article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the presence of long range memory in complex systems has been studied in the literature in a wide variety of fields with a hope of better understanding the dynamics of these systems and to make their evolution more predictive . in the simplest case of stationary and monofractal time
series the memory effects are defined with the use of two point autocorrelation function @xmath6 , where @xmath7 , @xmath8 are data in series , @xmath9 is its length , @xmath10 is an average taken over all data in the given signal and @xmath11 are series increments .
it is well known that @xmath12 reveals in this case the power law form in an asymptotic limit of large time lags @xmath13 @xcite @xmath14 where the scaling exponent @xmath4 @xmath15 is related with the level of memory in signal and describes at boundaries fully autocorrelated ( persistent ) signal for @xmath16 or completely uncorrelated increments for @xmath17 .
a direct calculation of correlation function and @xmath4 exponent gets more difficult in real signals when a noise or nonstationarities are present in data .
therefore , to avoid this problem , an alternative approach is often made where the fluctuations in the cumulated data , i.e. fluctuations in the series @xmath18 @xmath19 , are considered .
one measures the scaling hurst exponent @xmath20 @xcite of @xmath21 related to @xmath4 according to @xcite @xmath22 this way , any measurement of @xmath20 ( @xmath23 ) is mapped to the corresponding @xmath4 values .
the most prominent example of such indirect analysis of memory effects is the detrended fluctuation analysis ( dfa ) @xcite .
the scaling law from eq.(1 ) is replaced within dfa by the power law @xmath24 where @xmath25 is the averaged fluctuation of the signal around its local trend in time windows of fixed length @xmath26 .
one has in detailed quantitative description : @xmath27 where @xmath28 here @xmath29 $ ] stands for the number of non - overlapping boxes for which the detrending procedure is performed and @xmath30 is the polynomial trend fitted to initial data in @xmath31-th time window and then subtracted from these data .
an application of dfa is limited to systems whose memory properties are well described by a single scaling exponent .
such an approach is insufficient in so called multifractal systems @xcite@xcite , where small and large fluctuations follow different scaling laws ( see , e.g. @xcite@xcite ) .
one has to employ more general method of description in the latter case and a full understanding of issues connected with memory properties in multifractal data is still under debate .
the multifractal properties of time series are now extensively studied because of the omnipresence of multifractals in various phenomena in nature .
its presence is confirmed in empirical data from turbulence @xcite , astronomy @xcite , climate phenomena @xcite , physiology @xcite , text structure @xcite , physics @xcite up to finances @xcite@xcite , not covering many other publications on these phenomena .
it is worth to stress that the answer to questions regarding accuracy , applicability and reliability of multifractal measurements is crucial for the development of efficient simulation or forecasting methods @xcite .
one of the most frequently used techniques to quantify the multifractal properties in data series is multifractal detrended fluctuation analysis ( mf - dfa ) @xcite .
the mf - dfa method enables to eliminate polynomial trends in data what makes it superior to other multifractal oriented techniques like the structure function analysis @xcite or the higher - order autocorrelation functions @xcite .
the popular method to study multifractal properties is also the wavelet transform modulus maxima ( wtmm ) approach @xcite .
wtmm eliminates polynomial trends similarly to mf - dfa but is considerably more difficult to implement .
moreover , it may give biased outcomes and yields spurious multifractality more often @xcite .
the mf - dfa has been applied so far in diversified scientific problems , just to mention : seismology @xcite , cosmology @xcite , biology @xcite , meteorology @xcite , medicine @xcite , music @xcite , geophysics @xcite , and finances @xcite@xcite far from being exhaustive here .
let us remind that the main ingredient of mf - dfa is the @xmath32-weighted fluctuation @xmath33 of the time series signal around its local trend in a time window of size @xmath26 .
more precisely , it is defined as an @xmath32-th moment ( @xmath34 ) of @xmath25 already used in dfa , i.e. @xmath35^{q/2}\right\}^{1/q}\ ] ] for @xmath36 , and @xmath37\right\}\ ] ] for @xmath38 . the power law relation @xmath39 is expected in mf - dfa which forms a basis for the so called generalized hurst exponent @xmath40 @xcite .
eq.(8 ) coincides with eq.(3 ) for @xmath41 , hence @xmath42 returns the main hurst exponent value .
the multifractal properties may alternatively be presented as the multifractal spectrum @xmath43 @xcite , known also as hlder description .
both descriptions are linked together via relations @xcite @xmath44 one usually calculates the strength of multifractality by the spread of @xmath40 profile , i.e. @xmath45 , ( @xmath46 ) or by the width of multifractal spectrum @xmath47 , i.e. @xmath48 . both quantities coincide in the limit @xmath49 . in the case of stationary signal
the @xmath40 profile has a monotonic behavior @xcite and the spectrum @xmath47 has a form of inverted parabola .
its spread corresponds to the @xmath1 value . through this article two above descriptions
will be used .
it is worth to stress that multifractality is an interesting and perspective phenomenon for further investigation and applications . on the other hand ,
it is very subtle and generally difficult task to distinguish in many real situations between true multifractal properties of a system and the apparent multifractal phenomena resulting directly from the records based on mf - dfa .
recently , some new features connected with the so called spurious and corrupted multifractality were shown @xcite@xcite .
it turns out that the presence of additive white or color noise , short - term memory or periodicities in multifractal signals may significantly change its observed multifractal properties for all data lengths .
such contamination by additions of disturbances are typically found in many observational records in various complex systems : in physics , finance , physiology , geology , climate dynamics , network traffic , etc .
all real series contain also finite sets of data .
due to smaller statistics in short signals , the accidental fluctuations which are not related with multifractal properties contribute also to fluctuation function @xmath50 and may even dominate over fluctuations related with multifractality . on the other hand , the influence of large fluctuations in short time series can be suppressed since mf - dfa algorithm may not distinguish them from longer trends which are eliminated . in both cases
the observed @xmath0 spread is increased comparing with the corresponding value obtained for much longer signals .
this raises a question about confidence for multifractal findings which are naturally affected by finite size effects in autocorrelated or uncorrelated data even if they are not contaminated by the additional effects mentioned before .
the nonlinear transformations of primary data also significantly affect the registered multifractal properties usually magnifying them .
therefore , a link between true multifractal picture of the complex system and the observed quantitative properties which may suggest the presence of multifractality , even if it is absent in the system , is still an intriguing and not fully understood issue . it will obviously depend on the technique used to quantify multifractal phenomena . in this paper
we use mf - dfa as the commonly accepted technique to find multifractal properties of time series .
our goal is a deeper investigation of finite size effects ( fses ) influencing within mf - dfa method the observed multifractal features of time series .
the finite size effects in multifractality were mentioned already in ref . @xcite . since then several authors approached this problem ( see e.g. refs .
@xcite ) , however general search of this phenomenon has not been yet investigated .
the particular emphasis will be given to amplification of an apparent observed multifractal effects caused by the presence of long memory in signals and by basic transformations done on initial monofractal data .
these effects should be distinguished from the spurious or corrupted multifractality @xcite , where the given data series contaminated by various effects like noise , short - term memory , periodicities in signal , etc . , change the shape of multifractal profile @xmath40 or @xmath47 spectrum .
the latter effects will affect multifractality for all lengths of data @xmath9 , while fse should disappear if @xmath51 .
note , that one always deals in practice with finite set of data , so one should exactly know the level of confidence for results obtained within mf - dfa if they indicate a multifractality in the system , i.e. @xmath3 or @xmath52
. the preliminary quantitative analysis of this problem had already been made in our conference paper @xcite .
multifractal finite time series will contain the multifractal fse bias ( i.e. , the threshold spread of generalized hurst exponent values not related with multifractal properties of these series ) at least like a lot less complicated monofractal series of similar length and persistency level .
we will deal with simpler , involving fewer parameters and more clear for synthetic generation monofractal data that will serve as a tool to find the minimal multifractal fse bias threshold in all kind of signal .
we start with description of fse in monofractal series with the analysis of generalized hurst exponent @xmath40 . then this description is extended to an alternative approach of multifractal spectra in terms of @xmath1 in section 3 .
both sections provide detailed analysis of threshold values @xmath0 and @xmath1 below which the multifractal properties of complex system are only apparent .
the corresponding fse multifractal phenomena for linearly transformed data are discussed in section 4 .
an interesting and important from practical point of view phenomenon of multifractal bias generated solely by nonlinear transformations in monofractal data is described in section 5 .
one has to know the strength of of such bias to get the evidence of multifractal properties in primary not transformed data ( and in complex system in general ) if only volatile series of data from such systems are available .
section 6 .
summarizes semi - analytical results obtained in this article and contains final remarks and conclusions .
we start to describe quantitatively , in terms of generalized hurst exponents @xcite , the expected level of multifractal background existing due to the finite - size effects for uncorrelated data as well as for time series with long memory induced by the explicit form of autocorrelations . our approach is based on the fourier filtering method ( ffm ) @xcite that directly shapes the memory level in artificial data by the respective choice of autocorrelation exponent @xmath4 ( see eq.(1 ) ) . in order to check the accuracy of ffm for generation of time series with a given persistency level , we examined this procedure towards replication of the pre - assumed autocorrelation properties in artificially constructed series .
this efficiency is demonstrated in fig.1 .
the scaling resulting from eq .
( 8) was found also very satisfactory ( see fig.2 ) the performance of mf - dfa technique strictly depends on the power law scaling between @xmath53 order fluctuations @xmath33 and the box size @xmath26 ( see eq.(8 ) ) .
an exact extraction of the generalized hurst exponent @xmath40 requires well determined scaling range for the linear fit @xmath54 against @xmath55 , induced by the power law in eq.(8 ) .
the expected power law dependence for various lengths of the signal @xmath9 and for different values of deformation parameter @xmath32 is shown in fig.2 .
the scaling range from @xmath56 till @xmath57 , confirmed by plots in fig.2 , and the second degree polynomial trends @xmath30 were chosen to be used further on .
we considered the ensembles of numerically generated monofractal time series of length @xmath58 , @xmath59 with the pre - assumed autocorrelation exponent value @xmath60 , each containing @xmath61 independent realizations .
thus , the spread of @xmath4 exponents covered uniformly the range @xmath62 .
every obtained quantity has been averaged over such statistical ensemble .
the width @xmath0 of the generalized hurst exponents profile @xmath40 was limited to the edge values @xmath63 .
notice that the range @xmath64 is large enough for checking multiscaling properties of data from large to considerably very small fluctuations ( see fig.3 ) .
the key numerical data are presented in series of plots in figs.4 - 7 , where @xmath65 and @xmath66 , ( @xmath67 ) .
first , the uncorrelated artificial data were studied .
they have been generated from gaussian distribution ( @xmath17 ) , and also independently obtained as the shuffle of autocorrelated signals .
fig.4a shows a dependence of the edge values @xmath68 of the generalized hurst exponents @xmath40 for two different lengths of time series : @xmath69 and @xmath70 chosen to investigate short and long data sets respectively .
these plots not only confirm that shuffling procedure leads to the same results as gaussian distribution does ( what should be expected anyway ) , but also reveal the dependence of @xmath68 values on the data length .
the corresponding plots for the intermediate lengths are not shown for they look qualitatively the same , i.e. @xmath68 does not change with @xmath4 for shuffled data but it does depend on @xmath9 . this way , the spread @xmath0 of generalized hurst exponent versus length of data can be investigated for signals with no memory ( see fig.5a ) .
the data drawn in log - log scale convince about the power law dependence between @xmath71 and the length of data series @xmath9 @xmath72 with some real constants @xmath73 and @xmath74 .
next , keeping @xmath9 fixed , we turned to investigate the edge values for @xmath75 versus the autocorrelation exponent value @xmath4 for monofractal signals with long memory .
the examples of this dependence are shown in fig.6a for chosen lengths @xmath69 and @xmath76 .
the cases of remaining lengths ( not shown due to lack of space ) have been checked by us as well and look similarly .
all of them indicate an existence of the excellent linear relationship between @xmath75 and @xmath4 in the whole range of @xmath4 values .
therefore , we get @xmath77 where the coefficients @xmath78 and @xmath79 are functions of @xmath9 only . the boundary conditions , i.e. the form of @xmath80 and @xmath81 allow to specify the final shape of this relation .
the profile @xmath82 can not be found directly , because one is stuck with singularity in ffm method for @xmath83 .
therefore we made the extrapolation of the fitting lines @xmath84 versus @xmath4 up to the point @xmath85 for all lengths to get the collection of @xmath82 values .
it is shown in fig.6a .
the plot in log - log scale against the length of time series in fig.7a confirms that for fully persistent time series ( @xmath86 ) @xmath82 gets also the power law form : @xmath87 with some constants @xmath88 and @xmath89 to be determined from the fit . inserting eq.(10 ) and
eq.(12 ) into eq.(11 ) , we arrive with the final formula for @xmath90 @xmath91 in practical applications , the shape of the @xmath92 confidence level profile is important for the multifractal bias caused by fse and amplified by the presence of long memory in data
. it shall be given by the same formula but with different coefficients calculated on the basis of @xmath93 uncertainties @xmath94 , @xmath95 associated with the fit of @xmath96 and @xmath97 parameters , where @xmath98 corresponds to boundary conditions at @xmath83 and @xmath17 respectively .
one has to include also the standard deviation @xmath99 resulting from the series statistics .
the formulas given in eq.(10 ) and eq.(12 ) have to be replaced then by @xmath100 with @xmath101 taken for the particular @xmath92 confidence level . introducing the notation @xmath102 and @xmath103 one
gets the @xmath92 confidence level profile for the multifractal fse bias as @xmath104 the results of the best fit , done on the statistics of @xmath61 monofractal time series for all quoted coefficients with uncertainties entering eq.(15 ) and eq.(16 ) are indicated in table 1 . these findings may also be summarized as in fig.8 .
they show fse multifractal thresholds calculated at @xmath92 confidence level for @xmath0 spread in monofractal data series as a function of persistency level indicated by @xmath4 value and for variety of data lengths @xmath9 .
the dotted horizontal lines present thresholds connected entirely with fse , not affected by eventual autocorrelations in data .
the continuous sloped straight lines describe fse multifractal thresholds amplified by the presence of long - term memory in series and predicted by eq.(17 ) .
the true multifractality , related with the existence of variety of scaling exponents in infinite system , may occur for consecutive series of given length only when the spread in generalized hurst exponents enters the area above this line ( at @xmath92 confidence level ) .
this result clarifies and generalized quantitatively the statement from ref.@xcite on the existence of such threshold values . .collected results for coefficients of the fit describing the multifractal fse thresholds in persistent time series according to eqs.(13 ) and ( 17 ) .
the ensemble of @xmath61 simulations of time series was considered . [ cols="^,^,^,^,^,^,^,^",options="header " , ] the summary given by eqs .
( 27 ) and ( 28 ) for the dependence of the @xmath0 spread on persistency level @xmath4 for various lengths of data is presented in fig .
two different behaviors indicated for @xmath4 above and below crossover value are noticeable .
the transition area is visible due to the fact that two crossover values do not coincide ( @xmath105 ) .
in the first part of this paper we have shown qualitatively and quantitatively how the observed multifractal effects arise in monofractal series of finite uncorrelated or persistent data .
this kind of multifractality , called by us multifractal bias of finite size effect , should be clearly distinguished from the true multifractality caused by memory effects dependent on the time scale and thus related to different scaling properties of data at various time scales .
the quantitative findings presented here confirm and push further the study of the generalized hurst exponent spread suggested recently in @xcite , to be not indicative for multifractality if @xmath106 or @xmath107 are obtained .
although , found results refer to clearly synthetic monofractal series , they are important in practical applications , where in general , multifractal data are observed . results of the same methodology are affected in a similar manner for mono- and multifractal data and a similar level of the bias will appear in all multifractal data as well .
we considered two methods of description for multifractality , i.e. the generalized hurst exponent approach and the multifractal spectrum analysis based on the hlder exponent . in both cases the multifractal threshold of finite data was estimated by the spread of generalized hurst exponent or by the width of multifractal spectrum .
the threshold was shown to grow linearly with autocorrelation level in time series and to decay according to the power - law with series length .
we have estimated numerically the level of such apparent multifractal effect and we captured it in simple semi - analytical formulas .
additionally , in the second part of the article , we have discussed the multifractality revealed by the nonlinearly transformed monofractal data . in such a case , the finite size effects amplified by long - range autocorrelations in primary data are accompanied also by the fluctuation clusters arising from such nonlinear transformation of time series .
this last effect changes the scaling properties of various fluctuation sizes and does not vanish in the limit of infinite data length .
one must also keep in mind the influence introduced by nonlinear transformations , especially when only transformed data are searched and the goal are multifractal properties of the primary data series .
the result of multifractal effect can be significantly dimmed in there . in all cases
the ready to use formulas describing amount of multifractal fse threshold or a multifractal modification of time series properties after some simple nonlinear transformations were given .
we provided also results for statistical confidence at 95% level for all formulas , what enables the reliable estimation when they are used in practical applications .
we believe these formulas are general enough to be applied in any area of science when one is questioning the multiscale properties of data .
eventually , the predicted fse multifractal bias should be compared with examples of real multifractal records .
there is a common agreement that multifractality is a characteristic feature for financial markets , so we may first look at index series .
the multifractal features for various price indices were taken from the general study of ref .
@xcite . by applying formulas from eq.(17 ) to parameters of real index series revealed in @xcite
, we easy find that multifractality comes indeed as a result of scaling properties changing with the time scale for many world markets . however , there are markets where the observed multifractal features are generated mainly ( e.g. philippines , taiwan , germany ) or even entirely ( ireland ) by the finite size effects . in the case of philippines ,
taiwan , thailand , germany , spain and greece almost 80% of the observed multifractality in price indices is caused by multifractal fse bias . in ref .
@xcite it is reported that fse ( achieved for finite size surrogate data ) for djia time series counts for @xmath108 at @xmath109 , which agrees with the result of eq .
( 17 ) predicting @xmath110 at @xmath92 confidence level .
regarding other kinds of real data , e.g. in seismology , meteorology or geophysics , an importance of multifractal fse bias is also indisputable . the seismological records from refs .
@xcite suggest multifractal spread of generalized hurst exponent @xmath111 for @xmath112 at @xmath113 ( see fig.3 in @xcite ) or the spectrum width @xmath114 ( see fig.5 in @xcite ) , while the calculated bias from eqs.(17 ) and ( 20 ) counts for more than @xmath115 of this value in case of @xmath116 and @xmath117 in case of @xmath0 .
many records in meteorology and geophysics reveal high nonstationarity , so that authors focus on much narrower range of @xmath32 ( @xmath118 ) then used in this article .
the quoted multifractal quantitative features for records in meteorology ( see @xcite ) are reported very low for humidity , temperature , solar radiation and maximum squall ( @xmath119 ) and more significant for precipitation , wind speed and atmospheric pressure ( @xmath120 , @xmath121 , @xmath122 ) .
however , to make a proper comparison with the fse multifractal bias in such a case , one has to examine how formulas obtained in this article change when lower moment @xmath32 is used for evaluation .
the light on such problem is shed in coming publication @xcite .
thus , the multifractal bias ( or apparent multifractality ) may play significant role in the observed multifractal spread , especially when short and persistent time series are considered . in many cases
, it aggravates the separation between the main multiscaling phenomenon and the background ( bias ) of not multiscaling origin at all .
therefore , one should be especially careful drawing far reaching conclusions from the multifractal analysis based on interpretation of observed @xmath0 or @xmath1 values .
50 m. s. taqqu , v. teverovsky , and w. willinger , fractals 3 , 785 , ( 1995 ) .
fractals in science , springer , 2nd ed . , edited by a. bunde and s. havlin , springer , berlin , ( 1996 )
. h. e. hurst , trans .
( 1951 ) 770
. b. b. mandelbrot , j.r .
wallis , water resour .
res . 5 , no.2 , ( 1969 ) 321
. j. w. kantelhardt , e. koscielny - bunde , h. h. a. rego , s. havlin , a. bunde , physica a 295 ( 2001 ) 441 .
k . peng , s. v. buldyrev , s. havlin , m. simons , h. e. stanley , and a. l. goldberger , phys.rev.e 49 , 1685 ( 1994 ) .
k . peng , s. havlin , h. e. stanley , and a. l. goldberger , chaos 5 , 82 ( 1995 ) .
a. bunde , s. havlin , j. w. kantelhardt , t. penzel , j. h. peter , and k. voigt , phys .
85 , 3736 ( 2000 ) .
t. c. halsey , m. h. jensen , l. p. kadanoff , i. procaccia , b. i. shraiman , phys .
a 33 ( 1986 ) 1141
. h. g. e. hentschel , i. procaccia , physica d 8 ( 1983 ) 435 .
s. ghashghaie , w. breymann , j. peinke , p. talkner , and y. dodge , nature 381 , 767 ( 1996 ) .
r. n. mantegna and h. e. stanley , nature 383 , 587 ( 1996 ) .
b. b. mandelbrot , sci . am .
298 , 70 ( 1999 ) .
k. hu , p. ch .
ivanov , z. chen , p. carpena , h. e. stanley , phys .
e 64 ( 2001 ) 011114 .
z. chen , p. ch .
ivanov , k. hu , h. e. stanley , phys .
e 65 ( 2002 ) 041107 .
j. w. kantelhardt , s. a. zschiegner , e. koscielny - bunde , s. havlin , a. bunde , h. e. stanley , physica a 316 ( 2002 ) 87 . m. h. jensen , l. p. kadanoff , a. libchaber , i. procaccia , j. stavans , phys .
55 ( 1985 ) 2798 .
muzy , e. bacry , a. arneodo , phys .
( 1991 ) 3515 .
f. s. labini , m. montouri , l. pietronero , j. physique iv 8 ( 1998 ) pr-115 pr-118 .
k. ivanova , h.n .
shirer , e.e .
clothiaux , n. kitova , m.a .
mikhalev , t.p .
ackerman , m. ausloos , physica a 308 ( 2002 ) 518532 .
y. ashkenazy , d. r. baker , h. gildor , s. havlin , geophys .
res . lett . 30 ( 2003 )
p. c. ivanov , l. a. n. amaral , a. l. goldberger , s. havlin , m. g. rosenblum , z. r. struzik , h. e. stanley , nature 399 ( 1999 ) 461 .
m. ausloos , generalized hurst exponent and multifractal function of original and translated texts mapped into frequency and length time series , arxiv:1208.6174v1 [ physics.data-an ] c. amitrano , a. coniglio , p. meakin , m. zanetti , phys .
b 44 ( 1987 ) 4974
. h. e. stanley , p. meakin , nature 335 ( 1988 ) 405 .
a. fisher , l. calvet , b. mandelbrot , multifractality of deutschemark / us dollar exchange rates , cowles foundation discussion paper 1166 , 1997 .
k. ivanova , m. ausloos , eur .
j. b 8 ( 1999 ) 665 . k. ivanova , m. ausloos , physica a 265 ( 1999 ) 279 .
m. ausloos , k. ivanova , comp .
commun . 147
( 2002 ) 582 .
matteo , t. aste , m. m. dacorogna , physica a 324 ( 2003 ) 183 .
a. bershadskii , physica a 317 ( 2003 ) 591 .
f. ren , w .- x .
zhou , europhys .
84 ( 2008 ) 68001 .
j. feder , fractals , plenum , new york , 1989 .
barabasi , p. szepfalucy , t. vicsek , physica a 178 ( 1991 ) 17 . j.f .
muzy , e. bacry , a. arneodo , phys .
67 ( 1991 ) 3515 .
p. owicimka , j. kwapie , s. drod , phys .
e 74 ( 2006 ) 016103 .
l. telesca , v. lapenna , m. macchiato , physica a 354 ( 2005 ) 629 l. telesca , v. lapenna , tectonophysics 423 ( 2006 ) 115 m. s. movahed , f. ghasemi , s. rahvar , m. r. r. tabar , phys
e 84 ( 2011 ) 021103 p.
h. figueiredo , e. nogueira jr .
, m. a. moret , s. coutinho , physica a 389 ( 2010 ) 2090 s. dutta , j. stat .
mech . 12 ( 2010 ) p12021 i. t. pedron , journal of physics : conference series 246 ( 2010 ) 012034 f. liao , y .- k .
jan , jour . of rehab .
res . and develop . 48 ( 2011 )
787 g. r. jafari , p. pedram , l. hedayatifar , j. stat . mech . ( 2007 ) p04012 p. owicimka , j. kwapie , i. celiska , s. drod , r. rak , arxiv:1106.2902v1 [ physics.data-an ] f. a. hirpa , m. gebremichael , t. m. over , water resour . res .
46 ( 2010 ) w12529 k. matia , y. ashkenazy , and h. e. stanley , europhys .
61 ( 2003 ) 422 p. owicimka , j. kwapie , and s. drod , physica a 347 ( 2005 ) 626 j. kwapie , p. owicimka , and s. drod , physica a 350 ( 2005 ) 466 j. jiang , k. ma , and x. cai , physica a 378 ( 2007 ) 399 k. e. lee and j. w. lee , physica a 383 ( 2007 ) 65 .
g. lim , s. kim , h. lee , k. kim , and d .- i .
lee , physica a 386 ( 2007 ) 259 .
j. feder , fractals , new york , plenum press ( 1988 ) .
peitgen , h. jrgens , d. saupe , chaos and fractals , @xmath123nd ed .
springer ( 2004 ) .
j. ludescher , m. i. bogachev , j. w. kantelhardt , a. y. schumann , a. bunde , physica a 390 ( 2011 ) 2480 .
a. y. schumann , j. w. kantelhardt , physica a 390 ( 2011 ) 2637 .
d. gulich and l. zunino , physica a 391 ( 2012 ) 4100 .
lux , m. ausloos , market fluctuations i : scaling , multi - scaling and their possible origins , in : a. bunde , j. kropp , h .- j .
schellnhuber ( eds . ) , the science of disaster : scaling laws governing weather , body , stock - market dynamics , springer verlag , berlin , 2002 , pp .
s. drod , j. kwapie , p. owicimka , r. rak , europhys .
88 , 60003 ( 2009 ) w .- x .
zhou , chaos , solitons & fractals , 45 ( 2012 ) 147 d. grech and g. pamua , acta phys . pol . a 121 ( 2012 )
b-34 t. kalisky , y. ashkenazy , and s. havlin , phys .
e 72 , 011913 ( 2005 ) . c. tsallis , quim .
nova 17 ( 1994 ) 468 .
l. zunino , b.m .
tabak , a.figliola , d.g .
prez , m. garavaglia , o.a .
rosso , physica a 387 ( 2008 ) 6558 d. grech , g. pamua , arxiv:1307.3653 [ physics.data-an ] , @xmath124 , @xmath125 are shown in log - log scale for the generated data of length @xmath70 .
the lines present the fit to the desired power law dependence of eq.(1 ) , while error - bars show @xmath93 standard deviation following from the considered statistics of @xmath61 independent realizations .
the output @xmath4 values from the fit are found @xmath126 , @xmath127 , @xmath128 respectively . ]
-deformed fluctuations within mf - dfa .
results are presented for two different lengths of time series @xmath129 , @xmath130 , three autocorrelation parameters @xmath131 , @xmath124 , @xmath132 and @xmath133 ( from bottom to top ) .
all plots confirm the proposed scaling range from @xmath56 till @xmath134 . ] and @xmath1 fse multifractal thresholds for variety of data lengths @xmath135 , @xmath136 and @xmath137 .
the continuous sloped lines describe the minimal threshold for multiscaling to be observed at @xmath92 confidence level for series with autocorrelation exponent @xmath4 .
the respective values of hurst exponent are marked on top axis .
the dotted horizontal lines indicate the corresponding multifractal fse threshold resulting only from finite data length not affected by possible autocorrelations in data . ]
plots of @xmath68 edge values of multifractal profiles are shown for transformed persistent time series of various lengths .
an statistical ensemble of @xmath139 time series generated for each pair of length @xmath9 and persistency level @xmath4 was taken in calculations and the resulting statistical uncertainty is marked for each @xmath68 value . ] | we study quantitatively the level of false multifractal signal one may encounter while analyzing multifractal phenomena in time series within multifractal detrended fluctuation analysis ( mf - dfa ) .
the investigated effect appears as a result of finite length of used data series and is additionally amplified by the long - term memory the data eventually may contain .
we provide the detailed quantitative description of such apparent multifractal background signal as a threshold in spread of generalized hurst exponent values @xmath0 or a threshold in the width of multifractal spectrum @xmath1 below which multifractal properties of the system are only apparent , i.e. do not exist , despite @xmath2 or @xmath3 .
we find this effect quite important for shorter or persistent series and we argue it is linear with respect to autocorrelation exponent @xmath4 .
its strength decays according to power law with respect to the length of time series .
the influence of basic linear and nonlinear transformations applied to initial data in finite time series with various levels of long memory is also investigated .
this provides additional set of semi - analytical results .
the obtained formulas are significant in any interdisciplinary application of multifractality , including physics , financial data analysis or physiology , because they allow to separate the true multifractal phenomena from the apparent ( artificial ) multifractal effects .
they should be a helpful tool of the first choice to decide whether we do in particular case with the signal with real multiscaling properties or not .
@xmath5 * keywords * : multifractality , time series analysis , autocorrelations , multifractal detrended analysis , generalized hurst exponent , complex systems + * pacs : * 05.45.tp , 05.45.df , 89.75.da , 05.40.-a , 89.75.-k , 89.65.gh , 02.60.-x , 89.20-a |
@xmath0 pictoris is the most extensively studied of the young planetary systems discovered in the last decade and a half .
it is a bright southern hemisphere star ( type a5 v ) , located about 19.3 pc distant from the sun , with a systemic radial velocity of 20 km s@xmath12 ( for a review of the @xmath0 pictoris system , see @xcite ) .
it was observed in 1983 by the iras satellite to have a large excess of emission at infrared wavelengths .
this was referred to as the vega - like phenomenon and was identified as arising from an edge - on circumstellar ( hereafter cs ) dust disk , presumed to be associated with planetary formation @xcite .
it was soon determined through absorption spectroscopy that there was cs gas associated with the dust as well .
a large body of evidence has accumulated indicating that there are comet - like bodies present in the @xmath0 pictoris cs disk .
collisions between dust particles are expected to produce submicron fragments which should be expelled from the system by radiation pressure on time scales much shorter than any plausible stellar age .
thus there must be a secondary source of particles ; one model for the production of dust and gas in the cs disk focusses on evaporating comets and is called the orbiting evaporating bodies model ( oeb ) . in this picture , the comets orbit the star at several tens of au , and thus , the @xmath0 pictoris cs disk is a kind of `` gigantic multi - cometary tail with its natural constituents : gas and dust ''
spectra of @xmath0 pictoris show variable redshifted absorption features arising from gas infalling toward the star at high velocities ( and infrequently , blueshifted features as well ) ; these features are best attributed to the evaporation of star - grazing comets , called the falling evaporating bodies scenario ( feb ) @xcite .
also , gas at close to 20 km s@xmath12 ( zero radial velocity relative to the star ) is identified in all observations and is called the stable gas component .
this gas is difficult to understand , as modeling indicates that it should be expelled from the system by radiation pressure ; a continuous source for this gas is required .
neutral carbon and carbon monoxide have been observed in hst - ghrs uv absorption spectra of @xmath0 pictoris ; carbon monoxide is the only molecule detected in the cs disk to date @xcite . since co and can be dissociated and ionized by interstellar uv photons on time scales of the order of 200 years , both must be continuously replenished .
carbon monoxide , in particular , is difficult to reform after dissociation in the @xmath0 pictoris environment .
thus , the presence of these species indicates that a secondary source for this gas should exist , just as for the cs dust @xcite .
@xcite found the column densities of and co to be comparable , around @xmath13 @xmath3 ; since their rates of destruction are also comparable , this was taken as evidence that the is produced by photodissociation of co , which evaporates from comets orbiting at various distances and velocities .
however , the column density was determined from a heavily saturated multiplet and is therefore quite uncertain . in the hopes of further constraining the characteristics of and co in the @xmath0 pictoris system
, we have reinvestigated the transitions observed in the ghrs data , as well as some that were not seen due to the relatively low spectral resolution of ghrs compared with that of the stis high resolution echelle .
hst stis high resolution echelle spectra of @xmath0 pictoris were obtained on 1997 december 6 and 1997 december 19 , covering the wavelength range from 1459 to 2888 in six exposures each day .
table 1 shows the log of observations .
all the absorption features discussed in this paper appear in either the first or second data set listed for each day ( o4g001010/o4g002010 or o4g001020/o4g002020).the data were initially reduced and calibrated using the stsci iraf package _
calstis v1.8_. spectra with a signal - to - noise ratio of around 10 were achieved .
examination of the errors in the flux values showed that the error propagation calculation had not been performed correctly and that thestated errors were too small .
therefore the data were re - calibrated using _
calstis v1.9a _ , correcting the underestimate of the measurement errors .
the particular advantage of this data set over previous comparable ones is the very high spectral resolution achieved .
the instrumental line spread function using the e140h grating is well described by a gaussian with a fwhm of 1 pixel , corresponding to fwhm = @xmath14/220,000 @xcite . however , since the detector undersamples the line spread function , the effective resolution to separate two adjacent lines with this grating is only @xmath10 = 110,000 .
using the e230h grating , the fwhm is about 2 pixels , corresponding to fwhm = @xmath14/110,000 @xcite .
this previously undetected spin - forbidden transition @xmath15p@xmath16p@xmath17 at 1613.376 was clearly observed on both days . it is unsaturated , which allows for an improved determination of the column density of .
as there was no change in the line between the two days of observation , the spectra were averaged together to improve the s / n ; the result is shown in figure 1 .
unfortunately , the other two fine structure lines in the multiplet were not reliably detected .
atomic data were taken from @xcite and the continuum around the line was fit with a sixth - degree polynomial .
voigt line profiles were used to generate transmission functions , which were then convolved with a gaussian instrumental line spread function with fwhm = @xmath14/220,000 to create model spectra . the @xmath18 statistic between the model and the data
was then minimized to determine the best velocity centroid , @xmath19 , column density of in the ground level , @xmath20 , and doppler broadening parameter , @xmath21 ; error bars for these parameters were determined from the contours of @xmath18 .
since the line was unsaturated , the column density was also determined from the equivalent width of the line and @xmath21 and @xmath19 were determined from a simple gaussian fit to the line , as a check on the results of the @xmath18 minimization .
the central portion of the 1561 and 1657 multiplets , arising from the @xmath22p ground term , were heavily saturated in the stis data .
however , this allowed us to examine the smaller variable red and blueshifted features in the spectra which are the signatures of infalling comets in the cs disk . in figures 2 and 3 ,
one can easily see absorption lying at higher redshift in the december 6 data that is not visible in the december 19 data .
a model of the multiplet which contains only one velocity component is unable to reproduce the absorption dip at 1561.5 or the absorption at higher
redshift in the december 6 data . due to the fact that the multiplets were extremely saturated and that the multiple velocity components were blended together , we were not able to perform @xmath18 minimizations to find unique best values for the model parameters .
absorption features in our data set were analyzed and the velocities of their multiple velocity components found .
models of the @xmath141561 multiplet were constructed , containing the same number of components as were found in the features , with roughly the same velocities .
the parameters of the stable component at @xmath19 = 20 km s@xmath12 were set to those determined from the analysis of the unsaturated 1613.376 line .
the remaining parameters were then adjusted to find the best model by eye .
this model was compared to the 1657 multiplet to confirm that the values found in this way were reasonable . a small constant value ( @xmath23 erg s@xmath12 @xmath3 @xmath12 )
was subtracted from the 1657 data to bring the baselines of the totally saturated features to zero .
these lines arise from the excited , metastable @xmath5d state of the ground configuration and have not previously been observed in spectra of @xmath0 pictoris .
the @xmath5d level has a lifetime of about 4000 s , lying about 10,000 cm@xmath12 above the ground level ; c atoms in this state may be produced during the photodissociation of co in solar system comet comae @xcite .
the 1931 lines were analyzed using the same model generating and @xmath18 minimization procedures described in 3.1 ; the atomic data used were from @xcite .
although there may be some reason to suspect that the strong central absorption features near 1931.05 contain multiple velocity components , since this behavior is seen in lines arising from excited levels of other atomic species , the @xmath18 minimizations indicated that stable unique solutions containing more than one component in the central absorption features could not be found .
thus , best models were found using only one velocity component . however , the difficulty involved in modeling a saturated , blended line is such that there may well be other undetected components present , so models with two components in the central absorption feature were compared to the data by eye .
the 1463 spectra , shown in figure 5 , were so noisy that they served only to roughly confirm the values found from the 1931 line . a small constant flux value ( @xmath25 erg s@xmath12 @xmath3 @xmath12 )
was subtracted from the 1463 data to bring the baselines of the saturated lines to zero .
the ( 0 - 0 ) , ( 1 - 0 ) , and ( 2 - 0 ) bands of the fourth positive system of co ( @xmath26 ) did not vary between the two days of observation .
the spectra were therefore averaged together to improve the s / n ; the result appears in figure 6 . only a single velocity component was observed .
note the detection of bands arising from @xmath7co , the perturbation band @xmath27 ( 1 - 0 ) , not previously observed in spectra of @xmath0 pictoris , and the strong perturbation band @xmath28 ( 5 - 0 ) .
models were generated using wavelengths and oscillator strengths from @xcite .
energies of the ground state levels were calculated using the dunham coefficients from @xcite and lte assumed in order to determine the population of the rotational levels .
the parameters of the model are the rotational excitation temperature , @xmath29 , the doppler broadening parameter , @xmath21 , the column density of @xmath9co in the ground vibrational state , @xmath30 , the column density of @xmath7co in the ground vibrational state , @xmath31 , and the velocity centroid , @xmath19 .
@xmath18 minimization was then performed on all three bands simultaneously .
the ] @xmath141613.376 line showed only one velocity component at 20 km s@xmath12 , the systemic velocity of the star ; the nominal uncertainty in the absolute wavelength calibration of stis leads to an error in velocity determinations of about 1 km s@xmath12 @xcite .
this velocity and the fact that the line did not change between the two days of observation identifies the line as arising from stable gas .
the results of the @xmath18 minimizations for all the features analyzed appear in table 2 .
the column density for stable carbon in the ground level , @xmath22p@xmath32 , determined from @xmath18 minimization was found to agree with that determined from the equivalent width of the line .
similarly the @xmath19 and @xmath21 determined from @xmath18 minimization agreed with the values found from fitting a simple gaussian to the unsaturated line . using a 3-@xmath33 upper limit on the column density of in the @xmath22p@xmath34 level ( from the non - detection of the fine structure line at 1614.5068 ) and assuming lte , a firm upper limit of 100 k on the excitation temperature of @xmath22p carbon in the stable component was found . from analysis of the 1561 multiplet ,
discussed below , it was found that the excitation temperature of the stable component must be greater than about 50 k , or the multiplet could not be reasonably modeled .
this range in temperature allows us to determine that the total column density of stable in the @xmath22p ground term is @xmath35 @xmath3 .
this column density is more than an order of magnitude larger than the total @xmath22p column density found by @xcite from ghrs data taken in november 1994 , @xmath1@xmath36 @xmath3 .
either the abundance of stable has varied over the three years between observations or , as is more likely , the difficult modeling of the heavily saturated @xmath141561 multiplet in the ghrs data led to an inaccurate column density .
models with column densities of @xmath37 @xmath3 in the stable component were compared to the 1994 ghrs spectrum of the 1561 multiplet and found to fit the data equally well as models with the lower column density .
we found the doppler broadening parameter of the stable ( @xmath22p ) to be @xmath38 .
the previous work on the ghrs data found a doppler broadening parameter of 4.2 km s@xmath12 , which was much greater than the parameter found for the other atomic species and co @xcite .
our smaller @xmath21 value is equal to that found for co in our data ; we thus do not see any excess kinetic energy in the motions of the atoms .
the difficulty in modeling a saturated , blended multiplet is formidable , and the values determined for the multiple velocity components of ( @xmath22p ) from the 1561 multiplet are not reliable .
however , this analysis did confirm that models with column densities of @xmath37 @xmath3 in the stable component could reasonably fit the 1561 spectra .
we can also conclude that the total column density of ( @xmath22p ) in the variable components is about @xmath39 @xmath3 on december 6 and about @xmath13 @xmath3 on december 19 , to within an order of magnitude or so .
it is also clear that the variable features are generally better fit with higher excitation temperatures and larger @xmath21 parameters , but these values are very poorly determined . using models with only one velocity component in the strong central absorption features ,
two components arising from the excited @xmath5d level of were found in the december 6 data , only one in the december 19 data .
the best models are overplotted on the spectra in figures 4 and 5 .
note that the central absorption features do not appear at the same velocity on the two days , that they also do not appear at the systemic velocity of the star , and that the column density of ( @xmath5d ) changes significantly between the two days of observation .
this would seem to indicate that there is no stable component for ( @xmath5d ) .
models with two velocity components in the central absorption feature , which were compared to the data by eye , indicate that there could be a `` stable '' component which has the same column density on both days . however , it must be at about 22 - 23 km s@xmath12 , which is significantly different from the systemic velocity of @xmath0 pictoris .
these data are not able to conclusively determine the velocity structure present in the ( @xmath5d ) gas ; observation of unsaturated lines at similarly high resolution will be necessary .
[ [ co ] ] co ~~ the co bands showed no multiple velocity component structure and no change between the two days of observation ; the absorbing gas is thus entirely associated with the stable component at the systemic velocity of the star .
the best model is shown overplotted on the data in figure 6 . the rotational excitation temperature and the ratio of @xmath30 to @xmath31 show no significant change from the values found by @xcite .
@xmath40 = @xmath11 is quite small compared to typical values found in the ism , e.g. @xmath10 = @xmath41 for the diffuse clouds toward @xmath42 ophiuchi @xcite .
however , chemical fractionation at low kinetic temperatures can explain this unusual ratio , as discussed below .
the column densities of both @xmath9co and @xmath7co were found to be about 1/3 the values found from the ghrs data .
note that the column density of @xmath9co is about 50 times smaller than the column density of stable ( @xmath22p ) .
although co in solar system comet comae is photodissociated by solar fuv photons , the co in the @xmath0 pictoris cs disk is primarily destroyed by interstellar photons .
the dissociation energy for co is 11.1 ev ; thus co may only be dissociated by photons with wavelengths shortward of @xmath43 1100 .
the type a5 star @xmath0 pictoris lacks the strong fuv emission lines created in the sun s chromosphere ; thus it emits very little flux in the fuv . the only source for co - dissociating photons at @xmath0 pictoris is therefore the interstellar uv radiation field .
since co photodissociates primarily through discrete line absorptions , self - shielding can have a strong effect on the abundance of co in interstellar clouds @xcite .
however , this is unlikely to occur in the @xmath0 pictoris cs disk because of the very small transverse dimension of the disk .
thus , the photodissociation rate for @xmath9co in the @xmath0 pictoris disk should be equal to the unshielded value and should also be equal to the value for @xmath7co .
since @xmath7co is not selectively dissociated in this situation , its high abundance was explained by the reaction @xmath44 which favors the production of @xmath7co at gas kinetic temperatures below 35 k @xcite . assuming chemical equilibrium and that isotopic exchange is much more important than photodissociation for both @xmath9co and @xmath7co @xcite , @xmath45 the rotational excitation temperature of co often does not accurately describe the gas kinetic temperature in diffuse environments @xcite .
however , assuming a @xmath46 ratio , we may use the above expression to estimate the true gas kinetic temperature of the co. the average carbon isotopic ratio in the local ism is @xmath47 @xcite and the typical solar system value found in comets is @xmath48 @xcite . using the range @xmath49 , we find that the gas kinetic temperature of the carbon monoxide is 20 k @xmath50 25 k , indicating that the co gas is indeed colder than the stable gas .
this suggests that the and co are not located in the same regions of the disk . the assumption that isotopic exchange is more important than photodissociation should be reasonable .
for this assumption to apply , @xmath51 where @xmath52 is the unshielded photodissociation rate , @xmath53 @xcite , k@xmath54 is the forward reaction rate , @xmath55 at 80 k @xcite , and n@xmath56 is the volume density of @xmath57 .
the average volume density of c , @xmath58 , assuming that the carbon extends over a distance @xmath59 = 100 au .
further assuming that @xmath60 , as is likely , the right hand side of the above expression is greater than or equal to about @xmath61 , an order of magnitude larger than @xmath52 .
however , a more complete treatment of the relationship between @xmath62 and @xmath46 takes into account photodissociation but requires exact knowledge of the density of @xmath57 in the cs disk @xcite .
the much larger column density of ( @xmath22p ) compared to that of co leads us to believe that photodissociation of co can not be the only source of stable in the @xmath0 pictoris cs disk . since there is no evidence that co is produced by infalling comets ( no variability or red and blueshifted features ) , it has been postulated that the co gas slowly evaporates from the oebs at several tens of au from the star and is photodissociated to produce the stable @xcite .
obviously some portion of the gas must be produced directly from the febs ( the portion giving rise to the variable red and blueshifted absorption ) ; perhaps this gas is decelerated somehow and accumulates in the stable component before being destroyed by photoionization . in this scenario ,
the equilibrium column density of is @xmath63 where @xmath64 is the mean number of infalling comets per year , @xmath65 per year @xcite , @xmath66 is the total column density of gas produced by an infalling comet , and @xmath52 is the photoionization rate for , 0.004 yr@xmath12 .
this expression , which assumes that atoms are lost only through photoionization by interstellar uv photons , gives @xmath67 @xmath3 , which should easily be produced by infalling comets like the ones giving rise to the variable ( @xmath22p ) components in our data .
however , this is a rough treatment and the number of infalling comets per year varies significantly over time scales of a few years @xcite .
also , this treatment does not take into account reformation of by radiative recombination ; we can not take this process into account properly without knowledge of the amount of in the cs disk and the electron density .
considering the ( @xmath5d ) , when such atoms are produced by photodissociation of co , ( @xmath5d ) atoms must be produced also to conserve spin .
the minimum total photon energy needed for this dissociation is 14.33 ev , corresponding to a threshold wavelength of 865 .
since this threshold is below the lyman limit , there are virtually no interstellar uv photons capable of producing ( @xmath5d ) by photodissociation of co. thus , the ( @xmath5d ) atoms in the @xmath0 pictoris cs disk can not be produced by photodissociation of co by stellar or interstellar photons and must be produced by a collisional process involving ground state carbon atoms . since the energy of the @xmath5d state relative to the ground state is high , the collisional process must be a very energetic one and therefore is likely to be closely associated with the infalling bodies . consequently , with a short ( @xmath5d ) lifetime , there would be no stable component in this gas ; this behavior has not been previously seen in any constituent of the @xmath0 pictoris cs disk . but
this result is tentavive ; examination of an unsaturated line at high resolution is needed to determine the velocity structure of the ( @xmath5d ) gas .
the very high resolution , low scattered light contamination , and good order separation of this stis echelle data set has provided some clear advantages over previous observations of @xmath0 pictoris .
the rotational lines of co have been resolved , allowing for a much more precise determination of the physical parameters of the gas .
the column density of co is @xmath68 and the ratio @xmath69 is found .
the absence of transient red or blueshifted components in the high resolution co spectra supports the suggestion that this gas evaporates from cometary bodies orbiting far ( several tens of au ) from the star .
but the fact that the column density of co is only about 2% of the total column density of in the @xmath22p ground term implies that photodissociation of this co is not the primary source for gas .
it could perhaps be produced directly from infalling comets close to the star , but the mechanism by which it comes to zero velocity relative to the star and accumulates before being photoionized is unclear ( although see @xcite ) .
the ( @xmath5d ) gas may not have a stable component at 20 km s@xmath12 ; this unique species could prove to be a valuable tracer of feb activity in the @xmath0 pictoris cs disk . despite the advantages of this data set , our lack of success in modeling the heavily saturated multiplets indicates that in order to really determine the characteristics of the variable components of the gas , we need to observe an unsaturated line or multiplet , with an oscillator strength between that of the 1561 multiplet and that of the spin - forbidden 1613.376 line .
a number of suitable multiplets and lines lie in the fuv , shortward of @xmath43 1300 .
also , an unsaturated line arising from the @xmath5d level would allow us to confirm the velocity structure in this gas and to determine if the velocities of the @xmath5d gas components correspond with any of the velocities of the variable components in the @xmath22p gas .
three likely lines lie between 1311 and 1359 .
measurement of the densities of and would greatly help to unravel the carbon chemistry of the @xmath0 pictoris disk . again , potentially useful multiplets of these species lie in the fuv below 1340 .
thus , although this data set has vastly increased our knowledge about the important species and co in the @xmath0 pictoris disk , our understanding would probably benefit greatly from investigation of @xmath0 pictoris at shorter ultraviolet wavelengths .
we thank jason mcphate for his work on our absorption line profile codes and john debes for his work on the lines .
we also thank b - g andersson and our reviewer , x. tielens , for their fruitful comments .
this work is based on observations with the national aeronautics and space administration
european space agency hst obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 . support for this work at jhu was provided by grant go-07512.01 - 96a from the space telescope science institute .
beust , h. , lagrange - henri , a. m. , vidal - madjar , a. , & ferlet , r. 1990 , , 236 , 202 draine , b. t. 1978 , , 36 , 595 farrenq , r. , guelachvili , g. , sauval , a. j. , grevesse , n. , & farmer , c. b. 1991 , j. molec .
spectrosc . , 149 , 375 hibbert , a. , biemont , e. , godefroid , m. , & vaeck , n. 1993 , , 99 , 179 jewitt , d. c. , matthews , h. e. , owen , t. , & meier , r. 1997 , science , 278 , 90 jolly , a. et al .
1998 , , 329 , 1028 lagrange , a. m. et al .
1998 , , 330 , 1091 langer , w. d. , & penzias , a. a. 1993 , , 408 , 539 lecavelier des etangs , a. , vidal - madjar , a. , & ferlet , r. 1996 , , 307 , 542 lecavelier des etangs , a. 1998 , , 337 , 501 morton , d. c. 1991 , , 77 , 119 morton , d. c. & noreau , l. 1994 , , 95 , 301 sahu , k. et al . 1999 , stis instrument handbook v3.0 ( baltimore : stsci ) sheffer , y. , federman , s. r. , lambert , d. l. , & cardelli , j. a. 1992 , , 397 , 482 smith , b. a. , & terrile , r. j. 1984 , science , 226 , 1421 smith , d. , & adams , n. g. 1980 , , 242 , 424 tozzi , g. p. , feldman , p. d. , & festou , m. c. 1998 , , 330 , 753 van dishoeck , e. f. & black , j. h. 1988 , , 334 , 771 vidal - madjar , a. et al .
1994 , , 290 , 245 vidal - madjar , a. , lecavelier des etangs , a. l. , & ferlet , r. 1998 , , 46 , 629 wannier , p. , penprase , b. e. , & andersson , b - g 1997 , , 487 , l165 cccccc o4g001010 & e140h & @xmath70 & 1461 1663 & 07:17:58 & 900.0 + o4g001020 & e230h & @xmath71 & 1874 2146 & 07:40:26 & 80.0 + o4g001030 & e230h & @xmath72 & 1624 1896 & 07:48:20 & 678.8 + o4g001040 & e230h & @xmath73 & 2624 2895 & 08:48:28 & 360.0 + o4g001050 & e230h & @xmath73 & 2374 2646 & 09:02:20 & 360.0 + o4g001060 & e230h & @xmath73 & 2124 2396 & 09:16:12 & 288.0 + o4g002010 & e140h & @xmath70 & 1461 1663 & 19:51:29 & 900.0 + o4g002020 & e230h & @xmath71 & 1874 2146 & 20:13:57 & 80.0 + o4g002030 & e230h & @xmath72 & 1624 1896 & 20:21:51 & 678.8 + o4g002040 & e230h & @xmath73 & 2624 2895 & 21:25:45 & 360.0 + o4g002050 & e230h & @xmath73 & 2374 2646 & 21:39:37 & 360.0 + o4g002060 & e230h & @xmath73 & 2124 2396 & 21:53:29 & 288.0 + lcccc & @xmath74 & @xmath75 & @xmath76 & + december 6 & @xmath74 & @xmath77 & @xmath76 & 50 100 + & 26 & @xmath39 & 4 & @xmath78 + & 41 & @xmath39 & 10 & @xmath79 + & 57 & @xmath80 & 12 & @xmath79 + & & & & + december 19 & 11 & @xmath80 & 5 & @xmath81 + & @xmath74 & @xmath77 & @xmath76 & 50 100 + & 27 & @xmath13 & 5 & @xmath78 + december 6 & @xmath82 & @xmath83 & @xmath84 & + & @xmath85 & @xmath86 & @xmath87 & + & & & & + december 6 & @xmath88 & @xmath89 & @xmath90 & + & @xmath91 & @xmath92 & @xmath93 & + & @xmath94 & @xmath95 & @xmath93 & + & & & & + december 19 & @xmath96 & @xmath97 & @xmath98 & + & & & & + december 19 & @xmath99 & @xmath89 & @xmath90 & + & @xmath100 & @xmath101 & @xmath93 & + & @xmath74 & @xmath102 & @xmath103 & @xmath104 + & @xmath74 & @xmath105 & @xmath103 & @xmath104 + | high resolution fuv echelle spectra showing absorption features arising from and co gas in the @xmath0 pictoris circumstellar ( cs ) disk were obtained on 1997 december 6 and 19 using the space telescope imaging spectrograph ( stis ) .
an unsaturated spin - forbidden line of at 1613.376 not previously seen in spectra of @xmath0 pictoris was detected , allowing for an improved determination of the column density of at zero velocity relative to the star ( the stable component ) , @xmath1@xmath2 @xmath3 .
variable components with multiple velocities , which are the signatures of infalling bodies in the @xmath0 pictoris cs disk , are observed in the @xmath41561 and 1657 multiplets .
also seen for the first time were two lines arising from the metastable @xmath5d level of carbon , at 1931 and 1463 .
the results of analysis of the co @xmath6 ( 0 - 0 ) , ( 1 - 0 ) , and ( 2 - 0 ) bands are presented , including the bands arising from @xmath7co , with much better precision than has previously been possible due to the very high resolution provided by the stis echelle gratings .
only stable co gas is observed , with a column density @xmath8 . an unusual ratio of the column densities of @xmath9co to @xmath7co is found ( @xmath10 = @xmath11 ) .
the large difference between the column densities of and co indicates that photodissociation of co is not the primary source of gas in the disk , contrary to previous suggestion . |
gamma ray bursts ( grbs ) involve energetic , collimated relativistic outflows ( see * ? ? ?
* ; * ? ? ?
* for recent reviews ) .
the current consensus on short - duration grbs is that they are produced during the merger of a binary neutron star system or a neutron star - black hole system .
however , the detailed properties of the central engine which powers the burst remain uncertain . grb engine models often require both fast rotation and large magnetic fields , in order to produce a highly collimated jet .
however , the mechanism for producing a sufficiently strong large - scale field is uncertain , leading to the possibility that it might not be present .
for example , the blandford - znajek mechanism typically requires @xmath0 gauss in large - scale poloidal magnetic field to power a grb , but such large magnetic flux is not expected to be available in typical neutron star mergers .
it may instead be the case that the magnetic field in the central engine remains toroidal and/or small - scale , and is primarily responsible for providing angular momentum transport in the accretion disk . from an observational perspective , beaming - corrected grb energies are inferred to be @xmath1 erg @xcite , yet calculations of typical post - merger disk masses suggest @xmath2 erg @xcite , implying a very inefficient engine without a strong large - scale magnetic flux threading it ( contrary to what is often assumed , e.g. * ? ? ?
in fact , in the long grb context , even if one assumes the existence of strong large - scale poloidal magnetic fields , it can still be difficult for some engine models to produce a highly collimated outflow @xcite .
these concerns are severe enough that we take seriously the possibility that strong , large - scale magnetic fields may not be present in a short - duration grb central engine . in this study , we investigate an alternative picture , in which the engine itself does not need to be strongly collimated , but the outflow becomes a collimated jet after interaction with a dense surrounding medium .
this idea has been previously proposed to explain the formation of radio components in active galactic nuclei @xcite and is known to be an important part of the collimation process in long grbs @xcite .
the process of collimation from a wide engine to a narrow jet can be enhanced if the surrounding medium has an oblate geometry ( as also suggested by * ? ? ?
oblate density profiles are quite natural in grb scenarios , as most engine models require rapid rotation for operation .
under the right circumstances , even a spherically symmetric engine can produce a highly collimated outflow if the geometry of the surrounding medium is sufficiently oblate .
this is ultimately due the fact that a blastwave expanding in an asymmetric external medium will preferentially favor the direction of steepest density gradient @xcite .
thus , an oblate external medium can result in a prolate blastwave .
if the engine is persistent , continuing to push from the inside of this prolate bubble , the aspect ratio of the bubble will become increasingly exaggerated until the flow is collimated to within a sufficiently small opening angle . at this point
, recollimation shocks can take over @xcite and maintain the collimation until the jet completely breaks out of the surrounding cloud . in the present study
, this collimation process will be demonstrated numerically , in the context of short grbs .
numerical relativity simulations show that shortly after a neutron star merger , an oblate cloud of dynamical ejecta is present @xcite .
the idea that this cloud can provide a collimating cocoon has also recently been investigated @xcite , though in that study the parameter space of engines was not thoroughly explored and the surrounding cloud was assumed to be spherical . starting with this dynamical ejecta cloud as initial conditions , we investigate the collimation process for a range of engine parameters and including the oblate geometry , and find that very wide hydrodynamic central engines are capable of producing collimated jets .
this study performs a numerical integration of the equations of two - dimensional ( 2d ) axisymmetric relativistic hydrodynamics : @xmath3 @xmath4 where @xmath5 is proper density , @xmath6 is enthalpy density , @xmath7 is pressure , @xmath8 is the internal energy density , @xmath9 is the four - velocity , and @xmath10 is the minkowski metric ( the equations are expressed in units such that @xmath11 ) . the source terms @xmath12 and @xmath13 will be used to model the injection of mass , momentum , and energy by the central engine on small scales .
the equation of state is assumed to be radiation dominated : @xmath14 .
the equations are integrated using the moving - mesh hydrodynamics code , jet @xcite .
the jet code s moving mesh makes it effectively lagrangian , which allows for accurate resolution of highly supersonic and relativistic flows .
the inner and outer boundaries are also moved during the flow s evolution , so that the dynamic range covered at any given time is kept modest .
the spatial resolution is given by @xmath15 .
this value is not exact , as zones can expand and contract , but zones are actively refined or de - refined so as to maintain aspect ratios of order unity .
additionally , zones are more concentrated near the symmetry axis , so that resolution there is about twice as high ( @xmath16 ) .
initial conditions are modeled after the numerical output of a neutron star merger simulation @xcite .
it is an outflow described by the following : @xmath17 @xmath18 @xmath19 @xmath20 @xmath21 where @xmath22 , @xmath23 , @xmath24 , @xmath25 , and @xmath26 . unless explicitly stated otherwise , the characteristic scales are the initial radius of the cloud @xmath27 km , and the mass of the cloud @xmath28 , though the hydrodynamical equations are scale - invariant , and therefore the value of these two quantities is arbitrary , so long as all other dimensionful quantities are expressed in terms of these two variables ( as well as the speed of light , @xmath29 ) . in cgs units ,
the external density @xmath30 g/@xmath31 is much larger than typical ism densities of @xmath32 proton per @xmath31 , but this is irrelevant for the current study , as we are only concerned with the production of the jet and not its late - time deceleration .
the density @xmath33 is chosen to be sufficiently low so as to have negligible effect on the jet during the course of the calculation .
the parameter @xmath34 is the aspect ratio of the oblate cloud .
a hydrodynamic engine is injected on small scales , using the source terms @xmath12 and @xmath13 in equations ( [ eqn : claw1 ] ) and ( [ eqn : claw2 ] ) , as in @xcite . however , in contrast with that study , the engine in the present study will have a very wide half - opening angle , @xmath35 , or a full opening angle of about 60 degrees .
this is comparable to the opening angle of the high speed outflow seen in hydro simulations of radiatively inefficient black hole accretion disks ( e.g. * ? ? ?
* ) or mhd simulations without significant large - scale magnetic flux ( e.g. * ? ? ?
additionally , in this study , the injection radius @xmath36 is a fixed multiple of the inner boundary , @xmath37 .
the engine duration is set to @xmath38 seconds .
the engine produces a clean relativistic outflow , with an energy - to - mass ratio of @xmath39 , and a bulk lorentz factor @xmath40 .
the luminosity of the engine is given by @xmath41 where @xmath42 is the ratio of the energy deposited to the cloud rest mass .
several calculations were performed , varying the value of @xmath42 from @xmath43 to @xmath44 , implying a total energy of @xmath45 in the vicinity of @xmath46 - @xmath47 ergs ( again , the interpreted value of this energy depends on the choice of mass scale , @xmath28 ) .
the primary calculation assumes @xmath48 erg , and @xmath28 , implying @xmath49 .
short grb energies have been observationally inferred to be in the @xmath50 erg range @xcite .
alternatively , from theoretical models
, typical disk masses are of order @xmath51 @xcite , so that if this disk is powering the engine , the total energy available to the engine is @xmath52 erg , assuming @xmath53 efficiency converting the accretion power into an outflow .
the mass of the surrounding cloud is expected to be in the range of @xmath54 @xcite , and therefore the value of @xmath55 could take on values anywhere between @xmath56 and unity , comparable to the range explored in this study .
figure [ fig : pretty ] shows two example runs with @xmath57 erg .
the engine properties of these two runs are identical ; the only difference between the two is the shape of the surrounding cloud , parameterized by the aspect ratio , @xmath58 .
the left side of the figure shows a spherical setup with @xmath59 and the right side uses @xmath34 , a value chosen based on the aspect ratio of the dynamical ejecta found in binary neutron star merger simulations @xcite .
the @xmath59 calculation did not produce a collimated outflow , as the pressure from the cocoon by itself was not significant enough to keep the jet collimated .
the final outflow is roughly spherical ( see also figure [ fig : theta ] ) . in the oblate ( @xmath34 )
run , the outflow from the engine is initially only weakly collimated , but this asymmetry becomes more prominent as the bubble inflates , preferentially moving along the steepest density gradient .
as the engine pushes , this asymmetry becomes increasingly exaggerated until recollimation shocks are capable of keeping the jet collimated .
the flow then clears a passageway , through which a highly collimated jet can escape .
this collimated flow then breaks out of the cloud with a lorentz factor in the tens , after which post - shock acceleration can increase this lorentz factor into the hundreds .
the remaining figures show how the flow is distributed in angle and lorentz factor , and how the results depend on the energy injected by the central engine relative to the mass of the ambient medium ( figures [ fig : edep ] and [ fig : w1 ] ) and the injected opening angle of the outflow ( figure [ fig : thdep ] ) .
the terminal lorentz factor is measured via the specific enthalpy , @xmath60 , which is the terminal lorentz factor if the fluid element is allowed to expand and adiabatically convert all of its thermal energy to kinetic energy . in the oblate example shown in figure [ fig : pretty ] , about 30% of the energy has @xmath61 .
figure [ fig : theta ] shows the jet s distribution with latitude . for the @xmath34 example , @xmath62 of the energy in the jet is contained within @xmath63 .
the @xmath59 example is also included to show that the flow is roughly spherical and nonrelativistic if the surrounding medium is spherical ( @xmath64 for the @xmath59 model ) . looking back to the @xmath34 case , the dip at very small angles implies a less relativistic jet core ( this can also be seen in figure [ fig : pretty ] ) , but this is most likely an artifact of assuming axisymmetry .
the maximum terminal lorentz factor for this jet is around @xmath65 .
the average lorentz factor " @xmath66 ( bottom panel of figure [ fig : theta ] ) is calculated via the ratio of total energy to total mass at a given polar angle .
figure [ fig : edep ] shows how these results depend on the total energy injected by the engine relative to the mass of the cloud , @xmath67 . for sufficiently low energy injection
, @xmath68 is negligible compared to the kinetic energy in the dynamical ejecta , and therefore the engine is unsuccessful at producing a collimated outflow .
this leads to the low lorentz factors and low energy in relativistic material for low values of @xmath68 in figure [ fig : edep ] . for sufficiently large @xmath68 , the mass of the ejecta is not significant enough to collimate ( or otherwise affect ) the flow , and the outflow keeps the same opening angle it was given initially .
note that the limiting case in the high - energy limit is a jet with @xmath69 and @xmath70 , roughly the same jet parameters that were injected by the engine , implying that the surrounding cloud does little to affect the properties of the flow . the behavior in these two limits ( low and high energy ) imply an intermediate sweet spot " for the total energy output of the engine , which is optimal for producing a collimated , relativistic outflow .
for the model presented here , this sweet spot " is @xmath71 , which corresponds to @xmath68 in the few @xmath72 erg range for @xmath28 . this optimal energy
likely depends on the detailed structure of gas in the surrounding cloud . for our engine and cloud model , at this sweet spot , about @xmath73 of the energy is used to accelerate the relativistic material with @xmath61 , while the remaining @xmath74 is deposited in the slower - moving material in the dynamical ejecta .
figure [ fig : thdep ] shows the dependence of the final opening angle on @xmath75 , the half - opening angle of the engine ( for fixed @xmath76 ) . for sufficiently small @xmath75 , the outflow has @xmath77 , roughly independent of @xmath75
. however , above some critical threshold , @xmath78 begins increasing , so that the final outflow is wide when @xmath79 .
finally , figure [ fig : w1 ] compares the spherical ( @xmath59 ) and oblate ( @xmath34 ) ejecta models , to determine the extent to which oblateness affects collimation .
the spherical ejecta model has difficulty collimating such a wide engine , with a minimum @xmath80 .
the shape of the surrounding medium therefore plays an important role in the collimation process .
dynamical processes during neutron star mergers create an expanding oblate cloud of ejecta ( @xmath81 ) moving out at @xmath82 @xcite . we have demonstrated that this ejecta collimates outflows from short - duration grbs , even if the outflow has a large opening angle to begin with .
there is thus no need for a central engine that itself generates a highly collimated relativistic outflow .
this result is a variant of the well - known idea that hot flows can potentially self - collimate as they become supersonic @xcite .
we also show that the oblateness of the surrounding medium is an important ingredient to the dynamics .
the canonical view of short - duration grbs has been that they are powered by a relativistic jet analogous to radio - loud quasars or long - duration grbs @xcite .
one indication that this analogy breaks down is that the outflow energies in short grbs appear to be orders of magnitude below the energy available given remnant accretion disk masses of @xmath83 found in simulations of neutron star mergers @xcite .
this is in contrast to other jet - dominated sources .
this suggests that the black hole event horizon is not threaded by a dynamically important poloidal magnetic field .
our results are consistent with this general inference in that collimated outflows from short grbs need not originate from a central engine that in isolation produces a powerful collimated outflow .
some small - scale and/or toroidal magnetic field is likely still an important part of the engine s operation , but the field does not have to be responsible for collimating the flow to within @xmath84 , and the outflow does not have to be magnetically dominated .
a full opening angle of about @xmath85 appears sufficient for the surrounding medium to collimate the outflow from the central engine ( figures [ fig : pretty ] and [ fig : thdep ] ) .
this result removes some constraints from traditional short grb engine models , and also opens up new possibilities for what could be powering the central engine .
neutrino pair annihilation is more viable as an energy source if one does not require the engine itself to be collimated .
a @xmath85 opening angle is also typical of disk winds , meaning that a disk wind could be a potential engine model , providing that a very clean ( baryon - free ) wind could be launched from the disk .
the present study focused on the binary neutron star merger scenario for short grbs .
however , the same considerations apply to neutron star - black hole mergers , for which the dynamical ejecta is expected to be much more oblate ( e.g. * ? ? ? * ) . based on the results presented here
, we expect differences in the opening angles and lorentz factors of outflows from black hole - neutron star mergers , relative to binary neutron star mergers .
this result motivates several important future studies .
first , this same calculation should be carried out in full 3d , as it is possible that the collimation and ejection process are sensitive to 3d effects .
for example , the hollow jet core " seen in figure [ fig : theta ] is likely an artifact of the assumption of axisymmetry .
second , this calculation should be performed in the context of long grbs , i.e. a wolf - rayet progenitor which is oblate due to rotation ( * ? ? ? * in preparation ) .
it seems likely that such a system will exhibit similar behavior .
third , larger parameter surveys are warranted to study specifically the oblateness of the surrounding medium , and its effect on collimation .
it is also worth noting that figure [ fig : pretty ] shows a massive plug " or cork " in front of the jet , resulting in a top - heavy " flow , similar to what was seen in @xcite .
the properties of this cork and its impact on the jet dynamics should be investigated in more detail .
finally , it may be possible for analytical jet models @xcite to be modified to include the effect that oblate geometry can have on the properties of a jet breaking out of a surrounding medium .
this would be valuable for understanding what necessary properties an engine must have in order to produce a relativistic , collimated jet after breakout .
this work was supported in part by nsf grant ast-1205732 , nasa fermi grant nnx13ao93 g , a simons investigator award from the simons foundation , and the david and lucile packard foundation .
resources supporting this work were provided by the nasa high - end computing ( hec ) program through the nasa advanced supercomputing ( nas ) division at ames research center . | we use two - dimensional relativistic hydrodynamic numerical calculations to show that highly collimated relativistic jets can be produced in neutron star merger models of short - duration gamma ray bursts without the need for a highly directed engine or a large net magnetic flux .
even a hydrodynamic engine generating a very wide sustained outflow on small scales can in principle produce a highly collimated relativistic jet , facilitated by a dense surrounding medium which provides a cocoon surrounding the jet core .
an oblate geometry to the surrounding gas significantly enhances the collimation process .
previous numerical simulations have shown that the merger of two neutron stars produces an oblate , expanding cloud of dynamical ejecta .
we show that this gas can efficiently collimate the central engine power much as the surrounding star does in long - duration grb models . for typical short - duration grb
central engine parameters , we find jets with opening angles of order 10 degrees in which a large fraction of the total outflow power of the central engine resides in highly relativistic material .
these results predict large differences in the opening angles of outflows from binary neutron star mergers versus neutron star - black hole mergers . |
the static capacitance of a complex metallic structure is commonly considered to be a global quantity of the entire system and , consequently , is supposed to characterize the electronic state of the system rather insufficiently to deduce microscopic properties .
however this is not always the case .
for instance , recent experiments on polar oxide heterostructures with conducting laalo@xmath0/srtio@xmath0 interfaces suggest that the capacitance for these systems is a quantity which allows to draw distinct conclusions on the electronic interface state @xcite .
a special characteristic of this interface is that the electrons are subject to a sizeable spin - orbit coupling @xcite .
the question arises if spin - orbit coupling is a relevant control parameter for the capacitance of the interface . here , we address the fundamental issue whether spin - orbit coupling of the interface electrons allows to modify the capacitance of the heterostructure significantly , and we estimate the spin - orbit contribution for systems based on laalo@xmath0/srtio@xmath0 interfaces .
an enhancement of the capacitance is possible in the presence of a sufficiently large rashba - like spin - orbit coupling ( rsco ) , because the rsoc allows to alter the electronic band energies by controlling the charge carrier density .
such a rsoc appears to explain magnetotransport measurements at laalo@xmath0/srtio@xmath0 interfaces @xcite , and it has been shown within _ ab initio _ evaluations that a rsoc is generated by the electric field of the polar laalo@xmath0 layers on top of the srtio@xmath0 substrate @xcite . the rsoc lowers the band edge quadratically with the spin orbit coupling parameter @xmath1 ( see fig . [ dispersion ] ) .
this even allows for a negative compressibility assuming that @xmath1 depends on the electric field @xmath2 of the polar laalo@xmath0 layers which in turn is related to the charge density at the laalo@xmath0/srtio@xmath0 interface @xcite . with their spectacular result of a negative compressibility in the presence of a sizeable rsoc , the authors of refs .
@xcite suggest that this provides an intrinisic mechanism for the inhomogeneous phases observed at laalo@xmath0/srtio@xmath0 interfaces yet , other scenarios for phase separation at laalo@xmath0/srtio@xmath0 interfaces have also been investigated recently @xcite .
it is well known , that a negative electronic compressibility on one or both electrodes of a capacitor can lead to an enhancement of the capacitance beyond its geometrical value @xcite .
in fact , the geometric capacitance of a two - plate capacitor , @xmath3 , is modified by electronic contributions which are taken into account in the electronic compressibility of the plates @xmath4 , the derivative of the charge carrier density @xmath5 at plate @xmath6 with respect to the chemical potential @xmath7 . here
@xmath8 is the dielectric constant of the material between the two electrodes of area @xmath9 , and @xmath10 is the thickness of the dielectric .
the kinetic contributions to the capacitance were coined by luryi @xcite as quantum capacitance .
this notation may be extended to all electronic compressibility terms ( see ref . ) .
already in the early nineties , a capacitance enhancement of semiconductor heterostructures has been identified when the electronic compressibility of quantum hall systems was investigated @xcite .
the enhancement of the semiconductor capacitance could be traced quantitatively to the presence of quantum exchange effects in the electrodes .
the contribution from electronic correlations to the capacitance is not negligible for very dilute electron systems @xcite and for electronic systems close to half - filling @xcite .
however , the charge density at laalo@xmath0/srtio@xmath0 interfaces is in the intermediate range , neither as low as in the semiconductor systems @xcite nor close to half - filling .
therefore , the more sizeable capacitance enhancement of the oxide laalo@xmath0/srtio@xmath0 heterostructure @xcite and its negative compressibility @xcite can not be explained exclusively by exchange and correlation effects and other mechanisms might become relevant .
-space for @xmath11 ( left , only the lowest band is occupied ) and for @xmath12 ( right , both bands are occupied ) . ] to determine the thermodynamic electronic compressibility , two alternative procedures are conceivable : ( i ) one can either consider the interface subsystem with an _ external _ electric field which induces the rsoc .
in order to identify the electronic compressibility of the interface electrons , one has to keep the electric field fixed in thermodynamic derivatives . or
( ii ) , one evaluates the compressibility for the entire electronic system , for which the rsoc - generating field is an _ internal _ field .
a scheme in the sense of ( i ) has been suggested by the authors of ref .
there the electric field was set as a function of the interface charge carrier density @xmath13 and , then , the inverse compressibility was identified from the derivative of the chemical potential @xmath14 with respect to @xmath13 , keeping any dependence else fixed .
we disagree with this approach because there the electric field has to be kept fixed and therefore does not contribute to the compressibility @xcite . in this paper
we establish the second procedure ( ii ) and relate the electronic compressibility of the entire electronic system directly to the capacitance of the device consisting of the interface , dielectrics , and the surface of the laalo@xmath0 film or a metallic top electrode .
as indicated above , previous investigations focused on the interface subsystem @xcite , whereas we consider the entire system .
for realistic values of the rsoc , we identify a positive compressibility for the most part of the parameter space indicating that the system is thermodynamically stable . nevertheless , we find an increased capacitance with respect to its geometrical value , the capacitance enhancement being caused by the rsoc in the metallic interface . in this article
, we will investigate two different models , both building on a two - dimensional ( 2d ) metallic system in which a rsoc modifies the electronic state : first we analyze a capacitor with two metallic planes , where the top electrode is a 2d or 3d metal ( gate ) and the second plane is the interface between a polar and non - polar insulator .
the interface attains a metallic state with rsoc upon charge transfer from the top plate to the interface and the concomitant build - up of an internal electric field .
finally we study a two - layer system with the spin - orbit coupling controlled by the filling of three @xmath15-orbitals at the interface .
we will show that the multilayered system has a range of parameters where it has a positive total compressibility , but its capacitance is enhanced above its geometrical value on account of the spin - orbit coupling in the metallic interface plane . in this analysis
we adjust the parameters to those specific for laalo@xmath0/srtio@xmath0 ( lao / sto ) heterostructures in order to explore the feasibility of spin - orbit controlled capacitance enhancement with a polar dielectric material .
in a 2d free electron system subject to a rashba spin - orbit coupling ( rsoc ) @xmath16 , with coupling constant @xmath17 , the dispersion @xmath18 is given by ( see , for example , ref .
@xcite ) : @xmath19 where @xmath20 is the density of states for both spin directions in the absence of rsoc , @xmath1 defines the strength of the rsoc , and @xmath21 is the effective mass of the charge carriers .
the rsoc is made possible by the absence of inversion symmetry , caused by an electric field perpendicular to the two - dimensional layer containing the electrons with areal density @xmath13 .
the spin - orbit splitting of the dispersion , eq .
( [ rsocdispersion ] ) , is shown in fig .
[ dispersion ] .
if @xmath22 , only the lower band is occupied and the fermi surface is an annulus ; for @xmath23 the two occupied bands yield two concentric circles in @xmath24-space .
as the lowest energy @xmath25 of the dispersion is proportional to the strength of the rsoc , it decreases with increasing strength of the symmetry breaking electrical field .
thus a stronger electrical field lowers the chemical potential for constant electron density @xmath13 . in ref .
@xcite it was suggested that a negative compressibility can be obtained from exploiting this effect .
the inverse compressibility is obtained by the second derivative of the free energy @xmath26 per area with respect to the electron density _ with all external variables held fixed_. since the symmetry breaking electrical field ( sbef ) is an external variable in this scheme ( and the field generated by the electrons in the plane themselves is symmetric ) , @xmath1 is independent of @xmath13 in the thermodynamic derivatives , and there is no dependence of the compressibility on the rsoc .
here we propose an extended scheme in which the electric field from the polar film , and consequently the rsoc , is an internal parameter , i.e. , a quantity that depends on the electron density @xmath13 .
then the density dependence of @xmath27 has to be accounted for when taking the derivative of the energy for the compressibility .
in addition to the interface layer with a finite rsoc ( l0 ) , a second layer ( l1 ) , e.g. , the surface of the lao film in the lao / sto heterostructure , comprising electrons and a positive background charge , is placed parallel at a distance @xmath10 . eventually , both layers each of them electrically neutral are connected .
since typically both layers had different electrochemical potentials , electrons will be exchanged between l0 and l1 until both layers are on the same electrochemical potential .
it is of no relevance for the static capacitance how electronic charge is transferred between the plates .
however , an electric field , which is proportional to the transferred amount of charge , is generated .
this internal field breaks the inversion symmetry and thus generates rsoc in the respective layer .
it stands to reason that also the electrons in the surface layer l1 experience a rsoc because the same electric field acts on the electronic states in l1 . in the case of a bare rsoc @xmath28 in @xmath29 for free electrons with mass @xmath30 , momentum @xmath31 and spin @xmath32 in a plane subject to a transverse electrical field @xmath33
, a reasoning demanding an equivalent treatment of layers l0 and l1 with respect to the rsoc appears to be appropriate .
however , the bare rsoc with electric fields generated by the charged layers is orders of magnitude too small to have sizeable impact on the electronic reconstruction in the heterostructures . the resolution of this apparent unfeasibility of bare rsoc effects
has already been established for semiconductor heterostructures ( see the book by r. winkler @xcite for a comprehensive treatment ) and is implemented for the lao / sto interfaces as follows @xcite : the ti - ions provide an atomic soc , whereas an internal electric field which displaces the ti - ions slightly out of their in - plane positions , couples the three @xmath15 orbitals . as spin and orbital space couple , one has to diagonalize six bands .
the diagonalization of the 6-band system generates a soc of rashba type @xmath34 ( see eq .
( [ rsocdispersion ] ) ) even though only the lowest of the @xmath15 bands ( the @xmath35 ) may be partially occupied .
the rsoc parameter @xmath1 has to be estimated from electronic structure evaluations @xcite .
in fact , from magnetotransport measurements , a sizeable rsoc has been identified for the lao / sto interface electronic state ( @xmath36 for different backgate voltages ) , where the value of @xmath37 depends on the charge carrier density which is controlled by the internal electric field . in sec .
[ sec:3band ] we will address the 6-band situation explicitly . until then we consider an effective two - band situation with a rsoc that is controlled by the internal field that depends in turn on the charge density .
the actual set - up for the electronic surface states ( layer l1 ) is very different and soc effects have so far not been found to be relevant .
therefore we disregard rsoc in the surface layer l1 in our approach . through the mechanism outlined above , the strength of the induced rsoc , @xmath1 , becomes density dependent : we denote the helmholtz free energies of the single layers as @xmath38 and @xmath39 , calculated in the first stage without connection between l0 and l1 .
in fact , we calculate the single layer grand canonical energies and then use a legendre transformation to obtain the helmholtz free energies .
the total free energy of the connected system @xmath40 is the sum of the single layer energies and the electrostatic term @xmath41 .
the total energy @xmath42 is a function of the total density @xmath43 only , that is , the sum of both single layer densities : the dependence on the single layer densities is eliminated by the condition that @xmath13 has to minimize the total free energy , which is equivalent to the equality of the electrochemical potentials of both layers .
note that , as the layers are connected , @xmath44 becomes a function of the electric field generated by l1 and thus of @xmath45 and @xmath46 .
since both layers are coupled through the electric field and are supposed to be in electrical contact , only the inverse compressibility of the total system @xmath47 is a well - defined thermodynamic quantity . here
, @xmath48 is an appropriate legendre transform of @xmath49 with respect to the interface charge density @xmath13 : the thermodynamic variable is then the voltage bias @xmath50 between interface and surface ( see appendix [ appcompandcap ] ) .
the electronic compressibility derives from the second derivative of the electronic energy with respect to the volume . in the case of two - dimensional electron systems
( as in the considered heterostructure ) it is the area instead of the volume .
this translates into the second derivative with respect to the electronic areal density @xmath45 .
the second derivative in eq .
( [ kappa ] ) is to be taken with respect to the total dependence on @xmath45 , that is , also the @xmath45 dependence of the internal field in the polar film enters .
the positivity of @xmath51 then represents a stability criterion with respect to charge separation in the plane directions .
to approach a polar system with a surface and an interface to a non - polar insulator , we first present a model system that consists of three two - dimensional layers ( fig . [ electrostatic](a ) ) but only two of them comprise mobile electrons : the planes at the bottom and at the top have areal charge densities of @xmath52 and @xmath53 , respectively .
the electrons may be transferred between these two planes to minimize the free energy , while the plane sandwiched between them with a distance @xmath54 to the bottom and @xmath55 to the top holds immobile , fixed charges of density @xmath45 .
this charge density @xmath45 compensates the charge of the other two planes and therefore keeps the system neutral .
the dispersion of the electrons in the bottom layer ( l0 ) is given by eq .
( [ rsocdispersion ] ) , while the top layer ( l1 ) is described as an electron system with a density of states ( dos ) of @xmath56 . if @xmath57 , an electric field perpendicular to l0 exists which derives from the total charge of the two other layers .
this field induces the rsoc in the bottom layer and is proportional to the electron density @xmath13 of l0 .
the electrostatic energy per area of this layout is @xmath58 where @xmath59 are the dielectric constants of the material between the planes ( see appendix [ appelectrostaticenergy ] ) .
this arrangement is an effective model for the oxide heterostructure lao / sto , where polar layers with a distance @xmath60 and @xmath61 between them alternate ( fig .
[ electrostatic](b ) ) . due to electronic reconstruction ,
electrons from the surface relocate to the former empty layer at the lao / sto interface and occupy @xmath62 states @xcite .
we assume that the effective distances to the sandwiched layer are equal , i.e. , @xmath63 . in lao , the distances @xmath54 and @xmath55 are approximately equal but @xmath64 is not accurately known
. therefore we take a layer - independent dielectric constant @xmath8 for the lao film .
this simplification is of no qualitative consequences for the further evaluation ; the results depend on an effective distance @xmath65 .
there are various models for the density dependence of @xmath1 ( cf .
@xcite ) . in general ,
the strength of the rsoc @xmath1 is taken to be proportional to the sbef : @xmath66 .
the sbef @xmath2 is induced by the charge carrier density @xmath67 ( see fig . [ electrostatic](a ) ) and can be expanded as @xmath68 , where @xmath69 .
the factor @xmath70 accounts for the fact that the sbef is generated by the charge carrier density @xmath67 at the sandwiched layer and not by the electron density @xmath52 at the lao / sto interface l0 .
the field @xmath71 refers to an electric field that is already present in the absence of a charge density @xmath67 ; it may be produced by dipolar distortions in the lao - layer next to l0 . with this expansion of @xmath72
, one can also decompose @xmath73 a more adequate determination of @xmath74 will be given in the context of the 6-band model for the lao / sto interface in sec .
[ sec:3band ] . in order to implement the difference in chemical potentials of the electronic systems at the interface l0 and surface l1
, we introduce a term @xmath75 in the total free energy @xmath42 : @xmath76 the potential @xmath77 may be determined from electronic structure evaluations ; we use it phenomenologically to tune the electron density at layer l0 to the experimentally observed values .
the thermodynamic relation @xmath78 is used to find the function @xmath79 ( here we still consider @xmath80 , and @xmath81 ) .
since @xmath46 is limited to the interval @xmath82 $ ] , it is possible that the boundary values @xmath83 or @xmath84 yield the global minimum of @xmath85 . if neither all nor no charge resides at the bottom layer , an analytical expression for the stability of the total system can be derived ( appendix [ appcompandcap ] ) : @xmath86 we would like to point out that a system with @xmath87 can still be stabilized by a mechanism not related to the electronic system in l0 and l1 ( and thereby yield a positive total compressibility ) . in this case ,
an enhancement of the total electronic charge in l1 causes an amount of charge larger than the added charge to flow to the rsoc layer l0 .
the distribution of charge between l0 and l1 is controlled by the electric field which is in turn determined by the positive capacitance of the system . on the contrary , if the system can not be stabilized ( that is , the total compressibility is negative ) then it is expected to develop a phase separated ( inhomogeneous ) state in the plane . this latter scenario will not be investigated here but we will comment in our conclusions on the proposition @xcite of rsoc - driven phase separation .
we will not discuss the temperature dependence of the capacitance and therefore reduce our considerations to zero temperature in the following evaluations .
the free energy per area of the top two - dimensional layer ( l1 ) @xmath88 depends on the electron density transferred to the bottom layer .
here we assumed an interactionless electron system in l1 , characterized by a density of states @xmath56 . with @xmath89 ( that is , the mass of the charge carriers in l1 increases beyond limits , @xmath90 )
, a localized state can be generated technically .
the surface of the lao / sto heterostructure is insulating .
then , an additional metallic layer on top of the heterostructure produces a capacitor .
the top electrode can be a gold or yba@xmath91cu@xmath0o@xmath92 ( ybco ) film , which connects well with the topmost alo@xmath0 layer ( see appendix [ sec:3dlayer ] ) .
there , the capacitance increase is slightly smaller than for a 2d electrode due to the finite screening length in the 3d film but the qualitative results are the same as presented in this section for the 2d top electrode .
we re - emphasize that the top layer l1 must be included necessarily in order to render the correct electrostatics which then yields the proper total compressibility and capacitance .
however , as long as interaction or spin - orbit effects are negligible in l1 , the impact of l1 is determined by two parameters , only : the density of states @xmath56 at the surface electrode and the energetic displacement @xmath93 between interface and surface electronic levels .
the latter controls trivially the electronic density @xmath13 at the interface .
the former yields the kinetic term in the quantum capacitance and is known to produce a slightly increased effective distance between the electrodes ( see , e.g. , refs . ) .
consequently , the materials properties of the surface electrode do not affect the discussed physical nature of the heterostructure qualitatively , except for adjusting the carrier density at the interface .
the energy per area of the layer with rsoc ( with @xmath1 as density independent constant ) , calculated in appendix [ appenergysinglelayers ] , is @xmath94 when taking the density dependence of @xmath1 into account , we expand @xmath74 in @xmath13 ( cf .
( [ alphar ] ) ) and keep terms up to linear order in @xmath13 : @xmath95 one has to distinguish between the case where only the lower band of the rsoc split - bands is occupied , @xmath96 ( see fig . [ dispersion ] with @xmath22 ) , and the case where both bands are occupied , @xmath97 ( fig . [ dispersion ] with @xmath23 ) .
since @xmath98 corresponds to @xmath99 , the first case equals either @xmath100 or @xmath101 , with @xmath102 this means that also for large densities @xmath101 , all electrons are in the lower band , because the minimum energy of the bands , @xmath103 , decreases for increasing density . + with the total energy given by eqs .
( [ mlotorltotalenergy ] ) , ( [ mlotorlmetallicenergy ] ) and ( [ mlotorlrashbaenergy ] ) , we calculate the minimizing density @xmath46 numerically , and analytically for @xmath104 . we identify six different solutions for @xmath46 and determine the respective compressibility : + ( p0 ) : the bottom layer l0 is empty and the total compressibility is given by the compressibility of the top layer .
+ ( p1@xmath105 ) : this phase exists only for @xmath106 .
all electrons of the bottom layer are in the lower band .
the compressibility is calculated numerically according to eq .
( [ appcomp2 ] ) .
+ ( p2 ) : both bands of the bottom layer are occupied .
the compressibility is calculated numerically according to eq .
( [ appcomp2 ] ) .
+ ( p2@xmath107 ) : all electrons reside in the bottom layer , and both bands are occupied . if @xmath83 is inserted into the total energy @xmath108 and the second derivative with respect to @xmath45 is taken , the inverse compressibility is @xmath109 ( p1@xmath110 ) : all electrons of the bottom layer are in the lower band .
the compressibility is calculated numerically according to eq .
( [ appcomp2 ] ) .
+ ( p1@xmath107 ) : all electrons reside in the bottom layer , but only the lower band is occupied .
inserting @xmath83 into @xmath108 yields for the total compressibility : @xmath111 in fig .
[ compressibilityoverdal](a ) we display the inverse compressibility in the respective phases as a function of the effective plate distance @xmath112 and of @xmath113 for @xmath104 and a total charge of one electron per areal unit cell of the lao / sto structure .
the transition to a regime of negative compressibility is indicated by the transparent blue plane : the area below the blue plane exhibits negative compressibility , which can occur for all phases except p0 . for phase p2 , however , the negative compressibility regime is small , restricted to very small @xmath65 ( less than @xmath114 ) and large rsoc ( @xmath115 ) . for large effective plate distances both layers
are always occupied , since the electrostatic term is then dominating . in the reverse case
for small very @xmath65the electrostatic cost for transferring all electrons to one layer is negligible .
then , as @xmath116 decreases with increasing @xmath113 , only one band in l0 is occupied for strong rsoc .
for @xmath104 the phase boundary between p1@xmath110 and p1@xmath107 is given by @xmath117 this relation can be used to show analytically that @xmath118 is always negative .
one deduces from eq .
( [ kappap20 ] ) for phase p2@xmath107 that also @xmath119 . obviously , the phases p1@xmath107 and p2@xmath107 display a `` phase separation between the layers '' ( the index 0 indicates that all mobile charge carriers are in layer l0 ) . moreover , the negative compressibilities ( @xmath120 , @xmath119 ) signal that these phases are thermodynamically unstable and that a phase separation within plane l0 is expected . when no charge resides on the surface as it is the case in phases p1@xmath107 and p2@xmath107one could expected that the compressibilities of eqs .
( [ kappap20 ] ) or ( [ kappap10 ] ) are identical to the compressibilities calculated for one isolated layer in ref .
@xcite ( apart from the contribution of the electrostatic energy which straightforwardly produces the geometrical inverse capacitance ) .
this is , however , not the case : our results differ from those of ref .
@xcite , as explained in appendix [ appenergyvschempot ] .
this discrepancy can be traced back to the inclusion of a @xmath13-dependent rsoc @xmath74 either directly in the free energy or in a later stage in the chemical potential .
we calculate the inverse compressibility through the second derivative of the free energy with respect to @xmath121 whereas the authors of ref .
@xcite identify the inverse compressibility through a first derivative of the chemical potential .
the latter approach misses a non negligible contribution from the derivative of the energy with respect to the density in @xmath74 .
for the phase p1@xmath107 , for example , the latter approach results in an inverse compressibility that is three times smaller than our result for phase p1@xmath107 . moreover , those phases which do not have all electrons transferred from the surface l1 to the interface l0 , can not be addressed by the single layer scheme in ref .
@xcite , and we identify them from the free energy which comprises an @xmath13-dependent rsoc @xmath74 .
the application of a voltage bias @xmath50 between interface and surface , i.e. , between l0 and l1 , can be used to tune the electron density on the interface and hence also the phase of the system ( see fig .
[ nandcompoverv ] ) : as @xmath13 increases , so does the strength of the rsoc @xmath122 , and the band minimum is shifted to lower energies until only one band is partially occupied and the system is in state p1@xmath110 .
this process increases the compressibility that stays positive even if all charge is transfered to the interface ( p1@xmath107 ) .
this does not contradict the analytical analysis from above , which was for @xmath80 . as for the phases p0 and p1@xmath107 , note that @xmath123 is independent of @xmath124 , @xmath113 and @xmath125 , as no charge is in l0 , and that @xmath118 does not depend on the second electrostatic term in eq .
( [ mlotorltotalenergy ] ) , as no charge is in l1 .
this observation explains the sharp transitions in @xmath126 in fig .
[ nandcompoverv](a ) .
the differential capacitance @xmath127 is derived in appendix [ appcompandcap ] : @xmath128 the term @xmath129 yields the geometrical capacitance @xmath130 .
the total capacitance is enhanced above the geometrical value , if the sum of the inverse capacitances of the single layers is negative . in fig .
[ nandcompoverv ] we compare @xmath131 , the capacitance @xmath132 , the geometric capacitance @xmath133 , and @xmath134 , the capacitance for an interface without rsoc .
it is evident that the differential capacitance is enhanced above the capacitance of a system without rsoc for finite interface electron densities .
moreover , both @xmath131 and @xmath135 increase beyond @xmath133 for sufficiently large @xmath50 ( fig . [ nandcompoverv](b ) ) .
however , this does not signify a negative compressibility @xmath51 ( cf.fig .
