with parton model without any fragmentation functions distributions of charm photoproduction measured by e691 coll .
@xcite are more hard than the nlo parton model predictions @xcite , that can be considered @xcite as an argument for including a non - perturbative fragmentation function
. however , the @xmath36 slope of the calculated spectrum depends strongly on the charm quark mass , and the measured charm production cross section has very strange energy dependence @xcite . ] . moreover , the shapes of one - particle lo and nlo distributions practically coinside .
it means that instead of calculation the nlo contributions , it is enough to calculate only lo contributions , and rescale them using k - factor taken , say , from the calculated ratio of total cross sections . in the case of distribution over the total transverse momentum of the produced quark pair , or azimuthal correlations , the conventional nlo parton model with collinear approximation
can not describe the data .
the @xmath47 kick @xcite allows one to describe these data , however the problems with one - particle @xmath36-distributions appear , which can be solved by introducing the fragmentation function .
the last way should produce the problems in description of @xmath39-spectra .
moreover , it seems that the @xmath73 values should depend on the process and the kinematical regions .
another possibility to solve the problems of initial transverse momenta is to use semihard theory , accounting for the virtual nature of the interacting gluons , as well as their transverse motion and different polarizations .
it results in a qualitative differences with the lo parton model predictions @xcite . in ref .
@xcite the values of @xmath101 were calculated using phenomenological gluon structure functions , and the infrared contributions to @xmath101 were investigated in details .
the possible estimations of the shadow corrections in the processes of heavy flavour production can be found in refs .
@xcite .
i am grateful to m.g.ryskin and a.g.shuvaev for multiple discussions , to m.l.mangano for very useful critical comments and to e.m.levin who participated at the early stage of this activity .
i thank the organizing commitee of hera monte carlo workshop for financial support .
this work is supported by grant nato outr.lg 971390 .
99 p.nason , s.dawson and r.k.ellis . nucl.phys .
b303 ( 1988 ) 607 ; b327 ( 1988)49 .
l.v.gribov , e.m.levin and m.g.ryskin .
phys.rep . 100 ( 1983 ) 1 ; e.m.levin and m.g.ryskin . phys.rep . 189
( 1990 ) 267 .
m.g.ryskin , yu.m.shabelski and a.g.shuvaev .
c69 ( 1996 ) 269 .
j.c.collins , d.e.soper and g.sterman . nucl.phys .
b308 ( 1988 ) 833 .
ya.i.azimov , yu.m.shabelski and o.p.strogova .
57 ( 1994 ) 715 .
s.frixione , m.l.mangano , p.nason and g.ridolfi .
preprint cern - th/97 - 16 ( 1997 ) ; hep - ph/9702287 .
i.sarcevic , p.carruthers and q.gao .
d40 ( 1989 ) 3600 .
liu wenjie , o.p.strogova , l.cifarelli and yu.m.shabelski .
57 ( 1994 ) 900 .
m.i.adamovich et al .
wa92 coll .
b495 ( 1997 ) 3 .
g.a.alves et al .
e769 coll .
77 ( 1996 ) 2392 . beat coll .
, m.adamovich et al . phys .
b348 ( 1995 ) 256 .
c.peterson , d.schlatter,i.schmitt and p.zerwas .
d27 ( 1983 ) 105 .
s.catani , m.ciafaloni and f.hautmann .
b242 ( 1990 ) 97 ; nucl.phys .
b366 ( 1991 ) 135 . j.c.collins and
r.k.ellis . nucl.phys .
b360 ( 1991 ) 3 .
g.marchesini and b.r.webber . nucl.phys . b386 ( 1992 ) | the conventional nlo parton model is enough for the description of total cross sections and one - particle distributions . in the case of two - particle correlations ,
the collinear approximation has failed , and it is necessary to account for the transverse momenta of initial partons . the different possibilities to do this are discussed . * the role of incident parton transverse momenta in heavy quark hadroproduction * + yu.m.shabelski + petersburg nuclear physics institute , + gatchina , st.petersburg 188350 russia + e - mail : [email protected] + |
throughout the paper , by _ covering _ of a banach space @xmath0 we mean a family @xmath1 of proper subsets of @xmath0 ( @xmath2 any set of indices ) such that @xmath3 . by _ body _ in @xmath0 , we mean a nonempty proper subset of @xmath0 that is contained in the closure of its connected interior .