[ nandcompoverv](a))although the partial compressibility of the interface my be negative ( the latter is not a well - defined thermodynamic quantity ) . as elaborated in appendix [ appcompandcap ] , the connection between the compressibility of the system and its differential capacitance is @xmath136 where @xmath85 is the legendre transform of @xmath49 : @xmath137 . with the general assumptions that the total energy @xmath108 of the system is a functional of @xmath13 and @xmath138 exclusively , and
that @xmath13 is not a boundary value ( @xmath45 or 0 ) then the compressibility of the system is related to the differential capacitance through @xmath139 where both , @xmath51 and @xmath131 , are functions of @xmath45 and @xmath50 .
the first factor on the rhs is what one should expect for a single metallic electrode : the electronic compressibility is the capacitance of the electrode up to a factor @xmath140 . the further two factors with the derivative @xmath141 account for the charge redistribution between surface and interface which , of course , depends implicitly on the rsoc ( see appendix [ appcompandcap ] ) .
finally , we discuss the size of the experimental values of the control parameters .
we use @xmath142 for the effective mass of the electrons at the lao / sto interface .
this value results from the multi - band evaluation @xcite , addressed in the next section : a light mass @xmath143 and a heavy mass @xmath144 were identified from arpes measurements @xcite for the electrons of the @xmath145- , @xmath146- and @xmath147-bands .
illustrated in the suppl . of ref .
@xcite that , within the multiband model , the electrons of the lowest resulting band acquire an effective mass of @xmath142 close to the band minimum .
estimating @xmath124 and @xmath113 has the largest uncertainty : caviglia _ et al . _
@xcite concluded from their magnetotransport experiments that @xmath148 in the range of backgate voltages where the superconducting dome is formed . for a more negative bias
the interface becomes depleted and the rsoc is nearly constant with a low value of @xmath149 .
we neglect this rather small constant contribution with respect to the term linear in the bias .
the rsoc increases approximately linearly with bias in between values corresponding to the low - density foot and top of the superconducting dome ( compare fig . 3 in ref .
the difference in charge carrier density at foot and top of the dome @xcite is @xmath150 , so that @xmath151 we do not know the exact values for the density at which @xmath37 was measured .
therefore @xmath152 , as deduced from the experiment of caviglia _ et al . _
@xcite , might be somewhat smaller but certainly not more than one order of magnitude .
the free electrons at the lao / sto interface reside predominantly in the titanium @xmath153 orbitals of the first tio layer . for a refined description of the interface
we use a six - band model for the three spin - split @xmath15 orbitals @xcite . within this formalism , the emergence of a strong rashba - like soc is explained naturally through the breaking of inversion symmetry at the interface , which couples the atomic soc of the ti atoms to the inter - band hopping .
the hamiltonian can be written in the @xmath147 , @xmath154 , @xmath145 basis as : @xmath155 the first term is the intra - orbital hopping : @xmath156 @xmath157 is the density of states for light ( @xmath158 ) or heavy ( @xmath159 ) electrons and @xmath160 the splitting between the @xmath145- and the @xmath147- , @xmath154-orbitals .
the contribution from atomic spin - orbit coupling is @xmath161 @xmath162 where @xmath163 is the strength of the atomic soc , @xmath164 is the 2@xmath1652 identity matrix , and @xmath166 are the 2@xmath1652 pauli - matrices . due to the confining potential at the interface and its asymmetry , the @xmath10-orbitals are deformed and hence no longer orthogonal .
this allows ( anti - symmetric ) interorbital hopping between @xmath145- and @xmath146-orbitals along the @xmath167-direction and between @xmath145- and @xmath147-orbitals along the @xmath168-direction , presented by @xmath169 we assume that the magnitude of the hopping is proportional to the asymmetry inducing electrical field . in our modeling , the origin of this electric field is the ( positive ) charge on the surface that compensates for the electrons transferred from the surface to the interface system .
therefore we write : @xmath170 the term proportional to @xmath13 is generated through the asymmetry inducing electrical field , and the @xmath13-independent term @xmath171 can be traced to dipolar distortions in the lao - layer next to the interface .
the diagonalization , which can be done analytically only for @xmath172 , yields three spin split doublets .
@xmath173 , which originates from the deformation , vanishes at @xmath174 and the doubly degenerate energies are @xmath175 for a non - zero @xmath176 , the degeneracies are lifted .
moreover , crossings between different doublets are avoided @xcite ( see fig . [
multibandbands]a ) . with increasing @xmath176 , the splitting between the formerly degenerated bands becomes larger and the point of lowest energy in @xmath24-space moves to larger values of @xmath24 ( cf .
[ multibandbands](b ) ) . , @xmath177 , @xmath178 , @xmath179 , @xmath180 , @xmath181 , and @xmath182/@xmath183 , as for the right side of fig .
[ multibandbands ] .
the energy splitting is plotted from the @xmath184-point to the @xmath185-point ( continuous curves ) and to the @xmath186-point ( dashed curves ) . ]
the shape of the bands depends , analogously to the two - band model with rsoc ( cf .
( [ rsocdispersion ] ) ) , on the electron density .
we add a top gate as in sec .
[ sec : metalliclayer ] and perform the same minimization of the energy of the combined system with respect to @xmath13 for different applied voltages .
the results are shown in fig .
[ multibandncompcap ] , where we chose the light and heavy electron masses @xmath187 and @xmath177 , respectively ( in accordance with the arpes @xcite measurements for @xmath188 surfaces ) .
the energy splitting is taken to be @xmath179 , according to the x - ray measurements in ref .
@xcite . alternatively
, we also consider @xmath189mev , which is in better agreement with various ab initio evaluations ( see , e.g. , ref .
the atomic soc was estimated from transport experiments @xcite and ab - initio calculations @xcite to be @xmath19025mev ; we chose the lower boundary of 10mev .
the @xmath24-dependent energy splitting @xmath191 induced by @xmath176 in the lower and upper doublet increases linearly for small @xmath24 , i.e. , it is rashba - like : @xmath192 ( cf . fig . [ multibandsplitting ] ) .
this can be employed to fit the parameters @xmath171 and @xmath193 to the estimates of @xmath124 and @xmath113 made from experimental results in the previous chapter : by writing @xmath194 we identify @xmath195 and @xmath196 , where @xmath197 is obtained from fitting @xmath198 to the numerically calculated @xmath191 . for @xmath199 and @xmath178 this yields @xmath200 and we find @xmath201 and @xmath202 .
although this estimation for @xmath171 and @xmath193 neglects the dependence of @xmath203 on the angular direction in @xmath24-space , it gives a reasonable approximation for @xmath204 as long as the charge density is small , such that @xmath205 holds .
for larger densities , for which this proportionality fails around the fermi momentum ( i.e. , it is not rashba - like , cf .
[ multibandsplitting ] ) , the above estimation is no longer valid .
the capacitance and compressibility of the system is given mostly by the electrons from the lowest doublet , because for low densities the higher bands are unoccupied or considerably less occupied than the lowest two bands . for larger densities , on the other hand , the variation of the electron number upon changing @xmath50 or @xmath45 is largest in the lowest doublet (
since then the dos is larger in the lower doublet , cf .
[ multibandbands](b ) ) .
the differential capacitance @xmath131 for different gate voltages @xmath50 is displayed in fig .
[ multibandncompcap](b ) .
the difference of the chemical potentials in the interface and the metallic top layer , @xmath77 , was set to a value that gives a charge density @xmath206 for @xmath80 . for @xmath207 ,
i.e. , a density independent bandstructure , the capacitance increases with voltage ( see brown line in fig .
[ multibandncompcap](b ) ) until the fermi energy reaches a value where the deviation from a parabolic band structure becomes negligible and , consequently , the dos is constant .
notice that the capacitance derived from the microscopic model is smaller than the geometric capacitance , since the kinetic energy of the charge carriers in the electrodes adds a positive term to the inverse capacitance ( see , e.g. , ref .
@xcite ) .
a non - zero @xmath193 enhances @xmath131 , even above the geometrical capacitance if @xmath193 is sufficiently large . with increasing voltage @xmath50 , the density @xmath13 at the interface increases ( fig .
[ multibandncompcap](a ) ) , and hence @xmath176 .
this flattens the dispersion of the lower doublet and thereby enhances the effective mass at @xmath208 ( see fig .
[ multibandbands ] ) , which causes the increase of capacitance above the value for @xmath207 ( cf . fig [ multibandncompcap](b ) ) . for large @xmath176 ,
the energy minima of the two lowest bands are far from the zero point and @xmath174 is not occupied . in this regime
, @xmath191 is no longer linear in @xmath24 and the estimations for @xmath171 and @xmath193 are no longer valid .
this region is indicated by the dots of the differential capacitance in fig .
[ multibandncompcap](b ) . within the regime where the bands become flatter with increasing density
, negative compressibility may appear .
the inverse compressibility for different values of @xmath193 is shown in fig .
[ multibandcomp ] as a function of distance between the interface and surface . evidently , fig .
[ multibandcomp](a ) , which presents the inverse compressibility for the 6-band model corresponds to fig .
[ compressibilityoverdal](a ) for the 2-band model . for a small effective distance @xmath65 , all charge resides on the interface ( fig .
[ multibandcomp](b ) ) , the compressibility is negative , and @xmath209 increases nearly linearly with the distance .
this is understood by setting @xmath83 in eq .
( [ mlotorltotalenergy])the only term in the total energy with a dependence on @xmath65and taking the total second derivative with respect to @xmath45 . for large distance
the electrostatic energy dominates and the total energy is minimized by @xmath210 . inserting this into eq .
( [ mlotorltotalenergy ] ) yields half the gradient of the @xmath83 regime , see fig .
[ multibandcomp](a ) . in fig .
[ multibandcomp](a ) the inverse compressibility is displayed for @xmath211 , whereas in fig .
[ multibandcomp](c ) the potential is @xmath212 . for @xmath211 and @xmath213 , the lower band edge at the interface
is shifted by @xmath214 with respect to the band edge of the surface electrode .
then , the electrons tend to go to the interface if no other effect such as the rsoc interferes , provided that the electrostatic energy is small and the compressibility can be negative for sufficiently large @xmath193 . however , the behavior of @xmath209 as a function of @xmath65 changes drastically when we adjust @xmath77 to @xmath215 so that @xmath206 for @xmath216 and @xmath217 .
then @xmath13 is in the range of carrier densities established for the lao / sto interface with @xmath218 .
the inverse compressibility first increases linearly with @xmath65 because @xmath84 in the regime of small distances ( see fig .
[ multibandcomp](c ) ) .
there is a small offset due to the finite compressibility of the surface electronic state .
then , with a further rise in @xmath65 , the polar - catastrophe mechanism induces a finite electron density at the interface ( fig .
[ multibandcomp](d ) ) , and the inverse compressibility decreases sharply ( fig .
[ multibandcomp](c ) ) because a small but finite @xmath13 reduces the electrostatic term linearly with @xmath13 ( see eq .
( [ mlotorltotalenergy ] ) ) . eventually , @xmath209 increases towards its limiting linear behavior for large @xmath65 as in fig .
[ multibandcomp](a ) .
this analysis allows us to conclude that the compressibility stays strictly positive even for large values of the rsoc if the system parameters are adjusted within a range that is consistent with parameter values previously identified for the lao / sto heterostructure . for a thorough determination of the compressibility of the lao / sto heterostructure additional effects may be relevant :
these include oxygen vacancies , their distribution between the planes as well as the lateral extension of the two - dimensional electron liquid at the interface
. however , these effects influence primarily the charge carrier density . as the charge carrier density
is adjusted to the experimental values through a bias @xmath77 in our analysis , we expect that the evaluation of the compressibility remains valid .
low - dimensional charge states at interfaces of complex oxides have been explored intensively in recent years , and it is a combination of scientific and technological interest which drives the research in this field .
this is also valid for the work presented in this article : we explored the electronic charge redistribution in a heterostructure of a polar and a non - polar insulator , the paradigmatic realization of which has become the lao / sto interface of a polar lao film on the non - polar substrate sto . understanding the quantum phenomena that control the electronic reconstruction at the interface , and thereby also the charge redistribution between surface and interface , may provide a possible strategy to control capacitances by the electronic properties of the conducting electrodes as opposed to the standard search for dielectrics with high dielectric constant @xcite .
sizeable capacitance enhancements and negative electronic compressibility have been observed for lao / sto @xcite .
recently , these findings were related to a strong rsoc @xcite .
the rsoc is generated by an electric field that in turn is controlled by the electronic charge density @xmath219 at the interface .
in fact , a band structure , which depends on the density of the band electrons themselves , allows for a decreasing chemical potential with increasing charge carrier density assuming that the rsoc strength is sufficiently large .
such a dependence of the chemical potential on the carrier density represents a negative compressibility .
thus , instead of the standard explanation for a negative compressibility in low density semiconductor systems @xcite or in oxide heterostructures @xcite through exchange and correlation effects , it may be that spin - orbit effects play an essential role .
however , in order to settle the realization of a negative compressibility and the possible formation of an inhomogeneous charge - separated state , one has to investigate at least the compressibility of the complete electronic system ( in the case that all electronic subsystems are coupled ) .
the compressibility of other subsystems which are not comprised of mobile electrons may also contribute and keep the total compressibility positive . in this article
we not only included the mobile interface electrons , but also the surface charges with their respective field .
this is to be considered as a minimal setup for two reasons : ( 1 ) the rsoc is controlled by the electric field of the charges at the surface , and ( 2 ) this electric field and the charge carriers of the surface electrode contribute to the compressibility of the interface - surface system . a remarkable relation , which only involves the capacitance of the interface itself , @xmath220 , and the geometric capacitance @xmath3 of the interface - surface system , allows to determine if this system attains a negative compressibility : @xmath221 where @xmath222 is found from the second derivative of the interface free energy @xmath223 with respect to the interface electronic density @xmath13 .
this relation was derived in appendix [ appcompandcap ] under the conditions that the interface - surface capacitance @xmath224 is positive and that @xmath13 does not assume a boundary value ( that is , neither @xmath84 nor @xmath83 holds ) .
here , @xmath26 is the free energy of the entire heterostructure .
the relation ( [ negcomp1 ] ) is remarkable in two respects : first , the compressibility @xmath51 can become negative even when the capacitance @xmath131 is positive , and second , this stability criterion does not depend explicitly on the compressibility of the surface even though the total inverse compressibility @xmath209 contains a contribution from the surface electron system .
both statements can be traced back to the special interface - surface setup in which the capacitance is not equal to the total compressibility ( see eq .
( [ compressibilitycapacitance ] ) ) . the relation ( [ negcomp1 ] ) can be cast into a more intuitive form ( appendix [ appcompandcap ] ) : @xmath225 this inequality obviously expresses the fact that the system is stable if an additional overall electronic charge @xmath226 does not induce a charge transfer @xmath227 larger than the added overall charge .
the compressibility is negative and the electronic state may charge - separate only if @xmath228 . in order to be explicit
, we considered first a 2-band model for the interface electronic system ( secs .
[ sec : metalliclayer ] and appendix [ sec:3dlayer ] ) .
the bands are split by a rsoc that depends on the charge density of the surface electronic system .
the charge densities at interface and surface were calculated from the minimization of the total free energy .
this simple model allows to identify a series of electronic phases that depend on the occupation of the interface split bands .
the second derivatives of the free energy with respect to the total electronic density @xmath45 and the interface electronic density @xmath13 yield ( phase - dependent ) compressibility and capacitance , respectively .
a negative compressibility arises only for very small effective distances @xmath65 between interface and surface , and strong rsoc ( see fig . [
compressibilityoverdal](a ) ; for example for an rsoc - parameter @xmath229 the effective distance has to be @xmath230 ) . even for rsoc parameters twice as large as the largest rsoc estimated from magneto transport measurements for lao / sto
, the compressibility stays positive ( cf .
[ nandcompoverv](a ) ) , also for large but sensible values of a bias voltage @xmath50 .
this differs from the findings of refs .
the discrepancy originates most prominently from the electrostatic term @xmath231 in relation ( [ negcomp1 ] ) , which is entirely absent in refs .
the more complete treatment of our work yields a compressibility that is positive for the parameter set of the lao / sto heterostructure and therefore retains the stability of the electronic interface - surface system .
there is also a more technical difference : here , the compressibility is found from the second derivative of the free energy ( see relation ( [ kappa ] ) ) . in refs .
@xcite , however , the compressibility is calculated from the derivative of the chemical potential , which was identified from a model with fixed , density - independent rsoc . only for the derivative of the chemical potential , which yields @xmath209 ,
the rsoc was generalized to include the density dependence . evidently , the two approaches lead to diverse results even when the analysis is restricted to a single layer model without surface term ( see appendix [ appenergyvschempot ] ) . in this article
we addressed explicitly the dependence of the capacitance of the heterostructure on the rsoc .
as reiterated above , the differential capacitance @xmath131 is not the compressibility in this scheme .
the differential capacitance is technically found from the second derivative of the free energy with respect to the charge density @xmath13 at the interface .
so @xmath131 is an internal capacitance of the heterostructure , the capacitance between interface and surface .
a 3d conducting film may be deposited on the surface to have a metallic electrode ( gold or ybco , see appendix [ sec:3dlayer ] ) .
the kinetic contribution to the capacitance suppresses the capacitance below its geometrical value when exchange effects and rsoc can be neglected .
however , we find that a strong rsoc can enhance the capacitance with increasing gate bias to values beyond the geometric capacitance . hence , experiments can verify a rsoc through a capacitance enhancement .
a 6-band model is more realistic to investigate the capacitance of lao / sto heterostructures ( sec .
[ sec:3band ] ) .
most importantly , it allows to derive a rashba - like soc through the coupling of an atomic soc ( at the ti sites ) to the @xmath15 bands in the presence of field controlled inversion symmetry breaking at the interface .
the 2-band model is interpreted as an effective model for sufficiently small filling when only one @xmath15 orbital is occupied and the splitting of the bands is still linear in momentum @xmath24 , that is , `` rashba - like '' .
however , already for density independent rsoc and low filling , the capacitance as a function of bias voltage @xmath50 behaves differently than in the 2-band model .
the capacitance increases perceivably ( fig .
[ multibandncompcap](b ) ) , dashed brown line ) for the 6-band model whereas it is constant with respect to @xmath50 for the 2-band model in the absence of a density dependence of the rsoc . in order to find capacitances larger than the geometric capacitance @xmath232
, one has to include the density dependence of the band structure .
however , one has to realize that in the regime of stronger gate bias @xmath50 , a variable number of bands are occupied ( see fig .
[ multibandncompcap](a ) ) and the splitting of the soc band pairs is not anymore linear in @xmath24 ( see fig
. [ multibandsplitting ] , the middle band pair has anyway no linear splitting , even for @xmath233 ) .
then the capacitance as a function of @xmath50 depends strongly on the chosen parameter set ( @xmath15 band splitting , effective masses , soc , orbital deformation ) .
we could identify bias - induced capacitance enhancements of the order of 5% for parameter sets
compatible with the lao / sto heterostructure , but a 30% enhancement as observed in experiments @xcite appears to be beyond the present modelling .
it should be noted that those experiments are in a regime of very low charge carrier densities and , therefore , it is expected that electronic exchange is much more prominent , an effect which would have to be included . for a realistic value of the effective distance between interface and top gate , @xmath216 ,
the total compressibility of the system stays positive even for large values of the parameter @xmath193 that controls the density dependence of the hybridization between the @xmath145 orbital and the @xmath234 orbitals .
there is no instability which produces a phase separation on account of a rashba - like soc .
we portray the basic capacative properties of a polar heterostructure in fig .
[ multibandcompcapq ] by relating the compressibility , the relative value of the interface electronic densitiy and the differential capacitance to a variable polar charge density .
the value of @xmath235 for the lao / sto heterostructure is marked .
we assigned @xmath236 to the largest value of the rsoc that is still compatible with experiments ( see sec .
[ sec:3band ] ) .
the potential @xmath77 controls the shift of the interface @xmath15-bands with respect to the surface states : the lowest value of @xmath77 ( for each of the three sets of curves for given @xmath193 in fig .
[ multibandcompcapq ] ) adjusts the value of the interface electronic density @xmath13 to @xmath237 , which is in the physical range of electronic densities .
this corresponds to @xmath238 in fig .
[ multibandcompcapq](b ) .
it is obvious from fig .
[ multibandcompcapq](a ) that for @xmath235 and for the lao / sto values of @xmath77 , the compressibility is well in a positive range .
however , it is conceivable that for other heterostructures with a polarity of @xmath239 and large rsoc , that is sizeable @xmath193 , a phase separation is attainable with this mechanism and a considerable enhancement of the differential capacitance over the geometrical value can be achieved ( cf .
[ multibandcompcapq](c ) ) .
this work was supported by the deutsche forschungsgemeinschaft through trr 80 .
we are grateful to j. mannhart for his continuing support of our research .
the authors acknowledge helpful discussions with r. ashoori , s. graser , m. grilli , j. junquera , a. p. kampf , lu li , and v. tinkl .
for a single layer the strength of the rsoc @xmath1 depends on external fields only , so that @xmath1 is not a function of the density @xmath13 in that layer .
hence we obtain the relation between number of electrons @xmath240 and chemical potential via the energy dispersion , eq .
( [ rsocdispersion ] ) : @xmath241\notag\end{aligned}\ ] ] with the abbreviation @xmath242 and assuming @xmath243 ( otherwise the heaviside functions and hence the integrals are zero ) , we get @xmath244\end{aligned}\ ] ] in our analysis we have to distinguish between @xmath245 and @xmath22 , where both bands or one band is occupied , respectively : + @xmath246 @xmath247 the same distinction has to be made for the calculation of the free energy .
+ @xmath248 note that the @xmath249 term is zero , since the square root @xmath250 is always smaller than @xmath251 for negative @xmath14 . for @xmath23
this term yields the integral with the smaller upper bound : + @xmath252
we calculate the electrostatic energy of a system similar to a lao / sto multilayer , cf .
[ electrostatic ] : before electronic reconstruction positive and negative charged planes alternate in @xmath253-direction and the lowest layer is both electrical neutral and free of electrical fields . if an electron charge density @xmath219 is transferred from the top to the bottom layer ( fig .
[ electrostatic](a ) ) , an additional field proportional to @xmath13 is superimposed .
note that for the rsoc only the asymmetric fields are relevant , so that the field of the electrons in the bottom layer does not contribute to rsoc there . even if there is only a field in the upper half - space , the field important for soc _ in _ the plane is the field originating from the surface in the complete space .
+ the electrostatic energy of a system with @xmath254 alternating ( double ) layers can be calculated via the energy stored in the fields : the electric field @xmath2 of a planar charge density @xmath255 ( where @xmath256 is the electronic charge ) is given by @xmath257 and the energy per area between two plates with charge density @xmath258 and a distance @xmath259 between them is @xmath260 hence the energy per area of @xmath254 double layers with charge density @xmath261 in an external field , which is generated by a charge density @xmath262 , is @xmath263 fig .
[ electrostatic](b ) shows an equivalent circuit for @xmath264 .
dashed lines refer to @xmath265 . here
@xmath266 , @xmath216 , @xmath235 , and @xmath104 .
@xmath77 ensures that the density on the interface is @xmath267 ( see eq .
( [ totalenergy ] ) for the significance of @xmath77 ) .
the gold electrode has a thickness of @xmath268 , @xmath269 , @xmath270 and the capacitance @xmath271 for zero rsoc .
the system is in phase p2 for all plotted values of @xmath113 , except for the @xmath272 curve , where the system is in phase p@xmath273 to the right of the kink at @xmath274 .
the horizontal dashed brown line indicates the geometric capacitance.,scaledwidth=48.0% ] in some experimental setups , a three - dimensional ( 3d ) top gate is deposited on top of the lao / sto heterostructure ( see , for example , ref .
the electrode connects with the topmost @xmath275 layer and the electrons therein should be treated as a 3d electron system .
this setup requires to consider an inhomogeneous system with field penetration in the top electrode as in ref . .
we assume a space independent dos @xmath276 within the electrode on the ground that this does not modify the qualitative aspects ( in fact , modifications are tiny if the electronic system is not close to a van hove singularity ) .
then the kinetic term from the electrode free energy @xmath277 translates into a correction of the effective distance in the geometric capacitance : the effective distance @xmath65 is increased by @xmath278 , the screening length ( see refs . ) .
so instead of implementing an inhomogeneous electron system in the energy @xmath277 of the top electrode , one may cast the electrostatic energy into the form ( cf .
( [ mlotorltotalenergy ] ) ) : @xmath279 we assume here that the number of electrons in the 3d film is much larger than the number of unoccupied sites in the interface layer with rsoc , so that even though charge is transferred to the bottom layer , the volume density @xmath280 in the metal remains approximately unchanged .
this condition can always be fulfilled by a sufficiently thick metallic layer .
then the screening length is approximately constant characteristic of the used electrode material . in thomas - fermi theory @xmath278
depends on the volume electron density @xmath280 and the dielectric constant @xmath281 of the ion cores in the metal : @xmath282 where @xmath276 is the 3d dos at the fermi energy .
table [ tableoftf ] shows the screening lengths for @xmath283 , @xmath284 and @xmath285 .
.material properties and resulting screening lengths @xmath278 for gold , silver and ybco . here ,
@xmath280 is the volume electron density and @xmath281 the dielectric constant of the ion cores in the metal . [
cols="<,^,^,^,^",options="header " , ] [ tableoftf ] note that @xmath45 and @xmath280 are fixed , as they are determined by the polarity of the lao layers and the choice of the top gate film .
in contrast , the electron density @xmath13 is found from the minimization of the free energy . in fig .
[ goldybco ] the differential capacitance @xmath131 and @xmath132 are displayed for @xmath283 as top gate .
we assumed a screening length of @xmath286 for gold .
for the weakest rsoc , @xmath287 , the capacitance is not altered notably in comparison to the case without rsoc . for the largest value @xmath272
( see estimation in eq .
( [ approxalpha1 ] ) ) , we find a capacitance increase of approximately 8 % in a voltage range 0100 v. this is the capacitance increase with respect to the case of zero rsoc .
the increase with respect to the geometrical capacitance is 5 % for a gold top electrode .
different screening lengths cause sizeable differences in magnitude of the capacitances : the screening length in a @xmath285 film is @xmath288 , and the capacitance increase is approximately 3 % in the voltage range 0100 v. the increase with respect to the geometrical capacitance is only 1 % for the ybco top electrode .
in this section we assume that the electron density at the interface , @xmath13 , does not take one of the boundary values , 0 or @xmath45 .
moreover , we assume that the appropriate free energy is minimal with respect to @xmath79 ( see below ) .
the helmholtz free energy @xmath42 of the system is a function of the total electron density @xmath45 and the electron density on the interface @xmath13 ( cf .
( [ totalenergy ] ) ) : @xmath289 the partial derivative of @xmath49 with respect to @xmath13 is the voltage between the plates : @xmath290 this thermodynamic relation may be used to identify @xmath79 .
we introduce a legendre transformation of @xmath108 with respect to the variable @xmath13 : @xmath291 where @xmath85 depends on the total electron density @xmath45 and the external potential @xmath50 between interface and surface electrode .
the differential capacitance can be derived either from eq .
( [ appminimization ] ) @xmath292 where we suppress the label for the variable @xmath45 in @xmath131 , or from the legendre transform eq .
( [ appsecondfv ] ) : @xmath293 of course , the two capacitances from eqs .
( [ appcapacitance ] ) and ( [ appcapacitancev ] ) are identical in the sense that @xmath294 holds .
on the other hand , the inverse compressibility @xmath295 in dependence on an external voltage @xmath50 is the second total derivative of @xmath26 with respect to the total electron density : @xmath296\notag\\ & = \partial_q^2 f+2\partial_n\partial_q f \left(\partial_q n\right)+\partial_n^2 f\left(\partial_q n\right)^2\notag\\ & \qquad + \partial_n f\left(\partial_q^2n\right)\label{appcomp2}\end{aligned}\ ] ] note that @xmath297 is not a derivative with respect to a thermodynamic variable .
in fact , the thermodynamic variables are @xmath45 and @xmath50 for @xmath26 .
the partial derivative of @xmath26 with respect to @xmath13 is identified from the rhs of eq .
( [ appsecondfv ] ) .
the thermodynamic potential @xmath26 is minimal with respect to all values of the internal variable @xmath13 , a necessary condition for thermodynamic stability .
according to eqs .
( [ appsecondfv ] ) and ( [ appminimization ] ) , the last term on the right hand side of eq .
( [ appcomp ] ) is zero . under a further assumption we can derive more elaborate relations between differential capacitance and compressibility :
the dependence of the total free energy @xmath42 on the total electron density @xmath45 shall be exclusively of the form @xmath138 , so that also @xmath298 is valid .
this assumption is reasonable , since @xmath138 is the electron density in the surface layer , which should be the only density dependent quantity that determines the energy of the surface system .
hence we only demand the electrostatic energy not to have some unusual form .
note that , while @xmath299 is sufficient for the derivation below , only eq .
( [ appsecondassumption ] ) is necessary . under the validness of this relation
the compressibility becomes : @xmath300 we use the total derivative of the first condition eq .
( [ appminimization ] ) , @xmath301 to replace @xmath302 in the compressibility above : @xmath303 this relation between the compressibility and differential capacitance can be further analyzed to deduce a stability criterion , i.e. , the condition for positive compressibility . since we assumed that the density @xmath13 obtained from eq .
( [ appminimization ] ) yields a minimum of the free energy , @xmath304 the differential capacitance @xmath131 is positive , which is a necessary criterion for the stability .
note that the second partial derivatives of @xmath26 and @xmath108 ( with respect to @xmath13 and @xmath45 ) are equal .
hence there is an addtional condition to keep the compressibility positive : @xmath305 the system is unstable if adding of electrons causes either electrons to flow from the interface to the surface ( @xmath306 ) or more than the added electrons to flow to the interface ( @xmath307 ) . as we will show below , the first alternative does not occur in our model .
@xmath306 is in principle the reverse case of @xmath307 , i.e. , when the isolated surface system would display a negative compressibility .
+ furthermore , eq . ( [ appderivofmini ] ) can be solved for @xmath308 , @xmath309 this result inserted into the compressibility eq .
( [ appcompdn ] ) yields : @xmath310 we consider a system where the electrons in the surface electrode are not subject to a rsoc or other effects which may produce a negative compressibility .
correspondingly , we may assume that @xmath311 . as the electrostatic contribution to the free energy also generates a positive second derivative with respect to @xmath45 , the free energy @xmath26 has to obey @xmath312 since @xmath313 is negative ( eq . ( [ appsecondassumption ] ) with relation ( [ appf2q ] ) ) and
@xmath314 is positive in the considered regime , eq .
( [ appderivofmini ] ) shows that @xmath308 has to be positive in our model .
we consider the regime of positive capacitance
is it then possible to identify a negative compressibility of the complete electronic system ?
in fact , according to eq .
( [ appcompdiff ] ) and the positivity of @xmath315 , the compressibility is negative if @xmath316 this inequality does not involve properties of the surface system explicitly .
with the assumption that @xmath131 is positive , one finds that the compressibility of the heterostructure is negative , if the sum of the inverse compressibility of the interface system ( times @xmath317 ) and half the inverse geometric capacitance @xmath318 is smaller than zero .
we emphasize that the inverse capacitance is a sum , the terms of which are generated by the second derivatives of different contributions to the helmholtz free energy @xmath42 in eq .
( [ appcapacitance ] ) , consistent with previous analyses ( e.g. , ref .
the inverse compressibility is not the sum of these term : it is the asymmetric setup that causes @xmath319 .
however , the relation ( [ appnegcomp ] ) contradicts the identification of a negative compressibility in refs .
@xcite : there the authors missed the `` electrostatic compressibility '' @xmath318 entirely and the compressibilty of the interface electronic system was evaluated differently ( see appendix [ appenergyvschempot ] ) . in our model
the total free energy @xmath108 is of the form presented in eq .
( [ totalenergy ] ) , which reads with the abbreviation @xmath320 : @xmath321 due to the analytical form of @xmath322 , eq .
( [ appcompdiff ] ) yields a direct relation between the measured differential capacitance and the total compressibility : @xmath323\end{aligned}\ ] ] and the transfer of charge to the interface is : @xmath324 obviously , for the symmetric case without rsoc and with equal dos at the interface and surface electrode , @xmath325 , one finds @xmath326 and , consequently , @xmath327 . for effective interface - surface distances
@xmath328 , the electrostatic term @xmath329 dominates over the inverse compressiblities @xmath330 , and the charge transfer @xmath141 is close to @xmath331 for vanishing rsoc .
therefore , sizeable deviations of @xmath141 from @xmath331 can be indicative of a substantial rsoc in these systems . with the measurement of @xmath131 and knowledge of @xmath232 and @xmath56
one can determine @xmath141 and @xmath51 through eqs .
( [ appdndq ] ) and ( [ appkappac ] ) , respectively .
the electronic transfer @xmath141 is displayed in fig . [ multibanddndq ] as a function of the density dependent hybrization , which effectively controls the strength of the rsoc ( see sec .
[ sec:3band ] ) .
the limiting value of @xmath332 is reached , when for large rsoc all additional charge carriers are accumulated at the interface ( horizontal lines in fig .
[ multibanddndq ] ) .
( a ) , and the corresponding interface electronic density @xmath13 ( b ) , both with respect to @xmath193 , which tunes the density dependence of the rsoc . the density @xmath13 is controllel by @xmath77 . here , @xmath333 , @xmath177 , @xmath334 , @xmath216 , @xmath335/@xmath183 , @xmath178 , @xmath179 , and @xmath336 .
we find @xmath337 for @xmath338 .
the jump to the horizontal line with @xmath339 in ( a ) takes place when all electronic charge is accumulated at the interface , see ( b ) . ]
the first derivative of the energy with respect to the electron density @xmath13 yields the chemical potential while the second derivative generates the inverse compressibility @xmath209 .
we take the strength of the rsoc @xmath1 density dependent , and the density dependence is implemented already in the free energy
. however , caprara _ et al . _
@xcite determine the chemical potential from a model with given @xmath1 , then introduce the density dependence of @xmath1 at that level and calculate the inverse compressibility @xmath340 from the first derivate of the chemical potential with respect to @xmath13 .
here we compare the resulting compressibilities @xmath51 and @xmath341 of these two approaches .
therefor we use the abbreviation @xmath342 , which equals the lowest energy in the dispersion eq .
( [ rsocdispersion ] ) , and rewrite the energy eq .
( [ mlotorlrashbaenergy ] ) of a layer with rsoc : @xmath343 for density independent @xmath1 and hence density independent @xmath344 , the chemical potential @xmath14 is given by @xmath345 caprara _ et al . _
@xcite then set @xmath346 , and their inverse compressibility is @xmath347\\ \displaystyle \frac{1}{n_0}-2\zeta ' \end{array}\right.\end{aligned}\ ] ] here @xmath348 and @xmath349 .
in contrast , if we set @xmath346 already in the energy @xmath350 in eq .
( [ appevcpenergy ] ) : @xmath351 & , \mu<0\\ \displaystyle \mu_{\rm c}^>-\zeta'\left[2-\frac{4}{3}n_0\zeta\right ] & , \mu>0 \end{array}\right . .\end{aligned}\ ] ] for @xmath354 this simplifies to @xmath355 and @xmath356 as one can see from eqs .
( [ kappac ] ) and ( [ kappas ] ) , @xmath357 and @xmath358 differ by a factor of 3 .
the deviation for @xmath359 is less since @xmath360 dominates ( [ kappac ] ) and ( [ kappas ] ) for realistic values of @xmath13 , @xmath113 , and @xmath125 . the charge density , which generates the electric field in the rsoc , is not the interface charge carrier density @xmath13 the latter produces a symmetric electric field which does not entail a rsoc .
it is rather the compensating charge in the laalo@xmath0 layers and at the laalo@xmath0/srtio@xmath0 surface that generates the field which controls the rsoc in the interface layer
. one might be lead to the conclusion that this is a minor detail as the compensating or `` external '' charge is of the same magnitude as @xmath13 .
however , the thermodynamic expression for the compressibility does not allow to vary the external charge and , therefore , a negative compressibility seems to be difficult to realize in this scheme .
santander - syro , o. copie , t. kondo , f. fortuna , s. pailhs , r. weht , x. g. qiu , f. bertran , a. nicolaou , a. taleb - ibrahimi , p. le fvre , g. herranz , m. bibes , n. reyren , y. apertet , p. lecoeur , a. barthlmy , and m. j. rozenberg , nature * 469 * , 189 ( 2011 ) .
also bucheli _ et al .
_ consider a 6-band model in appendix b of ref .
however their spin - orbit hamiltonian is not hermitian , and the splitting between the two lowest bands and the second pair of bands is approximately 18 mev ( fig .
16 in ref .
this appears to be a different modelling .
rolf aidam , dissertation : _ untersuchung des epitaktischen wachstums dnner @xmath361-schichten und ihre anwendung in ferroelektrischen supraleitenden feldeffekttransistoren _ , forschungszentrum karlsruhe ( 1999 ) . | oxide heterostructures with polar films display special electronic properties , such as the electronic reconstruction at their internal interfaces with the formation of two - dimensional metallic states .
moreover , the electrical field from the polar layers is inversion - symmetry breaking and generates a rashba spin - orbit coupling ( rsoc ) in the interfacial electronic system .
we investigate the quantum capacitance of a heterostructure in which a sizeable rsoc at a metallic interface is controlled by the electric field of a surface electrode .
such a structure is , for example , given by a laalo@xmath0 film on a srtio@xmath0 substrate which is gated by a top electrode .
such heterostructures can exhibit a strong enhancement of their capacitance @xcite .
the capacitance is related to the electronic compressibility of the heterostructure , but the two quantities are not equivalent .
in fact , the transfer of charge to the interface controls the relation between capacitance and compressibility .
we find that due to a strong rsoc , the quantum capacitance can be larger than the classical geometric value .
however , in contrast to the results of recent investigations @xcite the compressibility does not become negative for realistic parameter values for laalo@xmath0/srtio@xmath0 and , therefore , we find that no phase - separated state is induced by the strong rsoc at these interfaces . |
madan and seneta @xcite introduced the univariate `` variance gamma '' ( @xmath0 ) process as a model for a financial asset price process with a special view to more accurate option pricing on the asset , beyond the standard geometric brownian motion ( @xmath1 ) model .
the @xmath0 model has proved to be outstandingly successful in this application , and is in common use by many financial institutions , as an alternative to the @xmath1 model , despite failing to cater for stochastic volatility or a leverage effect , or allowing for temporal dependence of absolute values and/or squares of log returns .
nevertheless , madan and seneta extended the @xmath0 model @xcite to a multi - asset version , again with a view to important applications in finance ( rainbow options " ) , by subordinating a multivariate brownian motion with a single univariate gamma process ( also see @xcite ) .
this construction leaves the marginal processes as univariate variance gamma processes .
but , components are dependent by virtue of the common time change .
semeraro @xcite generalises the multi - asset version of madan and seneta @xcite to allow for multivariate subordination .
this permits the dependence structure between asset prices to be modeled in a more flexible way .
the economic intuition behind multivariate subordination is that each asset may have an idiosyncratic risk with its own activity time and a common risk factor , with a joint activity time for all assets . in specific cases it is possible to maintain @xmath0 processes for each single asset ,
see @xcite and related applications in luciano and semeraro @xcite , though this may be sacrificed for more flexible dependence modeling , as in guillaume @xcite .
ballotta and bonfiglioni @xcite give an up - to - date account of modeling based on lvy processes in finance .
to summarize , a wide range of multi - asset models based on univariate or multivariate gamma subordination of a brownian motion has been proposed . however , there are still gaps in the literature concerning the characterization in general of classes of processes when the class is required to be stable under convolution .
further , theoretical results such as formulae for characteristic functions , lvy measures and , when possible , transition densities , are needed for a comprehensive description of key properties . as a prominent link between the real - world and the risk - neutral measure
so as to obtain an equivalent martingale measure , we advocate the esscher transform .
this approach has been rigorously investigated in the seminal papers of kallsen and shiryaev @xcite and eberlein , papapantoleon and shiryaev @xcite .
esche and schweizer @xcite show that the minimal entropy martingale measure for a multivariate lvy process preserves the lvy property because it corresponds to an esscher transform .
rschendorf and wolf @xcite provide explicit necessary conditions for the existence of esscher parameters for multivariate lvy processes and further show that the multivariate esscher parameter is unique if it exists .
tankov @xcite provides an introduction to the pricing theory in the context of exponential lvy processes .
the aim of the present paper is to contribute to filling the above gaps by presenting a general class of @xmath2-valued stochastic processes , constructed by subordinating multivariate brownian motion with a subordinator drawn from a suitable class of multivariate subordinators .
our intention is to lay out a systematic formulation suitable for future development . for the new processes
, we provide the formulae mentioned in the previous paragraph and link the real world and pricing measures by calculating esscher transforms . to illustrate the practical possibilities ,
we show how the explicit formulae can be used to price american and european multi - asset options .
the most general class of subordinators we consider is thorin s @xcite class of _ generalised gamma convolutions_. we call it the @xmath3 class of subordinators , and the process formed by subordinating brownian motion in @xmath2 with such a process we call a _ variance generalised gamma convolution _
( @xmath4 ) process .
grigelionis @xcite constructed such a vgg - class , which we call @xmath5 in the present paper .
the @xmath5 class contains madan - seneta s @xmath0 as well as the multivariate generalised hyperbolic distributions @xcite as special cases . complementing grigelionis @xmath5 class , we introduce the @xmath6 class of lvy processes .
our @xmath6 class includes a variety of previously derived models such as semeraro s @xmath7-processes @xcite and guillaume s process @xcite .
the general @xmath8 class extends the @xmath0 classes in a number of ways .
in particular , the @xmath4 classes allow for infinite variation and heavy tails .
figure [ diagram ] depicts the connections between the various subordinated classes .
@xmath9{000}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & { vgg}^{d,1}\\ & & & & \vdots&&\vdots\\ & & & & \gamma_l^d & { \mathrel{\rotatebox[origin = c]{000}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & { ggc}^{d}_l\\ & & & & & & \\\\ v\alpha\gamma^d & { \mathrel{\rotatebox[origin = c]{000}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & v{{\cal g}}\gamma^d & { \mathrel{\rotatebox[origin = c]{000}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & v\gamma_s^d & { \mathrel{\rotatebox[origin = c]{000}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & { vgg}^{d , d}\\\\ & & & { \mathrel{\rotatebox[origin = c]{340}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & v{{\cal m}}\gamma_s^d & { \mathrel{\rotatebox[origin = c]{20}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & \end{array}\ ] ] the thorin @xcite generalized gamma convolutions provide a very natural class of distributions on which to base our multivariate @xmath0 generalizations .
as we will show , they facilitate construction of a very general class of subordinators and corresponding multivariate lvy processes obtained as subordinated @xmath10dimensional brownian motions .
our new class complements @xcite , and contains a number of currently known versions of multivariate @xmath0 processes , and extends them significantly in a variety of important ways .
although rather technical in appearance , our approach is very much directed toward practical usage of the methodology .
explicit expressions for characteristic functions or laplace transforms , and lvy measures or densities , are derived and exhibited for all our processes .
this permits easy programming of option pricing routines as we demonstrate by an example , focusing in particular on the pricing of american style options on a bivariate underlying ; a thorny problem not often tackled in this context .
the @xmath4-class and some of its subclasses are shown to be invariant under esscher transformation , so the risk - neutral distribution constructed as the esscher transformation of a particular member is also in the @xmath4-class with the practical implication that it is possible to calculate prices under the risk - neutral measure and estimate model parameters by calibrating to market data . using these concepts , we set up a market model and show how an option based on multiple assets may be priced . for illustration
we restrict ourselves in this respect to a further subclass of the @xmath4-class which we term the @xmath11-processes .
models in this class have the virtue of allowing a quite general dependency structure between the coordinate processes . as an example
, we price best of and worst of european and american put options , using a tree - based algorithm .
the paper is organised as follows .
section [ secvg ] contains the theory . in subsection
[ subsecvggc ] , we revise the madan - seneta @xmath0 model and set out two major extensions : the variance - univariate @xmath12 and the variance - multivariate @xmath12 classes .
this necessitates recalling , first , some basic facts about gamma subordinators , and then outlining thorin s _ ggc_-class .
subsection [ subsecmoments ] gives some results on the moments and sample paths of the new processes , including a computation of blumenthal - getoor indices . in subsection [ subsecaltstelzerabreu ]
we investigate relations between our @xmath13-classes and various multivariate classes of infinitely divisible distributions , as introduced in prez - abreu and stelzer @xcite , which we call @xmath14-classes .
subsection [ subsecesscher ] calculates esscher transforms , stating the fact that both @xmath13-classes remain invariant .
subsection [ subsecvgamma ] introduces the variance-@xmath15-subclass on which we base the option pricing model in section [ secapplic ] .
finally , subsection [ subsecaltsubordclass ] collects further properties of our subordinator class , including comparisons with those occurring in the literature .
section [ secapplic ] contains applications . here
the market model is introduced , risk - neutral valuation is discussed , and in subsection [ subsecoption ] we price some cross - dependence sensitive options of both european and american types .
some illustrations of the kinds of dependencies the models allow are also given there .
proofs of subsections [ subsecvggc][subsecesscher ] and necessary methodological tools are relegated to section [ secproof ] , where polar decomposition of measures and subordination are briefly covered .
* preliminaries . *
@xmath2 is the @xmath10-dimensional euclidean space ; elements of @xmath2 are column vectors @xmath16 .
let @xmath17 denote the euclidean product and @xmath18 be the euclidean norm , and set @xmath19 for @xmath20 and @xmath21 .
for @xmath22 let @xmath23 .
@xmath24 denotes the indicator function . the dirac measure with total mass at @xmath25
is @xmath26 .
@xmath27 denotes the identity function .
@xmath28 is a @xmath10-dimensional lvy process if @xmath29 has independent and stationary increments , @xmath30 and the sample paths @xmath31 are cdlg functions
. the law of a lvy process @xmath29 is determined by its characteristic function via @xmath32 with lvy exponent , for @xmath33 , @xmath34 , @xmath35 here @xmath36 , @xmath37 is a symmetric nonnegative definite matrix , @xmath38 is a nonnegative borel measure on @xmath39 satisfying @xmath40 and @xmath41 is a given norm on @xmath2 .
we write @xmath42 whenever @xmath29 is a @xmath10-dimensional lvy process with canonical triplet @xmath43 ; @xmath44 refers to brownian motion with drift @xmath45 and covariance matrix @xmath46 .
paths of @xmath29 are of ( locally ) _ finite variation ( @xmath47 ) _ whenever @xmath48 and @xmath49 in this case , we write @xmath50 with @xmath51 denoting the drift of @xmath29 : @xmath52 .
a @xmath10-dimensional lvy process @xmath53 with nondecreasing components is called a @xmath10-dimensional _ subordinator _ , possibly with drift @xmath54 , written @xmath55 .
a general lvy process @xmath42 is a subordinator with drift @xmath51 if and only if @xmath50 with @xmath56 and @xmath38 being concentrated on @xmath57 .
we write @xmath58 and @xmath59 whenever @xmath60 and @xmath61 , respectively , where @xmath62 denotes the law of a random variable or stochastic process @xmath29 .
there is a correspondence between infinitely divisible distributions and lvy processes @xmath29 : for all @xmath33 the law of @xmath63 , @xmath64 , is infinitely divisible .
vice versa , any infinitely divisible borel probability measure @xmath65 on @xmath2 determines uniquely the distribution of a lvy process via @xmath66 .
this connection is used throughout the paper .
for instance , we write @xmath67 to indicate that @xmath53 is a subordinator with @xmath68 . see @xcite for basic properties of lvy processes and their applications in finance .
+ * subordination . * in @xcite various kinds of subordination
are introduced ( see subsection [ subsecsubord ] for details ) . in the present paper
, we will make use of two extreme cases : univariate and ( strictly ) multivariate subordination .
let @xmath69 be a @xmath10-dimensional lvy process .
@xmath29 serves as the subordinate .
given a univariate subordinator @xmath53 , independent of @xmath29 , define a @xmath10-dimensional lvy process , denoted @xmath70 , by setting @xmath71 in the sequel , we denote the law of @xmath70 by @xmath72 .
we refer to this type as _ univariate subordination _
( cf . section 6 in @xcite ) .
suppose @xmath29 has _
independent components _ @xmath73 .
let @xmath74 be a @xmath10-dimensional subordinator , independent of @xmath29 , and define a @xmath10-dimensional lvy process by setting @xmath75 the law of @xmath76 is denoted by @xmath77 .
[ countermultivarexample ] when dealing with strictly multivariate subordination , we have to restrict the class of admissible subordinates @xmath29 to lvy processes with
_ independent components_. this is necessary if we are to stay in the class of lvy processes .
for instance , let @xmath78 be a univariate standard bm .
then @xmath79 is a lvy process , but @xmath80 is not . *
gamma subordinator . *
denote by @xmath81 a gamma distribution with parameters @xmath82 , i.e. , a borel probability measure on @xmath83 , absolutely continuous with respect to lebesgue measure , defined by @xmath84 we write @xmath85 for a _ gamma process _
@xmath86 with parameters @xmath7 , @xmath87 , that is , @xmath88 is a univariate subordinator having marginal distributions @xmath89 , @xmath90 .
further , for @xmath91 , @xmath90 , recall that @xmath92 ( the first formula is well known , the second identity is frullani s integral , see p.16 & p.73 in @xcite ) . in particular ,
a gamma process has zero drift , and its lvy measure admits a lebesgue density @xmath93 .
for @xmath94 we have @xmath95=1 $ ] .
@xmath88 is then also called a _ standard gamma process _ , briefly @xmath96 . +
* madan - seneta @xmath97 process .
* madan and seneta @xcite ( for extensive investigations and reviews @xmath98 .
@xcite ) suggest subordinating brownian motion with a gamma process .
for the parameters of this model we assume @xmath99 , @xmath100 and @xmath21 , with @xmath46 being symmetric and nonnegative definite .
let @xmath101 be a @xmath10-dimensional brownian motion and @xmath102 be independent of @xmath103 .
a lvy process @xmath104 is a @xmath10-dimensional _ variance gamma @xmath105_-process with parameters @xmath106 whenever @xmath107 , which we write as @xmath108 ( a ) note that a @xmath109-process has zero drift and is of finite variation .
+ ( b ) the laplace transformation of @xmath104 takes on an explicit form , straight - forwardly derived from conditioning : for @xmath110 , @xmath111 with @xmath112 , @xmath113 ( c ) if @xmath46 is invertible , for @xmath90 the distribution of @xmath114 is absolutely continuous with respect to the lebesgue measure with the following density : @xmath115 further , still with @xmath116 , the canonical lvy measure of @xmath104 is absolutely continuous with respect to lebesgue measure and satisfies : @xmath117 here @xmath118 is the _ modified bessel function of the second kind _ ( see @xcite , their equations ( 3.471)9 and ( 8.469)3 ; also @xcite , their appendix ) . it is convenient to revise the following facts about the modified bessel function @xmath118 of the second kind : for @xmath119 , @xmath120 @xmath121 for large values of @xmath122 ( see @xcite , their equation ( 8.451)6 ) , @xmath123 we use a variant of @xmath118 , defined by @xmath124 [ remggcexplan ] we get from that the shifted gamma process @xmath125 , @xmath33 , with @xmath85 has laplace transform for @xmath126 @xmath127 for independent processes @xmath128 with @xmath129 , @xmath130 , introduce the associated shifted gamma processes by @xmath131 , @xmath33 , @xmath132 .
it follows from the independence that for @xmath126 @xmath133 where @xmath134 and @xmath135 is the discrete measure associated with the increasing function @xmath136 , @xmath137 .
using the representation in the last display and taking suitable limits in distribution , we arrive at the class of generalised gamma convolutions . in the multivariate setting in subsections [ subsecvgamma][subsecaltsubordclass ] , we shall employ finitely supported thorin measures , such as @xmath138 above , in the context of the @xmath15class .
* generalised gamma convolution subordinator ( @xmath139 ) . * the class of gamma distributions is not closed under convolutions . to extend the madan - seneta @xmath109-class it is convenient to use the subordinators corresponding to thorin s class @xcite of generalised gamma convolutions ( @xmath3 ) .
this is the smallest class of distributions that contains all gamma distributions , but is closed under convolution and weak convergence ( see @xcite ; for multivariate extensions see @xcite ) .
the class of ggc - distributions is a subclass of the self - decomposable distributions and , thus , the distributions are infinitely divisible .
a @xmath10-dimensional _ thorin measure _
@xmath140 is a borel measure on @xmath141 with @xmath142 ( throughout @xmath143 denotes the decomposition of an extended real number @xmath144 into positive and negative parts . )
a subordinator @xmath53 is a _
@xmath145-subordinator with parameters _
@xmath146 and @xmath140 , in brief @xmath147 , when @xmath140 is a @xmath10-dimensional thorin measure , @xmath148 and , for all @xmath149 , @xmath150 the distribution of a thorin subordinator is determined by @xmath146 and @xmath140 .
any thorin measure @xmath140 admits a polar representation @xmath151 relative to @xmath152 ( see lemma [ lempolarrec ] below ) .
this allows us to specify the corresponding lvy measure . for @xmath153 with @xmath151
we have @xmath154 with @xmath155 see @xcite , their theorem f and proposition 4.3 , respectively .
+ * variance univariate ggc @xmath156 . * as a first extension of the @xmath109-model , we review grigelionis
@xcite class .
grigelionis used univariate subordination @xmath157 and subordinated brownian motion with a univariate @xmath3-subordinator . for the parameters of his model we take @xmath99 , @xmath158 and @xmath21 , with @xmath46 being symmetric and nonnegative definite .
further , let @xmath140 be a univariate thorin measure .
let @xmath101 be a @xmath10-dimensional brownian motion , @xmath159 , independent of @xmath103 .
given such @xmath103 and @xmath53 , we call a lvy process of the form @xmath160 a @xmath10-dimensional _ variance univariate generalised gamma convolution @xmath161@xmath5@xmath162_-process with parameters @xmath163 .
we write this as @xmath164 the next theorem gives the characteristic function and lvy density .
part ( a ) ( our equation ) is proved in subsection [ subsecproofvggc ] .
part ( b ) occurs in @xcite ( see his proposition 3 ) .
( in , @xmath165\to{\mathbb{c}}$ ] denotes the principal branch of the logarithm . )
[ theogvg ] let @xmath166 .
+ ( a ) for all @xmath34 , @xmath33 , we have @xmath167 with @xmath168 ( b ) assume @xmath169 and @xmath170 .
then @xmath171 is absolutely continuous with respect to @xmath10-dimensional lebesgue measure on @xmath39 , where , for @xmath172 , @xmath173 besides this , we have @xmath174 in euclidean polar coordinates with lebesgue surface measure @xmath175 on @xmath176 ( for @xmath177 we interpret @xmath175 as the counting measure ) , for @xmath178 , @xmath179 , @xmath180 [ remvggcdvs1 ] in both classes , @xmath109 and @xmath5 , we subordinate a brownian motion with a single univariate subordinator .
thus the components of these processes must jump simultaneously . to allow the components to jump independently of each other we must use multivariate subordination of brownian motion .
this motivates our next step , the introduction of our @xmath6-class .
* variance multivariate ggc @xmath181 . *
next we give another modification of the @xmath109-model which is constructed by multivariate subordination @xmath182 .
for the parameters of this model we assume a @xmath10-dimensional thorin measure @xmath140 , @xmath99 , @xmath148 and @xmath21 , with @xmath183 having nonnegative entries .
( we impose on brownian motion the requirement to have independent components , so as to stay in the class of lvy processes , see remark [ countermultivarexample ] . ) let @xmath101 be a brownian motion .
let @xmath147 be independent of @xmath103 .
given such @xmath103 and @xmath53 , we call a lvy process of the form @xmath184 a @xmath10-dimensional _ variance multivariate generalised gamma convolution @xmath161@xmath6@xmath162_-process with parameters @xmath163 .
we write this as @xmath185 to state formulae for the characteristics of this process , define the outer @xmath186-products of @xmath187 and @xmath21 as @xmath188 we can decompose @xmath189 into semi - cones , where @xmath190 and @xmath191 are the unit coordinate vectors .
analogously , we can decompose @xmath192 into @xmath193 for @xmath194 .
let @xmath195 be the cardinality of @xmath196 .
we need a family of reference measures . with @xmath197 denoting the univariate lebesgue measure define @xmath198 as the product measure with the following factors : @xmath199 observe @xmath200 .
finally , to provide an analog of in euclidean polar coordinates , let @xmath201 denote the lebesgue surface measure on @xmath202 for @xmath203 .
( in the discrete case we interpret @xmath201 as counting measure ) .
the next theorem gives the characteristic function of @xmath104 and an expression for its lvy measure .
it is proved in subsection [ subsecproofvggc ] .
[ theogvgmulti]let @xmath204 . + ( a ) for all @xmath34 , @xmath33 , we have @xmath205 with @xmath206 ( b ) always , @xmath207 , with @xmath208 for @xmath209 . if @xmath210 then @xmath211 .
otherwise , if @xmath212 and @xmath213 then @xmath214 is absolutely continuous with respect to @xmath215 and , for @xmath216 , @xmath217 whereas in @xmath218 , for @xmath178 , @xmath219 , @xmath220 in subsection [ subsecaltsubordclass ] , we see that both @xmath4-classes support pure jump processes with infinite variation and infinite moments . in propositions [ propggcmoments][propvggcmoments ] ,
we provide conditions on the thorin measure that can be used to check local integrability of @xmath221 and @xmath171 as well as existence of moments ( for a proof see subsection [ subsecproofmoments ] ) .
set @xmath222 and @xmath223 .
[ propggcmoments ] let @xmath90 , @xmath224 , @xmath225 and @xmath153 . then : + ( a ) @xmath226 .
+ ( b ) @xmath227<\infty\;\quad\leftrightarrow\quad\int_{0<\|x\|\le 1 } { { \cal t}}({{\rm d}}x)/\|x\|^{p}<\infty$ ] . [ propvggcmoments ] let @xmath228 and @xmath229 .
then we have : + ( a ) let @xmath230 . if @xmath231 then @xmath232 if , in addition , @xmath233 then and are equivalent
. + ( b ) let @xmath234 . if @xmath235^d_*}{{\cal t}}({{\rm d}}x)/\|x\|^{p/2}_k<\infty&\mbox{when $ \mu=0$}\\ \int_{[0,1]^d_*}{{\cal t}}({{\rm d}}x)/\|x\|^{p}_k<\infty&\mbox{when $ \mu\neq 0 $ } \end{array}\right.\,\ ] ] then @xmath236<\infty\,.\ ] ] if , in addition , either @xmath237 , or @xmath238 and @xmath233 , then and are equivalent . [ remgigggcresavenue ] subordinating brownian motion with independent subordinators from the generalised inverse gaussian ( gig)-subordinators one obtains the class of generalised hyperbolic lvy processes ( @xcite for detailed accounts ) .
halgren @xcite identified the univariate gig distributions as generalised gamma convolutions , with the implication that the associated class of hyperbolic lvy processes @xcite forms a subclass of the @xmath5-class ( see @xcite and @xcite , his example 1 ) .
owing to proposition [ propvggcmoments ] , we may restrict our analysis to the subordinator class .
for @xmath239 the @xmath240-distribution has the following probability density on @xmath241 @xmath242 here @xmath243 is a normalising constant .
for @xmath244 the identity @xmath245 holds , and exponential moments @xmath246 $ ] are finite for @xmath247 , and this extends to other parameters @xmath248 as long as @xmath87 . for @xmath249 and @xmath250 observe @xmath251inv@xmath252 is the inverse gamma distribution with finite @xmath253-moments of order @xmath254 only .
we determine the blumenthal - getoor index @xcite as an indicator for activity of the associated @xmath255-subordinator , for potential applications see @xcite . in , note @xmath256 } x^q \pi_g({{\rm d}}x)=\alpha\int_0 ^ 1 x^{q-1 } e^{-\beta x}{{\rm d}}x$ ] is finite for all @xmath224 , forcing the blumenthal - getoor index of @xmath245 to degenerate to zero . for the remaining parameters , where @xmath257 , we compute the laplace transform of @xmath240 for @xmath258 as @xmath259 as follows from . observe @xmath260 as @xmath261 and thus @xmath262 as @xmath263 by a tauberian theorem ( see @xcite , p.75 ) with associated blumenthal - getoor index 1/2 .
[ remcmygasvgggcconjecture ] other possible extensions of the @xmath264-class comprise a range of possible sample path behaviour . in @xcite
the univariate @xmath265-processes have been identified to be subordinated brownian motions , and it is also known that the associated subordinator is a @xmath266-subordinator ( see @xcite , their example 8.2 ) . as perceived in @xcite ,
blumenthal - getoor indices of @xmath265-processes exhaust the whole of the interval @xmath267 . in particular , for any given @xmath268 there are @xmath3-subordinators with blumenthal - getoor index @xmath269 , the latter by part ( a ) of proposition [ propvggcmoments ] ( see @xcite for multivariate @xmath265-models . ) [ remnomomentsonlyfv ] with being straightforwardly verified , @xmath270 is a thorin measure on @xmath241 , where @xmath271 as @xmath272 for @xmath224 and @xmath273 , respectively , we get from proposition [ propggcmoments ] , that any associated @xmath274-subordinator has infinite @xmath253-moments , with blumenthal - getoor index equalling 1 . by proposition
[ propvggcmoments ] , also any @xmath275-process with @xmath233 has infinite @xmath253-moments , and its blumenthal - getoor index equals 2 . based on the polar decomposition of the lvy measure ( see lemma [ lempolarrec ] ) , prez - abreu and stelzer @xcite construct classes of self - decomposable distributions on cones generating classes of lvy processes of finite variation ( @xmath47 ) , surpassing subordinators , and including versions of gamma and @xmath12 processes .
+ * multivariate gamma process .
* we reproduce the model in @xcite ( their section 3 ) .
recall @xmath276 , and let @xmath277 be a borel function , and @xmath7 a finite borel measure on @xmath278 such that @xmath279 we refer to a @xmath10-dimensional lvy process @xmath29 as a @xmath280-process with parameters @xmath7 and @xmath281 , written as @xmath282 , whenever @xmath283 with @xmath284 assuming @xmath285 satisfying it is shown in @xcite that the rhs in defines a lvy measure ( see their proposition 3.3 ) .
much in the spirit of our subsection [ subsecmoments ] , the @xmath286-class carries processes with infinite moments , amongst other things ( see @xcite , their examples 3.14 , 3.15 and 3.16 ) .
+ * multivariate gamma subordinator . * recall @xmath287 , and let @xmath288 be a borel function , and @xmath7 a finite borel measure on @xmath289 such that is satisfied , but with @xmath278 replaced by @xmath289 .
we refer to a @xmath10-dimensional subordinator @xmath53 as a @xmath280-subordinator with parameters @xmath7 and @xmath281 , written as @xmath290 , whenever , for all @xmath291 , @xmath292 as follows from the frullani identity in , the rhs in matches @xmath293 and , thus , @xmath290 holds if and only if @xmath294 with @xmath221 as in , but with @xmath278 and @xmath38 replaced by @xmath289 and @xmath221 , respectively .
plainly , a @xmath295-subordinator is a @xmath286-process with nondecreasing components : @xmath296 . in the univariate case
, we have @xmath297 .
the connection with our @xmath12-class is @xmath298 .
+ the associated @xmath299-class is thus a subclass of our @xmath6-class . complementing this
, it is possible to contrive the @xmath299-class as a class of lvy processes associated with the _ matrix - gamma - normal _ class of @xcite ( see their subsection 5.3.2 ) . +
* multivariate gamma convolution process .
* in @xcite ( see their definition 4.4 and their subsection 4.4 ) it is also shown how to characterise the associated class of multivariate generalised gamma convolutions associated with cones .
it is possible to introduce a @xmath300-class of @xmath47-processes , extending the polar decomposition from @xmath141 to @xmath301 , as we illustrated in context of the inclusion @xmath302 . in particular
, we have @xmath303 .
we conclude this subsection by establishing further inclusions in figure [ diagram ] , postponing further investigations of this kind to subsection [ subsecaltsubordclass ] . +
* variance gamma ( @xmath304 ) revisited . *
assume @xmath305 such that @xmath306 with @xmath307 in .
if @xmath177 and @xmath308 then degenerates to , for @xmath178 , @xmath309 , @xmath310 and in we have @xmath311 comparable to the representation of univariate @xmath264-processes as a difference of independent gamma subordinators in @xcite . for @xmath312 and invertible @xmath46
we get from and that , for @xmath313 , as @xmath314 , @xmath315 the asymptotic equivalence in the last display does not match , erasing the possibility of @xmath104 being a @xmath286-process . within his @xmath5-class grigelionis
( see @xcite , his proposition 3 ) shows that any @xmath316-process is self - decomposable , provided either @xmath177 or @xmath317 and @xmath238 . assuming an invertible @xmath46 and imposing further moment conditions upon @xmath140 , which are duly satisfied for finitely supported thorin measures such as @xmath307
, he shows that his result is sharp : if @xmath318 and @xmath317 then a @xmath316-process can not be self - decomposable .
consequently , a @xmath319-process with @xmath320 and invertible @xmath46 can not be self - decomposable , let alone be an element of the @xmath321-classes . plainly , @xmath322 is completely monotone . by differentiating under the integral sign
we find that @xmath323 .
in particular , if @xmath317 , then @xmath324 is no longer completely monotone , and a @xmath325-processes with @xmath317 , @xmath238 and invertible @xmath46 can not be a @xmath326-process
. it would be interesting to provide a detailed study regarding the @xmath327 ( @xmath328)-classes and the @xmath329-classes , but this is beyond the scope of our present paper .
we use the notation , for one hence all @xmath90 , @xmath330 @xmath331 is a convex subset of @xmath2 , containing the origin ( see @xcite , p. 165 ) , similarly we introduce @xmath332 .
further , we need to introduce @xmath333 and a transformation @xmath334 we provide conditions on the thorin measure ensuring finiteness of exponential moments for the associated @xmath335-model ( see subsection [ subsecproofesscher ] for a proof ) .
[ propexpoggcvggc ] let @xmath90 , @xmath328 , @xmath336 , @xmath337 and @xmath338 . then : + ( a ) @xmath339 is a convex and compact subset of @xmath2 , and @xmath340 is a continuous function from @xmath341 into @xmath141 .
+ ( b ) @xmath342@xmath343 simultaneously , @xmath344 and @xmath345 ( c ) for @xmath342 the image measure of the restriction @xmath346 under the mapping @xmath340 , denoted by @xmath347 , is a well - defined thorin measure on @xmath141 .
+ ( d ) without restrictions on @xmath348 : @xmath349 , where @xmath350 assume that @xmath351 is a lvy process with respect to an underlying stochastic basis @xmath352 .
the _ esscher transform _ on @xmath353 with respect to @xmath104 is given by @xmath354 for @xmath33 and @xmath355 in it is well - known that @xmath356 $ ] defines a probability measure , equivalent to @xmath357 $ ] .
besides this , @xmath358 remains a lvy process under the new measure @xmath359 .
next we show that both @xmath4-classes are _ invariant _ under esscher transformations ( for a proof see subsection [ subsecproofesscher ] ) we provide more specific examples throughout the remaining part of the paper , more specifically see theorem [ theovmg ] , subsection [ subsecaltsubordclass ] and section [ secapplic ] . [ theoesschervggc ]
let @xmath360 , @xmath228 .
if @xmath361 with @xmath355 then @xmath362 and @xmath363 with @xmath364 and @xmath365 as in proposition [ propexpoggcvggc ] . in this subsection we restrict ourselves to _ finitely supported _ thorin measures and consider a corresponding subclass of @xmath6 . +
* @xmath367-subordinator .
* let @xmath368 , @xmath369 , @xmath370 having columns @xmath371 and let @xmath372 be independent standard gamma processes .
we call a @xmath10-dimensional subordinator @xmath53 an _ @xmath15-subordinator _ with parameters @xmath373 , @xmath374 , @xmath375 , briefly @xmath376 , provided @xmath377 next , we show that @xmath15-subordinators are @xmath378-subordinators , but having zero drift @xmath379 and finitely supported thorin measure : [ lemsubgamma ] let @xmath380 with @xmath381 .
then @xmath382 , where , simultaneously , @xmath383 and , for @xmath33 , @xmath342 , @xmath384 _ proof .
_ by , follows straightforwardly from .
part ( a ) of proposition is applicable to giving @xmath385 , with a similar superposition and intersection in place for @xmath386 and @xmath332 , respectively .
it remains to verify the formulae in , but with @xmath387 , @xmath388 , @xmath389 and @xmath390 . with @xmath391 and @xmath336 with @xmath392
calculate @xmath393 on the other hand , @xmath394 in which we can substitute the characteristic exponents of @xmath88 using . by , the resulting expressions match those in .
+ for the remaining part we review some properties of the gamma distribution .
[ lemgamma ] let @xmath395 .
let @xmath396 .
let @xmath397 be independent with @xmath398 for @xmath130 .
then we have @xmath399 as well as the equivalence ` ( i)@xmath400(ii ) ' , where : ( i ) @xmath401 for some @xmath402 ; ( ii ) @xmath403
. if ( i ) or ( ii ) is satisfied then @xmath404 and @xmath405 .
_ note @xmath406 holds if and only if both @xmath407 and @xmath405 , and the equivalence (i)@xmath400(ii) follows from the thorin representation of the gamma distribution .
+ we introduced the @xmath408 in , and its connection with the @xmath12-class was @xmath298 with the identity @xmath297 in the univariate case @xmath409 . as we clarify next , a given @xmath410-subordinator does not need to have gamma marginals nor does it need to be @xmath295-subordinator .
[ lemsubgammaiii ]
let @xmath411 with @xmath381 .
+ ( a ) then ( i ) @xmath400 ( ii ) , where ( i ) @xmath412 for some @xmath413 ; + ( ii ) there exists @xmath414 with @xmath415 such that @xmath416 for all @xmath417 with @xmath418 . + ( b ) ( i ) @xmath400 ( ii ) , where ( i ) @xmath419 for some @xmath285 ; + ( ii ) for all @xmath420 , the following implication holds @xmath421 in addition , if one of ( i ) or ( ii ) holds then we have @xmath422 and @xmath423 . also ,
if one of ( i ) or ( ii ) is satisfied then @xmath424 and @xmath425 @xmath426 . _
_ ( a ) follows from lemma [ lemgamma ] as we can decompose the @xmath427th component of @xmath53 in into a sum of @xmath373 univariate gamma subordinators .
+ ( b ) ` ( ii')@xmath428(i) : for @xmath429 and @xmath430 with euclidean norm @xmath431 , introduce a univariate subordinator @xmath432 by @xmath433 to determine the lvy measure @xmath434 of @xmath435 , let @xmath436 be a borel set and note @xmath437\\ & = & \pi_s(\{\tau\in[0,\infty)^d_*:\,\tau \in\{\alpha x:\,\alpha>0\}~\mbox{and}~{\left\langle x,\tau\right\rangle}\in a \})\,.\end{aligned}\ ] ] as @xmath438 for @xmath439 , we get from that the rhs in the last display matches @xmath440 substituting @xmath441 on the rhs of the last display , it follows from that either @xmath442 or @xmath443 .
thus prepared , let @xmath380 with @xmath444 as in for independent standard gamma subordinators @xmath128 .
let @xmath445 be defined as in , but with @xmath446 replaced by @xmath53 .
in particular , @xmath447 in addition , suppose @xmath448
. then @xmath449 must either be degenerate or a univariate gamma subordinator .
consequently , by lemma [ lemgamma ] , we must have @xmath450 for @xmath451 with @xmath452 .
the latter is equivalent to ( ii ) , completing the proof of ` ( ii')@xmath428(i). the proof of ` ( i')@xmath428(ii) is analogous . + * variance@xmath453 @xmath454 ) . * with parameters @xmath369 and @xmath455 as set for an @xmath15-subordinator , in addition ,
take @xmath456 and a _ diagonal _
matrix @xmath183 with nonnegative entries . whenever @xmath457 , with @xmath458 being independent and @xmath101 being brownian motion , while @xmath376 , we call @xmath104 a _ variance @xmath459 @xmath161vm@xmath280@xmath162-process _ , written in the following as @xmath460 for a generic case , where @xmath169 , we give formulae for the canonical lvy measure @xmath171 . to each column
@xmath461 we associate both a dimension @xmath462 by @xmath463 and a @xmath464-finite borel measure @xmath465 on @xmath2 as a product measure with the following factors @xmath466 for @xmath451 , we set @xmath467 the next theorem gives formulae for the lvy measure and laplace exponent of @xmath104 , which has finite variation ( recall ) and is invariant in form under esscher tranformations .
[ theovmg]assume @xmath468 with @xmath381 .
then : + ( a ) we have @xmath469 with @xmath470 . + ( b ) always , @xmath471 .
when , in addition , @xmath472 , then for all borel sets @xmath473 , @xmath474 ( c ) we have@xmath475 and , for @xmath110 and @xmath355 , @xmath476 and @xmath477 here @xmath478 , and @xmath479 has the following columns @xmath480 : @xmath481 _ proof .
_ ( a ) follows from lemma [ lemsubgamma ] and .
@xmath15-subordinators @xmath53 have zero drift .
also , @xmath482 is finite for @xmath15-subordinators @xmath53 .
thus , @xmath483 by part ( c ) of lemma [ lemunifunimultisub ] .
( a stronger result follows from part ( a ) in proposition [ propvggcmoments ] . ) in view of part ( a ) , the remaining parts ( b)-(c ) follow from theorems [ theogvgmulti]-[theoesschervggc ] as well as proposition [ propexpoggcvggc ] .
[ remwang ] if @xmath468 with @xmath381 then @xmath484 as an implication of . here
@xmath485 are independent processes with @xmath486 for @xmath451 .
it is , thus , possible to construct @xmath11-processes by superimposing independent madan - seneta @xmath109-processes . wang @xcite comes to similar conclusions , and constructs multivariate lvy processes with @xmath264-components by superimposing suitable @xmath109-processes , just as in the right hand - side of
. * transition densities . * for a subclass of @xmath15-class it is possible to obtain formulae for transition densities , as we illustrate next .
let @xmath487 and @xmath488 such that , simultaneously , @xmath489 with @xmath90 define @xmath490 the proof of the next result follows from a similar analysis as in section 48.3.1 in @xcite .
( we are unable to provide substantial simplification of the integral in which occurs by integrating the joint density of @xmath491 independent gamma random variables . however , using the results in @xcite , it is possible to expand the integral in terms of lauricella functions . ) [ lemtransitiondensmultgam ] let @xmath90 and @xmath492 with @xmath375 satisfying .
then @xmath493 admits a lebesgue density @xmath494 : for @xmath495 , @xmath496 next , we state such formulae for the associated @xmath11-model .
let @xmath497 and @xmath183 be the parameters of the underlying brownian motion . to ensure the existence of a lebesgue transition density for the @xmath11-process , we need to make the additional assumption that @xmath46 is invertible , @xmath498 all @xmath499
. with the help of and define @xmath500 further , for @xmath90 , we set @xmath501 [ theotransitiondensemvg ] assume @xmath502 with @xmath169 and .
then the law of @xmath114 admits a lebesgue density @xmath503 for @xmath504 : @xmath505 @xmath506 for @xmath507 . _
_ since we have @xmath508 for independent @xmath101 and @xmath509 , follows from lemma [ lemtransitiondensmultgam ] by conditioning @xmath510 on values of @xmath493 ( see part ( b ) of lemma [ lemunifunimultisub ] ) . in view of remark [ remwang ] , we may write @xmath511 for independent processes @xmath512 and @xmath513 . here
the @xmath10-dimensional process @xmath512 has _ independent _ components with the @xmath427th component being a @xmath514-process ( @xmath515 ) .
further , @xmath513 is a @xmath516-process .
the formula in follows from by convolution . in this subsection
we review subordinator classes as they occur in the literature and relate them to our formulations .
the @xmath145 , @xmath280 , and @xmath15-subordinator classes were introduced in subsections [ subsecvggc ] and [ subsecvgamma ] , respectively , and one of the connections was @xmath298 .
as defined in , in lemma [ lemsubgamma ] the @xmath15-class was identified to be the subclass of @xmath517-subordinators with drift @xmath379 , having finitely supported thorin measures @xmath140 . as we clarified in part
( b ) of lemma [ lemsubgammaiii ] a given @xmath15-subordinator does not need to be a @xmath280-subordinator .
two other classes , such as the ones introduced by semeraro @xcite and guillaume @xcite , are related to them as shown in figure [ diagramsubord ] .
( compare figure [ diagramsubord ] with figure [ diagram ] . ) in the univariate case , where @xmath177 , note that @xmath518 .
@xmath519{000}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & { { { \cal g}}\gamma}^d_s & \begin{array}{ccc } { \mathrel{\rotatebox[origin = c]{20}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & { { \cal m}}\gamma^d_s & { \mathrel{\rotatebox[origin = c]{340}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } \\&\vdots&\\ { \mathrel{\rotatebox[origin = c]{340}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } & \gamma^d_s & { \mathrel{\rotatebox[origin = c]{20}{$\boldsymbol{-\!\!\!-\!\!\!\blacktriangleright } $ } } } \end{array } & { ggc}^{d}_s \end{array}\ ] ] * prez - abreu and stelzer s @xmath520-subordinator . *
the @xmath295-class of subordinators based on @xcite was defined in . in the univariate case ,
we have observed @xmath297 . the connection with our @xmath12-class was @xmath298 .
let @xmath290 , @xmath336 , @xmath521 as in .
we get from part ( a ) of proposition [ propexpoggcvggc ] that , simultaneously , @xmath522 and holds with @xmath281 replaced by @xmath523 . in part ( c ) of proposition [ propexpoggcvggc ] , observe that @xmath524 .
consequently , the associated @xmath525-class of subordinated brownian motions is _ closed under the esscher _ transform in the interpretation of theorem [ theoesschervggc ] . +
* semeraro s @xmath526-subordinator .
* semeraro @xcite introduced another approach to multivariate gamma subordinators ( also see @xcite ) .
the parameters of this model are as follows : let @xmath527 such that , simultaneously , @xmath528 for all @xmath515 .
let @xmath529 be independent such that @xmath530 we refer to @xmath53 as an _ @xmath7-subordinator _ , in brief @xmath531 , provided @xmath532 with @xmath533 any @xmath534-subordinator @xmath53 admits standard gamma marginal distributions : @xmath535 . as a result ,
the associated @xmath536-processes , called @xmath537 in @xcite have @xmath264-marginal distributions .
we give an alternative representation of @xmath53 in .
introduce parameters @xmath487 and independent standard gamma subordinators @xmath538 , @xmath539 for @xmath540 , by setting @xmath541 and , with @xmath542 as above , @xmath543 for @xmath53 in we conclude that @xmath544 , where in our notation @xmath545 we show that the @xmath536 process is not closed under esscher transform by considering the following bivariate example . in part ( c ) of theorem [ theovmg ] , we have @xmath546 for @xmath547 , @xmath548 and @xmath549 but also , recalling , @xmath550 under the escher transform , the first component of @xmath53 reads @xmath551 for independent @xmath552 by .
however , as implied by lemma [ lemgamma ] , @xmath553 , and @xmath554 can not be a standard gamma process .
the associated @xmath536-class of subordinated brownian motions is thus not closed under the esscher transformation in the interpretation of theorem [ theoesschervggc ] . +
* guillaume s subordinator .
* guillaume @xcite extends semeraro s @xmath555-class as follows : let @xmath556 , @xmath557 .
let @xmath529 be independent such that @xmath558 we refer to @xmath53 as a _
@xmath559-subordinator _ , in brief @xmath560 , provided @xmath532 with @xmath561 . with @xmath542 as above ,
introduce independent standard gamma subordinators @xmath562 by setting @xmath563 we conclude that @xmath564 , where in our notation , @xmath565 and @xmath566 further , observe that @xmath567 by part ( c ) of theorem [ theovmg ] , the @xmath568-class of subordinated brownian motions is , thus , closed under the esscher transformation in theorem [ theoesschervggc ] .
unlike the @xmath555-subordinator , a given @xmath569-subordinator does not need to have gamma marginals , and we clarified this in part ( a ) of lemma [ lemsubgammaiii ] . by part ( b ) of lemma [ lemsubgammaiii ] ,
contemplating , a @xmath569-subordinator and , thus , any @xmath555-subordinator is also a @xmath295-subordinator , concluding settlement of our diagram in figure [ diagramsubord ] , also recalling the chain of inclusions @xmath570 .
we are primarily concerned with demonstrating how our @xmath11-subclass can be applied , in particular , to price multi - asset options .
the @xmath11-subclass , as we showed , contains other popular models , such as the multivariate vg @xcite , the semeraro @xmath537
@xcite , and guillaume s extension @xcite . in subsection [ subsecmarketmodel ] a market model using the @xmath11-process is introduced , and we give explicit formulae for the expected value of the @xmath427-dimensional log - price process and its covariance matrix , and for the expected value of the price process itself .
this allows us to tabulate values of these quantities for a specific parameter set which we will use to illustrate the results .
the corresponding densities are calculated using the formula for the characteristic function given in of theorem [ theogvgmulti ] and displayed in figure [ fig1 ] .
the parameters required to make the esscher transform an equivalent martingale measure linking the real world and risk neutral dynamics are derived in proposition [ esscherprop2 ] of subsection [ subsecriskneutral ] . as an example , pricing of four kinds of two - asset options , specifically , european and american best - of and worst - of put options , can then be operationalised as we demonstrate in subsection [ subsecoption ] .
the exact form of the lvy measure as given in theorem [ theovmg ] ( b ) is an essential ingredient here .
we employ the @xmath11-process to model the log - prices of risky assets of a financial market .
potentially latent risk factors are described by a process @xmath468-process with respect to a given stochastic basis @xmath352 .
the risk factors drive a @xmath427-dimensional price process @xmath446 with @xmath571 , for @xmath572 and @xmath573 , with @xmath427-dimensional log - price process @xmath574 given by @xmath575 where @xmath576 is the expected total return rate of the assets , @xmath577 is the dividend yield of the assets , @xmath578 with rows @xmath579 satisfying @xmath580 , @xmath581 , determines the factor loading of the corresponding log - return process , and @xmath582 is an adjustment vector given by @xmath583 such that @xmath584 , @xmath572 , @xmath585
. recall @xmath586 denotes the identity function .
proposition [ vmgmomentsprop ] gives formulae for the moments of @xmath587 and the explicit form of the adjustment vector @xmath588 .
[ rema ] the dependence structure of the risk factor process @xmath104 is limited , as @xmath46 has to be a diagonal matrix in order that we remain in the class of lvy processes .
the matrix @xmath589 maps those risk factors to specific asset prices and generates a richer and perhaps more realistic dependence structure , for similar arguments and setup see @xcite .
accordingly , @xmath590 and @xmath591 are _ not _ necessarily @xmath592-processes , but are of course lvy processes . [ vmgmomentsprop ] let @xmath33 and @xmath591 as in with @xmath381 . then : + ( a ) @xmath593 .
+ ( b ) _ cov_@xmath594{a}^\prime\ ,
t , \quad t \ge 0 $ ] .
+ ( c ) @xmath595 .
_ this follows by differentiating the laplace transform ( see in theorem [ theovmg ] ) .
.__expected value , volatility and correlation of @xmath596 for @xmath597 , @xmath598 , @xmath468 with parameters @xmath599 , @xmath600 , @xmath601 , @xmath602 , @xmath603 , @xmath604 , @xmath605 , @xmath606 .
_ _ [ cols="^,^,^,^,^,^",options="header " , ] the recombining multinomial tree calculation we use has probability weights derived from the lvy measure , as provided in theorem [ theovmg ] .
the option parameters are set to @xmath607 and @xmath608 .
the tree models the bivariate process @xmath609 directly , with an exponential transform to obtain the price process . at each node of the tree
the process branches on a regular rectangular @xmath610 grid .
the minimum step sizes are @xmath611 and @xmath612 for @xmath512 and @xmath513 respectively .
prices are then obtained to an accuracy of three significant digits .
the time increment is @xmath613 .
run times are reduced by truncating propagation of the tree in its spatial dimensions after one time increment . allowing the tree to grow further
does not affect the results .
the results are presented in table [ prices ] . as expected ,
put options prices are increasing in the exercise price @xmath614 .
also , the worst - of put option prices exceed the corresponding best - of put option prices , which is consistent with no - arbitrage . for out - of - the - money options , the early exercise premium is higher for the worst - of put compared to the best - of put .
the early exercise premium for at - the - money options is approximately similar in both cases . for in - the - money options ,
the early exercise premium is higher for the best - of put compared to the worst - of put .
the dependence parameter @xmath615 affects the option prices as expected .
the payoff of the best - of put increases the contingency that both price processes fall jointly , thus the option premium is increasing in @xmath615 .
the payoff of the worst - of put increases if at least one price process falls , thus the option premium is decreasing in @xmath615 .
for @xmath616 being @xmath464-finite measures , @xmath617 denotes the corresponding unique @xmath464-finite product measure .
the trace field of the @xmath10-dimensional borel field @xmath618 in @xmath619 is denoted by @xmath620 , and @xmath621 is the unit sphere for a given norm @xmath41 on @xmath2 .
we say that a borel measure @xmath622 is _ locally finite relative to _
@xmath623 , provided @xmath624 for all compact subsets @xmath625 .
let @xmath626 $ ] be a _ locally finite _ borel transition kernel relative to @xmath241 : simultaneously , @xmath627 is borel measurable ; @xmath628 is a borel measure , locally finite relative to @xmath241 .
there exists a unique measure @xmath629 $ ] , locally finite relative to @xmath630 satisfying @xmath631 for @xmath632 , @xmath633 ( for example see exercise 3.24 , chapter iii of @xcite ) .
define @xmath634 $ ] as the image of @xmath635 under @xmath636 . by construction
, @xmath637 is a locally finite borel measure relative to @xmath39 satisfing @xmath638 for nonnegative borel functions @xmath639 .
next we provide a polar decomposition on @xmath39 as a disintegration of @xmath640 for borel measures satisfying additional integrability conditions .
the result is directly applicable to lvy and thorin measures , as in lemma [ lempolarrec ] we may choose @xmath641 and @xmath642 in view of and , respectively .
we omit the proof .
it is possible to adapt the arguments in @xcite , lemma 2.1 , and @xcite , proposition 4.2 , respectively .
[ lempolarrec]assume @xmath643 for a borel measure @xmath622 on @xmath39 and a continuous function @xmath644
. then we have : + ( a ) @xmath622 is locally finite relative to @xmath39 with @xmath645 $ ] .
+ ( b ) there exists a pair @xmath646 such that , simultaneously , + ( i ) @xmath7 is a finite borel measure on @xmath647 ; + ( ii ) @xmath648 is a borel kernel , locally finite relative to @xmath241 ; + ( iii ) @xmath649 for all @xmath650 ; + ( iv ) @xmath651 .
+ ( c ) if @xmath652 is another pair , simultaneously satisfying ( i)(iv ) , then there exists a borel function @xmath653 such that @xmath654 and @xmath655 . for a borel measure @xmath659 on @xmath39 and @xmath660
, we define a borel measure @xmath661 on @xmath39 where @xmath662 for a borel @xmath473 . here @xmath663 and @xmath664 , for @xmath317 , @xmath665 . recalling ,
introduce @xmath666 and @xmath667 . when @xmath668 , @xmath669 , @xmath21 and @xmath659 is a borel measure on @xmath39 , we set @xmath670 , @xmath671 and @xmath672 .
recall .
[ lemunifunimultisub]let @xmath328 .
let @xmath674 .
let @xmath675 be independent of @xmath29 .
then we have : + ( a ) @xmath676 with @xmath677 ( b ) for all @xmath33 @xmath678 ( c ) if , in addition , @xmath679 and @xmath680^k _ * } \|t\|^{1/2}\,{{\rm d}}\pi_t(t)<\infty$ ] then @xmath681 . in part (
a ) of lemma [ lemunifunimultisub ] the dependence of @xmath53 enters into the formulae in a linear fashion . as a result ,
if a process @xmath29 is independently subordinated by a superposition of independent subordinators then it can be written ( in distribution ) as the sum of independent processes : let @xmath683 be independent with @xmath684 for @xmath417 .
let @xmath685 and @xmath686 .
then we have : + ( a ) @xmath687 with @xmath688 and @xmath689 .
+ ( b ) @xmath690 with @xmath691 , @xmath692 and @xmath693 .
+ ( c ) if @xmath694 are independent copies of @xmath29 , also being independent of @xmath695 , then @xmath696 . + ( d ) if , in addition , both @xmath697_*^k } \|t\|^{1/2}\,{{\rm d}}\pi_{t_l}(t)<\infty$ ] and @xmath698 , then @xmath699 and @xmath700 for all @xmath665 .
_ ( a ) is well known , but can alternatively be deduced from the laplace transformation .
( b ) follows from part ( a ) , owing to part ( a ) of lemma [ lemunifunimultisub ] .
( c ) follows from part ( b ) .
( d ) follows from part ( a ) as an implication of part ( c ) of lemma [ lemunifunimultisub ] .
[ subsecproofvggc ] _ proof of theorem [ theogvg ] .
_ ( a ) let @xmath701 where @xmath702 are independent with @xmath703 and @xmath101 . observe that extends to @xmath704 with @xmath705 .
this follows from schwarz s principle of reflection : the proof of theorem 24.11 of @xcite can be adapted to our situation .
let @xmath34 and set @xmath706 such that @xmath707
. now follows from via conditioning on @xmath493 : @xmath708=e[e^{- t_t \lambda_\theta}]=\exp\big\{-t a\lambda_\theta - t\int_{(0,\infty ) } \log[(x+\lambda_\theta)/x]\;{{\cal t}}({{\rm d}}x)\big\}\,.\ ] ] here the right hand - side matches the formulae in .
+ ( b ) is shown in @xcite ( his proposition 3.3 ) , whereas holds as with @xmath709 in ( @xmath178 , @xmath710 ) and any borel set @xmath711 ( see @xcite , his equation ( 4 ) ) we have @xmath712 + _ proof of theorem [ theogvgmulti ] . _ ( a ) we omit the proof as it is similar to the proof of part ( a ) of theorem [ theogvg ] .
+ ( b ) we decompose @xmath53 into a superposition of independent subordinators @xmath713 where @xmath714 , and @xmath715 with @xmath716 as in . here
@xmath717 for @xmath90 .
also , @xmath718 denotes the identity . by proposition [ propdecomp ] , we have @xmath719 where @xmath720 is a family of independent lvy processes with @xmath721 and @xmath722 with @xmath723 for @xmath203 . for @xmath203
we have @xmath724 . to see , suppose @xmath213 and @xmath725 with @xmath212 .
note @xmath723 with , using its polar representation , @xmath726 where @xmath727 is the quantity in . in view of lemma [ lemunifunimultisub ] , @xmath728 where @xmath729 is a @xmath10-dimensional standard normal vector . as both @xmath213 and @xmath730 , @xmath731 must be absolutely continuous with respect to @xmath215 , admitting the following density , for @xmath216 , @xmath732 here we set @xmath733 for @xmath734 . expanding a square in the exponent yields @xmath735 so as to evaluate the interior @xmath736-integral using the identity for the modified bessel function @xmath614 of the second kind .
for @xmath172 we get @xmath737^{\#j/4}\;\times\nonumber\\ & & { } \qquad\qquad k_{\#j/2}\big(\big\{2\tau\!+\!\sum_{j\in j}s_j\mu_j^2/\sigma_{jj}\big\}^{1/2}\,\|y\|_{j , s}\big)\,\frac{{{\cal k}}(s,{{\rm d}}\tau)\,\alpha({{\rm d}}s ) } { \prod_{j\in j}\sigma_{jj}^{1/2}s_j^{1/2 } } \,,\nonumber\end{aligned}\ ] ] where the rhs of matches . in , observe @xmath738 where the rhs matches .
this completes the proof of part ( b ) .
[ subsecproofmoments ] _ proof of proposition [ propggcmoments ] .
_ ( a ) let @xmath739 .
pick @xmath740 such that , for all @xmath741 , @xmath742 by
, we get from fubini s theorem and a simple substitution that @xmath743 in view of and , @xmath744 is finite if and only if @xmath745 is , completing the proof of ( a ) . + ( b ) let @xmath746 . pick @xmath740 such that , for all @xmath741 , @xmath747 using similar arguments as in the proof of part ( a ) , we get from , fubini s theorem and a simple substitution that @xmath748 in view of and , we see that @xmath749 is finite if and only if @xmath750 is , completing the proof of ( b ) .
+ _ proof of proposition [ propvggcmoments ] .
_ let @xmath751 .
let @xmath101 where @xmath46 is a diagonal matrix .
let @xmath752 be a standard normal vector , that is a vector with independent standard normal components .
for @xmath753 introduce @xmath754 . by self - similarity of @xmath755
we can write @xmath756 for all ( but fixed ) @xmath753 . for @xmath757 and @xmath758 the maximum norms
are denoted by @xmath759 and @xmath760 , respectively .
let @xmath761 be the operator norm of @xmath41 .
the equivalence of norms in finite dimensions applies to @xmath2 as well as @xmath762 : we find a common constant @xmath763 such that @xmath764 as well as @xmath765 .
further , with @xmath766 denoting the euclidean norm on @xmath2 there exists @xmath767 such that @xmath768 .
+ ( a ) let @xmath230 , and introduce a function @xmath769 $ ] by @xmath770}(\|b(s)\|)]\,,\qquad s\in[0,\infty)^d_*\,.\ ] ] ` @xmath428 ' : we get from the self - similarity of @xmath755 that @xmath771=e\big[\big\|\mu \diamond s+(\sigma\diamond
s)^{1/2}z\big\|^q\big]\,,\quad s\in[0,\infty)^d\ ] ] and , thus , with @xmath772\big)$ ] , @xmath773\big)\nonumber\\ & \le&2^{q\vee 1}c^q_\infty\;\big(\|\cdot\diamond s\|^q_\infty\,\|\mu\|^q+\|\cdot\diamond s^{1/2}\|_{\infty}^{q}\ , e[\|\sigma^{1/2}z\|^q]\big)\nonumber\\ & = & 2^{q\vee 1}c^q_\infty\;\big(\|s\|^q_\infty\,\|\mu\|^q+\|s\|^{q/2}_\infty\ , e[\|\sigma^{1/2}z\|^q]\big)\nonumber\\ & \le&c_1\big({{\bf 1}}_{0<\|s\|\le 1}\|s\|^{q/2}+{{\bf 1}}_{\|s\|>1}\|s\|^q\big)\,,\quad s\in [ 0,\infty)^d_*\,.\end{aligned}\ ] ] as @xmath774 is globally bounded by @xmath775 , there thus exists @xmath776 such that @xmath777 for @xmath778
. follows from this and the finiteness of @xmath779 .
( recall @xmath780 by lemma [ lemunifunimultisub ] . )
the proof is completed by part ( a ) of proposition [ propggcmoments ] .
+ ` @xmath428 ' : assume @xmath116 . for @xmath781
observe @xmath782 and @xmath783 such that , since @xmath784 , for any @xmath785,@xmath786 for @xmath787 and @xmath778 we get from girsanov s theorem that @xmath788 and , by self - similarity of @xmath755 , @xmath789\nonumber\\ & = & \exp\big\{-\|\lambda\|^2_{\sigma\diamond s}/2\big\}\ ; e\big[\exp\{\!-{\left\langle \lambda , b^*(s)\right\rangle}\}\;\|b^*(s)\|^q\;{{\bf 1}}_{0<\|b^*(s)\|\le 1}\big]\nonumber\\ & \ge&\|s\|^{q/2}\exp\{-\|s\|_\infty\|\lambda\|^2_{\sigma}/2\}\;h_0(s/\|s\|)\nonumber\\ & \ge&\|s\|^{q/2}\exp\{-c_\infty\|\lambda\|^2_{\sigma}/2\}\;h_0(s/\|s\|)\,,\quad \label{eqsmallmomentsvggcii}\end{aligned}\ ] ] for @xmath778 with @xmath790 .
( note @xmath791 for @xmath792 . ) here @xmath793 is defined by @xmath794\,.\ ] ] as @xmath795 , @xmath796 is lower semicontinuous on @xmath797 , almost surely ( indicator functions of open sets are lower semicontinuous ) . as a result
, @xmath798 is itself lower semicontinuous by fatou s lemma with a compact domain @xmath797 . in particular
, @xmath798 attains its global minimum at some @xmath799 .
note @xmath800 by .
to summarise , we get from that @xmath801 for @xmath778 with @xmath802 .
the proof is completed by an application of part ( a ) of proposition [ propggcmoments ] , using similar arguments as in the proof of @xmath428. + ( b ) let @xmath803 .
recall @xmath804 for independent @xmath103 and @xmath53 .
+ ` @xmath428 ' : set @xmath805 . as in the proof of part ( a ) note @xmath806&\le&c_4\;\big\{\|s\|^p\,\|\mu\|^p+\|s\|^{p/2}\ , e[\|\sigma^{1/2}z\|^p]\big\}\label{momentvggci}\\ & \le&c_4\;\big\{e[\|\sigma^{1/2}z\|^p]+\|s\|^p(\|\mu\|^p+e[\|\sigma^{1/2}z\|^p])\big\}\label{momentvggcii}\end{aligned}\ ] ] for @xmath753 .
if @xmath238 then it follows from that @xmath807=e[e[\|b(t(t))\|^p|t(t)]]\le c_4 e[\|\sigma^{1/2}z\|^p]\ ; e[\|t(t)\|^{p/2}]$ ] , and the lhs is finite provided @xmath808 $ ] is .
otherwise , if @xmath318 , we get from that @xmath809&\le & c_4 \big\{e[\|\sigma^{1/2}z\|^p]+e[\|t(t)\|^p]\big(\|\mu\|^p\!+\!e[\|\sigma^{1/2}z\|^p]\big)\}\,,\end{aligned}\ ] ] and the lhs is finite provided @xmath810 $ ] is . in view of part
( b ) of proposition [ propggcmoments ] , this completes the proof of ` @xmath428 ' .
+ ` @xmath428 ' : suppose @xmath238 .
define @xmath811 by @xmath812 $ ] .
employing similar arguments as in the proof of part ( a ) we find @xmath813 such that @xmath814 and , thus , @xmath815=e[\|b^*(s)\|^p]=\|s\|^{p/2}e[\|b^*(s/\|s\|)\|^p]\ge g(s_0 ) \|s\|^{p/2}\,,\ ] ] for @xmath778 .
this extends to @xmath816\ge g(s_0 ) \|s\|^{p/2}$ ] for @xmath753 , including the origin .
in particular , this implies the inequality @xmath817\ge g(s_0 ) e[\|t(t)\|^{p/2}]$ ] by conditioning on @xmath493 . in view of part ( b ) of proposition [ propggcmoments ] , this completes the proof for @xmath238 .
assume @xmath818 . by the equivalence of norms , we have @xmath819 for some @xmath820 such that , for @xmath753 , @xmath821 and , with @xmath822 , using the self - similarity of @xmath755 , @xmath823\\ & \le & c_6e[\|b(s)\|^p]+c_6e[\|b^*(s)\|^p]\le c_6e[\|b(s)\|^p]+c_7\|s\|^{p/2}\end{aligned}\ ] ] where @xmath824 $ ] .
thus , we can find @xmath825 such that @xmath826\ge \|s\|^p - c_7\|s\|^{p/2}\ge \|s\|^p/2 $ ] for @xmath827 with @xmath828 .
to summarise , we have @xmath829+r_0^p$ ] for @xmath753 and , thus , @xmath830+r_0^p$ ] , by conditioning , completing the proof of ` @xmath428 ' .
+ ( we omit the proof for @xmath831 being similar but simpler . ) [ subsecproofesscher ] _ proof of proposition [ propexpoggcvggc ] . _
( a ) let @xmath832 .
it is straightforwardly checked that @xmath833 is closed under taking convex combinations .
for @xmath834 we have @xmath835 and , thus , @xmath836 by the cauchy - schwarz inequality ( @xmath766 denotes the euclidean norm and @xmath837 is any constant with @xmath838 . ) thus , @xmath833 is a bounded subset of @xmath2 . in particular
, @xmath833 is a compact , as it is also a closed subset of @xmath2 .
continuity of @xmath340 is obvious .
for @xmath839 we have @xmath840 and thus @xmath841 , as desired .
+ ( b ) let @xmath336 .
we get from fubini s theorem and that @xmath842 consequently , if @xmath843 then @xmath844 . for the remaining part ,
assume @xmath845 , and choose @xmath740 such that for all @xmath741 @xmath846 note that @xmath847 in the right hand - side is finite in view of .
the proof of part ( a ) is completed by combining , and .
+ ( c ) suppose @xmath848 .
( by part ( b ) we have @xmath845 . ) in view of part ( a ) , @xmath849 is borel measurable . in particular , the image measure , denoted by @xmath850 , of the restiction of @xmath140 to @xmath851 under @xmath340 is a well - defined borel measure on @xmath39 with @xmath340 as in , note that there is a constant @xmath852 such that , for @xmath853 with @xmath854 , @xmath855 and , thus , by the transformation theorem , @xmath856 @xmath857 ( to show the inequality , split @xmath851 into @xmath858 and recall @xmath859 in . ) in view of and , the rhs in is finite , completing the proof .
+ ( d ) let @xmath336 , @xmath90 .
if @xmath751 , then we have @xmath860 for independent @xmath53 and @xmath101 .
conditioning on @xmath493 yields @xmath861 otherwise , if @xmath831 then @xmath862 . in either way
, this completes the proof of part ( d ) .
+ _ proof of theorem [ theoesschervggc ] .
_ let @xmath751 , @xmath90 , @xmath355 .
let @xmath863 as in .
as @xmath355 , we must have @xmath864 by part ( d ) of proposition [ propexpoggcvggc ] .
let @xmath379 .
adapting arguments from the proof of theorem 25.17 of @xcite , @xmath865 , we get from that , for @xmath866 with @xmath867 , @xmath868 let @xmath869 as in , but with @xmath870 replaced by @xmath269 .
observe @xmath871 , the latter by part ( b ) of proposition [ propexpoggcvggc ] .
note that @xmath872 for @xmath873 .
set @xmath478 .
since extends as well , we get from that , still with @xmath379 , @xmath874 next , apply the transformation theorem to @xmath875 and @xmath876 to see that the rhs of the last display matches , but with @xmath877 replaced by @xmath878 , respectively , where @xmath365 is the well - defined thorin measure in part ( c ) of proposition [ propexpoggcvggc ] , but with @xmath870 replaced by @xmath269 . according to , if @xmath879 it is possible to decompose @xmath880 into independent @xmath881 where @xmath882 and @xmath883 . using the independence , the proof is completed for @xmath751 by noting that @xmath884 the proof of the remaining case , where @xmath831 , is similar , but simpler .
this completes the proof of the theorem .
we thank the editor , the associated editor and two referees for a close reading of the paper and many helpful suggestions .
further , we take pleasure in thanking christian rau for his comments on an earlier version of this manuscript . 9 ( 2012 ) identifying the successive blumenthal - getoor indices of a discretely observed process
statist . _
* 40 * , 14301464 .
( 2004 ) _ lvy processes and stochastic calculus .
_ cambridge studies in advanced mathematics * 93*. cambridge university press .
( 1961 ) sample functions of stochastic processes with stationary independent increments . _ j. math .
mech . _ * 10 * , 493516 . (
1992 ) _ generalized gamma convolutions and related classes of distributions and densities . _ lecture notes in statistics * 76 * , springer , new york .
( 2001 ) application of generalized hyperbolic lvy motions to finance .
_ lvy processes : theory & applications .
_ birkhuser boston , boston , 319336 .
( 2008 ) esscher transform and the duality principle for multidimensional semimartingales .
probab . _ * 19 * , 19441971 .
( 2005 ) minimal entropy preserves the lvy property : how and why .
_ stochastic process .
_ * 115 * , 299327 .
( 2008 ) stationary - increment variance gamma and @xmath885 models : simulation and parameter estimation .
_ international statistical review _ * 76 * , 167 186 .
( 1979 ) self - decomposability of the generalized inverse gaussian and hyperbolic distributions . _
z. wahrsch .
gebiete _ * 47 * , 1317 .
( 2009 ) numerical methods for lvy processes .
_ finance stoch . _ * 13 * , 471 - 500 .
( 2010 ) multivariate variance gamma and gaussian dependence : a study with copulas . in _ mathematical and statistical methods for actuarial sciences and finance _ ,
corazza , m. & pizzi , r. ( eds ) , springer , berlin .
( 2014 ) infinitely divisible multivariate and matrix gamma distributions .
_ j. multivariate anal .
_ * 130 * , 155175 .
( 1990 ) on series representations of infinitely divisible random vectors .
probab . _ * 18 * , 405 - 430 .
( 2011 ) pricing and hedging in exponential lvy models : review of recent results . in _
paris - princeton lectures on mathematical finance 2010 _ , lecture notes in math . *
2003 * , 319359 , springer , berlin .
( 1977 ) on the infinite divisibility of the pareto distribution .
actuar . j. _ * 1 * , 3140 . | we unify and extend a number of approaches related to constructing multivariate madan - seneta variance - gamma model for option pricing . complementing grigelionis ( 2007 ) class ,
an overarching model is derived by subordinating multivariate brownian motion to a subordinator from thorin s ( 1977 ) class of generalised gamma convolutions .
multivariate classes developed by prez - abreu and stelzer ( 2014 ) , semeraro ( 2008 ) and guillaume ( 2013 ) are submodels .
the classes are shown to be invariant under esscher transforms , and quite explicit expressions for canonical measures are obtained , which permit applications such as option pricing using pides or tree based methodologies .
we illustrate with best - of and worst - of european and american options on two assets . _
2000 msc subject classifications : _ primary : 60g51 , 60f15 , 60f17 , 60f05 + secondary : 60j65 , 60j75 _ keywords : _ lvy process , variance - gamma , multivariate subordination , generalised gamma convolutions , thorin measure , esscher transformation , esscher invariance , superposition , option pricing . |
let @xmath5 be a complex manifold , and @xmath6 a complex submanifold .
a classical problem in complex geometry is to understand the possible deformations of @xmath7 in @xmath5 , i.e. the collection of complex submanifolds of @xmath5 that are close to " @xmath7 , in an appropriate sense . in the most well - behaved situation
, there exists a _ universal family _ of such deformation ; this is described , for example , in @xcite ( where kodaira calls it a maximal family " ) .
such a family consists of an auxiliary complex manifold @xmath8 with basepoint @xmath9 , and a complex submanifold @xmath10 , such that , for each @xmath11 , the fiber @xmath12 is a complex submanifold of @xmath13 ; the fiber over @xmath9 corresponds to @xmath7 itself . for every other family @xmath14
, there exists a neighborhood @xmath15 of the basepoint , and a unique ( pointed ) holomorphic map @xmath16 such that , for each @xmath17 we have @xmath18 ( regarded as complex submanifolds of @xmath7 ) . we may regard @xmath8 ( or more precisely the universal family ) as a ( local ) _ moduli space _ for the deformations of @xmath7 . for a given @xmath5 and @xmath7 , such a moduli space may or may not exist . at a formal level , the possible deformations of @xmath7 in @xmath5 may be encoded as a functor @xmath19 where @xmath20 is the category of local artin algebras over @xmath21 ( with residue field @xmath21 ) @xcite@xcite .
heuristically , given an artin algebra @xmath22 , we may view an element of @xmath23 as a family of complex submanifolds of @xmath5 deforming @xmath7 , which is parameterized by @xmath24 ( with basepoint the unique geometric point of @xmath24 , i.e. the maximal ideal ) .
in the situation that there is a well - defined moduli space @xmath8 as above , we may identify @xmath23 with the set of ( pointed ) maps @xmath25 ; for example , the value of @xmath26 on the so - called dual numbers " @xmath27/(\epsilon^2)$ ] gives the ( geometric ) tangent space of @xmath8 at @xmath28 .
appealing to the functor of points philosophy , we may view @xmath26 itself as a formal stand - in for the moduli space that can always be defined .
even when @xmath8 itself does not exist , the tangent space " @xmath29/(\epsilon^2))$ ] still has an interesting geometric interpretation .
we summarize this as a theorem ( although the result is a straightforward consequence of the definition of @xmath30 ) .
[ z tangent ] there is a one - to - one correspondence between elements of + @xmath29/(\epsilon^2))$ ] and holomorphic sections of the normal bundle @xmath31 .
in other words , there is a natural bijection @xmath32/(\epsilon^2))\cong h^0(z;\mathcal{o}_{nz}).\ ] ] in fact , by a general argument ( independent of theorem [ z tangent ] ) , @xmath29/(\epsilon^2))$ ] inherits the structure of a @xmath21-vector space ; theorem [ z tangent ] is then better formulated as an isomorphism of vector spaces . another way in which @xmath26 encodes the geometry of the putative moduli space involves the _ obstruction _ properties of @xmath26 .
given an artin algebra @xmath22 , a deformation @xmath33 , and a surjective map @xmath34 in @xmath20 , consider the problem of finding an extension of @xmath35 to @xmath36 , i.e. an element @xmath37 satisfying @xmath38 .
in general , there will be an obstruction to the existence of such an extension .
if a smooth moduli space @xmath8 exists , however , then it is always possible to solve this extension problem and we say @xmath26 is _
unobstructed_. from a different perspective , we might allow @xmath8 to be a more general ( possible singular ) type of space than a complex manifold ; the obstruction problem for @xmath26 then relates to the smoothness of @xmath8 at @xmath28 .
this explains the importance of the following well - known result @xcite@xcite .
[ z obstructions ] let @xmath6 be a complex submanifold satisfying the condition @xmath39 then the deformation functor @xmath40 is unobstructed .
compared to theorem [ z tangent ] , which has a very clear geometric interpretation , theorem [ z obstructions ] is perhaps harder to understand intuitively ( and is correspondingly more difficult to prove ) .
the goal of the present paper is to extend the above constructions and results as far as possible into the realm of _ generalized _ complex geometry @xcite@xcite .
a generalized complex structure on a manifold @xmath5 is an endomorphism of the vector bundle @xmath41 ( the direct sum of the tangent bundle and cotangent bundle ) , which squares to minus the identity , preserves the natural pairing of @xmath41 with itself , and satisfies an integrability condition defined with respect to the _ dorfman bracket _
@xmath42,\pounds(\xi)b-\iota(\eta)da).\ ] ] as the name suggests , an ordinary complex structure @xmath43 may be viewed as a particular example of a gc structure ; a symplectic structure on @xmath5 also gives an example , however , so the subject could just as well be called generalized symplectic geometry " . indeed
, a fundamental appeal of gc geometry is that it provides a framework for treating complex and symplectic geometry in a unified way .
the subject also admits a wide variety of interesting examples : these range from examples that may be viewed as hybrids ( or deformations ) of complex and symplectic manifolds , to more exotic structures that exhibit features not present in either of the two classical " cases @xcite . to motivate the generalization of complex submanifolds whose deformation theory we study in this paper , as well as to explain some of the issues that arise in generalizing from the case of complex submanifolds , consider the following situation .
let @xmath44 be a symplectic manifold , and consider the problem of constructing a well - behaved moduli space of _ lagrangian _ submanifolds of @xmath5 .
an issue that immediately arises not present in the complex case is the need to incorporate the so - called _ hamiltonian symmetries _ of @xmath44 .
recall that , to any smooth function @xmath45 we may associate its _ hamiltonian vector field _ , which is characterized ( up to a sign convention ) by the equation @xcite @xmath46
every such vector field is an infinitesimal symmetry of the symplectic structure ; in particular the flow @xmath47 generated by @xmath48 ( when it exists ) satisfies @xmath49 for each @xmath50 .
the symplectomorphisms of @xmath44 that arise in this way are known as _
hamiltonian symmetries_,\to \r$ ] . ]
and they form a subgroup of the group of all symplectomorphisms of @xmath44 .
when forming a moduli space of lagrangian submanifolds , it is custumary to identity two lagrangians @xmath51 if @xmath52 for some hamiltonian symmetry @xmath53 . this is necessary , for example , to have any hope of constructing a finite dimensional moduli space @xcite . from the point of view of mirror symmetry ( and string theory ) , it is actually more natural to consider a slightly different moduli space ; namely , the collection of all pairs @xmath54 , where @xmath6 is a lagrangian submanifold , and @xmath55 is a hermitian line bundle with a flat connection supported on @xmath7 @xcite@xcite . adopting terminology from physics @xcite , such an object
may be called a ( rank 1 ) _
lagrangian brane_. a nice property of this moduli space ( when it can be defined ) , is that it carries a natural complex structure .
an added subtlety is the need to keep track of the possible equivalences between line bundles with connection . on any gc manifold @xmath56 , a ( rank 1 ) _
gc brane _ is similarly defined as a submanifold @xmath6 , together with a hermitian line bundle with unitary connection @xmath55 supported on @xmath7 ; the pair @xmath57 must satisfy a certain condition defined with respect to @xmath58 . in the case of a complex manifold @xmath5 , the compatibility condition implies that @xmath7 must be a complex submanifold , and the curvature form @xmath59 of @xmath55 must be of type @xmath60 with respect to the induced complex structure on @xmath7 ( i.e. the connection must induce a holomorphic structure on @xmath55 ) @xcite@xcite .
in particular , any complex submanifold equipped with the trivial line bundle gives an example of a gc brane . in the symplectic case ,
the lagrangian branes described above give one type of example , but there are actually others as well .
these so - called coisotropic " a - branes were first introduced by kapustin and orlov in @xcite , motivated by considerations coming from string theory and homological mirror symmetry ; the fact that these more exotic objects arise naturally in the context of gc geometry is another appealing feature of this framework @xcite . in this paper we will study the formal deformation theory of gc branes . in analogy to the symplectic case described above , it will be important to incorporate the action of the _ generalized hamiltonian _ symmetries of @xmath56 on gc branes . on a gc manifold @xmath56 ,
one may associate to any _ complex_-valued function @xmath61 its _ generalized hamiltonian vector field _ , which is a section @xmath62 .
as in the symplectic case , @xmath63 is an infinitesimal symmetry of @xmath56 , and may ( at least locally ) be integrated to a family of symmetries ( a flow " ) .
such a symmetry consists of a pair @xmath64 , where @xmath65 ^ -{\cong } & } } x$ ] is a diffeomorphism and @xmath66 is a 1-form . roughly speaking
, the diffeomorphism component acts on a brane by pulling it back , and the 1-form component acts by changing the connection on the line bundle .
let describe the main constructions and results of the paper .
for each gc brane @xmath0 , we construct a functor in this paper , but an extension to @xmath20 is also possible . ]
@xmath67 encoding the formal deformations of @xmath0 ( definition [ brane deformation functor ] ) .
this functor is defined as the truncation of a certain formal groupoid @xmath68 ( definition [ full deformation groupoid ] ) , which encodes all the relevant notions of equivalence for deformations ( including those induced by generalized hamiltonian symmetries ) .
the construction of this functor is more intricate than in the case of complex submanifolds discussed above , and is done in several steps . as a first step in studying the deformations of a gc brane @xmath0
, we consider the value of the functor @xmath69 on the dual numbers @xmath70 ; these are the first - order deformations of @xmath0
. it was argued in @xcite that such first - order deformations should correspond to elements of a certain lie algebroid cohomology group associated to @xmath0 ; we recover their result in a rigorous framework .
[ brane tangent ] for every gc brane @xmath0 , there is a natural bijection between elements of @xmath71/(\epsilon^2))$ ] and elements of the lie algebroid cohomology group @xmath72 .
the calculation needed to prove this result is somewhat involved and different from the one in @xcite
so the fact that we reach the same conclusion may be regarded as a check on our definition [ brane deformation functor ] .
to explain the connection to theorem [ z tangent ] stated above , suppose @xmath6 is a complex submanifold , regarded as a gc brane @xmath73 .
there is a natural isomorphism @xmath74 where @xmath75 denotes the sheaf of holomorphic one - forms on @xmath7 .
the first component @xmath76 corresponds to the possible deformations of the submanifold @xmath7 , as discussed above .
the second component @xmath77 corresponds to deformations of the trivial holomorphic line bundle on @xmath7 . from a different perspective , if we denote by @xmath78 the coherent sheaf of @xmath79-modules determined by @xmath7 ( the push - forward of its structure sheaf ) , then there are natural isomorphisms @xmath80 for each @xmath81 .
in particular , first - order deformations of @xmath73 as a gc brane correspond to first - order deformations of @xmath78 as a coherent sheaf ( since these are parameterized by @xmath82 ) . to go deeper into the geometric content of the deformation functor ( and formal groupoid ) associated to a gc brane
, we next take a closer look at the relationship between symmetries of the gc manifold @xmath56 and the deformations of @xmath0 .
it is here we encounter the first fundamental difference from the complex and lagrangian cases described above . in the case of a complex submanifold @xmath7 of a complex manifold @xmath83 , for instance ,
every deformation of @xmath7 is locally induced by a ( local ) symmetry of @xmath83 .
to explain this , consider the case of 1st order deformations , which by theorem [ z tangent ] correspond to holomorphic sections of the normal bundle @xmath31 .
given such a section @xmath84 , for each @xmath85 it is always possible to extend @xmath84 to a holomorphic vector field @xmath86 on some neighborhood @xmath87 of @xmath88 ; heuristically , the corresponding deformation is then given near @xmath88 by flowing " @xmath7 along @xmath86 for an infinitesimal time .
an analogous construction in the case of an arbitrary @xmath89 brane is _ not _ always possible , however .
we call deformations that have this property _ locally trivializable_. although an arbitrary gc brane may have deformations which are not locally trivializable , we introduce a large class of branes , which we call _ leaf - wise lagrangian _ ( lwl ) branes , and prove that their deformations are always locally trivializable ( theorem [ essentially surjective ] ) . this is a non - trivial statement about the geometry of such branes , and relies on the existence of a local normal form ( which we prove in theorem [ lwl normal form ] )
. the final part of the paper is devoted to proving a version of theorem [ z obstructions ] for gc branes .
the result we prove applies only to those branes with locally trivializable deformations ( in particular , by proposition [ essentially surjective ] , to holds for lwl branes ) .
we prove the theorem using the machinery of ( bi)semicosimplicial differential graded lie algebras ( dglas ) , as developed in @xcite@xcite@xcite .
specifically , by adapting a construction given in @xcite , we construct , for each brane @xmath0 , a dgla @xmath90 ( depending on a choice ) .
we prove that @xmath90 governs the locally trivializable deformations of @xmath0 .
a concrete calculation involving @xmath90 then allows us to prove the following result .
[ brane obstructions ] let @xmath0 be a leaf - wise lagrangian brane on a gc manifold ( or more generally a brane with locally trivializable deformations ) .
if the lie algebroid cohomology group @xmath91 vanishes , then the functor @xmath69 is unobstructed .
this result can clearly be sharpened in various situations : for example , in the case of a lagrangian brane @xmath0 , it is not hard to see that the dgla @xmath90 is ( homotopy ) abelian , so that in this case @xmath92 is always unobstructed .
we leave the problem of strengthening theorem [ brane obstructions]as well as extending it to arbitrary branes for future work .
in [ courant algebroid ] through [ lwl submanifolds ] , we present the definitions and results in gc geometry that will be needed for the rest of the paper .
although much of this material is well - known , there are a few definitions and results that have not to the best of our knowledge appeared before .
this includes , for example , the notion of a leaf - wise lagrangian submanifold ( or brane ) , which we introduce in definition [ definition lwl submanifold ] .
our treatment of the symmetries of the standard courant algebroid is also somewhat non - standard . in
[ section formal ] , we recall the basic definitions and results about artin algebras and nilpotent lie algebras we will need to formulate the definition of the functor @xmath92 associated to a gc brane ( definition [ brane deformation functor ] ) . in particular , we introduce a formal " ( infinitesimal ) version of the group of symmetries of the courant algebroid , and establish some of its basic properties . in [ section deformations ] , we construct the formal groupoid @xmath68 of deformations of a gc brane ( definition [ full deformation groupoid ] ) ; the deformation functor @xmath93 of @xmath0 is formed by taking @xmath94 of this groupoid ( definition [ brane deformation functor ] ) . after defining the deformation functor ( and formal groupoid )
, we proceed in [ first order deformations ] to prove theorem [ brane tangent ] about first - order deformations . in
[ section functoriality ] we study the behavior of the deformation groupoid under equivalences between different branes , and in particular prove the invariance of the deformation functor under such equivalences . in [ lwl branes example ] we consider the deformations of leaf - wise lagrangian branes ; the main result of this section is theorem [ essentially surjective ] , which implies that every deformation of a lwl brane is locally induced by a symmetry of the ambient gc manifold . in
[ induced deformations]-[cosimplicial groupoids ] we investigate the relationship between deformations and symmetries in more detail .
the remainder of the paper is devoted to applying the theory of dglas to the deformation theory of gc branes . in
[ dgla theory ] , we recall some material from @xcite@xcite@xcite concerning dglas , the construction of the deligne groupoid , and semicosimplicial objects .
we then use this setup to construct a dgla @xmath90 governing the locally trivializable deformations of a gc brane @xmath0 . in the final section ( [ cohomology ] )
, we use this construction to prove theorem [ brane obstructions ] . finally , there is a short appendix in which we prove a technical result stated in the main body of the paper .
* acknowledgements : * i would like to thank florian schtz , michael bailey , andrei caldararu , uli bunke , and dan freed for useful discussions related to the subject matter of this paper .
i would also like to acknowledge the useful questions and feedback received in response to several talks i gave at the university of regensburg .
special thanks are also due to simone diverio for pointing me towards the reference @xcite ; this was in response to a question i posed on the mathoverflow website , and i would also like to thank domenico fiorenza and urs schreiber for useful answers .
let @xmath5 be a smooth manifold , with tangent bundle @xmath95 and cotangent bundle @xmath96 .
we denote by @xmath97 the direct sum @xmath41 .
we have the natural pairing @xmath98 as well as the dorfman bracket @xmath99,\pounds(\xi)b-\iota(\eta)da),\notag\end{aligned}\ ] ] which is closely related to the courant bracket @xcite . the vector bundle @xmath100 equipped with these structures , together with the projection @xmath101 have the structure of an _ exact courant algebroid _ @xcite . in this paper
we work with this standard " courant algebroid , leaving for future work the extension of our results to a more general setting .
let @xmath102 denote the group of diffeomorphisms of @xmath5 , and @xmath103 denote the opposite group @xmath104 . given a diffeomorphism @xmath105 ^ -{\cong } & } } x$ ] , we denote the corresponding element of @xmath103 by @xmath106 , so that multiplication in @xmath103 is given by @xmath107 .
there is a left action of @xmath103 on @xmath108 by pullback , where by definition the pull - back of a vector field @xmath84 by a diffeomorphism @xmath53 is the push - forward of @xmath84 by @xmath109 : @xmath110 this action is compatible with the differential graded commutative algebra structure on @xmath111 , the lie bracket on @xmath112 , and the operation of contracting a vector field with a differential form . by the formula ( [ dorfman bracket ] ) for the dorman bracket , it follows the action is also compatible with the courant algebroid structure on @xmath100 in the following sense : given any sections @xmath113 , and any @xmath114 , we have @xmath115 @xmath116 and @xmath117 a key feature of the courant algebroid @xmath118 is that it admits a group of symmetries which is larger than @xmath102 @xcite .
namely , given a closed 2-form @xmath119 , the bundle map @xmath120 given by @xmath121 is compatible with the dorfman bracket and preserves the pairing and the projection map .
in particular , any 1-form @xmath66 determines a symmetry of the courant algebroid by setting @xmath122 .
this motivates the following definition .
[ definition cg ] let @xmath123 denote the semi - direct product @xmath124 , i.e. as a set @xmath125 , and the group multiplication is given by the formula @xmath126 [ leftgroupaction ] there is a left action of @xmath123 on @xmath127 given as follows : for each @xmath128 , and each @xmath129 , we define @xmath130 moreover , for each @xmath131 , the identities ( [ sym 1 ] ) , ( [ sym 2 ] ) , and ( [ sym 3 ] ) hold . [ 2 group ] since the action of a group element @xmath132 on @xmath127 depends only on the diffeomorphism @xmath53 and the 2-form @xmath133 , one might ask why we did not define @xmath123 to be the semi - direct product of @xmath134 with the space of closed 2-forms on @xmath5 ( as is perhaps more standard , see for example @xcite ) .
the reason is that this group would not act on the collection of branes on @xmath5 . as discussed in the introduction ,
such a brane consists of a pair @xmath57 , where @xmath6 is a submanifold , and @xmath55 is a hermitian line bundle with unitary connection supported on @xmath7 . as discussed below in remark
[ geometric brane 2 ] , a diffeomorphisms acts on such a brane @xmath0 by pull - back , whereas an element of the form @xmath135 acts by altering the connection on @xmath55 ( by adding @xmath136 ) . from a more fundamental point of view , we should view @xmath123 as a stand - in for a certain 2-group @xmath137 ( a group object in categories ) .
an object of @xmath137 is a pair @xmath138 , where @xmath53 is a diffeomorphism of @xmath5 , and @xmath139 is a hermitian line bundle with unitary connection ( or equivalently , a principal @xmath140-bundle ) over @xmath5 .
the bundle component @xmath139 acts on a brane @xmath54 by first restricting @xmath141 to @xmath7 and then tensoring with @xmath55 .
this 2-group @xmath142 is the symmetry group of the trivial gerbe ( with connective structure ) on @xmath5 ; such symmetries naturally act on the category of branes on x. given an element @xmath143 , for example , we may view it as an object @xmath144 of @xmath142 , where @xmath139 is the trivial hermitian line bundle on @xmath5 with connection @xmath145 .
viewed in this way , a _
morphism _ between two elements @xmath146 and @xmath147 of @xmath123 ( with the same diffeomorphism component ) consists of a function @xmath148 satisfying @xmath149 roughly speaking , the elements of @xmath123 correspond to the objects of the identity component " of @xmath137 ; for the purposes of this paper these are all the symmetries we need . because we treat only the standard courant algebroid @xmath41 in this paper ( corresponding to the trivial gerbe on @xmath5 ) , the extra categorical structure of the 2-group will not be needed explicitly . in order to extend to the case of more general courant algebroids ,
however , it will almost certainly be necessary to replace @xmath123 with the relevant 2-group .
we now return to the main exposition , and prove proposition [ leftgroupaction ] .
let @xmath150 .
we have @xmath151 since for each
1-form @xmath66 its exterior derivative @xmath133 is closed , it follows that equations ( [ sym 1 ] ) , ( [ sym 2 ] ) , and ( [ sym 3 ] ) hold .
[ symmetry groupoid ] more generally , we may define a groupoid @xmath152 whose objects are smooth manifolds , such that a morphism from a manifold @xmath5 to a manifold @xmath153 is a pair @xmath146 , where @xmath154 ^ -{\cong } & } } x$ ] is a diffeomorphism and @xmath155 .
composition of morphisms in this groupoid is given by the formula ( [ cgmult ] ) .
for a fixed manifold @xmath5 , the group @xmath123 is then recovered as the automorphism group of @xmath5 .
note also that a morphism from @xmath156 in @xmath152 determines an isomorphism @xmath157 using the same formula ( [ group action ] ) . for each @xmath128 , the action ( [ group action ] ) on @xmath127
is induced by a bundle map @xmath158 covering the diffeomorphism @xmath159 ^ -{\cong } & } x$ ] .
in particular , for every bundle endomorphism @xmath160 , the map @xmath161 given by @xmath162 is induced by a unique bundle endomorphism @xmath158 , which we denote by @xmath163
. clearly , the map @xmath164 defines an action of @xmath123 on @xmath165 .
we next introduce a lie algebra that is the infinitesimal version of @xmath123 , in the same way that the infinitesimal version of @xmath166 is the lie algebra of vector fields on @xmath5 .
[ definition cg ] let @xmath167 denote @xmath112 , viewed as a real lie algebra with respect to the jacobi - lie bracket .
we define @xmath168 to be the semi - direct product @xmath169 , where the action of @xmath167 on @xmath170 is given by the lie derivative .
explicitly , as a vector space we have @xmath171 and the bracket is given by the formula @xmath172=([\xi,\eta],\pounds(\xi)b-\pounds(\xi)a)\ ] ] for each @xmath173 and @xmath174 . returning briefly to the discussion in remark [ 2 group ] , it is actually more natural to view elements of @xmath168 as objects of a certain lie 2-algebra , corresponding to the _ infinitesimal _ symmetries of the trivial gerbe on @xmath5 .
such symmetries were introduced and studied in the author s ph.d .
thesis @xcite .
as mentioned in remark [ 2 group ] , the fact that we deal only with the standard courant algebroid @xmath41 allows us to avoid the use of the full lie 2-algebra structure explicitly . even in this case , however , the reader may find the categorical perspective conceptually useful . note that , although the underlying vector space of @xmath168 is isomorphic to @xmath127 , the lie bracket ( [ ghatbracket ] ) is _ not _ the same as the dorfman bracket ( [ dorfman bracket ] ) ; in particular , the latter is not even a lie bracket @xcite . on the other hand ,
the two brackets are compatible in a certain sense , as explained in the following proposition .
[ courant action prop ] 1 .
there is a left action of the lie algebra @xmath168 on @xmath127 given by the same formula as the dorfman bracket ( [ dorfman bracket ] ) , i.e. for each @xmath175 and each @xmath176 we define @xmath177,\pounds(\xi)c-\iota(\tau)da).\ ] ] 2 .
let @xmath160 be a bundle map . for each @xmath178 , the map @xmath161 given by @xmath179 for all @xmath180 is induced by a unique bundle endomorphism @xmath181 , which we denote by @xmath182 .
the map @xmath183 defines a lie algebra action of @xmath168 on @xmath165 . to prove the first part
, we must show that for every @xmath184 , and every @xmath180 , we have @xmath185\cdot \alpha.\ ] ] writing @xmath186 , and @xmath176 , we calculate @xmath187,\pounds(\xi)\pounds(\eta)c-\pounds(\xi)\iota(\tau)db-\iota([\eta,\tau])da.\end{aligned}\ ] ] therefore @xmath188-[\eta,[\xi,\tau]],\pounds(\xi)\pounds(\eta)c-\pounds(\eta)\pounds(\xi)c-\pounds(\xi)\iota(\tau)db+\pounds(\eta)\iota(\tau)da-\iota([\eta,\tau])da+\iota([\xi,\tau])db . \end{aligned}\ ] ] by the jacobi identity for vector fields , we have @xmath189-[\eta,[\xi,\tau]]=[[\xi,\eta],\tau].\ ] ] to calculate the one - form part , recall the following identities : given vector fields @xmath190 , we have @xmath191=\pounds([\xi,\eta])$ ] , @xmath192)=[\pounds(\xi),\iota(\eta)]$ ] , and @xmath193 .
the one - form component of ( [ expression ] ) is therefore equal to @xmath194c-\pounds(\xi)\iota(\tau)db+\pounds(\eta)\iota(\tau)da-\pounds(\eta)\iota(\tau)da+\iota(\tau)\pounds(\eta)da+\pounds(\xi)\iota(\tau)db-\iota(\tau)\pounds(\xi)db \nonumber\\ & = \pounds([\xi,\eta])c+\iota(\tau)(\pounds(\eta)da-\pounds(\xi)db ) \nonumber\\ & = \pounds([\xi,\eta])c+\iota(\tau)(d\iota(\eta)da - d\iota(\xi)db ) \nonumber\\ & = \pounds([\xi,\eta])c-\iota(\tau)d(\pounds(\xi)b-\pounds(\eta)a).\end{aligned}\ ] ] combining ( [ calc 1 ] ) and ( [ calc 2 ] ) we see that ( [ expression ] ) is equal to @xmath195,\tau ) , ( \pounds([\xi,\eta])c-\iota(\tau)d(\pounds(\xi)b-\pounds(\eta)a))=[(\xi , a),(\eta , b)]\cdot ( \tau , c).\ ] ] this includes the proof of the first part of the proposition . as for the second part ,
let @xmath196 , @xmath197 , and @xmath198 .
we have @xmath199,\pounds(\xi)(fb)-\iota(f\eta)da ) \\ & = ( f[\xi,\eta]+\xi(f)\eta , f\pounds(\xi)b+\xi(f ) b - f\iota(\eta)da ) \\ & = f\xx\cdot\alpha+\xi(f)\alpha.\end{aligned}\ ] ] since @xmath200 is linear over @xmath201 , we see that @xmath202 this shows that the map @xmath161 defined by the right - hand side of equation ( [ end action ] ) is linear over functions , and therefore is induced by a well - defined section of @xmath165 . it is also easy to see that equation ( [ lieaction1 ] ) implies that @xmath203\cdot t\ ] ] holds for every @xmath204 .
recall that a vector field @xmath205 is called _ complete _ if it generates a flow @xmath206 given such a vector field @xmath207 , we define @xmath208 the ( left ) action of @xmath167 on @xmath108 by lie derivative is the infinitesimal version of the ( left ) action of @xmath166 on @xmath108 in the sense that for every complete @xmath209 , and every @xmath210 , we have @xmath211 where the derivative with respect to @xmath212 is defined point - wise . similarly , we say that an element @xmath213 is _ complete _ if the vector field @xmath84 is . in this case , define @xmath214 and @xmath215 we then have the following result , similar to ( * ? ? ?
* prop 2.3 ) .
[ exp dorfman ] let @xmath216 be complete . then 1 . for every @xmath217 , we have @xmath218 2 . for every every @xmath219 and every @xmath50
, we have @xmath220 where the derivative with respect to @xmath212 is defined point - wise .
given @xmath150 , we have @xmath221 we have @xmath222 where the second equality follows from the change of variables @xmath223 .
therefore @xmath224 substituting into equation ( [ multiplication ] ) we obtian the desired result @xmath225 given @xmath226 , consider @xmath227 we have @xmath228\ ] ] and @xmath229 since @xmath230 vanishes at @xmath231 , the leibnitz rule together with the fundamental theorem of calculus implies that implies that @xmath232 substituting equations ( [ t derivative 1 ] ) , ( [ t derivative 2 ] ) , and ( [ t derivative 3 ] ) into ( [ t derivative ] ) , we see that @xmath233 finally , combining ( [ exponential rule ] ) with ( [ t derivative 4 ] ) , easily verify the second part of the proposition .
a good reference for the material in this subsection is @xcite .
as described in the introduction , a _ generalized complex structure _ on a manifold @xmath5 is an endomorphism @xmath234 that preserves the pairing @xmath235 , satisfies @xmath236 , and satisfies a certain integrability condition . to describe this condition , note that , since @xmath58 squares to minus the identity , we may decompose @xmath237 where @xmath238 and @xmath239 are the @xmath240 and @xmath241 eigen - bundles of ( the complex - linear extension of ) @xmath58 , respectively .
we require that @xmath242 be involutive with respect to the dorfman bracket , i.e. we require that @xmath243 given a gc structure @xmath58 on @xmath5 , the restriction of the dorfman bracket to @xmath242 endows @xmath238 with the structure of a complex lie algebroid over @xmath5 , with anchor map given by projection @xmath244 . in particular , we may associate to @xmath58 its _ generalized dolbeault _
complex @xmath245^{\delta_l } & \cinf(\lambda^1l\uv ) \ar[r]^{\delta_l } & \cinf(\lambda^2l\uv ) \ar[r]^-{\delta_l } & \cdots}.\ ] ] for future reference , let us explicitly describe the first two differentials in this complex
. given @xmath61 ( viewed as a section of @xmath246 ) , the section @xmath247 is given by @xmath248 for every @xmath249 , where we recall @xmath244 is the projection ( anchor ) map .
given @xmath250 , for every @xmath251 we have @xmath252).\ ] ] a useful observation is that , since both @xmath238 and @xmath239 are maximally isotropic sub - bundles of @xmath253 ( with respect to the @xmath21-linear extension of the pairing ) , the pairing determines an isomorphism @xmath254 ^ -{\cong } & } l^{\vee}.\ ] ] [ complex ] any ordinary complex structure @xmath255 on @xmath5 determines a gc structure given by @xmath256 in this case we have @xmath257 and @xmath258 the generalized dolbeault complex is isomorphic to @xmath259 with differential the @xmath260-operator corresponding to the standard holomorphic structure on @xmath261 @xcite .
[ symplectic ] let @xmath262 be a symplectic structure on @xmath5 , viewed as an isomorphism @xmath263 .
this determines a gc structure given by @xmath264 in this case , we have @xmath265 the bundle map @xmath266 is an isomorphism of complex lie algebroids , and in particular the generalized dolbeualt complex is isomorphic to @xmath267 , the ordinary de - rham complex of @xmath5 with complex coefficients .
given a gc manifolds @xmath56 and @xmath268 , the product @xmath269 inherits a natural gc structure @xmath270 .
given a gc manifold @xmath56 with @xmath240 eigen - bundle @xmath271 , the _ type _ of @xmath58 at a point @xmath272 is the complex dimension of @xmath273 ; equivalently , it is the complex codimension of the projection @xmath274 .
the gc manifold is said to be _ regular _ at a point @xmath272 if the type of @xmath58 is constant on some neighborhood of @xmath275 .
an equivalent definition of the type can be given as follows .
decompose @xmath58 as @xmath276 where @xmath277 , @xmath278 , @xmath279 , and @xmath280 .
the fact that @xmath58 preserves the natural pairing and squares to minus the identity implies that @xmath281 is skew - symmetric , i.e. for every @xmath282 we have @xmath283 in fact , the integrability of @xmath58 implies that @xmath284 is a poisson structure on @xmath5 . defining @xmath285
, we easily check that @xmath286 preserves @xmath287 and in fact restricts to a complex structure on @xmath287 .
it is then easy to check that the type of @xmath58 is equal @xmath288 . in particular
, @xmath58 is regular at a point @xmath272 if and only if the poisson structure @xmath284 is regular at @xmath275 , i.e. if and only if @xmath289 is constant on a neighborhood of @xmath272 . as an example , given a symplectic manifold @xmath44 and a complex manifold @xmath290 of complex dimension @xmath81 , the product @xmath291 is of type @xmath81 at every point ; in particular , it is everywhere regular .
conversely , the _ generalized darboux theorem _
, due to gualtieri @xcite , says if @xmath275 is a regular point of an arbitrary gc manifold @xmath56 , then there exists some neighborhood of @xmath275 on which @xmath58 is equivalent to such a product .
let us give a precise statement of this result in a form convenient for our purposes .
endow @xmath292 with the standard symplectic structure @xmath293 and @xmath294 with the standard complex structure .
let @xmath295 be the product gc manifold .
we may then state ( a slight variation ) of the generalized darboux theorem ( * ? ? ?
* th . 4.35 ) .
[ generalized darboux ] let @xmath56 be a gc manifold of dimension @xmath296 , and let @xmath272 be a regular point such that @xmath58 is of type @xmath297 at @xmath275 .
then there exists a neighborhood @xmath298 of @xmath275 , a neighborhood @xmath299 of the origin in @xmath300 , a diffeomorphism @xmath301 ^ -{\cong } & } } u_0 $ ] taking @xmath275 to the origin , and a 1-form @xmath302 , such that @xmath303 let @xmath58 be a gc structure on @xmath5 , and consider the projection maps @xmath304 given by @xmath305 and @xmath306 given by @xmath307 also define @xmath308 by @xmath309 we easily check that @xmath310 so that if we regard @xmath100 as a complex vector bundle with complex structure @xmath311 , @xmath312 is a map of complex vector bundles .
we also note that , since @xmath313 and @xmath238 is isotropic , we have @xmath314 since @xmath315 ^ -{\cong } & } \bar{l}$ ] is an isomorphism of complex vector bundles , and @xmath239 pairs non - degenerately with @xmath238 , it follows that @xmath312 is an isomorphism .
[ generalized holomorphic vector field ] a section @xmath316 is a _ generalized holomorphic vector field _
if it satisfies @xmath317 we denote the space of all generalized holomorphic vector fields by @xmath318 .
the space of generalized holomorphic vector fields of course depends on the gc structure @xmath58 .
when necessary , will use the more precise notation @xmath319 to indicate which gc structure is appearing in the condition ( [ ghol ] ) .
the next proposition shows that we may view the space of generalized holomorphic vector fields as the infinitesimal symmetries of the gc structure @xmath58 .
[ holomorphic symmetries ] for each @xmath216 , we have @xmath320 if and only if @xmath321 in particular , if @xmath322 is complete then @xmath323 it is easy to see from this proposition that the subspace @xmath318 is closed under the lie bracket ( [ ghatbracket ] )
. the following lemma will be useful for the proof of the proposition . for each @xmath216
, we have @xmath324 if and only if @xmath325 suppose @xmath324 .
then for every @xmath326 we have @xmath327 so that @xmath328 so we see that @xmath329 .
conversely , suppose that @xmath330 .
since @xmath322 is real it follows that @xmath331 also . for arbitrary @xmath332
, we see that @xmath333 and @xmath334 therefore @xmath335 by the lemma , to prove the proposition , we must show that @xmath336 if and only if @xmath330 . since @xmath238 is a maximal isotropic subspace of @xmath253 , we have @xmath330 if and only if for every pair of sections @xmath337 we have @xmath338 we claim that the left - hand side of equation ( [ preservation condition ] ) is equal to @xmath339 , so that it vanishes for arbitrary @xmath340 and @xmath341 if and only if @xmath342 . to see this , recall the following two identities satisfied by the dorfman bracket ( see e.g. @xcite ) 1 .
for every @xmath343 @xmath344 2 .
for every @xmath345 @xmath346 using these we have @xmath347\ra)+\pi(w)\la\xx , v\ra \\ & = \pi(w)\la\xx , v\ra-\pi(v)\la\xx , w\ra-\la\xx,[w , v]\ra \\ & = \frac{1}{2}\delta_l(\mu(\xx))(w , v).\end{aligned}\ ] ] given a complex structure @xmath255 on @xmath5 , let @xmath348 be the induced gc structure described in example [ complex ] .
it is straightforward to check that a vector field @xmath349 and a 1-form @xmath350 determine a generalized holomorphic vector field @xmath351 on @xmath352 if and only if @xmath353 and @xmath354 note that condition ( [ holomorphic 1 ] ) is equivalent to requiring @xmath355 whereas condition ( [ holomorphic 2 ] ) is equivalent to requiring that @xmath356 be of complex type @xmath60 . given a symplectic structure @xmath262 on @xmath5 , let @xmath357 be the induced gc structure described in example [ symplectic ] . in this case ,
a vector field @xmath349 and 1-form @xmath350 determine a generalized holomorphic vector field @xmath351 on @xmath358 if and only if @xmath359 and @xmath360 thus , an infinitesimal symmetry of @xmath357 consists of an infinitesimal symplectomorphism @xmath84 together with a closed 1-form @xmath361 . given a symplectic manifold @xmath44 and a real - valued function @xmath45 , recall that the _ hamiltonian vector field _
associated to @xmath362 is defined ( up to a sign convention ) by @xmath363 this vector field is an infinitesimal symmetry of the symplectic structure in the sense that @xmath364 and the collection of all hamiltonian vector fields form a sub - algebra of @xmath112 .
we next introduce a generalization of this construction .
namely , given a gc manifold @xmath56 , for every _ complex_-valued function @xmath61 we will define infinitesimal symmetry @xmath63 of @xmath58 .
[ genhamdef ] let @xmath56 be a gc manifold .
for every smooth function @xmath61 , the _ generalized hamiltonian vector field _
@xmath365 associated to @xmath362 is given by @xmath366 the author learned of this construction ( for real - valued @xmath362 ) in ( * ? ? ?
* prop 6 ) , where it is shown that @xmath63 is an infinitesimal symmetry of @xmath56 .
we give a different proof of this result ( for complex - valued @xmath362 ) as part ( 1 ) of the following proposition . to the best of our knowledge , part ( 2 ) of proposition [ hamiltonian ] has not appeared before .
[ hamiltonian ] 1 . for every @xmath61
, the generalized hamiltonian vector field @xmath63 is an element of @xmath367 , i.e. satisfies @xmath368 2 .
the collection @xmath369 of generalized hamiltonian vector fields is a lie sub - algebra of @xmath367 .
given @xmath61 , one easily calculates using definition [ genhamdef ] and and ( [ mudef ] ) that @xmath370 where on the right - hand side @xmath362 is regarded as a section of @xmath371 .
the first part of the proposition therefore follows from the identity @xmath372 . to prove the second part
, we must show that given any @xmath373 , there exists @xmath374 such that @xmath375=\xx_h$ ] .
first , suppose that @xmath362 and @xmath376 are purely imaginary . writing @xmath377 and @xmath378 , we have @xmath379=[(0,df_i),(0,dg_i)]=0,\ ] ] so the result is trivial in this case .
next , suppose that @xmath362 is purely real and @xmath378 is purely imaginary .
decomposing the gc structure as @xmath380 we have @xmath379=[(p df , kdf_r),(0,dg_i)]=(0,\pounds(p df)dg_i).\ ] ] using the cartan formula for the lie derivative , we see that this is equal to @xmath381 the last case , where both @xmath362 and @xmath376 are purely real , is more difficult , and uses the integrability of @xmath58 in a crucial way
. it will be useful to rewrite the integrability condition for @xmath58 .
to do so , first recall the definition of the _ courant bracket _ on sections of @xmath97 , which is given by a formula closely related to that for the dorfman bracket @xcite : @xmath382_c=([\xi,\eta],\pounds(\xi)b-\pounds(\eta)a-\frac{1}{2}d(\iota(\xi)b-\iota(\eta)a)).\ ] ] given an almost generalized complex structure @xmath58 , we define the _ nijenhuis tensor _
@xmath383 of @xmath58 by the formula @xmath384_c-\j[\j a , b]_c-\j[a,\j b]_c-[a , b]_c\ ] ] for every pair of sections @xmath385 .
integrability of @xmath58 is equivalent to the vanishing of @xmath386 @xcite . returning to the proof of proposition [ hamiltonian ]
, the integrability of @xmath58 implies that , for every pair of real - valued functions @xmath362 and @xmath376 we have @xmath387_c-\j[\j(0,df),(0,dg)]_c-\j[(0,df),\j(0,dg)]_c-[(0,df),(0,dg)]_c.\end{aligned}\ ] ] in terms of the decomposition ( [ j decom ] ) of @xmath58 , we have @xmath388_c$]= @xmath389,\pounds(p df)kdg-\pounds(p dg)kdf-\frac{1}{2}d\iota(p df)dg+\frac{1}{2}d\iota(p dg)df).\ ] ] it is easy to check that @xmath390 must be skew - symmetric , in the sense that for every @xmath391 we have @xmath392
. therefore the expression ( [ eqn11 ] ) is equal to @xmath393,\pounds(p df)kdg-\pounds(p dg)kdf - d\iota(p df)dg).\ ] ] similarly , we calculate @xmath394_c & = \j(0,\pounds(p df)dg-\frac{1}{2}d\iota(p df)dg ) \notag\\ & = ( p\pounds(p df)dg-\frac{1}{2}p d\iota(p df)dg , k\pounds(p df)dg-\frac{1}{2}k d\iota(p df)dg ) \notag \\ & = ( p d\iota(p df)dg-\frac{1}{2}p d\iota(p df)dg , k d\iota(p df)dg-\frac{1}{2}kd\iota(p df)dg ) \notag \\ & = ( \frac{1}{2}p d\iota(p df)dg-,\frac{1}{2}k d\iota(p df)dg ) , \end{aligned}\ ] ] and @xmath395_c & = ( -p \pounds(p dg)df+\frac{1}{2}p d\iota(p dg)df , -k \pounds(p dg)df+\frac{1}{2}kd\iota(p dg)df)\notag \\ & = ( -p d\iota(p dg)df+\frac{1}{2}p d\iota(p dg)df ,- k d\iota(p dg)df+\frac{1}{2}k d\iota(p
dg)df ) \notag\\ & = ( -\frac{1}{2}p d\iota(p dg)df,-\frac{1}{2}k d\iota(p dg)df ) .\end{aligned}\ ] ] also , we easily see that @xmath396_c=0 $ ] .
adding the 1-form components of ( [ nij1 ] ) , ( [ nij2 ] ) , and ( [ nij3 ] ) ( and again using the skew symmetry of @xmath284 ) , the 1-form component of equation ( [ nij0 ] ) yields : @xmath397 or equivalently @xmath398 on the other hand , adding the vector field components of ( [ nij1 ] ) , ( [ nij2 ] ) , and ( [ nij3 ] ) we see that @xmath399 with equations ( [ nij4 ] ) and ( [ nij5 ] ) in hand , we calculate @xmath400 & = [ ( p(0,df),k(0,df),(p(0,df),k(0,df ) ) ] \notag \\ & = ( [ p(0,df),p(0,dg)],\pounds(p ( 0,df))k(0,dg)-\pounds(p ( 0,dg))k(0,df ) ) \notag \\ & = ( p d\iota(p dg)df , kd\iota(p dg)df)+(0,d\iota(p df)kdg).\end{aligned}\ ] ] if we define @xmath401 , we see that @xmath379=\xx_h.\ ] ] in the case that @xmath58 is induced by an ordinary complex structure @xmath255 , the generalized hamiltonian vector field associated to @xmath61 is given by @xmath402 suppose @xmath58 is induced by a symplectic structure @xmath262 .
given @xmath61 , write @xmath403 for real - valued functions @xmath404 .
the generalized hamiltonian vector field associated to @xmath362 is given by @xmath405 where @xmath406 is the ordinary hamiltonian vector field for the function @xmath407 given by formula ( [ ordinary hamiltonian ] ) .
the following definition is a special case of ( * ? ? ?
7.4 ) . a _ generalized submanifold _ of a smooth manifold @xmath5 is a pair @xmath408 , where @xmath6 is a submanifold , and @xmath409 is a closed 2-form .
[ geometric brane ] as discussed in the introduction , we will primarily be interested in a related structure , which we call ( following @xcite ) a ( rank 1 ) _ brane _ on @xmath5 .
such an brane is a pair @xmath57 , where @xmath6 is a submanifold , and @xmath55 is a hermitian line bundle with unitary connection supported on @xmath7 . in particular , by setting @xmath59 to be the curvature form of @xmath55 , every such @xmath0 determines a generalized submanifold ( which we sometimes call the _ underlying _ generalized submanifold of @xmath0 ) .
when we study the deformation theory of such branes later in the paper , it will actually be conventient to use a slightly different but essentially equivalent definition ( definition [ brane definition 1 ] ) given a submanifold @xmath6 , let @xmath410 denote the inclusion , and let @xmath411 and @xmath412 denote the restrictions to @xmath7 of the tangent and cotangent bundles of @xmath5 .
we then have the push - forward @xmath413 and pull - back @xmath414 [ gen tangent def ] ( * ? ? ?
7.5 ) the _ generalized tangent bundle _ @xmath415 of a generalized submanifold @xmath408 is the sub - bundle of @xmath416 whose fiber at @xmath85 is given by @xmath417 the generalized tangent bundle @xmath415 is a maximal isotropic sub - bundle of @xmath416 with respect to the ( restriction of ) the pairing @xmath235 .
it fits into an exact sequence @xmath418 & \ann(tz ) \ar[r ] & \tt(z , f ) \ar[r ] & tz \ar[r ] & 0}.\ ] ] when @xmath419 , there is a natural splitting of this sequence @xmath420 given by @xmath421 which exhibits @xmath415 as the direct sum @xmath422 .
[ definition k ] let @xmath423 denote the restriction map .
we define @xmath424 by @xmath425 in other words , @xmath426 consists of those section of @xmath100 which extend section of @xmath415 .
[ k closed under bracket ] the subspace @xmath427 is closed under the dorfman bracket .
let @xmath428 , @xmath429 be elements of @xmath430 . by the assumption that @xmath322 and @xmath431 are elements of @xmath430 , the vector fields @xmath84 and @xmath432 are tangent to @xmath7 , so there exist unique vector fields @xmath433 such that the restriction of @xmath84 to @xmath7 is equal to @xmath434 and
the restriction of @xmath432 to @xmath7 is equal to @xmath435 .
furthermore , we have @xmath436 and @xmath437 .
since @xmath84 and @xmath432 are both tangent to @xmath7 , their lie bracket @xmath438 $ ] is as well ; in fact its restriction to @xmath7 is equal to @xmath439 $ ] .
we have @xmath440,\pounds(\xi)w-\iota(\eta)dv),\ ] ] so we need to show that @xmath441)f.\ ] ] expanding the right hand side using the identity @xmath442)=[\pounds(\tau),\iota(\zeta)]$ ] and the cartan formula we obtain @xmath443)f & = \pounds(\tau)\iota(\zeta)f-\iota(\zeta)\pounds(\tau)f \\ & = \pounds(\tau)\iota(\zeta)f-\iota(\zeta)d\iota(\tau)f-\iota(\zeta)\iota(\tau)df \\ & = \pounds(\tau)\iota(\zeta)f-\iota(\zeta)d\iota(\tau)f \\ & = \pounds(\tau)i^*v-\iota(\zeta)di^*w\\ & = i^*(\pounds(\xi)v-\iota(\eta)dw).\end{aligned}\ ] ] the following definition is taken from @xcite .
the _ generalized normal bundle _ of the generalized submanifold @xmath408 is the quotient @xmath444
let @xmath445 denote the quotient map .
since @xmath446 is maximal isotropic , it follows that the there is a well - define pairing of @xmath447 with @xmath415 given by @xmath448 which identifies @xmath447 with the dual of @xmath415 .
[ gc submanifold ] ( * ? ? ?
7.6 ) let @xmath56 be a gc manifold .
a generalized submanifold @xmath408 of @xmath5 is _ compatible with @xmath58 _ if @xmath449 in the case we say that @xmath408 is a _ generalized complex submanifold _ of @xmath56 .
[ geometric brane 1 ] given a brane @xmath57 on @xmath5 , as described in remark [ geometric brane ] , we say that @xmath0 is compatible with @xmath58 if its underlying generalized submanifold @xmath408 is . in this case
, we simply call @xmath0 a _ generalized complex ( gc ) brane_. as mentioned above , the definition we use later ( definition [ gc brane definition 1 ] ) is actually slightly different . [ rem1]it will be convenient to recast the compatibility condition ( [ compatibility ] ) in a slightly different form . given a gc manifold @xmath56 , define @xmath450 by @xmath451we easily verify that @xmath452 is skew - symmetric and non - degenerate . furthermore
, since @xmath415 is a maximal isotropic sub - bundle of @xmath453 , it follows that @xmath415 is preserved by @xmath58 if and only if the restriction of @xmath452 to @xmath415 vanishes .
if we define @xmath454 then a generalized submanifold @xmath408 is compatible with @xmath58 if and only if @xmath455 [ symplectic sub ] in the case where @xmath456 comes from a complex structure @xmath255 on @xmath5 , it is shown in ( * ? ? ?
7.7 ) that a generalized submanifold @xmath408 is compatible with @xmath58 if and only if @xmath7 is a complex submanifold of @xmath5 , and @xmath457 is of type @xmath60 with respect to the induced complex structure on @xmath7 .
[ complex sub]suppose @xmath458 comes from a symplectic structure on @xmath5 . if @xmath6 is a lagrangian submanifold , then @xmath408 is compatible with @xmath58 if and only if @xmath419 .
conversely , a generalized submanifold of the form @xmath459 is compatible with @xmath58 if and only if @xmath7 is lagrangian . on the other hand , there exist more exotic examples where @xmath457 is non - zero , and @xmath7 is coisotropic but of dimension greater than @xmath460 ( * ? ? ?
7.8 ) let @xmath408 be a generalized submanifold of a manifold @xmath5 .
let @xmath153 be another manifold , @xmath461 , a diffeomorphism , and @xmath462 a 1-form .
let @xmath463 be the morphism in the groupoid @xmath152 introduced in remark [ symmetry groupoid ] .
consider the submanifold @xmath464 , and let @xmath465 denote the diffeomorphism induced by @xmath53 .
we define @xmath466 to be the generalized submanifold of @xmath153 given by @xmath467 [ action on gc submanifolds ] 1 .
@xmath468 2 . given a gc structure @xmath58 on @xmath5 , a generalized submanifold @xmath408 on @xmath5
is compatible with @xmath58 if and only if @xmath466 is compatible with the gc structure @xmath469 on @xmath153 .
first , let consider the case where @xmath376 is of the form @xmath470 for some @xmath66 ; we then have @xmath471 set @xmath472 . given @xmath473 , let @xmath474 ; we then have @xmath475 and @xmath476 . since @xmath477 ,
@xmath84 is tangent to @xmath7 . denoting the restriction of @xmath84 to @xmath7 by @xmath478 ,
we than have @xmath479 this implies that @xmath480 and we therefore have @xmath481 .
set @xmath482 and @xmath483 , so that @xmath484 .
let @xmath485 and @xmath486 the inclusion maps , and note that by construction we have @xmath487 and @xmath488 given @xmath85 , and @xmath489 , let @xmath490 denote the linear isomorphism induced by the pair @xmath64 ; in other words , for each section @xmath491 , we have @xmath492 part ( 1 ) of the proposition is equivalent to the statement that , for each @xmath85 we have @xmath493 given @xmath494 , by definition we have @xmath495 for some @xmath496 , @xmath497 .
let us write @xmath498 .
defining @xmath499 , the equality ( [ diff com ] ) implies that @xmath500 similarly , using ( [ diff com1 ] ) we see that @xmath501 this completes the proof of part ( 1 ) .
let @xmath502 denote @xmath503 .
using part ( 1 ) as well as proposition [ leftgroupaction ] we have @xmath504 on the other hand , we clearly have @xmath505 , so it follows that @xmath506 if and only if @xmath507 . recalling the discussion in remark [ rem1 ] , this proves part ( 2 ) of the proposition .
[ geometric brane 2 ] continuing with remarks [ geometric brane ] and [ geometric brane 1 ] , we may define a similar action of @xmath123 on the set of branes on @xmath5 . given
a brane @xmath57 and a diffeomorphism @xmath65 ^ -{\cong } & } } x$ ] , define @xmath508 where as above @xmath509 ^ -{\cong } & } } z$ ] denotes the diffeomorphism induced by @xmath53 .
given a 1-form @xmath66 , the group element @xmath510 acts on @xmath0 changing the connection @xmath511 on @xmath55 according to @xmath512 using proposition [ action on gc submanifolds ] , it is then easy to check that for any @xmath513 , if the brane @xmath0 is compatible with a gc structure @xmath58 then @xmath514 is compatible with the gc structure @xmath503 .
in particular , the symmetries of a fixed @xmath58 act on the set of @xmath89 branes on @xmath56 .
let @xmath408 be a gc submanifold of @xmath56 .
since @xmath58 restricts to an endomorphism of @xmath515 which squares to minus the identity , we have the decomposition @xmath516 where @xmath517 is the @xmath240-eigenbundle of @xmath518 and @xmath519 the @xmath241-eigenbundle . by construction
, @xmath517 is a sub - bundle of @xmath520 .
furthermore , as described in @xcite , the lie algebroid bracket on @xmath238 ( i.e. the restriction of the dorman bracket to @xmath238 ) induces a well - defined bracket on sections of @xmath517 , giving @xmath517 the structure of a complex lie algebroid over @xmath7 with anchor map the projection @xmath521 .
explicitly , given sections @xmath522 , choose sections @xmath523 that extend @xmath322 and @xmath431 , i.e. such that @xmath524 and @xmath525 .
then it is straightforward to check that @xmath526 lies in @xmath527 and is independent of the choice of extensions @xmath528 and @xmath529 .
we therefore have a well - defined bracket given by the formula @xmath530 just as we did for the lie algebroid @xmath238 , we construct from @xmath517 a differential complex @xmath531^{\delta_l } & \cinf(\lambda^1l\uv ) \ar[r]^{\delta_l } & \cinf(\lambda^2l\uv ) \ar[r]^{\delta_l } & \cdots}.\ ] ] recall that we defined a linear isomorphism @xmath532 on any gc manifold .
similarly , for any gc submanifold @xmath408 , we have a linear bijection @xmath533 characterized by the equation @xmath534 for any pair of sections @xmath535 and @xmath536 , where @xmath537 is the quotient map introduced above .
a _ generalized holomorphic section _ of the generalized normal bundle is a section @xmath538 satisfying @xmath539 the following proposition is an easy consequence of the definitions , and we omit the proof .
[ induced holomorphic ] let @xmath491 be a generalized holomorphic vector field on @xmath5 . then @xmath540 is a generalized holomorphic section of @xmath447 .
[ lwl submanifolds ] let @xmath58 be an a generalized complex structure on a manifold @xmath5 . as discussed in @xmath541 , @xmath58 induces a poisson structure @xmath390 .
let @xmath542 be the ( not necessarily constant rank ) distribution induced by @xmath284 .
the distribution @xmath543 inherits symplectic structure , i.e. a non - degenerate , skew - symmetric pairing @xmath262 ( which is closed " in an appropriate sense ) defined by @xmath544 where @xmath545 is any section with @xmath546 .
for example , when @xmath58 is induced by a symplectic structure as in example [ symplectic ] , then @xmath547 and the induced pairing @xmath262 is given by the original symplectic form .
[ definition lwl submanifold ] let @xmath408 be a gc submanifold of @xmath56 , such that @xmath58 is regular at each point of @xmath7 .
we say that @xmath408 is _ leaf - wise _ lagrangian if for each @xmath85 the intersection @xmath548 is a lagrangian subspace of @xmath549 with respect to the pairing ( [ sym dis ] ) .
[ standard brane ] let @xmath550 be the
standard " gc manifold that appeared in theorem [ generalized darboux ] .
introduce coordinates @xmath551 on @xmath552 , where @xmath553 are standard coordinates on @xmath554 , and @xmath555 are standard coordinates on @xmath294 . for a natural number @xmath556 , and let @xmath557 be the product of the lagrangian submanifold @xmath558 with the complex submanifold @xmath559 ; explicitly , @xmath560 is defined by the conditions @xmath561 .
we easily check that @xmath560 , equipped with the zero 2-form , is a leaf - wise lagrangian gc submanifold of @xmath562 .
the following result , giving a local normal form for lwl gc submanifolds , is a variation on the generalized darboux theorem ( theorem [ generalized darboux ] ) .
[ lwl normal form ] let @xmath408 be a lwl gc sub manifold of @xmath56 , where @xmath5 is of dimension @xmath296 , @xmath7 is of dimension @xmath563 , and @xmath58 is of type @xmath297 at each point of @xmath7 .
then for each @xmath85 , there exists a neighborhood @xmath298 of @xmath88 in @xmath5 , a 1-form @xmath302 , a neighborhood @xmath299 of the origin in @xmath300 , and a diffeomorphism @xmath564 such that 1 .
@xmath565 and @xmath566 , 2 .
@xmath567 , and 3 .
@xmath568 , where @xmath569 is the pull - back ( restriction ) map . for simplicity write @xmath570 and @xmath571 . by the generalized darboux theorem [ generalized darboux ]
, we may assume without loss of generality that @xmath5 is an open subset of @xmath572 , and @xmath573 .
decompose that tangent space of @xmath574 at the origin as @xmath575 define @xmath576 , and let @xmath577 be the projection of @xmath578 onto @xmath579 . by assumption
, @xmath580 is a lagrangian subspace of @xmath581 .
@xmath582 is a complex subspace of @xmath583 .
introduce the notation @xmath584 , @xmath585 , so that we have a decomposition @xmath586 .
for every @xmath587 , there exists @xmath588 , @xmath589 and @xmath590 such that @xmath591 .
we have @xmath592 since by assumption this is again an element of @xmath593 , we must have @xmath594 therefore @xmath595 so we see that @xmath596 .
recall that the set of linear symplectomorphisms of @xmath554 acts transitively on the set of lagrangian subspaces , and that the set of complex linear automorphisms of @xmath597 acts transitively on the set of complex subspaces of fixed dimension .
therefore , by applying a linear change of coordinates on @xmath572 , we may assume without loss of generality that @xmath580 is the subspace spanned by @xmath598 , and @xmath582 is the subspace spanned by @xmath599 .
in particular , if we let @xmath600 denote the projection map , then we see that @xmath601 is an isomorphism at the origin .
we may therefore choose a neighborhood @xmath602 of the origin , such that the restriction of @xmath603 to @xmath604 give a diffeomorphism from @xmath604 to @xmath605 .
introduce the alternate coordinates @xmath606 given by @xmath607 and @xmath608 .
these coordinates are compatible with the decomposition @xmath609 , where @xmath610 are coordinates on @xmath611 and @xmath612 are coordinates on @xmath613 . in particular
, the restriction of @xmath614 give coordinates on @xmath604 , and we may write @xmath615 for unique smooth functions @xmath616 .
introduce the notation @xmath617 , and define @xmath618 by @xmath619 by construction , we have @xmath620 choose an @xmath621-ball @xmath622 around the origin such that @xmath623 define @xmath624 @xmath625 and let the diffeomorphism @xmath626 be the restriction of @xmath627 to @xmath298 .
let @xmath628 be the restriction of @xmath457 to @xmath604 .
by shrinking @xmath604 if necessary , we can find @xmath629 such that @xmath630 .
let @xmath631 be the projection , and define @xmath632 ; if we denote by @xmath633 , then clearly @xmath634 .
finally , let @xmath635 be the restriction of @xmath636 to @xmath298 .
consider the quadruple @xmath637 . by construction
, we have @xmath638 , @xmath639 , and @xmath640 .
therefore , the proof of theorem [ lwl normal form ] will be complete if we can show that @xmath641 defining @xmath642 and @xmath643 , we will show that @xmath644 which clearly implies the condition ( [ pullback of gc ] ) .
let @xmath645 be the sub - bundle spanned by @xmath646 and @xmath647 the sub - bundle spanned by @xmath648 .
the following facts are clear by inspection : 1 .
@xmath649 . furthermore
, both @xmath287 and @xmath543 are maximal isotropic so the pairing gives an identification @xmath650 and @xmath651 .
2 . @xmath652 and @xmath653 .
3 . @xmath654 .
let @xmath655 be the set of vector fields on @xmath572 of the form @xmath656 for real constants @xmath657 .
define also @xmath658 and @xmath659 to be the subspaces of elements @xmath660 satisfying @xmath661 for all @xmath662 .
[ basic facts ] 1 .
@xmath663 , 2 . @xmath664 .
3 . @xmath665 , @xmath666 .
4 . for every @xmath667 ,
we have @xmath668 , and @xmath669 . 5 .
every @xmath670 is complete , and satisfies @xmath671 every @xmath670 is of the form @xmath672 whereas every element @xmath673 is of the form @xmath674 we calculate @xmath675=-\eta^i\frac{\partial\xi^i}{\partial x^i}\frac{\partial}{\partial y^i},\ ] ] @xmath676 and @xmath677 using ( [ brack1 ] ) , ( [ brack2 ] ) , and ( [ brack3 ] ) , we see that for @xmath322 given by ( [ r form ] ) and @xmath431 by ( [ s form ] ) the dorfman bracket @xmath678 is equal to @xmath679 this verifies part ( 1 ) of the lemma .
part ( 2 ) is verified by a similar computation , which we omit .
part ( 3 ) follows from an easy calculation using the explicit form of @xmath680 given by ( [ jsymplectic ] ) and ( [ jcomplex ] ) .
for example , for @xmath681 we have @xmath682 part ( 4 ) follows from combining parts ( 1 ) , ( 2 ) , and ( 3 ) . to prove part ( 5 ) , recall that @xmath683 is complete if and only its vector field component @xmath684 is complete , i.e. if and only if it generates a well - defined flow @xmath685 for all @xmath50 . if @xmath667 , its vector field component is of the form @xmath686 , for which the flow is explicitly given by the formula @xmath687 given @xmath688 , for each @xmath689 define @xmath690 .
it follows from proposition [ exp dorfman ] that the map @xmath691 is characterized by the conditions @xmath692 , and @xmath693 we claim that , if @xmath431 is an element of either @xmath694 or @xmath695 , then @xmath696 and in particular @xmath697 to see this , note that @xmath692 , and that the left - hand side of ( [ deriv ] ) is equal to @xmath698 for all @xmath689 . if @xmath699 , then by part ( 4 ) of the lemma @xmath700 , so that both sides of ( [ deriv ] ) vanish .
if @xmath701 , then @xmath702 and therefore @xmath703 , so that the right - hand side of ( [ deriv ] ) is equal to @xmath704 and again we see that equation ( [ deriv ] ) holds . by inspection , @xmath705
is clearly determined by its action @xmath706 , so that part ( 5 ) of the lemma follows from equation ( [ int1 ] ) .
consider the vector field on @xmath707 given by @xmath708 .
it is easy to check that @xmath709 in @xmath710 .
furthermore , since @xmath711 and @xmath712 are both elements of @xmath694 , it follows from lemma [ basic facts ] that @xmath713 so that @xmath714 . therefore ,
defining @xmath715 , we see that @xmath716 therefore , to complete the proof of the proposition , it will be sufficient to show that @xmath717 actually , we will prove the equivalent statement @xmath718 . since @xmath719
, it follows from lemma [ basic facts ] that @xmath720 vanishes if and only if @xmath721 since @xmath722 and @xmath665 , this is equivalent to @xmath723 which is in turn equivalent to @xmath724 it follows from proposition [ action on gc submanifolds ] that @xmath725 since @xmath726 and @xmath408 is compatible with @xmath58 , it follows that @xmath727 on the other hand , since for each @xmath728 the quantity @xmath729 is invariant under the flow generated by the vector field @xmath84 , it follows that if @xmath729 vanishes on @xmath7 then it must vanish on all of @xmath730 , so we see that @xmath731
the remainder of the paper will be devoted to studying the deformation theory of generalized complex branes .
as discussed in the introduction , our first task will be to define , for every gc brane @xmath0 , a functor @xmath732 encoding the formal deformations of @xmath0 ; we do this in definition [ brane deformation functor ] . before we can formulate definition [ brane deformation functor ] , however , it will be necessary to introduce a framework that includes certain formal " versions of geometric structures studied in [ courant algebroid]-[lwl submanifolds ] .
we therefore begin by recalling some basic facts about local artin algebras , as well as nilpotent lie algebras and their exponentiation .
we then use this theory to recast some of the definitions and results about gc geometry presented in the previous sections .
recall that a ring is called _ artinian _ if it satisfies the descending chain condition on ideals .
let @xmath733 denote the category of local , unital artinian @xmath734-algebras with residue field @xmath734
. for a discussion of artin algebras and their use in deformation theory , see @xcite@xcite .
the basic facts stated below about @xmath733 may be found in these references .
every @xmath735 may be decomposed , as a real vector space , as @xmath736 where @xmath737 is the unique maximal ideal .
furthermore , @xmath738 is finite dimensional over @xmath734 , and coincides with the set of nilpotent elements of @xmath739 .
any @xmath734-algebra homomorphism @xmath740 for @xmath741 necessarily satisfies @xmath742 . for each natural number @xmath743 , the @xmath734-algebra @xmath744/(\epsilon^{n+1})\ ] ]
is an object of @xmath733 , with maximal ideal @xmath745 . given @xmath735 ,
a _ small extension _ of @xmath739 is a pair @xmath746 , where @xmath747 is an object of @xmath733 , and @xmath748 is a surjective map of @xmath734-algebras , such that the kernel of @xmath603 is a principal ideal @xmath749 satisfying @xmath750 .
continuing with the previous example , for each natural number @xmath743 we have a small extension @xmath751/(\epsilon^{n+2})\to \r[\epsilon]/(\epsilon^{n+1}).\ ] ] the following basic result will enable us to prove a number of results by inducting on small extensions @xcite .
[ induct on small extensions ] any surjective map in @xmath733 can be factored through a finite sequence of small extensions . in particular , for any @xmath735 , the quotient map @xmath752 can be factored through a finite sequence of small extensions .
let @xmath753 be a real lie algebra .
then @xmath753 determines a functor from @xmath733 to the category of real lie algebras , given by @xmath754 we will refer to such a functor as a _ lie algebra over @xmath733_. as a slight abuse of notation , we will sometimes use the same symbol @xmath753 " to refer both to the lie algebra itself and the lie algebra over @xmath733 it determines ; the meaning will always be clear from context . [ over art ] more generally , for any category @xmath755 , we may define an object of @xmath755 over @xmath733 to be functor @xmath756 , and a morphism ( over @xmath733 ) between such objects to be a natural transformation .
for example , a real vector space @xmath757 determines two different vector spaces over @xmath733 , given respectively by @xmath758 or @xmath759 a representation of a group @xmath760 over @xmath733 on a vector space @xmath757 over @xmath733 consists of a representation of @xmath761 on @xmath762 for each @xmath735 , the collection of which must be suitably compatible with morphisms @xmath763 in @xmath733 .
the groups over @xmath733 we consider in this paper all have the property that their value on the trivial artin algebra @xmath764 is the trivial group .
we will refer to such a group as a _ formal group over @xmath733 _ , or more succinctly as a _
formal group_. to each real lie algebra @xmath753 , recall that we can associate a formal group over @xmath733 as follows . for each @xmath735 ,
the lie algebra @xmath765 is nilpotent , and may therefore be exponentiated " it to form a group @xmath766 . by definition , there is a canonical bijective map from the underlying set of @xmath767 to the underlying set of @xmath766 , which we write as @xmath768 . for each pair of elements @xmath769 , the product of @xmath770 with @xmath771
is defined using the baker - campbell - hausdorff formula @xmath772+\cdots},\ ] ] where the formally infinite sum @xmath773\cdots$ ] terminates after a finite number of terms due to the nilpotence .
this construction @xmath774 is functorial : every homomorphism of lie algebras @xmath775 induces a homomorphism of formal groups @xmath776 , given for each artin algebra @xmath735 and each @xmath777 by the formula @xmath778 .
we record here for later use some well - known facts about this construction .
[ exp prop ] let @xmath753 be a real lie algebra with corresponding formal group @xmath779 . 1
. every action of @xmath753 on a real vector space @xmath757 induces an action of @xmath779 on @xmath757 ( viewed as a vector space @xmath780 over @xmath733 ) .
this action is given by the exponential formula : for each artin algebra @xmath735 , each @xmath777 , and each @xmath781 , we set@xmath782 2
. consider the action of @xmath779 on @xmath753 obtained by exponentiating the adjoint representation of @xmath753 on itself , i.e. for every artin algebra @xmath739 and every @xmath783 have @xmath784+\frac{1}{2}[x,[x , y]]+\cdots.\ ] ] then for every @xmath769 , we have the identity @xmath785 3 .
fix @xmath735 .
every group homomorphism @xmath786 is of the form @xmath787 for a unique @xmath777 .
conversely , for each @xmath777 , the formula ( [ 1parsubformal ] ) determines a homomorphism @xmath788 .
fix a smooth manifold @xmath5 .
as in [ courant algebroid ] , we denote by @xmath167 the real lie algebra of vector fields on @xmath5 , and by @xmath789 the semi - direct product lie algebra introduced in definition [ definition cg ] . following the convention introduced above , we may also regard @xmath167 and @xmath168 as a lie algebras over @xmath733 , which assign to each @xmath735 with maximal ideal @xmath737 the nilpotent lie algebras @xmath790 and @xmath791 , respectively .
more generally , we have sheaves of lie algebras over @xmath733 , which assign to each open set @xmath87 and each @xmath735 the lie algebras @xmath792 and @xmath793 .
we also introduce the notation for differential forms @xmath794 and @xmath795 and similarly @xmath796 @xmath797 @xmath798 and @xmath799 a potentially confusing point about this notation : while @xmath112 and @xmath167 both denote the space of vector fields on @xmath5 , given @xmath735 the spaces @xmath800 and @xmath801 are not the same .
this corresponds to the different roles played by the space of vector fields in our development .
consider the formal group @xmath802 , whose elements may be viewed as infinitesimal diffeomorphisms of @xmath5
. the action of @xmath167 on the spaces @xmath803 and @xmath112 via the lie derivative may be exponentiated ( as described in proposition [ exp prop ] ) to an action of @xmath802 .
unsurprisingly , this action is compatible with the relevant structures on @xmath803 , and @xmath112 , such as the wedge product of differential forms , and the contraction operation between vector fields and forms .
this is summarized in the following proposition , whose proof is deferred to the appendix .
[ formal courant action ] fix an artin algebra @xmath735 . for each @xmath804 , each @xmath805 , and each @xmath806 we have @xmath807 @xmath808 @xmath809 and @xmath810)=[g(\xi),g(\eta)].\ ] ] turning next to the formal group @xmath811 , we may similarly exponentiate the action of @xmath168 on @xmath127 ( described in proposition [ courant action prop ] ) to give an action of @xmath811 . using proposition [ formal courant action ] )
, we may then easily prove the following .
[ action ] for each @xmath735 , the action of @xmath812 on @xmath813 is compatible with the ( a - linear extensions of the ) dorfman bracket ( [ dorfman bracket ] ) , the natural pairing @xmath235 , and the projection @xmath814
. it will be convenient to have a formal version of proposition ( [ exp dorfman ] ) , which gives an explicit formula for the exponential map @xmath815 in terms of the subgroups @xmath816 and @xmath817 ; introducing the notation @xmath818 for @xmath170 ( regarded as an abelian lie algebra ) , the latter group may be denoted by @xmath819 .
adapting the same notation used in the statement of proposition ( [ exp dorfman ] ) , for each @xmath820 and each @xmath821 , we define @xmath822 where the precise meaning of the right - hand side is as follows : for each @xmath823 and @xmath824 , the expression @xmath825 defines a polynomial in @xmath826 , i.e. an element of @xmath827 $ ] .
the integral with respect to @xmath826 in ( [ formal a def ] ) is then defined algebraically : for each natural number @xmath297 and each @xmath828 we set @xmath829 [ exponential ] for all @xmath830 , we have @xmath831 given @xmath832 , consider the map @xmath833 given by @xmath834 by proposition [ exp prop ] , for each @xmath820 and each @xmath824 we have @xmath835 the same calculation used in the proof of proposition [ exp dorfman ] may then be used to show that @xmath836 is a group homomorphism .
therefore , by part ( 2 ) of proposition [ exp prop ] , it follows that @xmath53 is of the form @xmath837 for a unique @xmath838 .
note that @xmath839 where @xmath840 is an element of @xmath841 $ ] ( depending on @xmath84 and @xmath842 ) , so that by the baker - campbell - hausdorff formula we have @xmath843 where @xmath844 is an element of @xmath841 $ ] . comparing equations ( [ poly1 ] ) and ( [ poly2 ] )
, we see that @xmath845 and @xmath150 . by construction , we have inclusions of @xmath816 and @xmath819 as sub - groups of @xmath846 .
proposition [ exponential ] implies that these subgroups ( which have trivial intersection ) generate @xmath847 , so that @xmath846 may be identified with the semi - direct product @xmath848 ( this is of course a special case of a more general result about the exponentiation functor ) . in particular
, we have the following corollary , recorded here for later use .
[ exp formula cor ] for each @xmath820 , the linear map @xmath849 mapping @xmath850 is a bijection . next , let @xmath58 be a gc structure on @xmath5
. recall from [ gc structures ] the lie algebra @xmath367 of generalized holomorphic vector fields ( definition [ generalized holomorphic vector field ] ) , which are the infinitesimal symmetries of @xmath58 , as well as the sub - algebra @xmath851 of generalized hamiltonian vector fields ( definition [ genhamdef ] ) .
for each @xmath735 we may then define the nilpotent lie algebra @xmath852 as well as the subalgebra @xmath853 . by inspection of the exponential formula ( [ exp action ] ) , we immediately see the following [ formal holomorphic symmetries ] for each @xmath854 we have @xmath855
using the framework introduced in the last section , we now to turn to the problem of defining for each @xmath89 brane @xmath0 a functor @xmath856 encoding the formal deformations of @xmath0 . as mentioned in the introduction , this functor ( given in definition [ brane deformation functor ] ) ,
will be constructed in terms of a formal groupoid @xmath68 over @xmath733 ( definition [ full deformation groupoid ] ) , which encodes the appropriate notions of equivalence between deformations of @xmath0 . given a brane @xmath57 , we may simultaneously deform both @xmath7 and @xmath55 . as a preliminary to studying the general situation
, we first consider the deformations of @xmath55 only , i.e. the deformations of a hermitian line bundle with unitary connection over a fixed space .
let @xmath5 be a smooth manifold , and @xmath857 an open cover of @xmath5 .
we denote by @xmath858 the @xmath81-fold intersection @xmath859 .
the set @xmath860 consists of triples @xmath861 , where @xmath862 is an open cover of @xmath5 , @xmath863 is a collection of real - valued functions satisfying @xmath864 and @xmath865 is a collection of ( real ) 1-forms satisfying @xmath866 on the intersections @xmath867 .
[ gluing bundles ] given an element @xmath868 as above , we may construct a hermitian line bundle with unitary connection on @xmath5 by gluing .
the line bundle is constructed by gluing trivial ( hermitian ) line bundles on each @xmath869 using transition functions @xmath870 .
the unitary connection is then locally specified on each @xmath871 by @xmath872 furthermore , for each manifold @xmath5 , there exists an open cover @xmath873 such that every hermitian line bundle with unitary connection is isomorphic to one of this form .
more formally , the gluing construction may be extended to an equivalence of categories ( defined on objects as above ) .
for the most part , we will work with a fixed open cover @xmath874 and omit it from the notation , i.e. we will simply write @xmath875 to specify an element of @xmath860 .
an element @xmath876 determines a closed 2-form @xmath877 , which is given by @xmath878 on each open set @xmath871 . given an open set @xmath87
, there is a natural restriction map @xmath879 , which we denote by @xmath880 for @xmath881 .
[ groupoid over art ] recall from remark [ over art ] , that for any category @xmath755 we defined an _ object of @xmath755 over @xmath733 _ to be a functor @xmath756 .
a similar definition can be made when @xmath755 is a 2-category , for example the 2-category @xmath882 of ( small ) groupoids .
for example a ( strict ) _ groupoid over @xmath733 _ may be defined as a ( strict ) functor @xmath883 ; explicitly , this consists of a groupoid @xmath884 for every @xmath735 , and for every homomorphism @xmath885 a functor @xmath886 . for our present purposes
, we require the composition of these functors to be strictly compatible with composition in @xmath733 .
the groupoids over @xmath733 we consider will all have the property that they map the trivial artin algebra @xmath764 to a trivial groupoid , i.e. a groupoid with a single object and a single morphism .
we will refer these as _ formal groupoids _ ( over @xmath733 ) .
the notion of a ( strict ) functor between formal groupoids is defined in the obvious way .
we will sometimes drop the modifier formal " when it is clear from context . to streamline the exposition
, we will usually define specific formal groupoids ( and functors between them ) by simply describing their value on each @xmath735 .
the additional structure needed to make things completely precise will always be clear from the context .
[ herm def ] given @xmath887 , we define a formal groupoid @xmath888 ( over @xmath733 ) as follows . for each @xmath735 , an object of @xmath889 , which we call an _ a - deformation _ of @xmath55 ,
is a pair @xmath890 , where a. each @xmath891 , and is of the form @xmath892 for @xmath893 , b. each @xmath894 is of the form @xmath895 for @xmath896 , c. on each overlap @xmath897 we have @xmath898 or equivalently @xmath899 , d. on each 3-fold intersection @xmath900 we have @xmath901 or equivalently @xmath902
. given @xmath739-deformations @xmath890 and @xmath903 of @xmath55 , an _ isomorphism _
@xmath904 is a collection @xmath905 such that 1 .
@xmath906 ( @xmath907 ) and 2 .
@xmath908 ( @xmath909 ) .
+ the composition of isomorphisms is given by addition . for any @xmath739-deformation of @xmath55 ,
the identity isomorphism is given by @xmath910 .
[ bundle def sheaf ] given an open set @xmath87 , let us introduce the notation @xmath911 for every inclusion of open subset @xmath912 , we then have a natural restriction functor ( between formal groupoids ) @xmath913 altogether , we obtain a ( strict ) sheaf of formal groupoids on @xmath5 sending @xmath298 to the formal groupoid @xmath914 .
[ pi zero bundle]every @xmath739-deformation of @xmath55 is isomorphic to one of the form @xmath915 , i.e. to one with undeformed transition functions @xmath916 . given @xmath917 , by part ( ii ) of definition [ herm def ] we see that @xmath918 the functions @xmath919 therefore define a cech cocycle . since the sheaf @xmath920 of smooth , real - valued functions on @xmath5 admit partitions of unity ,
so does the sheaf @xmath921 .
we may therefore choose @xmath922 satisfying @xmath923 on the overlaps @xmath867 .
this implies that @xmath924 is isomorphic to @xmath925 .
[ group action on category ] let @xmath755 be a groupoid , and @xmath760 a group .
a _ strict left action of @xmath760 on @xmath755 _ is a collection of functors @xmath926 satisfying the following : 1 . for every @xmath927 , we have @xmath928 .
the functor @xmath929 induced by the identity element @xmath930 is the identity functor . a strict _ right _
action is defined similarly , except that for every @xmath927 we require that @xmath931 .
given a formal group @xmath760 and a formal groupoid @xmath755 , we may define an action of @xmath760 on @xmath755 to consist of an action of @xmath761 on @xmath932 for each @xmath735 , subject to some obvious compatibility relations . given a strict left action of @xmath760 on @xmath755
, we will use the notation @xmath933 for each @xmath513 and each object @xmath934 .
similarly , given a morphism @xmath935 in @xmath755 we denote @xmath936 by @xmath937 .
similarly , if @xmath760 acts on @xmath755 on the right , we will use the notation @xmath938 and @xmath939 .
we now describe an action of @xmath811 ( the formal symmetries of the courant algebroid ) on @xmath888 .
[ bundle action ] given @xmath735 , @xmath940 , and @xmath941 , let @xmath942 given an isomorphism @xmath943 in @xmath944 , define @xmath945 we then have the following easy result , the proof of which is omitted . [ bundle action]definition [ bundle action ] determines a strict left action of @xmath811 on @xmath888 .
we now turn to the deformations of gc branes . for each gc brane @xmath0 on a gc manifold @xmath56
, we will define a formal groupoid @xmath946 encoding the infinitesimal deformations of @xmath0 and their equivalences ( definition [ full deformation groupoid ] ) . by passing to equivalence classes of deformations ,
we construct from @xmath0 a functor ( definition [ brane deformation functor ] ) @xmath947 we will see that in many situations it is necessary to work with the formal groupoid @xmath948 itself , and not the functor @xmath92 . the construction of @xmath68 will be given in several steps .
first , we ignore the gc structure @xmath58 and define a formal groupoid @xmath949 that does not encode any compatibility condition with respect to @xmath58 .
we then define a sub - groupoid @xmath950 whose objects are those deformations compatible with @xmath58 .
the formal group @xmath951 of symmetries of @xmath56 acts on @xmath950 ; in particular , there is an action of @xmath952 , the formal generalized hamiltonian symmetries . incorporating this action leads to the formal groupoid @xmath948 .
[ brane definition 1 ] let @xmath5 be a smooth manifold . a ( rank 1 ) _ brane _ on @xmath5 is a pair @xmath57 , where @xmath6 is a smooth submanifold and @xmath55 is an element of @xmath953 .
we denote the collection of all such branes on @xmath5 by @xmath954 .
given such a brane @xmath54 , applying the gluing construction described in remark [ gluing bundles ] to @xmath55 determines a gc brane in the sense of remark [ geometric brane ] .