a covering of @xmath0 is called a _ tiling _ of @xmath0 whenever its members ( _ tiles _ ) are closed bodies with pairwise disjoint interiors . a covering @xmath4 of @xmath0
is said to be _ point - finite _ if no point of @xmath0 belongs to infinitely many members of @xmath4 ; the ( possibly infinite ) _ order _ of @xmath4 is the supremum of those @xmath5 in @xmath6 such that there exist @xmath5 members of @xmath4 with a common point .
a covering @xmath4 of @xmath0 is said to be _ locally finite _ if every point in @xmath0 has a neighborhood that meets only finitely many members of @xmath4 ( equivalently , if every compact subset of @xmath0 meets only finitely many members of @xmath4 ) .
it is easy to produce examples of point - finite coverings ( even by convex bodies , even tilings ) that are not locally finite : for an exhaustive discussion of this topic in the general setting of spaces of any dimension see the nice paper @xcite by v. klee ; see also @xcite and @xcite . a great contribution to the study of coverings of infinite - dimensional banach spaces
was given in 1961 by h. h. corson in his classical paper @xcite , motivated by topological reasons .
the main result of that paper states that , if a banach space @xmath0 contains some infinite - dimensional reflexive subspace , then @xmath0 admits no locally finite covering by bounded convex sets . such a result has been recently improved in several directions by v. p. fonf and the second author .
in fact in @xcite it is proved that , if every compact subset of @xmath0 meets only finitely many members of some covering of @xmath0 by bounded @xmath7-closed subsets , then @xmath0 is @xmath8-saturated . moreover , in order that there exist a ( algebraically ) finite - dimensional compact set that meets infinitely many members of any covering @xmath4 of @xmath0 by closed convex bounded ( in short ccb ) sets , it is enough for @xmath0 to contain an infinite - dimensional separable dual space ( see @xcite ) . in particular ,
in this case even a segment exists that meets infinitely many members of @xmath4 whenever the members of @xmath4 are ccb rotund or smooth bodies ( see @xcite ) , or simply ccb bodies if @xmath4 is a tiling and @xmath0 itself is ( infinite - dimensional ) reflexive ( see @xcite ) . despite these results ,
an interesting question asked by corson in @xcite still remains unanswered . after proving his theorem
, he essentially asks whether some infinite - dimensional reflexive space exists that admits a point - finite covering by ccb bodies .
note that , without any assumption on the infinite - dimensional banach space @xmath0 , even locally finite tilings by ccb bodies can be exhibited , like the `` lattice '' tiling of @xmath8 by suitable translates of the unit ball ( see @xcite for a significant characterization of the separable banach spaces admitting such tilings as those being isomorphically polyhedral ) .
moreover , in special non - separable spaces even ( not locally finite ) tilings of order 1 can be found by ccb bodies : see the surprising construction given by v. klee in @xcite of a tiling of @xmath9 for suitable @xmath10 by pairwise disjoint translates of the closed unit ball .
the aim of this paper is to answer corson s question in the affirmative .
in fact we prove the following * theorem * _ every banach space @xmath0 admits a tiling of order 2 by closed convex bounded bodies . _
our proof is obtained by combining in a suitable way two main ideas .
the first one allowed v. p. fonf , a. pezzotta and the second author to prove in @xcite that any banach space can be tiled by ccb bodies . via the second one ,
a. h. stone constructed in @xcite a tiling of order 2 of @xmath11 , @xmath5 any natural number , by ccb bodies . throughout the paper we use standard geometry of banach spaces notation as in @xcite .
all the banach spaces under consideration are assumed to be real .
for finite - dimensional spaces a construction can be found in @xcite ( the bodies produced there are also uniformly bounded ) , so we can work only in the infinite - dimensional setting .
@xmath0 is a subset of @xmath17 which is @xmath15norming for some @xmath18 .
moreover , @xmath19 is the smallest cardinal number @xmath20 such that there exists a norming set @xmath21 for @xmath0 with @xmath22 .
* step 1 * let us begin producing a special covering @xmath27 by pairwise disjoint convex bounded bodies of @xmath28 , @xmath10 any infinite set . given any family @xmath29\ } _ { \nu \in \gamma } $ ] of bounded closed non trivial real intervals indexed on @xmath10 with @xmath30 , we say that the ccb body following stone ( see @xcite ) , if in ( [ box ] ) exactly for one value of @xmath33 the closed interval @xmath34 $ ] is replaced by the left - open interval @xmath35 $ ] , we say that the corresponding set @xmath36 is a lidless box " .
our construction of @xmath27 is totally inspired by the work that was already done in proving proposition 1.6 in @xcite . practically , we repeat that construction just replacing boxes by lidless boxes .
it is worthwhile to present it again , since it will be used in step 2 and some more work on it will be done in step 3 .
a picture giving an idea of how the construction works can be found in that paper .
the collection @xmath42 covers @xmath43 and its members are pairwise disjoint .
in fact , trivially any point in @xmath44 belongs to no member of @xmath27 different from @xmath45 .
moreover , let @xmath46 with @xmath47 and * step 2 * let us show how the covering @xmath27 that we have built in step 1 plays a crucial role in providing a covering @xmath59 by pairwise disjoint convex bounded bodies for any banach space .
we need to recall here under the chief headings what has been done in @xcite , sect .