[ definition par def ] given a manifold @xmath5 , and a brane @xmath955 , let @xmath956 be the following formal groupoid
. 1 . for each @xmath735 , an object of @xmath957 is a pair @xmath958 , where 1 .
@xmath959 is of the form @xmath960 for some @xmath823 , where @xmath961 denotes the pull - back map for the inclusion @xmath410 .
@xmath924 is an object of @xmath962 , as in definition [ herm def ] .
+ we will refer to such an object as an _ @xmath739-deformation of @xmath0_. 2 . given two @xmath739-deformations @xmath958 and @xmath963 , an _ equivalence _
@xmath964 ^ -{\cong } & } } \hb'$ ] is a pair @xmath965 , where 1 .
@xmath966 satisfies @xmath967 and 2 .
@xmath968 is a morphism @xmath969 in the groupoid @xmath970 .
3 . given equivalences @xmath971
^ -{\cong } & } } \hb'$ ] and @xmath972 ^ -{\cong } & } } \hb''$ ] , their composition is defined by @xmath973 ^ -{\cong } & } } \hb''.\ ] ] for any @xmath739-deformation @xmath974 of @xmath0 , the _ identity _ isomorphism @xmath975 is given by @xmath976 .
let us examine the composition in @xmath977 in a little more detail ( and in particular verify that it is well - defined ) .
suppose @xmath965 is an equivalence from @xmath978 to @xmath979 and @xmath980 is an equivalence from @xmath981 to @xmath982 . by definition
, we have @xmath983 , @xmath984 , @xmath968 is an equivalence @xmath985 and @xmath986 is an equivalence @xmath987 . defining @xmath988 , it follows that @xmath989 .
furthermore , @xmath990 is an equivalence from @xmath991 , so that the composition @xmath992 is an equivalence from @xmath993 ; therefore @xmath994 does in fact define a morphism in @xmath977 from @xmath974 to @xmath995 .
writing the composition as on the left - hand side of ( [ comp identity ] ) makes it easy to check its associativity using proposition [ bundle action ] .
[ right action par ] there is a strict right action of the formal group @xmath811 on @xmath996 , given as follows : 1 . for each @xmath735 ,
each @xmath997 , and each @xmath998 , define @xmath999 where we recall that by definition @xmath1000 2 . for each equivalence
@xmath1001 ^ -{\cong } & } } \hb_2 $ ] , define @xmath1002 to be given by the same pair @xmath1003 , regarded as an equivalence @xmath964 ^ -{\cong } & } } \hb'$ ] .
first , let us check that if @xmath965 is an equivalence @xmath1004 ^ -{\cong } & } } \hb_2 $ ] , then the same pair @xmath1003 does in fact define an equivalence @xmath1005 ^ -{\cong } & } } \hb_2\cdot g$ ] .
let us write @xmath1006 , @xmath1007 , @xmath1008 , and similarly write @xmath1009 with @xmath1010 , and @xmath1011 with @xmath1012 and @xmath1013 .
note that since @xmath1014 , we have @xmath1015 .
furthermore , since @xmath1016 is an equivalence from @xmath1017 in @xmath970 , it follows that @xmath1018 is an equivalence from @xmath1019 to @xmath1020 .
but we have @xmath1021 so that @xmath1022 is a morphism from @xmath1023 in @xmath1024 . thus , we see that @xmath1025 does indeed define an equivalence from @xmath1026 to @xmath1027 , as claimed .
since @xmath376 acts trivially on morphisms , the functoriality property @xmath1028 holds trivially .
furthermore , by construction , the identity element of @xmath811 acts as the identity functor . therefore ,
to finish the proof of the proposition we just need to check that @xmath1029 hold for each object @xmath1030 and each @xmath1031 .
let @xmath1032 and @xmath1033 , @xmath1034 .
we have @xmath1035 since @xmath1036 , we see that the right - hand side of ( [ ggprime ] ) is equal to @xmath1037 .
every @xmath1038 determines a generalized submanifold @xmath408 on @xmath5 , where @xmath457 is the curvature form of @xmath55 .
in particular , we have the _ generalized tangent bundle _ @xmath1039 and the _
generalized normal bundle _
@xmath1040 , which by definition are the generalized tangent bundle and generalized normal bundle of @xmath408 ( as defined in
[ gc submanifolds ] ) .
similarly , we denote by @xmath1041 the space @xmath426 introduced in definition [ definition k ] , and by @xmath1042 the lie algebroid cohomology groups associated to @xmath408 ( described in [ lie alg cohomology ] ) .
recall also the notation @xmath1043 [ gc brane definition 1 ] let @xmath56 be a gc manifold .
a ( rank 1 ) _ gc brane _ on @xmath56 is an element @xmath1044 as in definition [ brane definition 1 ] , such that the underlying generalized submanifold @xmath408 of @xmath0 is compatible with @xmath58 .
we denote the collection of all such gc branes by @xmath1045 .
[ compatibility1 ] given @xmath1046 , and an artin algebra @xmath735 , an element @xmath1047 is said to be__compatible _ _ with @xmath58 with respect to @xmath0 if @xmath1048 where @xmath1049 is the @xmath739-linear extension of pairing defined by ( [ q def ] ) , @xmath1050 and @xmath1051 .
[ well - defined ] let @xmath1052 , and let @xmath1033 and @xmath1053 be elements of @xmath812 such that 1 .
@xmath1054 and 2 .
@xmath1055 .
then @xmath376 is compatible with @xmath58 with respect to @xmath0 if and only if @xmath1056 is .
suppose we are given @xmath1057 such that @xmath1058 ; this is equivalent to requiring that @xmath1059 .
suppose also that we are given @xmath1060 that is tangent to @xmath7 , with @xmath1061 ; this means that @xmath1062 for every @xmath1063 .
we claim that in this case @xmath1064 is still tangent to @xmath7 with @xmath1065 . to see this
, note that for every @xmath1066 we have @xmath1067 next , suppose we are given @xmath1068 such that @xmath1069 and @xmath1070 .
we claim that in this case @xmath1071 to see this , let @xmath1072 be an arbitrary element of @xmath1073 . by definition , this means that @xmath432 is tangent to @xmath7 , say with @xmath1074 , and @xmath1075 .
write @xmath1076 , with @xmath1077 and @xmath1078 we have already seen that @xmath1079 is tangent to @xmath7 with @xmath1080 . furthermore , we see that @xmath1081 so that @xmath1082 , as claimed . to complete the proof , let @xmath1083 and @xmath1053 with @xmath1054 and @xmath1055 .
defining @xmath1084 , we see that @xmath1085 defining @xmath86 by the equation @xmath1086 and @xmath1087 , we see that @xmath1088 , and that @xmath1089 and @xmath1090 as shown above this implies @xmath1071 therefore , we have @xmath1091 since @xmath1092 , in particular both @xmath86 and @xmath1093 are tangent to @xmath7 , so that it follows that @xmath1094 . combining this with the equality ( [ q equal ] )
, we obtain the desired result . with proposition
[ well - defined ] in hand , we are ready to define what it means for a deformation of gc brane to be compatible with the gc structure .
let @xmath1052 , and let @xmath958 be an @xmath739-deformation of @xmath0 as defined in definition [ definition par def ] .
choose @xmath1095 such that @xmath960 .
also , for each @xmath1096 , let @xmath1097 be an open set such that @xmath1098 and choose @xmath1099 such that @xmath1100 .
set @xmath1101 , and define @xmath1102 [ compatibility def ] we say that @xmath1103 is _ compatible _ with @xmath58 if each @xmath1104 ( chosen as above ) is compatible with @xmath1105 with respect to @xmath1106 .
we denote by @xmath1107 the formal subgroupoid of @xmath1108 , whose objects for each @xmath735 are the @xmath739-deformations compatible with @xmath58 .
it is clear from proposition [ well - defined ] that the condition in definition [ compatibility def ] is well - defined , i.e. does not depend on the particular choices of @xmath84 , @xmath1109 or @xmath1110 .
[ brane def sheaf ] similarly to the situation described in remark [ bundle def sheaf ] , the formal groupoid @xmath950 extends in a natural way to a presheaf of formal groupoids on @xmath5 .
the fact that this is actually a sheaf ( satisfies the descent condition ) is proposition [ descent property ] .
[ compatibility under isomorphism ] let @xmath1111 be isomorphic elements of @xmath977 . then @xmath974 is compatible with @xmath58 if and only if @xmath981 is compatible with @xmath58 .
write @xmath958 and @xmath1112 ( with the same underlying submanifold @xmath6 ) with @xmath1113 and @xmath1114 , and let @xmath965 be an isomorphism @xmath1115 . as in definition [ compatibility def ] ,
let @xmath1109 be open sets in @xmath5 with @xmath1116 , and choose @xmath820 and @xmath1117 such that @xmath960 and @xmath1118 .
define @xmath1119 , where @xmath1120 is the restriction of @xmath84 to @xmath1121 .
according to definition [ compatibility def ] , @xmath974 is compatible with @xmath58 if and only if , for each @xmath1096 we have @xmath1122 where @xmath1123 and @xmath1124 are defined here with respect to @xmath1121 .
let @xmath1125 be an extension of @xmath1126 , and define @xmath1127 by @xmath1128 . then @xmath1129 furthermore , we have @xmath1130 which implies that @xmath1131 choose @xmath1132 such that @xmath1133 , and @xmath1134 such that @xmath1135
. then @xmath1136 set @xmath1137 and @xmath1138 , where @xmath1139 is the restriction of @xmath1140 to @xmath1121 .
then @xmath981 is compatible with @xmath58 if and only if for each @xmath1096 we have @xmath1141 we calculate @xmath1142 we have @xmath1143 defining @xmath1144 we have @xmath1145 note that @xmath1146 is an element of @xmath1147 , so that by lemma [ k closed under bracket ] we have @xmath1148 using the fact that @xmath1149 acts trivially on @xmath813 , and writing @xmath1150 , we see that @xmath1151 since @xmath432 is tangent to @xmath7 , it follows that
@xmath1152 so we see that condition ( [ compat1 ] ) is indeed equivalent to condition ( [ compat2 ] ) .
[ action of formal symmetries ] given @xmath1153 and @xmath1154 , the deformation @xmath1155 is compatible with @xmath58 if and only if @xmath974 is .
in particular , the action of @xmath811 on @xmath956 restricts to give a well - defined right action of the formal subgroup @xmath951 on @xmath1107 .
given @xmath1156 compatible with @xmath58 , choose @xmath1109 , @xmath1157 , and @xmath84 , as in definition [ compatibility def ] , and set @xmath1158 . by definition , for each @xmath1096 we have @xmath1159 given @xmath1160 , write @xmath1161 .
then we have @xmath1162 and @xmath1163 with @xmath1164 . defining @xmath1165 by @xmath1166
, we see that the compatibility of @xmath981 with @xmath58 is equivalent to the conditions @xmath1167 letting @xmath1168 , we see that @xmath1169 since by assumption @xmath376 is a symmetry of the gc structure @xmath58 , it follows that the left - hand sides of ( [ compat1 ] ) and ( [ compat2 ] ) are equal .
recall from definition [ genhamdef ] and proposition [ hamiltonian ] the subalgebra @xmath851 of generalized hamiltonian vector fields on @xmath56 .
we correspondingly have the formal subgroup@xmath1170 of formal generalized hamiltonian symmetries .
[ full deformation groupoid ] given @xmath1052 , let @xmath946 be the following formal groupoid ( over @xmath733 ) .
for each @xmath735 , the groupoid @xmath1171 has the same objects as @xmath1172 . given objects @xmath974 and @xmath981 , a morphism in @xmath1173 from @xmath974 to @xmath981 is a pair @xmath1174 , where @xmath1175 and @xmath1016 is a morphism in @xmath1172 from @xmath974 to @xmath1176 .
given @xmath1177 and @xmath1178 , the composition is defined by @xmath1179 the identity morphism of any @xmath974 is the pair @xmath1180 , where @xmath1181 denotes the identity morphism of @xmath974 considered as an object in @xmath1172 .
one easily checks using proposition [ right action par ] that the composition given by ( [ composition ] ) is well - defined , unital , and associative .
[ brane deformation functor ] the _ deformation functor _ @xmath1182 of @xmath1183 is given by @xmath1184
in @xcite , an argument was given that first - order deformations of a gc brane @xmath0 should correspond to elements of the lie algebroid cohomology group @xmath72 . in our framework
, first - order deformations of @xmath0 are encoded as elements of @xmath1185(\epsilon^2))$ ] ; therefore , a natural expectation motivated by @xcite is that elements of @xmath1186(\epsilon^2))$ ] should correspond to classes in @xmath72 . in this section ,
we verify this by unpacking definition [ brane deformation functor ] in the special case @xmath1187 . on the hand hand , this result ( stated as theorem [ brane tangent ] in the introduction ) can be regarded as a rigorous verification of the result obtain in @xcite .
going in the other direction , we may view theorem [ brane tangent ] as a check that our definition [ brane deformation functor ] is a reasonable one from a geometric point of view
. we will actually prove theorem [ brane tangent ] as part of a slightly better statement , given below as proposition [ first order induced ] .
namely , we will construct an explicit map @xmath1188 , and prove that it is both well - defined and bijective .
we now turn to the construction .
let @xmath1046 be a gc brane , with generalized tangent bundle @xmath1039 . recall that , since @xmath0 is compatible with @xmath58 , the gc structure @xmath58 preserves @xmath1189 , and
we define @xmath1190 to be the @xmath240-eigenbundle , which is a complex sub - bundle of @xmath1191 .
recall also that the generalized normal bundle @xmath1040 is defined as the quotient of @xmath1192 by @xmath1039 ( with quotient map @xmath1193 ) . since @xmath1039 is a maximal isotropic subbundle of @xmath1192 ,
the pairing on @xmath1192 gives a well - defined non - degenerate pairing of @xmath1039 with @xmath1194 .
furthermore , since @xmath58 preserves @xmath1039 , it determines a well - defined complex structure @xmath1195 on @xmath1040 ; regarding @xmath1040 as a complex vector bundle with complex structure @xmath1196 , we have an isomorphism of complex vector bundles @xmath1197 given by @xmath1198 for every section @xmath322 of @xmath1192 .
given @xmath1199 , choose a section @xmath1200 extending @xmath322 , i.e. such that @xmath524 . as described in proposition [ right action par ] , the generalized vector field @xmath528 determines an object of @xmath1201 given by @xmath1202 the deformation @xmath1203 depends only on @xmath322 , i.e. not on the choice of the extension @xmath528 .
suppose @xmath1204 satisfies @xmath1205 , i.e. @xmath432 vanishes at each point of @xmath7 .
then for every @xmath1206 , we have @xmath1207 this implies that , for every @xmath1208 , we have @xmath1209 therefore , given two different extensions @xmath1210 of @xmath84 , we have @xmath1211 this in turn implies that @xmath1212 is equal to @xmath1213 .
it is also clear that @xmath1214 depends only on @xmath1215 , so that @xmath1214 is independent of the choice of extension .
since the deformation depends only on @xmath322 , we will use the notation @xmath1216 . note that the section @xmath322 also determines a section @xmath1217 as well as a section @xmath1218 .
[ first order induced ] 1 .
the deformation @xmath1219 is compatible with @xmath58 if and only if @xmath1220 , i.e. if and only if @xmath1221 is a generalized holomorphic section of @xmath1040
. 2 . given @xmath1222 such that both @xmath1221 and @xmath1223 are generalized holomorphic sections of @xmath1040 , the deformations @xmath1224 and @xmath1225 are isomorphic in @xmath1226 if and only if the elements @xmath1227,[\mu q(\xx')]\in h^1(\b)$ ] are equal .
in other words , there is a well - defined injective map @xmath1188 mapping @xmath1227 $ ] to the equivalence class of @xmath1224 .
every element of @xmath1226 is isomorphic to one of the form @xmath1224 for some @xmath1228 . in other words ,
the injective map @xmath1229 ^ -{\cong } & } } \df_{\b}(\dn)$ ] is actually a bijection .
let us begin with part ( 1 ) . by definition [ compatibility def ]
, we see that @xmath1224 will be compatible with @xmath58 if and only if @xmath1230 is compatible with @xmath58 in the sense of definition [ compatibility1 ] , i.e. if and only if we have @xmath1231 since @xmath1232 , this is equivalent to @xmath1233 by definition , this means that for every @xmath1234 we have @xmath1235 suppose condition ( [ compatibility cond1 ] ) holds , and let @xmath1236 be arbitrary sections .
choose sections @xmath1237 which extend @xmath1238 , i.e. such that @xmath1239 and @xmath1240 .
since @xmath1241 , condition ( [ compatibility cond1 ] ) implies that @xmath1242 making use of the identities ( [ dorfman1 ] ) and ( [ dorfman2 ] ) satisfied by the dorfman bracket , we calculate the the right - hand side of ( [ compatibility condit2 ] ) is equal to @xmath1243 where in the second line the expression @xmath1244 denotes the lie algebroid bracket of the sections @xmath1245 , as defined in ( [ brane bracket ] ) .
thus , we see that if @xmath1224 is compatible with @xmath58 , then @xmath1221 is a generalized holomorphic section of @xmath1040 .
conversely , suppose we have @xmath1220 .
given @xmath1246 , define @xmath1247 by @xmath1248 and @xmath1249 , and note that @xmath1250 . using the above calculation ( [ compatibility condit3 ] ) , we see that @xmath1251 in particular , this implies @xmath1252 holds for arbitrary @xmath1246 ; as previously mentioned , this condition is equivalent to the compatibility of @xmath1224 with @xmath58 .
this completes the proof of part ( 1 ) of proposition [ first order induced ]
. we will prove part ( 2 ) of proposition [ first order induced ] via a series of lemmas .
[ coboundary isomorphism ] suppose we are given @xmath1253 such that @xmath1254 .
then there exists an isomorphism @xmath1255 ^ -{\cong } & } } \hb^{\xx'}$ ] in @xmath1256 .
let us write @xmath57 , with @xmath1257 , and also @xmath351 and @xmath1258 .
also , let @xmath1259 be a choice of extension of @xmath322 and @xmath1260 be a choice of extension of @xmath1261 . by definition
, the assumption @xmath1254 implies that @xmath1262 is a section of @xmath1039 .
thus , we have that @xmath1263 is tangent to @xmath7 , and @xmath1264 where @xmath1265 and @xmath457 is the curvature form of @xmath55 . defining @xmath1266
, we claim that @xmath1267 is an isomorphism @xmath1255 ^ -{\cong } & } } \hb^{\xx'}$ ] in @xmath1201 .
recall that @xmath1268 and @xmath1269 according to definition [ definition par def ] , we need to check three conditions to show that @xmath627 does indeed define an isomorphism from @xmath1270 : 1 .
@xmath1271 , 2 .
@xmath1272 3 .
@xmath1273 condition ( 1 ) is equivalent to @xmath1274 , which holds by construction . to check condition
( 2 ) , first note that , since we are working over @xmath1187 , we have @xmath1275 on the other hand , we calculate that @xmath1276 to check condition ( 3 ) , first note that it is equivalent to @xmath1277 we then calculate @xmath1278 [ restrict function ] for any smooth function @xmath61 , we have the equality @xmath1279 where @xmath365 is the generalized hamiltonian vector field associated to @xmath362 . for any section @xmath1280 ,
the vector field component @xmath1281 is tangent to @xmath7 .
therefore , if @xmath61 is any smooth function , we have @xmath1282 writing @xmath403 , recall that the corresponding generalized hamiltonian vector field is given by the formula @xmath1283 given a section @xmath1280 , we then calculate @xmath1284 returning to the proof of part ( 2 ) of proposition [ first order induced ] , suppose @xmath1227=[\mu q(\xx')]\in h^1(\b)$ ] .
then there exists a smooth function @xmath1285 such that @xmath1286 choose a smooth function @xmath1287 that extends @xmath362 , i.e. such that @xmath1288 . by lemma
[ restrict function ] , we have @xmath1289 or equivalently @xmath1290 this implies that @xmath1291 lemma [ coboundary isomorphism ] therefore implies that @xmath1224 is isomorphic to @xmath1292 in @xmath1293 , so that by definition @xmath1224 and @xmath1225 are isomorphic in @xmath1294 .
therefore @xmath1224 and @xmath1225 determine the same element of the set @xmath1295 to finish the proof of part ( 2 ) of proposition [ first order induced ] , it remains to show that if @xmath1219 and @xmath1296 are isomorphic in @xmath1226 , then @xmath1227=[\mu q ( \xx')]\in h^1(\b)$ ] .
let @xmath1297 be an isomorphism from @xmath1219 to @xmath1296 in @xmath1226 ; here @xmath1298 is an element of @xmath1299 , and @xmath1300 ^ -{\cong } & } } \hb^{\xx'}\cdot z\ ] ] is an isomorphism in @xmath1293 .
write @xmath351 , @xmath1258 , and@xmath1301 , a calculation similar to that used in the proof of lemma [ coboundary isomorphism ] implies the following three conditions : @xmath1302 @xmath1303 and @xmath1304 define @xmath1305 by @xmath1306 . using the assumption that @xmath1307 , together with condition ( [ condition22 ] ) , we see that on each overlap @xmath1308 we have @xmath1309 so that there exists a unique function @xmath1310 with @xmath1311 on each @xmath871 . using the cartan formula @xmath1312 ,
we may then rewrite condition ( [ condition33 ] ) as @xmath1313 choosing a smooth function @xmath1314 that extends @xmath1315 , equation ( [ hi condition ] ) is equivalent to @xmath1316 defining @xmath1317 , we have @xmath1318 equation ( [ condition11 ] ) together with equation ( [ hi condition1 ] ) imply that @xmath1319 is a section of @xmath1039 , so that @xmath1320 and therefore @xmath1321 where in the last equality we used lemma [ restrict function ] .
therefore , we see that @xmath1322=[\mu q(\xx)]\in h^1(\b),\ ] ] as claimed .
finally , part ( 3 ) of proposition [ first order induced ] follows easily from proposition [ pi zero bundle ] .
in this section we explain how various equivalence between gc branes @xmath0 and @xmath1323 induce equivalences between the corresponding formal deformation groupoids .
to do so , it will be convenient to rephrase the definition of the formal groupoid @xmath946 from the action of @xmath1324 on @xmath1107 in terms of a general construction .
[ definition action groupoid ] let @xmath755 be a groupoid , and @xmath760 a group with a strict right action on @xmath755 . we define a new category @xmath1325 , called the ( right ) _ action groupoid of @xmath760 acting on @xmath755 _ , which has the same objects as @xmath755 , and such that for pair of objects @xmath1326 we have @xmath1327 composition is defined by @xmath1328 for each object @xmath275 , the identity morphism from @xmath275 to itself in @xmath1325
is defined to be @xmath1329 , where @xmath1330 is the identity morphism in @xmath755 , and @xmath930 is the identity group element .
it is an easy exercise using definition [ group action on category ] to check that the composition given by equation ( [ compo def ] ) is well - defined , associative , and unital .
given a formal group @xmath760 ( over @xmath733 ) with a right action on a formal groupoid @xmath755 , we similarly define a formal groupoid @xmath1325 , which associates to each @xmath735 the groupoid @xmath1331 defined as above . for example , translating definition [ full deformation groupoid ] into this language , we see that @xmath1332 the proof of the following proposition follows easily from the definitions , and is omitted .
[ induced equivalence ] let @xmath755 be a groupoid equipped with a strict right action of the group @xmath760 , and @xmath1333 a groupoid equipped with a strict right action of a group @xmath1334 .
let @xmath1335 ^ -{\cong } & } } g'$ ] be a group isomorphism , and let @xmath1336 be a functor that intertwines the group actions in the sense that , for every @xmath934 and @xmath513 we have @xmath1337 , and for every @xmath1338 in @xmath755 we have @xmath1339 .
then there is a well - defined functor @xmath1340 which agrees with @xmath1341 on objects , and which maps a morphism @xmath1342 in @xmath1325 to @xmath1343 in @xmath1344 .
furthermore , @xmath1340 is an equivalence if and only if @xmath1341 is .
this result has an immediate extension to the case that @xmath755 and @xmath1333 are formal groupoids and @xmath1345 are formal groups .
we now return to our main topic .
let @xmath57 and @xmath1346 be two elements of @xmath1045 supported on the same submanifold @xmath7 .
furthermore , suppose that @xmath55 and @xmath1347 are defined with respect to the same open cover @xmath873 of @xmath7 .
write @xmath875 and @xmath1348 .
suppose we are given @xmath1349 satisfying @xmath1350 and @xmath1351 we say that @xmath1352 is an equivalence @xmath1353 ^ -{\cong } & } } \l'$ ] .
[ invariance 1 ] the equivalence
@xmath1354 ^ -{\cong } & } } \l'$ ] induces an equivalence of formal groupoids @xmath1355 ^ -{\cong } & } } \dt^{\b'}(x,\j)\ ] ] defined as follows
. given @xmath1356 and @xmath1357 , write @xmath1358 , and define @xmath1359 given a morphism @xmath1360 in @xmath1361 , we define @xmath1362 first , note that it is clear by inspection that the right - hand side of ( [ onobjects ] ) defines an @xmath739-deformation of @xmath1323 .
next , we must check that if @xmath1363 is a morphisms from @xmath1364 to @xmath1365 , then @xmath1366 as defined by equation ( [ gamma mor ] ) does in fact define a morphisms from @xmath1367 to @xmath1368 .
writing @xmath1369 and @xmath1370 , by definition [ definition par def ] , this means that @xmath1371 @xmath1372 and @xmath1373 we need to check that these same equations hold when @xmath1374 is replaced by @xmath1375 , @xmath1376 is replaced by @xmath1377 , and @xmath1378 is replaced by @xmath1379 .
equation ( [ psi1 ] ) is unchanged under these replacements , and thus continues to hold automatically .
we calculate @xmath1380 so we see that the primed " version of equation ( [ psi2 ] ) holds .
finally , we have @xmath1381 so the primed " version of equation ( [ psi3 ] ) also holds .
this finishes the proof that @xmath1366 does in fact define an isomorphism from @xmath1367 to @xmath1368 .
we need to verify that @xmath1382 is functorial , i.e. we must check that , given isomorphisms @xmath1383 and @xmath1384 that @xmath1385 .
if we write @xmath1386 and @xmath1387 , then by definition [ definition par def ] we have @xmath1388 and therefore @xmath1389 on the other hand , we calculate @xmath1390 to complete the proof , consider @xmath1391 , which is an equivalence from @xmath1392 .
it is clear by inspection that @xmath1393 is a ( strict ) inverse for @xmath1382 ; in particular , @xmath1382 is an equivalence of groupoids .
the equivalence @xmath1354 ^ -{\cong } & } } \l'$ ] induces an equivalence of formal groupoids @xmath1394 ^ -{\cong } & } } \de^{\b'}(x,\j),\ ] ] where @xmath1395 is defined as in proposition [ induced equivalence ] . in particular , this induces a well - defined natural isomorphism of functors
@xmath1396 ^ -{\cong } & } } \df_{\b'}.\ ] ] let @xmath1397 be an open cover of @xmath7 with indexing set @xmath1398 . recall that a _ refinement _ of @xmath873 is a pair @xmath1399 , where @xmath1400 is an open cover of @xmath7 , and @xmath1401 is a map of indexing sets such that for each @xmath1402 we have @xmath1403 given @xmath1404 defined with respect to the open cover @xmath873 , we obtain a new element @xmath1405 defined with respect to the open cover @xmath1406 .
explicitly @xmath1407 with @xmath1408 and @xmath1409 . in particular , if @xmath57 is a gc brane on @xmath56 , then the refinement @xmath1399 determines @xmath1410 , which is again a gc brain on @xmath56 .
furthermore , there is a natural functor of formal groupoids @xmath1411 [ invariance 2 ] the functor @xmath1412 is an equivalence of formal groupoids . in particular , it induces a natural isomorphism of functors @xmath1396 ^ -{\cong } & } } \df_{\sigma^*\b}.\ ] ] fix @xmath735 , and let @xmath1413 . using propositions [ pi zero bundle ] and , it follows that @xmath974 is isomorphic to an object of the form @xmath1414 for some @xmath1415 ; by definition , @xmath1416 is compatible with @xmath58 as in definition [ compatibility1 ] .
it is also clear that @xmath1417 .
this shows that @xmath1418 is essentially surjective . on the other hand ,
it is also clear that @xmath1419 is fully faithful , and therefore an equivalence .
finally , let @xmath0 be a gc brane on @xmath56 , and let @xmath350 be a 1-form .
this determines a new gc structure @xmath1420 .
writing @xmath57 , define @xmath1421 explicitly , if we write @xmath1257 , we have @xmath1422 using proposition [ action on gc submanifolds ] , it follows that @xmath1423 . for each @xmath735 , let us define @xmath1424 as follows : given @xmath1425 , we define @xmath1426 [ b transform on deformations ] 1 .
@xmath1427 is an object of @xmath1428 for each @xmath1429 .
2 . for each equivalence
@xmath1430 ^ -{\cong } & } } \hb_2 $ ] in @xmath1431 , @xmath1016 also determines an equivalence @xmath1432 ^ -{\cong } & } } \tilde{\phi}\da^u(\hb_2)$ ] in @xmath1433 , which we denote by @xmath1434 .
3 . the map taking @xmath1435 to @xmath1436 and @xmath1437 in @xmath1438 to @xmath1439 defines an equivalence of formal groupoids @xmath1440 ^ -{\cong } & } } \dt^{\b}(x,\j').\ ] ] the proof of part ( 2 ) is a straightforward computation ( similar to that in the proof of proposition [ invariance 1 ] ) , and is therefore omitted . to verify part ( 1 ) , let @xmath1109 be a collection of open sets in @xmath5 chosen as in definition [ compatibility def ] . also choose @xmath1441 such that @xmath960 and @xmath1442 such that @xmath1443 . according to definition [ compatibility def ]
, the compatibility of @xmath974 with @xmath58 means that , for each @xmath1096 we have @xmath1444 by construction , @xmath1445 is given by @xmath1446 where we have defined @xmath1447 and @xmath1448 . therefore , to verify that @xmath981 is compatible with @xmath502 , we must check that for each @xmath1096 we have @xmath1449 we have @xmath1450 also , by proposition [ action on gc submanifolds ] we have @xmath1451 , so that the left hand side of ( [ comp2 ] ) is equal to @xmath1452 therefore , we see that condition ( [ comp1 ] ) is equivalent to condition ( [ comp1 ] ) , so the assumption that the former is satisfied implies that the latter is as well .
this completes the proof of part ( 1 ) . to see why part ( 3 ) of the proposition is true , first note that the functoriality of @xmath1453 follows trivially from its definition
also , by inspection @xmath1453 is a bijection on both objects and morphisms , and is in particular an equivalence of categories .
indeed , the functor @xmath1454 is easily verified to be the inverse of @xmath1455 . given @xmath66 and @xmath1456 , we have @xmath1457=(0,-\pounds(\xi)u)\ ] ] and @xmath1458=0.\ ] ] the adjoint action of @xmath1459 on @xmath1460 is therefore given by @xmath1461 if @xmath58 is a gc structure on @xmath5 , we clearly have @xmath1462 .
@xmath1463 recall from proposition [ exp prop ] the identity @xmath1464 therefore , we need to show that for every @xmath1465 there exists @xmath1466 such that @xmath1467 . given @xmath1468 , recall from definition [ genhamdef ] that @xmath1469 therefore , setting @xmath1470 we have @xmath1471 if @xmath377 is purely imaginary , we see from ( [ adjoint ] ) that @xmath1472 if @xmath362 is real , write @xmath1473 , then we see that @xmath1474 on the other hand , we have @xmath1475 let @xmath1476 be the isomorphism of formal groups given for each @xmath735 and @xmath1477 by @xmath1478 for fixed @xmath1477 , let @xmath1479 be the functor @xmath1480 determined by the right action of @xmath1481 on @xmath1482 , and similarly let @xmath1483 be the functor @xmath1484 . @xmath1485 .
the result follows from an explicit calculation very similar to that used in the proof of proposition [ right action par ] .
therefore , using the construction described in proposition [ induced equivalence ] , we may extend the functor @xmath1486 to a functor @xmath1487 .
furthermore , since @xmath1488 is an equivalence , proposition [ induced equivalence ] implies the following result .
the functor @xmath1487 is an equivalence of formal groupoids .
let @xmath83 be a complex manifold , and @xmath65 ^ -{\cong } & } } x$ ] a diffeomorphism satisfying @xmath1489 , i.e. a symmetry of @xmath83 . given any complex submanifold @xmath6 , the submanifold @xmath1490 will again be compatible with @xmath255 : symmetries of @xmath83 act on its collection of complex submanifolds . similarly , for every artin algebra @xmath739 and every holomorphic vector field @xmath1491
, the action of the formal symmetry @xmath1492 on @xmath7 induces a deformation of @xmath7 , i.e. an element @xmath35 of @xmath1493 .
since applying the inverse @xmath1494 to @xmath35 gives the trivial deformation ( z itself ) , we say that @xmath35 is trivializable .
of course , not every deformation is trivializable in this way ; however , for each point @xmath85 , we can find an open set @xmath87 containing @xmath88 , so that every deformation of @xmath1495 in @xmath1496 is induced by a formal symmetry of @xmath1496 .
similarly , given a gc brane @xmath1497 and a formal symmetry @xmath1498 for some @xmath1356 , we saw in proposition [ action of formal symmetries ] that @xmath1416 induces an @xmath739-deformation @xmath1499 of @xmath0 . motivated by the complex case described above , a natural question is whether every @xmath739-deformation is locally isomorphic to one of this form .
[ definition locally extendible ] let @xmath1500 be a gc brane .
we say that @xmath0 has _
deformations if for each @xmath1356 , and each @xmath739-deformation @xmath974 of @xmath0 , there exists a formal symmetry @xmath1501 such that @xmath974 is isomorphic to @xmath1499 in @xmath1502 . similarly , we say that @xmath0 has _ locally trivializable deformations _ if for each @xmath85 there exists an open set @xmath87 containing @xmath88 , such that @xmath1503 has trivializable deformations .
examination of examples shows that unlike in the complex case discussed above not every gc brane has locally trivializable deformations .
however , in this section we will prove the following result .
[ essentially surjective ] let @xmath1052 be a leaf - wise lagrangian brane ( that is , a brane whose underlying gc submanifold is leaf - wise lagrangian as in definition [ definition lwl submanifold ] ) .
then @xmath0 has locally trivializable deformations .
let @xmath57 be a lwl brane , and @xmath85 .
given a neighborhood @xmath87 of @xmath88 and a 1-form @xmath635 , let @xmath1504 and @xmath1505 . by proposition
[ action on gc submanifolds ] , we have @xmath1506 .
suppose that for every @xmath735 and @xmath1507 we can find @xmath1508 such that @xmath974 is isomorphic to @xmath1509 . then by applying proposition [ b transform on deformations ] , together with the fact that @xmath1510 it follows that same is true when @xmath1323 is replaced by @xmath0 and @xmath502 is replaced by @xmath58 .
using theorem [ lwl normal form ] , it is therefore sufficient to consider the case where @xmath5 is a neighborhood of the origin in @xmath1511 , that @xmath1512 for natural numbers @xmath1513 ( with @xmath556 ) , and where the curvature form @xmath457 of @xmath55 is zero ; here @xmath1514 is submanifold described in example [ standard brane ] .
furthermore , by applying propositions [ invariance 1 ] and [ invariance 2 ] , it is enough to consider the case that @xmath55 is the trivial element of @xmath953 .
the rest of the proof will be similar to the proof of theorem [ lwl normal form ] , and we begin by recalling some of the notation introduced there .
we introduce coordinates @xmath551 on @xmath1515 , where @xmath1516 are coordinates on @xmath554 and @xmath1517 are ( real ) coordinates on @xmath1518 .
we then relable these to obtain new coordinates @xmath1519 , with @xmath1520 and @xmath1521 , defined by @xmath1522 and @xmath1523 in terms of these coordinates , @xmath7 is defined by the equations @xmath1524 for @xmath1525 .
let @xmath958 be an object of @xmath1526 .
let @xmath1527 be the projection map @xmath1528 , and @xmath1529 the pull - back homomorphism .
there exists a unique @xmath1530 such that @xmath1531 suppose that @xmath1532 is a small extension in @xmath733 with kernel @xmath1096 , and let @xmath1533 be an @xmath734-linear splitting of @xmath312 .
inductively , assume that the result holds for the artin algebra @xmath747 .
write @xmath1534 , and let @xmath1535 , then exists a unique @xmath1536 such that @xmath1537 .
let @xmath1538 , i.e. for each @xmath1539 we have @xmath1540 define @xmath1541 such that for every @xmath1542 we have @xmath1543 define @xmath1544 .
noting that @xmath1545 lies in the center of @xmath1546 , we see that @xmath1547 furthermore , we claim that @xmath432and therefore @xmath1126is a derivation . to see this
, note that by proposition [ action ] @xmath1548 is a homomorphism ( actually automorphism ) of @xmath287-algebras .
therefore we have @xmath1549 by comparing this with the equation @xmath1550 we see that indeed @xmath1551 for every @xmath1552 .
therefore @xmath1553 is the identity .
since @xmath974 is isomorphic in @xmath1554 to @xmath1555 , we may assume without loss of generality
that @xmath1556 in terms of the coordinates @xmath1557 , this means that @xmath1558 and @xmath1559 let us write @xmath1560 , and define @xmath1561 @xmath960 . by definition , there exists some @xmath1562 such that @xmath1563 defining @xmath1564 by @xmath1565 , we see that the desired result is equivalent to the equation @xmath1566 introducing the notation @xmath1567 for the collective coordinates @xmath1568 on @xmath1569 , by assumption we have @xmath1570 for each @xmath1571 ; we want to show that this implies the identity @xmath1572 for arbitrary @xmath1573 .
let us expand @xmath1574 , noting that the component functions @xmath1575 are elements of @xmath1576 .
for every function @xmath1573 , we have @xmath1577 in particular , if we can show that each component function @xmath1575 satisfies @xmath1578 , then @xmath1579 for every @xmath1573 ; in this case we say that the _ restriction of @xmath1580 to @xmath7 vanishes_. inductively , for each @xmath1581 we then have @xmath1582 for every @xmath1573 , which in turn implies that @xmath1583 .
therefore , it is sufficient to show that , if @xmath1580 is any element of @xmath801 such that @xmath1570 holds for each @xmath1571 , then each of the component functions @xmath1575 satisfies @xmath1578 ( equivalently , the restriction of @xmath1580 to @xmath7 vanishes ) .
inductively , assume that this result holds for some @xmath1584 , and let @xmath1532 is a small extension with kernel @xmath1096 .
define @xmath1585 .
since @xmath1586 and @xmath1587 , it follows that @xmath1588 for each @xmath1571 , so by the inductive hypothesis the restriction of @xmath1589 to @xmath7 vanishes .
let @xmath1533 be a linear splitting of @xmath312 , and define @xmath1590 , which by construction is an element of @xmath1591 .
it is easy to see that the vanishing of @xmath1589 along @xmath7 implies that @xmath1592 vanishes along @xmath7 as well , so that @xmath1593 . for each @xmath1573
we therefore have @xmath1594 expanding @xmath1595 , and noting that the component functions satisfy @xmath1596 , we then have @xmath1597 therefore @xmath1598 vanishes along @xmath7 , so that @xmath1599 does as well .
let us write @xmath1600 for @xmath1601 .
define @xmath1602 , and @xmath1603 .
the proof of theorem [ essentially surjective ] then follows from the following lemma .
[ useful lemma ] 1 .
@xmath1604 and 2 .
@xmath1605 . as in the proof of theorem
[ lwl normal form ] , we introduce @xmath645 as the sub - bundle spanned by @xmath646 and @xmath647 as the sub - bundle spanned by @xmath648 . in that proof , we saw that 1 .
@xmath649 . furthermore
, both @xmath287 and @xmath543 are maximal isotropic so the pairing gives an identification @xmath650 and @xmath651 .
@xmath1606 and @xmath1607 .
3 . @xmath1608 .
we also introduced the space @xmath655 of vector fields on @xmath572 of the form @xmath656 for constants @xmath657 , and defined @xmath658 and @xmath659 to be the subspaces of elements @xmath660 satisfying @xmath661 for all @xmath662 . via straightforward calculation , we proved the following lemma , which we state here again for convenience .
[ rslemma ] 1 .
@xmath1609 2 .
@xmath1610 3 .
with respect to the isomorphism @xmath1611 induced by the pairing , we have @xmath1612 by construction , both @xmath711 and @xmath712 are elements of @xmath694 , so that we have @xmath1613 using proposition [ exponential ] , we see that @xmath1614 from which the first part of lemma [ useful lemma ] easily follows .
it remains to show that @xmath1615 since @xmath670 , it follows from lemma [ rslemma ] that @xmath1616 , and also that @xmath1617 .
since the pairing identifies @xmath287 with @xmath1618 , we see that @xmath1619 vanishes if and only if @xmath1620 this is equivalent to @xmath1621 which is in turn equivalent to @xmath1622 since @xmath1623 , it follows from the discussion in remark [ rem1 ] that @xmath1624 on the other hand , for each @xmath1625 the function @xmath1626 is independent of the @xmath1627 coordinates , so if it vanishes on @xmath7 it vanishes on all of @xmath5 .
we therefore conclude that @xmath1621 which finishes the proof of the lemma .
in this section , we continue the study of induced ( trivializable ) deformations . fix a gc brane @xmath1052 . for each @xmath735 and @xmath1628
, we introduce the notation @xmath1629 in this section we will define a formal groupoid @xmath1630 , whose objects for @xmath735 are the elements of @xmath1631 .
we then extend the map @xmath1632 to a functor of formal groupoids @xmath1633 which we prove is fully faithful ; by construction , it is essentially surjective ( and hence an equivalence ) precisely when @xmath0 has trivializable deformations . in order to define the groupoid @xmath1630
we will need some preliminary definitions .
let @xmath1634 be the following lie algebra : as a vector space @xmath1635 , and the bracket is given by @xmath1636=([\xi,\eta],\xi(g)-\eta(f)+\iota(\eta)\iota(\xi)f).\ ] ] the jacobi identity for the bracket ( [ r bracket ] ) follows from the fact that @xmath457 is closed .
consider the formal group @xmath1637 .
for each @xmath735 , by proposition [ exp prop ] we see that for every @xmath1638 and every @xmath1639 we have @xmath1640 we also have the following result ; the proof is identical to the proof of proposition [ exponential ] , and is therefore omitted .
[ exponential r ] for each @xmath735 and @xmath1641 , we have @xmath1642 the map @xmath1643 given by @xmath1644 is a lie algebra homomorphism
. along with the cartan formula @xmath193 , recall also the identity @xmath192)f=[\pounds(\xi),\iota(\eta)]f$ ] . using these
, we calculate @xmath1645 $ ] to be @xmath1646 \\ = & ( [ \xi,\eta],\pounds(\xi)\iota(\eta)f-\pounds(\xi)dg-\pounds(\eta)\iota(\xi)f+\pounds(\eta)df)\\ = & ( [ \xi,\eta],\iota([\xi,\eta])f+\iota(\eta)\pounds(\xi)f - d\iota(\xi)dg-\pounds(\eta)\iota(\xi)f+d\iota(\eta)df)\\ = & ( [ \xi,\eta],\iota([\xi,\eta])f+\iota(\eta)d\iota(\xi)f - d\iota(\eta)\iota(\xi)f-\iota(\eta)d\iota(\xi)f - d(\xi\cdot g-\eta\cdot f))\\ = & ( [ \xi,\eta],\iota([\xi,\eta])f - d(\xi\cdot g-\eta\cdot f+\iota(\eta)\iota(\xi)f ) ) \\ = & \mu(([\xi,\eta],\xi\cdot g-\eta\cdot f+\iota(\eta)\iota(\xi)f ) ) \\ = & \mu([(\xi , f),(\eta , g)]).\end{aligned}\ ] ] let @xmath0 be a gc brane on @xmath56 , with underlying gc submanifold @xmath408
. denote by @xmath1647 the lie sub - algebra of @xmath367 consisting of elements @xmath1648 such that @xmath84 is tangent to @xmath7 .
this lie algebra comes equipped with a restriction homomorphism @xmath1649 [ definition kkk ] let @xmath1650 be the fiber - product of lie algebras @xmath1651 in other words , an element of @xmath1650 is a pair @xmath1652 , where @xmath1653 , @xmath1654 , such that @xmath1655 and @xmath1656 the lie bracket is defined component - wise . by construction
, @xmath1650 comes equipped with three homomorphisms : 1 .
@xmath1657 2 .
@xmath1658 and 3 .
@xmath1659 , where @xmath1660 .
we will denote the group homomorphisms induced by these lie algebra homomorphisms by the same symbols .
[ definition tilde extended ] let @xmath1661 be the following formal groupoid : for each @xmath735 , the objects of @xmath1662 are elements of the group @xmath1631 .
a morphism from @xmath1628 to @xmath1663 is an element @xmath1664 such that @xmath1665 given @xmath1666 and @xmath1667 such that @xmath1668 and @xmath1669 , the composition @xmath1670^{y ' } & x'')}\circ ( \xymatrix { x \ar[r]^y & x'})\ ] ] is defined by @xmath1671^{y'y } & x''},\ ] ] where we note that @xmath1672 for every @xmath1628 , the identity morphism @xmath1673 is the group identity element @xmath1674 . for each @xmath1628 , denote @xmath1675 given @xmath1676 and @xmath1677 such that @xmath1668 , we will construct @xmath1678 write @xmath1679 , @xmath1680 , @xmath1681 , where @xmath1682 and @xmath1683 . by proposition [ exponential ] , we have @xmath1684 where @xmath1685 the equation @xmath1668 implies that @xmath1686 and @xmath1687 furthermore , by definition of the lie algebra @xmath1688 , we have @xmath1689 and @xmath1690 [ sigma morphism lemma ] @xmath1691 with @xmath1692
is an equivalence from @xmath1693 to @xmath1694 in @xmath1502 .
write @xmath1695 and @xmath1696 . according to definition [ definition par def ]
, we need to check that @xmath1697 and that @xmath1698 is an isomorphism in @xmath1699 from @xmath985 . by construction
, we have @xmath960 and @xmath1700 . combining equation ( [ first condition ] ) with equation ( [ third condition ] )
, we see that ( [ iso condition 1 ] ) is indeed satisfied .
next , note that by construction @xmath1701 and @xmath1702 therefore , we also have @xmath1703 according to definition [ herm def ] , therefore , @xmath1698 is an isomorphism in @xmath970 from @xmath985 if and only if we have @xmath1704 and @xmath1705 by equation ( [ g definition ] ) we have @xmath1706 which verifies condition ( [ iso condition 2 ] ) . to verify condition ( [ iso condition 3 ] ) , we calculate @xmath1707 note that , in going from the first line to the second we used the cartan formula for the lie derivative , together with ( [ fourth condition ] ) ; in going from the second to the third we used that @xmath1708 , @xmath1709 , and ( [ zeroth condition ] ) ; in going from the third to the fourth we used ( [ second condition ] ) ; and in going from the fourth line to the last we used ( [ third condition ] )
. [ functoriality prop ] the map sending @xmath1628 to @xmath1710 and @xmath1711 to @xmath1712 is functorial ; that is , for every @xmath1667 and @xmath1666 such that @xmath1668 and @xmath1669 , we have @xmath1713 suppose @xmath1667 and @xmath1714 satisfy @xmath1668 and @xmath1669 .
defining @xmath1715 , we may rewrite the condtion ( [ functoriality1 ] ) in the slightly altered form @xmath1716 writing @xmath1717 , @xmath1718 , and @xmath1719 , by definition [ definition par def ] equation ( [ functoriality ] ) means that @xmath1720 and that @xmath1721 holds for each @xmath1096 . using formula ( [ r multiplication ] ) , we see that the conditions ( [ fun cond 1 ] ) and ( [ fun cond 2 ] ) may be combined into the equations @xmath1722 in the groups @xmath1723 , where by definition we have @xmath1724 and @xmath1725 .
define @xmath1726 , @xmath1727 , and @xmath1728 , and also @xmath1729 , @xmath1730 , and @xmath1731 . using the definition ( [ g definition ] ) together with lemma [ exponential r ]
, we see that ( [ group prop ] ) may be rewritten as @xmath1732 by hypothesis , we have @xmath1715 ; applying the homomorphism @xmath1733 and restricting to @xmath869 , this implies that @xmath1734 holds in @xmath1723 . on the other hand ,
if we define @xmath1735 by @xmath1736 we see that @xmath1737 , @xmath1738 and @xmath1739 .
it is easy to check that @xmath1740 is a homomorphism of lie algebras ; therefore equation ( [ group prop 2 ] ) follows from ( [ group prop 2 ] ) by applying the group homomorphism @xmath1741 . using lemma [ sigma morphism lemma ] and proposition [ functoriality prop ] in hand
, we may now give the following definition .
[ definition sigma tilde ] the functor @xmath1742 is defined as follows . for each @xmath735 and each @xmath1628 , we have @xmath1743 on morphisms @xmath1744 is defined as described in lemma [ sigma morphism lemma ] .
[ fully faithful ] the functor @xmath1745 is fully faithful .
we first prove that @xmath1745 is faithful .
given @xmath735 , let @xmath1746 be elements of @xmath1631 , and let @xmath1666 be elements satisfying @xmath1668 and @xmath1747 .
in particular , this implies that @xmath1748 .
let us write @xmath1749 , where @xmath1750 and @xmath1751 , such that @xmath1752 and @xmath1753 .
then @xmath1754 is of the form @xmath1755 , where @xmath1756 . by construction
, we have @xmath1757 if and only if , on each @xmath869 we have @xmath1758 by lemma [ exp formula cor ] , this holds if and only if @xmath1759 for each @xmath1096 , or equivalently if @xmath1760 .
this in turn holds if and only if @xmath1761 .
therefore we see that @xmath1744 is faithful . to show that @xmath1744 is full ,
suppose we are given @xmath1676 and a morphism @xmath1003 from @xmath1762 .
define @xmath1763 writing @xmath1679 , @xmath1680 , and @xmath1764 , we have @xmath1765 , and @xmath1766 . writing @xmath1767 , we have @xmath1768 and @xmath1769 the fact that @xmath1003 is an equivalence from @xmath1693 to @xmath1694 is then equivalent to the following three conditions : a. @xmath1770 , or equivalently @xmath1771 , which implies that @xmath432 is tangent to @xmath7 and @xmath1772 .
b. @xmath1773 and c. @xmath1774 write @xmath1775 , so that @xmath1776 and therefore @xmath1777 we see from condition ( iii ) that @xmath1778 or equivalently @xmath1779 also write @xmath1780 , so that @xmath1781 defining @xmath1782 , equation ( [ int equat ] ) implies that @xmath1783 therefore , by lemma [ exp formula cor ] we have @xmath1784 on the other hand , using condition ( ii ) , we see that on each overlap @xmath867 we have @xmath1785 which implies ( again using lemma [ exp formula cor ] ) that @xmath1786 , so there is a unique function @xmath1787 such that @xmath1311 . by equation ( [ lie equation ] )
, it follows that @xmath1788 is an element of @xmath1688 , and by construction @xmath1789 this completes the proof that @xmath1744 is full .
given an open set @xmath87 , denote @xmath1790 and @xmath1791 , as well as @xmath1792 . [ equivalence of induced ]
let @xmath57 be a @xmath89 brane with locally trivializable deformations .
then for every @xmath85 , there exists a neighborhood @xmath87 of @xmath88 such that @xmath1793 is an equivalence .
in particular , this applies to lwl branes ( by proposition [ essentially surjective ] ) . given @xmath1046 with locally trivializable deformations and @xmath85 , by definition there exists a neighborhood @xmath87 of @xmath88 such that for every @xmath735 the functor @xmath1794 is essentially surjective . by proposition [ fully faithful ] , it is also fully faithful , and therefore an equivalence .
the proof of the following proposition , which we omit , is a straightforward consequence of definition [ group action on category ] and definition [ definition tilde extended ] there is a strict right action of @xmath952 on @xmath1661 defined as follows : given @xmath1628 and @xmath1175 , we define @xmath1795 to be the product @xmath1796 ( recall that @xmath1797 is a subgroup of @xmath1631 ) .
given a morphism @xmath1798 , we define @xmath1799 , now regarded as a morphism from @xmath1800 . for the following definition , recall definition [ definition action groupoid ] .
let @xmath1630 be the action groupoid @xmath1801 .
explicitly , for @xmath735 an object of @xmath1802 is an element of @xmath1631 .
given @xmath1676 , a morphism from @xmath275 to @xmath1803 is a pair @xmath1804 such that @xmath1805 .
composition is given by group multiplication , i.e. @xmath1806 it is clear that the functor @xmath1807 is compatible with the right actions of @xmath1797 on both formal groupoids .
therefore , it follows from proposition [ induced equivalence ] that @xmath1808 naturally extends to a functor @xmath1809 .
furthermore , @xmath1810 is an equivalence if and only if @xmath1808 is an equivalence .
this leads to the following corollary .
let @xmath57 be a leaf - wise lagrangian brane on a gc manifold @xmath56 ( or more generally a brane with locally trivializable deformations ) .
then for every @xmath85 , there exists a neighborhood @xmath87 of @xmath88 such that @xmath1811 is an equivalence of formal groupoids .
let @xmath0 be a gc brane on @xmath56 . as discussed in remark [ brane def sheaf ] , the formal groupoid @xmath950 extends in a natural way to a presheaf of formal groupoids on @xmath5 .
in particular , for any open cover @xmath1812 of @xmath5 , we may define the groupoid of _ descent data _ for @xmath1813 with respect to @xmath1814 .
this is described explicitly in the following definition .
[ glued deformations ] let @xmath1815 be the following formal groupoid
. for each @xmath735 , an object of @xmath1816 is pair @xmath1817 , where 1 .
each @xmath1818 is an object of @xmath1819 , 2 .
each @xmath1820 ^ -{\cong } & } } \hb\sa$ ] is an isomorphism , and 3 . on triple overlaps we have @xmath1821 .
a morphism from @xmath1817 to @xmath1822 is a collection @xmath1823 such that on each overlap we have @xmath1824 . by construction
, there is a natural restriction functor @xmath1825 explicitly , given @xmath1826 and @xmath1827 , we define @xmath1828 by setting @xmath1829 for each @xmath1571 , and @xmath1830 on each @xmath1831 . given a morphism @xmath1832 , we define @xmath1833 [ descent property]the restriction functor @xmath1834 is an equivalence of formal groupoids .
fix @xmath735 . by inspection ,
it is clear that @xmath1835 is faithful , i.e. for each pair of objects @xmath958 and @xmath1112 in @xmath1836 the induced map @xmath1837 is injective . on the other hand ,
suppose we are given a morphism @xmath1838 . by definition
, on each open set @xmath1839 we have an equivalence @xmath1840 , and on each overlap we have @xmath1841 .
it is straightforward to see from definition [ definition par def ] that this implies the existence of @xmath1842 whose restriction to each @xmath1839 is equal to @xmath1843 .
thus we see that @xmath1844 is fully faithful .
to finish the proof , we will show that @xmath1844 is also essentially surjective . given @xmath1845 ,
write @xmath1846 .
in particular , part ( 3 ) of definition [ glued deformations ] , implies that on each @xmath1847 we have @xmath1848 [ nonabelian coboundary ] there exist @xmath1849 such that on @xmath1850 we have @xmath1851 by proposition [ induct on small extensions ] , we may proceed by induction on small extensions . thus , suppose the result holds for some @xmath1584 with unique maximal ideal @xmath1852 , and let @xmath1532 be a small extension .
denote the kernel of @xmath312 by @xmath1096 .
choose a linear splitting @xmath1853 of @xmath312 .
since @xmath1854 holds in @xmath1855 , we also have @xmath1856 in @xmath1857 . by the inductive hypothesis
, we can find @xmath1858 such that @xmath1859 we must then have @xmath1860 for some @xmath1861 since @xmath1862 lies in the center of @xmath1863 , it is easy to see from equation ( [ almost ] ) that on triple overlaps @xmath1864 , so that we can choose @xmath1865 satisfying @xmath1866 .
setting @xmath1867 , it follows from equation ( [ almost ] ) that @xmath1868 returning to the proof of proposition [ descent property ] , choose @xmath1869 as in lemma [ nonabelian coboundary ] , and define @xmath1870 , @xmath1871 , and @xmath1872 note that , by lemma [ compatibility under isomorphism ] , for each @xmath1571 we have @xmath1873 . by construction
, we have equivalences @xmath1874 if we define @xmath1875 , then by construction @xmath1876 , and @xmath1877 defines an isomorphism in @xmath1878 from @xmath1879 .
furthermore , by construction we have @xmath1880 where @xmath1881 in particular , on the overlaps @xmath1850 we have @xmath1882 , so there exists a unique @xmath959 such that each @xmath1883 is given by restricting @xmath1884 . for each @xmath1571 write @xmath1885 . using proposition [ pi zero bundle ]
, we may assume without loss of generality that each @xmath1886 is of the form @xmath1887 , i.e. has undeformed transition functions @xmath1888 . therefore ,
if we write @xmath1889 , it follows from definition [ definition par def ] and definition [ herm def ] that on each @xmath1890 we have @xmath1891 , so that there is a well - defined function @xmath1892 whose restriction to each @xmath1893 is equal to @xmath1894 .
furthermore , we see that on @xmath1895 we must have @xmath1896 so we may choose @xmath1897 such that @xmath1898 .
if we define @xmath1899 , @xmath1900 and @xmath1901 then @xmath1902 defines an element of @xmath1903 , which by construction is isomorphic to @xmath981 , and hence also to @xmath974 . also by construction
, there exists a unique @xmath1904 such that @xmath1905 .
this completes the proof that @xmath1844 is essentially surjective .
we saw in proposition [ right action par ] that there is a strict right action of the formal group @xmath951 on @xmath950 .
similarly , for each open set @xmath87 , we have a strict right action of @xmath951 on @xmath1906 constructed using the restriction homomorphism @xmath1907 together with the action of @xmath1908 on @xmath1906 .
we then have the following easy result , the proof of which is omitted .
there is a strict right action of @xmath951 on @xmath1815 defined as follows : for each @xmath735 , given @xmath1154 and @xmath1909 , we define @xmath1910 similarly , given an isomorphism @xmath1911 in @xmath1912 , we define @xmath1913 let @xmath1914 be the formal groupoid @xmath1915 explicitly , for each @xmath735 @xmath1912 has the same objects as @xmath1912 .
a morphism from @xmath1879 in @xmath1916 is a pair @xmath1297 , where @xmath1175 , and @xmath1341 is a morphism in @xmath1912 from @xmath974 to @xmath981 .
composition is given by @xmath1917 .
clearly , the restriction functor @xmath1918 is compatible with the actions of @xmath952 on both formal groupoids . using proposition [ induced equivalence ]
, we may then extend @xmath1835 to a functor @xmath1919 furthermore , combining proposition [ induced equivalence ] and proposition [ descent property ] , we arrive at the following result .
[ restriction equivalence ] the restriction functor @xmath1920 is an equivalence of formal groupoids .
the remainder of the paper will be devoted to constructing a dgla governing the deformation theory of branes with locally trivializable deformations . to do so
, we will adapt techniques developed in @xcite , as well as @xcite@xcite . as a first step , in this section we review the framework of semicosimplicial groupoids and the descent groupoid construction .
we will use this to reformulate the definitions introduced in the previous section more systematically .
this will enable us to formulate and prove some results which would be much more cumbersome otherwise .
most of the following discussion including the notation is taken from @xcite .
let @xmath1921 be the category whose objects are the finite ordinal sets @xmath1922=\{0,1,\cdots , n\}$ ] for @xmath1923 , and whose morphisms are order - preserving injective maps among them . recall that a _ semicosimplicial object _ in a category @xmath755 is a functor @xmath1924 .
this may be pictured as a diagram @xmath1925\ar@<-.4ex>[r ] & a_1 \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & a_2\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \dots},\ ] ] where each @xmath1926)$ ] is an object of @xmath755 .
the _ coface maps _ @xmath1927 ( for @xmath1928 ) satisfy a number of relations determined by the combinatorial structure of @xmath1921 .
more generally , given a ( strict ) 2-category @xmath755 , we may similarly define a ( strict ) semicosimplicial object in @xmath755 : this consists of a diagram of the form ( [ simplicial diagram ] ) , where the entries are objects of @xmath755 and the arrows are 1-morphisms , which we require to satisfy cosimplicial relations on the nose .
the example we will need is when @xmath755 is the 2-category of formal groupoids
( over @xmath733 ) . [ example nerve groupoid ]
let @xmath1929 be a sheaf on a topological space @xmath5 , valued in a category ( or strict 2-category ) @xmath755 .
given an open cover @xmath1812 of @xmath5 , we may define a semicosimplicial object @xmath1930 in @xmath755 , called the _
nerve _ of @xmath1929 . for each @xmath1923
we take @xmath1931 the coface maps @xmath1932 are constructed using the restriction maps of the sheaf @xmath1929 : for each @xmath1923 and each @xmath1928 , we have @xmath1933 where for each inclusion of open subset @xmath1934 the notation @xmath1935 denotes the restriction map .
[ definition descent groupoid ] ( * ? ? ?
2.6 ) given a semicosimplicial groupoid @xmath1936 , the _ groupoid of descent data _
@xmath1937 is defined as follows : 1 . an object of @xmath1937 is a pair @xmath1938 with @xmath517 an object of @xmath1939 , and @xmath1940 a morphism in @xmath1941 from @xmath1942 , such that the equation @xmath1943 holds in @xmath1944 .
a morphism from @xmath1938 to @xmath1945 is a morphism @xmath1946 in @xmath1939 , such that the following diagram commutes : @xmath1947^{m } \ar[d]_{\partial_0a } & \partial_1l \ar[d]^{\partial_1a } \\
\partial_0l ' \ar[r]^{m ' } & \partial_1l'.}\ ] ] [ nerve groupoid 2 ] continuing example [ example nerve groupoid ] , let us describe explicitly the groupoid @xmath1937 in the case that @xmath1948 is the nerve of a strict presheaf of groupoids on @xmath5 .
an object is a pair @xmath1949 , where @xmath1950 , and the @xmath1951 are morphisms @xmath1952 ( where both objects have implicitly been restricted to @xmath1953 ) .
the morphisms must satisfy @xmath1954 on the triple intersections @xmath1955 .
a morphism from @xmath1956 to @xmath1957 is a collection @xmath1958 of morphisms @xmath1959 that satisfy @xmath1960 on overlaps @xmath1831
. definition [ definition descent groupoid ] and remark [ nerve groupoid 2 ] extend immediately to the case of formal groupoids . for example ,
in the case of the sheaf of formal groupoids @xmath68 associated to a gc brane @xmath0 , the formal groupoid in definition [ glued deformations ] with respect to an open cover @xmath1814 is precisely the descent groupoid for the nerve of @xmath68 . given a gc brane @xmath1052 , recall from [ induced deformations ] the formal groupoid @xmath1961 described in definition [ definition tilde extended ] .
this groupoid may also be refined to a sheaf of formal groupoids over @xmath5 , which assigns to each open subset @xmath87 the formal groupoid @xmath1962 .
let @xmath1963 be the groupoid formal of descent data associated to the nerve of the sheaf of groupoids @xmath1964 with respect to an open cover @xmath1814 .
explicitly , for each @xmath735 an object of @xmath1965 is a pair @xmath1966 where 1 .
@xmath1967 is an element of @xmath1968 , 2 .
@xmath1969 is an element of @xmath1970 , 3 . on each overlap
@xmath1831 we have @xmath1971 , 4 . on each triple overlap @xmath1955
we have @xmath1972 a morphism from @xmath1966 to @xmath1973 in @xmath1965 is a collection @xmath1974 , where 1 .
@xmath1975 , 2 .
on each @xmath1839 we have @xmath1976 , 3 . on each double overlap @xmath1831
we have @xmath1977 .
the composition of morphisms is given by group multiplication , i.e. @xmath1978 the identity morphisms is given by the identity group elements .
we then have the following easy result , whose proof is omitted . for each @xmath735 ,
each object @xmath1979 , and each @xmath1980 , let @xmath1981 for each morphism @xmath1982 let @xmath1983 , regarded as a morphism from @xmath1984 to @xmath1985 .
then this defines a strict right action of @xmath952 on @xmath1963 .
[ definition of extended glued ] let @xmath1986 be the formal groupoid associated to the right action of @xmath952 on @xmath1963 , as described in definition [ definition action groupoid ] .
the map taking a formal semicosimplicial groupoid to its formal groupoid of descent data is functorial in a natural sense @xcite .
namely , there is a ( strict ) 2-category of semicosimplicial formal groupoids , and the descent construction defines a 2-functor from this category to the two - category of formal groupoids .
for example , a map from a semicosimplicial formalgroupoid @xmath1948 to another one @xmath1987 by definition consists a collection of functors @xmath1988 that are compatible with the boundary maps .
such a map induces a functor @xmath1989 .
[ equivalence of descent groupoids]@xcite let @xmath1990 be a map of semicosimplicial formal groupoids , such that for each @xmath81 the functor @xmath1988 is an equivalence .
then @xmath1989 is also an equivalence .
let @xmath755 and @xmath1991 be ( strict ) sheaves of formal groupoids on @xmath5 , and let @xmath1992 be a strict map of such sheaves , i.e. a collection of functors @xmath1993 that are strictly compatible with the restriction functors .
then @xmath1994 induces a map of semicosimplicial formal groupoids from the nerve of @xmath755 to the nerve of @xmath1991 , which in turn induces a functor @xmath1995 of the associated descent groupoids .
furthermore , if for each open subset @xmath87 the functor @xmath1993 is an equivalence of groupoids , then then induced functor @xmath1995 is also an equivalence .
let @xmath1996 be the functor induced by the map of sheaves described in definition [ definition sigma tilde ]
. by construction , the functor @xmath1997 is compatible with the actions of @xmath952 on both formal groupoids .
therefore , using proposition [ induced equivalence ] , we may extend @xmath1745 to a functor from @xmath1998 to @xmath1999 .
[ definition extended nerve ] @xmath2000 is the functor @xmath2001 induced from @xmath2002 using proposition [ induced equivalence ] .
[ theorem extended and lwl ] let @xmath0 be a leaf - wise lagrangian gc brane on @xmath56 ( or more generally a brane with locally trivializable deformations ) .
then there exists an open cover @xmath1814 of @xmath5 such that @xmath2003 is an equivalence of formal groupoids .
furthermore , by theorem [ restriction equivalence ] , the formal groupoids @xmath2004 and @xmath68 are also equivalent . by corollary
[ equivalence of induced ] , we there exists an open cover @xmath1812 of @xmath5 such that , for each @xmath1571 , the functor @xmath2005 is an equivalence .
it then easily follows from proposition [ equivalence of descent groupoids ] that @xmath2006 is an equivalence as well .
the proof then follows by applying proposition [ induced equivalence ] .
for simplicity , in the following we work over the field @xmath734 . a differential graded lie algebra ( dgla ) consists of a cochain complex @xmath2007 , together with a cochain map ( the bracket " ) @xmath2008:\g\ub\otimes \g\ub\to \g\ub\ ] ] of degree zero , which is skew - symmetric ( in the graded sense ) , and satisfies a graded version of the jacobi identity @xcite .
[ lie algebra as dgla ] any lie algebra may be regarded as a dgla concentrated in degree zero .
conversely , given an arbitrary dgla @xmath2007 , the restriction of the bracket to @xmath2009 gives it the structure of a lie algebra .
for any dgla @xmath2007 , the set of _ maurer - cartan elements _ of @xmath2007 are defined as @xmath2010=0\}.\ ] ] if @xmath2007 is nilpotent , the group @xmath2011 is well - defined ( sometimes called the _ gauge group _ ) .
this group has a left action on the set of maurer - cartan elements of @xmath2007 given by the formula @xmath2012^n}{(n+1)!}([y , x]-dy).\ ] ] given an arbitrary dgla @xmath2007 ( not necessarily nilpotent ) , one may define a formal groupoid @xmath2013 over @xmath733 , known as the _ deligne groupoid_. this is defined for each @xmath735 ( with unique maximal ideal @xmath737 ) as the action groupoid @xmath2014 passing to equivalence classes , we obtain a functor @xmath2015 , which assigns to every @xmath735 the set of maurer - cartan elements of @xmath2016 modulo the action of @xmath2017 . as in example
[ lie algebra as dgla ] , let @xmath753 be a lie algebra , regarded as a dgla concentrated in degree @xmath28 .
then for each @xmath735 , the deligne groupoid @xmath2018 may be identified with the groupoid @xmath2019 by definition , the groupoid ( [ deligne for lie algebra ] ) has a unique object @xmath2020 , and the set of morphisms from @xmath2020 to itself are the elements of the group @xmath2021 , with composition given by group multiplication .
consider a functor @xmath2022 .
let @xmath2023 be a surjective map in @xmath733 , and @xmath2024 .
one says that @xmath275 can be _ extended _ to an element of @xmath2025 if there exists @xmath2026 such that @xmath2027 . in general there will be an obstruction to the existence of such an extension .
the functor @xmath457 is called _ unobstructed _ , however , if for every surjective map @xmath2028 the induced map @xmath2029 is also surjective . the fact that every surjective map in @xmath733 can be factored through a sequence of small extensions ( proposition [ induct on small extensions ] ) implies that @xmath457 is unobstructed if and only if @xmath2029 is surjective for every small extension @xmath2028 .
[ h2 and obstructions]@xcitelet @xmath753 be a dgla with @xmath2030 then the functor @xmath2031 is unobstructed .
the notation and terminology in this section ( mostly ) follows @xcite .
let @xmath2032\ar@<-.4ex>[r ] & v_1 \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & v_2\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \dots}\ ] ] be a semicosimplicial cochain complex , i.e. a semicosimplicial object in the category of cochain complexes ( differential graded vector spaces ) .
as in [ cosimplicial groupoids ] , for each @xmath2033 , we denote by @xmath2034 the differential on the cochain complex @xmath2035 , and we denote by @xmath2036 for @xmath2037 the coface maps . from @xmath2038 we may construct a cochain complex @xmath2039 , called the _ total complex _ or _ totalization _ of @xmath2038 . as a graded vector space , @xmath2039 is equal to the direct sum @xmath2040 $ ] , where for any graded vector space @xmath2041 and any integer @xmath81 , the shifted space @xmath2042 $ ] is defined by @xmath2043)^i = w^{i+k}$ ] . for each @xmath297 ,
let @xmath2044 define @xmath2045\to \bigoplus_{n>0 } v_n[-n]\ ] ] and @xmath2046\to \bigoplus_{n\geq0 } v_n[-n].\ ] ] the differential on @xmath2039 is then defined to be the sum @xmath2047 for any category @xmath755 , the collection of semicosimplicial objects in @xmath755 themselves form the objects of a category : a morphism @xmath2048 of semicosimplicial objects in @xmath755 is by definition a collection of morphisms @xmath2049 in @xmath755 that commute with the coface maps .
the totalization construction sending a semicosimplicial cochain complex @xmath2038 to its total complex @xmath2039 extends in an natural way to a functor ( from the category of semicosimplicial cochain complexes to the category of cochain complexes ) .
next , let @xmath2050 be a semicosimplicial dgla .
if we apply the above construction to the semicosimplicial cochain complex underlying @xmath2050 , there is no natural way to give the resulting cochain complex @xmath2051 a dgla structure .
there is an alternate ( functorial ) construction , however , that takes as input a semicosimplicial dgla @xmath2050 , and gives as output a dgla @xmath2052 , known as the _ thom - whitney totalization of @xmath2050_. we will not need the explicit form of this construction , only its existence and the properties summarized in propositions [ cochain iso1 ] , [ descent for deligne ] , [ bi chain iso ] , and [ double descent ] below . [ cochain iso1 ] ( * ? ? ?
* lemma 2.5 ) let @xmath2050 be a semicosimplicial dgla . then the cochain complexes @xmath2052 and @xmath2051 are quasi - isomorphic . since @xmath2053 is a dgla , we may apply the deligne construction described above to construct the formal groupoid @xmath2054 .
on the other hand , applying the deligne construction term by term to @xmath2050 produces a semicosimplicial formal groupoid @xmath2055\ar@<-.4ex>[r ] & \del_{\g_1 } \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & \del_{\g_2}\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \dots}\ ] ] we may then form its descent groupoid @xmath2056 .
[ descent for deligne ] ( * ? ? ? * thm .
2.6 ) there is a natural equivalence of formal groupoids @xmath2057 ^ -{\cong } & } } \desc(\g\ut).\ ] ] next , let @xmath2058 be a bisemicosimplicial cochain complex , which is a diagram @xmath2059\ar@<-.6ex>[u ] \ar[u ] \ar@<+.4ex>[r]\ar@<-.4ex>[r ] & v_{1,1 } \ar@<+.6ex>[u]\ar@<-.6ex>[u ] \ar[u ] \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & v_{2,1 } \ar@<+.6ex>[u]\ar@<-.6ex>[u ] \ar[u]\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \cdots \\ v_{0,0 } \ar@<+.4ex>[u]\ar@<-.4ex>[u ] \ar@<+.4ex>[r]\ar@<-.4ex>[r ] & v_{1,0 } \ar@<+.4ex>[u]\ar@<-.4ex>[u]\ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & v_{2,0 } \ar@<+.4ex>[u]\ar@<-.4ex>[u]\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \cdots}\ ] ] we may succinctly define @xmath2058 as a functor from the category @xmath2060 to the category of cochain complexes .
each individual row @xmath2061 in the above diagram itself forms a semicosimplicial cochain complex , and we alternatively view @xmath2058 as a semicomisimplical object in the category of semicosimplicial cochain complexes : @xmath2062\ar@<-.4ex>[r ] & v_{\bullet,1 } \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & v_{\bullet,2 } \ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \dots}\ ] ] from this point of view , it is clear how to generalize the totalization functor to the bisemicosimplicial case : first , form the total complex of each row , yielding a semicosimplicial cochain complex @xmath2063\ar@<-.4ex>[r ] & \tot(v_{\bullet},1 ) \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & \tot(v_{\bullet},2 ) \ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \dots}\ ] ] we may then applying the totalization procedure again to form the cochain complex @xmath2064 , which we will simply denote by @xmath2065 .
similarly , if @xmath2066 is a bisemicosimplicial dgla , we may iterate the thom - whitney totalization procedure to construct a dgla @xmath2067 [ bi chain iso ] for any bisemicosimplicial dgla @xmath2066 , the cochain complexes @xmath2068 and @xmath2069 are quasi - isomorphic .
one may similarly define a strict bisemicosimplicial formal groupoid @xmath2070 to be a diagram of the form [ square diagram ] , where the entries are groupoids and coface maps are required to satsify the same compatibility conditions ( on the nose ) . to define the descent groupoid of @xmath2070
, we first form the descent groupoid @xmath2071 of each row , resulting in a semicosimplicial groupoid @xmath2072 .
we may then form its formal groupoid of descent data @xmath2073 , which we denote simply by @xmath2074 .
[ double descent ] let @xmath2066 be a bisemicosimplicial dgla , and let @xmath2075 be the bisemicosimplicial formal groupoid formed by applying the deligne construction term by term to @xmath2066
. then there is a natural equivalence of formal groupoids between @xmath2076 and @xmath2077 .
we now apply the theory described in the previous section to the deformation theory of generalized complex branes .
specifically , for every brane @xmath1052 and open cover @xmath1814 of @xmath5 , we will construct a dgla @xmath2078 . using results from the previous section
, we then prove that the formal groupoid @xmath2079 is equivalent to the formal @xmath2080 introduced in definition [ definition extended nerve ] . for @xmath0 a lwl brane
, it follows from theorem [ theorem extended and lwl ] that @xmath2079 is also equivalent to @xmath68 . in particular
, there is a natural isomorphism of functors @xmath2081 ^ -{\cong } & } } \df_{\b}.\ ] ] as a concrete application of this construction we prove theorem [ brane obstructions ] , stated in the introduction .
this generalizes the well - known result in complex geometry that the obstructions to deforming a complex submanifold @xmath7 of a complex manifold @xmath5 are contained in the sheaf cohomology group @xmath2082 .
let @xmath0 be a gc brane on a gc manifold @xmath56 , and let @xmath1812 be an open cover of @xmath5 .
consider the following diagram of lie algebras @xmath2083\ar@<-.4ex>[r ] & \prod\t(u_{\alpha\beta } ) \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & \prod\t(u_{\alpha\beta\gamma } ) \ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \cdots \\
\h(x)\oplus\prod\kkk(u\sa ) \ar@<+.4ex>[u]\ar@<-.4ex>[u ] \ar@<+.4ex>[r]\ar@<-.4ex>[r ] & \h(x)\oplus\prod\kkk(u\sab ) \ar@<+.4ex>[u]\ar@<-.4ex>[u]\ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & \h(x)\oplus\prod\kkk(u\sabc ) \ar@<+.4ex>[u]\ar@<-.4ex>[u]\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \cdots}\ ] ] the maps in the diagram are defined as follows : the top row is the semicosimplicial dgla which is the nerve ( with respect to the open cover @xmath1814 ) of the sheaf @xmath2084 of dglas on @xmath5 , as described in example [ example nerve groupoid ] .
the bottom row is the direct sum of the nerve ( with respect to @xmath1814 ) of the sheaf @xmath2085 , and the semicosimplicial dgla @xmath2086\ar@<-.4ex>[r ] & \h(x ) \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & \h(x)\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \dots}\ ] ] with all maps equal to the identity .
next we describe the vertical maps , which we denote by @xmath2087 for @xmath2088 .
given @xmath2089 , we have @xmath2090 and @xmath2091 it is easy to see that the diagram ( [ v diagram ] ) , extended upwards by zero , defines a bisemicosimplicial dgla , which we denote by @xmath2092 .
[ concrete description ] the formal groupoid @xmath2093 is equivalent to @xmath2094 for each @xmath735 , the bisemicosimplicial groupoid @xmath2095 may be identified with @xmath2096\ar@<-.4ex>[r ] & { * } //\prod e^{\t\da(u_{\alpha\beta } ) } \ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & { * } //\prod e^{\t\da(u_{\alpha\beta\gamma } ) } \ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \cdots \\ { * } //e^{\h\da(x)}\times\prod
e^{\kkk\da(u\sa ) } \ar@<+.4ex>[u]\ar@<-.4ex>[u ] \ar@<+.4ex>[r]\ar@<-.4ex>[r ] & { * } //e^{\h\da(x)}\times\prod
e^{\kkk\da(u\sab ) } \ar@<+.4ex>[u]\ar@<-.4ex>[u]\ar@<+.6ex>[r]\ar@<-.6ex>[r ] \ar[r ] & { * } //e^{\h\da(x)}\times\prod
e^{\kkk\da(u\sabc ) } \ar@<+.4ex>[u]\ar@<-.4ex>[u]\ar@<+1ex>[r]\ar@<-1ex>[r]\ar@<+.3ex>[r]\ar@<-.3ex>[r ] & \cdots}\ ] ] let us label these groupoids as @xmath2097 , where @xmath2098 is the row index , and @xmath2099 is the column index . consider first the descent groupoid @xmath2100 of the bottom row .
recall from definition [ definition descent groupoid ] , that an object of @xmath755 is a pair @xmath1938 where @xmath517 is an object of @xmath2101 and @xmath1940 is a morphism in @xmath2102 from @xmath2103 to @xmath2104 , which satisfies the associativity condition @xmath2105 we must have @xmath2106 ( since this is the only object in @xmath2101 ) , and @xmath1940 must be an element @xmath2107 ; the associativity condition ( [ assoc1 ] ) is equivalent to @xmath2108 so we see that @xmath2109 must be trivial , and the collection @xmath2110 must satisfy the nonabelian cocycle condition @xmath2111 . to simplify the notation , we will denote the object @xmath2112 of @xmath2113 by @xmath2110 . spelling out part ( 2 ) of definition [ definition descent groupoid ]
, we see that a morphism in @xmath2114 from @xmath2110 to @xmath2115 consists of a pair @xmath2116 , where @xmath88 is an arbitrary element of @xmath1797 , and @xmath2117 are subject to the condition @xmath2118 . similarly , an object of @xmath2119 consists of @xmath2120 satisfying @xmath2121 .
a morphism in @xmath2122 from @xmath2123 to @xmath2124 consists of @xmath2125 satisfying @xmath2126 . to complete the construction of @xmath2127 , we then then apply the descent groupoid construction in the vertical direction : @xmath2128\ar@<-.4ex>[r ] & \cc_1}).\ ] ] here , @xmath2129 is the functor taking and object @xmath2110 in @xmath2113 to the trivial object @xmath2130 , and a morphism @xmath2131 to the morphism @xmath2132 from the trivial object to itself .
we also see that @xmath2133 takes @xmath2110 to @xmath2134 , and morphism @xmath2135 to the morphism @xmath2136 . therefore , an object of @xmath2127 is a pair @xmath2137 with @xmath2138 satisfying @xmath2139 , and @xmath2140 satisfying @xmath2141 .
a morphism in @xmath2127 from @xmath2137 to @xmath2142 is a pair @xmath2116 with @xmath2143 , @xmath2144 such that @xmath2118 and @xmath2145 .
comparing this with the definition [ definition of extended glued ] of @xmath2146 , we arrive at the desired result
. introduce the notation @xmath2147 . combining proposition [ concrete description ] and proposition [ double descent ]
, we have the following result .
[ dgla for extended ] there is an equivalence of formal groupoids @xmath2148 finally , corollary [ dgla for extended ] with theorem [ theorem extended and lwl ] we arrive at the following result . [ dgla for lwl ] let @xmath0 be a leaf - wise lagrangian gc brane on @xmath56 ( or more generally a brane with locally trivializable deformations ) .
then there exists an open cover @xmath1814 of @xmath5 such that @xmath68 is equivalent to @xmath2079 .
in particular , the deformation functor @xmath2149 is isomorphic to the functor @xmath92 associated to @xmath0 as described in definition [ brane deformation functor ] .
let @xmath1052 be a lwl brane ( or more generally , a brane with locally trivializable deformations ) , and let fix an open cover @xmath1814 of @xmath5 , chosen as in theorem [ dgla for lwl ] .
let @xmath2150 be the bisemicosimplicial dgla constructed in
[ construction of the dgla ] , and @xmath2151 the dgla @xmath2152 .
we may also form the cochain complex @xmath2153 , which by proposition [ bi chain iso ] is quasi - isomorphic to ( the underlying cochain complex of ) @xmath238 ; in particular the cohomology groups of @xmath2154 and @xmath238 are isomorphic .
as state in [ dgla obstructions ] , the deformation functor @xmath2155 is unobstructed if the cohomology group @xmath2156 vanishes . by theorem [ dgla for lwl ] , the condition @xmath2157 also implies that the functor @xmath92 is unobstructed .
the following theorem makes use of this fact by relating the group @xmath2158 to the more familiar lie algebroid cohomology group @xmath91 associated to @xmath0 , which were described in [ lie alg cohomology ] .
there is an injective linear map @xmath2159 in particular , we obtain theorem [ brane obstructions ] as a corollary . using the explicit construction of the total complex @xmath2160 given in
[ bisemi ] , we see that @xmath2161 @xmath2162 and @xmath2163 the differential @xmath2164 from @xmath2165 to @xmath2166 is given by @xmath2167 and from @xmath2168 to @xmath2169 by @xmath2170 with @xmath2171 we wish to construct a map @xmath2172 .
let @xmath2173 be a cocycle representing a class @xmath2174\in h^2(c\ub)$ ] . from ( [ d1 ] )
, we see that every element of @xmath2168 is cohomologous to one with @xmath573 , so without loss of generality we may assume @xmath2175 is of this form .
write @xmath2176 , with @xmath2177 and @xmath2178 .
we claim that @xmath2179 is cohomologous to a cocycle of the form @xmath2180 . to see this , note using ( [ d1 ] ) that that the condition @xmath2181 implies that @xmath2182 therefore , using a partition of unity we may find functions @xmath2183 such that @xmath2184 .
let @xmath2185 be a choice of functions extending @xmath2186 .
note that @xmath2187 is the generalized hamiltonian vector field associated to the complex - function @xmath2188 , so in particular @xmath2189 .
furthermore , by construction @xmath2190 is an element of @xmath2191 .
writing @xmath2192 using ( [ d1 ] ) we see that @xmath2193 .
therefore , without loss of generality we may assume that @xmath2175 is of the form @xmath2194 by definition [ definition kkk ] , we see that @xmath2195 , so that @xmath2196 therefore , if we define @xmath2197 , then using ( [ d2 ] ) we see that @xmath2198 choose @xmath2199 such that @xmath2200 . by proposition [ induced holomorphic ] ,
the fact @xmath2201 implies that @xmath2202 .
it follows there exists a unique section @xmath2203 such that on each open set @xmath2204 we have @xmath2205 . by construction
we have @xmath2206 , and we define @xmath2207)=[\zeta]\in h^2(l)$ ] . to show that @xmath1341 is well - defined
, we must check that the class @xmath2208 $ ] does not depend on the choice of @xmath1869 or the representative @xmath2175 for the class @xmath2174 $ ] .
first , suppose that we have another collection @xmath2209 that also satisfy @xmath2210 , and that @xmath2211 such that the restriction of @xmath1589 to each @xmath2212 is equal to @xmath2213 .
then there is a unique section @xmath2214 such that @xmath2215 .
we then have @xmath2216 , which implies that @xmath2217=[\zeta]\in h^2(\b)$ ] .
this verifies that @xmath2208 $ ] does not depend on the choice of @xmath1869 . to see that @xmath2208 $ ] is also independent of the choice of @xmath2175 ,
suppose that @xmath2218 is different cocycle such that @xmath2219=[c]$ ] in @xmath2220 .
given @xmath2221 such that @xmath2222 , by a nearly identical argument to that used above , we may assume that @xmath2223 is of the form @xmath2224 .
we have @xmath2225 since @xmath2226 , it follows that @xmath2227 given @xmath2228 satisfying @xmath2229 , define @xmath2230 .
it follows from ( [ diff1 ] ) that @xmath2231 .
furthermore , since @xmath2232 is a generalized holomorphic vector field , it follows that @xmath2233 , so we see that @xmath2234 .
this finishes the demonstration that @xmath1341 is well - defined . to show that @xmath1341 is injective ,
suppose that @xmath2207)=[\zeta]=0 $ ] in @xmath91 , say @xmath2235 for some @xmath2236 , where @xmath2237 and @xmath1869 are as above .
then on each @xmath2238 we have @xmath2239 it is not hard to show ( using , for example , the arguments in [ lwl branes example ] in the special case @xmath2240/(\epsilon^2 $ ] ) ) that there exists @xmath2241 with @xmath2242 .
therefore we also have @xmath2243 , which implies that @xmath2244 . setting @xmath2245 using the formula ( [ d1 ] )
we see that @xmath2246 .
we therefore conclude that @xmath2174=0 $ ] in @xmath2220 .
_ proof of proposition [ action ] _ : consider the following general situation : @xmath753 is a nilpotent lie algebra acting on vector spaces @xmath2247 , we have a bilinear map @xmath2248 sending @xmath2249 and for each @xmath2250 we have @xmath2251 note that @xmath2253 and more generally @xmath2254 therefore @xmath2255)&= \sum_{k=0}^{\infty}\frac{1}{k!}\sum_{j=0}^k\frac{k!}{j!(k - j)!}\xi^j(x)*\xi^{k - j}(y ) \nonumber \\ & = \sum_{j , l=0}^{\infty } \frac{1}{j!l!}\xi^j(x)*\xi^l(y ) \nonumber \\ & = \sum_{j=0}^{\infty}\frac{1}{j!}\xi^j(x)*\sum_{l=0}^{\infty}\frac{1}{l!}\xi^l(y ) \nonumber \\ & = e^{\xi}(x)*e^{\xi}(y).\end{aligned}\ ] ] 1 . in the case that @xmath2257 , the action of @xmath801 is by the lie bracket , and @xmath2020 the lie bracket .
2 . in the case that @xmath2258 , the action of @xmath801 is by the lie derivative , and @xmath2020 is the wedge - product .
3 . in the case
where @xmath2259 with @xmath801 acting by the lie bracket , @xmath2260 with @xmath801 acting by the lie derivative , and @xmath2020 is the contraction operation of vector fields with differential forms . to see that @xmath1492 is compatible with the exterior derivative , note that ( using the cartan formula for the lie derivative ) , we have @xmath2261 iterating , we see that for each @xmath2262 we have @xmath2263 , and therefore @xmath2264 m. abouzaid , m. boyarchenko , _ local structure of generalized complex manifolds _ , j. symplectic geom .
43 - 62 , 2006 .
r. bandiera , m. manetti , _ on coisotropic deformations of holomorphic submanifolds _ , arxiv:1301.6000 [ math.ag ] k. fukaya , _ floer homology and mirror symmetry i _ , available at [ https://www.math.kyoto-u.ac.jp/ fukaya / mirror1.pdf ] k. fukaya , _ floer homology for families - a report of a project in progress _ , available at [ https://www.math.kyoto-u.ac.jp/ fukaya / familyy.pdf ] d. fiorenza , m. manetti , e. martinengo , _ simicosomplicial dglas in deformation theory _ ,
communications in algebra vol .
40 , 2012 , 2243 - 2260 m. gualtieri , _ generalized complex geometry , _ oxford university dphil thesis , [ arxiv : math/0401221 ] m. gualtieri , _ generalized complex geometry , _ princeton annals of mathematics , vol- ume 174 , issue 1 ( 2011 ) , 75 - 123 . m. gualtieri _
branes on poisson varieties _ ,
arxiv:0710.2719 k. kodaira , _ a theorem of completeness of characteristic systems for analytic families of compact submanifolds of complex manifolds _ , annals of mathematics , second series , vol .
75 , no1 ( jan . 1962 ) , pp .
146 - 162 .
p. koerber , l. martucci , _ deformations of calibrated d - branes in flux generalized complex manifolds _ , journal of high energy physics 12 ( 2006 ) 062 [ hep - th/0610044 ] | we investigate the formal deformation theory of ( rank 1 ) branes on generalized complex ( gc ) manifolds .
this generalizes , for example , the deformation theory of a complex submanifold in a fixed complex manifold . for each gc brane @xmath0 on a gc manifold @xmath1
, we construct a formal ( pointed ) groupoid @xmath2 ( defined over a certain category of real artin algebras ) that encodes the formal deformations of @xmath3 .
we study the geometric content of this groupoid in a number of different situations . using the theory of ( bi)semicosimplicial differential graded lie algebras ( dglas )
, we construct for each brane @xmath3 a dgla @xmath4 that governs the locally trivializable " deformations of @xmath3 . as a concrete application of this construction ,
we prove an unobstructedness result . |
mechanical models for tumor growth are used extensively in recent years for the prediction of cancer evolution based on imaging analysis .
such models are based on the assumption that the growth of the tumor is mainly limited by the competition for space .
mathematical modeling , analysis and numerical simulations together with experimental and clinical observations are essential components in the effort to enhance our understanding of the cancer development . the goal of this article is to make a further step in the investigation of such models by presenting a convergent explicit finite difference scheme for the numerical approximation of a hele - shaw - type model for tumor growth and by providing its detailed mathematical analysis .
even though the main focus in the present work is on the investigation of the evolution of the proliferating cells , it provides a mathematical framework that can potentially accommodate more complex systems that account for the presence of nutrient and drug application
. this will be the subject of future investigation @xcite . in the present context
the tissue is considered as a multi - phase fluid and the ability of the tumor to expand into a host tissue is then primarily driven by the cell division rate which depends on the local cell density and the mechanical pressure in the tumor .
the dynamics of the cell population density @xmath0 under pressure forces and cell multiplication is described by a transport equation @xmath1 where @xmath2 represents the number density of tumor cells , @xmath3 the velocity field and @xmath4 the pressure of the _
tumor_. @xmath5 is a bounded domain in @xmath6 , @xmath7 .
the pressure law is given by @xmath8 where @xmath9 .
following @xcite , we assume that growth is directly related to the pressure through a function @xmath10 which satisfies @xmath11 the pressure @xmath12 is usually called _ homeostatic pressure . _ here , and in what follows , for simplicity we let @xmath13 for some @xmath14 . the continuous motion of cells within the tumor region , typically due to proliferation , is represented by the velocity field @xmath15 given by an alternative to darcy s equation known as _ brinkman s equation _ @xmath16 @xmath17 where @xmath18 is a positive constant describing the viscous like properties of tumor cells and @xmath4 is the pressure given by .
relation consists of two terms .
the first term is the usual darcy s law , which in the present setting describes the tendency of cells to move down pressure gradients and results from the friction of the tumor cells with the extracellular matrix . the second term ,
on the other hand , is a dissipative force density ( analogous to the laplacian term that appears in the navier - stokes equation ) and results from the internal cell friction due to cell volume changes .
a second interpretation of relation is the tumor tissue can be viewed as fluid like . " in other words , the tumor cells flow through the fixed extracellular matrix like a flow through a porous medium , obeying brinkman s law .
( accounting for the friction among the tumor cells themselves . ) and is analogous to the laplacian term that appears in the navier - stokes equation .
darcy s law describes the tendency of cells to move down pressure gradients accounting for the friction of the tumor cells with the extracellular matrix . on the other hand , the dissipative force density in brinkman s law results from the internal cell friction due to cell volume changes representing the friction among the tumor cells themselves.the laplace term accounts for the friction among the tumor cells themselves . the resulting model , governed by the transport equation for the population density of cells , the elliptic equation for the velocity field and a state equation for the pressure law , now reads at a first look , appears as an over damped force balance .
a second interpretation of this relation states that the tumor tissue is fluid like " and that the tumor cells flow through the fixed extracellular matrix like a flow through a porous medium , obeying brinkman s law .
@xmath19 we complete the system with a family of initial data @xmath20 satisfying ( for some constant @xmath21 ) @xmath22 the objective of this work is to establish the global existence of weak solutions to the nonlinear model for tumor growth by designing an efficient numerical scheme for its approximation and by showing that this scheme converges when the mesh is refined .
the main ingredients of our approach and contribution to the existing theory include : 1 . the introduction of a suitable notion of solutions to the nonlinear system consisting of the transport equation and the brinkman regularization .
2 . the construction of an approximating procedure which relies on an artificial vanishing viscosity approximation and the establishment of the suitable compactness in order to pass into the limit and to conclude convergence to the original system ( cf .
section [ s3 ] , lemma 3.7 ) .
3 . the design of an efficient numerical scheme for the numerical approximation of the nonlinear system - .
the proof of the convergence of the numerical scheme . in the center of the analysis lies the proof of the strong convergence of the cell densities .
this is achieved by establishing the weak continuity of the _ effective viscous pressure _ in the spirit of lions @xcite ( cf . section [ s4 ] , lemma 4.8 ) .
the design of numerical experiments in order to establish that the finite difference scheme is effective in computing approximate solutions to the nonlinear system ( cf .
section 5 ) . for relevant results on the analysis and the numerical approximation of a two - phase flow model in porous media
we refer the reader to @xcite .
related results on the numerical approximation of compressible fluids employing the weak compactness tools developed by of lions @xcite in the discrete setting have been established by karper _
_ @xcite and gallout _ et al .
_ @xcite .
relevant work on the mathematical analysis of mechanical models of hele - shaw - type have been presented by perthame _
the analysis in @xcite establishes the existence of traveling wave solutions of the hele - shaw model of tumor growth with nutrient and presents numerical observations in two space dimensions .
the present article is according to our knowledge the first article presenting rigorous analytical results on the global existence of general weak solutions to hele - shaw - type systems .
a different approach yielding results on the global existence of weak solutions to a nonlinear model for tumor growth in a general moving domain @xmath23 without any symmetry assumption and for finite large initial data is presented in @xcite .
the approach introduced in @xcite relies on the _ penalization _ of the boundary behavior , diffusion and viscosity .
the work by donatelli and trivisa in @xcite establishes the global existence of weak solutions to a nonlinear system modeling tumor growth in a general moving domain @xmath23 without any symmetry assumption and for finite large initial data .
but in contrast to the present nonlinear system , the transport equation for the evolution of cancerous cells in @xcite has a source term which is linear with respect to cell density .
relevant results on nonlinear models for tumor growth governed by the darcy s law for the evolution of the velocity field are presented by zhao @xcite based on the farmework introduced by friedman _
_ @xcite .
the objective of the present work is to establish the global existence of weak solutions to the nonlinear model for tumor growth - by designing an efficient numerical scheme of its approximation and by showing that this scheme converges when the mesh is refined .
the main ingredients of our approach include : 1 . the introduction of a suitable notion of solutions to the system consisting of the transport equation and the brinkman regularization .
2 . the design of an efficient numerical scheme for the numerical approximation of the nonlinear system - with the aid of a finite difference scheme .
the proof of the convergence of the numerical scheme . in the heart of the analysis lies the establishment of the strong convergence of the cell densities by establishing the weak continuity of the _ effective viscous pressure_. 4 . the design of numerical experiments in order to establish that the finite difference scheme is effective in computing approximate solutions to the nonlinear system - .
the paper is organized as follows : section [ s1 ] presents the motivation , modeling and introduces the necessary preliminary material .
section [ s2 ] provides a weak formulation of the problem and states the main result .
section [ s3 ] is devoted to the global existence of solutions via a vanishing viscosity approximation . in section [ s4 ]
we present an efficient finite difference scheme for the approximation of the weak solution to system on rectangular domains and section [ s5 ] is devoted to numerical experiments . a discretized aubin - lions lemma and some technical lemmas are presented in appendices a and b respectively .
for @xmath24 , @xmath25 , we will denote by @xmath26 and @xmath27 the gradient and divergence in the spatial direction in @xmath5 .
[ d2.1 ] let @xmath5 a bounded domain in @xmath6 , @xmath7 , which is either rectangular or has a smooth boundary @xmath28 and @xmath29 a finite time horizon .
we say that @xmath30 is a weak solution of problem - supplemented with initial data @xmath31 satisfying ( [ ic ] ) provided that the following hold : @xmath32 @xmath33 represents a weak solution of - on @xmath34 , i.e. , for any test function @xmath35\times{\mathbb{r}}^d ) , t>0 $ ] , the following integral relations hold @xmath36 in particular , @xmath37 we remark that in the weak formulation , it is convenient that the equations hold in the whole space @xmath38 provided that the densities @xmath39 are extended to be zero outside the tumor domain .
@xmath32 brinkman s equation holds in the sense of distributions , i.e. , for any test function @xmath40 satisfying @xmath41,\ ] ] the following integral relation holds for a.e .
@xmath42 $ ] , @xmath43 and @xmath44 almost everywhere .
all quantities in are required to be integrable , and in particular , @xmath45;h^{2}({\omega})).$ ] the main result of the article now follows .
[ t2.2 ] let @xmath46 be a bounded domain with smooth boundary @xmath28 , @xmath47 .
assume that the initial data @xmath48 with @xmath49 and that @xmath50 is of the form .
then the problem - , admits a weak solution in the sense specified in definition [ d2.1 ] .
the following two remarks are now in order . in section [ s3 ] ,
such a solution is obtained as the limit of the vanishing viscosity approximations @xmath51 of to as @xmath52 . in section [ s4 ] ,
such a solution is obtained in the case of a rectangular domain , as the limit of the sequence of approximations @xmath53 computed by the numerical scheme as @xmath54 .
in this section we prove theorem [ t2.2 ] by constructing an approximating scheme which relies on the addition of an artificial vanishing viscosity approximation @xmath55 where @xmath56 is a smoothnened version of @xmath57 , that is @xmath58 for a smooth function @xmath59 with compact support , and a bounded domain @xmath60 with smooth boundary or alternatively the @xmath61-dimensional torus @xmath62 , and we establish its convergence to the nonlinear system at the continuous level . for simplicity , we assume @xmath63 and homogeneous neumann boundary conditions for @xmath64 and @xmath65 ( if the domain is a torus @xmath62 we can also use periodic boundary conditions ) . for every @xmath66 , the parabolic - elliptic system admits a unique smooth solution @xmath67 .
the proof of this result relies on classical arguments ( cf .
ladyzhenskaya @xcite ) , namely by employing the contraction mapping principle and the regularity of the initial data one can show the existence of a unique solution @xmath67 defined for a small time @xmath68 then one derives apriori estimates establishing that the solution does not blow up and in fact is defined for every time . finally , a bootstrap argument yields the smoothness of the solution .
the remaining part of this section aims to establish the necessary compactness of the approximate sequence of solutions @xmath69 we start by proving that @xmath64 are uniformly bounded independent of @xmath70 and nonnegative : [ lem : linfne ] if @xmath71 uniformly in @xmath70 , then for any @xmath72 , the functions @xmath73 are uniformly ( in @xmath70 ) bounded and nonnegative , specifically , @xmath74 first we notice that if @xmath65 has a maximum at a point @xmath75 , then @xmath76 and therefore @xmath77 .
similarly , if it has a minimum at a point @xmath75 , it will satisfy @xmath78 and therefore @xmath79 .
if @xmath65 attains a strict maximum on the boundary , i.e. , there is a point @xmath80 such that @xmath81 for any other @xmath82 , we apply hopf s lemma , e.g. @xcite , to the function @xmath83 which satisfies @xmath84 which has a strict maximum at the point @xmath75 . if @xmath85 , then @xmath86 and otherwise hopf lemma gives @xmath87 where we have denoted the boundary normal @xmath88 , this contradicts the homogeneous boundary conditions . in a similar way
we show that @xmath89 ( applying hopf s lemma to @xmath90 and hence @xmath91 we rewrite the evolution equation for @xmath64 using the equation for the potential @xmath65 , @xmath92 now assume @xmath93 is a point , where @xmath94 reaches its maximum ( and therefore also @xmath95 reaches a maximum )
. then @xmath96 and @xmath97
. hence @xmath98 by , the second term on the right hand side is nonpositive and since @xmath99 for @xmath100 , we get @xmath101 hence @xmath64 will decrease and if initially @xmath102 , this implies that @xmath103 for any later time @xmath104 . to show the nonnegativity of @xmath64 , we integrate the evolution equation for @xmath64 , @xmath105 on the other hand , multiplying the same equation by a regularized version of the sign function , integrating and then passing to the limit in the approximation , we have @xmath106 subtracting the two equations from one another , and using that @xmath107 , @xmath108}|{{{\bm{g}}}}(s)|\int_{{\omega}}\big||{n}_{\varepsilon}|-{n}_{\varepsilon}\big| dx . \end{split}\ ] ] now using grnwall s inequality and that @xmath109 by assumption , we obtain @xmath110 and thus that @xmath111 almost everywhere . next we prove a simple lemma on the regularity of @xmath65 .
[ lem : we ] we have that @xmath112;w^{2,q}({\omega})),\ ] ] for any @xmath113 uniformly in @xmath70 and @xmath114 uniformly in @xmath70 as well .
we square the equation for @xmath65 and integrate it over the spatial domain and then use integration by parts , @xmath115 by the previous lemma [ lem : linfne ] , we have that @xmath116 is uniformly bounded in @xmath70 and therefore that the left hand side of the above equation is bounded and that @xmath117;h^2({\omega}))$ ] . using a calderon - zygmund inequality ( e.g. ( * ? ? ?
* thm . 9.11 . ) ) , we obtain @xmath118 ; w^{2,q}({\omega}))$ ] for all @xmath119 . by the sobolev embedding theorem
, this implies that in particular @xmath120 .
the second claim follows from and the uniform bound on the pressure proved in lemma [ lem : linfne ] .
to prove strong convergence of the approximating sequence @xmath121 , it will be useful to derive entropy inequalities for @xmath64 .
to this end , the following lemma will be useful : [ lem : fn ] let @xmath122 be a smooth convex , nonnegative function and denote @xmath123 . then @xmath124 satisfies the following identity @xmath125 where @xmath126 with @xmath127 a constant independent of @xmath70 .
in particular , this implies that @xmath128 with @xmath129\times { \omega})$ ] and @xmath130 ; w^{-1,2}({\omega}))$ ] .
the identity follows after multiplying the evolution equation for @xmath64 , , by @xmath131 and using chain rule . integrating the inequality in space and time , we obtain @xmath132 the right hand side is bounded by the assumptions on the initial data and the @xmath133-bounds proved in lemmas [ lem : linfne ] and [ lem : we ] .
this implies .
therefore the right hand side of is contained in @xmath134 . using for the third term on the left hand side
, we conclude that it is contained in @xmath135 ; h^{-1}({\omega}))$ ] . the second term on the left hand side is contained in @xmath136;w^{-1,2}({\omega}))$ ] .
hence @xmath128 with @xmath129\times { \omega})$ ] and @xmath130 ; w^{-1,2}({\omega}))$ ] and in particular , @xmath137;w^{-1,1^*}({\omega}))$ ] by the sobolev embedding ( @xmath138 ) .
the preceeding lemma implies that the time derivative of the approximation of the pressure @xmath139 where @xmath140 is uniformly bounded in @xmath135\times { \omega})$ ] and @xmath141 in @xmath135 ; h^{-1}({\omega}))$ ] .
hence @xmath142 where @xmath143;h^1({\omega}))$ ] solves @xmath144 and @xmath145;w^{1,r}({\omega}))$ ] , @xmath146 solves @xmath147 ( see ( * ? ? ?
6.1 ) for a proof of the second statement ) .
hence @xmath148;w^{1,r}({\omega}))$ ] for any @xmath146 .
the estimates of the previous ( sub)sections allow us to pass to the limit @xmath52 in a subsequence , still denoted @xmath149 , and conclude the existence of limit functions @xmath150\times { \omega } ) , \
, 1\leq q<\infty,\\ p_{\varepsilon}{\rightharpoonup}{{\overline{p}}}\geq 0,&\quad \mathrm{in}\ , l^q([0,t]\times { \omega } ) , \ , 1\leq q<\infty,\end{aligned}\ ] ] where @xmath151 and @xmath152\times { \omega})$ ] .
using aubin - lions lemma for @xmath65 and @xmath153 , we obtain strong convergence of a subsequence in @xmath154\times{\omega})$ ] for any @xmath155 to limit functions @xmath156\times { \omega})$ ] .
moreover , from the estimates in lemma [ lem : we ] we obtain that @xmath45\times{\omega})\cap l^{\infty}([0,t];w^{2,q}({\omega}))$ ] .
hence we have that @xmath157 satisfy for any @xmath158 , @xmath159 where @xmath160 is the weak limit of @xmath161 . to conclude that the limit @xmath162 is a weak solution of
, we need to show that @xmath64 converges strongly and therefore in the limit @xmath163 and @xmath164 . for this purpose
, we combine a compensated compactness property ( lemma [ lem : effve ] ) with a monotonicity argument .
we will also make use of the following lemma which was proved in a more general version in @xcite : [ lem : renormalized ] let @xmath165\times{\omega})$ ] and @xmath166;h^1({\omega}))$ ] with @xmath167\times{\omega})$ ] satisfy @xmath168 in the sense of distributions .
then for all continuously differentiable functions @xmath169 , @xmath170{\operatorname{div}}{{\bm{u}}},\ ] ] in the sense of distributions .
we let @xmath171 be a smooth , radially symmetric mollifier , i.e. @xmath172 and @xmath173 , with @xmath174 and denote for @xmath175 , @xmath176 .
then we choose as a test function in @xmath177 , with @xmath178 is compactly supported in @xmath179 where @xmath180 includes all the points @xmath181 in @xmath5 which have distance @xmath182 and do a change of variables : @xmath183 integrating in @xmath184 , this becomes @xmath185 we define @xmath186 and @xmath187 and choose as a test function @xmath188 for a smooth @xmath189 compactly supported in @xmath179 ( which is possible since @xmath190 is smooth and bounded thanks to the convolution . ) .
then we can rewrite the last identity using chain rule as @xmath191{\operatorname{div}}{{\bm{u } } } + b'({n}_\delta ) r_\delta \right)\phi \
, dxdt}.\end{gathered}\ ] ] where @xmath192 . by (
* lemma 2.3 ) , we have that @xmath193 in @xmath194 and thanks to the properties of the convolution that @xmath195 almost everywhere as well as @xmath196 a.e . when @xmath197 .
thus we obtain that in the limit @xmath197 , @xmath2 satisfies @xmath198{\operatorname{div}}{{\bm{u } } } \right)\phi \ , dxdt}.\ ] ] which is exactly in the sense of distributions . applying lemma [ lem : renormalized ] for the weak limit @xmath2 in with @xmath199
, we obtain that @xmath2 satisfies @xmath200 for any test functions @xmath201 . on the other hand , from for @xmath199 we obtain after integrating in space and time @xmath202 passing to the limit @xmath52 in this inequality , we have @xmath203 where @xmath204 denotes the weak limit of @xmath205 and @xmath206 and @xmath207 are the weak limits of @xmath208 and @xmath209 respectively .
letting @xmath210 in this inequality , we obtain , thanks to the boundedness of the integrand on the right hand side , @xmath211 on the other hand , since @xmath199 is convex , we have @xmath212 and hence @xmath213 .
we now choose smooth test functions @xmath214 approximating @xmath215}(t)$ ] , where @xmath216 $ ] , in inequality and then pass to the limit in the approximation to obtain the inequality @xmath217 subtracting from , we have @xmath218 now using the explicit expression of @xmath219 , , the first term on the right hand side can be estimated as follows : @xmath220 where we have used ( * ? ? ?
* lemma 3.35 ) , which implies @xmath221 , for the first inequality .
to estimate the second term on the right hand side , we use that @xmath222 is bounded thanks to lemma [ lem : we ] and that @xmath212 by the convexity of @xmath223 .
hence @xmath224 for the last term , we use the following lemma , [ lem : effve ] the weak limits @xmath225 of the sequences @xmath121 satisfy for smooth functions @xmath226 , @xmath227 where @xmath228 , @xmath229 , @xmath230 are the weak limits of @xmath231 , @xmath232 and @xmath233 respectively .
applying this lemma to the second term in with @xmath234 , we can estimate it by @xmath235 using that @xmath236 ( cf .
thus , @xmath237 hence grnwall s inequality implies @xmath238 by convexity of the function @xmath223 we also have @xmath239 almost everywhere and so @xmath240 almost everywhere in @xmath34 .
therefore we conclude that the functions @xmath64 converge strongly to @xmath2 almost everywhere and in particular also @xmath241 which means that the limit @xmath157 is a weak solution of the equations .
we multiply the equation for @xmath65 by @xmath232 and integrate over @xmath5 , @xmath242 passing to the limit @xmath52 , we obtain @xmath243 on the other hand , using the smooth function @xmath232 as a test function in the weak formulation of the limit equation @xmath244 and passing to the limit @xmath52 , we obtain @xmath245 combining the last identity with , we obtain .
we consider the problem in two space dimensions in a rectangular domain , for simplicity we use @xmath246 ^ 2 $ ] , the generalization to other rectangular domains as well as three space dimensions is straightforward but more cumbersome in terms of notation , for this reason we restrict ourself to a square two dimensional domain here . for simplicity , we will also assume @xmath63 in the brinkman law in .
we let @xmath247 the mesh width , and @xmath248 the time step size .
we will determine the necessary ratio between @xmath249 and @xmath248 later on .
for @xmath250 , where @xmath251 , @xmath249 chosen such that @xmath252 is an integer , we denote grid cells @xmath253\times ( ( j-1)h , jh]$ ] with cell midpoints @xmath254 .
in addition , we denote @xmath255 , @xmath256 , where @xmath257 for some final time @xmath29 . the approximation of a function @xmath258 at grid point @xmath259 and time @xmath260 will be denoted @xmath261 .
we also introduce the finite differences , @xmath262 and define the discrete laplacian , divergence and gradient operators based on these , @xmath263 for ease of notation , we also let @xmath264 and @xmath265 denote the discrete velocities in the transport equation , specifically , given @xmath266 , we let @xmath267 given @xmath268 at time step @xmath269 , we define the quantities @xmath270 at the next time step by [ eq : expscheme1 ] @xmath271 where @xmath272 and the fluxes @xmath273 , @xmath274 are defined by @xmath275 we use homogeneous neumann or periodic boundary conditions for both variables : @xmath276 the initial condition we approximate taking averages over the cells , @xmath277 in the following , we will prove estimates on the discrete quantities @xmath278 obtained using the scheme . we therefore define the piecewise constant functions @xmath279\times{\omega},\\ \ ] ] where @xmath280 .
we first prove that @xmath281 stays nonnegative and uniformly bounded from above .
[ lem : linfn_h ] if @xmath282 uniformly in @xmath247 and the timestep @xmath248 satisfies the cfl condition @xmath283 ( where @xmath284 ) , then for any @xmath72 , the functions @xmath285 are uniformly ( in @xmath247 ) bounded and nonnegative , specifically , defining @xmath286 , we have for all @xmath287 , @xmath288 the proof goes by induction on the timestep @xmath269 .
clearly , by the assumptions , we have @xmath289 . for the induction step
we therefore assume that this holds for timestep @xmath290 and show that it implies the nonnegativity and boundedness at timestep @xmath291 .
we first show that the @xmath292 are bounded in terms of the @xmath293 .
to do so , let us assume it has a local maximum @xmath294 in a cell @xmath295 , for some @xmath296
. then @xmath297 ( if @xmath298 or @xmath299 , then because of the neumann boundary conditions , the forward / backward difference in direction of the boundary is zero and thus the previous inequality is true as well ) .
hence @xmath300 therefore , @xmath301 similarly , at a local minimum @xmath294 of @xmath302 , we have @xmath303 and hence @xmath304 which implies @xmath305 thus , @xmath306 now we rewrite the scheme as @xmath307 where @xmath308\\ \alpha_{i , j}^{(2),m}&= { \delta t}\,{{{\bm{g}}}}(p_{i , j}^m)+\frac{{\delta t}}{h}\left[u_{i+1/2,j}^m - u_{i-1/2,j}^m+v_{i , j+1/2}^m - v_{i , j-1/2}^m\right]\\ \beta_{i , j}^m&=\frac{{\delta t}}{2h } \left ( u_{i+1/2,j}^m+| u_{i+1/2,j}^m|\right)\\ \zeta_{i , j}^m&=\frac{{\delta t}}{2h}\left(|u_{i-1/2,j}^m|-u_{i-1/2,j}^m\right)\\ \eta_{i , j}^m&=\frac{{\delta t}}{2h } \left ( v_{i , j+1/2}^m+| v_{i , j+1/2}^m|\right)\\ \theta_{i , j}^m&=\frac{{\delta t}}{2h}\left(|v_{i , j-1/2}^m|-v_{i , j-1/2}^m\right)\\ \end{split}\end{aligned}\ ] ] we note that @xmath309 , and that under the cfl - condition , also @xmath310 . hence , assuming that @xmath311 for all @xmath312 , we have @xmath313 we proceed to showing the boundedness of @xmath281 .
thanks to the cfl - condition , we have @xmath314 moreover , @xmath315 . using the induction hypothesis that @xmath316 for all @xmath312 and the nonnegativity of @xmath281 which we have just proved , we can estimate @xmath317 : @xmath318 we can rewrite and bound @xmath319 using the equation for @xmath292 , , @xmath320 where we have used for the first inequality , that @xmath321 for some intermediate value @xmath322 $ ] , with @xmath323 , for the second inequality and the cfl - condition for the last inequality . now going back to and inserting this there , we obtain , @xmath324 if @xmath325 then @xmath326 and hence the expression in is bounded by @xmath327 . on the other hand , if @xmath328 , we can bound it by @xmath329 where we used the definition of @xmath327 for the last equality .
this proves that @xmath330 for all @xmath312 if the same holds already for the @xmath331 .
the estimates in the proof of the previous lemma are very coarse and therefore one can use a much larger cfl - condition than in practice . also note that @xmath332 when @xmath333 .
[ lem : wh ] we have that @xmath334;l^2({\omega})),\ ] ] uniformly in @xmath247 , where @xmath335 and @xmath336 and @xmath337 uniformly in @xmath247 as well . to obtain the @xmath338-estimates , we square the equation for the potential @xmath302 , and sum over all @xmath312 , @xmath339 using summation by parts and that @xmath340 satisfies either periodic or homogeneous neumann boundary conditions
, we obtain @xmath341 from the previous estimates , we know that @xmath342\times{\omega})$ ] uniformly in @xmath247 and therefore also uniformly bounded in any other @xmath343-space , which implies together with the above identity , that @xmath344\times { \omega})$ ] . that @xmath302 is uniformly bounded follows from and the uniform bound on @xmath281 which was proved in the previous lemma [ lem : linfn_h ] . using this and the uniform boundedness of the pressure , we conclude by that also @xmath345 is uniformly bounded . using the discrete gagliardo - nirenberg - sobolev inequality , ( * ? ? ?
3.4 ) , we obtain that @xmath346;l^q({\omega}))$ ] for @xmath347 .
to prove strong convergence of the approximating sequence @xmath348 , it will be useful to derive entropy inequalities for @xmath281 .
to this end , the following lemma will be useful : [ lem : fn ] let @xmath122 be a smooth convex function and assume that @xmath248 satisfies the cfl - condition @xmath349 denote @xmath350 and @xmath351 a piecewise constant interpolation of it as in .
then @xmath261 satisfies the following identity @xmath352 + \frac{h}{4 } d^-_2 \left[f'({n}_{i , j}^m)|v_{i , j+1/2}^m| d^+_2 { n}_{i , j}^m\right]\\ \label{eq : t3 } & \hphantom{=}+\frac{h}{4 } d^+_1 \left[f'({n}_{i , j}^m)|u_{i-1/2,j}^m| d^-_1 { n}_{i , j}^m\right ] + \frac{h}{4 } d^+_2 \left[f'({n}_{i , j}^m)|v_{i , j-1/2}^m| d^-_2 { n}_{i , j}^m\right]\\ \label{eq : t4 } & \hphantom{=}-\frac{h^2}{4}d^-_1\left [ f''(\widetilde{n}_{i+1/2,j}^m ) u_{i+1/2,j}^m |d^+_1 { n}_{i , j}^m|^2\right]\\\label{eq : t5 } & \hphantom{=}-\frac{h^2}{4}d^-_2\left [ f''(\widetilde{n}_{i , j+1/2}^m ) v_{i , j+1/2}^m |d^+_2 { n}_{i , j}^m|^2\right]\\\label{eq : t6 } & \hphantom{= } -\frac{h}{4}f''(\widehat{{n}}_{i-1/2,j}^m ) |u_{i-1/2,j}^m||d^-_1 { n}_{i , j}^m|^2-\frac{h}{4}f''(\widehat{{n}}_{i , j-1/2}^m ) |v_{i , j-1/2}^m||d^-_2 { n}_{i , j}^m|^2\\ \label{eq : t7 } & \hphantom{= } -\frac{h}{4}f''(\widehat{{n}}_{i+1/2,j}^m ) |u_{i+1/2,j}^m||d^+_1 { n}_{i , j}^m|^2-\frac{h}{4}f''(\widehat{{n}}_{i , j+1/2}^m ) |v_{i , j+1/2}^m||d^+_2 { n}_{i , j}^m|^2\\ \label{eq : t8 } & \hphantom{=}+(f'({n}_{i , j}^m){n}_{i , j}^m - f_{i , j}^m ) \delta_h { w}_{i , j}^m+f'({n}_{i , j}^m ) { n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m)\\ \label{eq : t9 } & \hphantom{=}+\frac{{\delta t}}{2 } f''(\widetilde{{n}}^{m+1/2}_{i , j } ) |d_t^+ { n}_{i , j}^m|^2 , \ ] ] where @xmath353 $ ] , @xmath354 $ ] and @xmath355 $ ] and where the term is uniformly bounded and the terms and satisfy @xmath356 in particular , this implies that the piecewise constant interpolation @xmath357 is of the form @xmath358 where @xmath359\times { \omega})$ ] and @xmath360 ; w^{-1,q}({\omega}))$ ] for any @xmath361 if @xmath362 and for @xmath363 if @xmath364 , uniformly in @xmath247 .
we first rewrite the scheme for @xmath331 as @xmath365 + \frac{h}{2 } d^-_2 \left[|v_{i , j+1/2}^m| d^+_2 { n}_{i , j}^m\right]\\ & \hphantom{=}+{n}_{i , j}^m \delta_h { w}_{i , j}^m+{n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m ) . \end{split}\end{aligned}\ ] ] then , using the taylor expansion , @xmath366 where @xmath367 $ ] , we can write @xmath368 where @xmath369 and @xmath370 are intermediate values . hence
, multiplying equation by @xmath371 , it becomes @xmath372 -\frac{h}{4}f''(\widehat{{n}}_{i-1/2,j}^m ) |u_{i-1/2,j}^m||d^-_1 { n}_{i , j}^m|^2\\ & \hphantom{=}+\frac{h}{4 } d^-_2 \left[f'({n}_{i , j}^m)|v_{i , j+1/2}^m| d^+_2 { n}_{i , j}^m\right ] -\frac{h}{4}f''(\widehat{{n}}_{i , j-1/2}^m ) |v_{i , j-1/2}^m||d^-_2 { n}_{i , j}^m|^2\\ & \hphantom{=}+\frac{h}{4 } d^+_1 \left[f'({n}_{i , j}^m)|u_{i-1/2,j}^m| d^-_1 { n}_{i , j}^m\right ] -\frac{h}{4}f''(\widehat{{n}}_{i+1/2,j}^m ) |u_{i+1/2,j}^m||d^+_1 { n}_{i , j}^m|^2\\ & \hphantom{=}+\frac{h}{4 } d^+_2 \left[f'({n}_{i , j}^m)|v_{i , j-1/2}^m| d^-_2 { n}_{i , j}^m\right ] -\frac{h}{4}f''(\widehat{{n}}_{i , j+1/2}^m ) |v_{i , j+1/2}^m||d^+_2 { n}_{i , j}^m|^2\\ & \hphantom{=}+f'({n}_{i , j}^m){n}_{i , j}^m \delta_h { w}_{i , j}^m+f'({n}_{i , j}^m ) { n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m)\\ & = \frac{{\delta t}}{2 } f''(\widetilde{{n}}^{m+1/2}_{i , j } ) |d_t^+ { n}_{i , j}^m|^2\\ & \hphantom{=}+ \frac{1}{2 } d^-_1\left ( u_{i+1/2,j}^m \left(f_{i , j}^m+f_{i+1,j}^m\right)\right ) + \frac{1}{2 } d_2 ^ -\left ( v_{i , j+1/2}^m \left(f_{i , j}^m + f_{i , j+1}^m\right)\right)\\ & \hphantom{=}+\frac{h}{4 } d^-_1 \left[f'({n}_{i , j}^m)|u_{i+1/2,j}^m| d^+_1 { n}_{i , j}^m\right ] + \frac{h}{4 } d^-_2 \left[f'({n}_{i , j}^m)|v_{i , j+1/2}^m| d^+_2 { n}_{i , j}^m\right]\\ & \hphantom{=}+\frac{h}{4 } d^+_1 \left[f'({n}_{i , j}^m)|u_{i-1/2,j}^m| d^-_1 { n}_{i , j}^m\right ] + \frac{h}{4 } d^+_2 \left[f'({n}_{i , j}^m)|v_{i , j-1/2}^m| d^-_2 { n}_{i , j}^m\right]\\ & \hphantom{=}-\frac{h^2}{4}d^-_1\left [ f''(\widetilde{n}_{i+1/2,j}^m ) u_{i+1/2,j}^m |d^+_1 { n}_{i , j}^m|^2\right]\\&\hphantom{=}-\frac{h^2}{4}d^-_2\left [ f''(\widetilde{n}_{i , j+1/2}^m ) v_{i , j+1/2}^m |d^+_2 { n}_{i , j}^m|^2\right]\\&\hphantom{= } -\frac{h}{4}f''(\widehat{{n}}_{i-1/2,j}^m ) |u_{i-1/2,j}^m||d^-_1 { n}_{i , j}^m|^2-\frac{h}{4}f''(\widehat{{n}}_{i , j-1/2}^m ) |v_{i , j-1/2}^m||d^-_2 { n}_{i , j}^m|^2\\ & \hphantom{= } -\frac{h}{4}f''(\widehat{{n}}_{i+1/2,j}^m ) |u_{i+1/2,j}^m||d^+_1 { n}_{i , j}^m|^2-\frac{h}{4}f''(\widehat{{n}}_{i , j+1/2}^m ) |v_{i , j+1/2}^m||d^+_2 { n}_{i , j}^m|^2\\ & \hphantom{=}+(f'({n}_{i , j}^m){n}_{i , j}^m - f_{i , j}^m ) \delta_h { w}_{i , j}^m+f'({n}_{i , j}^m ) { n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m).\end{aligned}\ ] ] which implies . in particular , for @xmath223 , this becomes @xmath373-\frac{h^2}{2}d^-_2\left [ v_{i , j+1/2}^m |d^+_2 { n}_{i , j}^m|^2\right]\\&\hphantom{=}+\frac{h}{2 } d^-_1 \left[{n}_{i , j}^m|u_{i+1/2,j}^m| d^+_1 { n}_{i , j}^m\right]+\frac{h}{2 } d^-_2 \left[{n}_{i , j}^m|v_{i , j+1/2}^m| d^+_2 { n}_{i , j}^m\right ] \\ & \hphantom{=}+\frac{h}{2 } d^+_1 \left[{n}_{i , j}^m|u_{i-1/2,j}^m| d^-_1 { n}_{i , j}^m\right ] + \frac{h}{2 } d^+_2 \left[{n}_{i , j}^m|v_{i , j-1/2}^m| d^-_2 { n}_{i , j}^m\right]\\ & \hphantom{=}-\frac{h}{2 } |u_{i-1/2,j}^m||d^-_1 { n}_{i , j}^m|^2-\frac{h}{2}|v_{i , j-1/2}^m||d^-_2 { n}_{i , j}^m|^2\\ & \hphantom{= } -\frac{h}{2 } |u_{i+1/2,j}^m||d^+_1 { n}_{i , j}^m|^2-\frac{h}{2}|v_{i , j+1/2}^m||d^+_2 { n}_{i , j}^m|^2\\ & \hphantom{=}+ f_{i , j}^m \delta_h { w}_{i , j}^m+2 f_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m ) , \end{split}\end{aligned}\ ] ] we estimate the first term on the right hand side of the inequality inserting , @xmath374 + \frac{h}{2 } d^-_2 \left[|v_{i , j+1/2}^m| d^+_2 { n}_{i , j}^m\right]\biggr|^2 \\ & \hphantom{\leq } + 2\bigl|{n}_{i , j}^m \delta_h { w}_{i , j}^m+{n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m)\bigr|^2\\ & \leq 4\biggl|\frac{1}{2 } u_{i+1/2,j}^m d^+_1 { n}_{i , j}^m+\frac{1}{2 } u_{i-1/2,j}^m d^-_1 { n}_{i , j}^m+\frac{h}{2 } d^-_1 \left[|u_{i+1/2,j}^m| d^+_1 { n}_{i , j}^m\right]\biggr|^2\\ & \hphantom{\leq } + 4\biggl|\frac{1}{2 } v_{i , j+1/2}^m d^+_2 { n}_{i , j}^m+\frac{1}{2 } v_{i , j-1/2}^m d^-_2 { n}_{i , j}^m+\frac{h}{2 } d^-_2 \left[|v_{i , j+1/2}^m| d^+_2 { n}_{i , j}^m\right]\biggr|^2\\ & \hphantom{\leq } + 2\bigl|{n}_{i , j}^m \delta_h { w}_{i , j}^m+{n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m)\bigr|^2\\ & \leq 8 \big| u_{i+1/2,j}^m d^+_1 { n}_{i , j}^m\big|^2 + 8\big| u_{i-1/2,j}^m d^-_1 { n}_{i , j}^m\big|^2 + 8 \big| v_{i , j+1/2}^m d^+_2 { n}_{i , j}^m\big|^2\\ & \hphantom{\leq } + 8\big| u_{i , j-1/2}^m d^-_2 { n}_{i , j}^m\big|^2 + 2\bigl|{n}_{i , j}^m \delta_h { w}_{i , j}^m+{n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m)\bigr|^2\\ & \leq 8\max_{i , j } |{\nabla}_h { w}_{i , j}^m|\bigl\ { | u_{i+1/2,j}^m|\ , |d^+_1 { n}_{i , j}^m\big|^2+| u_{i-1/2,j}^m|\ , |d^-_1 { n}_{i , j}^m\big|^2\\ & \hphantom{\leq 8\max_{i , j } |{\nabla}_h { w}_{i , j}^m|\bigl\{}+| v_{i , j+1/2}^m|\ , |d^+_2 { n}_{i , j}^m\big|^2+| v_{i , j-1/2}^m|\ , |d^-_2 { n}_{i , j}^m\big|^2\bigr\}\\ & \hphantom{\leq } + 2\bigl|{n}_{i , j}^m \delta_h { w}_{i , j}^m+{n}_{i , j}^m { { { \bm{g}}}}(p_{i , j}^m)\bigr|^2\end{aligned}\ ] ] thus if we assume that @xmath248 satisfies the cfl - condition , we have @xmath375 now summing over all @xmath312 , multiplying with @xmath376 and using the latter inequality , we obtain @xmath377 where @xmath127 is a constant independent of @xmath249 , thanks to the @xmath133-bounds on @xmath281 and @xmath345 obtained in lemma [ lem : linfn_h ] and [ lem : wh ] .
this implies that @xmath378 and therefore using hlder s inequality and the uniform @xmath133-bounds on @xmath281 , . using summation by parts
, we realize that the other terms , are in @xmath136;w^{-1,q}({\omega}))$ ] for @xmath379 where @xmath380 if @xmath381 and any finite number greater than one if @xmath362 .
the preceeding lemma implies that the forward time difference of the approximation of the pressure @xmath382 is of the form @xmath383 where @xmath359\times { \omega})$ ] and @xmath360 ; w^{-1,q}({\omega}))$ ] for any @xmath361 if @xmath362 and for @xmath363 if @xmath364 , uniformly in @xmath247 . using this , we have that @xmath384 where @xmath385 and @xmath386 solve @xmath387 by lemma [ lem : elll1 ]
, we have @xmath388 ; l^q({\omega}))$ ] for @xmath389 and by standard results , @xmath386 , @xmath390 ; l^2({\omega}))$ ] . hence @xmath391 ; l^q({\omega}))+ l^\infty([0,t ] ; l^2({\omega}))$ ] .
the estimates from lemma [ lem : wh ] imply that the velocity @xmath392;l^{2^*}({\omega}))$ ] uniformly in @xmath247 , @xmath380 or any number in @xmath393 if @xmath362 , using the sobolev embedding theorem . using an inverse inequality , we can bound it in the @xmath394-norm as follows : @xmath395 thus the time step size @xmath248 is of order @xmath396 . in practice
a linear cfl - condition seems to work well though .
the estimates of the previous ( sub)sections allow us to pass to the limit @xmath54 in a subsequence still denoted @xmath249,@xmath397\times { \omega } ) , \ , 1\leq q<\infty,\\ p_h{\rightharpoonup}\overline{p}\geq 0,&\quad \mathrm{in}\ , l^q([0,t]\times { \omega } ) , \ , 1\leq q<\infty,\end{aligned}\ ] ] where @xmath398 and @xmath399\times { \omega})$ ] . using the `` discretized '' aubin - lions lemma [ lem : a - l - lem ] for @xmath302 and @xmath400
, we obtain strong convergence of a subsequence in @xmath154\times{\omega})$ ] for any @xmath155 in the case of @xmath302 and @xmath401 in the case of @xmath400 ( @xmath380 if @xmath381 and any finite number greater than or equal to one if @xmath362 ) , to limit functions @xmath156\times { \omega})$ ] .
moreover , from the estimates in lemma [ lem : wh ] we obtain that @xmath45\times{\omega})\cap l^{\infty}([0,t];h^2({\omega}))$ ] .
hence we have that @xmath157 satisfy for any @xmath402\times{\omega})$ ] , @xmath403 where @xmath160 is the weak limit of @xmath404 . to conclude that the limit @xmath162 is a weak solution of
, we proceed as in the previous section [ s4 ] and show that @xmath281 in fact converges strongly : first , we recall that the limit @xmath2 satisfies . on the other hand , from ,
we obtain ( under the cfl - condtion ) @xmath405-\frac{h^2}{2}d^-_2\left [ v_{i , j+1/2}^m |d^+_2 { n}_{i , j}^m|^2\right]\\ & \hphantom{=}+\frac{h}{2 } d^-_1 \left[{n}_{i , j}^m|u_{i+1/2,j}| d^+_1 { n}_{i , j}^m\right]+\frac{h}{2 } d^-_2 \left[{n}_{i , j}^m|v_{i , j+1/2}| d^+_2 { n}_{i , j}^m\right ] \\ & \hphantom{=}+\frac{h}{2 } d^+_1 \left[{n}_{i , j}^m|u_{i-1/2,j}| d^-_1 { n}_{i , j}^m\right ] + \frac{h}{2 } d^+_2 \left[{n}_{i , j}^m|v_{i , j-1/2}| d^-_2 { n}_{i , j}^m\right]\\ & \hphantom{=}+|{n}_{i , j}^m|^2 \delta_h { w}_{i , j}^m+2|{n}_{i , j}^m|^2 { { { \bm{g}}}}(p_{i , j}^m ) , \end{split}\end{aligned}\ ] ] considering this inequality in terms of the piecewise constant functions @xmath281 , @xmath302 and @xmath406 , multiplying it with a nonnegative @xmath407-test function @xmath178 , integrating and then passing to the limit @xmath54 , we obtain ( using the bounds , the weak convergence of @xmath281 and @xmath406 and the strong convergence of @xmath302 and @xmath408 ) , @xmath409 where @xmath204 denotes the weak limit of @xmath410 and @xmath206 and @xmath207 are the weak limits of @xmath411 and @xmath412 respectively . adding and , we have @xmath413 we now choose smooth test functions @xmath214 approximating @xmath215}(t)$ ] , where @xmath216 $ ] , in this inequality and then pass to the limit @xmath414 to obtain @xmath415 by convexity of @xmath223 , we have @xmath212 , on the other hand , the discrete @xmath338-entropy inequality , , implies @xmath416 which gives , passing to the limit @xmath54 , @xmath417 letting @xmath210 , the second term on the right hand side vanishes ( as the integrand is bounded ) , and we obtain @xmath418 we deduce that @xmath419 almost everywhere and that therefore the second term on the left hand side of is zero .
we have already estimated the first two terms on the right hand side of in and . to bound the other term , we use a discretized version of lemma [ lem : effve ] : [ lem : effv ] the weak limits @xmath225 of the sequences @xmath420 satisfy for any smooth function @xmath226 , @xmath421 where @xmath228 , @xmath229 , @xmath230 are the weak limits of @xmath422 , @xmath423 and @xmath424 respectively . applying this lemma to the last term in with @xmath234 , we can estimate it by @xmath235 using again that by exercise 3.37 in @xcite , @xmath236 .
thus , @xmath237 grnwall s inequality thus implies @xmath425 by convexity of the function @xmath223 we also have @xmath426 almost everywhere and hence @xmath427 almost everywhere in @xmath34 .
therefore we conclude that the functions @xmath281 converge strongly to @xmath2 almost everywhere , thus also @xmath241 and so the limit @xmath157 is a weak solution of the equations .
we multiply the equation for @xmath302 by @xmath423 and integrate it over the spatial domain @xmath5 , @xmath428 passing to the limit @xmath54 in the last equation , we obtain @xmath429 on the other hand , using @xmath430(x)$ ] , where @xmath431 is a smooth mollifier converging to a dirac measure at zero when @xmath432 is sent to zero , as a test function in the weak formulation of the limit equation @xmath244 and passing first to the limit @xmath197 and then @xmath54 , we obtain @xmath433 combining the last identity with , we obtain .
to test the scheme in practice , we compute approximations for the following two examples . as a first example , we consider the initial data @xmath434 on the domain @xmath435 ^ 2 $ ] and @xmath436 with pressure law @xmath437 and @xmath438 and @xmath439 . strictly speaking ,
these are not homogeneous neumann boundary conditions , but since the gradient of @xmath57 near the boundary is very small , this works well in practice . [ cols= " < , > " , ] the approximations computed at times @xmath440 are shown in figure [ fig : eg2 ] .
the interface between the area with maximum cell density and zero cell density seems to be sharper than in the previous example , this appears to be caused by the pressure law with the higher exponent @xmath441 .
further tests with higher and lower exponents confirmed that assertion .
[ lem : a - l - lem]let @xmath442 be a piecewise constant function defined on a grid on @xmath443 , @xmath5 a bounded rectangular domain , satisfying @xmath444 for some @xmath445 , uniformly with respect to @xmath247 and @xmath446 where @xmath447 is a first order linear finite difference operator , and @xmath448 are piecewise constant functions , satisfying uniformly in @xmath247 , @xmath449 for some @xmath450 .
then @xmath451 in @xmath452 . denote @xmath453 a piecewise linear interpolation of @xmath454 in space piecewise constant in time and similarly , let @xmath455 , @xmath456 and @xmath457 piecewise linear interpolations of @xmath458 , @xmath351 and @xmath459 respectively in space and piecewise constant in time such that @xmath460 by ladyshenskaya s norm equivalences ( * ? ? ?
* ff ) , we have @xmath461 where the right hand sides are bounded by assumptions and . since @xmath462 for @xmath463 , we have that @xmath464;w^{-1,s}({\omega}))$ ] for @xmath463 and hence thanks to this and , we obtain @xmath465 uniformly with respect to the discretization parameter @xmath247 .
thus we can apply the version ( * ? ? ?
* theorem 1 ) of the aubin - lions lemma to find that up to a subsequence @xmath466 in @xmath452 and the limit @xmath467 . by (
* lemma 3.2 . ,
this implies that also @xmath451 in @xmath452 ( and @xmath468 ) .
if the @xmath454 in lemma [ lem : a - l - lem ] are of the form @xmath469 for some @xmath470 piecewise constant function , this lemma implies that @xmath471 in @xmath472 , again applying ( * ? ? ?
* lemma 3.2 . ,
in this section , we prove the following lemma : [ lem : elll1 ] let @xmath454 solve the difference equation @xmath473 with homogeneous neumann boundary conditions , where @xmath447 is a diagonal positive definite @xmath474-matrix with entries @xmath475 and @xmath476 uniformly in @xmath247 , @xmath82 , @xmath5 is a rectangular domain in @xmath6 and @xmath477 uniformly in @xmath247 .
we have denoted @xmath478 and @xmath479 ( or alternatively @xmath480 and @xmath481 ) .
then @xmath482 where @xmath483 , for a constant @xmath127 independent of @xmath247 .
the proof of this lemma will be a ( simplified ) finite difference version of the proof of theorem 2.1 in @xcite .
but before proving the lemma , we need to introduce some notation .
[ not : marcinkiewicz ] for any @xmath484 , we denote by @xmath485 the marcinkiewicz space with norm defined by @xmath486 the marcinkiewicz spaces are continuously embedded in @xmath487 for any @xmath488 , @xcite : @xmath489 moreover , we need the trunctation operator @xmath490 defined as follows : [ not : truncation ] let @xmath491 be a real number . then we define the truncation operator @xmath492 by @xmath493 it will be convenient in the proof to use the following tuple notation for the finite difference approximations : [ not : tuple ] we denote @xmath494 , @xmath495 , @xmath496 the number of cells in the @xmath497th spatial direction , a @xmath61-dimensional tuple and and @xmath498 the approximation in cell @xmath499\times\dots\times ( ( i_d-1)h , i_d h]$ ] .
the piecewise constant function @xmath454 can be written as @xmath500 we also need the following auxilary result : [ lem : truncl2 ] let @xmath454 solve the difference equation under the assumptions of lemma [ lem : elll1 ] .
then @xmath501 for some constant @xmath127 independent of @xmath247 .
given @xmath491 , we multiply equation by @xmath502 and integrate over the domain @xmath5 . after changing variables in the integrals ,
we obtain @xmath503 the right hand side can be bounded by @xmath504 using hlder s inequality .
the left hand side , we can rewrite and estimate as follows @xmath505\right)\right)\cdot { \nabla}_h s_k(u_h ) + c_h \left(u_h - s_k(u_h)\right ) s_k(u_h)\ , dx\\ & \quad \geq \eta
\|{\nabla}_h s_k(u_h)\|^2_{l^2({\omega})}+\nu \|s_k(u_h)\|^2_{l^2({\omega } ) } \\ & \quad\quad+ \int_{{\omega } } \left(a_h \left({\nabla}_h \left[u_h - s_k(u_h)\right]\right)\right)\cdot { \nabla}_h s_k(u_h ) + c_h \left(u_h - s_k(u_h)\right ) s_k(u_h)\ , dx . \end{split}\ ]
] @xmath506 is either zero or has the same sign as @xmath502 .
therefore @xmath507 and @xmath508 in order to prove that the other term is positive as well , we will show that @xmath509 the proof of this fact consists of boring case distinctions and is exactly analoguous for @xmath510 , therefore we will do it only for @xmath511 and omit writing the tuple index @xmath512 .
then we have @xmath513 the potential reader is welcome to check that these are all the possible cases and that each of the terms on the right hand side is nonnegative .
thus we have that @xmath514 which implies together with the estimate on the right hand side of first , we note that by the discrete gagliardo - nirenberg - sobolev inequality , ( * ? ? ?
3.4 ) , @xmath515 where @xmath380 if @xmath381 and any number with @xmath516 if @xmath362 , and where @xmath21 is a constant depending on @xmath517 but not on @xmath247 . by lemma [ lem : truncl2 ] , we can bound the right hand side and obtain therefore @xmath518 now we define the set @xmath519 by @xmath520 we have @xmath521 and therefore , using , @xmath522 for @xmath523 ( which is @xmath524 if @xmath381 ) since the choice of @xmath491 was arbitrary .
now denote @xmath525 where @xmath526 . informally speaking ,
the cells in @xmath527 have a neighbor cell which is contained in @xmath519 .
we have @xmath528 by .
now let @xmath529 , @xmath491 and decompose @xmath530 hence @xmath531 on @xmath532 and the cells bordering the set , we have @xmath533 and therefore @xmath534 .
hence we can estimate the size of the second set in the above inequality , @xmath535 where we have used chebyshev inequality for the last step .
now we can estimate the size of the set @xmath536 using once more , @xmath537 choosing @xmath538 , we obtain @xmath539 if @xmath381 , we have @xmath540 and so @xmath541 for @xmath542 .
for @xmath362 , since @xmath543 is an arbitrary finite positive number , we can achieve the same . using the embedding of the marcinkiewicz spaces ,
, we obtain the claim of the lemma .
the work of k.t . was supported in part by the national science foundation under the grant dms-1211519 .
the work of f.w . was supported by the research council of norway , project 214495 liqcry .
f.w . gratefully acknowledges the support by the center for scientific computation and mathematical modeling at the university of maryland where part of this research was performed during her visit in fall 2014 . o. a. ladyzhenskaya . .
second english edition , revised and enlarged . translated from the russian by richard a. silverman and john chu
. mathematics and its applications , vol .
2 . gordon and breach , science publishers , new york - london - paris , 1969 .
l . lions . , volume 3 of _ oxford lecture series in mathematics and its applications_. the clarendon press , oxford university press , new york , 1996 .
incompressible models , oxford science publications . | mechanical models for tumor growth have been used extensively in recent years for the analysis of medical observations and for the prediction of cancer evolution based on imaging analysis .
this work deals with the numerical approximation of a mechanical model for tumor growth and the analysis of its dynamics .
the system under investigation is given by a multi - phase flow model : the densities of the different cells are governed by a transport equation for the evolution of tumor cells , whereas the velocity field is given by a brinkman regularization of the classical darcy s law .
an efficient finite difference scheme is proposed and shown to converge to a weak solution of the system .
our approach relies on convergence and compactness arguments in the spirit of lions @xcite .
[ multiblock footnote omitted ] |
srf researchers have begun a significant effort to develop alternative materials to niobium , superconductors that could offer higher accelerating gradients @xmath2 and/or lower surface resistances @xmath3 at a given temperature .
there are several promising candidates , but most of them suffer from two potential liabilities .
first , as shown in table [ tab : mats ] , they have relatively small first critical fields @xmath0 , the magnetic field at which it becomes energetically favorable for a vortex to be inside the superconductor .
second , they have relatively small coherence lengths @xmath4 . vortex penetration is prevented at fields significantly above @xmath0 by an energy barrier , but surface defects on the order of @xmath4 can reduce this barrier .
these materials have @xmath4 on the order of a few nm , compared to tens of nm for niobium , making even very small defects a potential vulnerability . as a result
, there has been significant concern in the srf community over whether vortex dissipation will occur if these materials are exposed to fields that bring them into the metastable state between @xmath0 and @xmath1 , the superheating field at which the energy barrier is reduced to zero for an ideal surface .
.material properties of niobium and three promising alternative srf materials .
the penetration depth @xmath5 is calculated using eqn 3.131 in @xcite .
the correlation length @xmath4 is calculated using the equations in @xcite .
for nb a rrr of 100 was assumed .
for mgb@xmath6 , @xmath5 and @xmath4 are not calculated , as the experimental values are given in the reference . for calculations ,
@xmath7 is used , where @xmath8 is the flux quantum @xcite .
@xmath0 for nb found from power law fit to numerically computed data from @xcite and for strongly type ii materials is found from eqn 5.18 in @xcite .
@xmath1 for nb is found from @xcite and for others calculated from @xmath9 ( valid only for strongly type ii materials near @xmath10 ) @xcite .
nb data from @xcite , nb@xmath11sn data from @xcite , nbn data from @xcite , and mgb@xmath6 data from @xcite .
note that the two gap nature of mgb@xmath6 may require more careful analysis than is performed here . [
cols="^,>,>,>,>",options="header " , ] a. gurevich proposed @xcite a method to avoid the potentially vulnerable metastable state altogether . pointing to the enhancement of parallel @xmath0 in films with thickness @xmath12 smaller than the penetration depth @xmath5 , he suggested coating a niobium cavity with alternating layers of insulator ( i ) and thin film superconductor ( s ) .
with such a sis structure , he proposed it might be possible to take advantage of the high @xmath1 and low @xmath3 of the alternative superconductors used in the thin films without the disadvantage of their small @xmath0 .
srf researchers have been putting significant effort into developing sis multilayers , and they are producing excellent work @xcite@xcite@xcite@xcite@xcite@xcite@xcite . in this work , we will start by showing that in practice there is no enhancement of @xmath0 for sis films we will show that in fact @xmath0 is zero for such a structure .
next we will study the superheating field of sis structures and show that for a homolaminate , @xmath1 is always lower than the bulk value , and for a heterolaminate , only a small increase in @xmath1 is possible .
following this , we show that sis structures can not be used above @xmath1 for srf applications , as the heating would be unmanageable .
we then consider sis multilayer films for dc and at low frequencies , and show that they may be effective in screening large fields by setting up a gradient in the phase of the order parameter .
finally , we consider the outlook for alternative materials for srf cavities and show that recent developments give reason for strong optimism for bulk films .
tinkham @xcite defines @xmath0 as the field at which `` the gibbs free energy [ has ] the same value whether the first vortex is in or out of the sample . '' for a sis , the sample under consideration should be the full structure @xcite .
stejic et al .
@xcite calculate the gibbs free energy of a vortex in a thin film superconductor immersed in a parallel external field .
they show that @xmath0 of the film is enhanced relative to the bulk value , according to @xmath13 where @xmath8 is the flux quantum , @xmath14 and @xmath15 . however , if a sis structure is used to screen nb srf cavities , the geometry is quite different than that of an isolated film .
how does stejic s expression for @xmath0 change when the film is screening a bulk superconductor ? in this case , it will have a b - field gradient across it , which will affect the free energy .
we can use the same formalism as stejic to calculate the gibbs free energy for this case , and use it to find @xmath0 and @xmath1@xcite . consider a single layer sis structure , as shown in fig .
[ fig : geometry ] .
a strongly type ii superconducting film of thickness @xmath12 , penetration depth @xmath16 , and coherence length @xmath17 is separated from a bulk superconductor with with penetration depth @xmath18 by an insulating film of thickness @xmath19 .
the superconducting film is screening the bulk from a parallel magnetic field with amplitude @xmath20 .
the screened field inside the bulk region has amplitude @xmath21 . in our geometry ,
the x - axis is perpendicular to the film , pointing into it , with origin at the interface with the exterior .
the z - axis is aligned with the magnetic field .
stejic shows that the gibbs free energy of a vortex in a superconductor can be determined from the value of two magnetic fields evaluated at the vortex location @xmath22 : the meissner - screened external field @xmath23 and the field generated by the vortex in the film @xmath24 .
@xmath25 @xmath23 can be found by minimizing the free energy in the structure when no vortex is present .
this procedure gives : @xmath26 where @xmath21 is given by @xmath27^ { -1 } \label{eq : bi}\ ] ] stejic gives a relatively simple expression for @xmath24 for the case when @xmath15 , but this would restrict us to very thin films . to study the full range of thicknesses
, we turn to the more general expression from shmidt @xcite ( this expression assumes @xmath28 ) , which agrees with stejic s expression for very small films : @xmath29 we can check our procedure by choosing @xmath30 , such that the film behaves as a bulk supercondcutor .
this calculation is shown in the top plot of fig .
[ fig : gplots ] .
@xmath31 when the free energy outside the superconductor is equal to that when a vortex is deep in the bulk .
@xmath32 when the barrier to flux penetration is reduced to zero ( this plot is very similar to the one from bean and livingston s 1963 paper @xcite ) .
sn , ( center ) a 50 nm nb@xmath11sn thin film , and ( bottom ) a sis structure with a 50 nm nb@xmath11sn film on a nb@xmath11sn bulk .
@xmath0 is the smallest field at which there is a position inside the structure where the free energy for a vortex is smaller than the value outside .
@xmath1 is the field at which the energy barrier to vortex penetration disappears .
the top and center plots show the @xmath0 enhancement for a thin film compared to a bulk .
the bottom plot shows that for a sis structure @xmath33 .
the expression for the thin film @xmath0 is not valid for the sis structure because it assumes that the first stable vortex position will be at the center of the film .
however , for the sis structure , the first stable vortex position occurs on the side of the film adjacent to the insulating layer.,title="fig:",scaledwidth=48.0% ] sn , ( center ) a 50 nm nb@xmath11sn thin film , and ( bottom ) a sis structure with a 50 nm nb@xmath11sn film on a nb@xmath11sn bulk .
@xmath0 is the smallest field at which there is a position inside the structure where the free energy for a vortex is smaller than the value outside .
@xmath1 is the field at which the energy barrier to vortex penetration disappears .
the top and center plots show the @xmath0 enhancement for a thin film compared to a bulk .
the bottom plot shows that for a sis structure @xmath33 .
the expression for the thin film @xmath0 is not valid for the sis structure because it assumes that the first stable vortex position will be at the center of the film .
however , for the sis structure , the first stable vortex position occurs on the side of the film adjacent to the insulating layer.,title="fig:",scaledwidth=48.0% ] sn , ( center ) a 50 nm nb@xmath11sn thin film , and ( bottom ) a sis structure with a 50 nm nb@xmath11sn film on a nb@xmath11sn bulk .
@xmath0 is the smallest field at which there is a position inside the structure where the free energy for a vortex is smaller than the value outside .
@xmath1 is the field at which the energy barrier to vortex penetration disappears .
the top and center plots show the @xmath0 enhancement for a thin film compared to a bulk .
the bottom plot shows that for a sis structure @xmath33 .
the expression for the thin film @xmath0 is not valid for the sis structure because it assumes that the first stable vortex position will be at the center of the film .
however , for the sis structure , the first stable vortex position occurs on the side of the film adjacent to the insulating layer.,title="fig:",scaledwidth=48.0% ] we can study a single thin film ( not in a sis structure ) by setting @xmath34 in eqn .
[ eq : bm ] .
this calculation is shown in the center plot of fig .
[ fig : gplots ] ( the free energy outside the film is subracted from each of the plots for clarity ) . in this case
, there is no bulk , so the first location at which the free energy drops below the external value at high fields is in the center of the film
. this would be the stable position for a single vortex above @xmath0 .
both @xmath0 and @xmath1 are much higher for the film than the bulk . finally , we plot the free energy of vortex in a single sis structure in the bottom plot of fig . [
fig : gplots ] .
in contrast to the previous case , only one side of the thin film is exposed to the external magnetic field .
the field at the other side is smaller due to screening by film . since @xmath24 = 0 at the edges of the film , eqn .
[ eq : gibbs ] shows that the free energy in the insulating layer is lower than the free energy outside .
the film provides screening at any finite @xmath20 below the second critical field , so for @xmath35 , the energetically favorable configuration is for flux to be trapped in the insulating layer . as we explain below
, this implies that in practice for the sis structure , @xmath0 is zero
. why is eqn .
[ eq : bc1thin ] describing the enhancement of @xmath0 in a lone thin film not applicable for the sis structure ?
this expression assumes that the first stable vortex position will occur in the center of the film .
it predicts when the free energy at the center of the film will dip below the value of the free energy in the exterior .
however , for the sis structure , the free energy at the insulator side of the film will dip below the exterior value at fields much smaller than this .
[ fig : gplots ] shows that at moderate fields , when @xmath21 is below @xmath0 of the bulk superconductor , there is no stable position for a vortex in either the bulk or the film . in effect
, both superconductors are below their individual @xmath0 , so it is not immediately obvious if @xmath0 of the overall structure is important .
let us consider the implications of it being energetically favorable for a vortex to pass through the film , and have its flux trapped in the insulating layer .
once the flux is trapped in this way , it is non - dissipative under rf fields ( unlike a vortex , which has a normal conducting core ) .
however , as the vortex penetrates through the film to the insulator , dissipation occurs due to drag , and later we will show that this dissipation is too strong to be tolerable for srf applications .
therefore it is the @xmath0 of the sis structure that is important , not that of the individual superconductors . above @xmath0 ,
the structure is in a metastable state : only the energy barrier of the film prevents quench - inducing vortex penetration .
we have shown that the @xmath0 of sis structures is always zero .
however , they may still offer a significant advantage over bulk films if they have a higher superheating field . in fig .
[ fig : gplots ] , the barrier height is indicated for fields between @xmath0 and @xmath1 .
we can find the superheating field easily using the same free energy calculations to find the field at which the barrier is reduced to zero . in fig .
[ fig : bshvsdmulti ] , the maximum applied field at which both the film and the bulk are exposed to fields below their respective @xmath1 is plotted as a function of film thickness . various materials and insulator thicknesses are considered @xcite . of both the film and the bulk as a function of film thickness for various film materials in a sis structure with nb .
the effect of varying the insulator thickness @xmath19 is shown for the nb@xmath11sn film , as is the effect of splitting the film thickness @xmath12 over 5 equally thick multilayers with thin separating insulators.,scaledwidth=48.0% ] consider the curve for the nb@xmath11sn / insulator / nb@xmath11sn structure .
calculations show that for a homolaminate like this one a homolaminate being a sis structure in which the film is the same material as the substrate the film always reaches its @xmath1 before the bulk , and the thinner the film , the lower its @xmath1 .
the highest @xmath1 occurs when the film is so thick that it approximates a bulk superconductor .
this can be understood in terms of the forces on a vortex ( which can be derived from eqn .
[ eq : gibbs ] ) , as shown in fig .
[ fig : vortex ] .
@xmath23 , the meissner - screened external field , pushes the vortex into the film due to the gradient in the field .
the boundary condition imposed by @xmath24 , the magnetic field of the vortex , can be satisfied by an image antivortex outside of the boundary , which creates a force that pulls the vortex out of the film @xcite .
assuming that the insulating layers are very thin ( this is optimal , as will be shown later ) , then as eqn [ eq : bi ] shows , if @xmath36 , then the force due to @xmath23 is the same no matter how thick the film is .
however , there are two image antivortices contributing to @xmath24 : one for the external boundary , and one for the boundary at the insulating layer . for a thin film
, the image antivortex at the insulating layer will have a significant attractive force , pulling the vortex into the film . since this is the only change in the forces on the vortex as the film thickness changes
, a homolaminate will only have its @xmath1 decrease as the film becomes significantly smaller than a bulk .
note that this same logic applies no matter how many times the superconductor is divided a homolaminate sis multilayer with many thin film layers
will still experience a reduced @xmath1 .
only one curve in figure [ fig : bshvsdmulti ] is a homolaminate ; the rest are heterolaminates , in which the film material is different than the bulk . for a very thin heterolaminate
, the film does not provide much screening for the bulk , and @xmath21 reaches the bulk s @xmath1 before the thin film barrier disappears .
a very thick film behaves as a bulk , and reaches that material s bulk @xmath1 while @xmath21 is still relatively small .
however , there is a small range of film thicknesses where the film and the bulk can simultaneously be close to their respective @xmath1 . in this case , the @xmath1 of the sis structures is somewhat higher than the bulk value of the film material , but the potential gain in @xmath1 is relatively small , and the gain decreases as the thickness of the insulating layer increases .
the @xmath1 gain can be explained with a rough argument invoking the superfluid velocity @xmath37 : @xmath38 where @xmath39 and @xmath40 are the effective cooper pair mass and charge , @xmath41 is the magnetic vector potential , and @xmath42 is the gradient in the phase of the order parameter . if the superconductor is in the meissner state , @xmath43 and the magnitude of @xmath41 can be approximated by @xmath44 , where the integration starts deep in the bulk superconductor where @xmath45 and proceeds outwards .
if the bulk superconductor is a material with a relatively small @xmath5 , such as niobium , then at the film @xmath46 and therefore @xmath37 will be smaller than if the bulk material were a large-@xmath5 material , such as the alternative materials under consideration .
as a result , the superconductors in the sis structure are relatively `` unstressed , '' and can screen a larger external field .
this same argument explains why it is optimal for @xmath19 to be as small as possible this keeps @xmath46 as small as possible .
however , the gain in @xmath1 can not be multiplied by adding more films of the same material . as with the homolaminate , @xmath23 ( and the vector potential ) will be unchanged by splitting up the superconducting film into separate layers with thin insulators between them , but @xmath24 will have a stronger influence on the vortex , pulling it into the film .
this detrimental effect is visible in the last line of figure [ fig : bshvsdmulti ] , showing the calculation for a series of 5 nb@xmath11sn films , each with thickness @xmath47 , on a nb bulk , with thin insulating films between the superconductors .
note that for the material parameters used , which came from experimental references , the theoretical superheating fields of nbn and mgb@xmath6 are smaller than that of nb . in this study
, calculations were not performed for the ss structure , a superconducting film deposited directly onto a different superconductor , without an insulating layer .
the calculation would be significantly more difficult , but such a structure may offer a stronger performance improvement for appropriate choices of the thickness of the film layer and the superconducting materials .
gurevich states @xcite that `` thin film coating significantly decreases vortex dissipation at @xmath48 , '' where @xmath49 is the vortex penetration field .
this raises the question of whether it might be possible to use a sis structure above @xmath49 , in a regime in which vortices pass through the film each half cycle , bringing trapped flux into and out of the insulating region .
gurevich gives an expression for the power dissipated per area : @xmath50 where @xmath51 .
consider a @xmath52 nm nb@xmath11sn film in a sis structure with a nb@xmath11sn bulk and a very thin insulator , exposed to a field @xmath20 that is 1 mt higher than the film s @xmath1 . in this case , @xmath53 w/@xmath54 .
for a single cell tesla cavity , even this very small excess over @xmath1 generates approximately 4 kw of heat , far too much to be feasible for srf applications .
[ eq : pdiss ] shows , the power dissipated by vortex penetration is proportional to frequency . for low frequency ac and dc applications ,
vortex penetration could be tolerated . in this case ,
sis structures may offer a distinct advantage .
[ eq : vs ] shows , the superfluid velocity , a measure of the `` stress '' in a superconductor , is reduced in a large @xmath46 field by a gradient in the phase of the order parameter @xmath55 . even though the @xmath46 field at the film is tied to that of the bulk ( since the @xmath46 field is continuous across the insulating gap ) , the insulating gap does allow the opportunity to decouple @xmath55 in the film from that in the bulk .
this is accomplished by passing vortex lines through the superconducting film , which would incur large dissipation at high frequencies , but is permissable at low frequencies . with several layers , and an appropriate @xmath55 maintained in each of them to compensate for the @xmath46 field
, it should be possible to screen even very large fields while maintaining a relatively small @xmath37 in each of the layers , so that they remain in the meissner state .
the primary motivation for trying to fabricate sis structures was the promise of enhancing the @xmath0 of low-@xmath4 alternative materials .
it was suggested that above @xmath0 , dissipation would occur due to vortex dissipation @xcite .
there was some experimental data supporting this idea , as nb@xmath11sn cavities coated by wuppertal researchers showed increasing surface resistance with field , an effect that onset near the @xmath0 of nb@xmath11sn @xcite@xcite .
however , it was not clear if this performance degradation was fundamental and potentially related to vortex dissipation or if it was related to some other loss mechanism that could be ameliorated .
this year , a nb@xmath11sn cavity was coated and tested at cornell . vertical test data and fits to material parameters show that the cavity clearly exceeds @xmath0 without a signficant increase in surface resistance @xcite@xcite .
the @xmath0 value calculated from the material parameters was confirmed by @xmath56-sr measurements on a witness sample @xcite .
furthermore , a niobium cavity was prepared and tested at cornell after receiving a furnace treatment that gave it a very small mean free path .
as a result , the niobium had a very small @xmath4 , similar to that of the alternative materials under consideration .
it too had a performance with minimal increase in surface resistance , reaching fields significantly higher than the @xmath0 value determined from fits to material parameters @xcite@xcite .
in this paper , we have shown that contrary to suggestions that sis structures enhance @xmath0 , in fact they reduce it to zero .
in addition , it was shown that the @xmath1 of an sis structure is only marginally larger than the bulk value and only for a small parameter space , and that using a multilayer only decreases @xmath1 of the film .
it was also shown that sis structures exhibit unmanageable levels of heating above @xmath1 at high frequencies .
therefore , it seems that sis structures are not beneficial for srf applications
. however , they may be useful in dc and low frequency applications , where it should be possible to set up a gradient in the phase of the order parameter in the thin films , allowing them to screen very large fields . based on the results of this study
, the authors of this paper recommend that srf researchers developing alternative materials concentrate their efforts on bulk films .
bulk films are quite simple to fabricate compared to sis films , but they offer a similar ideal srf performance . and although we have shown that it is not possible to augment @xmath0 with sis structures , there is still great promise for alternative materials . because of the recent experiments showing that low - surface - resistance operation above @xmath0 is possible with cavities made from short coherence lengths superconductors , we now know that the potential of bulk films has not yet been realized . c. antoine , phys .
st - ab 13 , 12 ( 2010 ) .
t. tajima et al . , aip conf .
1435 , 297 ( 2012 ) .
d. beringer et al .
, ieee trans .
app . superc .
23 , 3 ( 2013 ) .
valente - feliciano et al . , thpo074 ( presented at the srf conference , chicago , usa , 2011 ) .
x xi 2009 supercond .
22 043001 t. proslier et al . , weioc03 ( presented at the srf conference , paris , france , 2013 ) .
d. oates et al .
, ( presented at the thin films srf workshop , newport news , usa , 2012 ) .
one definition of @xmath0 might be the field at which there is a stable position for a vortex inside a superconductor .
however , in our definition , the stable position for the flux might be inside the insulating region , in which case there will be no vortex , as a vortex has a normal conducting core in a superconductor . it will be shown in this paper that our definition is the relevant one , because if vortices move from the external region across the superconductor to become trapped flux in the insulating region , a very large amount of heating will occur . in other words , above @xmath0 as we have defined it , the film is in a metastable state , and only an energy barrier prevents strongly dissipative flux penetration .
these calculations use the london model , valid only for strongly type ii materials .
it neglects non - linearity , so at high fields , it is only valid qualitatively , and likely overestimates @xmath1 . in fact , @xmath1 calculated in this paper is the field at which the vortex penetration barrier disappears , an upper bound to the actual @xmath1 .
the latter is calculated by a linear stability analysis , similar to that done in @xcite for bulk superconductors , that will be presented elsewhere .
however the @xmath1 found in this paper does seem to be a good approximation : the @xmath1 for bulk nb@xmath11sn in fig .
[ fig : gplots ] is very close to the calculated value from tab . [
tab : mats ] . s. posen and m. liepe ,
tup087 ( presented at the srf conference , paris , france , 2013 ) .
m. liepe and s. posen , weioa04 ( presented at the srf conference , paris , france , 2013 ) .
a. grassellino et al . , tup029 ( presented at the srf conference , paris , france , 2013 ) .
d. gonnella and m. liepe , tup027 ( presented at the srf conference , paris , france , 2013 ) . for calculations of @xmath0 from material parameters ,
ginzburg landau theory is applied outside its validity , so corrections on the order of a few percent will apply .
for example , using gl theory at low temperatures is known to underestimate the bulk superheating eld ( see catelani and sethna , prb 78.224509 ) . | the sis structure a thin superconducting film on a bulk superconductor separated by a thin insulating film was propsed as a method to protect alternative srf materials from flux penetration by enhancing the first critical field @xmath0 . in this work ,
we show that in fact @xmath0 = 0 for a sis structure .
we calculate the superheating field @xmath1 , and we show that it can be enhanced slightly using the sis structure , but only for a small range of film thicknesses and only if the film and the bulk are different materials . we also show that using a multilayer instead of a single thick layer is detrimental , as this decreases @xmath1 of the film .
we calculate the dissipation due to vortex penetration above the @xmath1 of the film , and find that it is unmanageable for srf applications .
however , we find that if a gradient in the phase of the order parameter is introduced , sis structures may be able to shield large dc and low frequency fields .
we argue that the sis structure is not beneficial for srf cavities , but due to recent experiments showing low - surface - resistance performance above @xmath0 in cavities made of superconductors with small coherence lengths , we argue that enhancement of @xmath0 is not necessary , and that bulk films of alternative materials show great promise . |
though the standard model ( sm ) has been verified to be correct times by times , many new physics beyond standard model are proposed to solve both experimental and aesthetical problems , such as neutrino masses , @xmath1 anomalous magnetic movement problem or hierarchy problem , etc .
many new models introduce vector like particles ( vlp ) @xcite whose right handed and left handed components transform in the same way under the weak su(2)@xmath2u(1 ) gauge group .
the extension is acceptable because the anomalies generated by the vlps cancel automatically , and vector quarks can be heavy naturally .
vlps also arise in some grand unification theories .
for example , in order to explain the little hierarchy problem between the traditional gut scale and string scale , a testable flipped @xmath3 model are proposed in ref .
@xcite in which the tev - scale vlps were introduced @xcite .
such kind of models can be constructed from the free fermionic string constructions at the kac - moody level one @xcite and from the local f - theory model @xcite .
however when we do the flavor physics with doublet vlps in these models @xcite , a problem always appears when we are dealing with the mass spectrum of quarks and leptons .
let us start with the sm in which all fermion masses come from the yukawa couplings .
after the spontaneously gauge symmetry breaking , we can get two separate mass matrices @xmath4 for the up and down type fermions .
the mass eigen states are obtained after the diagonalization @xmath5 where @xmath6 $ ] , @xmath7 $ ] .
the physical measurable parameters are @xmath8 and the so called ckm matrix @xmath9 since @xmath4 come from separate yukawa couplings , we can always set one of the matrices diagonal , for example @xmath10 , and use the ckm matrix to get the yukawa couplings @xmath11 for the calculation in flavor physics . note that @xmath12 is the vacuum expectation value ( vev ) of the higgs , and @xmath13 is a random unitary matrix
. such a trick can not be used in case of the participation of a vector doublet , namely @xmath14 with gauge charge @xmath15 and @xmath16 with gauge charge @xmath17 , resulting bilinear term in the lagrangian @xmath18 it is clear that in the model , there are the same input parameters in the matrices @xmath4 @xmath19 the mass matrices for up and down type quarks are related to each other .
therefore , we can not set one of the matrices diagonal and the ckm matrix can not be got easily .
the shooting method is always used to treat such an obstacle .
random @xmath10 and @xmath20 are generated to meet the requirements after diagonlization : the mass of eigen state and the measurements of elements of ckm matrix .
however this is too much time consuming , and precise solution for diagonalization is almost unavailable .
although this is just a numerical problem , when one treats the vlp contributions to the flavor physics seriously , diagonalization of quark matrices will be the first and important step . in this paper
, we will first propose a general method to solve the obstacle in models with vector like quark doublets .
as its application , we will study rare b decay @xmath0 in the sm with one vector like quark doublet .
the paper is organized as follows .
we show the detail of the trick in section 2 .
the simple application to @xmath0 process , including quark mass spectrum , feynman rules and the wilson coefficients , as well as the numerical analysis for calculation of @xmath0 is shown in section 3 .
a summary is given in section 4 .
firstly , we address the problem clearly on how to deal with the diagonalization of @xmath21 matrix @xmath10 and @xmath20 : @xmath22 in which @xmath23 are the diagonal mass matrices for up and down type quark , respectively .
note that @xmath24 should be greater than 3 and the first three elements in the matices should be the three generations of quark multiplates in the sm , other elments with @xmath25 are the new multiplates introduced in new physics beyond the sm
. then we have @xmath26 the last line of the two matrices has the same parameters except the last elements . considering that there are some same parameters in @xmath10 and @xmath20 , we find that a very simple way is to add two matrices in eq .
( [ eq : mvud ] ) @xmath27 the left side of the equation is @xmath28 obviously , the mass inputs from bilinear terms vanish .
we can denote the matrix in the form as @xmath29 in which @xmath30 , @xmath31 , @xmath32 are @xmath33 , @xmath34 and @xmath35 matrices correspondingly . to prepare for the diagonalization , we chose the diagonal mass matrix elements of quarks @xmath36 , @xmath37 and a matrix @xmath38 , which are determined partly by experimental measurements as input parameters @xmath39 note that above @xmath40 are unitary matrices , but @xmath41 is not an ordinary ckm matrix @xmath42 which is non - unitary in this case .
detailed dicussion will be shown in the following section .
what we need to do for the next is to generate a unitary matrix @xmath43 . in the similar way we denote @xmath43 as @xmath44 both sides of eq .
( [ eqvud ] ) times the matrix @xmath43 , we can get @xmath45 from above equation , we can get the last line of @xmath43 simply by inputting @xmath46 and random @xmath47 , @xmath13 : @xmath48 where @xmath49 is a unit vector in @xmath24 dimension .
next we use the unit vector to generate total @xmath43 . since @xmath30 and @xmath31 are random matrix , @xmath43 can be random too .
the unit vector @xmath50 of @xmath43 can be determined as @xmath51 it is clear that the vector is orthogonal to @xmath52 and normalized to 1 .
then we use the first three elements of @xmath53 and @xmath50 to generate @xmath54 : normalize the algebraic complements of first line of the @xmath55 matrix .
step by step , we can finally get ( @xmath56 , @xmath57 , @xmath58 , @xmath50 ) and form a special @xmath59 @xmath60 from above steps , we can see that ( @xmath56 , @xmath61 , @xmath58 , @xmath50 ) can be rotated into any other orthogonal @xmath62 vectors to construct random matrix @xmath30 and @xmath31 , only @xmath52 must be kept unchanged .
therefore , a general unitary matrix can be realized by timesing a unitary @xmath21 matrix @xmath63 , @xmath64 in which @xmath65 is a @xmath66 unitary matrix .
we finish the work by @xmath67 at this stage , we would like to summarize our method here * step 1 : chose ( @xmath68 ) and @xmath38 and generate random unitary matrices @xmath47 and @xmath13 as the inputs for the model ; * step 2 : determine the last line of matrix @xmath69 as @xmath70 and normalize it into a unit vector @xmath71 .
* step 3 : use the unit vector @xmath71 to generate other @xmath62 unitary vectors ( @xmath56 , @xmath61 , @xmath58 , @xmath50 ) , and form a special @xmath72 @xmath73 * step 4 : generate a @xmath62 unitary matrix @xmath65 to form a unitary matrix @xmath63 which is @xmath74 then , a general @xmath43 is obtained by @xmath75 * step 5 : use these equations @xmath76 to get the inputs for the flavor physics .
we can see that by this trick we can skip the inputs of the bilinear mass terms @xmath77 . in physical analysis ,
the mass of eigen states @xmath78 in the vlp models are inputs freely . @xmath79 and @xmath80
can be generated randomly , @xmath81 and @xmath43 can also be scanned the most generally if we vary @xmath63 randomly .
thus the method can do the most general scan in the parameter space of mass matrices in the models with vlps for the numerical studies , which will be shown in the following section .
cc .a simple extension of the standard model with one vector like quarks doublet [ cols="^,^",options="header " , ] in the model with three generation quarks , the ckm matrix unitarity is already used in the calculations of the loop - level fcnc induced rare b decays . for consistency , in numerical analysis
the constraints on ckm matrix element are not from processes occurred at loop level , such as rare b decays , but from tree - level processes shown in table [ tab : c1b ] @xcite .
since there are no tree - level measurements of @xmath82 , @xmath83 now , we use above inputs and the unitarity to get @xmath84 unitary matrix at first .
the method is that we scan @xmath85 randomly ( keeping @xmath86 ) in range listed in table [ tab : c1b ] , then we define two parameters @xmath87 and solve them by the equations @xmath88 @xmath89 are got by the unitarity relation with @xmath85 and @xmath90 .
after that we times the @xmath84 unitary matrix with three matrices @xmath91 in which @xmath92 and @xmath93 , to generate a @xmath94 unitary matrix @xmath95 .
@xmath96 are got by the eq .
( [ eq : vkm ] ) .
all the corresponding elements should satisfy the experiment bound list in table [ tab : c1b ] and @xmath82 , @xmath83 ( @xmath97 which is consistent with the fitting results in ref .
@xcite ) can be got too . with this
inputs in hand , the first task is to check the scale of the mass parameter of model , such as @xmath98 , @xmath99 . from the @xmath100 vertexes in eq .
( [ ud_z1 ] ) and eq .
( [ ud_z2 ] ) , we can see that in order to keep gauge universality of quarks , the tail terms in the feynman rules must be much smaller than the sm like terms , namely @xmath101 .
thus in the numerical studies we require @xmath102 note that though these elements are greater than @xmath103 ( parameter in the wolfenstein parameterization @xcite ) , they are much smaller than the product of @xmath104 ( almost equals @xmath105 ) , thus the requirements are suitable for indicating the contraints from the deviation from unitarity . since the scanning in the parameter space is freely , we set @xmath106 ( mass of top quark plus 1000 gev ) and scan @xmath107 in the range of @xmath108 ( mass of bottom quark plus 1000 gev ) , and @xmath109 randomly ( ignoring the cp phases ) .
@xmath110 is defined by @xmath111 the result for @xmath110 versus @xmath112 is shown in the fig .
[ fig2 ] which checks the mass input of vector doublet .
we can see that @xmath110 increases as @xmath107 growing up .
however @xmath110 is much smaller than @xmath113 and @xmath107 .
small mixings lead to parameter @xmath98 which determine the mixing between sm quarks and vector like quarks are also suppressed .
this is in agreement with that the deviation from unitarity is suppressed by the ratio @xmath114 where @xmath115 denotes generically the standard quark masses , which is a typical result of vlp models .
@xcite the second task is to check the vlp contribution to @xmath0 .
we find the wilson coefficient of fcnc operator @xmath116 is not so suppressed as the mixing .
the new contributions , from the terms of the last line in eq.([fc7 ] ) are suppressed by the mixing , whereas the terms of first line are almost the same as the sm .
the enhancement comes mainly from the @xmath117 terms which are got from the goldstone loop in @xmath118 transition .
( the right diagram in the first line and diagrams in the second line of the fig .
[ fig1 ] ) in order to show the enhancement clearly , we define two factors @xmath119 in which @xmath120 denotes the deviation from the unitarity of @xmath55 ckm matrix , while @xmath121 shows the enhancement of the contribution from vector like particles .
@xmath121 is in fact got from the coefficient of first term in the second line of analytical expression of @xmath122 in eq .
( [ fc7 ] ) when @xmath123 .
it will be changed into exactly @xmath120 in case of the sm4 .
note that other terms with @xmath117 can give enhancement too , we chose factor @xmath121 for a typical demonstration since it seems that it will be suppressed by @xmath113 .
results are shown in the fig [ fig3 ] in which the left panel shows @xmath121 and @xmath120 versus @xmath110 while the right panel shows @xmath124 versus @xmath121 . from the left panel , we can see that though @xmath120 increase as @xmath110 increases , it is still much smaller than @xmath125 , implying that deviation of unitarity are negligible .
however the factor @xmath121 can be enhanced up to order @xmath126 by the increase of @xmath110 . from the right panel
, we can see that @xmath121 enhances @xmath127 up to a value much larger that the result of the sm .
the reason for the enhancement mainly comes from the new type of yukawa couplings .
combining of eqs .
( [ massdiag ] , [ ud_yl ] , [ ud_yr ] ) , one can get similar form compared with the sm4 @xmath128,\\ \frac{g^g_r(4,3)}{m_{b } } & = & -u_{ckm4}^{\ast 43 } \label{eqgrvb } -\frac{1}{m_{b } } \left[\sum_{m=1}^3 \left(m_{q}^{\ast m}u_{u}^{\ast m4}z_{d}^{43}+m_d^{\ast m}u_{u}^{\ast44}z_{d}^{m3}\right ) + ( y^*_{vd}-y_{vu}^\ast)u_{u}^{\ast44}z_{d}^{43}v\right].\end{aligned}\ ] ] since @xmath129 , @xmath130 , one can easily obtain that @xmath131 the suppression of @xmath132 ( order of @xmath114 ) in eq .
( [ eqgrvb ] ) are enhanced by terms with factor such as @xmath133 , etc .
, resulting @xmath134 thus the term @xmath135 satisfying the unitary constraint @xmath136 is enhanced greatly by heavy vlps , then the factor leads the enhancement to @xmath127 .
this is different from those in the sm4 in which the contribution from the fourth generation can be neglected . in the numerical scan ,
we vary @xmath137 and @xmath138 randomly , keeping the constraints of @xmath139 , scan @xmath113 and @xmath107 in the range of @xmath140 . apart from the ckm limits
, we use the @xmath141 process to constrain parameter space .
the branching ratio of @xmath0 is normalized by the process @xmath142 : @xmath143.\label{bsg}\ ] ] here @xmath144 , and @xmath145 is the phase - space factor in the semi - leptonic b decay .
the method of running of the operators from @xmath146 scale to @xmath147 scale can be found in ref .
we use the following bounds on the calculation @xcite @xmath148 the numerical results show that the @xmath149 is much smaller than @xmath150 , therefore we do not present the formula of @xmath151 here . the branching ratio as a function of @xmath152 is shown in fig .
[ fig4 ] , from which we can see that @xmath153 can be enhanced much greater than the experiment bound .
then the measurements of fcnc process can give a stringent constraint on the vector like quark model , especially when the masses of vector quark are much greater than the electro - weak scale .
a few remarks should be addressed : * there is one point of view on the unitarity of the ckm matrix which is that the @xmath154 ordinary quark mixing matrix is regarded as nearly unitary , deviation from unitarity is suppressed by heavy particle in the new physics beyond the sm .
in other word , one admits that the extended ckm matrix elements exist , they approach to zero while mass scale of the new physics approaches to infinity .
all the new physical effects should decouple from the flavor sector and what should be checked is that if @xmath55 unitariry is consistent in all kinds of flavor processes .
* another point of view is that , as in the sm case , the @xmath154 ordinary quark mixing matrix elements are only extracted by experiments in the measurements of tree and loop level precesses .
the unitarity should be checked , experiment measurements on the elements of matrix can be used as the constraints to the new physics beyond the sm . in the numerical analysis ,
the elements of ckm matrix are regarded as inputs .
thus what should be done is to scan the parameter space generally under these constraints , no prejudice should be imposed .
then the enhancement effect in @xmath0 will be more clear .
large parameter space is excluded by the measured branching fraction of @xmath141 as shown in fig .
the enhancement effect of the vlps can be seen in fig .
[ fig5 ] in which the left panel shows the enhancement factor @xmath121 versus @xmath113 while the right panel shows @xmath120 versus @xmath113 . from the right panel we can see that deviation from unitarity are very small and almost irrelevant with @xmath113 since we are doing a general scan of @xmath137 and @xmath138 .
however as we see from the left panel , as @xmath113 increases up , @xmath155 measurement will constrain the enhancement factor and then constrain the input parameter of @xmath113 . in all , the enhancement can be summarized as that when mass of vector like particle increases up , it will increase the mass parameter @xmath152 thus give an enhancement factor under very small deviation from unitarity
. this should be a special point when we do the study on the vector like quark models .
in the model with vector doublets , there exist bilinear terms in the lagrangian , making the general scan of the yukawa coupling very difficult . in this paper ,
we show a trick to deal with the scan .
our scan method are exactly and the more efficient .
we use the trick to study a very simple extension of the sm with vector like quarks .
we studied one of the most important rare b decay @xmath141 process in which we found that even the deviations from the unitarity of quark mixing matrix are small , the enhancement to rare b decay from vlps are still significant .
the enhanced effect is an important feature in the vector like particle model . in this work
we just show the scan method , the key point of the enhancement and how stringent constraints on the parameter space from @xmath141 measurements .
what should be done includes models like extension of the sm with vlps , two higgs doublets models @xcite or supersymmetry models @xcite .
such effect should be checked in all kinds of rare decays such as inclusive process @xmath156 and exclusive processes @xmath157 , @xmath158 and @xmath159 mixing _ et .
al . _ the detailed studies on the parameter space including other rare b decays and new models will appear in our future work .
this work was supported by the natural science foundation of china under grant numbers 11375001 and by talents foundation of education department of beijing .
* the loop functions for calculating the wilson coefficients at the matching scale are the following g. cvetic , rev .
phys . * 71 * , 513 ( 1999 ) [ hep - ph/9702381 ] ; h. j. he , t. m. p. tait and c. p. yuan , phys .
d * 62 * , 011702 ( 2000 ) [ hep - ph/9911266 ] ; h. j. he , c. t. hill and t. m. p. tait , phys .
d * 65 * , 055006 ( 2002 ) [ hep - ph/0108041 ] .
j. cao , l. shang , w. su , f. wang and y. zhang , arxiv:1512.08392 [ hep - ph ] .
f. wang , w. wang , l. wu , j. m. yang and m. zhang , arxiv:1512.08434 [ hep - ph ] .
j. jiang , t. li , d. v. nanopoulos and d. xie , phys .
b * 677 * , 322 ( 2009 ) ; nucl .
b * 830 * , 195 ( 2010 ) .
j. jiang , t. li and d. v. nanopoulos , nucl .
b * 772 * , 49 ( 2007 ) .
i. antoniadis , j. r. ellis , j. s. hagelin and d. v. nanopoulos , phys .
b * 208 * , 209 ( 1988 ) [ addendum - ibid .
b * 213 * , 562 ( 1988 ) ] ; phys .
b * 231 * , 65 ( 1989 ) .
j. l. lopez , d. v. nanopoulos and k. j. yuan , nucl .
b * 399 * , 654 ( 1993 ) . c. beasley , j. j. heckman and c. vafa , jhep * 0901 * , 058 ( 2009 ) ; jhep * 0901 * , 059 ( 2009 ) ; r. donagi and m. wijnholt , adv .
phys . * 15 * , no . 5 , 1237 ( 2011 ) [ arxiv:0802.2969 [ hep - th ] ] ; adv .
* 15 * , no .
6 , 1523 ( 2011 ) [ arxiv:0808.2223 [ hep - th ] ] .
t. li , d. v. nanopoulos , w. y. wang , x. c. wang and z. h. xiong , jhep * 1207 * , 190 ( 2012 ) [ arxiv:1204.5326 [ hep - ph ] ] .
t. li , w. wang , x. c. wang and z. h. xiong , arxiv:1506.04987 [ hep - ph ] .
a. j. buras , m. misiak , m. mnz and s. pokorski , nucl .
b * 424 * , 374 ( 1994 ) .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * , 010001 ( 2012 ) .
o. eberhardt , a. lenz and j. rohrwild , phys .
d * 82 * , 095006 ( 2010 ) [ arxiv:1005.3505 [ hep - ph ] ] .
l. wolfenstein , phys .
lett . * 51 * , 1945 ( 1983 ) .
a. j. buras , m. e. lautenbacher and g. ostermaier , phys .
d * 50 * , 3433 ( 1994 ) [ hep - ph/9403384 ] .
j. charles _ et al . _
[ ckmfitter group collaboration ] , eur .
j. c * 41 * , 1 ( 2005 ) [ hep - ph/0406184 ] .
p. langacker and d. london , phys
* d38 * ( 1988 ) 886 .
f. j. botella , g. c. branco and m. nebot , jhep * 1212 * , 040 ( 2012 ) [ arxiv:1207.4440 [ hep - ph ] ] ; a. k. alok , s. banerjee , d. kumar , s. u. sankar and d. london , phys .
d * 92 * , no . 1 , 013002 ( 2015 )
[ arxiv:1504.00517 [ hep - ph ] ] ; a. k. alok , s. banerjee , d. kumar and s. uma sankar , nucl .
b * 906 * , 321 ( 2016 ) [ arxiv:1402.1023 [ hep - ph ] ] .
b. grinstein , r. p. springer and m. b. wise , nucl .
b * 339 * , 269 ( 1990 ) .
w. altmannshofer , a. j. buras , s. gori , p. paradisi and d. m. straub , nucl .
b * 830 * , 17 ( 2010 ) [ arxiv:0909.1333 [ hep - ph ] ] . | in the models with vector like quark doublets , the mass matrices of up and down type quarks are related .
precise diagonalization for the mass matrices becames an obstacle in the numerical studies . in this work
we propose a diagonalization method at first . as its application , in the standard model with one vector like quark doublet we present quark mass spectrum , feynman rules for the calculation of @xmath0 .
we find that i ) under the constraints of the ckm matrix measurements , the mass parameters in the bilinear term are constrained to a small value by the small deviation from unitarity ; ii ) compared with the fourth generation extension of the standard model , there is an enhancement to @xmath0 process in the contribution of vector like quark , resulting a non - decoupling effect in such models . |
the jaynes - cummings model ( jcm ) , proposed in 1963 @xcite , constitutes an excellent theoretical approach to describe analytically the interaction of a two level atom with a single mode of a quantized radiation field .
the field frequency may belong either to the optical domain or to the microwave one . in the first case the researchers use common atoms @xcite whereas in second case they use ( highly excited ) rydberg atoms @xcite .
the issue was also extended to other systems , as ( i ) in nanocircuits operating in microwave domain , either through the substitution of the atom by a copper - pair box ( cpb ) and the field by a nanomechanical resonator in nanocavities @xcite ; ( ii ) or the cpb inside a chip @xcite ; ( iii ) substituting the atom by quantum a dot embedded in a photonic - crystal ad ; ( iv ) using spin in quantum - dot arrays @xcite , etc . in spite of its simplicity
the jcm gives exact solutions of the schrdinger equation in many examples that occur in such physical systems .
the jcm has been employed in the study of various fundamental quantum aspects involving the matter - radiation . to give some examples we mention : collapse and revival of the atomic inversion @xcite ; the rabi frequency of oscillation for a given atomic transition acted upon by a light field @xcite ; nonclassical statistical distributions of light fields @xcite , antibunching effect @xcite ; squeezed states @xcite , and others .
an alternative model that maintains various characteristics of the jcm and offers advantages in certain situations was proposed by buck - sukumar in 1981 , abbreviated as bsm @xcite .
it is called intensity - dependent jcm , since it substitutes the jcm interaction @xmath0 by another interaction that includes the number operator @xmath1 in this way : @xmath2 with @xmath3 and @xmath4 in the previous expressions @xmath5 ( @xmath6 ) stands for annihilation ( creation ) operator , @xmath7 ( @xmath8 ) is lowering ( raising ) operator , ( @xmath9 ) is the number operator , and @xmath10 stands for the atom - field coupling .
this model also leads to analytical solution of the schrdinger equation .
it has been argued that its physical simulation in laboratory could be implemented via matrices of waveguides @xcite ; optical analogies of quantum systems realized in waveguide arrays have recently impacted the field of integrated optical structures @xcite . in particular , susy photonic lattices can be used to provide phase matching conditions between large number of modes allowing the pairing of isospectral crystals @xcite . in spite of its apparent theoretical nature
the bsm has attracted the attention of various researchers in the quantum optical community .
shanta , sivakumar , ahmed , rlm , yang , cardimona , yang1,yang2,dukelsky , reeta , buzek , valverde , sayed , cordeiro . in 1992 p. shanta ,
s. chaturvedi , and v. srinivasana ( scs - model ) proposed an extension of the intensity - dependent jcm @xcite .
this model interpolates between the jcm and the bsm . in this approach
the authors assumed the modified hamiltonian , @xmath11 where @xmath12 is the number operator and the operators @xmath5 , @xmath6 are quons operators satisfying the the commutation relation @xmath13 ; @xmath14 is a c - number restricted to the interval @xmath15.$ ] accordingly , quons would stand for particles intermediate between bosons ( @xmath16 ) and fermions ( @xmath17 ) .
the authors then use specific connections between the operator @xmath12 and @xmath5 and @xmath6 and prove that the scs model interpolates between the bsm and jcm in the limits @xmath16 and @xmath18 respectively , with @xmath14 playing the role of the interpolating parameter . however , although being a creative approach , here we will not take it forward because we are restricting ourselves to photonic field , not to quons .
according to ref .