2 , where detailed proofs are available . let @xmath21 be a norming set for @xmath0 with @xmath63 ; passing to the equivalent norm @xmath64 , we may assume that @xmath21 is @xmath65norming . with the aid of zorn s lemma , for some ordinal @xmath66 with @xmath67 we construct a totally ordered set of pairs @xmath68 ( which , in some sense , apes a biorthogonal system ) such that now , let us consider the covering @xmath27 of @xmath43 that has been constructed in step 1 .
it is obvious that @xmath80 has nonempty interior relative to @xmath81 .
moreover , it can be easily seen ( see @xcite ) that , for fixed @xmath82 and @xmath83 , the point @xmath84 is an interior point of @xmath85 , so it is an interior point of @xmath86 relative to @xmath81 too .
* step 3 * following a. h. stone ( see @xcite ) , we now produce a refinement @xmath90 of @xmath27 of order 2 , that turns out to be a tiling of @xmath43 by ccb bodies . to do that , it is enough to express each lidless box @xmath91 as a countable union of boxes , in such a way that any point of @xmath92 belongs to at most two of them .
* step 4 * finally , we have only to follow the same procedure used in step 2 , just replacing @xmath27 by @xmath103 when defining the isomorphic embedding @xmath104 .
it remains only to prove that , for any value of @xmath105 , also the ccb body @xmath106 has nonempty interior relative to @xmath81 .
this is clear for @xmath107 . for @xmath108 , it is easy to show that the segment @xmath36 of @xmath81 , having the origin and @xmath109 as its endpoints , meets the interior of @xmath110 relative to @xmath81 .
in fact set * remark * clearly our construction can not lead anyway to a tiling @xmath4 uniformly bounded from above or from below , i.e. we can not get that the members of @xmath4 are uniformly bounded or that there exists some positive @xmath118 such that all of them contain some ball of radius @xmath118 .
so , while the choice of coefficients @xmath119 in @xcite was suggested by the possibility to get all members of @xmath4 uniformly bounded from below , in our step 1 it has been made just in order to refer quickly to that paper for those proofs that have been omitted here . it is worthwhile to notice that in @xcite ( prop .
2.8 ) it is suggested the way to obtain , in any banach space with the radon - nykodym property , a tiling uniformly bounded from above just applying a cutting procedure to each member of a given tiling in a straightforward transfinite way .
this can obviously be done on the members of our tiling @xmath4 .
after that , to each slice that we have obtained , we can apply a new cutting procedure as described in step 3 , just referred to the linear continuous functional we used to produce the slice .
v. p. fonf and j. lindenstrauss : _ some results on infinite - dimensional convexity _ , israel j. math .
* 108 * ( 1998 ) , 1332 .
v. p. fonf , a. pezzotta and c. zanco : _ tiling infinite - dimensional normed spaces _ , bull .
london math .
* 29 * ( 1997 ) , 713 - 719 .
j. lindenstrauss : _ weakly compact , their topological properties and the banach spaces they generate _ , symposium on infinite - dimensional topology , ann . of math
* 69 * ( ed .
r. d. anderson , princeton university press , princeton , nj , 1973 ) , 235 - 273 .
w. b. johnson and j. lindenstrauss : _ basic concepts in the geometry of banach spaces _ , handbook of the geometry of banach spaces , volume 1 , w.b .
johnson and j. lindenstrauss editors , north - holland elsevier , 2001 . | we answer in the affirmative the following question raised by h. h. corson in 1961 : `` is it possible to cover every banach space x by bounded convex sets with nonempty interior in such a way that no point of x belongs to infinitely many of them ? ''
actually we show the way to produce in every banach space x a bounded convex tiling of order 2 , i.e. a covering of x by bounded convex closed sets with nonempty interior ( tiles ) such that the interiors are pairwise disjoint and no point of x belongs to more than two tiles . to appear in israel j. math .
_ 2000 mathematics subject classification _ : 46b20 , 52a45 . _
key words and phrases _ : coverings of banach spaces , point - finite coverings , tilings . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.