@xcite there are other nonlinear models in this context , but they treats the coupled system only approximately vad another type of intensity - dependent jcm , was proposed in 2002 by s. sivakumar @xcite , named here as sivakumar model ( sm ) .
this model also interpolates between the jcm and the bsm via the following hamiltonian , @xmath19 where @xmath20 and @xmath21 stand respectively for annihilation @xmath22 and creation @xmath23 operators .
the change from @xmath5 to @xmath24 aims to get a convenient deformed algebra for various theoretical applications , as in group theory , field theory , and others .
as established in @xcite , for @xmath25 one has the heisenberg - weyl algebra generated by @xmath26 and for @xmath27 one finds the @xmath28 algebra . for _ _ _ _ all values of @xmath29 the algebra is closed , @xmath30=2\hat{k}_{0},\text { } [ \hat{k}_{0},\hat{k}^{\dagger } ] = k\hat{k}^{\dagger } , \text { } [ \hat{k}_{0},\hat{k}]=-k\hat{k } , \label{asm2}\ ] ] with @xmath31 we note some resemblance between the hamiltonian in eq .
( [ asm1 ] ) and that given by the bsm for @xmath32 as pointed out by the authors , the bsm is only reached when the mean photon number of the field satisfies the condition @xmath33 @xmath34 leading the term @xmath35 to an approximate form of bsm @xmath36 @xmath37 a somewhat ` similar ' model , also intensity - dependent , was proposed in 2014 by rodrgues - lara @xcite , named here as rodrgues - lara model ( rlm ) , constituting a generalization of bsm since it substitutes the operator @xmath38 of the bsm by the operator @xmath39 .
the rlm recovers the bsm in the limit @xmath40 but it includes the counter - rotating terms , due to the form of the interaction hamiltonian , @xmath41 where the decomposition @xmath42 explains the appearance of counter - rotating terms @xmath43 and @xmath44 as well known , separately they do not conserve energy . also , due to the inclusion of the counter rotating terms , this model puts a restriction on the average number of photons . in this report
we present a generalized hamiltonian that provides a continuous and exact interpolation between various hamiltonian models , including the jcm , bsm , sm , and rlm .
the plan of the paper is as follows . in sec .
[ sec - model ] we briefly discuss this class of hamiltonian , showing its interpolating property . in sec .
[ basic ] we obtain the solution of the schrdinger equation in this extended scenario . in sec .
sec - apli we give some applications , in the sec .
[ c1 ] we calculate mandel parameter .
[ cc ] contains comments and the conclusion .
the hamiltonian described by the jcm , widely referred to as the jcm in the rotating wave approximation , is given in the form , @xmath45where @xmath46 stands for the field frequency , @xmath47 is the atomic frequency , and @xmath48 stands for atom - field coupling .
now , our mentioned class of interpolating hamiltonians is obtained substituting @xmath49 by @xmath50 given by @xmath51 where @xmath52 and @xmath53 for @xmath54 and @xmath55 here it is easily seen that the hamiltonian in eq.([a1a ] ) interpolates between the various interaction models of hamiltonians , as follows : * the jaynes - cummings model ( jcm ) @xcite for @xmath56 and @xmath57 * the buck - sukumar model ( bsm ) @xcite for @xmath58 and @xmath59 * the sivakumar model ( sm ) @xcite for @xmath60 and @xmath57 * the rodrgues - lara model ( rlm ) @xcite for @xmath58 and @xmath61 some basic properties involving these atomic and field operators are , @xmath62 & = & \pm 2\hat{\sigma}_{\pm } , \text{\ } [ \hat{\sigma}_{+},\hat{\sigma}_{-}]=\hat{\sigma}_{z } , \label{cc1 } \\ \lbrack \hat{a},\hat{a}^{\dagger } ] & = & 1,\text { } [ \hat{a},\hat{n}]=\hat{a},\text { } [ \hat{a}^{\dagger } , \hat{n}]=-\hat{a}^{\dagger } , \end{aligned}\ ] ] @xmath63=\hat{r},\text { } [ \hat{r}^{\dagger } , \hat{n}]=-\hat{r}^{\dagger } , \text { } [ \hat{r},\hat{r}^{\dagger } ] = 2\hat{r}_{0}=\delta + \xi + 2\xi \hat{n } , \label{cc3}\ ] ] with @xmath64 thus we have a closed algebra in this scenario , @xmath65=2\hat{r}_{0},\text { } [ \hat{r}_{0},\hat{r}^{\dagger } ] = \xi \hat{r}^{\dagger } , \text { } [ \hat{r}_{0},\hat{r}]=-\xi \hat{r}.\ ] ] the eq.([a1a ] ) can be rewritten in the form , @xmath66 where , @xmath67 with @xmath68 @xmath69 next we can use the eqs.(cc1 ) and ( [ cc3 ] ) to show that @xmath70 and @xmath71 are constant of motion , namely , @xmath72=[\mathscr{h},\hat{\mathscr{h}}_{i}]=[\hat{\mathscr{h}}_{a},\hat{\mathscr{h}}_{i}]=0.\]]all essential dynamic properties contained in a state of the atom - field system described by any of the previous interpolating hamiltonians , can also be described by the interpolating hamiltonian proposed here , @xmath71 , considering that @xmath70 contributes only for general phase factors , usually not relevant .
let us consider a simple example assuming the system in resonance , @xmath73@xmath74now , to analyze the time evolution of the coupled atom - field system we solve the time dependent schrdinger equation using the hamiltonian in eq.(ac ) , @xmath75 we can write the formal solution of eq .
( [ a3a ] ) as , @xmath76 where @xmath77 , is the ( unitary ) evolution operator .
next , using the expression @xmath78and decomposing the above sum in their even and odd terms , plus the use of the two following relations @xmath79 and , @xmath80with @xmath81 , we get the evolution operator in a convenient form for systems involving a two - level atom,@xmath82with @xmath83 and @xmath84 , @xmath85 given above
. we will assume the entire system initially decoupled , @xmath86 the atom in its ground state @xmath87 and the field in arbitrary state @xmath88 so , the wavefunction describing the atom - field system for arbitrary times is obtained from equation @xmath89 with @xmath90 given in eq .
( [ alfa ] ) .
after an algebraic procedure we find , @xmath91 paradoxical evolution of average number of photons[sec - apli ] ------------------------------------------------------------ the paradox concerned with the time evolution of the average number of photons , discussed by luis @xcite , used the jcm . here we treat this paradox for the various interpolating hamiltonians mentioned above .
this is obtained directly from our hamiltonian by an appropriate choice of the pair @xmath92 and @xmath93 .
the mean number of photons of the field is calculated as,@xmath94where @xmath95 _ _ _
_ is the density operator . in this section
we study the dynamic behavior of the average number of photons , @xmath96 @xmath97 and @xmath98 where @xmath99 , \label{ng}\ ] ] @xmath100 , \label{ne}\ ] ] @xmath101 where @xmath102 @xmath103 @xmath104 @xmath105 regardless of the types of interpolations , i.e. , jcm @xmath106 bsm@xmath106 sm @xmath106 rlm , and eventually others obtained by varying the pair @xmath92 and @xmath93 _ _ , _ _ the essential features of the paradox discussed in ref .
@xcite remains for all these interpolation models .
now , for small times , the following relation is valid , irrespective of the interpolating model .
@xmath107 in the plots of fig .
( [ figura_1 ] ) we have assumed the initial field in a coherent state , assuming the average number of photons @xmath108 here we have used mathematical expressions more general than those in eqs .
( [ ng ] , [ ne ] and [ n ] ) , hence the following plots are not restricted to small times__. _ _ we observe in fig .
( figura_1 ( a ) ) the occurrence of the mentioned paradox , which starts immediately and remains up to @xmath109 for the jcm ; in fig . ( [ figura_1 ] ( b ) )
@xmath110 for the bsm ; in fig .
( figura_1 ( c ) ) @xmath111 for the sm ; and in fig .
( figura_1 ( d ) ) @xmath112 for the rlm .
( solid curve ) @xmath113 ( dashed curve ) and @xmath114 ( dotted curve ) , for an initial coherent state with @xmath115 a ) jcm , b ) bsm , c ) sm and d ) rlm.,width=302,height=264 ] hence , these results show that the paradox raised by _ _
luis luis using the jcm happens no matter what kind of hamiltonian model used within the class considered here .
a quantized photon field with sub - poissonian statistics is characterized when the variance is smaller than the average number of photons , namely : @xmath116 @xmath117 the opposite chacterizes a super - poissonian photon field and if @xmath118 the photon field exhibits poissonian statistics , characterizing all coherent states .
the mandel s parameter tells us what kind of statistics the field displays @xcite ; it is given by the relation,@xmath119 so , when @xmath120 the field is super - poissonian ; when @xmath121 it is sub - poissonian ; and poissonian for @xmath122 . in fig .
( [ figura_2 ] ) , we represent our hamiltonian in eq .
( [ cv2 ] ) interpolating between the four hamiltonians : jcm , bsm , sm and rlm .
, for an initial coherent state with @xmath115 a ) jcm , b ) bsm , c ) sm and d ) rlm.,width=302,height=264 ] fig.([figura_3 ] ) exhibits various plots of the mandel parameters in these different models of hamiltonian .
the various plots show that , by conveniently adjusting the pair of parameters @xmath92 and @xmath93 in the present model hamiltonian we can interpolate continuously from the jcm to the bsm , the sm , and the rlm . in these interpolations we have observed in which way the mandel parameter modifies during the time evolutions , as shown in fig.([figura_3 ] ) ,
plots ( a ) , ( b ) , and ( c ) ; also , this interpolation occurs in a softly way , from the jcm to bsm .
the same happens for the interpolation from the jcm to the sm , shown in fig.([figura_3 ] ) , plots ( d ) , ( e ) , and ( f ) ; and also from the jcm to the rlm , fig.([figura_3 ] ) , plots ( g ) , ( h ) , and ( i ) . for an initial coherent state with @xmath123 ; interpolating from jcm to bsm a ) for @xmath124 and @xmath125 b ) for @xmath126 and @xmath127 c ) for @xmath128 and @xmath129 ; interpolating from jcm to sm d ) for @xmath130 and @xmath131 e ) for @xmath126 and @xmath132 ; f ) for @xmath133 and @xmath132 ; interpolating from jcm to rlm g ) for @xmath130 and @xmath134 ; h ) for @xmath126 and @xmath135 i ) for @xmath133 and @xmath136 , for an initial coherent state ; a ) with @xmath115 para @xmath128 and @xmath137 b ) with @xmath115 para @xmath128 and @xmath138 c ) with @xmath115 para @xmath128 and @xmath135 d ) with @xmath139 para @xmath128 and @xmath137 e ) with @xmath139 para @xmath128 and @xmath138 f ) with @xmath139 para @xmath128 and @xmath140 ] we can note in fig.([figura_4 ] ) that , when we compare the case where the system state has a small average excitation @xmath141 with those having larger values of @xmath141 , the mandel parameters for different hamiltonians differ sensitively from each other for small values of @xmath141 , the region where the quantum nature of the system state is more evident .
contrarily , for larger values @xmath141 the corresponding plots are very similar . in theses
examples we are analizing the mandel parameter close to bsm , fig.(figura_4 a ) and d ) ) , with other close to sm , fig.([figura_4 ] b ) , and e ) ) , and another close to rlm , fig.([figura_4 ] c ) and f ) ) .
this shows a great sensitivity of the system to the parameters @xmath92 and @xmath93 in the quantum regime of small numbers , as usually expected .
in addition , for small values of @xmath141 the field state exhibts a greater sub - poissonian effect .
we have proposed a ( two parameters ) interpolating hamiltonian .
it allows one to extend from ( a ) the jcm , ( b ) the bsm , ( c ) the sm , and ( d ) the rlm .
this new hamiltonian employs the basic operators @xmath52 , @xmath142 , and @xmath143 which form a closed algebra . as mentioned before
, it contains all essential dynamic properties contained in a state of the atom - field system described by the previous interpolating hamiltonians . to give an example we have verified that , essentially , the results found in the paradox discussed by a. luis @xcite in the jcm remains in the scenario of this extended hamiltonian ( see fig .
( [ figura_1 ] ) ) , no matter the chosen extension , say : from ( a ) to ( b ) , from ( a ) to ( b ) , from ( a ) to ( c ) , and from ( a ) to ( d ) .
we have also calculated the mandel parameter to obtain the evolution of the statistical properties of the system state and their time evolution when we pass from our interpolating model to another after appropriate choices of the pair @xmath144 in these time evolutions we have highlighted the influence of the average excitation @xmath145 when large or small , upon the statistical properties of the system . from what we have learned in quantum optics , concerning the degradation caused by decoherence effects affecting quantum states @xcite , for practical purposes this result would lead us to give priority to states with smaller excitations , the quantum region of small numbers , where some types of interpolating hamiltonians have problems @xcite .
we thank the brazilian funding agencies cnpq and fapeg for the partial supports .
t. yoshie , a. scherer , j. hendrickson , g. khitrova , h. m. gibbs , g. rupper , c. ell , o. b. shchekin , and d. g. deppe ) , _ vacuum rabi splitting with a single quantum dot in a photonic crystal nanocavity , _ nature * 432 * , 200 - 203 ( 2004 ) m. brune , _ _
schmidt - kaler , a. maali , j. dreyer , e. hagley , j. m. raimond , and s. haroche , _ quantum rabi oscillation : a direct test of field quantization in a cavity , phys _ , rev . lett . *
76 * , 01800 ( 1996 ) .
b. m. rodrguez - lara , f. soto - eguibar , a. z. crdenas , and h. m. moya - cessa , _ a classical simulation of nonlinear jaynes cummings and rabi models in photonic lattices _
express * 21 * , 12888 ( 2013 ) .
m. brune , e. hagley , j. dreyer , x. matre , a. maali , c. wunderlich , j. m. raimond , and s. haroche , _ observing the progressive decoherence of the meter in a quantum measurement , _ phys .
lett . * 77 * , 4887 ( 1996 ) . | we introduce a new hamiltonian model which interpolates between the jaynes - cummings model and other types of such hamiltonians .
it works with two interpolating parameters , rather than one as traditional . taking advantage of this greater degree of freedom
, we can perform continuous interpolation between the various types of these hamiltonians . as applications we discuss a paradox raised in literature and compare the time evolution of photon statistics obtained in the various interpolating models . the role played by the average excitation in these comparisons
is also highlighted . |
the dust content of a spiral galaxy disk can be mapped either by the characteristic dust emission in the far - infrared ( fir ) and sub - mm regimes , or by using the attenuation of known background sources . both techniques have seen recent significant improvements in accuracy and sensitivity , with complementary results shedding light on the dusty interstellar medium in spiral disks . the emission from the interstellar dust in the disks of spiral galaxies has been characterized with increasing accuracy by several infrared space missions ( iras , iso and , recently , _ spitzer _ ) , as well as the sub - mm observations of scuba on the jcmt .
the improvements in spatial resolution and wavelength coverage have led to significant insight into the temperature components of the dust in spiral disks , and into the relation between dusty clouds and star - formation .
the fir emission from spiral galaxies has revealed that the dust can be described by two dominant thermal components : warm ( 25 k @xmath0 t @xmath0 100 k ) and cold ( t @xmath0 25 k ) .
both the warm and cold components can be found in spiral arms , and there is a smooth disk of cold dust between these arms .
most of the dust mass in the spiral disk is cold ( see the review articles by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
fir and sub - mm observations of galaxies find indications ( e.g. , * ? ? ?
* ; * ? ? ?
* ) or direct evidence of cold dust disks extending beyond the stellar disk @xcite .
the studies of edge - on spirals by @xcite and @xcite quantified the radial profile as a scalelength of the dust that is 40% larger than that of the starlight .
the contribution of the cold ism ( t @xmath0 25 k ) to the overall emission of spiral disks has been difficult to constrain because of the degeneracy between dust temperature and mass . hence , the cold ism s relation with hi and their relative distribution remain unknown .
_ spitzer _ observations , mostly from the spitzer infrared nearby galaxy survey ( sings , * ? ? ?
* ) , have already contributed greatly to the understanding of spiral disks .
the relations between the tracers of cold dust ( 70 and 160 micron emission ) and star - formation , both obscured ( 24 micron ) and unobscured ( uv and h@xmath2 emission ) have already been studied with this multi - wavelength survey in several canonical galaxies and their substructure : the starburst m51 @xcite , the grand design spiral m81 @xcite , the rings of ngc 7331 @xcite and m31 @xcite , the superwind in m82 @xcite , and the dwarf ngc 55 @xcite . @xcite
discuss the sed of all sings galaxies over all available wavelengths .
@xcite find ample evidence for dust in all the sings galaxies , with the gas - to - dust ratio related to the metallicity .
they find no evidence for very cold ( t @xmath0 10k ) dust , however .
these studies find ample evidence of cold dust throughout the optical disks of spirals but , interestingly , also outside them in various places : on the edge of the optical disk @xcite , outside m82 s superwind @xcite and extending beyond the stellar disk @xcite .
parallel to these investigations of dust emission has gone an observational effort to quantify the absorption by dust in spiral disks using known background sources . @xcite proposed using occulting galaxy pairs for this purpose .
nearby occulting galaxy pairs were initially investigated with ground - based data , both images @xcite and spectra @xcite .
subsequently , with the hubble space telescope ( hst ) , a more detailed picture of dust in these nearby disks emerged @xcite .
the results of these studies are that extinction is gray when measured over disk sections greater than 100 pc , but resembles the galactic extinction law at smaller scales those that can only be resolved with hst .
arms are found to be more opaque than the general disk , and some evidence suggests that the dust disk is a fractal , similar to the hi disk .
@xcite investigated the use of the calibrated number of distant galaxies seen though the foreground disk in hst images .
the calibrated counts of distant galaxies have been explored further in the previous papers in this series @xcite .
both the occulting galaxy technique and counts of distant objects yield very similar opacities for disks and spiral arms @xcite . in recent years
, models have been developed to explain the spectral energy distribution ( sed ) of edge - on spirals spanning wavelengths ranging from the uv to the fir .
( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
+ three scenarios have been proposed to explain the discrepancy between the apparent absorption in uv and optical wavelengths and the emission of dust in the fir and sub - mm regimes : * a young stellar population is embedded in the dense plane of the disk .
this is proposed by @xcite and corroborated by @xcite .
the embedded young stars pump the fir emission radiated by the dust plane . * a strongly clumped dusty medium .
the clumping would lead to underestimate the dust mass from optical extinction in edge - on systems @xcite ; the dust mass would also be underestimated by a uv to fir sed @xcite . * a different emissivity of the cold dust grains higher than canonical in the fir and sub - mm .
a change of emissivity has been proposed for denser ism regions @xcite or , alternatively , for the lower density regions of the disk @xcite . in all three of these scenarios ,
the clumpiness of the dusty ism is an important factor . while in some models
the large scale structure of the dusty ism has been somewhat constrained ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the small - scale geometry ( `` clumpiness '' ) of the cold ism remains unknown .
therefore , an estimate of the prevalent dusty cloud size for spiral disks would provide a constraint for the sed models of spiral disks .
given that sed and extinction techniques are sensitive to different characteristics of the dusty clouds in the spiral disk , a comparison between the optical depth derived from these two methods has the potential to figure out the structure of the dusty ism . here
, we compare the @xmath3-band optical depths for a section of the spiral disk , one derived from a sed model of the _ spitzer _ fluxes ( `` average '' ) , and one determined from the number of distant galaxies found in an hst image ( `` apparent '' ) .
the term `` average '' refers to what the optical depth proportional to the dust mass , uniformly distributed over the disk section .
the term `` apparent '' means the effective optical depth of the disk section for a background uniform light source ( original definitions from * ? ? ?
the term `` opacity '' is used throughout our previous papers for the apparent optical depth measured over a section of the disk for its whole height .
this paper is structured as follows : [ sec : data ] discusses the _ spitzer _ and hst data used . in [ sec : analysis ] , the two different methods to derive optical depths are presented .
we discuss the relation between dust geometry and galaxy counts in [ sec : csize ] . in
[ sec : results ] , we present the derived optical depths ; [ sec : model ] presents a simple geometric model to interpret the results , and [ sec : concl ] lists our conclusions and future work .
the data for this paper come from two archives , the hst archive and the fourth data release @xcite of the _ spitzer infrared nearby galaxy survey _ ( sings , * ? ? ? * ) .
there is an overlap of fourteen galaxies between the sings sample and that of @xcite .
two of these , ngc 3621 and ngc 5194 , have two wfpc2 exposures analyzed in @xcite .
the hst / wfpc2 data reduction is described in @xcite .
the reasoning behind the hst sample selection from the archive is explained in
[ ssec : sfm ] . the overlap between hst and _ spitzer _ data is illustrated in figure [ fig : overlap ] , with the wfpc2 footprint projected on the 24 micron _ spitzer _ images .
only the solid angle covered by the wf chips is used for further analysis ( the pc chip is excluded ) . the infrared array camera ( irac ) mosaic
is made with the custom _ singsdither _ script , by m. regan , that combines the scan images into a single mosaic using the `` drizzle '' algorithm ( fourth data release notes , * ? ? ?
the multiband imaging photometer for _ spitzer _ ( mips ) data products are calibrated , sky subtracted mosaics in all three bands , reduced as described in @xcite .
the basic instrument parameters , pixel scale , and adopted psf fwhm for the seven main _ spitzer _ imaging modes are summarized in table [ table : data ] .
two parallel estimates of the optical depth of disks are used in this paper : first , the apparent optical depth of the spiral disks is determined from the number of distant galaxies identified in the hst / wfpc2 images , calibrated with the `` synthetic field method '' ( sfm ) .
secondly , the optical depth of the same section of the spiral disks is derived from the dust surface density , which is a result of the sed model fit to the _ spitzer _ fluxes using the model from @xcite . in principle , the number of distant galaxies seen through a spiral disk is a function of the dust extinction , as well as the crowding and confusion by the foreground disk .
initial applications of the number of distant galaxies as an extinction tracer were on the magellanic clouds @xcite , but they lacked accuracy .
the `` synthetic field method '' ( sfm ) was developed by @xcite to correct an extinction measurement based on the number of distant galaxies in an hst image , for the effects of crowding and confusion by the foreground spiral disk .
the `` synthetic field method '' follows a series of steps .
first , the number of distant galaxies in an hst science field is determined .
second , a series of simulated ( `` synthetic '' ) fields are made . in each of these fields , a typical background ( e.g. , the hubble deep field )
is first dimmed by a gray screen and added to the science field .
third , the added distant objects are identified in these synthetic fields .
the fourth step is to measure the relationship between the number of these identified synthetic distant galaxies and background dimming . from this relation and the original number of actual distant galaxies found in the science field
, an average opacity can be found .
it is important to remake the synthetic fields for each science field because the crowding and confusion issues are unique in each case .
an additional uncertainty in the resulting average extinction measurement is the cosmic variance in the intrinsic number of distant galaxies behind the foreground disk .
the uncertainty due to cosmic variance can be estimated from the two - point correlation function of distant galaxies and folded into the poissonian error .
the cosmic variance uncertainty is of the same order as the poisson statistical error for small numbers ( for a complete discussion on the uncertainties of the sfm , see * ? ? ?
therefore , single - field sfm measurements remain uncertain , but a meaningful conclusion can be drawn from a combined set of science fields . we have applied this method successfully on archival wfpc2 data .
@xcite present the average radial opacity profile of spiral disks and the effect of spiral arms .
the spiral arms are more opaque and show a strong radial dependence , while the more transparent disk shows a flat profile .
@xcite compare hi radial profiles to the opacity ones , and conclude that no good relation between disk opacity and hi surface density radial profiles can be found .
however , @xcite find that the sub
mm profile from @xcite generally agrees with their opacity measurements of m51 .
@xcite compare the relation between surface brightness and disk opacity ; this relation is strong in the spiral arms , but weak in the rest of the disk .
@xcite predicted , based on simulated data , that the optimum distance for the application of the sfm with current hst instruments is approximately that of virgo .
the identification of background galaxies suffers in closer disks , as the stellar disk becomes more resolved , compounding confusion .
this optimum distance , combined with the availability of deep hst / wfpc2 images from the cepheid extragalactic distance scale key project , resulted in the sample presented in @xcite .
@xcite confirmed the results from @xcite using this sample , with foreground disks spanning distances between 3.5 and 35 mpc .
a selection effect of the key project is that the majority of the hst science fields is concentrated on spiral arms and exclude the centers of the galaxies . here , we present average extinction values for the whole wfpc2 field of view , minimizing the uncertainties to the extent possible . in our initial papers , we did not apply an inclination correction to the optical depths because the correction depends strongly on the dust geometry ( see the discussion in * ? ? ?
however , in this paper we assume a simple dust model in [ sec : model ] .
the appropriate inclination correction @xmath4 has been applied to the points in figures [ fig : fit ] and [ fig : n_val ] , and in table [ table : a ] . the uncertainties in the tables and figures reflect the 1-sigma confidence levels produced by the combination of the poisson error and the cosmic variance of background galaxies . the average disk optical depth is derived from the _ spitzer _ observations .
first , the surface brightnesses within the wf chips footprint are measured ( see figure [ fig : overlap ] ) .
second , these are converted into a dust surface density using an sed model .
third , this surface density is translated into an @xmath3-band optical depth .
all the irac and mips data are convolved to the poorest resolution of the 160 micron observations ( see table [ table : data ] ) , and the pixel scale is set to 9 " .
this is done with the _ gauss , wcsmap _ and _ geotran _ tasks , under iraf .
subsequently , the overall flux is measured in the wfpc2 field of view ( figure [ fig : overlap ] ) . because the l - shaped aperture is a highly unusual one , the aperture correction remains uncertain , but not negligible , since the fwhm at 160 micron ( @xmath5 ) is of the order of the aperture diameter ( @xmath6 in an l - shape ) .
published aperture corrections for the irac instruments @xcite overestimate the correction for extended objects @xcite .
here we use the aperture corrections for extended sources from @xcite for irac fluxes , and from @xcite for the mips fluxes ( see also table [ table : data ] ) .
table [ table : flux ] gives the average surface brightnesses for the seven _
channels in the field - of - view of the three wf chips of the wfpc2 array .
the uncertainties are derived from the variance in the sky .
generally , the surface brightnesses agree with the results presented by @xcite for the entire disks . the second step is to convert these surface brightnesses to a dust surface density .
initially , we fitted only the mips fluxes with two blackbodies and derived surface densities from these @xcite .
however , a more rigorous treatment of the ir fluxes can be done with a sed model , such as the one presented in @xcite .
this model uses the physics of grain heating and reradiation , and a model distribution of grain sizes and types .
the updated version from @xcite has been fit to the data , and the results are shown in figure [ fig : spec ] .
the derived stellar and dust surface brightnesses , dust surface density , and mean temperature are presented in table [ table : model ] .
dust surface densities are between 0.1 and @xmath7 , with mean temperatures between 14.6 and 17.8 k. the mean dust temperatures are obtained from the mean radiation scaling , @xmath8 , ) of the local interstellar radiation field to calculate the irradiation that the grains see .
@xmath8 is the average of this scaling distribution . ] in the @xcite model ( @xmath9 in their equation 18 ) .
the model uses a distribution of temperatures and grain sizes .
therefore the mean temperature is an indication of the thermal equilibrium point of the bulk of the dust .
most of the dust is cold ( @xmath10 ) .
these results are an obvious improvement over a simple single - temperature fit , but the model parameters in table [ table : model ] are still not fully constrained ; disagreement between data and model at the pah peak at 8 micron could be an effect of metallicity or the presence of a bright hii region .
the 70 and 160 @xmath11 m fluxes hint at colder or more dust in the disk ( figure [ fig : spec ] ) .
there are three caveats to the fits : ( 1 ) a lack of sub mm data , ( 2 ) the single color isrf used and ( 3 ) averaging over different types of emission regions , i.e. , hii regions and the general disk . additional sub - mm data of comparable quality , needed to better constrain the model and especially the cold dust emission , will not be available until scuba2 starts operations on the jcmt and the launch of _
herschel_. at present , comparable - quality sub - mm maps are available only for ngc 5194 @xcite and ngc 7331 @xcite .
@xcite discuss sed models with and without sub - mm data .
the second caveat in the derivation of dust mass is the assumption of a constant color for the interstellar radiation field ( isrf ) illuminating the emitting dust . in reality
, dust grains deeper in a dust structure will encounter radiation field that is not only dimmed but also reddened and , hence , will contribute less flux to the fir emission .
locally , the isrf will also depend on the age of the nearby stellar population .
due to the reddening of the isrf deeper in the cloud , dense clouds could contain more dust mass in their centers than inferred from just the fir emission .
additional sub - mm observations will help resolve this uncertainty in the sed optical depth .
@xcite discuss fits of the model to the total fluxes of the sings galaxies with and without additional sub - mm data .
they find that the fir estimate underestimates of 1.5 times the dust mass more ( 5 out of 17 cases , notably ngc 3627 and 7331 of our sample ) than overestimates it ( only m51 of the 17 ) .
the third caveat is that the model values are an average over many different types of ism regions , each with a different heating mechanism , dust structure and composition ( e.g. , photo - dissociation regions , cirrus , and star - forming regions in spiral arms ) .
the @xcite model s assumptions hold better for some regions than for others , but we use the results as `` typical '' for these disks .
the relative contribution of pah emission to the sed is a function of ism geometry as well as irradiation and composition ( e.g. , * ? ? ?
* ; * ? ? ?
together with a better - constrained fir / sub - mm sed one could constrain ism geometry solely from the relative contributions to the sed .
@xcite discuss the application of the @xcite model to whole disks of the sings galaxies .
the dust surface density is translated into an average optical depth with the absorption cross section per unit dust mass , @xmath12 , and grain albedo from @xcite for the johnson @xmath3-band ( 865.5 . ) : @xmath13 .
these optical depth values are presented in table [ table : a ] .
it continues to be difficult to constrain dusty cloud geometry from models of either extinction or emission . in this section ,
we review what can be learned , solely from the sfm measurements , about the geometry of the extincting medium . in @xcite ,
two indications that the dust disk is clumpy are identified : ( 1 ) the average color of the distant galaxies is independent of the disk opacity implied by their number , and ( 2 ) the measurement of disk opacity is independent of inclination . the lack of a relation between the average opacity of the disk and the average color of the detected galaxies can be explained by two scenarios : ( 1 ) some of the background galaxies are completely blocked by large clouds , and some are not . the color measurement is done on background objects that do not suffer from extinction and , hence , reddening . alternatively , ( 2 ) _ all _ background galaxies are dimmed by clouds smaller than the projected distant galaxies . consequently , some of the distant galaxies are dimmed enough to drop below the detection threshold .
any detected galaxy s color is , however , measured from mostly unreddened flux .
however , we note that a relation between the reddening and derived extinction from the distant galaxies is difficult to detect because ( 1 ) the spread in colors of distant galaxies is substantial and ( 2 ) color is measured from the detected galaxies automatically the least dimmed whereas opacity is measured form the _ missing _ galaxies .
the inclination effect on the apparent optical depth from the number of distant galaxies is minimal ( see * ? ? ? * and figure 3 ) . assuming a thin layer of optically thick clouds , the apparent optical depth of the disk , measured from the number of distant galaxies , is dominated by the apparent filling factor of clouds .
the projected filling factor does not change much with inclination : a flat cloud covering 40% of a certain disk section still covers 40% of the inclined section .
only when the height of the cloud becomes important when the inclination is closer to edge - on , does the apparent filling factor change .
this explanation for the lack of an inclination effect in the opacity profiles does not depend on the size of the clouds .
it could be a single , large , cloud or many small ones in the plane of the disk .
however , the optical depth values in @xcite are from different sections of the disks although generally centered on a spiral arm and the effect of small inclination differences could well have been masked by comparing different regions in the disks .
in @xcite we did not apply an inclination correction because it depends on the assumed dust geometry . in this paper
we do assume a dust geometry and hence make an inclination correction ( [ sec : model ] ) .
the simulations in the sfm assume a gray screen , an uniform unclumped dust layer with opacity equal in the v and i bands .
the sfm opacity measurements in this paper are based on such simulations . in @xcite , we ran a series of simulations on ngc 1365 to characterize the effect of average cloud cross - section on the number of distant galaxies observable through a disk .
figure [ fig : sfmsize ] shows the effect of cloud size , expressed in pixels , on the simulated relation between average opacity and number of distant objects . in each simulation , we fix a cloud size and vary their number to increase disk opacity . an ensemble of unresolved clouds is effectively the gray screen . for clouds resolved with hst , more than 2 pixels in our data , after the drizzle reduction . ] , the relation between opacity ( cloud filling factor ) and number of distant objects becomes much shallower .
the same number of distant galaxies observed would then imply a much _ higher _ opacity of the disk .
because the sfm ( calibrated with a gray screen ) generally agrees well with measurements from overlapping galaxies @xcite , it seems unlikely that the disk s opacity is predominantly due to large , resolved clouds .
a pixel of @xmath14 at the distance of ngc 1365 ( 18 mpc ) is 4 pc in linear size .
it is therefore implied that the structure of the ism responsible for the disk opacity measured with the sfm varies in optical depth on scales of @xmath1510 pc or less . from the sfm measurements alone ,
the cloud geometry is impossible to determine . only when additional information is used e.g . ,
the general agreement with the occulting galaxy technique it favors small ( unresolved ) scales for the clouds .
therefore , to constrain cloud geometry , information from two different techniques needs to be combined .
table [ table : a ] presents the optical depth estimates from the synthetic field method and the sed model @xcite , for the wfpc2 field - of - view .
the optical depths range between 0.1 and 3.5 magnitudes in the @xmath3-band .
the measurements are for different parts of different spiral disks ( figure [ fig : overlap ] ) , explaining in part the range in values .
the optical depth estimates presented here may appear high for the johnson @xmath3-band , compared with other extinction estimates ( e.g. , those from inclination effects or reddening ) , but these are ( 1 ) for the entire height of the disk , and ( 2 ) generally centered on a spiral arm .
typical extinction values in the @xmath3-band are several tenths of a magnitude for a dust screen in front of the stellar spiral disk ( e.g. , * ? ? ?
two of the derived sfm opacities are negative , possibly the effect of an overdensity of distant galaxies behind the target galaxy .
the discrepancy in ngc5194 - 1 may be due to misidentification of background galaxies , as they are difficult to identify in this field .
figures [ fig : fit ] and [ fig : n_val ] show the values of disk opacity by both methods , over the same section of the disk .
both methods generally agree within the uncertainties of the measurements .
the agreement is better than our initial estimate from a blackbody fit to the mips fluxes in @xcite .
the general agreement and the mean temperature of the dust ( table [ table : model ] ) imply that most of a disk s opacity is due to the cold dust in the disk .
the relation between the apparent and average @xmath3-band optical depth measurements the first from the number of distant galaxies in hst images and the second derived from the _ spitzer _ sed could reveal the nature of the prevalent structure in the ism . the _ a priori _ assumptions are that ( 1 ) all dust structure is transparent to the fir emission from which the dust surface density is estimated in the sed model , and ( 2 ) the _ entire _ volume of the cloud emits in the fir .
we adopt _ model c _ from @xcite , in which a randomly distributed series of clumps covers the area .
these authors define two optical depths : ( 1 ) the typical optical depth ( @xmath16 ) , i.e. , the average optical depth that is proportional to the dust mass ; and ( 2 ) the apparent optical depth ( @xmath17 ) , or the optical depth if a uniform layer would cover the area .
our two measurements of optical depth sed and sfm correspond to these two optical depths , average and apparent .
an optical depth based on the sed depends on the dust mass within the area , and hence corresponds to @xmath16 .
the sfm optical depth is the apparent optical depth averaged over the area , and hence @xmath17 .
these two optical depths need not be the same , and their relation is an indication of how clumped the medium is .
let us assume a number of small dust structures with a height ( @xmath18 ) , an average grain cross - section ( @xmath19 ) , and a grain emissivity ( @xmath20 ) .
the grain number density in the clouds is denoted by @xmath21 , and the average number of clouds in a line - of - sight is @xmath22 .
we assume all clumps have the same optical depth @xmath23 : @xmath24 the average optical depth ( @xmath16 ) is then : @xmath25 and the apparent optical depth can be derived if one assumes a gaussian distribution of the number of clouds along the possible lines - of - sight and sum the contributions of all clouds ( see equations 15 and 17 - 19 in * ? ? ?
* ) : @xmath26 with asymptotic values for both optically thin and thick clouds : @xmath27 @xmath28 we note that the inferred dust mass in the disk ( @xmath29 ) is proportional to the number of clouds ( @xmath22 ) and the cloud optical depth ( @xmath23 ) .
the ratio of the apparent over the average optical depth is : @xmath30 leaving only the optical depth of the clouds ( @xmath23 ) , and hence the grain density ( @xmath21 ) and the cloud size ( @xmath18 ) as the variables .
our fit to the relation between sfm and sed optical depth estimates only has @xmath23 as the variable ( figure [ fig : fit ] ) .
we can now use equation [ eq : rat ] to derive @xmath23 from a fit to the optical depths from sed ( @xmath16 ) , and from the number of distant galaxies ( @xmath17 ) .
we want to answer three questions .
are the clouds in the disks typically optically thin or thick ?
if all disks are equal , what is the implied cloud optical depth ? how many clouds lie typically along a given line - of - sight ? optically thin clouds ( @xmath31 ) result in a ratio of optical depths close to unity : @xmath32 optically thick clouds ( @xmath33 ) result in a ratio of : @xmath34 figures [ fig : fit ] and [ fig : n_val ] show how a majority of the data exhibit a ratio of order unity .
the cold dust mass in the sed model could be better constrained with additional sub - mm information @xcite
. however , the ratios are , within the errors , consistent with optically thin clouds in most of the disks .
figure [ fig : fit ] shows the fit to the ratios of apparent to average optical depths , with @xmath23 as the single fit parameter , as per equation [ eq : rat ] . for simplicity , we assume here that all disks are made up of similar clouds , and that there is no difference between arm and disk regions .
the negative sfm measurements are excluded from the fit .
the best fit is for a cloud optical depth of @xmath350.4 with , on average , 2.6 clouds along the line - of - sight ( if the two negative points are included , the values change , respectively , to @xmath23 = 0.56 and n = 1.9 ) .
the inferred value of @xmath23 is likely to be a mean between higher values in spiral arms and much lower values in the disk .
optically thick disks have more clouds along the line of sight , while optically thin disks harbor a single cloud .
reasonable values for the parameters in equation [ eq : tc ] are : @xmath36 , @xmath37 and @xmath38 with @xmath39 @xcite ( e.g. , * ? ? ?
@xmath40 was fixed at 2 in the @xcite models . ] .
the value of 0.4 for @xmath23 implies a cloud height , @xmath18 , of @xmath15 60 pc !
much larger clouds could be resolved in extinction maps of these disks based on stellar reddening . in the case of ngc 3627 ,
ngc 5194 , ngc 6946 and ngc 7331 , there is a clear spiral arm in the reddening map @xcite ; the other reddening maps are smooth or do not extend out to cover the whole wfpc2 pointing .
a more likely scenario is that there is an inverse relation between the cloud density @xmath21 and scale @xmath18 .
such a relation can be seen in giant molecular clouds ( gmc ) of our own galaxy ( e.g , * ? ? ?
* ) . in this case
, the single optically thin cloud can be replaced by smaller optically thick ones .
we note that the largest gmcs are of the order of 60 pc .
the value of 60 pc clouds appears in contradiction to the implied size in the sfm calibration ( @xmath1510 pc ) .
however , the typical cloud size can be 60 pc and still the disk opacity can change over smaller scales , if several partially overlapping clouds are seen in projection . the inverse relation between scale and density
would also help make the two scales compatible .
our data are consistent with optically thin clouds ( @xmath1 ) and their average opacity value implies a typical cloud size that is unresolved with _
spitzer _ in our galaxies .
the above fit to the relation between @xmath17 and @xmath16 indicates that , on average , more than one cloud is needed along the line - of - sight in most disks ( @xmath41 ) .
however , the assumption was that @xmath23 had a single value for all the disks . in
[ ssec : tc ] , we argued that the ratio between @xmath17 and @xmath16 implied the @xmath23 is optically thin .
it logically follows that optically thick disks must have more than a single cloud along the line - of - sight .
figure [ fig : n_val ] illustrates the effect of number of clouds along the line - of - sight ( @xmath22 ) on the relation between @xmath17 and @xmath16 , when @xmath23 is freely increased from 0 to @xmath42 ( the lines have been drawn according to equations [ eq : tm ] and [ eq : tap ] ) .
we note that all disks are consistent with many clouds along the line - of - sight including the optically thin ones , also because they lie in the optically thin cloud ( @xmath43 , @xmath44 ) regime .
there are many refinements to be made to the simple model presented here .
some improvements for future comparisons between these two measurements of optical depth are : ( 1 ) a distribution of cloud sizes for both the sfm calibration , as well as in the model explaining the ratio between sfm and sed optical depths .
a model distribution can be taken from observations of gmcs in our own galaxy and nearby ones @xcite .
the cross - section distribution could be used in sfm measurements in the future .
a cross - section distribution can only be applied , if the foreground disk is at a single , fixed distance ; only one counts though a single foreground galaxy are used .
there are three face - on spirals with enough solid angle in hst imaging as well as additional _
spitzer _ data : m51 , m81 and m101 .
( 2 ) the sed model can be much better constrained with additional sub - mm observations , the opportunities for which will expand dramatically in the near future ( _ scuba2 _ on the jcmt and _ herschel _ satellite ) .
( 3 ) the effects of grand design spiral arms and galactic radius could be identified in a single disk ; the comparison sfm and sed is not made for different sections of the disks combined .
( 4 ) a future sed model can take into account the reddening of the interstellar radiation field , as it penetrates the ism .
this would require a comprehensive treatment of the ism structure in addition to its temperature , composition and irradiation .
to constrain models of the spiral disk s energy budget with typical values for the size of dusty clouds in the ism , we compare two techniques to extract the average and apparent optical depths of a section of spiral disk . from the comparison between sfm and sed results ,
we conclude the following : * the sfm s calibration alone implies projected cloud scales predominantly unresolved with hst ( of the order of 10 pc in ngc 1365 , see [ sec : csize ] and figure [ fig : sfmsize ] ) . *
the dust responsible for the disk s opacity is predominantly cold ( t @xmath0 25 k , table [ table : model ] ) . * the average and apparent optical depths of these disk sections , measured from sed and sfm respectively , generally agree ( figure [ fig : fit ] and [ fig : n_val ] ) .
this implies generally optically thin clouds ( @xmath45 , [ ssec : tc ] ) . *
the fit to the ratio between apparent and average optical depth measurements , @xmath46 , indicates a cloud optical depth , @xmath23 , of 0.4 , more than a single cloud along the line - of - sight , and a cloud size of @xmath15 60 pc .
if several partially overlapping clouds are seen in projection through the disk , the disk s opacity will change over smaller scales , consistent with conclusion 1 . *
optically thick disks appear to have more than a single cloud along the line - of - sight ( figure [ fig : n_val ] ) and optically thin disks may have several clouds as well .
future work using counts of distant galaxies through a foreground disk could be used to find cold dust structures at larger galactic radii , provided a sufficiently large solid angle has been imaged with hst / acs s superb resolution .
notably , the acs data on m51 , m81 and m101 are very promising for such an analysis . _
spitzer _ observations of these nearby disks are also available , making a similar comparison between sed and apparent optical depth possible for portions of these disks .
the typical cloud scale for spiral arms or disk sections or as a function of galactic radius could then be found .
the scuba-2 instrument has recently been installed on the james clerk maxwell telescope .
a project with scuba-2 to map the sings galaxies in two sub - mm bands will improve future sed modelling of these spiral disks significantly over the sed models presented here .
this work is based in part on archival data obtained with the _ spitzer _ space telescope , which is operated by jpl , caltech , under a contract with nasa .
this work is also based on observations with the nasa / esa hubble space telescope , obtained at the stsci , which is operated by the association of universities for research in astronomy ( aura ) , inc .
, under nasa contract nas5 - 26555 .
the authors would like to thank t. jarrett , for making his aperture corrections of extended sources available to us at an early stage , and erik rosolowsky , for useful discussions on local group cloud sizes .
we would like to thank maarten baes for discussion of the motivation .
we would also like to thank george bendo , erik hollenback and kristen keener for their comments on earlier drafts of this paper .
, g. j. , dale , d. a. , draine , b. t. , engelbracht , c. w. , kennicutt , jr .
, r. c. , calzetti , d. , gordon , k. d. , helou , g. , hollenbach , d. , li , a. , murphy , e. j. , prescott , m. k. m. , & smith , j .- d .
t. 2006 , , d. , kennicutt , r. c. , bianchi , l. , thilker , d. a. , dale , d. a. , engelbracht , c. w. , leitherer , c. , meyer , m. j. , sosey , m. l. , mutchler , m. , regan , m. w. , thornley , m. d. , armus , l. , bendo , g. j. , boissier , s. , boselli , a. , draine , b. t. , gordon , k. d. , helou , g. , hollenbach , d. j. , kewley , l. , madore , b. f. , martin , d. c. , murphy , e. j. , rieke , g. h. , rieke , m. j. , roussel , h. , sheth , k. , smith , j. d. , walter , f. , white , b. a. , yi , s. , scoville , n. z. , polletta , m. , & lindler , d. 2005 , , 633 , 871 , d. a. , bendo , g. j. , engelbracht , c. w. , gordon , k. d. , regan , m. w. , armus , l. , cannon , j. m. , calzetti , d. , draine , b. t. , helou , g. , joseph , r. d. , kennicutt , r. c. , li , a. , murphy , e. j. , roussel , h. , walter , f. , hanson , h. m. , hollenbach , d. j. , jarrett , t. h. , kewley , l. j. , lamanna , c. a. , leitherer , c. , meyer , m. j. , rieke , g. h. , rieke , m. j. , sheth , k. , smith , j. d. t. , & thornley , m. d. 2005 , , 633 , 857 , d. a. , de paz , a. g. , gordon , k. d. , hanson , h. m. , armus , l. , bendo , g. j. , bianchi , l. , block , m. , boissier , s. , boselli , a. , buckalew , b. a. , buat , v. , burgarella , d. , calzetti , d. , cannon , j. m. , engelbracht , c. w. , helou , g. , hollenbach , d. j. , jarrett , t. h. , kennicutt , r. c. , leitherer , c. , li , a. , madore , b. f. , martin , d. c. , meyer , m. j. , murphy , e. j. , regan , m. w. , roussel , h. , smith , j. d. t. , sosey , m. l. , thilker , d. a. , & walter , f. 2007 , , 655 , 863 , b. t. , dale , d. a. , bendo , g. , gordon , k. d. , smith , j. d. t. , armus , l. , engelbracht , c. w. , helou , g. , kennicutt , jr . , r. c. , li , a. , roussel , h. , walter , f. , calzetti , d. , moustakas , j. , murphy , e. j. , rieke , g. h. , bot , c. , hollenbach , d. j. , sheth , k. , & teplitz , h. i. 2007 , , 663 , 866 , c. w. , gordon , k. d. , bendo , g. j. , prez - gonzlez , p. g. , misselt , k. a. , rieke , g. h. , young , e. t. , hines , d. c. , kelly , d. m. , stansberry , j. a. , papovich , c. , morrison , j. e. , egami , e. , su , k. y. l. , muzerolle , j. , dole , h. , alonso - herrero , a. , hinz , j. l. , smith , p. s. , latter , w. b. , noriega - crespo , a. , padgett , d. l. , rho , j. , frayer , d. t. , & wachter , s. 2004 , , 154 , 248 , c. w. , kundurthy , p. , gordon , k. d. , rieke , g. h. , kennicutt , r. c. , smith , j .- d .
t. , regan , m. w. , makovoz , d. , sosey , m. , draine , b. t. , helou , g. , armus , l. , calzetti , d. , meyer , m. , bendo , g. j. , walter , f. , hollenbach , d. , cannon , j. m. , murphy , e. j. , dale , d. a. , buckalew , b. a. , & sheth , k. 2006 , , 642 , l127 , k. d. , bailin , j. , engelbracht , c. w. , rieke , g. h. , misselt , k. a. , latter , w. b. , young , e. t. , ashby , m. l. n. , barmby , p. , gibson , b. k. , hines , d. c. , hinz , j. , krause , o. , levine , d. a. , marleau , f. r. , noriega - crespo , a. , stolovy , s. , thilker , d. a. , & werner , m. w. 2006 , , 638 , l87 , k. d. , prez - gonzlez , p. g. , misselt , k. a. , murphy , e. j. , bendo , g. j. , walter , f. , thornley , m. d. , kennicutt , jr .
, r. c. , rieke , g. h. , engelbracht , c. w. , smith , j .- d .
t. , alonso - herrero , a. , appleton , p. n. , calzetti , d. , dale , d. a. , draine , b. t. , frayer , d. t. , helou , g. , hinz , j. l. , hines , d. c. , kelly , d. m. , morrison , j. e. , muzerolle , j. , regan , m. w. , stansberry , j. a. , stolovy , s. r. , storrie - lombardi , l. j. , su , k. y. l. , & young , e. t. 2004 , , 154 , 215 , k. d. , rieke , g. h. , engelbracht , c. w. , muzerolle , j. , stansberry , j. a. , misselt , k. a. , morrison , j. e. , cadien , j. , young , e. t. , dole , h. , kelly , d. m. , alonso - herrero , a. , egami , e. , su , k. y. l. , papovich , c. , smith , p. s. , hines , d. c. , rieke , m. j. , blaylock , m. , prez - gonzlez , p. g. , le floch , e. , hinz , j. l. , latter , w. b. , hesselroth , t. , frayer , d. t. , noriega - crespo , a. , masci , f. j. , padgett , d. l. , smylie , m. p. , & haegel , n. m. 2005 , , 117 , 503 , b. w. , meyer , m. , regan , m. , calzetti , d. , gordon , k. d. , smith , j. d. , dale , k. , engelbracht , c. w. , jarrett , t. , thornley , m. d. , bot , c. , buckalew , b. a. , kennicutt , r. c. , & gonzlez , r. a. 2007 , , _ submitted _
, j. l. , fazio , g. g. , allen , l. e. , ashby , m. l. n. , barmby , p. , deutsch , l. k. , huang , j. s. , marengo , m. , megeath , s. t. , melnick , g. j. , pahre , m. a. , patten , b. m. , smith , h. a. , wang , z. , willner , s. p. , hoffmann , w. f. , pipher , j. l. , forrest , w. j. , mcmurtry , c. w. , mccreight , c. r. , mckelvey , m. e. , mcmurray , jr . , r. e. , moseley , s. h. , arendt , r. g. , mentzell , j. e. , marx , c. t. , fixsen , d. j. , tollestrup , e. v. , eisenhardt , p. r. , stern , d. , gorjian , v. , bhattacharya , b. , carey , s. j. , glaccum , w. j. , lacy , m. d. , lowrance , p. j. , laine , s. j. , nelson , b. o. , reach , w. t. , stauffer , j. r. , surace , j. a. , wilson , g. , & wright , e. l. 2004 , in microwave and terahertz photonics . edited by stohr , andreas ; jager , dieter ; iezekiel , stavros .
proceedings of the spie , volume 5487 , pp .
77 - 92 ( 2004 ) . , ed . j. c. mather , 7792 , r. c. , armus , l. , bendo , g. , calzetti , d. , dale , d. a. , draine , b. t. , engelbracht , c. w. , gordon , k. d. , grauer , a. d. , helou , g. , hollenbach , d. j. , jarrett , t. h. , kewley , l. j. , leitherer , c. , li , a. , malhotra , s. , regan , m. w. , rieke , g. h. , rieke , m. j. , roussel , h. , smith , j. t. , thornley , m. d. , & walter , f. 2003 , , 115 , 928 , p. g. , kennicutt , jr .
, r. c. , gordon , k. d. , misselt , k. a. , gil de paz , a. , engelbracht , c. w. , rieke , g. h. , bendo , g. j. , bianchi , l. , boissier , s. , calzetti , d. , dale , d. a. , draine , b. t. , jarrett , t. h. , hollenbach , d. , & prescott , m. k. m. 2006 , , 648 , 987 , m. w. , thornley , m. d. , bendo , g. j. , draine , b. t. , li , a. , dale , d. a. , engelbracht , c. w. , kennicutt , r. c. , armus , l. , calzetti , d. , gordon , k. d. , helou , g. , hollenbach , d. j. , jarrett , t. h. , kewley , l. j. , leitherer , c. , malhotra , s. , meyer , m. , misselt , k. a. , morrison , j. e. , murphy , e. j. , muzerolle , j. , rieke , g. h. , rieke , m. j. , roussel , h. , smith , j .- d .
t. , & walter , f. 2004 , , 154 , 204 -band , from sed ( @xmath16 ) and sfm ( @xmath17 ) .
the best fit with equation [ eq : rat ] to the ratios of these values is also shown ( @xmath47 ) .
the average value for number of clouds along the line - of - sight , @xmath22 , is then 2.6 .
if the negative sfm values are included in the fit , the cloud optical depth rises to @xmath48 and the average @xmath22 becomes 1.9 . ] and @xmath17 , from equations [ eq : tm ] and [ eq : tap ] , for fixed values of the average number of clouds along the line - of - sight ( n = 0.5 , 1 , 2 , 3 , 4 ) .
the cloud optical depth @xmath23 is left to vary proportionally to @xmath16 ( @xmath49 , equation [ eq : tm ] ) .
hence maximum cloud optical depths are 6 , 3 , 1.5 , 1 and 0.75 respectively .
if optically thin clouds are made up of optically thin clouds , it seems necessary that more than one cloud will lie along the line - of - sight . ]
l l l l irac ( 3.6 ) & @xmath50 & @xmath51 & 0.9 + irac ( 4.5 ) & @xmath50 & @xmath51 & 0.9 + irac ( 5.8 ) & @xmath50 & @xmath51 & 0.7 + irac ( 8.0 ) & @xmath50 & @xmath51 & 0.75 + mips ( 24 ) & @xmath52 & @xmath53 & 1.16 + mips ( 70 ) & @xmath54 & @xmath55 & 1.2 + mips ( 160 ) & @xmath56 & @xmath5 & 1.4 + l l l l l l l l ngc 925 & 0.17 @xmath57 0.04 & 0.19 @xmath57 0.08 & 0.12 @xmath57 0.05 & 3.79 @xmath57 2.65 & 0.33 @xmath57 0.06 & 4.76 @xmath57 0.44 & 21.92 @xmath57 0.91 + ngc 2841 & 0.39 @xmath57 0.04 & 0.33 @xmath57 0.09 & 0.88 @xmath57 0.59 & 3.42 @xmath57 2.43 & 0.46 @xmath57 0.04 & 4.54 @xmath57 0.34 & 34.27 @xmath57 0.58 + ngc 3031 & 0.56 @xmath57 0.03 & 0.39 @xmath57 0.06 & 0.25 @xmath57 0.07 & 2.07 @xmath57 1.41 & 0.23 @xmath57 0.04 & 3.74 @xmath57 0.37 & 23.73 @xmath57 1.38 + ngc 3198 & 0.13 @xmath57 0.02 & 0.12 @xmath57 0.04 & 0.11 @xmath57 0.04 & 2.28 @xmath57 1.20 & 0.32 @xmath57 0.05 & 3.48 @xmath57 0.36 & 20.09 @xmath57 0.46 + ngc 3351 & 0.55 @xmath57 0.05 & 0.45 @xmath57 0.16 & 0.33 @xmath57 0.07 & 5.81 @xmath57 4.14 & 1.11 @xmath57 0.06 & 12.54 @xmath57 0.54 & 43.55 @xmath57 0.68 + ngc 3621 - 1 & 0.24 @xmath57 0.01 & 0.12 @xmath57 0.01 & 0.40 @xmath57 0.05 & 1.27 @xmath57 0.37 & 0.70 @xmath57 0.05 & 9.68 @xmath57 0.46 & 41.38 @xmath57 0.87 + ngc 3621 - 2 & 0.15 @xmath57 0.01 & 0.17 @xmath57 0.01 & 0.24 @xmath57 0.05 & 1.08 @xmath57 0.37 & 0.66 @xmath57 0.05 & 11.04 @xmath57 0.46 & 36.81 @xmath57 0.87 + ngc 3627 & 0.97 @xmath57 0.04 & 0.68 @xmath57 0.06 & 1.20 @xmath57 0.06 & 6.72 @xmath57 2.89 & 5.80 @xmath57 0.06 & 49.58 @xmath57 0.61 & 152.63 @xmath57 0.95 + ngc 4321 & 0.55 @xmath57 0.02 & 0.39 @xmath57 0.04 & 0.64 @xmath57 0.06 & 4.45 @xmath57 2.23 & 2.62 @xmath57 0.05 & 25.12 @xmath57 0.46 & 100.53 @xmath57 0.75 + ngc 4536 & 0.11 @xmath57 0.04 & 0.15 @xmath57 0.07 & 0.72 @xmath57 3.45 & 3.64 @xmath57 2.89 & 0.19 @xmath57 0.05 & 2.35 @xmath57 0.52 & 10.97 @xmath57 0.60 + ngc 4559 & 0.05 @xmath57 0.03 & 0.04 @xmath57 0.01 & 0.06 @xmath57 0.05 & 1.51 @xmath57 1.05 & 0.10 @xmath57 0.05 & 2.65 @xmath57 0.38 & 11.80 @xmath57 0.53 + ngc 4725 & 0.11 @xmath57 0.02 & 0.09 @xmath57 0.05 & 0.28 @xmath57 0.20 & 1.76 @xmath57 1.29 & 0.16 @xmath57 0.05 & 1.59 @xmath57 0.35 & 12.50 @xmath57 0.49 + ngc 5194 - 1 & 0.52 @xmath57 0.01 & 0.32 @xmath57 0.01 & 0.80 @xmath57 0.05 & 2.61 @xmath57 0.22 & 2.64 @xmath57 0.04 & 28.61 @xmath57 0.38 & 124.68 @xmath57 1.36 + ngc 5194 - 2 & 0.56 @xmath57 0.01 & 0.34 @xmath57 0.01 & 0.88 @xmath57 0.05 & 2.76 @xmath57 0.22 & 3.26 @xmath57 0.04 & 36.35 @xmath57 0.38 & 144.30 @xmath57 1.36 + ngc 6946 & 0.43 @xmath57 0.03 & 0.27 @xmath57 0.01 & 0.71 @xmath57 0.09 & 2.31 @xmath57 0.48 & 2.92 @xmath57 0.07 & 28.88 @xmath57 0.53 & 80.83 @xmath57 2.13 + ngc 7331 & 0.14 @xmath57 0.01 & 0.10 @xmath57 0.01 & 0.10 @xmath57 0.05 & 1.39 @xmath57 0.88 & 0.23 @xmath57 0.04 & 3.49 @xmath57 0.67 & 17.05 @xmath57 0.79 + ngc 0925 & 0.8 & 0.2 & 0.2 & 15.7 & 2.8 & 0.70 & 10 & 0.8 & 5.8 & u0.70_1e7_mw3.1_20 + ngc 2841 & 1.8 & 0.3 & 0.5 & 14.6 & 4.15 & 0.70 & 10 & 0.8 & 1.0 & u0.70_1e7_mw3.1_20 + ngc 3031 & 2.7 & 0.2 & 0.3 & 14.6 & 2.32 & 0.50 & 1.0 & 0.5 & 1.5 & u0.50_1e6_mw3.1_30 + ngc 3198 & 0.6 & 0.2 & 0.3 & 15.0 & 3.73 & 0.50 & 10 & 0.6 & 7.5 & u0.50_1e7_mw3.1_30 + ngc 3351 & 2.7 & 0.5 & 0.4 & 16.2 & 1.9 & 0.80 & 1.0 & 1.0 & 1.6 & u0.80_1e6_mw3.1_10 + ngc 3621 - 1 & 0.7 & 0.5 & 0.4 & 16.2 & 2.45 & 1.00 & 1.0 & 1.0 & 2.6 & u1.00_1e6_mw3.1_60 + ngc 3621 - 2 & 0.7 & 0.4 & 0.2 & 16.9 & 10.91 & 1.20 & 10 & 1.3 & 7.3 & u1.20_1e7_mw3.1_30 + ngc 3627 & 3.7 & 2.1 & 1.1 & 17.0 & 2.11 & 1.00 & 10 & 1.4 & 24.5 & u1.00_1e7_mw3.1_30 + ngc 4321 & 2.2 & 1.1 & 1.0 & 16.0 & 1.91 & 0.70 & 10 & 0.9 & 15.6 & u0.70_1e7_mw3.1_30 + ngc 4536 & 0.5 & 0.1 & 0.1 & 15.7 & 2.32 & 0.80 & 10 & 0.8 & 2.7 & u0.80_1e7_mw3.1_60 + ngc 4559 & 0.2 & 0.1 & 0.1 & 15.7 & 1.76 & 0.80 & 1.0 & 0.8 & 0.0 & u0.80_1e6_mw3.1_30 + ngc 4725 & 0.5 & 0.1 & 0.2 & 14.6 & 3.01 & 0.50 & 1.0 & 0.5 & 0.0 & u0.50_1e6_mw3.1_60 + ngc 5194 - 1 & 1.6 & 1.4 & 1.2 & 16.0 & 2.33 & 0.80 & 1.0 & 0.9 & 9.9 & u0.80_1e6_mw3.1_50 + ngc 5194 - 2 & 1.8 & 1.7 & 1.4 & 16.0 & 1.83 & 0.80 & 1.0 & 0.9 & 11.7 & u0.80_1e6_mw3.1_40 + ngc 6946 & 1.4 & 1.2 & 0.5 & 17.8 & 0.65 & 1.50 & 1.0 & 1.9 & 20.5 & u1.50_1e6_mw3.1_50 + ngc 7331 & 0.6 & 0.2 & 0.2 & 15.4 & 2.15 & 0.70 & 10 & 0.7 & 4.5 & u0.70_1e7_mw3.1_30
+ l l l l l l l ngc 0925 & @xmath58 & -0.2 & 0.6 & 0.3 & -0.63 + ngc 2841 & @xmath59 & 0.4 & 1.4 & 0.7 & 0.53 + ngc 3031 & @xmath60 & 0.4 & 0.9 & 0.5 & 0.88 + ngc 3198 & @xmath61 & 0.3 & 0.7 & 0.4 & 0.97 + ngc 3351 & @xmath62 & 1.1 & 1.0 & 1.0 & 1.06 + ngc 3621 - 1 & @xmath63 & 1.1 & 1.0 & 0.5 & 2.05 + ngc 3621 - 2 & @xmath64 & 0.5 & 0.7 & 0.4 & 1.38 + ngc 3627 & @xmath65 & 1.0 & 3.1 & 1.7 & 0.61 + ngc 4321 & @xmath66 & 1.7 & 2.8 & 2.2 & 0.76 + ngc 4536 & @xmath67 & 0.4 & 0.3 & 0.1 & 2.58 + ngc 4559 & @xmath68 & 0.1 & 0.3 & 0.1 & 0.39 + ngc 4725 & @xmath61 & 0.5 & 0.5 & 0.4 & 1.38 + ngc 5194 - 1 & @xmath69 & -0.4 & 3.3 & 3.0 & -0.12 + ngc 5194 - 1 & @xmath70 & 1.2 & 3.9 & 3.5 & 0.34 + ngc 6946 & @xmath71 & 0.8 & 1.3 & 1.0 & 0.76 + ngc 7331 & @xmath72 & 0.2 & 0.5 & 0.3 & 0.62 + | the quantity of dust in a spiral disk can be estimated using the dust s typical emission or the extinction of a known source . in this paper
, we compare two techniques , one based on emission and one on absorption , applied on sections of fourteen disk galaxies .
the two measurements reflect , respectively the average and apparent optical depth of a disk section .
hence , they depend differently on the average number and optical depth of ism structures in the disk .
the small scale geometry of the cold ism is critical for accurate models of the overall energy budget of spiral disks .
ism geometry , relative contributions of different stellar populations and dust emissivity are all free parameters in galaxy spectral energy distribution ( sed ) models ; they are also sometimes degenerate , depending on wavelength coverage . our aim is to constrain typical ism geometry .
the apparent optical depth measurement comes from the number of distant galaxies seen in hst images through the foreground disk , calibrated with the `` synthetic field method '' ( sfm ) .
we discuss what can be learned from the sfm measurement alone regarding ism geometry .
we measure the ir flux in images from the _ spitzer _ infrared nearby galaxy survey in the same section of the disk that was covered by hst .
a physical model of the dust is fit to the sed to estimate the dust surface density , mean temperature , and brightness in these disk sections .
the surface density is subsequently converted into the average optical depth estimate .
the two measurements generally agree and the sed in order model finds a mostly cold dust ( t @xmath0 25 k. ) .
the ratios between the measured average and apparent optical depths of the disk sections imply optically thin ( @xmath1 ) clouds in these disks .
optically thick disks , are likely to have more than a single cloud along the line - of - sight . |
evidence for highly ionized outflowing material is seen in the rest - frame ultra - violet spectral energy distributions ( seds ) of many quasars .
outflows are most prominent in broad absorption line quasars ( balqsos ) , where extensive line absorption by matter at outflow velocities up to 30000 is seen as a broad trough blueward of the rest - frame emission from the quasar @xcite .
balqsos themselves are observed to make up some 20per cent of the population ( e.g. * ? ? ?
* ) and the intrinsic fraction of balqsos is even higher at redshifts @xmath5 @xcite .
the ubiquitous presence of outflows with velocities up to 12000 , visible via their absorption in the 1548.19,1550.77 doublet @xcite suggests the estimate that 60per cent of quasars possess outflows in the review by @xcite may be conservative .
while the evidence for the presence of outflows associated with the majority of quasars is established , the origin , dynamics and composition of such outflows are poorly constrained .
models based on radiation driven flows @xcite , magnetohydrodynamic driven winds @xcite and thermal winds @xcite all remain viable and it is possible that more than one mechanism is important .
the quantity of material expelled from the central region of the quasar , and the energy it carries , has important implications for determining the lifetime of the quasar , the composition of the interstellar medium ( ism ) and galactic feedback mechanisms . @xcite
provide a recent review of the key physical quantities involved .
information about the nature of outflows , and the importance of radiative driving in their generation , can be gained from an analysis of the observed velocity structures present in the outflowing material .
radiation - driving , and particularly the importance of acceleration via the interaction of photons with bound electrons in the outflowing material , is perhaps most familiar as the mechanism for producing winds from hot stars .
@xcite presented the theory for stars at a time when spectroscopic observations of quasars were just beginning to provide observational constraints on the properties of gaseous material associated with the objects @xcite .
compared to the situation with hot stars , the geometry of outflows in quasars is likely more complex and there are key differences in the physical properties of the outflowing material , notably the much lower particle densities .
the theory behind radiatively - driven winds for quasars was developed by arav and coworkers in the 1990s @xcite .
the importance of radiation - driving for outflows in balqsos was stressed @xcite and @xcite proposed the detection of the ` ghost of ly@xmath6 ' as unambiguous evidence for radiation - driving as the dominant source of acceleration in balqso winds .
the phenomenon is postulated to arise in outflows where absorption of ly@xmath6 photons by -ions at a velocity of @xmath05900(where , in the frame of the outflowing -ions , strong resonant absorption occurs ) results in the rapid acceleration of material to greater velocities , reducing the amount of material present .
the overall absorption opacity at @xmath05900is thereby reduced and an excess of emission is visible ( e.g. see figure 1 of @xcite ) . for a tracer - ion in the flow , such as ,
the manifestation of the phenomenon is a reduction in the absorption at @xmath05900 , leading to an apparent local excess of continuum emission ; hence the ` ghost of ly@xmath6 ' name .
the in - depth consideration of the phenomenon by @xcite was motivated in part by possible differences in the spectra of non - balqsos and balqsos presented by @xcite and the subsequent investigation of @xcite . as discussed by @xcite ,
very specific conditions are necessary for the ` ghost of ly@xmath6 ' to be detected .
an outflow , including significant absorption , extending to @xmath710000 , must be present in a quasar that possesses both strong and relatively narrow intrinsic ly@xmath6 emission .
the ghost signature is predicted to be most obvious for quasars where the far - ultraviolet spectrum of the quasar is weak , suppressing any contribution to the acceleration of the outflow from additional absorption due to resonance lines with wavlengths @xmath81000 .
observational evidence for the presence of the ghost of ly@xmath6 has proved elusive , even when a systematic approach , using the power of the very large sample of homogeneous , moderate resolution and moderate signal - to - noise ratio spectra of quasars @xcite from the sloan digital sky survey ( sdss ) are employed @xcite . in this paper
, we undertake a complementary search , focused on the presence of ` line - locked ' absorption features in the spectra of individual quasars with extended narrow line absorption .
such line - locked absorption is also accepted as evidence that radiative line - driving plays an important role in quasar outflow dynamics @xcite .
the principle behind line - locking in radiatively - driven outflows is that , under certain circumstances , material closer to the origin of the outflow can shield material further out , resulting in a synchronising of the outflow velocities .
specifically , if a distant ` cloud ' is being accelerated outwards by line - driving arising from resonant absorption of photons in a particular ionic species , the reduction in the flux received by this cloud due to shielding by an inner cloud , moving with a relative velocity equivalent to the separation of two prominent absorption lines , could cause the clouds to become locked together in velocity @xcite .
simulations of the structure , dynamics and origin of quasar outflows also indicate that , for radiatively driven flows , line - locking could be prevalent @xcite .
quasar outflows are almost certainly quite complex and simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ) predict the presence of significant structure on a variety of scales , involving fragmentation and clumping , some of which may be highly non - spherical ( e.g. filamentary or pancake - like ) .
our use of the cloud-terminology above , and later in the paper , is as a shorthand for some form of modulation in density within the flow , along the line - of - sight to the illuminating source .
the potential significance of observing a clear signature of line - locking has been recognised @xcite for much of the period since the discovery of quasars .
evidence for the presence of line - locking signatures have , however , been largely confined to the identification of ` picket fence'-type absorption configurations in a relatively small number of individual quasars .
the availability of the sdss spectroscopic quasar sample @xcite now allows quantitative statistical investigation of outflow absorption signatures to be undertaken .
we focus on the properties of absorption due to triply - ionized carbon , specifically the doublet transition 1548.19,1550.77 , which is the most readily observable high - ionization absorption feature in the sdss spectra that appears longward of the effects of the ly@xmath6-forest ( below @xmath01200 ) .
the nature of absorption in balqsos and non - balqsos is dramatically different and we analyse samples of balqsos and non - balqsos separately .
initially we began the investigation using an auto - correlation calculation , designed to be sensitive to the presence of many , individually difficult to detect , faint absorption features .
more specifically , we looked for correlations between ` absorber pixels ' as a function of velocity separation , within each quasar spectrum , without the need to pre - identify individual , discrete absorbers .
the auto - correlation method was devised primarily for balqsos , which represent a subset of the quasar population where outflows of material are very definitely present , yet the complex absorption profiles of the bal - troughs mean that searches for discrete absorbers might not represent the optimum approach . in practice ,
such a reservation turned out to be unfounded and , as a result , we employed the familiar approach of identifying discrete absorption features and then investigating the incidence of absorber ` pairs ' as a function of velocity separation .
the discrete - absorber method turns out to be highly effective for the investigation of the properties of individual , relatively narrow , -absorbers present in both the non - balqso and balqso quasar samples .
the paper is structured as follows .
first , we present the sdss quasar sample in section [ sec : quasarsample ] .
the absorption line identification scheme is described in section [ sec : discabs ] .
statistics of the population of -absorbers as a function of outflow velocity are presented in section [ sec : ll_stats ] . in section 5 ,
we create and analyse composite quasar spectra as a function of outflow velocity , allowing an investigation into the physical properties of the absorbers and the host quasars .
line - locking is shown to be almost ubiquitous among narrow -absorbers in quasar outflows with multiple absorbers .
, a discussion of the interpretation of the results , including exploratory simulations using cloudy , is presented in section [ sec : discuss ] , before the main conclusions of the paper are summarised in section [ sec : conclude ] .
the base quasar sample consisted of 42273 objects with redshifts @xmath9 from the dr7 quasar catalogue of @xcite , where the lower redshift limit was imposed in order for the blue - wing of the emission line to lie within the sdss spectra .
above redshift @xmath10 the number of quasars in the sdss catalogue drops rapidly , and the majority are fainter than their lower redshift counterparts resulting in a significant reduction in the signal - to - noise ratio ( s / n ) of the spectra .
the quasars possess absolute magnitudes in the range -23.0@xmath11@xmath12@xmath11 - 29.5 @xcite but the sample is predominantly ( 79per cent ) made up of objects with -25.0@xmath11@xmath12@xmath11 - 27.0 .
the spectra were all processed through the sky - residual subtraction scheme of @xcite , which results in significantly improved s / n at observed wavelengths @xmath137200 , contributing to the estimation of more accurate quasar redshifts from @xcite , which are employed throughout . for 3581 quasars in the sample ,
the sdss dr7 release includes multiple spectra of high quality .
for these quasars , the spectra were co - added using inverse variance weighting to produce a combined spectrum with improved s / n .
all spectra were also corrected for the effects of galactic extinction using the sdss measurement , taken from the dust maps of @xcite , and the milky way extinction curve of @xcite .
the sample was culled to remove objects with spectra of low s / n in which the majority of moderate equivalent width absorption features are not readily detectable .
specifically , spectra were retained in the sample if the average s / n per pixel in the continuum redward of the emission , at 1600 - 1650 , exceeded s / n=5.0 . to eliminate objects possessing spectra with extended wavelength regions missing , spectra with fewer than 3500 pixels ( defined by the sdss - flag ngood )
were also excluded .
application of the two criterion resulted in the removal of 8477 spectra , leaving a sample of 33796 quasars .
samples of non - balqsos and balqsos were defined using the classifications of @xcite for quasars present in the sdss dr6 release .
the small fraction of quasars present only in dr7 were classified via visual inspection .
the resulting sub - samples of non - balqsos and balqsos contained 31142 and 2654 quasars respectively .
the sdss spectra are supplied on a logarithmic wavelength scale , with pixels of @xmath14 width . when defining reference wavelengths , or generating composite spectra in the quasar rest - frame , nearest - pixel values were employed and no interpolation or rebinning was performed at any stage .
discrete absorption features were identified using the following scheme . a ` continuum ' is defined for each quasar via the application of a simple 41pixel median filter .
the ` difference ' spectrum , to be searched for absorption features , is then obtained by subtracting the continuum from the original quasar spectrum .
the absorption line search uses a matched - filter technique ( e.g. * ? ? ?
* ) with template gaussian profiles of full width at half maximum ( fwhm ) = 160(the resolution of the sdss spectra ) , 200and 240 .
templates with lines centred at half - pixel ( i.e. 34.5 ) intervals are employed , allowing the absorber centroids to be located to better than @xmath1534.5 . at each half - pixel ,
the template giving the minimum @xmath16 value is determined and candidate absorbers are selected by applying a detection - threshold of s / n@xmath173 , where the noise is calculated using the array of variance values included with each sdss spectrum . the low value of the s / n - threshold is chosen deliberately to ensure the inclusion of as many real features as possible , albeit at the expense of the presence of a population of spurious ` detections ' .
the combination of the finite resolution of the quasar spectra and the intrinsic velocity widths of absorption systems precludes the detection of pairs of absorbers with small velocity separations .
two unresolved absorbers must be at least @xmath05-pixels ( @xmath0345 ) apart to be detected as distinct features and the ` exclusion ' interval becomes somewhat larger for features that are resolved , i.e. possess intrinsic velocity widths @xmath18100 .
a doublet was defined from the presence of two absorption features separated by [email protected] pixels .
the doublet velocity separation of 500(i.e .
7.24 pixels in the sdss spectra ) is such that absorbers with intrinsic velocity widths of up to @xmath0200are detectable , hence the maximum template velocity width of 240used for absorber detection .
real doublets with larger observed velocity widths no longer appear as two separate , individual , absorbers and are thus not included in the doublet catalogue . , identified in the sample of 31142 non - balqsos .
the majority of absorbers with @xmath19 are expected to arise in intervening galaxy haloes , unrelated to the quasars .
the excess of apparently outflowing absorbers is present out to at least @xmath20 .
an excess of absorbers with a relatively low velocity - dispersion ( @xmath21including remaining quasar redshift errors ) centred on the quasar systemic velocity is also evident from employing the improved quasar redshifts of @xcite .
, width=317 ] the rest - frame equivalent width ( ew@xmath22)-limit reached depends on the s / n of the individual quasar spectra . the absorber sample is essentially complete for the ew@xmath22-interval @xmath23 . at lower ew@xmath22 ,
the sample becomes increasingly incomplete ; the number of absorbers detected reaches a maximum at ew@xmath22=0.3 , below which the number of systems detected declines precipitately towards the minimum ew included in the non - balqso sample of ew@xmath22=0.1 . above ew@xmath22=1.0 the sample also rapidly becomes incomplete because of the maximum velocity width ( fwhm = 240 ) of the detection templates , imposed to ensure that the components of the doublets remain resolved . for the balqso sample the catalogue of narrow absorbers is less complete because of the presence of significant absorption , of substantial optical depth , over much larger velocity intervals .
less than 1 percent of the doublets were excluded from the non - balqso sample by applying the ew@[email protected] , and , given the difficulty of defining a meaningful ew - estimate for many of the narrow absorbers lying within broader absorption features , no ew@xmath22-limit was applied to the absorber catalogue from the balqso sample .
investigation of the number of line - locked systems within the absorber catalogue is differential in nature and determination of the exact completeness of the absorber sample as a function of ew@xmath22 does not affect the results presented .
the key factor is that the relative probabilities of detecting doublets and triplets ( i.e. potentially line - locked systems ) are well determined .
none of the results presented in the paper change , other than the exact number of absorbers involved , if the absorber sample is restricted to minimum ew@xmath22 thresholds of up to 0.4 .
results from the full sample , to a minimum ew@xmath22=0.1 , are presented to maximise the number of absorbers involved and hence the s / n of the results for both doublet and triplet absorbers .
the absorber sample has properties very similar to other absorber samples derived recently from sdss spectra ( e.g. * ? ? ?
* ) , although the scheme for continuum determination and feature finding are somewhat different .
the apparent optical depth method @xcite leads to an estimate of the minimum column densities reached of @xmath24 . .
the absence of absorbers at large @xmath25 for quasar redshifts @[email protected] results from the minimum observed - frame wavelength of @xmath03810 included in the sdss spectra.,width=317 ] the outflow / inflow velocity of a absorber , detected at wavelength @xmath27 in a rest - frame quasar spectrum , is parametrized by @xmath28 , where @xmath29 is the velocity of light and 1548.19 is the rest - frame wavelength of the blue component of the doublet .
an outflowing absorber has positive @xmath25 and the absorber search was undertaken for @xmath30 , i.e. to @xmath3130000 . at larger @xmath25 ,
the presence of absorption due to @xmath32 1393.8,1402.8 increases the number of spurious absorber detections , particularly in the balqso sample .
the significantly improved errors in the quasar redshifts @xcite reduces the number of apparently infalling absorbers ( with negative @xmath25 ) dramatically .
absorbers are still , however , detected with negative @xmath25 due to i ) remaining quasar redshift errors , ii ) the presence of ` associated ' absorbers that are not participating in an outflow , and iii ) [ potentially ] infalling absorbers .
the distribution of 18908 absorbers in the non - balqso sample as a function of @xmath25 is shown in fig .
[ f : betadist ] and the distribution as a function of the quasar redshift and @xmath25 is shown in fig .
[ f : qz_beta ] . throughout the paper ,
the statistics given for absorbers apply to those with @xmath33 , unless otherwise stated .
@xcite drew attention to the shape of the absorber distribution at small @xmath25 ( their figs . 5 and 6 ) .
the twin - peak , or , equivalently , reduced number of outflow absorbers at @xmath34@xmath82000(@xmath25@xmath8@xmath35 ) is confirmed at high significance in fig .
[ f : betadist ] .
the apparent two - component nature of the absorber velocity distribution is perhaps most evident in fig .
[ f : qz_beta ] , where the much larger number of absorbers and the effect of the improved redshift accuracy for the quasars , confirms the presence of a significant absorber population centred on the systemic quasar redshift and the presence of an apparent maximum in the outflow absorber population at @xmath00.00667 .
analysis of the different absorber populations is beyond the scope of the results presented here , but further investigation of absorbers employing the significant statistical power of the new sdss dr10 quasar sample will be forthcoming .
our absorption line detection routine resulted in samples of 16762 ( 2637 ) doublets ( @xmath36 ) from the 31142 ( 2654 ) non - balqso ( bal ) spectra respectively . in fig .
[ f : abs_comp ] we show the arithmetic - mean composite absorption - line spectrum of 5098 absorbers with @xmath37 identified in the non - balqso sample .
the individual quasars containing the absorbers were normalised using a high - s / n quasar composite ( see section [ sec : ll_prop ] for details ) and combined , after shifting the normalised - spectra to the rest - frame of the absorbers .
numerous additional absorption features from other ions associated with the systems , covering a range of ionization potential , are visible in the high - s / n composite spectrum . with the -absorber sample in hand
it is straightforward to search for potentially line - locked systems via the distribution of separations from -doublets to other ( individual ) absorbers blueward of the doublet .
the presence of line - locked systems would lead to an excess of separations of absorbers at @xmath014.5 pixels , indicating the presence of three , equally - spaced , absorption features , henceforth ` triples ' .
figure [ f : pixlag ] shows the distribution of -doublet to absorber separations for doublets with @xmath38 identified in the non - balqso sample .
the separation is measured from the position of the blue ( 1548.19 ) line in each doublet .
the distribution is dominated by the 4579 pairs with separations of [email protected] pixels ; the doublets themselves . at larger pixel - lags @xmath0165 and @xmath0332pixels ) , the detection of 1608.5 and 1670.8 absorption respectively , associated with some of the absorbers ( see fig .
[ f : abs_comp ] ) demonstrates the effectiveness of the absorption - line identification scheme .
evidence for some line - locked systems can be seen from the small , but significant , excess of @xmath087@xmath1513 absorbers around @xmath014pixels .
the exact number of doublet - absorber pairs at larger separations depends primarily on the velocity - separations of the doublets from the bulk of the outflow - related absorbers present at @[email protected] in the quasars ( fig .
[ f : betadist ] ) .
the broad excess of doublet - absorber detections centred at @xmath7250 pixels results from the increased number of such pairs .
considering the number of doublet to absorber separations for doublets at smaller values of @xmath25 reveals a clear detection of an excess of absorbers at @xmath014.5pixels , indicating the potential presence of line - locked systems in both the non - balqso and balqso samples .
figure [ f : pixlag_2 ] shows the histograms of doublet to absorber separations for doublets with @xmath39@xmath8@xmath25@xmath40 in the 31142 non - balqso and 2654 balqso samples .
excesses of @xmath0800 ( 220 ) absorber - pairs at @xmath014.5 pixels for the non - balqso ( balqso ) samples are present .
the number of doublets is much smaller in the balqso sample but no other significant excess of doublet - absorber pairs is evident at other velocity separations in either of the samples .
the absorber - identification scheme described in the previous section , while effective , is essentially blind to the intrinsic nature of the absorber systems found .
in addition to the outflowing absorbers that are the focus of the investigation , there are four additional sources of doublets in the absorber catalogue : ( a ) spurious systems due to the limited s / n of the quasar spectra and hence of the absorber detections , ( b ) apparent systems resulting from the presence of one , or two , real narrow gaussian - like absorption features arising from species other than , ( c ) real absorbers due to intervening gaseous systems that are not directly related to the quasars but possess redshifts that apparently place them within 30000 of the quasar redshift , ( d ) real absorbers resulting from gaseous systems that are directly related to the presence of the quasar , e.g. within the quasar host - galaxy halo or associated with galaxies in the same group or cluster in which the quasar resides .
the incidence of line - locking of interest is that relative to the proportion of systems in outflows and a statistical correction to the observed number of absorbers is desirable for each of the four ` contaminant ' populations .
( i.e. at velocities of @xmath41@xmath1315000 ) identified in the 31142 non - balqso sample
. the distribution is dominated by the 4579 detections of the 1550.5 line in each doublet at [email protected] pixels i.e. the doublets themselves ( shown in the smaller enclosed plot ) . at larger pixel - lags , ( @xmath0165 and 332pixels )
, the detection of 1608.5 and 1670.8 absorption respectively , associated with some of the doublets ( see fig .
[ f : abs_comp ] ) , demonstrates the effectiveness of the absorption - line identification scheme .
evidence for the presence of some line - locked systems can be seen from the small , but significant , excess of @xmath090 absorbers around @xmath014pixels.,width=317 ] the number of apparent doublets , due to the coincidence of two spurious gaussian absorbers with the velocity separation of a doublet , was estimated by performing a search for gaussian ` emission ' features in the same quasar spectra ( both the non - balqso and balqso samples ) . from the resulting catalogues of low s / n emission features , a total of [email protected] and [email protected] spurious doublets were ` identified ' per 0.01 interval in @xmath25 for the non - balqso and balqso samples respectively .
the expected number of triple - systems resulting solely from the coincidence of three spurious absorber detections is essentially zero .
the majority of both doublets and triplets that are not real arise from the coincidence of genuine , usually weak , absorption features present at the velocity separations used to define doublets or triplets .
the probability of finding a narrow gaussian - like absorber at a given velocity separation from a location chosen at random within the quasar spectra is straightforward to determine empirically .
specifically , on average the probability is equal to 0.52 ( 0.76 ) per cent , per pixel , at small velocity separations ( @xmath11500 ) for the non - balqsos ( balqsos ) . taking the sample of gaussian absorbers in the non - balqso sample
, there is therefore an approximately 1/200 chance that another gaussian absorber will be found in a single pixel at a specified velocity separation .
a doublet was defined by requiring the presence of a second gaussian absorber within @xmath151.0pixels of the predicted 500doublet separation and thus 1.04 ( 1.53 ) per cent of the doublets are predicted to be spurious .
it is worth noting that the contamination of the sample of doublets due to effects arising from both the finite s / n of the spectra and the presence of unrelated absorbers ( i.e. correction factors ` a ' and ` b ' above ) is thus very small .
triplet absorbers were defined by requiring the presence of a third gaussian absorber within @xmath151.5pixels of the predicted location of a line - locked absorber associated with each doublet1.5pixel definition was determined from the observed velocity - width evident for absorbers associated with the doublets , e.g. the 1608 and 1671 lines in fig [ f : abs_comp ] , much of which results from the low s / n of the individual absorber detections .
other than the exact number of systems defined , none of the results in the paper are sensitive to the precise definition of doublets or triplets in velocity space . ] .
the number of spurious triples is therefore equal to 1.52 ( 2.29 ) per cent _ of the number of doublets _ identified in the non - balqso ( balqso ) samples . as a consequence ,
the correction to the observed number of triplet absorbers is fractionally quite large , particularly at small values of @xmath25 , where the number of outflow doublets is significant .
potential higher - order systems , consisting of four or more absorbers separated by 7.24-pixels , were also sought .
the number of spurious systems includes a component ( exactly analogous to the calculation for the triplets ) equal to 1.52 ( 2.29 ) per cent of the number of triplets identified in the non - balqso ( balqso ) samples . in addition , the probability that two real doublets , separated by 14.5-pixels , occur is not insignificant .
again , the number of such instances expected is readily calculated from the statistics of the number of doublets identified per @xmath25-interval .
the number of intervening absorbers present in the absorber catalogue was calculated using the results from @xcite .
for the ew@xmath22-interval @xmath42 , where the absorber sample is close to complete , the slope of the power - law model for the ew frequency distribution ( @xmath43 ) of @xcite , @xmath44 , with @xmath45=3.49 and @xmath6=-2.62 , is in excellent agreement with the observed @xmath46 in the non - balqso sample . calculating the redshift - path for detection of absorbers in the 31142 quasars of the non - balqso sample as a function of @xmath25 produces a predicted number of intervening absorbers . for the @xmath25-interval @xmath47 ,
the @xcite model predicts 1329@xmath15140 intervening absorbers , compared to the total of 2063 absorbers detected .
thus , 64.4@xmath158per cent of the detected absorbers with @xmath48 are estimated to be intervening .
a correction for the number of intervening absorbers for the full sample was made by assuming that the observed excess of associated systems in the @xmath49 interval is independent of absorber rest - frame ew@xmath22 ; leading to a predicted number of intervening absorbers [ in the full ew@xmath22-range ] of 723@xmath1558 ( 62@xmath155 ) per 0.01-interval in @xmath25 for the entire non - balqso ( balqso ) sample .
the ew@xmath22-extrapolation involved is small ( @xmath080per cent of the entire absorber sample possess ew@[email protected] ) and the key results reported below are insensitive to the exact level of the correction for intervening absorbers because the sample is dominated by outflowing absorbers at @[email protected] .
the same correction for the presence of intervening absorbers was applied to the balqso sample , taking into account the smaller total redshift path . at small @xmath25 values ( @[email protected] )
there is evidence ( fig .
[ f : betadist ] ) for a population of ` associated'-absorbers ( ` d ' above ) that do not participate in any outflow present . without additional information it is not straightforward to derive an accurate estimate of the number of such associated - absorbers as a function of @xmath25 .
the exclusion of the @[email protected] absorbers results in a very crude , partial , correction by removing absorbers that , given the level of the quasar redshift uncertainties , are unlikely to form part of any outflow .
the statistics of the line - locked systems at small @xmath25 values thus represent lower - limits to the fraction of such systems present in any outflows .
tables [ tab : lstat_nbal ] ( non - balqsos ) and [ tab : lstat_bal ] ( balqsos ) include the observed and corrected numbers of doublets and triplets as a function of @xmath25 .
the figures in tables [ tab : lstat_nbal ] and [ tab : lstat_bal ] for @[email protected] demonstrate unambiguously that approximately 10per cent of absorbers associated with quasar outflows are line - locked . given the very small number of individual line - locked absorber systems reported in the literature
the result was not anticipated . also of potential interest
is the presence of line - locked narrow absorbers in the balqso sample , with a frequency very similar to that in the non - balqso sample .
[ cols="^ , > , > , > , > , > , > , > " , ] quasars from the non - balqso sample ( section [ sec : quasarsample ] ) , without detected absorption , were employed to provide a catalogue of ` control quasars ' , from which high - s / n composite quasars were generated . for each quasar with a detected absorber ( absorber quasar ) , the 100 quasars in the control sample with the closest absolute magnitudes , @xmath12 , within @xmath500.1 of the absorber quasar , are identified .
selection of quasars with very similar @xmath12 minimises any systematic differences in emission line to continuum ratios @xcite , while the narrow redshift interval maximises the rest - wavelength range in common with the absorber quasar .
the quasars are normalised to possess the same integrated flux , using the rest - wavelength interval 1650 - 2250 .
the normalised spectra are then combined , employing a simple arithmetic mean at each rest - wavelength , to produce a composite spectrum .
the composite is normalised to the integrated flux of the absorber quasar and divided into the absorber quasar .
the result is essentially a continuum - divided spectrum with the quasar signal removed and the absorption systems remaining .
small , large - scale , systematic differences in the slopes of the seds of the absorber and control quasars , consistent with the presence of dust at the level of @[email protected] mag , associated with the absorbers or the quasars that possess absorbers , are still evident in the absorber composites .
the implications for the dust content of the absorbers are discussed below ( section [ s : res_qseds ] ) but , to compare the absorber properties of the composites , any slope differences have been removed .
specifically , the composites have been normalised to unity using a large - scale median - filter derived continuum ( filter - scale of 251 pixels , @xmath0100 ) . fig .
[ f : comp_ds ] shows the results of the composite - absorber spectrum - generation scheme applied to the sample of quasars that possess only a single doublet in the interval 0.0@xmath11@[email protected] .
composite spectra are constructed as a function of @xmath25 .
the number of individual doublets is large , resulting in composites with a high s / n .
the single - doublet composite spectra are overplotted on the -triplet composite spectra in fig .
[ f : comp_tds ] .
the figure demonstrates the reality of the line - locked systems unambiguously , through the presence of two absorber signatures ( for multiple absorber species ) separated by the 500 -doublet interval .
transitions resulting from hydrogen , magnesium , nitrogen and silicon are all present .
the relative strengths of the twin absorption systems ( see for example the 1393,1402 panels ) is consistent with the relatively low contamination of the observed triplet sample , predicted from the statistics presented in table [ tab : lstat_nbal ] .
figure [ f : comp_tdsbals ] shows the equivalent plot for the balqso sample .
the large - scale depressions in the ` continua ' evident in the figure are due to the presence of the extended bal - troughs in the quasar spectra . again ,
even with the poorer s / n , the signature of the twin absorption systems making up the -triplets are clearly visible in the @xmath52 1216 , 1238,1242 and 1393,1402 transitions .
the s / n for the 2796,2803 absorption is particularly poor but even here , the presence of twin absorber systems can be made out at small @xmath25 .
[ f : comp_tds ] and [ f : comp_tdsbals ] provide a visual representation of some of the main results of the paper : i ) a large fraction of intrinsically narrow ( @xmath53@xmath1200 ) absorption systems , present in outflows extending to at least @xmath1820000 , are line - locked , with velocity differences equivalent to the separation of the 1548.19,1550.77 absorption doublet , ii ) the gas associated with the narrow absorption systems produces strong absorption due to species with a wide range of ionization potential ( 15.0ev ( ) to 97.9 ( ) in figs .
[ f : comp_tds ] and [ f : comp_tdsbals ] ) , iii ) line - locked absorption systems are present in both non - balqso and balqso samples and the dependence of the absorber properties as a function of outflow velocity , @xmath25 , is similar .
the presence of species over an even greater ionization potential range is illustrated by the composite absorption - line spectra of triplet absorbers and doublets in the non - balqso sample where the absorber redshifts allow the rest - frame spectra to extend down to @xmath71000 ( fig .
[ f : comp_ovi ] ) .
the sample of 36 triplet absorbers and 448 doublet absorbers are from quasars with 2.8@xmath11@[email protected] , allowing the high - ionization 1031.9,1037.6 doublet ( ionization potential 138ev ) to be included in the sdss spectra .
the s / n of the triplet - composite is not high but the presence of ` double ' , , and absorbers , separated by the 500 -doublet interval confirms that the absorbers consist of gas , at velocities coincident to @xmath869 ( i.e. one pixel in the sdss spectra ) , with physical conditions capable of producing absorption due to species with a very extended range of ionization potentials .
the systematic trends in the properties of both the doublet and triplet absorbers , notably the strength of the high- and low - ionization species as a function of @xmath25 seen in fig .
[ f : comp_tds ] , are quantified in table [ tab : absrat ] .
the ratio of the ew of @xmath54 to @xmath22 decreases with increasing @xmath25 , while that of @xmath55 to @xmath22 increases . for the single doublets , it is necessary to perform a correction to the observed ew - ratios to allow for the presence of a significant number of intervening absorbers that are unrelated to outflows .
the ew - ratios for a composite spectrum constructed from a sample of 2200 absorbers with 0.15@xmath8@[email protected] were used to provide the values for intervening absorption systems .
the ew - ratios for the single -doublet absorber - composite in the interval 0.05@xmath8@[email protected] are indistinguishable from the intervening absorber - composite , as expected , given intervening absorbers dominate the sample of single doublets at @[email protected] .
estimates of the : and : ew - ratios in the single outflowing absorbers as a function of @xmath25 are given in table [ tab : absrat ] .
the calculation assumes that the fraction of intervening single doublet absorbers in each @xmath25 interval is given by the predicted numbers from section [ ssec : ll_stats_orig ] , after the statistical removal of the number of quasars with multiple intervening absorbers and intervening plus outflow absorbers that contribute to the multiple - doublet sample ( table [ tab : qstat_nbal ] ) .
for @[email protected] the s / n of the triplet - absorber composites is very low and results for the larger interval 0.02@xmath8@[email protected] are also given .
the systematic changes in the : and : ew - ratios as a function of @xmath25 for the single - doublet and triplet absorbers in the outflows are consistent within the errors .
the uncertainties in the ew - ratios for all but the very weak -absorption at large @xmath25 are dominated by uncertainties in the exact placement of the ` continua ' , necessary for the measurements .
the poor s / n in the absorber - composites for the balqso - sample precludes a similar quantitative comparison but inspection of fig .
[ f : comp_tdsbals ] suggests that the same trends as seen in the non - balqso sample are present=0.01 intervals ( for both the non - balqso and balqso samples ) , even before correction for the fraction of intervening absorbers , is too small to provide significant constraints on the ew - ratio trends as a function of @xmath25 , although the low - s / n composites are consistent with the results for the single absorbers and the triplets . ] .
interpretation of any differences in the overall shape of the ultra - violet seds of the quasars with and without outflowing absorbers is not straightforward , given the presence of populations of both associated and intervening absorbers in the same quasars . to test for the presence of dust associated with the outflow absorbers ,
composite host - quasar spectra were constructed for samples of triplet and single doublet absorbers in the non - balqso sample that lie in the interval 0.0067@xmath8@[email protected] ( i.e. 2000 - 7000 , where few ` associated'-absorbers are present ) .
a large sample of quasars , with the same redshift and absolute magnitude distribution , from the sample of quasars with no detected absorbers was used to provide the ` control ' sed .
dividing each absorber - composite by the control - composite and fitting the resulting ratio ( over the wavelength interval 1600 - 3300 ) with an smc - like extinction curve gives @[email protected] and [email protected] for the doublet- and triplet - composites respectively .
the uncertainties are dominated by the intrinsic dispersion among the constituent quasar seds .
the measurements are consistent with absorbers in the specified outflow velocity - range possessing a small dust content and the triplet sample consisting of pairs of such doublets .
the result strongly suggests that dust associated with the outflow absorbers is responsible for the small differences in the slope of the quasar ultraviolet seds but the interpretation is not unambiguous .
@xcite investigated balqso ultraviolet seds and correlations with the high - ionization absorption properties , including outflow - velocity , @xmath25 .
@xcite draw particular attention to the anti - correlation between the strength of the @xmath56 1640 recombination emission feature in the quasar spectra and the increasing @xmath25 to which the bal - troughs extend .
the strength of is a measure of the far - ultraviolet quasar sed ( capable of ionizing helium ) .
a strong far - ultraviolet sed in the quasars results in highly ionized outflowing material and line - driven acceleration ( e.g. via ) is thus reduced . as a consequence , material does not reach as extreme outflow velocities as in outflows associated with quasars possessing weaker far - ultraviolet seds . quantifying
the ew of the relatively weak @xmath521640 emission line is difficult . as described in their section 2 , @xcite choose to define a continuum by extrapolating from a point redward ( at 1720 ) of , producing ew - values of up to 10 .
variations in the overall shape of the quasar sed also result in some dispersion in the inferred ews .
the continuum - definition procedure is very different from that often employed , which results in -ew measurements of @xmath00.5 , a factor of ten smaller ( e.g. * ? ? ?
we have adopted the more traditional continuum - definition scheme , using a linear interpolation between the continuum levels at 1620 and 1650 . the -ew in composite spectra made from samples of @xmath01000 quasars from the no - absorber non - balqso control sample is 0.62@xmath57 . dividing the absorber sample at @xmath25=0.02 , to define
` close ' ( @xmath86000 ) and ` far ' ( @xmath136000 ) absorption , produces sub - samples of single doublet ( sd ) , multiple doublet ( md ) and triplet ( t ) absorbers that produce composite quasar spectra with adequate s / n to measure the relative -emission line strength to high accuracy .
the ew measurements for the close / far composites in the sd , md and t samples are 0.70/0.57 , 0.60/0.44 , 0.71/0.44 respectively .
the continuum - definition procedure is well - defined and identical for each composite spectrum .
the ratios of the ews in the close to far samples are determined to an accuracy of better than 10 per cent and the three close / far ratios , of 1.23 , 1.36 and 1.61 , in the same sense ( stronger -ew for absorbers with relatively low outflow velocities ) as that determined by @xcite is potentially significant .
the caveat is that the dominant uncertainties on the ratios for the md and t samples is the small number of quasars in the ` far'-sub - samples and a more robust measurement is required to establish the systematic ew - difference definitively .
@xcite provide a compilation of quasar properties for the sdss dr7 quasars , including radio detections and estimates of the quasar luminosity relative to the eddington luminosity .
the non - balqso quasar sample can be divided into sub - populations with i ) no absorbers , ii ) outflow absorbers , iii ) line - locked outflow absorbers . while there are some weak trends that derive from the detectability of absorbers as a function of the s / n in the sdss spectra , there are no significant differences among the sub - populations in a ) radio - detection fraction , b ) quasar luminosity or c ) luminosity relative to the eddington luminosity . given that a substantial fraction of all quasars show evidence for outflows ( e.g. * ? ? ?
* ) and that such a large fraction of quasars with outflows show evidence for the presence of line - locked absorbers , the lack of significant differences in the distribution of such properties is perhaps not surprising . the focus of the investigation has been the presence of absorbers line - locked at the 500 separation of the 1548,1550 doublet . the absorber detection procedure described in section [ sec : discabs ] requires absorbers to exhibit significant absorption but the ` triplet'-identification scheme is sensitive to any excess of doublets present at line - locked velocity separations of other absorption species , e.g. .1238,1242 ( 962 ) or 1393,1402 ( 1932 ) .
inspection of fig .
[ f : pixlag ] and fig .
[ f : pixlag_2 ] demonstrates that no additional excess of absorber pairs is detectable at any separation [ that ca nt be attributed to pairs from different species in the same absorber system , such as the and features seen in fig .
[ f : pixlag ] ] . in fact
, the doublet catalogue does not provide much sensitivity to the detection of doublets line - locked at velocity separations @xmath181100 because there are so few quasars that possess pairs of doublets that are separated by such velocities within individual outflows .
specifically , within the @xmath25-range 0.0@xmath8@[email protected] there are only 1543 doublet outflow - pairs with velocity separations @xmath131100 .
almost exactly two - thirds of the pairs are due to intervening - outflow absorber and intervening - intervening absorber coincidences , leaving just @xmath0500 doublet outflow - pairs . as a result , to be detectable
, a line - lock at a velocity @xmath131100 in the non - balqso sample would have to include @xmath050per cent of the doublet pairs at a given velocity .
no evidence for any such line - locks is seen . the number of doublets present at separations where line - locking due to 1238,1242 at 962occurs is larger .
the velocity separation is only half an sdss - spectrum pixel different from the velocity separation of any trios of line - locked doublets . the confusion with pairs of doublets separated by @xmath0500 and a small fraction of genuine quadruples due to three line - locked doublets ( fig .
[ f : quad_abs ] ) is such that , again , the constraint on the fraction of line - locked doublets is weak .
information on the presence of other line - locked doublet pairs will improve when larger samples of absorbers are analysed ( e.g. the sdss iii quasar samples ) but from the sdss dr7 spectra the constraints on the prevalence of line - locking at @xmath131100 are weak .
the presence of line - locked narrow absorbers in the majority of highly ionized outflows associated with luminous quasars will likely provide powerful constraints on the acceleration mechanism responsible for the outflows and the physical conditions present in the flows .
a detailed investigation of models is beyond the scope of this paper but some initial calculations , intended to illustrate several concepts , are presented below .
the intention is not to attempt to model any particular outflow clouds in detail , a task that will be addressed in subsequent papers by exploring an extensive parameter space .
the radiative acceleration is given by @xcite @xmath58 where @xmath59 is the gas opacity per nucleon , @xmath60 is the mean density per nucleon , and the integral is over the incident radiation field @xmath61 .
the gas opacity @xmath59 is related to the optical depth , @xmath62 , which we use throughout because it is directly observed , by @xmath63 where @xmath64 is the nucleon density and @xmath65 is the cloud depth along the line of sight . in practice , the clouds considered below are fairly homogeneous , so @xmath62 and @xmath59 are nearly proportional to one another .
two dependencies are revealed by equation [ eqn : radaccel ] .
first , the metallicity @xmath66 enters through the ratio @xmath67 because heavy elements provide most of the opacity but little of the mass - this ratio is proportional to @xmath66 .
we assume conventional metallicities and do not explore consequences of changing @xmath66 in this paper .
second , the sed of the agn enters through @xmath61 .
we assume that the ratio @xmath68 is constant as a first approximation although other seds are considered in section [ sec : agn_shape ] .
version 13.03 of cloudy , most recently described by @xcite , is employed to simulate conditions within clouds .
the results given in table 3 of @xcite are used as a guide to set up a simple baseline photoionization model .
it is assumed that the cloud is irradiated by the standard active galactic nuclei ( agn ) sed built into cloudy .
gas found near the centres of agn , or massive galaxies generally , has metallicities significantly above solar .
therefore , adopting a conservative approach , metallicity , @xmath66 , is set at three times higher than solar . in practice , the default solar abundance for all elements heavier than helium is increased by a factor of three , with nitrogen increased by a further factor of three to account for secondary nitrogen production . a representative column density of @xmath3()=10@xmath692 and a microturbulent velocity of 50are assumed , along with a hydrogen density of @xmath70 @xmath71 .
this turbulence is supersonic with a mach number of @xmath72 , meaning that turbulent pressure will dominate over gas pressure .
these parameters are typical of the structures studied in @xcite and produce an optical depth in the doublet of a few , close to the observed value and optimised to produce the maximum acceleration due to .
the cloud is roughly isothermal with a temperature of @xmath73@xmath720000k , a typical value .
photoionization simulations of the absorbers can not determine the density of the cloud , only its ionization parameter @xmath74 ( i.e. the number of ionizing photons per atom ) .
the results above ( section [ s : res_qseds_dust ] ) suggest that each cloud has an internal reddening of @xmath75 mag .
the grain optical properties , the ratio of total to selective extinction @xmath76 , and the dust to gas ratio , are unknown .
@xcite find @xmath77 to be typical within our galaxy .
if the dust to gas ratio scales with the metallicity @xmath66 , then the hydrogen column density corresponding to the deduced reddening is @xmath78 a series of models are computed , where the ionization parameter , @xmath74 , varies and the cloud thickness is adjusted to keep @xmath3()=10@xmath692 .
figure [ fig : varyu ] shows the total hydrogen column density as a function of @xmath74 .
( h ) is shown as a function of the dimensionless ionization parameter @xmath74 .
enhanced abundances , a typical agn sed , and a microturbulent velocity of 50were assumed .
the measured reddening of 0.008 mag / cloud suggests log @xmath74 @xmath79,width=302 ] the deduced reddening and assumed grain properties correspond to log @xmath80 for @xmath66=3 .
we adopt this abundance and value of @xmath74 in the following work .
cloudy includes a high - resolution fine opacity grid that is used for multi - grid calculations of line acceleration and overlap @xcite .
the predicted line optical depths in this fine grid are shown in fig .
[ fig : opticaldepths ] .
40 , @xmath52143 , @xmath52150 and @xmath52172 .
, width=302 ] individual lines are well resolved on this fine grid , which has @xmath81 . in a highly ionized gas with enhanced abundances ,
the largest opacities are due to resonance lines of li - like , he - like and h - like ions of o , c and n. the and resonance lines are also significant .
there are an especially large number of lines at short wavelengths , @xmath82 .
the lines grow sparser at longer wavelengths and the doublet is the longest - wavelength feature .
the details of the model do not affect these qualitative features .
clouds will tend to be accelerated up to line - locked velocities if line radiative acceleration is the dominant force acting on the clouds , and if the mutual self - shielding of overlapping strong lines produces a local minimum in the acceleration .
the easiest way to search for such overlaps is by computing the line optical - depth autocorrelation function ( acf ) .
the optimal situation for line - driven radiative acceleration is when the clouds possess an optical depth , @xmath62 , of order unity for the line species .
the momentum transfer from a line species essentially saturates once the optical depth is a few ( there are hardly any photons remaining to be absorbed ) and before calculating the acf a ceiling of @xmath83 was applied to the line optical depths shown in fig .
[ fig : opticaldepths ] . the results of an autocorrelation analysis of the opacity distribution ,
is shown in the top panel of fig .
[ fig : acorr ] . with a maximum optical depth of @xmath84=2 .
the lower panel shows the autocorrelation for the same lines but weighted by the @xmath85 factor , approximating the form of the incident agn sed ( see section [ sec : agn_shape ] ) ; the doublet becomes the strongest feature below @xmath01000 .
the positions of the strong , , and doublet pairs are indicated in the lower panel .
the acfs are normalised such that they have a value of unity at zero - velocity and the features of interest are the relative heights of the peaks at velocities greater than 200(the minimum velocity shown on the x - axis above ) .
the axis at the top shows the pixel - lag , to allow direct comparison with the observational data shown in figs . [
f : pixlag ] and [ f : pixlag_2].,width=302 ] the autocorrelation results shown extend to velocity - separations of @xmath77000and , as discussed in section [ s : res_qseds_noll ] , the observational constraints on the presence of line - locks at separations @xmath181000are poor .
that said , the form of the acf at lower velocities is not initially encouraging , as the 500signal due to the doublet is immersed among a number of other line - lock signatures with similar velocities , due primarily to line species in the far ultraviolet part of the agn sed . for reasons related to the finite resolution of the sdss - spectra and the velocity - separation of the doublet ( section [ sec : discabs ] ) identification of clouds with detectable doublets , line - locked at separations @xmath8500is
not possible . at velocities
@xmath13500however , there is no reason why pairs of doublets would not be found with velocity separations in the range 500 - 1000 .
such pairs would manifest themselves as an excess number of absorber detections above the background for pixel lags @xmath015 - 30 in the left - hand panel of fig .
[ f : pixlag_2 ] .
any such population present is clearly at a much lower incidence than is the case for absorbers line - locked at the doublet separation .
the well - defined narrow peak of the third absorption component evident in fig .
[ f : pixlag_2 ] and the essentially identical velocity - widths of the two intrinsic components making up each line - locked triplet ( fig .
[ f : comp_tds ] ) also provide evidence against the presence of a significant population of additional line - locked absorbers with velocity separations 500 - 1000 . in principle , other line - pairings , with velocity separations of @xmath0500 , could be contributing to the peak at 500that we have ascribed to . removing the -doublet from the absorption - line opacity table , however , demonstrates that dominates the contribution to the 500peak completely . a focus on factors that result in an enhanced doublet signal , consistent with the observations
, therefore appears worthwhile .
the analysis above assumes a uniform weighting of the lines , ignoring the shape of the incident agn sed responsible for accelerating the gas .
the sed used to compute figs .
[ fig : opticaldepths ] and [ fig : acorr ] ( top panel ) is a generic agn sed , deduced in part from observations of intrinsic emission lines .
the sed that drives the absorbing clouds may not be the same .
the central regions of an agn are believed to have a cylindrical geometry with the emitting clouds along the equator .
it is likely that we are viewing the agn from regions near the axis of the cylinder since we have a relatively unobstructed view of the sed . to examine the effects of various seds upon the auto correlation function the sed
may be parametrized as a single power - law .
the approach is only valid in the broadest sense , but has the advantage of simplicity and serves to show the dependencies of the acf on the illuminating sed .
assume therefore , that the agn sed can be fitted , over the wavelength range shown in fig .
[ fig : opticaldepths ] , with a power - law as a function of frequency , @xmath86 , @xmath87 .
the photon - number flux is then @xmath88 .
assume further that lines are microturbulently broadened to a velocity width @xmath89 ( 50is adopted in the baseline model ) .
the line width in frequency units is @xmath90 , where @xmath29 is the velocity of light , and the number of photons within an absorption line is @xmath91 .
each photon has momentum @xmath92 , so the photon momentum transferred to the clouds is @xmath93 .
the main effect of varying the slope of the sed is to change the weighting given to lines at different wavelengths .
steep slopes give greater weight to lines at longer wavelength .
we adopt a power - law index of @xmath6=-1.7 , typical for the sdss - quasars in the sample ( e.g. * ? ? ?
the optical depths shown in fig . [ fig : opticaldepths ] were rescaled by a factor @xmath94 and normalised to give a maximum optical - depth of @xmath95 for the 1548 line .
the doublet optical depth for the sample has to be of this order for the lines to be detected .
the resulting acf is shown in the lower panel of fig .
[ fig : acorr ] .
the steeper agn sed results in the long - wavelength line pairs , due to , and , becoming stronger .
in particular , the differences at velocities @xmath11000 , compared to the unweighted case , are significant and the peak due to the doublet line - lock is much more pronounced and more consistent with the observations .
future work should investigate how the acf depends on the sed .
sophisticated agn seds including the effects of a range in @xmath96 now exist @xcite .
these could be combined with the acf approach taken in this paper to investigate how the properties of the outflow depend on other agn parameters .
line radiative acceleration must be important if line locking is to occur .
our cloudy calculations include a self - consistent treatment of acceleration by both lines and continua @xcite .
line self - shielding and overlap with other lines is fully treated using the fine opacity grid shown in fig .
[ fig : opticaldepths ] .
the database cloudy employs for resonance lines ( transitions arising from the ground state ) should be fairly complete and such transitions dominate the line - dependent acceleration .
the situation contrasts with that applying to stellar winds where @xcite stressed that millions of lines arising between excited states can contribute to the acceleration in a stellar atmosphere .
this is not the case in the low density agn absorber material considered here , where most atoms will be in the ground state . in an o - star wind ,
collisional processes resulting from the high densities create a substantial population of atoms in excited states , so subordinate lines ( transitions between two excited states ) are important
. there are _ many _ more permitted subordinate lines , which are included in the cloudy calculations , but their influence is found to be unimportant .
[ fig : opticaldepths ] shows all lines with significant optical depth .
[ fig : accelldepth ] shows the acceleration across the fiducial cloud .
the upper curve shows the total radiative acceleration while the lower curve includes only lines .
the radiative acceleration is predominantly due to lines across the entire cloud .
the acceleration is greatest at the illuminated face of the cloud , a depth of zero .
the acceleration decreases as lines become optically thick and are self - shielded . as shown by (
* equation 3.48 ) , the radiative acceleration decreases when the line optical depth becomes significant . the outward acceleration has fallen by about a factor of two at the shielded face of the cloud . although lines contribute 89 per cent of the total radiative acceleration , the doublet is not an important contributor to this total .
the @xmath52629.7 line contributes over two thirds of the total acceleration , with lesser contributors from the and doublets .
the doublet contributes only about 1 per cent of the total .
nearly all of the acceleration due to continuum absorption is due to photoelectric opacity , with electron scattering being negligible .
the exploratory investigation has been successful in showing that line - driving is capable of accelerating ` clouds ' with properties given in section [ sec : clouds ] . in the context of the observations , it has also been possible to show that the line - locking signature due to exists as a local maximum in velocity space . in terms of the spectrum of potential line - locks
, the 500feature represents the lowest velocity feature present and the idea that material in an outflow might ` lock ' naturally at the first significant peak in the line autocorrelation function ( fig .
[ fig : acorr ] ; lower panel ) is attractive .
what remains unclear , however , is why material in outflows appears to maintain the 500separation when the contribution of the -doublet represents such a small fraction of the total acceleration .
line - driven winds are famously unstable @xcite .
the decreasing acceleration means that the cloud will be compressed , becoming a pancake @xcite . eventually , the the geometry shown in fig .
[ fig : accelldepth ] will become rayleigh - taylor unstable and will fragment over scales similar to the line shielding length .
this brings in rich dynamical aspects which merit further more detailed investigation .
[ fig : accelldepth ] gives hints to the origin of the line locking phenomenon .
the properties of the cloud were taken from previous investigations and were not ` fine tuned ' for this work . had the column density been significantly smaller than assumed , the lines would have been optically thin ( and hence undetectable in this study ) and the cloud would also have been stable . had the column density been significantly larger , the lines would have been far more optically thick and self - shielded , with the acceleration falling precipitously as a consequence
. such a cloud would fragment over scales corresponding to roughly optical depth unity in the driving lines .
in fact , for these assumed properties , the cloud has precisely these properties .
could the currently visible , line - locked , small column density clouds be the result of radiative instabilities in a cloud which was originally considerably larger ?
such a scenario might explain why the line - locking signature for the narrow absorbers is also present among the balqso sample .
alternatively , in a model in which bal features are due to winds above the accretion disk , seen only for special viewing angles , the narrow absorption systems studied in this paper could result from the outwardly accelerated interstellar medium of the host galaxy , visible for a much larger range of viewing angle .
investigation of the statistics of strong , narrow doublet absorbers present in outflows seen in a large sample of luminous quasars ( -25.0@xmath1@xmath12@xmath1 - 27.0 ) demonstrates that line - locked absorber doublets are an almost ubiquitous feature of outflows . from the samples of @xmath031000 non - bal and @xmath02600 bal quasars it
is found that : @xmath97 approximately two thirds of quasars with multiple absorption systems in outflows extending to @xmath712000possess line - locked absorbers , visible as absorber triplets .
@xmath97 the line - locked absorbers are present in both the non - balqso and balqso samples with comparable frequencies , suggesting a common origin for the acceleration of the clouds traced by the narrow absorbers in non - bal and bal quasars .
@xmath97 line - locked systems are present in outflows with velocities reaching at least 20000 .
@xmath97 there are no detectable differences in the absorber properties and the dust content of single doublets and line - locked doublets .
@xmath97 the gas associated with both single and line - locked narrow absorption systems produces strong absorption due to species with a wide range of ionization potential ( 15ev ( ) to 98 ( ) and even 138ev ( ) ) .
@xmath97 both single and line - locked absorber systems show strong systematic trends in their ionization as a function of outflow velocity @xmath25 . the ratio of the ew of @xmath54 to @xmath22 decreases with increasing @xmath25 , while that of @xmath55 to @xmath22 increases .
@xmath97 the current sample of absorbers from the sdss dr7 quasars does not provide useful constraints on the prevalence of line - locks with velocity separations @xmath181000 , including those due to and at 964and 1654respectively .
@xmath97 simulations employing cloudy demonstrate that a rich spectrum of line - locked signals at various velocities may be expected due to significant opacities from resonance lines of li - like , he - like and h - like ions of o , c and n , along with contributions from and resonance lines .
@xmath97 computing the line optical - depth autocorrelation function ( acf ) for the full agn sed spectrum for parameters taken from published models provides a quantitative way to assess the importance of overlapping lines .
@xmath97 many strong contributors to the acf are found , particular those from , and doublets but , in the general case , the doublet lock is not the strongest .
@xmath97 the strength of particular line - locks depends on the shape of the illuminating agn sed , which provides the photons responsible for the acceleration at the wavelength of each absorber species .
the line - lock can be made to appear significantly stronger by adopting an agn sed shape in the ultraviolet to near x - ray range that is consistent with current knowledge of the agn sed .
@xmath97 the cloudy simulations confirm that line driving is the dominant acceleration for clouds with a column of @xmath3()@xmath010@xmath692 and @xmath3(h)@xmath010@xmath42 .
the outward acceleration falls by a factor of two across our fiducial cloud , showing that the flow is rayleigh - taylor unstable .
such a flow will fragment over scales similar to that observed the point where the acceleration has fallen by a moderate factor .
could the clouds we observe be the result of such fragmentation of clouds that were originally significantly larger ?
the observational results presented open up a potentially important new constraint for i ) understanding the origin of the accelerating mechanism for high - ionization outflows in quasars and ii ) the physical conditions pertaining within the flow .
observationally , improved constraints on the frequency of line - locks due to other species will be possible through the analysis of the forthcoming release ( dr11 ) of the sdss iii quasar catalogue .
modelling outflows is a challenging , many - parameter , problem but the dependence of line - lock signatures from different species on the form of the agn sed is a particularly promising direction for future work , on which we intend to embark
. the improved statistics from the sdss iii quasar sample will also provide detailed determinations of the outflow cloud properties over an extended range in @xmath25 .
we thank the anonymous referee for a constructive and thoughtful report and fred hamann for comments and discussions .
bob carswell s input , over an extended period , is much appreciated .
pch acknowledges support from the stfc - funded galaxy formation and evolution programme at the institute of astronomy .
jta acknowledges the award of an stfc ph.d .
studentship and an arc super science fellowship .
gjf acknowledges support by nsf ( 1108928 , 1109061 , and 1412155 ) , nasa ( 10-atp10 - 0053 , 10-adap10 - 0073 , nnx12ah73 g , and atp13 - 0153 ) , and stsci ( hst - ar-12125.01 , hst - ar- 13245 , go-12560 , hst - go-12309 , and go-13310.002-a ) , and is grateful to the leverhulme trust for support via the award of a visiting professorship at queen s university belfast ( vp1 - 2012 - 025 ) .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
hewett p. c. , foltz c. b. , 2003 , aj , 125 , 1784 hewett p. c. , irwin m. j. , bunclark p. , bridgeland m. t. , kibblewhite e. j. , he x. t. , smith m. g. , 1985 , mnras , 213 , 971 hewett p. c. , wild v. , 2010 , mnras , 405 , 2302 scargle j. d. , 1973 , apj , 179 , 705 schlegel d. j. , finkbeiner d. p. , davis m. , 1998 , apj , 500 , 525 schneider d. p. , et al . , 2007 , aj , 134 , 102 schneider d. p. , et al .
, 2010 , aj , 139 , 2360 shen y. , et al . , 2011 , apjs , 194 , 45 shlosman i. , vitello p. a. , shaviv g. , 1985 , apj , 294 , 96 simon l. e. , hamann f. , 2010 , mnras , 409 , 269 tripp t. m. , lu l. , savage b. d. , 1997 , apjs , 112 , 1 urry c. m. , padovani p. , 1995
, pasp , 107 , 803 vanden berk d. e. , et al . , 2001 ,
aj , 122 , 549 weymann r. j. , morris s. l. , foltz c. b. , hewett p. c. , 1991 , apj , 373 , 23 wild v. , hewett p. c. , 2005 , mnras , 358 , 1083 wild v. , et al . , 2008 , mnras , 388 , 227 williams r. e. , strittmatter p. a. , carswell r. f. , craine e. r. , 1975 , apj , 202 , 296 vilkoviskij e. y. , irwin m. j. , 2001 , mnras , 321 , 4 vilkoviskij e. y. , lovelace r. v. e. , pavlova l. a. , romanova m. m. , yefimov s. n. , 2006 , ap&ss , 306 , 129 | an analysis of @xmath019500 narrow ( @xmath1200 ) @xmath21548.2,1550.8 absorbers in @xmath034000 sloan digital sky survey quasar spectra is presented .
the statistics of the number of absorbers as a function of outflow - velocity shows that in approximately two - thirds of outflows , with multiple absorbers present , absorbers are line - locked at the 500velocity separation of the absorber doublet ; appearing as ` triplets ' in the quasar spectra .
line - locking is an observational signature of radiative line driving in outflowing material , where the successive shielding of ` clouds ' of material in the outflow locks the clouds together in outflow velocity .
line - locked absorbers are seen in both broad absorption line quasars ( bals ) and non - bal quasars with comparable frequencies and with velocities out to at least 20000 .
there are no detectable differences in the absorber properties and the dust content of single doublets and line - locked doublets .
the gas associated with both single and line - locked absorption systems includes material with a wide range of ionization potential ( 14 - 138ev ) .
both single and line - locked absorber systems show strong systematic trends in their ionization as a function of outflow velocity , with ionization decreasing rapidly with increasing outflow velocity .
initial simulations , employing cloudy , demonstrate that a rich spectrum of line - locked signals at various velocities may be expected due to significant opacities from resonance lines of li- , he- and h - like ions of o , c and n , along with contributions from and resonance lines .
the simulations confirm that line driving can be the dominant acceleration mechanism for clouds with @xmath3()@xmath010@xmath42 .
[ firstpage ] galaxies : active - quasars : absorption lines |
the cosmological constant is the simplest solution to the problem of the present - day acceleration of the universe .
it is consistent with all the available observational tests , from the constraints imposed by cmb maps @xcite to bao picks @xcite to sn ia constraints @xcite and the estimates of fraction of baryons in galaxy clusters through x - ray detection @xcite .
nevertheless , there is an uncomfortable feeling from the fact that @xmath10 is not satisfactory explained as a vacuum state of some field theory : the energy scale observed for @xmath10 and predicted by quantum field theory badly disagree @xcite .
this is one of the reasons why alternative explanations for the present - day cosmic acceleration have been put out since 1998 , when it was first detected @xcite .
these proposals can be classified into two broad categories , each one related to a modification of a different side of the einstein field equations of general relativity @xcite .
the modified matter approach introduces a negative pressure ( that counteracts gravity ) through the energy momentum tensor @xmath11 on the right - hand side of einstein equations .
three types of models constitute this approach , namely quintessence models @xcite , k - essence solutions @xcite and perfect fluid models ( such as the chaplygin gas @xcite ) .
the quintessence and k - essence models are based on scalar fields whose potential ( in the former case ) or the kinetic energy ( in the later one ) drive the accelerated expansion .
the perfect fluid models use exotic equations of state ( for the pressure as a function of the energy density ) to generate negative pressure .
the weakness of k - essence and quintessence proposals lies in the fact that they require a scalar field to be dominant today , when the energy scale is low .
this requirement demands the mass of such a field @xmath12 to be extremely small .
one would then expect long range interaction of this field with ordinary matter , which is highly constrained by local gravity experiments .
the difficulty with the perfect fluid models is the lack of an interpretation to the exotic equations of state based on ordinary physics .
the second approach , called modified gravity , affects the left hand side of einstein equations @xcite .
the subtypes of models in this category include braneworld scenarios @xcite , scalar - tensor theories @xcite , non - riemannian geometries @xcite , @xmath13 gravity @xcite and @xmath14 theories @xcite . in the braneworld models there
exists an additional dimension through which the gravitational interaction occurs and by which the possibility of a cosmic acceleration is accommodated .
scalar - tensor theories couple the scalar curvature @xmath0 with a scalar field @xmath12 in the action @xmath15 , allowing enough degrees of freedom for an accelerated solution .
non - riemannian geometries are another alternative way to explore the geometrization of the gravitational interaction , but allowing new degrees of freedom describing different geometric structures than curvature .
the @xmath13 gravity corresponds to models resulting from non - linear lagrangian densities dependent on @xmath0 ; @xmath14 theories build their cosmological models on manifolds whose connection is equipped with torsion instead of curvature
. all these mechanisms of generating the present - day cosmic dynamics are generally called dark energy models .
our aim with the present paper is to contribute with the study of dark energy in the branch of modified gravity .
we explore the cosmological consequences of a particular model in the context of the @xmath16 gravity testing the model against some of the observational data available .
@xmath16 gravity is a theory based on an action integral whose kernel is a function of the ricci scalar @xmath0 and terms involving derivatives of @xmath0 .
@xcite is perhaps an early example of what we call @xmath16 gravity .
a more recent example is the ( non - local ) higher derivative cosmology presented in refs .
these works intend to eliminate the big bang singularity through a bounce produced by a theory based on an action containing the regular einstein - hilbert term plus a string - spired term .
this last ingredient contains an infinite number of derivatives of @xmath0 , a requirement needed to keep this modified gravity theory ghost - free .
references @xcite are @xmath17-gravity prototypes concerned not with the past - incomplete inflationary space - time but with the cosmological constant problem . in this regard
, they try to address a similar issue that interests us in the present work : the cosmological acceleration nowadays .
few years ago , we proposed a gauge theory based on the second order derivative of the gauge potential @xmath18 following utiyama s procedure @xcite .
this led to a second field strength @xmath19 ( depending on the field strength @xmath20 ) and allowed us to obtain massive gauge fields among other applications , such as the study of gauge invariance of the theory under @xmath21 and @xmath22 groups when @xmath23 . in a subsequent work ,
gravity was interpreted as a second order gauge theory under the lorentz group , and all possible lagrangians with linear and quadratic combinations of @xmath20 and @xmath24 or the riemann tensor , its derivatives and their possible contractions , from the geometrical point of view were built and classified within that context @xcite . in a third paper we selected one of those quadratic lagrangians for investigating the cosmological consequences of a theory that takes into account @xmath0- and @xmath25-terms @xcite . that lagrangian appears in the action @xmath15 bellow
( [ action ] ) and can be understood as a particular example of an @xmath16 gravity : @xmath26 where @xmath27 ( once we take @xmath28 ) , @xmath2 is the coupling constant , @xmath29 is the covariant derivative , and @xmath30 is the lagrangian of the matter fields . as usual , @xmath31 . ) proposed here but their argument is based on different motivations . ]
an important feature of @xmath15 is the presence of ghosts ( see @xcite and references therein ) .
however , here this is not a serious issue because ( [ action ] ) should be considered as an _ effective action _ of a fundamental theory of gravity , this last one a ghost - free theory .
there are some consistent proposals of extended theories of gravity in the literature .
a good example is the string - inspired non - local higher derivative gravitation model presented in ref .
the ghost - free action of that model is:@xmath32 where @xmath33 is the planck mass ( @xmath34 , @xmath24 is the gravitational constant ) , @xmath35 is the mass scale where the higher - derivative terms become important and @xmath36 in this context , our higher - order model must be treated as a low energy limiting case of ( [ s string ] ) where the only relevant contribution of eq .
( [ f string ] ) for the present - day universe is the first term of the series , i.e. @xmath37 , @xmath38 . in this approximation ,
( [ s string ] ) is equivalent to the action integral ( [ action ] ) of our @xmath16-cosmology up to a surface term .
the equivalence of both models leads to a relation between the coupling constant @xmath2 and the parameter @xmath39
@xmath40 @xmath41 , namely @xmath42 taking the variation of ( [ action ] ) with respect to the metric tensor @xmath43 ( cf . the metric formalism ) gives the field equations @xmath44 with@xmath45 + \frac{1}{2}\nabla_{\mu}r\nabla_{\nu}r\nonumber\\ & -r_{\mu\nu}\square r - g_{\mu\nu}\square\left [ \square r\right ] -\frac { 1}{4}g_{\mu\nu}\nabla_{\rho}r\nabla^{\rho}r~ , \label{h_mu_nu}\ ] ] where @xmath46 and @xmath47 is the ricci tensor .
a four dimensional spacetime with homogeneous and isotropic space section is described by the friedmann - lematre - robertson - walker ( flrw ) line element@xmath48 ~ , \label{ds2 flrw}\ ] ] where the curvature parameter @xmath49 assumes the values @xmath50 and @xmath51 is the scale parameter . substituting the metric tensor from eq.([ds2 flrw ] ) along with the energy - momentum tensor of a perfect fluid in co - moving coordinates,@xmath52 ( @xmath53 is the density and @xmath54 is the pressure of the components fulling the universe ) , into the gravity field equations eq.([einsteineq ] ) leads to the two higher order friedmann equations:@xmath55@xmath56 where@xmath57@xmath58 which are written in terms of the hubble function @xmath59,@xmath60 and where we have set @xmath61 as is largely favored by the observational data .
the dot ( such as in @xmath62 ) means derivative with respect to the cosmic time @xmath63 and for any arbitrary function of time @xmath64 , @xmath65 i.e. , @xmath66 is the fifth - order derivative of the hubble function with respect to the cosmic time .
( [ fried i ] ) and ( [ fried ii ] ) reduce to the usual friedmann equations if @xmath67 . from this system of modified friedmann equations
one can obtain the equation expressing the covariant conservation of the energy momentum tensor , @xmath68 , which reads@xmath69 in ref.@xcite , we considered simultaneously the set formed by eq .
( [ rho dot ] ) and a combination of eqs.([fried i ] ) and ( [ fried ii ] ) , namely@xmath70 the system of eqs.([rho dot])-([fried ] ) was solved perturbatively with respect to the coupling constant @xmath2 for pressureless matter .
using the observational data available , we have determined the free parameters of our modified theory of gravity and the scale parameter as a function of the time .
the plot of @xmath71 indicates that the decelerated regime peculiar to the universe dominated by matter evolves naturally to an accelerated scenario at recent times .
this makes it promising to perform a more careful analyzes of our @xmath72 model making a complete maximum likelihood fit to all parameters . in this paper
we will solve the system of eqs.([rho dot])-([fried i ] ) exactly , without using any approximation technique .
section [ sec - model ] presents the script to obtain the complete solution to the higher order cosmological equations .
another improvement we shall do in the present work ( when comparing with ref.@xcite ) is to use the observational data available to fully test our model . as far as we know , this is the first time this is done for a higher derivative theory of gravity . in section [ sec - datafit ]
we construct the likelihood function to be maximized in the process of fitting the model parameters using data from sn ia observations @xcite and from the measurements of the mass fraction of gas in clusters of galaxies @xcite .
the results of this fit are given in section [ sec - results ] , where we show the present - day numerical values for the deceleration parameter @xmath73 @xcite,@xmath74 the _ jerk _ function @xmath75,@xmath76 and also of the _ snap _ @xmath77,@xmath78 these last two functions are given in terms of higher - order time derivatives of the scale factor and it is only natural that they show up here . in section [ sec - behaviour ]
we perform an investigation of the dynamical evolution of the model , comparing its behaviour towards the past with the standard solutions for single component universes .
moreover , we show that our @xmath72-model accommodates a future big rip or a rebouncing scenario depending on the values of the coupling constant @xmath2 .
the rebouncing event is the point where the universe attains its maximum size of expansion and from which it begins to shrink .
our final comments about the features of our model and its validity are given in section [ sec - conclusion ] .
the solution of our higher - order model shall be obtained from the integration of the pair of eqs.([fried i ] ) and ( [ rho dot ] ) .
the definitions@xmath79 and@xmath80 enables us to express eq .
( [ fried i ] ) as@xmath81 and the conservation of energy - momentum eq.([rho dot ] ) as@xmath82 which is readly integrated:@xmath83 it is standard to take @xmath84 .
moreover , substituting the solution ( [ omega_m(a ) ] ) into eq .
( [ fried h b ] ) , leads to a fifth - order differential equation in @xmath85:@xmath86 with@xmath87 in order to solve eq .
( [ fried a b ] ) , we define the non - dimensional variables@xmath88 which are related to the hubble function , deceleration parameter , jerk and snap .
in fact , when these variables are calculated at the present time , @xmath89 , they give precisely the well known cosmological parameters : @xmath90 , @xmath91 , @xmath92 and @xmath93 . in the set of variables ( [ x])-([s ] ) ,
one variable is essentially the derivative of the previous one , therefore they can be used to turn the fifth - order differential equation eq.([fried a b ] ) into five first order differential equations to be solved simultaneously .
following the method of characteristics , we choose the scale factor @xmath85 as the independent variable of the model ( instead of the cosmic time @xmath63 ) , obtaining the system@xmath94{l}\frac{d\mathcal{h}}{dz}=-\frac{1}{\left ( 1+z\right ) ^{2}}\frac { q}{\mathcal{h}}\\ \frac{dq}{dz}=-\frac{1}{\left ( 1+z\right ) ^{2}}\frac{j}{\mathcal{h}}\\ \frac{dj}{dz}=-\frac{1}{\left ( 1+z\right ) ^{2}}\frac{s}{\mathcal{h}}\\ \frac{ds}{dz}=\frac{-1}{\left ( 1+z\right ) ^{2}}f_{3}\left ( z;\mathcal{h},q , j , s\right ) \end{array } \right .
~ , \label{system(z)}\ ] ] with@xmath95 where we have used the relationship between the redshift and the scale factor,@xmath96 the system of eq.([system(z ) ] ) is solved numerically , which is indeed a far more simpler task than dealing directly with eq .
( [ fried a b ] ) .
this leads to a function @xmath97 ( essentially equal to @xmath62 ) for each set @xmath98=\{q_{0},j_{0},s_{0},b,\omega_{m0}\}$ ] of cosmological parameters .
this function is used to obtain an interpolated function for the luminosity distance @xmath99\right ) $ ] of our model defined bellow which is to be fitted to the data from sn ia observations and from @xmath100-gas measurements .
in order to relate the proposed model to the observational data , one has to build the luminosity distance @xmath101 @xcite,@xmath102 our task is to solve the system eq.([system(z ) ] ) for obtaining @xmath103 , then to integrate numerically the non - dimensional luminosity distance,@xmath104\right ) = \frac{h_{0}}{c}d_{l}\left ( z;[\mathcal{m}]\right ) ~,\;\;\;\;\;\;\left ( c=1\right ) \label{dh(z)}\ ] ] for each set of cosmological parameters @xmath98=\{q_{0},j_{0},s_{0},b,\omega_{m0}\}$ ] .
function @xmath105 appears in the definition of the distance modulus,@xmath106 where @xmath107 is the reference distance of @xmath108 pc .
observational data give @xmath109 of the supernovae and their corresponding redshift @xmath110 . according to the likelihood method the probability distribution of the data@xmath111 |\left\ { z_{i},\mu_{r_{i}}\right\ } \right ) = n_{snia}e^{-\frac{\chi_{snia}^{2}\left ( \mathcal{[m]}\right ) } { 2 } }
\label{l snia}\ ] ] ( @xmath112 is a normalization constant ) is the one that minimizes @xmath113\right ) = { \displaystyle\sum\limits_{i } } \frac{\left ( \mu_{r_{i}}-5\log d_{h}\left ( z_{i},[\mathcal{m}]\right ) -5\log_{10}\left ( \frac{c}{h_{0}d_{0}}\right ) \right ) ^{2}}{\sigma_{\mu _ { i}}^{2}+\sigma_{ext}^{2}}~ , \label{chi2 snia}\ ] ] where @xmath114 is the uncertainty related to the distance modulus @xmath115 of the @xmath116-th supernova . the data set used in our calculations is composed with the 557 supernovae of the union 2 compilation @xcite .
uncertainty @xmath117 is given by@xmath118 i.e. , it is the combination of the error @xmath119 in the magnitude due to host galaxy peculiar velocities ( of order of @xmath120),@xmath121 and the uncertainty from gravitational lensing @xmath122 which is progressively important at increasing sn redshifts .
further detail can be found in refs .
@xcite , for example .
measurements of the fraction of the mass of gas in galaxy clusters in the x - ray band of the spectrum constrain efficiently the ratio @xmath123 and exhibit a lower dependence on the values of other cosmological parameters @xcite . is the non - dimensional baryon density parameter . ]
according to ref.@xcite , the mass fraction of gas @xmath124 relates to @xmath123 as@xmath125 ^{1,5}~ , \label{f_gas}\ ] ] where @xmath126 is the angular distance for @xmath61,@xmath127\right ) } ~. \label{da}\ ] ] the term @xmath128 is the angular distance for the standard @xmath10cdm model , for which @xmath129 , @xmath130 and @xmath131mpc@xmath132 , i.e.@xmath133 ^{1/2}}. \label{da lcdm}\ ] ] the other factors in eq .
( [ f_gas ] ) are defined bellow : * @xmath134 is a constant due to the calibration of the instruments observing in x - ray .
following ref .
@xcite , we take @xmath135 . *
constant @xmath18 measures the angular correction on the path of an light ray propagating in the cluster due to the difference between @xmath10cdm model and the model that we want to fit to the data . ref .
@xcite argues that for relatively low red - shifts ( as those involved here ) this factor can be neglected , i.e. taken as @xmath136 .
* @xmath137 accounts for the influence of the non - thermal pressure within the clusters .
according to ref.@xcite , one may set @xmath138 .
our choice is @xmath139 .
* @xmath140 is the bias factor related to the thermodynamical history of the inter - cluster gas . once again due to arguments given in ref.@xcite
, this factor can be modeled as @xmath141 where @xmath142 and @xmath143 .
we set @xmath144 .
* parameter @xmath145 describes the fraction of baryonic mass in the stars .
is not the snap , defined in eq .
( [ snap ] ) . even at risk of confusion
, we decided to maintain the notation used in ref .
@xcite uses @xmath146 and @xmath147 . here
@xmath148 and we will set @xmath149 .
analogously to eq.([h70 ] ) , we define@xmath150 that will be required bellow . *
the baryon parameter @xmath151 is set to@xmath152 following the constraints imposed by the primordial nucleosynthesis on the ratio @xmath153 @xcite . the uncertainties in the set of parameters @xmath154 are propagated to build the uncertainty of @xmath155 .
it is worth mentioning that the quantities @xmath156 have gaussian uncertainties . on the other hand , the parameters @xmath157 present a rectangular distribution
, e.g. @xmath137 has equal probability of assuming values in the interval @xmath158 $ ] .
the uncertainty of the parameters obeying the rectangular distribution is estimated as the half of their variation interval divided by @xmath159 .
maybe it is useful to remark that in the case of @xmath160 , we adopted @xmath161 .
moreover , @xmath162 .
@xcite provides data for the fraction of gas @xmath124 and its uncertainty @xmath163 of 42 galaxy clusters with redshift @xmath110 .
these data can be used to determine the free parameter of our model by minimizing @xmath164\right ) = { \displaystyle\sum\limits_{i } } \frac{\left ( f_{gas}^{i}-f_{gas}^{\lambda cdm}\left ( z,[\mathcal{m}]\right ) \right ) ^{2}}{\sigma_{f_{i}}^{2 } } \label{chi2 f - gas}\ ] ] notice that the dependence of @xmath165 on the set @xmath166 $ ] of cosmological parameters occurs through the function @xmath155 given by eq.([f_gas ] ) , i.e. ultimately through @xmath167 \right ) $ ] eq . [ da ] .
function @xmath168 is related to the distribution @xmath169
\right ) = n_{gas}e^{-\frac { \chi_{gas}^{2}\left ( \mathcal{[m]}\right ) } { 2 } } \label{l gas}\ ] ] which should be maximized . the majority of papers that present cosmological models tested through sn ia data fit see e.g. refs .
@xcite perform an analitical marginalization with respect to @xmath170 . on the other hand , constraint analyses using @xmath100-gas data adopt a fixed value for @xmath170 .
our basic reference on @xmath100-gas treatment , namely ref.@xcite , choose @xmath171mpc@xmath132 . as we will consider sn ia and @xmath100-gas data sets in the same footing , it is convenient to adopt @xmath171mpc@xmath132 as a fixed quantity in our calculations ( and not the most up - to - date value of @xmath172mpc@xmath132 found in @xcite ) .
we want to maximize the probability distributions given in eqs.([l snia ] ) and ( [ l gas ] ) simultaneously , which is done by maximixing@xmath173 \right ) = \mathcal{l}_{snia}\left ( \left [ \mathcal{m}\right ] \right ) \mathcal{l}_{gas}\left ( \left [ \mathcal{m}\right ] \right ) ~ , \label{l}\ ] ] or minimizing@xmath174\right ) = \chi_{snia}^{2}\left ( [ \mathcal{m}]\right ) + \chi_{gas}^{2}\left ( [ \mathcal{m}]\right ) ~. \label{chi2}\ ] ]
the parameters to be ajusted are of two types .
the kinematic parameters are @xmath176 ; are precisely the values of @xmath177 calculated at the present time @xmath89 . ]
the cosmological parameters are @xmath178 .
maximization of the likelihood function @xmath179 \right ) $ ] eq .
( [ l ] ) determines all kinematic parameters and one cosmological parameter , namely @xmath180 .
parameter @xmath181 is not fixed by statistical treatment of the data , and the reason for that is evident in the plot of fig .
[ figb ] .
\right ) $ ] as a function of the absolute value of the coupling parameter @xmath181 .
the other parameters @xmath182 were fixed in their best - fit values.,width=566 ] the curve of @xmath183 is not gaussian but rather shows a flat region at high values of @xmath184 .
indeed , values of @xmath184 greater than @xmath185 are almost equally probable and more favorable than the @xmath186 .
this shows it is not possible to determine @xmath181 by maximizing its likelihood function .
instead , we choose four values of @xmath181 to proceed the analysis leading to the determination of the free parameters of the model .
these values are @xmath187 , so that we have two values of @xmath181 in the beginning of the plateau ( @xmath188 ) and two of them far from the beginning of the plateau ( @xmath189 ) .
these choices are only partially arbitrary : we have done preliminary calculations with other values of @xmath184 ( @xmath190 ) and the results for the best - fit calculations of @xmath180 are almost the same as those obtained by using @xmath189 .
the standard statistical analysis leads to the best - fit values of the parameters shown in table [ table - parameters ] .
[ c]||c||c||c||c||c||@xmath181 & @xmath191 & @xmath192 & @xmath160 & @xmath3 + @xmath193 & @xmath194 & @xmath195 & @xmath196 & @xmath197 + @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath202 + @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath197 + @xmath207 & @xmath194 & @xmath195 & @xmath208 & @xmath197 + it is clear from table [ table - parameters ] that our higher order cosmological model presents an acceleration at present time : the deceleration parameter @xmath191 is negative for all the values of the coupling constant @xmath181 .
this is a robust result indeed once we have got negative values for @xmath191 with @xmath209 confidence level .
both sn ia and @xmath100-gas data sets give information about astrophysical objects at relatively low redshifts ( @xmath210 ) .
that is the reason for the high values of the uncertainties on the parameter @xmath192 and @xmath160 .
positive values of @xmath192 mean that the tendency is the increasing of the rate of cosmic acceleration .
because the jerk parameter @xmath192 does not exhibit the same sign for all values of @xmath181 , it is not possible to argue in favor of an overall behavior for the cosmic acceleration today .
this statement is also applicable to the snap @xmath160 obtained with @xmath211 .
however , the uncertainties in the value of @xmath160 for @xmath212 do indicate that it is positive within 68% cl .
the negative value of @xmath160 for @xmath213 is also guaranteed within the same cl .
this difference in the sign of @xmath160 for @xmath212 and @xmath213 leads to distinct dynamics for the scale factor towards the future , as we shall discuss in section [ sec - behaviour ] .
the best - fits for @xmath3 are statistically the same for all the chosen values of @xmath181 .
this is linked to the fact that the @xmath100-gas data produce a sharp determination of this parameter .
notice that the value @xmath3 of our model is compatible with that of @xmath10cdm model as reported in ref.@xcite within 95% cl .
inspection of table [ table - parameters ] indicates that the gaussian - like distributions centered in each parameter of the set @xmath180 for @xmath214 overlap the distributions centered in @xmath180 for @xmath215 ; this overlap of distribution curves occurs within 68% cl of each distribution .
moreover , this overlap of distributions is greater than the one observed in the case of the sets @xmath180 for @xmath212 and @xmath213 , which occurs within 95% cl .
hence , the greater the value of @xmath184 , the greater the overlap of distribution and the less important is our choice of the value of @xmath181 .
this emphasizes what was said before about the choice of values for @xmath181 .
the confidence regions ( cr ) in the @xmath216 and @xmath217 planes for each one of the four values of @xmath181 are shown in figs .
[ regbminus01]-[fig - regionsbplus01 ] . and @xmath216 planes for @xmath214 from sn ia and @xmath100-gas data sets .
the inner contour contains the 68% cr , whilst the outer curves confines the 95% cr.,title="fig:",width=302 ] and @xmath216 planes for @xmath214 from sn ia and @xmath100-gas data sets .
the inner contour contains the 68% cr , whilst the outer curves confines the 95% cr.,title="fig:",width=302 ] and @xmath217 planes for @xmath212.,title="fig:",width=302 ] and @xmath217 planes for @xmath212.,title="fig:",width=302 ] and @xmath218 planes for the value @xmath213 of the coupling constant.,title="fig:",width=302 ] and @xmath218 planes for the value @xmath213 of the coupling constant.,title="fig:",width=302 ] and @xmath219 for @xmath215.,title="fig:",width=302 ] and @xmath219 for @xmath215.,title="fig:",width=302 ] the shape of the confidence regions for @xmath220 indicates that the deceleration parameter @xmath191 and the jerk @xmath192 are correlated parameters ; this occurs for all the values of @xmath181 .
the four plots @xmath216 point to an accelerated universe at recent times within 95% cl , but the change in the acceleration rate takes positive and negative values with equal probability .
conversely , the shape of the plots of @xmath219 in figs .
[ regbminus01][fig - regionsbplus01 ] indicate that @xmath191 and @xmath3 are weakly correlated parameters .
notice that the difference between the maximum and the minimum values of @xmath3 in the range of the plots @xmath221 is only 0.05 with 95% cl : this confirms that the parameter @xmath3 is sharply determined by the best - fit procedure ( something that is consistent with the fact that we have employed @xmath100-gas data to build the likelihood function ) . in order to quantify the correlation between the parameters @xmath222 and @xmath223 , the covariances @xmath224 , @xmath225 and correlation coefficients @xmath226 , @xmath227 were evaluated according to ref .
the mean values and their respective standard deviations used to calculate the covariances and correlation coefficients were those given in table [ table - covariances ] . the same table shows the values obtained for the covariances and correlation functions for each value of @xmath181 taken in this paper . [ c]||c||c||c||c||c||c||c||c||c||@xmath181 & @xmath191 & @xmath192 & @xmath3 & @xmath160 & @xmath224 & @xmath226 & @xmath225 & @xmath227 + @xmath193 & @xmath228 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 + @xmath198 & @xmath236 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 + @xmath203 & @xmath244 & @xmath245 & @xmath230 & @xmath246 & @xmath247 & @xmath248 & @xmath234 & @xmath249 + @xmath207 & @xmath250 & @xmath229 & @xmath230 & @xmath251 & @xmath252 & @xmath233 & @xmath234 & @xmath235 + the values of the correlation coefficient presented in table [ table - covariances ] confirm what was stated before : the parameters @xmath191 and @xmath192 are strongly ( negatively ) correlated while @xmath191 and @xmath3 are weakly correlated for all values of b considered here .
comparison of the results presented in table [ table - covariances ] and table [ table - parameters ] shows that there is no statistical difference between considering the mode and a confidence interval of 68% ( table [ table - parameters ] ) from taking the mean and standard deviation .
this emphasizes the quasi - gaussian behaviour of the probability function .
we can investigate the evolution of our higher order model in time by plotting the function @xmath253 as a function of @xmath110 .
this is done setting the parameters at their best - fit results for each value of @xmath181 .
there is agreement between our model
regardless the value of @xmath181 and the standard @xmath10cdm model in the region @xmath254 . in particular ,
the curves of @xmath97 for @xmath214 and @xmath215 almost coincide : the difference between them is only noticeable at @xmath255 .
the transition from a decelerated phase to an accelerated expansion in the higher order model occurs at @xmath256 for all values of @xmath181 . towards the future ( @xmath257 )
the redshift takes negative values , @xmath258 . in this region
our model present different dynamics depending on the value of the coupling constant . for @xmath214 , @xmath212 and @xmath215
, there is an onset of an ever growing acceleration eventually leading to a _ premature big rip _ , respective to the one taking place in the @xmath259cdm model . on the other hand , the velocity in the higher order model with @xmath213 black curve in fig .
[ fig - behaviour ] reaches a maximum value at @xmath260 , the acceleration then changes the sign , leading to a positive deceleration parameter .
this deceleration is sufficiently intense to produce a _ rebouncing _
, where the universe reaches a maximum size and then begins to shrink .
this happens at @xmath261 .
this difference between big rip and rebouncing scenarios deserves a more careful statistical analysis .
we shall consider this in a broader space of parameters , and not using only the best - fit values for the parameters of our model , as we have done so far .
a careful analysis shows that the deceleration parameter @xmath191 influences weakly on the possible future regimes of our model .
the parameters that decisively decide in favor of a big rip scenario or a rebouncing event are the jerk @xmath192 and the snap @xmath160 .
[ fig - jerksnap ] displays the @xmath262 confidence regions plots for @xmath211 . the straight red lines passing through the confidence regions of the plots fig .
[ fig - jerksnap ] separate the sector of the space of parameters where the rebouncing occurs ( lower part ) from that region of values corresponding to a big rip ( upper part ) .
a numerical analysis shows that 39% of the @xmath263 confidence region allows for the rebouncing scenario in the higher order model with @xmath214 and 42% of the confidence region for the model with @xmath215 leads to rebouncing .
a similar analysis shows that the model with @xmath212 produces a premature big rip for all the values of the pair @xmath262 in the 95% confidence region ; converselly , the model with @xmath213 engenders rebouncing within the 95% cr . at last ,
it is interesting to explore the features of the model for @xmath264 . extrapolating the results presented in fig .
[ fig - behaviour ] we verified that our model _ does not reproduce _ the decelerated phase typical of a matter dominated friedmann universe .
besides , for @xmath265 the higher order model exhibits a behaviour compatible with @xmath266 where the equation of state is @xmath267 , i.e. the equation for a perfect fluid in the standard friedmann cosmology . on the other hand , a linear perturbative analysis shows that the perturbations in gravitation and matter fields do not grow for the equation of state with @xmath268 as given in ( [ w interval ] ) .
therefore , the model proposed here can not describe the structure formation in the early universe .
in this paper we have studied the phenomenological features of a particular modified gravity model built with an action integral dependent on the ricci scalar @xmath0 and also on the contraction @xmath269 of its covariant derivatives .
the friedmann equations resulting from the field equation and the flrw line element involve sixth - order time - derivatives of the scale factor @xmath85 .
we solved those differential equations numerically but without any approximation assuming a null cuvature parameter ( @xmath270 ) and pressureless matter .
our higher order model depends on five parameters , namely @xmath166 = \left\ { q_{0},j_{0},s_{0},\omega_{m0};b\right\ } $ ] .
the present day value of the hubble constant is taken as given and not constrained by data fit
the reason for that being consistency with the @xmath100-gas data analysis .
four of the five parameters @xmath166 $ ] are determined by the best - fit to the data for the sn ia @xmath271 and fraction of gas in galaxy clusters @xmath272 available from refs.@xcite and @xcite , respectively .
the shape of the likelihood function @xmath273 in terms of @xmath181 shows that it is not possible to determine this parameter by maximization .
we were forced to select some values of @xmath181 and proceed to the best - fit of the other parameters in the set @xmath166 $ ] .
it was clear that the @xmath3 is statistically independent of the initial conditions , regardless the values of @xmath181 and the other best - fit parameters .
in addition the sign of both @xmath192 and @xmath160 depend on @xmath181 , so that the behaviour of the model towards the future is not unique .
the confidence regions for the pair of parameters @xmath220 , @xmath219 and @xmath262 were built .
they showed that @xmath191 and @xmath192 are strongly correlated parameters , whereas @xmath191 and @xmath3 are weakly correlated quantities .
this was confirmed by the results in table [ table - covariances ] .
the plot of the confidence region for the pair @xmath274 is relevant to understand the future evolution of our higher - order model , which can exhibit either a premature big rip or a rebouncing depending on the value of @xmath181 and on the values of @xmath192 and @xmath160 .
if additional data for high @xmath110 objects were used , as gamma ray bursts , for instance , they might reduce the uncertainties in our estimations of the jerk and the snap , and help us to decide if the rebouncing scenario is favored with respect to the big rip , or conversely .
a positive feature of the particular @xmath16-model presented here is its consistency with the standard @xmath10cdm dynamics at recent times .
maybe this is anticipated because the model has a great number of free parameters to be fit to the data .
the high number of kinematic parameters ( i.e. , those involving acceleration and its first and second derivatives ) is a consequence of the model being of the fifth - order in derivatives of @xmath275 , a fact that could always make room for agreement with the observations .
however , the freedom for bias is not unlimited .
even with all these free parameters there is a constraint to the dynamics of the model towards the past . besides this consistency with the standard @xmath10cdm dynamics ( at recent times )
the model does not reproduce a large scale structure formation phase .
this shows that just fitting a model to actual data is not enough to falsify or prove the model .
general essential characteristic must be present too , as the instabilities allowing the large scale structure formation .
therefore , the extra term presented in the action ( [ action ] ) should be ruled out as viable alternative for dark energy .
this paper is dedicated to prof .
mario novello on the ocasion of his 70th birthday .
rrc thanks fapemig - brazil ( grant cex
apq04440 - 10 ) for financial support .
camm is grateful to fapemig - brazil for partial support .
lgm acknowledges fapern - brazil for financial support .
d. n. spergel _
et al_. , first year wilkinson microwave anisotropy probe ( wmap ) observations : determination of cosmological parameters , _ astrophys .
j. suppl . _
* 148 * ( 2003 ) , 175 ; e. komatsu _ et al_. , seven - year wilkinson microwave anisotropy probe ( wmap ) observations : cosmological interpretation , _ astrophys .
_ * 192*:18 ( 2011 )
. planck collaboration : p. a. r. ade _ et al_. , planck 2013 results . i. overview of products and scientific results , arxiv : 1303.5062 ( 2013 ) ; planck collaboration : p. a. r. ade _
et al_. , planck 2013 results .
cosmological parameters , arxiv : 1303.5076 ( 2013 ) .
d. j. eisenstein _
et al_. , detection of the barion acoustic peak in the large - scale correlation function of sdss luminous red galaxies , _ astrophys .
j. _ * 633 * ( 2005 ) , 560 ; w. j. percival _
et al . _ ,
baryon acoustic oscillations in the sloan digital sky survey data release 7 galaxy sample , _ mon . not .
_ * 401 * ( 2010 ) , 2148 .
_ , the 2df galaxy redshift survey : power - spectrum analysis of the final dataset and cosmological implications , _ mon . not
* 362 * ( 2005 ) , 505 ; f. beutler , the 6df galaxy survey : baryon acoustic oscilations and the local hubble constant , _ mon . not .
astron . soc_.
* 416 * ( 2011 ) , 3017 .
a. g. riess _
el al . _ ,
type ia supernova discoveries at @xmath279 from the hubble space telescope : evidence for past deceleration and constraints on dark energy evolution , _ astrophys .
j. _ * 607 * ( 2004 ) , 665 ; a. g. riess _
el al . _ , new hubble space telescope discoveries of type ia supernovae at @xmath279 : narrowing constraints on the early behavior of dark energy , _ astrophys .
j. _ * 659 * ( 2007 ) , 98 ; n. suzuki _
_ , the hubble space telescope cluster supernova survey : v. improving the dark energy constraints above @xmath279 and building an early - type - hosted supernova sample , _ astrophys . j. _ * 746 * ( 2012 ) , 85 . | this paper analyses the cosmological consequences of a modified theory of gravity whose action integral is built from a linear combination of the ricci scalar @xmath0 and a quadratic term in the covariant derivative of @xmath0 .
the resulting friedmann equations are of the fifth - order in the hubble function .
these equations are solved numerically for a flat space section geometry and pressureless matter .
the cosmological parameters of the higher - order model are fit using sn ia data and x - ray gas mass fraction in galaxy clusters .
the best - fit present - day @xmath1 values for the deceleration parameter , jerk and snap are given .
the coupling constant @xmath2 of the model is not univocally determined by the data fit , but partially constrained by it .
density parameter @xmath3 is also determined and shows weak correlation with the other parameters .
the model allows for two possible future scenarios : there may be either a premature big rip or a rebouncing event depending on the set of values in the space of parameters .
the analysis towards the past performed with the best - fit parameters shows that the model is not able to accommodate a matter - dominated stage required to the formation of structure .
@xmath4instituto de cincia e tecnologia , universidade federal de alfenas.rod .
jos aurlio vilela ( br 267 ) , km 533 , n@xmath511999 , cep 37701 - 970 , poos de caldas , mg , brazil.@xmath6instituto de fsica terica , universidade estadual paulista.rua bento teobaldo ferraz 271 bloco ii , p.o .
box 70532 - 2 , cep 01156 - 970 , so paulo , sp , brazil.@xmath7escola de cincia e tecnologia , universidade federal do rio grande do norte . campus universitrio , s / n - lagoa nova , cep 59078 - 970 , natal , rn , brazil.@xmath8inst .
de fomento e coordenao industrial , departamento de cincia e tecnologia aeroespacial .
praa mal .
eduardo gomes 50 , cep 12228 - 901 , so jos dos campos , sp , brazil .
@xmath9 instituto tecnolgico de aeronatica , departamento de cincia e tecnologia aeroespacial .
praa mal .
eduardo gomes 50 , cep 12228 - 900 , so jos dos campos , sp , brazil . * keywords : * gauge theory , cosmic acceleration , higher order gravity , cosmology.**pacs:**98.80.-k , 11.15.-q . |
in this paper we study some generalizations of the gromov - thurston @xmath4 theorem @xcite .
informally , the @xmath4 theorem states that `` most '' dehn fillings of a hyperbolic @xmath5manifold with cusps admit negatively curved metrics .
moreover , these negatively curved metrics are close approximations of the metric on the original cusped manifold .
group - theoretically , the fundamental groups of these fillings are relatively hyperbolic ( defined below ) .
if all cusps of the original manifold were filled , the fundamental group of the filling is hyperbolic ; if moreover the original manifold was finite volume , the filling is an aspherical closed @xmath5manifold . in the world of coarse geometry , the correct generalizations of fundamental groups of hyperbolic manifolds with and without cusps are _ relatively hyperbolic _ and _ hyperbolic _ groups , respectively .
these generalizations were introduced by gromov in @xcite .
we give a quick review of the definitions : a metric space is _
@xmath6hyperbolic _ for @xmath7 if all its geodesic triangles are @xmath6thin , meaning each side of the triangle is contained in the @xmath6neighborhood of the other two .
a space is _ gromov hyperbolic _ if it is @xmath6hyperbolic for some @xmath6 .
groups which are hyperbolic or relatively hyperbolic are those which have particular kinds of actions on gromov hyperbolic spaces .
if @xmath8 acts by isometries on a gromov hyperbolic space @xmath9 properly and cocompactly , then @xmath8 is said to be _
a _ geometrically finite _ action of a group @xmath8 on a gromov hyperbolic space @xmath9 is one which satisfies the following conditions : 1 .
@xmath8 acts properly and by isometries on @xmath9 .
2 . there is a family @xmath10 of disjoint _ horoballs _
( sub - level sets of so - called _ horofunctions _ ) , preserved by the action .
the action on @xmath11 is cocompact .
the quotient @xmath12 is quasi - isometric to a wedge of rays .
( see @xcite for more details . )
if @xmath8 acts on @xmath9 geometrically finitely , then @xmath8 is said to be _ relatively hyperbolic _ , relative to the _ peripheral _ subgroups , which are the stabilizers of individual horoballs .
( typically the list of peripheral subgroups is given by choosing one such stabilizer from each conjugacy class . ) in @xcite and @xcite , group - theoretic analogues of the @xmath4 theorem were proved .
these theorems roughly state that if one begins with a relatively hyperbolic group , and kills off normal subgroups of the peripheral subgroups which contain no `` short '' elements , then the resulting group is also relatively hyperbolic , and `` close '' to the original group in various ways . from this one recovers a weaker group - theoretic version of the @xmath4 theorem , as the fundamental group of a hyperbolic @xmath5manifold is relatively hyperbolic , relative to the cusp groups ; the operation of dehn filling acts on fundamental groups by killing the cyclic subgroups generated by the filling slopes . in the context of the original @xmath4 theorem , we begin with a group which has such an action on @xmath13 , which is not only @xmath6hyperbolic , but which is actually cat@xmath3 .
( see section [ s : prelim ] for the definition of cat@xmath14 for @xmath15 . for
now , just note that cat@xmath3 implies gromov hyperbolic . )
moreover , the group acts cocompactly on a `` neutered '' @xmath13 , which is cat@xmath2 with isolated flats .
the fundamental group of the filled manifold acts properly and cocompactly on a cat@xmath3 @xmath5manifold . for @xmath16
, we say that a group is _
cat@xmath14 _ if it acts isometrically , properly , and cocompactly on some cat@xmath14 space
. we will say that a group is _
relatively cat@xmath3 _ if it has a geometrically finite action on a cat@xmath3 space .
a group is _ cat@xmath2 with isolated flats
_ if it acts isometrically , properly , and cocompactly on some cat@xmath2 space with isolated flats .
( it follows from a result of hruska and kleiner @xcite that horoballs can always be added to such a space to make it @xmath6hyperbolic for some @xmath6 .
in particular , a group which is cat@xmath2 with isolated flats is relatively hyperbolic . )
it is natural to wonder to what extent the main results of @xcite and @xcite can be strengthened in the presence of these stronger conditions .
if one performs relatively hyperbolic dehn filling on a relatively cat@xmath3 group , is the filled group always relatively cat@xmath3 ?
if one performs relatively hyperbolic dehn filling on a cat@xmath2 with isolated flats group , is the filled group always cat@xmath2 with isolated flats ?
these questions form a part of the motivation for this paper ; we answer them positively in some special cases .
the utility of these answers is illustrated by using the additional cat@xmath2 structure of the quotient groups to obtain information about the filled groups which is less accessible from the coarse geometric viewpoint . in section
[ filling ] , we describe the main results . in section
[ s : prelim ] , we recall the definition of warped product metric and an important result of alexander and bishop .
we also discuss the space of directions and the logarithm map from a cat@xmath2 space to the tangent cone at a point . in section [
s : basic ] , we extend original construction of the @xmath4 theorem to our setting , using doubly warped product constructions to give locally cat@xmath2 and locally cat@xmath3 models for our fillings .
section [ s : visbound ] contains the `` morse theoretic '' arguments which give information about the boundary at infinity of the fundamental groups of our fillings . in section [ s : questions ] , we pose some questions .
fujiwara started working on this project during his visit at max planck institute in bonn in the fall of 2005 .
he has benefited from conversation with potyagailo .
the joint project with manning started when he visited caltech in april 2006 .
he thanks both institutions for their hospitality .
he is partially supported by grant - in - aid for scientific research ( no .
19340013 ) . manning thanks ric ancel for explaining how to prove claim [ horiz ] in the proof of proposition [ p : pinch ] , and thanks noel brady and daniel groves for useful conversations .
he was partially supported by nsf grants dms-0301954 and dms-0804369 .
we first establish some notation : [ def.cone ] let @xmath17 be a flat torus of dimension @xmath18 , and let @xmath19 be a totally geodesic @xmath20dimensional submanifold in @xmath17 , where @xmath21 .
the torus @xmath17 has euclidean universal cover @xmath22 ; let @xmath23 be a component of the preimage of @xmath19 in @xmath17 .
the parallel copies of @xmath24 are the leaves of a fibration @xmath25 .
this fibration covers a fibration @xmath26 of @xmath17 over some @xmath27torus @xmath28 .
we define the _ partial cone _
@xmath29 to be the mapping cylinder of @xmath30 , i.e. @xmath31 / \sim,\ ] ] where @xmath32 if @xmath33 .
we refer to @xmath34 as the _ boundary _ of @xmath29 , and to @xmath35 as the _
core_. the space @xmath36 is the cone on @xmath17 . in the case that @xmath17 is a metric product @xmath37 , then @xmath29 can be canonically identified with @xmath38 . ( although the identification is not canonical in general , @xmath29 is always homeomorphic to @xmath39 for some torus @xmath28 of complementary dimension to that of @xmath19 . in particular ,
the core @xmath40 is always homeomorphic to a torus of dimension @xmath41 . )
the partial cone @xmath29 is a manifold with boundary only when @xmath42 .
let @xmath43 be a hyperbolic @xmath44manifold .
a _ toral cusp _ of @xmath43 is a neighborhood @xmath45 of an end of @xmath43 so that 1 .
@xmath45 is homeomorphic to the product of a torus with the interval @xmath46 $ ] , and 2 .
the induced riemannian metric on @xmath47 is flat . for the rest of the section we make the following : [ standing ] @xmath43 is a hyperbolic @xmath44manifold of finite volume containing disjoint toral cusps @xmath48 , and no other ends .
let @xmath49 be the manifold with boundary obtained by removing the interiors of the cusps .
let @xmath50 be the components of @xmath51 .
[ d : filling ] for each component @xmath52 of @xmath51 choose an embedded totally geodesic torus @xmath53 of dimension @xmath54 .
form @xmath55 as in the previous definition .
we use the canonical identification @xmath56 of @xmath57 with @xmath52 to form the space @xmath58 we say that @xmath59 is _ obtained from @xmath43 by filling along @xmath60_. we refer to the cores of the @xmath55 as the _ filling cores_. if no @xmath53 contains a geodesic of length @xmath4 or less , we say that @xmath59 is a _ @xmath4filling of @xmath43_. the space @xmath59 described above is homeomorphic to a manifold if and only if every filling core has dimension exactly @xmath61 . in any case
@xmath59 is a pseudomanifold of dimension @xmath44 , and the complement of the filling cores is homeomorphic to @xmath43 . on the level of fundamental groups
, @xmath62 is obtained from @xmath63 by adjoining relations representing the generators of some direct summands of the cusp subgroups ; namely , for each @xmath64 , one adjoins relations corresponding to the generators of @xmath65 .
our main theorem says that under hypotheses exactly analogous to those used in the @xmath4 theorem , a space obtained by filling as above can be given a nice locally cat@xmath2 metric ; moreover , if all the filling cores are zero or one dimensional , the metric can be chosen to be locally cat@xmath3 .
we remind the reader that , under the condition [ standing ] , @xmath43 is assumed to be a hyperbolic @xmath44manifold with only toral cusps .
[ t : filling ] let @xmath66 and assume the condition [ standing ] .
if @xmath59 is a @xmath4filling of @xmath43 , then there is a complete path metric @xmath0 on @xmath59 satisfying the following : 1 .
[ fill1 ] the path metric @xmath0 is the completion of a path metric induced by a negatively curved riemannian metric on the complement of the filling cores . 2 .
[ fill2 ] the path metric @xmath0 is locally cat@xmath2 .
[ fill3 ] if for every @xmath67 , @xmath53 has codimension at most @xmath68 in @xmath69 , then @xmath0 is locally cat@xmath70 for some @xmath71 .
the metric constructed in theorem [ t : filling ] is identical to the original hyperbolic metric , away from a neighborhood of the filling cores .
we will prove the following in the course of showing theorem [ t : filling ] .
[ p : hyper ] let @xmath66 and assume the condition [ standing ] , and let @xmath59 be a @xmath4filling of @xmath43 .
let @xmath72 be as described in definition [ d : filling ] .
choose @xmath73 so that the @xmath74neighborhood of @xmath75 is embedded in @xmath43 .
let @xmath76 be the complement in @xmath43 of the @xmath77neighborhood of @xmath75 .
we thus have inclusions @xmath78 , and @xmath79 .
the riemannian metric described in theorem [ t : filling ] is equal to the hyperbolic metric , when restricted to @xmath76 .
let @xmath80 be the fundamental group of some @xmath4filling of @xmath43 .
it follows from theorem [ t : filling ] that @xmath81 is cyclic .
one can ask what the gromov norm of a generator is , and define this number to be the _ simplicial volume _ of @xmath59 ( or of @xmath8 ) . in case @xmath59
is a manifold , this definition is the same as gromov s in @xcite . in the setting of the original @xmath4 theorem , both @xmath43 and @xmath59 are hyperbolic manifolds , and it was shown by thurston that the simplicial volume of @xmath59 is bounded above by the hyperbolic volume of @xmath43 , divided by a constant @xcite .
we extend thurston s result to higher dimensions in @xcite .
[ r : specialcases ] special cases of theorem [ t : filling ] appear in * the gromov - thurston @xmath4theorem @xcite , in which @xmath82 and each filling core is a circle , * schroeder s generalization of the @xmath4 theorem to @xmath83 , with filling cores equal to @xmath84dimensional tori @xcite ( this is the case in which every @xmath53 is one - dimensional , and the only case in which the filling @xmath59 is a manifold . ) , and * mosher and sageev s construction in @xcite of @xmath44dimensional hyperbolic groups which are not aspherical manifold groups ; here the filling cores are points .
the construction of mosher and sageev is based on a discussion in ( * ? ? ?
* a.vi ) , and the examples constructed by mosher and sageev answered a question which appeared in an early version of bestvina s problem list @xcite .
the new examples of the current paper fall in some sense _ between _ those of schroeder , and those of mosher - sageev .
our construction gives the following corollary , which can be deduced from theorem [ t : filling ] and proposition [ p : iflats ] .
[ c : cwif ] let @xmath66 and assume the condition [ standing ] .
let @xmath59 be a @xmath4filling of @xmath43 , and let @xmath8 be the fundamental group of @xmath59 .
then @xmath8 is cat@xmath2 with isolated flats . if each @xmath53 has dimension @xmath18 or @xmath61 , then @xmath8 is cat@xmath3 .
thus theorem [ t : filling ] says that given a cusped hyperbolic @xmath44manifold with @xmath66 , there are many ways to fill its cusps to get a space whose universal cover is cat@xmath70 for @xmath71 .
if @xmath85 , there are many ways to fill its cusps to get a space whose universal cover is cat@xmath2 with isolated flats .
the following proposition says that different fillings generally have different fundamental groups : [ p : different ] let @xmath66 and assume the condition [ standing ] . for each @xmath86 choose a filling @xmath87 of @xmath43 , and suppose that the fillings @xmath88 satisfy : for each @xmath89 and each element @xmath90 of @xmath91 , the set @xmath92 is finite .
then for any @xmath20 , there are only finitely many @xmath86 so that @xmath93 .
we argue by contradiction .
the hypotheses imply that all but finitely many fillings @xmath88 are @xmath4fillings . by passing to a subsequence
we may suppose that all the @xmath94 are isomorphic to some fixed @xmath8 .
applying corollary [ c : cwif ] and ( * ? ? ?
* theorem 1.2.1 ) , @xmath8 is relatively hyperbolic with respect to a collection of free abelian subgroups . if @xmath95 we have a sequence of surjections @xmath96 obtained by killing subgroups @xmath97 of the peripheral subgroups of @xmath98 corresponding to the tori @xmath99 .
since the injectivity radii of these tori are going to infinity , the length ( with respect to a fixed generating set for @xmath98 ) of a shortest nontrivial element of @xmath100 is also going to infinity .
it follows from corollary 9.7 of @xcite that the _ stable kernel _
@xmath101 is trivial .
work of groves then shows that @xmath98 is either abelian or splits over an abelian group ( * ? ? ?
* proposition 2.11 ) .
but @xmath98 is the fundamental group of a finite volume hyperbolic manifold of dimension at least @xmath5 , so it can not split over a virtually abelian subgroup ( see , e.g. ( * ? ? ?
* theorem 1.3.(i ) ) )
. we will show ( corollary [ c : notpdn ] ) that no @xmath4filling of an @xmath44dimensional hyperbolic manifold with filling cores of dimension @xmath102 has the same fundamental group as a closed hyperbolic manifold ( or even a closed aspherical manifold ) for @xmath85 .
this is in sharp contrast to the situation for hyperbolic @xmath5manifolds . combining corollary
[ c : notpdn ] and proposition [ p : different ] immediately yields the following : [ t : examples ] let @xmath85 and assume the condition [ standing ] .
there are infinitely many @xmath4fillings of @xmath43 whose fundamental groups are non - pairwise - isomorphic , torsion - free , @xmath44dimensional hyperbolic groups , each of which acts geometrically on some @xmath44dimensional cat@xmath3 space , and none of which is a closed aspherical manifold group . if @xmath103 , there are infinitely many @xmath4fillings of @xmath43 whose fundamental groups are non - pairwise - isomorphic , torsion - free , @xmath44dimensional relatively hyperbolic groups , each of which acts geometrically on some @xmath44dimensional cat@xmath2 space with isolated flats , and none of which is a closed aspherical manifold group . flats of any dimension up to @xmath104 occur in infinitely many such fillings . in @xcite , januszkiewicz and
wiatkowski introduce _ systolic groups _
, in part as a means of giving examples of non - manifold @xmath18dimensional hyperbolic groups for all @xmath18 ( see @xcite for other approaches ) .
recently , osajda has shown that systolic groups are never simply connected at infinity @xcite .
it follows from our analysis of the group boundary in section [ s : visbound ] that the high - dimensional groups obtained in theorem [ t : examples ] are simply connected at infinity whenever no filling cores are points .
in particular , these groups are not systolic ( corollary [ c : notsys ] ) .
for general results and terminology for cat@xmath14 spaces we refer to bridson and haefliger @xcite . for convenience , we remind the reader of the definition .
let @xmath15 , and let @xmath105 be the complete simply connected surface of constant curvature @xmath106 .
let @xmath107 be the diameter of @xmath105 .
now let @xmath9 be a length space .
( see @xcite for more on length spaces . )
if @xmath108 is a geodesic triangle in @xmath9 whose perimeter is bounded above by @xmath109 , with corners @xmath110 , @xmath111 , and @xmath112 , then there is a _ comparison triangle _ @xmath113 ; i.e. , the corners of @xmath114 can be labelled @xmath115 , @xmath116 , and @xmath117 , and there is a map from @xmath118 to @xmath114 which is an isometry restricted to any edge of @xmath118 , taking @xmath110 to @xmath115 , and so on ( see figure [ f : compare ] ) .
[ f : compare ] if this map never decreases distance on any triangle @xmath118 whose perimeter is bounded above by @xmath109 , we say that @xmath9 is _ cat@xmath14_. put informally , @xmath9 is cat@xmath14 if its triangles are `` no fatter '' than the comparison triangles in @xmath105 . if every point in @xmath9 has a neighborhood which is cat@xmath14 , then we say that @xmath9 is _ locally cat@xmath14_. if @xmath9 is locally cat@xmath2 , then we say that @xmath9 is _
nonpositively curved_. for @xmath16 , the universal cover of a complete , locally cat@xmath14 space is cat@xmath14 , by the cartan - hadamard theorem .
we take this definition from @xcite .
given two complete length spaces @xmath119 and @xmath120 , and a continuous nonnegative @xmath121 , we can define the _ warped product of @xmath9 with @xmath122 , with warping function @xmath123 _ to be the length space based on the following method of measuring lengths of paths . given any closed interval @xmath124
$ ] and any lipschitz path @xmath125 , with @xmath126 , the quantities @xmath127 and @xmath128 are well defined almost everywhere on @xmath129 ( see @xcite ) .
define the length of @xmath90 , @xmath130 to be @xmath131 the quotient map @xmath132 is injective exactly when @xmath123 has no zeroes . if @xmath133 then @xmath134 is identified to a point in @xmath135 .
often we will have @xmath136 , or @xmath137 for @xmath138 , and @xmath123 depending only on @xmath139 . in this case , we will usually write @xmath140 as @xmath141 .
for example : * the euclidean plane is isometric to @xmath142 .
the subset @xmath143 is a circle of radius @xmath144 . *
the hyperbolic space @xmath145 is isometric to @xmath146 .
the subset @xmath147 is a horosphere .
we will often abuse notation by using points in @xmath148 to refer to points in @xmath140 , even when the quotient map is not injective .
for example the cone point of the euclidean cone @xmath149 might be referred to as @xmath150 for some @xmath151 , or just as @xmath152 . in case
the metrics @xmath153 and @xmath154 are induced by riemannian metrics @xmath155 and @xmath156 and @xmath157 is smooth , the warped product metric on @xmath140 is equal to the path metric induced by a riemannian metric @xmath158 , given at a point @xmath159 by the formula @xmath160 curvature bounds on warped product spaces are studied by alexander and bishop in @xcite , where quite general sufficient conditions are given for a warped product to have curvature bounded from above or below . in order to state the results from @xcite which we need , we first recall some definitions .
let @xmath161 .
let @xmath162 be the set of solutions to the differential equation @xmath163 .
a function @xmath123 satisfying @xmath164 for any function @xmath158 in @xmath162 agreeing with @xmath123 at two points which are sufficiently close together is said to _ satisfy the differential inequality @xmath165 in the barrier sense .
_ let @xmath166 , and let @xmath28 be a geodesic space .
a continuous function @xmath167 is _
@xmath162convex _ if its composition with any unit - speed geodesic satisfies the differential inequality @xmath165 in the barrier sense .
we collect here some useful facts about @xmath162convexity : [ convex ] 1 .
a function is @xmath168convex if and only if it is _ convex _ in the usual sense .
[ c - fort ] a non - negative @xmath162convex function is also @xmath169convex for any @xmath170 .
* theorem 1.1(1a ) ) if @xmath28 is cat@xmath3 and @xmath171 , then @xmath172 is a @xmath173convex function on @xmath28 .
[ c - dist ] ( * ? ? ?
* theorem 1.1(3a ) ) if @xmath174 is a convex subset of a cat@xmath2 space @xmath28 , then @xmath175 is a convex function on @xmath28 . 5 .
[ c - sinh ] ( * ? ? ?
* theorem 1.1(4a ) ) if @xmath174 is a convex subset of a cat@xmath3 space @xmath28 , then @xmath176 is a @xmath173convex function on @xmath28 .
we now state a special case of a more general result in @xcite .
[ t : ab ] ( * ? ? ? * theorem 1.1 ) let @xmath177 , and let @xmath178 .
let @xmath28 and @xmath179 be complete locally compact cat@xmath180 and cat@xmath181 spaces , respectively .
let @xmath182 be @xmath162convex .
let @xmath183 .
suppose further that either 1 .
@xmath184 and @xmath185 , or 2 .
@xmath9 is nonempty and @xmath186 whenever @xmath187\to b$ ] is a shortest ( among all paths starting in @xmath9 ) unit - speed geodesic from @xmath9 to a point @xmath188 .
then the warped product @xmath189 is cat@xmath180 .
suppose that @xmath9 is a length space , and @xmath190 .
recall that the _ space of directions at @xmath191 _ is the collection of geodesic segments issuing from @xmath191 , modulo the equivalence relation : @xmath192\sim[p , y]$ ] if the alexandroff angle between @xmath192 $ ] and @xmath193 $ ] at @xmath191 is zero .
the alexandroff angle at @xmath191 makes the space of directions , @xmath194 , into a metric space .
if @xmath9 is a riemannian manifold , then @xmath194 is always a round sphere of diameter @xmath30 .
the _ tangent cone at @xmath191 _ is the euclidean cone on the space of directions at @xmath191 : @xmath195 we will sometimes just write @xmath196 and @xmath197 if the space @xmath9 is understood .
if @xmath9 is a riemannian manifold near the point @xmath191 , then the tangent cone @xmath197 can be canonically identified with the tangent space of @xmath9 at @xmath191 . in our setting @xmath9
will always be geodesically complete and locally compact . with these assumptions ,
if @xmath198 is any open neighborhood of @xmath190 , then @xmath199 is isometric to the pointed gromov - hausdorff limit of spaces @xmath200 isometric to @xmath198 with the metric multiplied by factors @xmath201 going to infinity ( see for example ( * ? ? ?
* theorem 9.1.48 ) ) .
alexander and bishop @xcite observe that the space of directions at a point in a warped product is determined by the spaces of directions in the factors together with infinitesimal information about the warping functions .
if @xmath167 is a convex function and @xmath202 , then @xmath123 has a well - defined `` derivative '' @xmath203 at the point @xmath191 ; this derivative is linear on each ray .
[ d : sphjoin ] let @xmath174 and @xmath28 be metric spaces .
we may give the product @xmath204\times a\times b$ ] a pseudometric as follows .
let @xmath205 and @xmath206 .
the distance @xmath207 is given by @xmath208 where @xmath209 means @xmath210 . the _ spherical join _
@xmath211 of @xmath174 and @xmath28 is the canonical metric space quotient of @xmath204\times a\times b$ ] with the above metric .
points in @xmath211 can be written as triples @xmath212 with @xmath213 $ ] , @xmath214 and @xmath215 . if @xmath216 , the third coordinate can be ignored , and if @xmath217 , the second coordinate can be ignored .
@xcite[p : dirwarp ] let @xmath28 , @xmath179 , and @xmath123 be as in theorem [ t : ab ] .
let @xmath218 , and let @xmath219 . if @xmath220 , then there are isometries @xmath221 and @xmath222 if @xmath223 , there are isometries @xmath224 and @xmath225 let @xmath9 be a uniquely geodesic metric space and let @xmath190 . in general
, there is no _ exponential map _ defined on the entire tangent cone @xmath197 .
however , there is a well - defined _ logarithm map _ from @xmath9 to @xmath197 @xcite .
let @xmath9 be a uniquely geodesic metric space , and let @xmath226 be the tangent cone at a point @xmath190 .
if @xmath227 , then @xmath228 is the equivalence class of @xmath229)$ ] in @xmath197 .
if the metric is only locally uniquely geodesic , then the logarithm map can still be defined in a neighborhood of each point .
the existence of a local exponential map is limited by the extent to which the logarithm is injective and surjective .
[ d : linj ] the _ log - injectivity radius at @xmath191 _
, @xmath230 , is be the supremum of those numbers @xmath231 for which @xmath232 is well - defined and injective on the ball of radius @xmath231 about @xmath191 .
a neighborhood of @xmath191 on which @xmath232 is injective is called a _ log - injective neighborhood of @xmath191_. similarly , the _ log - surjectivity radius at @xmath191 _ ,
@xmath233 , is the supremum of those numbers @xmath231 for which @xmath234\times_r\sigma_p\subset c_p$ ] is contained in the image of @xmath232 .
in general , neither @xmath230 nor @xmath233 is positive .
recall that a space has the _ geodesic extension property _
if any locally geodesic segment terminating at @xmath111 can be extended to a longer locally geodesic segment terminating at some @xmath235 .
a complete cat@xmath2 space satisfying the geodesic extension property has @xmath236 .
[ obs : exp ] if @xmath237 , then there is an _ exponential map _
@xmath238 which is a homeomorphism onto the open @xmath144neighborhood of @xmath191 .
more generally there is always an exponential map @xmath239 for any log - injective neighborhood @xmath198 of @xmath191 , and this map is a homeomorphism onto its image .
a subset @xmath240 of a uniquely geodesic space is _ convex _ if for any two points in @xmath240 , the geodesic connecting them is contained in @xmath240 .
the subset @xmath240 is _ strictly convex _ if it is convex and if the frontier of @xmath240 contains no non - degenerate geodesic segment .
[ l : dconvex ] let @xmath9 be a proper uniquely geodesic space satisfying the geodesic extension property .
suppose that @xmath241 is a compact strictly convex set in @xmath9 containing @xmath191 in its interior , and suppose that @xmath241 is contained in a log - injective neighborhood of @xmath191 .
it follows that the frontier of @xmath241 is homeomorphic to @xmath196 .
since under the hypotheses both the space of directions @xmath196 and the frontier of @xmath241 ( which we denote here by @xmath242 ) are compact hausdorff spaces , it suffices to find a continuous bijection from one to the other .
let @xmath243 be the map which sends a point @xmath110 to the equivalence class of @xmath244 $ ] . since @xmath245 , this map is well - defined .
since @xmath9 is a proper uniquely geodesic space , geodesic segments vary continuously with their endpoints ( * ? ? ?
* i.3.13 ) .
it follows that the map @xmath246 is continuous .
we next show that @xmath246 is surjective .
let @xmath247 , and choose a geodesic segment beginning at @xmath191 with direction @xmath248 .
since @xmath9 satisfies the geodesic extension property , this geodesic segment can be extended to a point @xmath110 not in @xmath241 . some point on the geodesic
@xmath192 $ ] must be in the frontier of @xmath241 , and so @xmath248 is in the image of @xmath241 .
finally we show that @xmath246 is injective .
choose @xmath247 , and suppose that @xmath249 for some @xmath250 in @xmath242 . without loss of generality ,
suppose that @xmath251 .
since @xmath241 is strictly convex , the geodesic from @xmath110 to @xmath111 must be contained in @xmath241 but can not be contained entirely in @xmath242 .
thus there is some point @xmath252 $ ] so that @xmath112 is in the interior of @xmath241 . by the hypotheses and observation [ obs : exp ]
there is an open set @xmath253 containing @xmath254 so that the exponential map @xmath255 is a homeomorphism from @xmath198 to its image .
the point @xmath256 is contained in some basic open subset of @xmath198 of the form @xmath257 where @xmath240 is an open set in @xmath196 containing @xmath248 .
but since @xmath241 is convex , @xmath254 must also contain the union of the segments beginning at @xmath191 and ending in @xmath257 ; in other words , @xmath254 must contain @xmath258 .
in particular , @xmath241 contains an open neighborhood of any point on the geodesic between @xmath191 and @xmath112 , and thus @xmath110 must be an interior point of @xmath241 , which is a contradiction .
in this section we prove theorem [ t : filling ] by constructing suitable warped product metrics on the `` partial cones '' @xmath55 ( see definition [ def.cone ] ) so that those metrics are compatible with the hyperbolic metric on @xmath259 .
the metric will be shown to be cat@xmath2 by applying theorem [ t : ab ] near the singular part , and by computing the sectional curvatures in the riemannian part .
we first make an observation about warped products , whose proof we leave to the reader .
[ l : tcs ] let @xmath138 be connected , and let @xmath260 , and let @xmath179 be a geodesic metric space .
suppose that @xmath123 has a local minimum at @xmath261 . 1 .
@xmath262 is a convex subset of @xmath263 .
2 . let @xmath264 , and @xmath265 .
the shortest path from @xmath266 to @xmath267 in @xmath268 is @xmath269\times \{x\}$ ] .
we next give some applications of theorem [ t : ab ] .
[ l : step1 ] let @xmath45 be complete and cat@xmath270 for @xmath271 $ ] .
the warped product @xmath272 is cat@xmath270 , and @xmath273 is convex . since @xmath274 is @xmath275 convex , it is @xmath276convex , by [ convex ] .. the zero set @xmath277 is empty , and @xmath278 , so condition ( 1 ) of theorem [ t : ab ] is verified with @xmath279 .
it follows that @xmath28 is cat@xmath270 .
since the warping function @xmath274 has a minimum at @xmath280 , lemma [ l : tcs ] implies that @xmath281 is convex .
[ p : double ] suppose that @xmath45 is complete and cat@xmath270 for @xmath282 , and suppose that @xmath179 is cat@xmath283 .
the space @xmath284 is cat@xmath270 , and @xmath285 is convex . by lemma [ l : tcs ] , the set @xmath286 is a convex subset of @xmath287 .
on @xmath28 , the function @xmath288 is the distance to @xmath122 . by lemma [ l : step1 ] and
useful fact [ convex ] . , the function @xmath289 is convex on @xmath28 suppose @xmath290 .
the function @xmath291 is the composition of an increasing convex function with a convex function , hence it is convex .
if @xmath292 , then @xmath291 is @xmath275convex by [ convex ] .. in either case , @xmath291 is @xmath293convex , and so we can try to apply theorem [ t : ab ] to the warped product @xmath294 we must verify condition ( 2 ) of theorem [ t : ab ] , since @xmath295 is nonempty .
let @xmath296 .
we can apply the second part of lemma [ l : tcs ] to see that a shortest unit speed geodesic @xmath297 from @xmath179 to @xmath298 is of the form @xmath299 , with domain @xmath300 $ ] .
we thus have @xmath301 as required .
finally , we show that @xmath302 is convex .
if @xmath303\to w$ ] is any rectifiable path with both endpoints in @xmath268 , and @xmath304 , we note that @xmath305 is a shorter path with the same endpoints .
it follows that any geodesic with both endpoints in @xmath268 must lie entirely in @xmath268 . using the convention that @xmath306 is a point , the euclidean space @xmath307 is cat@xmath2 for all @xmath18 , and is cat@xmath3 for @xmath308 and @xmath309 .
we thus have : [ c : singular ] let @xmath19 be a complete cat@xmath283 space , and let @xmath20 be a nonnegative integer .
the warped product @xmath310 is complete and cat@xmath2 . if @xmath311 , then @xmath312 is cat@xmath3 .
[ l : completion ] let @xmath19 be a flat manifold .
the warped product @xmath310 is the metric completion of the riemannian manifold @xmath313 the manifold @xmath314 is clearly dense in @xmath312 , so it suffices to show that @xmath314 includes isometrically in @xmath312 .
let @xmath315 and @xmath316 be two points in @xmath314 , and let @xmath317 .
let @xmath90 be a lipschitz path in the product @xmath318 nearly realizing the distance between @xmath319 and @xmath320 in @xmath312 , so that the length in @xmath312 of @xmath90 is at most @xmath321 .
we will replace @xmath90 by a path @xmath322 in @xmath314 so that the length of @xmath322 exceeds the length of @xmath90 by at most @xmath323 , with @xmath324 . letting @xmath325 tend to zero , the lemma will follow . for @xmath326 in the domain of @xmath90
, we have @xmath327 .
if @xmath328 is positive for all @xmath326 , then @xmath90 stays inside @xmath314 , and there is nothing to prove .
we therefore assume that @xmath329 for @xmath326 in a ( possibly degenerate ) interval @xmath330 $ ] . by proposition [ p : double ]
, @xmath331 is convex , so we have @xmath332 for any @xmath333 $ ] .
the coordinates of @xmath90 are uniformly continuous in @xmath326 , so we can find small positive @xmath334 and @xmath335 so that @xmath336 and all the following are satisfied : @xmath337 let @xmath338\to \e^k$ ] be a constant - speed geodesic from @xmath339 to @xmath340 , and let @xmath341 $ ] be a constant - speed geodesic from @xmath342 to @xmath343 .
let @xmath322 be given by @xmath344 as the reader may check , the difference between the length of @xmath90 and the length of @xmath322 is at most @xmath345 . letting @xmath325 tend to zero ,
we have established the lemma .
if @xmath19 is a riemannian manifold , the warped product from corollary [ c : singular ] is a riemannian manifold in a neighborhood of any @xmath346 .
it follows that the space of directions at @xmath191 is a sphere .
the next lemma describes the space of directions at a non - manifold point .
[ l : dirjoin ] let @xmath312 be as in corollary [ c : singular ] .
the space of directions @xmath347 at a point @xmath348 of @xmath312 is isometric to the spherical join @xmath349 . proposition [ p : dirwarp ]
can be applied twice , as follows .
let @xmath350 , and apply the second part of proposition [ p : dirwarp ] to deduce @xmath351 next , since @xmath352 , we can apply the first part of proposition [ p : dirwarp ] to deduce that @xmath353 [ l : gep ]
suppose @xmath19 is a complete flat riemannian manifold with injectivity radius bigger than @xmath30 .
then @xmath354 has the geodesic extension property .
since @xmath312 is cat@xmath2 by corollary [ c : singular ] , geodesics are the same as local geodesics . away from @xmath355 ,
the space @xmath312 is riemannian , so geodesics can be extended in @xmath356 . at a point @xmath191 of @xmath240 , lemma [ l : dirjoin ] implies that the space of directions @xmath196 is isometric to @xmath349 .
suppose @xmath297 is a geodesic segment terminating at @xmath191 , and let @xmath357=(\phi,\alpha,\theta)\in \sigma_p$ ] be the direction of @xmath297 , where @xmath213 $ ] , @xmath358 , and @xmath359 .
since @xmath19 has injectivity radius bigger than @xmath30 , there is some @xmath360 with @xmath361 .
it is straightforward to check that the alexandrov angle between @xmath357 $ ] and @xmath362 is @xmath30 .
letting @xmath363 be any geodesic segment starting at @xmath191 , with direction @xmath362 , we see that @xmath363 geodesically extends @xmath297 past @xmath191 .
we next argue that the space @xmath312 described in corollary [ c : singular ] has log - injective neighborhoods at the singular points , in the sense of definition [ d : linj ] .
it suffices to show that geodesic segments emanating from a point in the singular set can only make an angle of zero if they coincide on an initial subsegment .
we first find the direction of an arbitrary geodesic segment emanating from a point in the singular set .
[ l : direction ] let @xmath19 be a flat manifold of injectivity radius larger than @xmath30 .
let @xmath364 be a point in @xmath365 , and let @xmath366 be some other point of @xmath312 ; that is , @xmath367 and @xmath368 for some @xmath369 , @xmath370 and @xmath371 and @xmath372 .
let @xmath297 be a geodesic from @xmath364 to @xmath366 . if @xmath373 , then the direction of @xmath297 at @xmath364 ( in the spherical join coordinates of definition [ d : sphjoin ] ) is @xmath374 .
otherwise let @xmath375 be the direction of a geodesic in @xmath45 from @xmath370 to @xmath376 ; the direction of @xmath297 is @xmath377 where @xmath378 if @xmath379 the lemma is obvious . otherwise , we identify @xmath380 with a geodesic @xmath90 in @xmath45 passing through @xmath370 and @xmath376 , setting @xmath381 .
the warped product @xmath382 contains the geodesic from @xmath364 to @xmath366 .
the map @xmath383 takes @xmath384 isometrically onto the hyperbolic half plane in the upper half - space model consisting of points with nonnegative real part .
the map @xmath246 takes @xmath364 to @xmath67 and @xmath366 to some point of modulus at least @xmath68 .
the geodesic @xmath90 is sent to the positive imaginary axis .
it is an exercise in hyperbolic geometry to verify that if the geodesic from @xmath385 to @xmath386 makes an angle of @xmath387 with @xmath388 , then @xmath389 .
the lemma follows .
[ c : loginj ] let @xmath390 lie in @xmath391 .
any open neighborhood of @xmath191 is log - injective at @xmath191 .
let @xmath368 and @xmath392 be points of @xmath354 , and suppose that @xmath393 in the tangent cone at @xmath191 .
let @xmath394 be the unique geodesic from @xmath191 to @xmath395 , for @xmath396 .
since @xmath397 , the geodesic segments @xmath398 and @xmath399 have the same length and the same direction @xmath400 as described in lemma [
l : direction ] .
in particular , the segments @xmath401 $ ] and @xmath402 $ ] in @xmath45 must have the same direction at @xmath403 .
the log - injectivity radius at any point in euclidean space is infinite , so these geodesic segments are subsets of a single line in @xmath45 . as in the proof of lemma
[ l : direction ] , the geodesics @xmath398 and @xmath399 both lie in a subset of @xmath312 isometric to a hyperbolic halfplane .
since a hyperbolic half - plane has infinite log - injectivity radius , @xmath398 and @xmath399 must coincide .
if one is merely interested in nonpositive curvature , and not log - injectivity , there is more flexibility in the choice of warping functions .
suppose that @xmath19 is a flat manifold .
it is not hard to show that the space @xmath404 is cat@xmath2 . on the other hand
, @xmath405 does not have log - injective neighborhoods at points @xmath348 .
the following proposition can be proved by a simple if tedious computation involving christoffel symbols .
see @xcite for a proof in case both flat factors are @xmath68dimensional .
[ p : sectional ] let @xmath406 and @xmath407 be flat manifolds , and let @xmath138 .
let @xmath408 , @xmath409 be positive smooth functions on @xmath129 .
the warped product manifold @xmath410 has sectional curvatures at @xmath411 which are convex combinations of the following functions , evaluated at @xmath326 : @xmath412 let @xmath413 . if @xmath414 , then the term @xmath415 can be ignored . if @xmath416 , then all terms involving @xmath417 and its derivatives can be ignored .
it follows from this proposition that if @xmath408 and @xmath409 are positive , convex , and increasing , the warped product @xmath418 is nonpositively curved .
if @xmath408 , @xmath409 , and their first and second derivatives are all bounded between two positive numbers @xmath419 and @xmath420 , then the curvature of @xmath418 is bounded between @xmath421 , and @xmath422 .
we will construct smooth convex increasing functions @xmath123 and @xmath158 on @xmath423 $ ] interpolating between the exponential function @xmath424 near @xmath425 , and the functions @xmath291 and @xmath274 near @xmath426 .
we will apply the following result of agol , proved in @xcite .
* lemma 2.5)[l : agol ] suppose @xmath427 where @xmath428 and @xmath429 are @xmath430 on @xmath431 , and @xmath432 , @xmath433 .
then we may find @xmath434 functions @xmath435 on @xmath431 for @xmath317 such that 1 .
there is a @xmath436 such that @xmath437 and @xmath438 for @xmath439 , @xmath440 for @xmath441 , and 2 .
@xmath442 @xmath443 agol s lemma allows us to smoothly interpolate between functions which agree up to first order at some point .
we wish to interpolate between pairs of functions ( either @xmath291 and @xmath424 or @xmath274 and @xmath424 ) which do not agree anywhere up to first order . to solve this difficulty ,
we introduce a third function which interpolates between the pair : [ l : interp ] let @xmath444 , @xmath445 .
suppose @xmath446 , and let @xmath447 $ ] be an interval containing @xmath448 ( where the graphs of @xmath449 and @xmath450 meet ) .
there is a @xmath451 and a smooth function @xmath452 on @xmath447 $ ] so that : 1 .
@xmath453 , 2 .
@xmath454 , and 3 .
@xmath455 on @xmath447 $ ] .
one way to do this is to start with the circle @xmath456 and let @xmath457 be the unique affine map taking @xmath458 to @xmath459 , @xmath460 to @xmath461 , and @xmath462 to the point of intersection of the graphs of @xmath449 and @xmath450 .
the resulting ellipse @xmath463 is tangent to the graph of @xmath449 at @xmath459 , and to the graph of @xmath450 at @xmath461 .
the part of the ellipse between these two points and closest to the union of the graphs of @xmath449 and @xmath450 is the graph of a function @xmath452 satisfying the requirements of the lemma .
see figure [ fig : affine ] for an illustration .
[ fig : affine ] [ p : warpfg ] for any @xmath73 there is a @xmath7 , @xmath451 , and a pair of smooth functions @xmath464 and @xmath465 satisfying : 1 . for @xmath466 , @xmath467 and @xmath468 , 2 . for @xmath469 , @xmath470 , and 3 . for @xmath471 , @xmath472 , and for @xmath473 , @xmath474 .
we first define @xmath475 functions @xmath476 and @xmath477 with the above properties .
we then invoke agol s gluing lemma ( lemma [ l : agol ] ) to obtain smooth functions .
let @xmath450 be the equation of the tangent line to @xmath478 at @xmath479 , and note that @xmath480 .
since @xmath481 , the line @xmath450 intersects the tangent lines at @xmath280 to both @xmath274 and @xmath291 somewhere in the interval @xmath482 . by continuity
, there is some @xmath483 so that the tangent lines to @xmath274 and @xmath291 at @xmath484 still hit @xmath450 somewhere in the interval @xmath482 .
in particular , if @xmath449 is the tangent line to @xmath291 at @xmath484 , we may apply lemma [ l : interp ] to obtain a function @xmath485 on @xmath486 with @xmath487 , @xmath488 , and @xmath489 everywhere , for some @xmath490 .
we define @xmath491 ( see figure [ f : interp ] ) .
choosing some positive @xmath492 , we may twice apply lemma [ l : agol ] ( at @xmath484 and at @xmath479 ) to obtain a smooth function @xmath493 satisfying : 1 .
@xmath494 outside the intervals @xmath495 and @xmath496 , 2 .
@xmath497 on @xmath495 , and 3 .
@xmath498 on @xmath499 .
( here @xmath500 can be replaced with any number less than @xmath68 . )
a similar argument can be used to construct @xmath158 .
lemma [ l : interp ] can be used to construct a function @xmath501 on the interval @xmath486 with @xmath502 , @xmath503 , and @xmath504 everywhere .
we can thus assemble a @xmath475 function : @xmath505 applying lemma [ l : agol ] twice to @xmath477 and using the same @xmath325 as above we obtain a function @xmath158 satisfying : 1 .
@xmath506 outside the intervals @xmath495 and @xmath496 , 2 .
@xmath507 on @xmath495 , and 3 .
@xmath508 on @xmath499 . the functions @xmath123 and @xmath158 satisfy the conclusion of the proposition for @xmath509 , and @xmath510 .
[ t : nonpos ] let @xmath123 and @xmath158 be as in proposition [ p : warpfg ] , and let @xmath19 be flat manifold with injectivity radius bigger than @xmath30 .
the warped product @xmath511\times_{g(r)}\e^k\times_{f(r)}t\ ] ] is riemannian away from @xmath512 .
the space @xmath268 is complete and cat@xmath2 , and is cat@xmath70 for some @xmath71 if @xmath311 .
since @xmath513 and @xmath514 are positive and smooth for @xmath471 , the space @xmath268 is riemannian away from @xmath515 . since @xmath516 , the space @xmath268 is simply connected , so it suffices to check the curvature locally .
near the singular set @xmath517 , the @xmath268 is locally isometric to a subset of the space @xmath312 described in corollary [ c : singular ] .
it follows that @xmath268 is locally cat@xmath2 ( and locally cat@xmath3 if @xmath311 ) in a @xmath6neighborhood of @xmath517 .
away from @xmath517 , we may estimate derivatives , and apply proposition [ p : sectional ] .
since @xmath123 , @xmath158 , and their first and second derivatives are all positive and continuous on the interval @xmath518 $ ] , there is a positive lower bound @xmath106 valid for all the quantities @xmath519 on this interval .
proposition [ p : sectional ] implies that the sectional curvatures are thus bounded above by @xmath520 , away from a @xmath6neighborhood of @xmath517 . putting the local pictures together
, @xmath268 is locally ( and hence globally ) cat@xmath70 if @xmath311 , and cat@xmath2 otherwise .
if @xmath43 is a hyperbolic manifold , and @xmath521 is a closed horospherical neighborhood of a toral cusp , then @xmath45 is isometric to a warped product @xmath522\times_{e^r}n,\ ] ] where @xmath523 with the induced flat riemannian metric .
( to see this , recall that if we identify @xmath145 with @xmath146 , then sets of the form @xmath147 are concentric horospheres .
any discrete group @xmath524 of parabolic isometries of @xmath145 can be realized as a group preserving these horospheres .
the quotient by the action is then of the form @xmath525 for some flat manifold @xmath179 .
if @xmath524 is a maximal parabolic subgroup of a torsion - free lattice @xmath98 , some subset of the form @xmath526\times_{e^r}f$ ] embeds in @xmath527 by rescaling @xmath179 to another flat manifold @xmath528 , this subset looks like @xmath46\times_{e^r}f'$ ] . )
if @xmath45 is embedded in @xmath43 , then there is some @xmath73 so that @xmath529 , the closed @xmath74neighborhood of @xmath45 , is still embedded .
we have @xmath530\times_{e^r}n.\ ] ] it is convenient to reparameterize @xmath529 as @xmath531\times_{e^{r-1}}n.\ ] ] corresponding to any geodesic submanifold @xmath19 there is a cover @xmath532 of @xmath17 , isometric to @xmath533 , where @xmath20 is the codimension of @xmath19 in @xmath17 .
likewise , there is a cover @xmath534 of @xmath529 corresponding to @xmath19 , so that @xmath535\times_{e^{r-1}}\tilde{n}\\ & = & ( -\infty,1+\lambda]\times_{e^{r-1}}\e^k\times_{e^{r-1 } } t .\end{aligned}\ ] ] the group @xmath536 acts , preserving the product structure , by isometries on @xmath534 , with quotient @xmath529 .
if the warping functions are replaced by any other warping functions , @xmath537 will still act by isometries .
we will use the functions @xmath123 and @xmath158 defined in proposition [ p : warpfg ] to define a new metric , first on @xmath538\times \e^k\times t\ ] ] and then on the space @xmath539 , which is homeomorphic to @xmath29 .
[ t : metricplug ] let @xmath43 be a hyperbolic @xmath44manifold , with an embedded toral cusp @xmath540\times_{e^r}n,\ ] ] and suppose @xmath541 is a totally geodesic @xmath27dimensional torus with injectivity radius greater than @xmath30 .
choose @xmath73 so that the @xmath74neighborhood of @xmath45 is still embedded in @xmath43 , with boundary @xmath542 .
there is then a nonpositively curved metric on the partial cone @xmath29 so that the following hold : 1 . [ mplug1 ] @xmath29 is riemannian away from its core .
[ mplug2 ] there is a riemannian isometry between the @xmath77neighborhood of @xmath543 , and the @xmath77neighborhood of @xmath542 in @xmath544 , which takes @xmath545 to a torus @xmath546 , isotopic in @xmath544 to @xmath19 .
[ mplug3 ] if @xmath311 , then @xmath29 is locally cat@xmath70 for some @xmath71 . corresponding to @xmath541 there is a cover @xmath547 so that @xmath532 is isometric to @xmath548 .
likewise , there is a corresponding cover of @xmath529 , which we parametrize as @xmath549\times_{e^{r-1}}\e^k\times_{e^{r-1}}t.\ ] ] let @xmath123 and @xmath158 be the functions from proposition [ p : warpfg ] , and let @xmath538\times_{g(r)}\e^k\times_{f(r)}t\ ] ] by theorem [ t : nonpos ] , @xmath550 is cat@xmath2 , and is cat@xmath70 for some @xmath71 if @xmath311 .
the group @xmath536 acts freely and properly discontinuously by isometries on @xmath550 , with quotient @xmath384 homeomorphic to @xmath29 .
thus the quotient is locally nonpositively curved , locally cat@xmath70 if @xmath311 , and riemannian away from the core @xmath551 note that if we write @xmath552 for @xmath553 , then there are @xmath77neighborhoods of @xmath554 and @xmath555 which are canonically isometric .
this isometry descends to an isometry between @xmath77neighborhoods of @xmath556 and @xmath557 , which takes @xmath558 to @xmath559 as described in the statement of the theorem . since we obtain a compact space with the same metric near the boundary as a cusp by this operation , this is sometimes called cusp closing ( cf.@xcite ) .
the difference in our setting is that we are not generally closing the cusp `` as a manifold '' or even as an orbifold .
rather , we show how to close a cusp `` as a pseudomanifold '' , with a certain curvature control .
a reverse operation , cusp opening or `` drilling '' , has also been studied .
( see for example @xcite for applications to hyperbolic @xmath5manifolds . ) for example , in @xcite a cusp is produced under a certain curvature control after removing a totally geodesic codimension two submanifold in a closed hyperbolic manifold .
this construction also used doubly warped products and was generalized in @xcite ( cf .
@xcite ) .
let @xmath66 .
suppose @xmath43 is a finite volume hyperbolic @xmath44manifold , with horospherical cusps @xmath48 embedded in @xmath43 .
suppose each of those cusps is isometric to @xmath46\times n_i$ ] for a flat torus @xmath52 .
choose @xmath73 so that the @xmath74neighborhood of @xmath560 is still embedded in @xmath43 .
for each @xmath67 , let @xmath53 be a totally geodesic torus in @xmath52 with injectivity radius larger than @xmath30 , so that the space @xmath59 from definition [ d : filling ] is a @xmath4filling of @xmath43 .
we will describe a metric on @xmath59 satisfying the conclusions of theorem [ t : filling ] and [ p : hyper ] .
fix @xmath64 .
the @xmath74neighborhood of the cusp @xmath561 is embedded , and disjoint from the @xmath74neighborhood of all the other cusps .
give @xmath55 the nonpositively curved metric from theorem [ t : metricplug ] .
for @xmath562 $ ] , let @xmath563 be the closure of the complement of the union of the @xmath326neighborhoods of the cusps @xmath48 .
thus @xmath564 is the compact manifold @xmath565 from definition [ d : filling ] . by our choice of @xmath74 ,
each @xmath566 is an embedded submanifold with boundary , and each @xmath566 is isotopic to @xmath564 .
let @xmath567 be an open @xmath77neighborhood of the boundary of @xmath55 . according to theorem [ t : metricplug ] . , there is an isometry @xmath568 for each @xmath67 , so that the space @xmath569 is homeomorphic to the filling @xmath59 by a homeomorphism taking @xmath570 isometrically to @xmath571 .
it is clear that proposition [ p : hyper ] holds for this metric on @xmath59 , and @xmath572 .
conclusion of theorem [ t : filling ] holds by construction ; indeed , every non - singular point of @xmath122 has a neighborhood which is either isometric to a neighborhood in @xmath43 or to a neighborhood in the non - singular part of some @xmath573 , endowed with the metric from theorem [ t : metricplug ] .
this metric is riemannian by theorem [ t : metricplug ] .. the fact that the metric on @xmath122 is the completion of the riemannian metric on the non - singular part can be deduced from lemma [ l : completion ] .
every point in @xmath122 has a neighborhood isometric either to a neighborhood of a point in @xmath43 , which has curvature bounded above by @xmath574 by assumption , or in @xmath573 , which is nonpositively curved by theorem [ t : metricplug ] .
conclusion of theorem [ t : filling ] is established .
if for every @xmath67 , the torus @xmath53 has codimension at most @xmath68 in @xmath52 , then the spaces @xmath55 have curvature bounded above by @xmath520 for some @xmath71 , by theorem [ t : metricplug ] .. conclusion of theorem [ t : filling ] follows .
let @xmath59 be a @xmath4filling of the hyperbolic @xmath44manifold @xmath43 .
in this section we concentrate on the case in which , for at least one @xmath64 , the torus @xmath53 has dimension at most @xmath575 . in this case
the metric constructed in theorem [ t : filling ] is nonpositively curved but not locally cat@xmath70 for any positive @xmath106 .
this is because the filling core @xmath576 of @xmath55 is an isometrically embedded @xmath20torus , where @xmath577 .
the metric described in theorem [ t : filling ] lifts to a metric on the universal cover @xmath9 of @xmath59 . a convex subset of a geodesic metric space which is isometric to @xmath578
is called a _ flat _ , if @xmath579 .
the preimage of @xmath576 in @xmath9 is a union of @xmath20dimensional flats .
conversely , every flat in @xmath9 is in the preimage of some filling core , since the metric on @xmath59 is riemannian and negatively curved away from the cores , by theorem [ t : filling ] .. in @xcite , hruska and kleiner give a definition of isolated flats ( their ( if1 ) ) , which is easier to check than the equivalent definition from @xcite .
the next two definitions are taken from @xcite .
[ d : hcobs ] let @xmath9 be a proper metric space , and let @xmath580 be the set of closed subsets of @xmath9 , and let @xmath581 be the hausdorff distance on @xmath580 .
let @xmath582 , let @xmath227 , and let @xmath231 and @xmath325 be positive real numbers .
define @xmath583 the _ topology of hausdorff convergence on bounded sets _ is the smallest topology on @xmath580 so that all such sets are open .
let @xmath9 be a cat@xmath2 space , and let @xmath98 act properly discontinuously , cocompactly , and isometrically on @xmath9 .
the space @xmath9 has _ isolated flats _ if it contains a family of flats @xmath584 satisfying : * _ equivariance : _ the set of flats @xmath584 is invariant under the action of @xmath98 . * _ maximality : _ there is a constant @xmath28 such that every flat in @xmath9 is contained in a @xmath28neighborhood of some flat @xmath585 . * _ isolation : _ the set @xmath586 is a discrete subset of @xmath587 , the set of all flats in @xmath9 with the topology of hausdorff convergence on bounded sets . we prove that our @xmath9 has isolated flats .
[ p : iflats ] let @xmath59 be a @xmath4filling of the hyperbolic manifold @xmath43 , endowed with the nonpositively curved metric from theorem [ t : filling ] . if @xmath9 is the universal cover of @xmath59 , then @xmath9 has isolated flats .
we let @xmath586 be the collection of all flats in @xmath9 .
maximality and equivariance are clear .
we show the isolation property .
let @xmath588 and @xmath589 be elements of @xmath586 .
any path from @xmath588 to @xmath589 must project to a path from some filling core @xmath590 to some ( possibly the same ) filling core @xmath591 , which passes through @xmath592 , where @xmath76 is as in the statement of proposition [ p : hyper ] .
such a path must have length at least @xmath593 .
let @xmath594 , and let @xmath265 .
choose any @xmath471 , and let @xmath595 be defined as in definition [ d : hcobs ] .
clearly @xmath596 , so @xmath198 is an open neighborhood of @xmath179 . on the other hand ,
since any path from @xmath179 to another flat has length at least @xmath593 , no other flat is in @xmath198 , so @xmath179 is an isolated point of @xmath586 .
the flat @xmath179 was arbitrary , so @xmath586 is discrete .
we suppose that @xmath597 is obtained by filling a hyperbolic @xmath44manifold with @xmath598 toral cusps , and that the tori @xmath53 all satisfy the hypotheses of theorem [ t : filling ] .
it follows ( corollary [ c : cwif ] ) that the universal cover @xmath9 of @xmath59 is a cat@xmath2 space . in particular , since @xmath9 is cat@xmath2 , the visual boundary @xmath599 is a boundary for @xmath600 in the sense of bestvina @xcite .
the _ shape _ of such a boundary is an invariant of @xmath8 ( * ? ? ?
* proposition 1.6 ) .
( since @xmath9 is either cat@xmath3 or cat@xmath2 with isolated flats , the _ homeomorphism type _ of the boundary is an invariant of @xmath8 , by a result of hruska @xcite . ) in particular , the reduced integral ech cohomology groups of @xmath599 are invariants of @xmath8 , since they are shape invariants . by another result of bestvina ,
these cohomology groups are the same as the ( reduced ) cohomology groups of @xmath8 with @xmath601 coefficients ( ( * ? ? ?
* proposition 1.5 ) , cf .
@xcite for the case when @xmath8 is word hyperbolic ) .
the particular results we need from @xcite can be summarized : [ t : bestvina ] @xcite if @xmath8 is the fundamental group of a compact non - positively curved space @xmath122 with universal cover @xmath9 , and @xmath268 is the visual boundary of @xmath9 , then @xmath602 moreover , if @xmath8 is a poincar duality group of dimension @xmath18 , then @xmath268 is a ech cohomology sphere of dimension @xmath61 . in our setting
, we do not find the homeomorphism type of @xmath268 , but we give enough information to determine the shape of @xmath268 . if @xmath9 is any cat@xmath2 space , and @xmath227 ,
then the visual boundary of @xmath9 is homeomorphic to an inverse limit of metric spheres around @xmath110 ( * ? ? ?
* ii.8.5 ) . in order to describe the visual boundary
, it therefore suffices to describe 1 .
the spheres of finite radius about a fixed point , and 2 . the projection maps toward the fixed point between these spheres .
for the remainder of this subsection , we suppose that @xmath43 is a hyperbolic @xmath44manifold , and that @xmath59 is a filling satisfying the hypotheses of theorem [ t : filling ] , so that @xmath603 are the filling cores .
let @xmath9 be the universal cover of @xmath59 .
if @xmath174 is a component of the preimage in @xmath9 of a filling core in @xmath59 , we say that @xmath174 is a _ singular point _ , _ singular geodesic _ , or _
singular flat _ , depending on whether @xmath174 is @xmath280 , @xmath68 , or @xmath20dimensional for @xmath604 .
the union of all singular points , geodesics , and flats in @xmath9 is called the _ singular set _ , or @xmath517 .
[ d : astandard ] if @xmath174 is a component of the singular set , then by the construction of section [ s : basic ] , @xmath174 has a regular neighborhood in @xmath9 isometric to @xmath605\times_{\cosh(t)}a\times_{\sinh(t)}t^k\ ] ] for some flat torus @xmath606 with injectivity radius strictly larger than @xmath30 .
we call this a _ standard neighborhood of _ @xmath174 of radius @xmath6 . in this section
the emphasis is on filling cores which are positive dimensional , but all statements we give remain true , if the following conventions are kept in mind : 1 .
the @xmath280disk is a point .
2 . the @xmath574sphere and the @xmath574disk are both the empty set @xmath607 .
the spherical join of the empty set with any metric space @xmath608 is taken to be @xmath609 , where @xmath610 .
if all cores are assumed to be zero - dimensional , the results here merely recapitulate those of @xcite .
the space @xmath9 is a manifold away from the singular set @xmath517 . even near the singular set
, @xmath9 is quite well behaved .
[ l : nicelocal ] there exists a constant @xmath6 so that @xmath611 for all @xmath612 . for @xmath613
we have @xmath614 . if @xmath615 for some singular component @xmath174 of @xmath517 , and @xmath174 has a standard @xmath6neighborhood as in definition [ d : astandard ] , then @xmath616 , by corollary [ c : loginj ] . since there are finitely many orbits of components of @xmath174 in @xmath9 , some @xmath6 works for all such @xmath191 .
let @xmath617 , let @xmath317 , and let @xmath17 be an open @xmath325neighborhood of @xmath517 in @xmath9 .
the space @xmath618 is a complete nonpositively curved manifold with boundary . if @xmath619 , then the @xmath231ball around @xmath191 lies in the interior of this manifold , and so the logarithm @xmath232 restricted to this ball is injective .
we will see in lemma [ l : reverseshadow ] that if @xmath620 , then @xmath621 , so actually @xmath622 .
[ l : gepx ] @xmath9 has the geodesic extension property .
away from the singular set , @xmath9 is a riemannian manifold of nonpositive curvature , so local geodesics can be extended in @xmath623 .
a local geodesic terminating on @xmath517 can be extended by lemma [ l : gep ] . for the remainder of the section
, we fix a point @xmath624 , and a constant @xmath7 as in lemma [ l : nicelocal ] . a _ sphere centered at @xmath625 _
is a metric sphere @xmath626 for some @xmath471 .
we write @xmath627 for the closed ball of radius @xmath231 centered at @xmath625 , and write @xmath628 for the interior of @xmath627 . for small enough @xmath231
, @xmath629 is homeomorphic to an ordinary sphere @xmath630 and @xmath627 to an ordinary @xmath44dimensional ball .
describing the topology for larger @xmath231 is the main goal in this subsection .
recall that an _ enr _ , or _
euclidean neighborhood retract _ , is any space homeomorphic to a retract of an open subset of some euclidean space .
equivalently , an enr is any locally compact , metrizable , separable , finite dimensional anr .
[ l : enr ] suppose that @xmath631 for any component @xmath174 of the singular set @xmath517 .
then the metric sphere @xmath629 is an enr , and @xmath632 is an @xmath18manifold which is open and dense in @xmath629 .
we first note that @xmath629 is finite dimensional , as it is embedded in the @xmath44dimensional space @xmath9 .
it is obviously locally compact , metrizable , and separable , so we need only establish that @xmath629 is an anr to prove the first part of the lemma . a theorem of dugundji @xcite states that a finite dimensional metric space is an anr if and only if it is locally contractible .
we therefore examine the local behavior of @xmath629 .
let @xmath633 .
there is some @xmath634 so that @xmath191 is contained in the interior of a log - injective ball centered at @xmath635 .
there are two cases , depending on whether @xmath191 is in the singular set @xmath517 or not .
suppose first that @xmath620 , and let @xmath636 .
let @xmath637 be a point on the geodesic from @xmath191 to @xmath625 , so that @xmath638 .
the point @xmath635 is no closer than @xmath639 from @xmath517 , so the log - injectivity radius at @xmath635 is at least @xmath640 , by lemma [ l : nicelocal ] .
it follows that @xmath191 is in a log - injective ball about @xmath635 . in the second case , @xmath191 is contained in some component @xmath174 of the singular set . by the hypothesis of the lemma @xmath641
; let @xmath642 satisfy @xmath643 .
choose @xmath637 on the geodesic from @xmath191 to @xmath112 so that @xmath644 , where @xmath6 is as in lemma [ l : nicelocal ] . since @xmath174 is convex , @xmath645 .
applying lemma [
l : nicelocal ] , the log - injectivity radius at @xmath635 is at least @xmath6 , and so @xmath191 is contained in a log - injective ball centered at @xmath635 .
let @xmath198 be the log - injective ball around @xmath635 given by the claim , and let @xmath28 be a slightly smaller closed ball around @xmath635 so that @xmath191 is still in the interior of @xmath28 .
the intersection @xmath646 is a strictly convex set containing @xmath635 in its interior , so that @xmath241 is contained in a log - injective neighborhood of @xmath198 .
moreover , @xmath9 has the geodesic extension property by lemma [ l : gepx ] .
we can therefore apply lemma [ l : dconvex ] to obtain a homeomorphism @xmath647 from the frontier of @xmath241 to the space of directions @xmath648 at @xmath635 .
this space of directions is always either a round @xmath18sphere or ( using lemma [ l : dirjoin ] ) an @xmath18pseudomanifold isometric to the join of a sphere and a torus .
in particular , it is locally contractible
. let @xmath649 be the radius of the ball @xmath28 , and choose a positive @xmath650 .
the open @xmath325ball around @xmath191 intersects the frontier of @xmath241 in a set @xmath651 , which is an open neighborhood of @xmath191 in @xmath629 .
using the homeomorphism @xmath246 to the locally contractible space @xmath648 , we may find an open contractible @xmath652 containing @xmath191 . to see that @xmath653 is a manifold , note that , in the argument just made , if @xmath620 , then @xmath654 , and @xmath246 takes a neighborhood of @xmath191 in @xmath629 to an open set in the sphere @xmath630 .
even for @xmath620 , the local description makes it clear that @xmath191 is in the closure of the set of manifold points of @xmath629 . for any @xmath231 ,
there is a map @xmath655 from the exterior of @xmath629 to @xmath629 defined by geodesic projection toward @xmath625 : if @xmath110 is any point lying further than @xmath231 from @xmath625 , then the geodesic from @xmath110 to @xmath625 intersects @xmath629 in the unique point @xmath656 .
[ d : good ] a _ shell _ is a set of the form @xmath657 a shell @xmath658 is _ good _ if it satisfies all of the following : 1 .
[ reg ] if @xmath6 is the constant fixed after lemma [ l : nicelocal ] , then @xmath659 .
[ goodint ] whenever @xmath174 is an @xmath660dimensional component of @xmath517 and @xmath661 , then @xmath662 is either empty or a ( topological ) sphere of dimension @xmath663 .
[ embed ] the map @xmath664 restricted to @xmath665 is an embedding
. the following observation will be useful later .
[ l : subgood ] suppose that @xmath666 . if @xmath658 is good , then either @xmath667 is good , or @xmath668 for some component @xmath174 of the singular set @xmath517 .
components of the singular set have very nice intersections with good shells : [ l : intersect ] suppose @xmath658 is a good shell .
let @xmath174 be a component of the singular set of dimension @xmath660 which intersects @xmath658 nontrivially .
if @xmath174 intersects @xmath658 but not @xmath669 , then @xmath670 is a disk of dimension @xmath660 .
otherwise @xmath671 is a sphere of dimension @xmath663 and there is a homeomorphism @xmath672 from @xmath670 to @xmath673 .
distance to the basepoint @xmath674 is a proper strictly convex function on @xmath675 .
if @xmath174 does not intersect @xmath669 , then @xmath670 is a sublevel set of this strictly convex function , so @xmath670 is strictly convex .
since @xmath676 , the set @xmath670 has nonempty interior ( as a subset of @xmath174 ) , so it is an @xmath660disk . now suppose that @xmath174 intersects @xmath669 as well as @xmath677 .
let @xmath635 be the closest point on @xmath174 to @xmath625 .
we define the first coordinate of @xmath678 by geodesic projection toward @xmath635 ; the second coordinate is the distance to @xmath625 , affinely rescaled to lie in @xmath679 $ ] .
explicitly , @xmath680 where @xmath112 is the unique point in @xmath669 on a geodesic from @xmath191 to @xmath635 .
the convexity of @xmath174 and of the distance function make this map a homeomorphism .
the next lemma says that there are plenty of good shells .
[ l : good ] for every @xmath471 there is some @xmath317 so that the shell @xmath681 is good .
let @xmath6 be the smallest radius of a standard neighborhood of a component of the singular set of @xmath9 .
the distance between two distinct components is thus greater than @xmath682 .
it follows that if @xmath683 , then @xmath684 can not send points from different singular components to the same point in @xmath685 .
let @xmath686 be the collection of components of the singular set @xmath517 , and note that the set @xmath687 is a discrete subset of @xmath688 .
we may therefore choose @xmath689 so that neither @xmath690 nor @xmath691 is in @xmath692 .
condition is obviously satisfied for any such @xmath325 .
we will show that and are as well .
suppose @xmath693 intersects @xmath694 nontrivially for @xmath695 , and let @xmath696 be the intersection of @xmath174 with the ball of radius @xmath697 around @xmath625 .
since @xmath698 , the set @xmath28 has non - empty interior ( as a subset of @xmath174 ) , and since both @xmath174 and the metric are convex , @xmath28 is a strictly convex subset of @xmath174 .
lemma [ l : dconvex ] implies that @xmath699 is an @xmath700sphere , so condition of definition [ d : good ] is satisfied . finally , suppose that that @xmath681 does not satisfy condition of definition [ d : good ] .
there is then some @xmath701 and two points @xmath110 , @xmath702 so that @xmath703 . since @xmath703 , the geodesics @xmath704
$ ] and @xmath705 $ ] must coincide on some ( maximal ) initial segment @xmath706 $ ] of length at least @xmath690 . there are two cases : * case 1 : * suppose that @xmath707 . without loss of generality ,
suppose that @xmath708 , and rechoose @xmath110 so that @xmath110 is the closest point on @xmath704\cap a=\{x\}$ ] . since @xmath174 is convex , the geodesic @xmath709 $ ] lies entirely in @xmath174 , so @xmath709 $ ] represents a point in @xmath710 .
since @xmath705 $ ] is geodesic , the geodesic sub - segments @xmath709 $ ] and @xmath711 $ ] must have alexandroff angle @xmath30 at @xmath110 .
this implies that @xmath711 $ ] also lies in @xmath710 , but since @xmath174 has a standard neighborhood , this implies that @xmath711 $ ] , and hence @xmath625 , lies in @xmath174 . but
this contradicts the fact that @xmath625 lies outside the singular set . * case 2 : * suppose @xmath712 . the directions in @xmath713 corresponding to @xmath714 $ ] and @xmath715 $ ] are both at distance @xmath30 from the direction corresponding to @xmath716 $ ] .
since the logarithm map is injective near @xmath112 , these directions are all distinct in @xmath713 . in particular
, @xmath713 is not a round sphere of diameter @xmath30 and so @xmath112 must lie in some component of the singular set @xmath717 .
since @xmath718 , we must have @xmath719 . replacing either @xmath110 or @xmath111 by @xmath112
, we may derive a contradiction as in case 1 .
this establishes condition .
to understand the difference between the two boundary components of a good shell , we have to understand how geodesics can be continued through a piece of the singular set . in a simply connected riemannian manifold of non - positive curvature , geodesics ( in the sense of length - minimizing paths ) can always be continued uniquely . at a point in a singular space
whose space of directions is not a round sphere of diameter @xmath30 ( or a point which does not have a log - injective neighborhood ) this uniqueness generally fails .
the next lemma gives a description of the collection of geodesics of a fixed length starting at a point and passing through some particular component of the singular set .
[ l : reverseshadow ] let @xmath174 be a component of the singular set isometric to @xmath720 , and let @xmath721 .
choose @xmath144 so that @xmath722 if @xmath268 is the set of endpoints of geodesics of length @xmath144 starting at @xmath625 and meeting @xmath174 , then @xmath268 is homeomorphic to @xmath723 , where @xmath724 is a flat @xmath20dimensional torus minus the interior of a closed embedded @xmath20dimensional ball . by the construction of section [ s : basic ]
, there is a standard neighborhood of @xmath174 which we may identify with the warped product : @xmath605\times_{\cosh(r)}a\times_{\sinh(r)}t\ ] ] for some flat @xmath20torus @xmath19 whose injectivity radius is greater than @xmath30 . let @xmath725 be the closest point to @xmath625 , and let @xmath726 be arbitrary . via proposition [ p : dirwarp ] , we may identify the space of directions @xmath727 with @xmath728 .
let @xmath729 be the canonical projection onto the @xmath19 factor .
finally , let @xmath730 send a point @xmath731 to the unique point in @xmath174 on the geodesic from @xmath112 to @xmath625 .
we can then define a map @xmath732\times \sigma_a(a)\times t / \sim$ ] by @xmath733 , \pi_t([\psi(z),z])\right),\ ] ] abusing notation in the second and third factor by writing a geodesic segment with a given direction instead of the direction itself . because geodesics are continuous as their endpoints are varied
, @xmath387 is a continuous map onto some closed subset of @xmath728 .
note here that the first two factors are essentially the polar coordinates centered at @xmath403 of @xmath734 , reparametrized in the radial direction . to check injectivity of @xmath387
, we therefore only need to check injectivity of @xmath387 restricted to @xmath735 for @xmath736 . if @xmath737 , there is nothing to check , so it suffices to verify the following . for @xmath738 with @xmath739
, @xmath387 restricted to @xmath735 is an embedding , and @xmath735 is homeomorphic to @xmath19 minus an open ball of radius @xmath30 . for any @xmath740 , the geodesic from @xmath625 to @xmath112
is composed of two subsegments @xmath741 $ ] and @xmath742 $ ] . regarded as directions in @xmath727 , these subsegments must subtend an angle of @xmath30 .
conversely , any direction which is alexandroff angle @xmath30 from @xmath741 $ ] is the direction of the geodesic continuation of @xmath743 $ ] on to some unique @xmath731 .
( uniqueness follows from log - injectivity at @xmath744 of the @xmath6neighborhood of @xmath174 , in which @xmath268 is contained . ) to sum up , @xmath745 is an embedding restricted to @xmath735 , with image @xmath314 , the set of directions whose angle with @xmath741 $ ] is exactly @xmath30 . identifying @xmath727 with @xmath728 ,
suppose that @xmath741 = ( \phi,\alpha,\theta)$ ] in the standard join coordinates .
because @xmath746 , @xmath747 .
standard trigonometric identities can be used to show that those @xmath748 which make an angle of @xmath30 with @xmath741 $ ] are exactly those with @xmath749 , @xmath750 , and @xmath751 .
it follows that @xmath752 is @xmath19 minus an open ball of radius @xmath30 around @xmath248 .
it follows that the image of @xmath387 is homeomorphic to the join of @xmath753 with @xmath19 minus a ball of radius @xmath30 .
since @xmath387 was a homeomorphism onto its image , and @xmath753 is an @xmath700sphere , the lemma is proved .
the space @xmath754 described in lemma [ l : reverseshadow ] is a pseudomanifold with boundary homeomorphic to @xmath755 .
[ d : connectsum ] let @xmath756 and @xmath757 be spaces containing @xmath758 and @xmath759 which are open , connected , dense , and homeomorphic to oriented @xmath18manifolds .
then the _ connect sum _
@xmath760 is the space obtained by choosing small closed @xmath18balls @xmath761 and @xmath762 , choosing an orientation reversing homeomorphism @xmath763 , and gluing the exteriors together thus : @xmath764 for @xmath765 there is a ( unique up to homotopy ) map @xmath766 which takes @xmath767 to itself by the identity map , and sends @xmath768 onto @xmath567 by a degree one map .
we refer to @xmath769 as _ the map which pinches @xmath770 to a point .
_ we remark that the operation of connect sum described in definition [ d : connectsum ] is associative , so that the connect sum of three or more spaces is well - defined . moreover , if @xmath771 , then the map which pinches @xmath772 to a point is homotopic to the composition of the map which pinches @xmath28 to a point with the map which pinches @xmath241 to a point .
[ p : pinch ] suppose @xmath658 is a good shell , and let @xmath773 be the components of the singular set which intersect @xmath658 but not @xmath669 .
if the dimension of @xmath774 is @xmath775 for each @xmath67 , then there is a homotopy equivalence @xmath387 from @xmath677 to the connect sum @xmath776 , where @xmath777 moreover , the projection map @xmath778 fits into a homotopy commutative triangle @xmath779_{p_a^b}\ar[rr]^\phi & & s_a\mathrel{\ # } j\ar[dl]^{q_{s_a}}\\ & s_a & } } \ ] ] where @xmath780 is the map which pinches @xmath781 to a point .
one complicating point here is that near the components of the singular set , the projection map @xmath782 is not a local homeomorphism .
this is true even if @xmath669 and @xmath677 intersect the same components of the singular set . to deal with this issue
, we replace @xmath669 and @xmath677 with homotopy equivalent quotients @xmath783 and @xmath784 so that @xmath785 , chosen so that @xmath782 induces a map @xmath786 , and so that @xmath787 is the map which pinches @xmath781 to a point .
the maps @xmath672 from lemma [ l : intersect ] can be patched together to give an embedding @xmath788 whose image is the union of the components of @xmath665 which intersect both boundary components of the shell .
since @xmath658 is a good shell , the map @xmath664 restricts to an embedding of @xmath789 , and so @xmath790 is also an embedding .
we next give decompositions @xmath791 and @xmath792 of @xmath669 and @xmath677 into closed sets .
a set in @xmath791 is either a point outside the image of @xmath790 , or it is an arc of the form @xmath793 for some @xmath794 .
a set in @xmath792 is either a point @xmath112 so that @xmath795 lies outside the image of @xmath790 , or it is the preimage of one of the arcs of the form @xmath796 . arguing as in lemma [ l : reverseshadow ] , each element of @xmath792 is either a point or homeomorphic to the cone on a closed subset of a torus .
we leave it to the reader to check that the decompositions @xmath791 and @xmath792 are upper semicontinuous .
it follows that @xmath783 and @xmath784 are hausdorff , and therefore compact metric spaces .
the projection map from @xmath669 sends elements of @xmath792 to elements of @xmath791 , and thus induces a continuous map @xmath787 from @xmath784 to @xmath783 .
we have the following commuting square : @xmath797^{q_b}\ar[d]^{p_a^b } & q_b\ar[d]^{p_q } \\
s_a \ar[r]^{q_a } & q_a}}\ ] ] note that the map @xmath787 is a homeomorphism away from the image in @xmath783 of @xmath798 the proposition now reduces to two claims : [ vert ] the space @xmath784 is homeomorphic to the connect sum @xmath799 , and @xmath787 is homotopic to the map which pinches @xmath781 to a point .
[ horiz ] the horizontal maps in are homotopy equivalences .
as noted above , @xmath787 is a homeomorphism away from @xmath800 where the @xmath774 are those components of the singular set which intersect @xmath658 but not @xmath669 .
a small regular neighborhood in @xmath669 of each @xmath801dimensional disk @xmath802 is a ball @xmath803 ( again using lemma [ l : intersect ] ) , so this implies that @xmath784 is homeomorphic to the connect sum of @xmath783 with some collection of spaces @xmath804 where each @xmath805 is equal to @xmath806 , capped off with a ball .
we must describe the topology of the spaces @xmath806 .
the topology of @xmath807 is given in lemma [ l : reverseshadow ] : each such space is homeomorphic to a join @xmath808 with @xmath809 and @xmath524 homeomorphic to a @xmath54torus with a small open ball removed .
it follows that @xmath805 is homeomorphic to the join of a @xmath54torus with an @xmath810sphere . putting the pieces together ,
we see that @xmath784 is homeomorphic to a connect sum of @xmath783 with the space @xmath781 described in the statement , and projection from @xmath784 to @xmath783 is homotopic to the map which pinches @xmath781 to a point .
we first claim that the quotient spaces @xmath783 and @xmath784 are enrs .
indeed , lemma [ l : enr ] implies that the metric spheres @xmath669 and @xmath677 are enrs , so each of @xmath783 and @xmath784 is a quotient of an enr by a decomposition into closed contractible ( in particular cell - like ) sets . a theorem of lacher ( * ? ? ?
* corollary 3.3 ) says that if such a quotient is metrizable and finite dimensional , then it is an enr .
we have already shown that @xmath783 and @xmath784 are metrizable .
moreover , there is an obvious decomposition of each into an open @xmath18manifold and a union of closed annuli and disks of strictly lower dimension .
( only annuli are needed for @xmath783 . ) applying the addition theorem from dimension theory ( see , e.g. @xcite ) , we obtain the finite dimensionality of @xmath783 and @xmath784 .
it now follows from another theorem of lacher ( * ? ? ?
* theorem 1.2 ) that the quotient maps are homotopy equivalences .
the proposition now follows from the two claims .
[ t : cohomology ] let @xmath66 .
let @xmath43 be a hyperbolic @xmath44manifold with @xmath598 toral cusps @xmath48 , let @xmath565 be a manifold with boundary obtained from @xmath43 by removing open horospherical neighborhoods of the cusps , and let @xmath60 be totally geodesic tori in @xmath811 satisfying the hypotheses of theorem [ t : filling ] .
let @xmath649 be the maximum of the dimensions of @xmath60 .
let @xmath812 be obtained from @xmath43 by filling along @xmath60 , and let @xmath813 .
the cohomology of @xmath8 with @xmath601 coefficients satisfies : @xmath814 let @xmath122 be the universal cover of @xmath9 . by theorem [ t :
filling ] , the space @xmath122 is cat@xmath2 . as pointed out in @xcite , the visual boundary @xmath815 is a boundary for @xmath8 , and so we have @xmath816 for all @xmath635 .
we thus compute the ech cohomology of @xmath815 . as at the beginning of section [ s : inverselimit ] , choose some @xmath625 in @xmath122 , not in the singular set , and write @xmath629 for the metric sphere of radius @xmath231 centered as @xmath110 . as already noted , the visual boundary of @xmath122 is homeomorphic to the inverse limit of the system of spheres centered at @xmath110 , i.e. , @xmath817 where @xmath818 is projection along geodesics toward @xmath625 .
in fact , any sub - collection of spheres @xmath819 with @xmath820 determines the inverse limit . by lemmas [
l : subgood ] and [ l : good ] , we may therefore choose the sequence @xmath821 so that each shell @xmath822 is good .
let @xmath823 .
we have @xmath824 and so @xmath825 for each @xmath67 , we apply proposition [ p : pinch ] to compute @xmath826 .
let @xmath827 be the space @xmath781 from the proposition , applied with @xmath828 and @xmath829 , and let @xmath830 be the largest dimension of any torus @xmath831 occurring in @xmath832 .
elementary algebraic topology shows that @xmath826 is injective , and that @xmath833 is free abelian of nonzero rank in dimension @xmath834 for @xmath835 and zero in all other dimensions .
note that @xmath836 for infinitely many @xmath86 .
it follows that the ( reduced ) ech cohomology of @xmath837 is @xmath838 applying theorem [ t : bestvina ] gives the desired result .
[ c : notpdn ] let @xmath43 , @xmath60 satisfy the hypotheses of theorem [ t : cohomology ] .
the group @xmath839 is a poincar duality group if and only if every @xmath53 has dimension one . if every @xmath53 has dimension one , then @xmath59 is a non - positively curved ( and hence aspherical ) closed manifold of dimension @xmath840 , and so @xmath62 must be a poincar duality group of dimension @xmath840 .
( the special case in which every @xmath53 has dimension one was already handled by schroeder in @xcite . )
if some @xmath53 has dimension different from one , then theorem [ t : cohomology ] implies that @xmath8 has cohomological dimension @xmath840 , but @xmath841 for some @xmath842 , and so @xmath8 can not be an @xmath44dimensional poincar duality group , by the second part of theorem [ t : bestvina ] .
the proof that our examples are not manifold groups is essentially the same as that used in @xcite for the case in which every cusp is coned off completely .
the difference in our setting is that we get a wider variety of behavior at infinity .
[ p : simpconn ] let @xmath59 be a @xmath4filling of an @xmath44dimensional hyperbolic manifold @xmath43 . unless some @xmath53 is @xmath18dimensional , the universal cover @xmath9 of @xmath59 is simply connected at infinity .
we retain the notation of subsection [ s : inverselimit ] . to show simple connectivity at infinity
, it suffices to show that for every @xmath471 there is some @xmath843 so that any loop in @xmath844 is null - homotopic in @xmath845 .
accordingly , let @xmath471 , and choose @xmath846 as in the proof of theorem [ t : cohomology ] , so that each shell @xmath822 is good and so that @xmath847
. choose some @xmath67 so that @xmath848 .
we claim that @xmath849 suffices .
let @xmath850 be a loop .
it is easy to show that @xmath9 deformation retracts ( holding @xmath851 fixed ) to @xmath851 , so @xmath90 is homotopic in the complement of @xmath627 to a loop @xmath322 in @xmath852 .
induction together with proposition [ p : pinch ] shows that @xmath852 is a connect sum of joins of tori and spheres . moreover , unless some @xmath53 is @xmath18dimensional , each such join is simply connected , so @xmath852 is simply connected .
it follows that @xmath322 is null - homotopic in @xmath852 , so @xmath90 is null - homotopic in the complement of @xmath627 . in @xcite
, it is shown that systolic groups are never simply connected at infinity .
[ c : notsys ] let @xmath80 , where @xmath59 is a @xmath4filling of an @xmath44dimensional hyperbolic manifold , and suppose that no @xmath53 is @xmath18dimensional .
then @xmath8 is not systolic .
suppose we are given a group @xmath8 and a collection of subgroups @xmath853 with respect to which @xmath8 is relatively hyperbolic .
suppose further that normal subgroups @xmath854 ( filling kernels ) are given , and @xmath855 is the normal closure in @xmath8 of these subgroups .
the group @xmath856 is said to be a _ hyperbolic filling of @xmath8 _ if if @xmath8 acts properly and cocompactly on a cat@xmath2 space with isolated flats , then it is hyperbolic relative to the flat stabilizers @xcite .
these flat stabilizers are virtually abelian , and so there are many candidates for the filling kernels as above .
if @xmath8 acts properly and cocompactly on a cat@xmath2 space with isolated flats , does every hyperbolic filling of @xmath8 act properly and cocompactly on some cat@xmath2 space with isolated flats ?
what can be said about the dimension of this space ? if @xmath8 acts properly and cocompactly on a cat@xmath2 space with isolated flats ( or geometrically finitely on a cat@xmath3 space ) , and @xmath862 is a hyperbolic filling of @xmath8 which is hyperbolic , does @xmath8 act properly and cocompactly on some cat@xmath3 space ?
what can be said about the dimension of this space ? in the current paper we have given answers to both of these questions in the special case that @xmath8 is the fundamental group of a hyperbolic @xmath44manifold with toral cusps and the filling kernels are direct summands of the ( free abelian ) peripheral subgroups .
it should be possible to appeal to haefliger s theory of nonpositively curved orbispaces and extend our results to the case in which the filling kernels are arbitrary subgroups of the peripheral subgroups .
focusing on fillings of the fundamental group @xmath8 of a fixed hyperbolic manifold @xmath43 with toral cusps , other kinds of question arise .
theorem [ t : examples ] says that there are infinitely many isomorphism types of hyperbolic fillings of @xmath43 . on the other hand ,
the discussion in section [ s : visbound ] shows that these fillings have boundary whose shape is determined by the dimensions of the filling tori .
[ q : diff ] let @xmath66 .
let @xmath43 be a hyperbolic @xmath44manifold with toral cusps , and let @xmath863 .
let @xmath864 be the collection of fundamental groups of @xmath4fillings @xmath59 so that @xmath865 how many quasi - isometry types are there in @xmath864 ? the case of classical hyperbolic dehn filling of @xmath5manifolds is @xmath866 and @xmath867 .
the @xmath4 theorem together with the positive solution to the geometrization conjecture by perelman @xcite implies that , in this case , every group in @xmath864 is quasi - isometric to @xmath13 .
m. gromov .
asymptotic invariants of infinite groups . in _
geometric group theory , vol .
2 ( sussex , 1991 ) _ , volume 182 of _ london math .
lecture note ser .
_ , pages 1295 .
cambridge univ . press ,
cambridge , 1993 . | we give new examples of hyperbolic and relatively hyperbolic groups of cohomological dimension @xmath0 for all @xmath1 ( see theorem [ t : examples ] ) .
these examples result from applying cat@xmath2/cat@xmath3 filling constructions ( based on singular doubly warped products ) to finite volume hyperbolic manifolds with toral cusps .
the groups obtained have a number of interesting properties , which are established by analyzing their boundaries at infinity by a kind of morse - theoretic technique , related to but distinct from ordinary and combinatorial morse theory ( see section [ s : visbound ] ) .
[ multiblock footnote omitted ] |
cryptography is created to satisfy the people s desire of transmitting secret messages . with the development of the quantum computation , especially the proposal of shor s algorithm@xcite , the base of the most important classic cryptographic scheme was shocked .
but at the same time , the principles of quantum mechanics have also shed new light on the field of cryptography as these fundamental laws guarantee the secrecy of quantum cryptosystems .
any intervention of an eavesdropper , eve , must leave some trace which can be detected by the legal users of the communication channel .
all kinds of quantum key distribution ( qkd ) schemes , such as bb84 protocol@xcite , b92 protocol@xcite , and the epr scheme@xcite have been proposed .
recently , quantum cryptography with @xmath2-state systems was also introduced@xcite
. experimental research on qkd is also progressing fast , for instance , the optical - fiber experiment of bb84 and b92 protocols have been realized up to 48 km@xcite , and qkd in free space for b92 scheme has been achieved over 1 km distance@xcite . in paper@xcite , lior and
lev presented a quantum cryptography based on orthogonal states firstly .
then there is quantum cryptographic scheme involving truly two orthogonal states @xcite .
the basic technic is to split the transfer of one - bit information into two steps , ensuring that only a fraction of the bit information is transmitted at a time .
then the non - cloning theorem of orthogonal states@xcite guarantee its security .
based on the impossibility of cloning nonorthogonal mixed states , the no - cloning theorem of orthogonal states says that the two ( or more ) orthogonal states @xmath3 of the system composed of @xmath4 and @xmath5 can not be cloned if the reduced density matrices of the subsystem which is available first ( say @xmath4 ) @xmath6 $ ] are nonorthogonal and nonidentical , and if the reduced density matrices of the second subsystem are nonorthogonal .
it is a very surprising result since it means that entanglement is not vital for preventing cloning of orthogonal states . in the case of a composite system made of two subsystems , if the subsystem are only available one after the other , then there are various cases that orthogonal states can not be cloned
. for the multi - state systems , bennett _ et al . _
have shown that there are orthogonal product pure states in the @xmath7 hilbert space and proved that these states may have some degree of nonlocality without entanglement@xcite .
there was also experimental demonstration of three mutually orthogonal polarization states@xcite , where biphotons are used as multi - state systems .
we propose the general conditions for the orthogonal product states of the multi - state composite systems to be used in qkd , then present a qkd scheme with the orthogonal product states of @xmath7 system which has several distinct features , such as high efficiency and great capacity .
the generalization to the @xmath8-state systems , and eavesdropping is analyzed where a peculiar limitation , @xmath9 for the success probability of an efficient eavesdropping strategy is found as _ _ @xmath8 _
_ becomes large enough .
in the present qkd scheme with orthogonal product states in the @xmath1 hilbert space , the transmission processing is same as the qkd scheme with common orthogonal states@xcite .
the information is encoded in the holistic state of the two particles , and these two particles are sent separately to ensure that any eavesdropper can not hold both particles at the same time . since only orthogonal product state are employed , operations on one subsystem have no effect to the other . there are some basic conditions for any set of orthogonal product states in the @xmath8-state composite systems to be used in the present qkd scheme : for any density matrix of any subsystem , @xmath10 there must be at least one @xmath11 which is both nonidentical and nonorthogonal to @xmath12 ( @xmath13 represents subsystem @xmath4 or @xmath5 . @xmath14 and @xmath15 represent different states of the set . ) then from the point of any subsystem@xcite , the standard non - cloning theorem @xcite is satisfied , this guarantees the security of the protocol .
what s more , we can transmit @xmath16 bits information , double the value of existing qkd protocol with usual orthogonal states@xcite .
it is evident that this is the maximal information that can be transmitted by the @xmath1 system . for @xmath17 system , there is obviously no such orthogonal product states that satisfy the orthogonal states cryptography conditions .
the reason is that if @xmath18 are nonidentical and nonorthogonal to @xmath19 , then @xmath20 can not be orthogonal to @xmath21 next , we consider the @xmath7 system .
a general set of orthogonal product states in this hilbert space is as following @xmath22 where @xmath23 @xmath24 @xmath25 @xmath26 @xmath27 @xmath28 @xmath29 @xmath30 are complex number and @xmath31 this set of states has been proved to have some degree of nonlocality without entanglement , when @xmath32@xcite .
for the general case , no satisfying proof for the existing of the nonlocality has been found yet .
but if they satisfy the conditions mentioned above , they can still be used in this qkd scheme .
the process of this qkd scheme is as follows :
alice prepares two particles @xmath4 and @xmath5 randomly in one of the nine orthogonal product states shown above and sends particle @xmath4 to bob , when bob receives it , he informs alice through open classical channel .
then alice sends out particle @xmath5 .
when particle @xmath4 and @xmath5 are both in the hand of bob , he makes a collective orthogonal measurement under the basis of eqs .
( 1 ) to determine which state the two - particle system has been prepared . after a sequence of this procedure , they can share a random bit string , which is the raw keys .
2 . in order to find possible eavesdropping , alice and
bob randomly compare some bits to verify whether the correlations have been destroyed .
if it is true with as high probability as they require , it can be believed that there is no eavesdropper and all of the rest results can be used as cryptographic keys .
otherwise , all the keys are discarded and they must be redistributed .
what is vital to this scheme is that alice sends the second particle only when the first one reaches bob to eliminate the possibility of any eavesdropper to possess the two particles at the same time .
this protocol has some distinct features .
as all raw keys except a small portion chosen for checking eavesdroppers is usable , it is very efficient ( nearly @xmath33 ) , and it has large capacity since @xmath34 bits information is transmitted by a @xmath7 system .
we first consider one efficient eavesdropping strategy . in this strategy , eve measures the first particle from alice and sends it to bob .
she measures the second particle corresponding to the measurement result of the first one and sends it to bob .
it has been found that with this strategy , eve has the success eavesdropping probability as high as @xmath35 for qkd with orthogonal states in the @xmath17 hilbert space@xcite .
the eavesdropping is as following , eve intercepts particle @xmath4 and makes an orthogonal measurement in the basis @xmath36 .
suppose particle @xmath4 is found in state @xmath37 , eve knows that the two - particle states of @xmath4 and @xmath5 is @xmath38 , @xmath39 with probability @xmath40 respectively , or @xmath41 , @xmath42 with probability @xmath43 and @xmath44 respectively .
then she sends it to bob .
when particle @xmath5 comes , she intercepts it too and measures it in the basis @xmath45 and then sends it to bob . if eve sees that particle @xmath5 is in @xmath46 or @xmath47 , then obviously she knows that the two - particle state is @xmath38or @xmath39 . in this case
she is lucky enough to conceal from alice and bob for the two - particle state is not disturbed . and if eve finds particle @xmath5 in state @xmath48 , then the two - particle state is @xmath41 or @xmath42 , which has collapsed to @xmath49 , then bob has the partial probability of @xmath50 or @xmath51 to find the two - particle state in @xmath41 or @xmath52 respectively .
so it is clear that for the case eve measures the particle @xmath4 in the state @xmath53 , the probability for he to eavesdrop without being detected is @xmath54 analyzed in the same way , the total probability that eve eavesdrops the key information without being detected is @xmath55 then it is evident that @xmath56 gets the minimal value of @xmath57 when @xmath58 .
we depict this set of states in the @xmath7 hilbert space in a visual graphical way as figure 1 shows .
the four dominoes represent the four pairs of states that involve superposition of the basis states @xmath59 , @xmath60 or @xmath61 .
it is obvious that this figure is @xmath62-fold rotation symmetric , and we will show later that this symmetry is one of the basic requirements for there may be two symmetrical eavesdropping strategies . from this figure
we can obviously see that all the states included in one row , where particle @xmath4 is in basis states , can be eavesdropped without being detected in this strategy .
but for all other states , there is only certain probability , the quartic of probability amplitude of the superposition states of particle @xmath4 in the basis states , for eve s success eavesdropping .
then the total probability is @xmath56 .
now let s consider the case of @xmath1 system .
since we only utilize orthogonal product states , and any superposition of @xmath8 basis states that covers all the grids of any row ( any column ) in the graphic depiction will surely be distinguished by eve which will have no use in the present qkd scheme .
in other words , any set of states including such superposition state will violate the above conditions , then the possible superposition states can just be the ones of less than @xmath8 basis states which cover no more than @xmath63 grids of the figures .
the @xmath64 system can be depicted in figure 2 .
we can see that the only difference between the figures of the @xmath0 and @xmath64 systems is that there are four states in the center of the @xmath64 system , which can be eavesdropped without being detected .
then generalized straightforwardly , the @xmath65 system can be analyzed in the similar way as the @xmath66 system .
thus here we take the @xmath66 system for example . @xmath67 .
the figure 3 depicts a set of orthogonal product states of the @xmath68 system , which is generalized straightforwardly from @xmath7 system . for any complete set of orthogonal product states , the success probability for eve to eavesdropping without being detected
is @xmath69 /25,\ ] ] where @xmath70 is the probability amplitude of the superposition states of particle @xmath4 in the basis states and @xmath71 for any @xmath72 @xmath73 and @xmath14 , and @xmath74 means that particle @xmath5 is in the basis state @xmath74 , @xmath73 denotes dominoes in line @xmath74 , @xmath75 denotes the superposition states in domino @xmath73 , and @xmath76 denotes the basis states that superposition state @xmath75 involves , @xmath56 is eve s success probability for the @xmath7 system and @xmath77 is the probability for states in row @xmath78 and @xmath79 . we know from the graph that the dominoes in column @xmath80 and @xmath81 cover @xmath62 grids in total ( if @xmath82 grids are covered , all the states in this column can be eavesdropped without being detected . ) , and we have shown that @xmath83 . it can be proved easily that @xmath84 reaches its minimal value @xmath85 when @xmath56 gets its minimal value and only one domino is in column @xmath80 or @xmath86 and @xmath87 .
this corresponds to the set of states depicted in figure 3 .
then for the @xmath88 system , the success probability is @xmath89 /(2n+1)^2.\ ] ] the minimal probability can be obtained similarly : @xmath90 /(2n+1)^2=1/2+(1 + 4n)/2(2n+1)^2,\ ] ] where @xmath91 .
we can deduce immediately that this value approaches @xmath92 when @xmath8 gets large enough . of course , there are other ways of plotting the graph symmetrically .
but this set of states is of the most secure .
in fact , the value of @xmath93 can be deduced straightforwardly from the symmetry of the plot . for any vertical domino contributes @xmath94 to this value , any state in the horizontal dominoes and the state in the center each contributes @xmath94 .
due to the symmetry , there are at least @xmath95 vertical or horizontal dominoes and @xmath96 /2 $ ] states in the horizontal dominoes . for the @xmath65 system
, the same result can be reached .
that is to say there is a limitation in the probability of the success eavesdropping when the hilbert space becomes large enough .
and it is evident that in this strategy only particle @xmath4 may be demolished , and particle @xmath5 is not infected at all .
the function of the operation to @xmath5 which is depended on the result of the measurement on @xmath4 is just to extract more information .
eve may adopt the complementary eavesdropping strategy , in which eve try to eavesdrop some information by intercepting and operating only on the second particle @xmath5 , which may cause demolition to it .
then for the set of states in @xmath1 systems , whose graphic depictions are @xmath62-fold rotation symmetric , the probability to eavesdrop some information without being detected is equal to that of the first strategy , i.e. , @xmath97 . but for those states without such symmetry , it can be verified that one of the success probabilities for the complementary strategies is larger than @xmath97 .
so we employ the symmetric states in the present scheme .
of course , there are other strategies , for example , she can hold up the first particle @xmath4 and send out a substitute particle @xmath98 to bob .
when @xmath5 comes , she makes a collective measurement under the two - particle orthogonal basis , then sends out a particle @xmath99 in the state of @xmath5 . in this strategy ,
eve can eavesdrop the information entirely , but the probability for she to pass the checking process is only @xmath100 , which tends to zero , for the state of particle @xmath98 is randomly chosen from a @xmath8 hilbert space .
we have proposed the general conditions for the orthogonal product states to be used in qkd , then presented a qkd scheme with the orthogonal product states of @xmath7 system which has several distinct features , such as high efficiency and great capacity .
the generalization to the @xmath8-state systems , and eavesdropping is analyzed where a peculiar limitation , @xmath9 for the success probability of an efficient eavesdropping strategy is found as _ _
@xmath8 _
_ becomes large enough . | the general conditions for the orthogonal product states of the multi - state systems to be used in quantum key distribution ( qkd ) are proposed , and a novel qkd scheme with orthogonal product states in the @xmath0 hilbert space is presented .
we show that this protocol has many distinct features such as great capacity , high efficiency .
the generalization to @xmath1 systems is also discussed and a fancy limitation for the eavesdropper s success probability is reached .
pacs number(s ) : 03.67.dd , 03.65.bz |
the magnetization dynamics can be described by means of the landau - lifshitz equation @xmath57 where @xmath23 is the magnetization vector , @xmath22 is the gyromagnetic ratio and @xmath58 is an effective field . in the hamiltonian formalism proposed by slavin and tiberkevich @xcite , eq .
( [ eq : ll ] ) is recast as a function of the complex amplitude @xmath5 defined through a holstein - primakoff transformation @xmath59 where @xmath60 is the magnetization component parallel to @xmath61 , @xmath62 and @xmath63 are perpendicular to @xmath61 , and @xmath64 is the saturation magnetization density . by expanding the resulting equation in taylor
series , the linear dynamics can be written as @xmath65 where @xmath11 is the hamiltonian of the system . in the main text
, we generalize eq .
( [ eq : linears ] ) to an array of complex amplitudes @xmath6 , so that @xmath11 becomes a matrix .
here , we outline the expressions for the hamiltonian matrices for the field contributions specified in the main text .
we assume that the anisotropy field in the magnetic elements is dominated by shape ; this is certainly the case for the permalloy islands that are commonly used .
the demagnetizing factors in thin films are defined as @xmath66 , @xmath53 , and @xmath67 in the @xmath68 , @xmath69 , and @xmath70 directions , respectively [ see eq .
( [ eq : hp ] ) ] .
computing the energy functional for every island leads to the diagonal hamiltonian matrices @xmath71}{4}\mathcal{i},\end{aligned}\ ] ] where @xmath72 is the @xmath73 identity matrix .
the external field is considered to be homogeneous throughout the spin ice structure , with magnitude @xmath74 and an arbitrary direction in space . to second order in @xmath6 ,
the hamiltonian takes a diagonal form with terms proportional to the stable magnetization direction of each island . in other words ,
only fields parallel to each magnetic element easy axis will affect linear spin waves . since we consider thin films , only @xmath75 and @xmath76 survive , and we are left with the matrices @xmath77 where @xmath78 is the @xmath73 zero matrix
. the derivation for the dipolar field is outlined in the main text .
the hamiltonian matrices obtain for the ground state are @xmath79 the elements of the above hamiltonian matrices contain contributions between a particular island and every other island in the structure but itself .
consequently , we can divide them in two terms : @xmath80 containing them sum between island @xmath25 at cell @xmath26 and every island in cell @xmath28 ; and the elements @xmath81 containing the component - wise products between the islands in cell @xmath26 . using the notation where @xmath31 is the distance between island @xmath25 in cell @xmath26 and island @xmath27 in cell @xmath28 and @xmath82 is the distance between islands @xmath25 and @xmath27 in island @xmath26
, these terms along the cartesian direction are defined as @xmath83e^{-i\vec{q}\vec{r}_{jk}},\\ \label{eq : appsytt } s_{\hat{y},\tau\tau ' } & = & \sum_{k,\tau'\neq\tau}\left[\frac{3(\vec{r}_{jk,\tau\tau'}\cdot\hat{y})^2}{\vec{r}_{jk,\tau\tau'}^5}-\frac{1}{\vec{r}_{jk,\tau\tau'}^3}\right]e^{-i\vec{q}\vec{r}_{jk}},\\ \label{eq : appsztt } s_{\hat{z},\tau\tau ' } & = & \sum_{k,\tau'\neq\tau}\left[-\frac{1}{\vec{r}_{jk,\tau\tau'}^3}\right]e^{-i\vec{q}\vec{r}_{jk}},\\ \label{eq : appsxjk } s_{\hat{x},jk}^{jk } & = & \left[\frac{3(\vec{r}_{jk}\cdot\hat{x})^2}{\vec{r}_{jk}^5}-\frac{1}{\vec{r}_{jk}^3}\right],\\ \label{eq : appsyjk } s_{\hat{y},jk}^{jk } & = & \left[\frac{3(\vec{r}_{jk}\cdot\hat{y})^2}{\vec{r}_{jk}^5}-\frac{1}{\vec{r}_{jk}^3}\right],\\ \label{eq : appszjk } s_{\hat{z},jk}^{jk } & = & \left[-\frac{1}{\vec{r}_{jk}^3}\right].\end{aligned}\ ] ] the summation on the cross direction @xmath84 can be similarly divided into two contributions , defined as @xmath85e^{-i\vec{q}\vec{r}_{jk}},\\ \label{eq : appsbjk } s_{c , jk}^{jk } & = & \left[\frac{3(\vec{r}_{jk}\cdot\hat{x})(\vec{r}_{jk}\cdot\hat{y})}{\vec{r}_{jk}^5}\right].\end{aligned}\ ] ] finally , the diagonal matrix @xmath38 in eq .
( [ eq : hd21s ] ) can be labeled from @xmath68 to @xmath86 , taking the values @xmath87 in the case of the remanent state , we note that the components of the hamiltonian are @xmath73 matrices with a similar form as the matrices in the vortex state hamiltonian .
in fact , taking the first two rows and columns of the above matrices and considering the structure s translation vector for a remanent state , @xmath88 , leads to the correct hamiltonian matrices .
intra - island exchange interactions between non - collinear spins are important to correctly describe the dynamics of spin ices , especially the modes that arise because of internal degrees of freedom . for our analytical model
, we consider each island to be composed of @xmath89 macrospins interacting with their nearest neighbors by using an effective , discrete heisenberg hamiltonian model .
the exchange hamiltonian matrix blocks defined above will now have a dimension @xmath90 to take into account the internal spins in each island .
consequently , the elements of the hamiltonian matrices above must be also expanded by replacing each of them by an @xmath91 block .
furthermore , we can introduce an arbitrary direction for each spin , so that the complex amplitude of a particular spin is @xmath92\hat{x}\nonumber\\ & + & \left[\left(1 - 2|a|^2\right)\sin{\theta}+\sqrt{1-|a|^2}\left(a+a^*\right)|\cos{\theta}|\right]\hat{y}\nonumber\\ & - & i\sqrt{1-|a|^2}\left(a - a^*\right)\hat{z},\end{aligned}\ ] ] where @xmath93 is the angle with respect to the @xmath55 axis , and the absolute values represent the isotropic nature of deviations from the magnetic elements easy axis .
we consider @xmath94 , i.e. , one bulk and two edge spins .
for the particular example of the vortex square ice , the exchange hamiltonian takes the form @xmath95 where @xmath96 indicates a @xmath97 zero matrix , and @xmath98 is expressed as a function of the exchange constant @xmath99 and the array of spin angles @xmath100 .
for example , for the case where all spins are collinear in a single magnetic element , the matrix @xmath98 takes the form @xmath101 the dominant hamiltonian matrices described above can be extended simply by completing the diagonal terms in @xmath102 and @xmath103 and calculating the summations between the new spins of _ different _ islands in @xmath104 .
non - collinear spins can be also easily included in the model by computing the products originating from the definition of eq .
( [ eq : agen ] ) .
in the main text , we explored the effect of an applied field along the @xmath55 direction .
varying the angle of such a field provides the means to explore the symmetry of the square ices . as discussed in the main text , c and
s states are energetically stable in the nanoislands . a vortex state with c - state magnetic elements
has a four - fold symmetry while both vortex and remanent states with s - state magnetic elements have a two - fold symmetry .
this can be readily shown by calculating the angle dependence of the spin waves at the @xmath49 point with an applied field of @xmath105 oe .
figure [ fig1s](a)-(b ) clearly displays these symmetries for each case , focusing on the m2 and m3 bulk modes for clarity . on the other hand ,
the remanent state in an onion state or in a macrospin approximation has a two - fold symmetry with elements magnetized at 90 degrees .
the resulting angle dependence shown in fig .
[ fig1s](c ) follows these symmetries as well .
such an angle dependence represents a valuable tool to experimentally manipulate the magnon spectra , and to infer the magnetic configuration of square ices .
coupled with the field tunability , it is possible to unambiguously determine the dominant static state throughout the structure .
( color online ) angle dependence of spin wave bands using an applied field of magnitude @xmath105 oe . panels ( a ) and ( b ) shows both the parallel modes for a c and s vortex state , exhibiting their different symmetries .
the remanent state is shown in panel ( c ) , also consistent with their symmetry . ]
as specified in the main text , full - scale micromagnetic simulations implemented a computational system containing eight islands .
the system was discretized into a mesh of size 1.25 nm@xmath106 nm@xmath107 nm and the magnetostatic interactions calculated using fast fourier transform with periodic boundary conditions applied in the plane .
the system was first set in an approximate local equilibrium state with each island homogeneously magnetized along their easy axis , approximating the vortex or remanent states .
these initial configurations were then relaxed by integrating the landau - lifshitz - gilbert equation for the micromagnetic spins using a dimensionless damping of @xmath108 .
the magnetization of the islands in a vortex ( remanent ) states then relaxed into an onion ( s ) state .
the relaxed configuration was then subjected to a uniform external field pulse of magnitude approximately 10 oe in the @xmath109 direction for 50 ps and the full landau - lifshitz - gilbert equation integrated in time steps of 0.25 ps for 10 ns using a damping of @xmath110 .
magnetization configurations were sampled every 25 ps ; the average magnetization at each time slice was fourier transformed to yield 1d spectra of the magnetization components as functions of frequency , and the sequence of 2d time slices was fourier transformed to yield full 2d amplitude and phase maps for each frequency .
these calculations were performed for constant external fields of 0 , 50 , and 100 oe along the ( 1,0 ) direction for the vortex and remanent states . | artificial square spin ices are structures composed of magnetic elements arranged on a geometrically frustrated lattice and located on the sites of a two - dimensional square lattice , such that there are four interacting magnetic elements at each vertex . using a semi - analytical approach , we show that square spin ices exhibit a rich spin wave band structure that is tunable both by external magnetic fields and the configuration of individual elements .
internal degrees of freedom can give rise to equilibrium states with bent magnetization at the edges leading to characteristic excitations ; in the presence of magnetostatic interactions these form separate bands analogous to impurity bands in semiconductors .
full - scale micromagnetic simulations corroborate our semi - analytical approach .
our results show that artificial square spin ices can be viewed as reconfigurable and tunable magnonic crystals that can be used as metamaterials for spin - wave - based applications at the nanoscale .
spin waves , or magnons , are fundamental excitations in magnetic thin films and nanostructures . because of their potential applications in information technology @xcite and computation @xcite , means to control magnon dispersion and band gap
have been studied intensively over the past few decades .
the term _ magnonics _ has been coined to describe this field of study @xcite .
one pathway to control magnon dispersions is to construct magnonic crystals @xcite that are metamaterials with a spatial modulation of the magnetic properties on length scales comparable to relevant magnonic wavelengths @xcite .
patterned thin magnetic films @xcite or topographically modulated thin films have been used to manipulate the magnon spectra @xcite .
this approach is similar to super - lattices in photonics and , fundamentally , to the crystal structure of semiconductors .
a paradigm that is the focus of recent investigation consists in actively modifying the band structure of magnonic crystals @xcite .
this has been achieved to date by use of meander - type structures @xcite and , more recently , via heating @xcite in one - dimensional ferromagnets .
artificial spin ices @xcite are another class of structures based on an organized array of nanosized magnetic elements that have been shown to support a wealth of static , dynamic , and emergent magnetic phenomena @xcite .
artificial spin ices are geometrically _ frustrated _ : the geometry of the elements and the lattice are such that all interaction energies can not be simultaneously minimized .
examples of artificial spin ices are the square ice @xcite , and the kagome ice @xcite
. the square ice is composed of magnetic stadium - shaped nanoislands positioned on the sites of a two - dimensional square lattice with lattice constant @xmath0 , fig .
[ fig1](a ) , and obeys the `` ice rules '' in which low - energy states are characterized by the magnetization in two islands pointing into a vertex and out of the vertex in the two other nanoislands .
dynamically , correlated excitations are supported in spin ices because of the magnetostatic interactions between magnetic islands @xcite .
because of their intrinsic periodicity and wealth of static states , artificial spin ices offer interesting opportunities as programmable magnonic crystals to control the magnon dispersion and band gap @xcite .
the resonant mode spectrum of square ices has been studied numerically by means of micromagnetic simulations , demonstrating the observable effects of magnetic defects @xcite .
more recently , a detailed numerical study has shown that edge modes arising from the internal degrees of freedom equally have observable consequences in the resonant spectrum in sufficiently thick nanoislands @xcite .
in fact , edge modes efficiently couple neighboring nanoislands , influencing the collective oscillations @xcite .
this is reminiscent of impurity states in semiconductors that locally modify the energy landscape and give rise to shallow electronic bands @xcite .
recent experimental results have explored the excitation spectrum of artificial spin ices @xcite but the existence and dependencies of the band diagram in square ices has not been explored to date . to close the gap between the fields magnonics and artificial spin ices , we examine square ices from the perspective of magnonics , including bands arising from the edge modes as well as the bulk modes . in this rapid communication , we study long - range dipolar - mediated two - dimensional magnon dispersion in square ices in the spirit of a tight - binding model .
in contrast to similar procedures on simpler structures @xcite , we account for the internal degrees of freedom resulting from edge modes in the nanoislands .
consequently , we are able to calculate the magnon dispersion as a function of local equilibrium states as well as its field tunability , including edge mode bands .
our semi - analytical approach provides enough degrees of freedom to qualitatively estimate the band structure of an extended square ice lattice while being computationally tractable .
( color online ) ( a ) square ice lattice with lattice constant @xmath0 , and with magnetic stadia of width @xmath1 , length @xmath2 , and thickness @xmath3 .
the stable magnetization directions ( black arrows ) of the magnetic elements in a unit cell are shown for the ( b ) ground ( vortex ) and ( c ) remanent state ( the gray - colored stadia are not part of the unit cell and are shown here for clarity ) .
edge states have two stable configurations as ( d ) c and ( e ) s states . ]
we focus on the small - amplitude excitations in two energetically stable configurations of a square ice , namely the vortex and remanent states , fig .
[ fig1](b ) and ( c ) , respectively .
the vortex state is the ground state of the system , achieved by thermal relaxation @xcite , and the remanent state can be obtained by saturating the system in an external field along the @xmath4 direction , and then slowly removing the external field , letting the system relax . in each configuration , the magnetization can bend close to the nanoisland edges @xcite , providing a local `` impurity '' . in square ices ,
two stable edge configurations satisfy the minimization of dipolar fields at the ground state , resulting in c and s states @xcite , fig .
[ fig1](d)-(e ) .
the small - amplitude dynamics in square ices can be approached semi - analytically using a hamiltonian formalism @xcite .
the same approach has been used and shown to be accurate in many dynamical regimes to date @xcite . in this formalism , the landau - lifshitz equation of motion describing conservative magnetization dynamics is cast as a function of a complex amplitude @xmath5 , using a holstein - primakoff transformation . by expanding the resulting equation in taylor series , the linear dynamics for an ensemble of complex amplitudes @xmath6 can be generally expressed ( see supplementary material ) as @xmath7 where the dagger denotes the complex transpose , @xmath8 is an array of @xmath9 complex amplitudes @xmath10=[a_1 , ... , a_n , a_1^ * , ... , a_n^*]$ ] and @xmath11 is the @xmath12 hamiltonian . the right - hand - side of eq .
( [ eq : linear ] ) includes terms up to second order in @xmath6 , corresponding to linear excitations .
beacuse of the lattice perodicity , propagating waves are bloch waves with a time dependence gvien by @xmath13 .
this allows us to reduce eq .
( [ eq : linear ] ) to an eigenvalue problem by means of colpa s grand dynamical matrix @xcite @xmath14 from which we obtain the eigenvalues @xmath15 , and the eigenvectors @xmath16 . due to the complex conjugate definition of @xmath8
, we observe that @xmath17 and @xmath18 , leading to conjugate eigenvalues in eq .
( [ eq : eigen ] ) .
the hamiltonian is related to the magnetic field @xmath19 via @xmath20 , where @xmath21 is the energy functional , @xmath22 is the gyromagnetic ratio , @xmath23 is the magnetization vector , and @xmath24 is the saturation magnetization . we consider field contributions from shape anisotropy , dipolar interactions , and intra - element exchange as well as an external applied field .
each field contribution can be reduced to a hamiltonian matrix as detailed in the supplementary material .
of particular importance are the dipolar interactions , which are the only source of inter - element coupling in our framework and the concomitant magnon dispersion .
the dipolar energy between a nanoisland @xmath25 in cell @xmath26 and all the other nanoislands @xmath27 in cells @xmath28 can be expressed as @xmath29,\end{aligned}\ ] ] where @xmath30 is the volume of the magnetic element and @xmath31 is the translation vector between the nanoisland @xmath25 in cell @xmath26 and the nanoisland @xmath27 in cell @xmath28 .
considering the bloch wave @xmath32 , where @xmath33 is the wave vector , it is possible to recast eq .
( [ eq : dipolar ] ) for the unit cell in terms of the lattice ( @xmath34 where @xmath35 ) and cross - direction ( @xmath36 ) summations . as an example , the resulting hamiltonian matrices for the ground state in the absence of exchange interactions are [ eq : hd ] @xmath37 where @xmath38 is a diagonal matrix containing inter - island interactions ( the expressions for @xmath38 and the lattice summations are shown in the supplementary material ) .
the reduction of the dipolar field to hamiltonian matrices is a key result of this work .
the magnon dispersion can be numerically calculated by solving the eigenvalue problem of eq .
( [ eq : eigen ] ) .
we consider a square ice composed of permalloy stadia with dimensions @xmath39
nm@xmath40 @xmath41 nm@xmath40 @xmath42 nm , saturation magnetization @xmath43 ka / m , and center - to - center separation of @xmath44 nm . exchange is implemented as an additional degree of freedom in a nanoisland divided by three equidistant spins coupled by the constant @xmath45 , which parametrizes the exchange in permalloy @xmath46 , where @xmath47 nm is the lattice constant and @xmath48 pj / m is the exchange stiffness .
this approximation for the exchange interaction is applicable for the low - energy sector of the magnon bands , as demonstrated below by the good quantitative agreement with full - scale micromagnetic simulations .
( color online ) band structure of the vortex state .
the insets show the magnetization vector configuration of the unit cell for each band , showcasing their excitation symmetry . ]
it is instructive to consider first the band structure neglecting internal degrees of freedom , or `` macrospin '' approximation . a typical band structure for the macrospin vortex state
is shown in fig .
[ fig2 ] .
there are four bands consistent with the available degrees of freedom in the system , one for each island . from the corresponding eigenvectors , it is possible to identify the location and symmetry of each mode .
a snapshot of the magnetic configurations at the @xmath49 point for each band ( labeled from m1 to m4 ) are shown above fig .
[ fig2 ] .
we notice that m1 has pair of islands in phase and a phase difference of @xmath50 between each pair , whereas m4 represents a mode with all islands excited in phase .
furthermore , m1 ( m4 ) has positive ( negative ) group velocity .
m2 and m3 are close in energy and consist of modes with a pairwise phase difference of @xmath51 .
note that the pairwise difference make these bands non - degenerate , resulting in anti - crossings close to the @xmath49 and m points .
these latter two modes form narrow bands that separate away form the @xmath49 point , and establish a band gap reaching @xmath52 mhz between the @xmath49 and x point of m1 and m2 , respectively .
bands effectively touch at the @xmath49 and @xmath53 points .
however , we did not observe band inversion in any calculation .
we now include exchange interactions in our framework . by dividing each magnetic island into three equidistant spins , we now have access to 12 bands . in the ground state ,
three configurations are stable : homogeneous or onion @xcite , c , and s states .
the corresponding band diagrams are shown in fig .
[ fig3 ] .
the additional degrees of freedom give rise to lower frequency bands identified as edge modes ( black dashed lines ) , also showing anti - crossing behavior .
we observe that the bulk modes ( blue lines ) maintain their qualitative features .
however , the bandgaps are enhanced due to the additional energy incorporated into the system .
furthermore , the particular magnetic configuration quantitatively modifies the band diagram , indicating that edge bending can be compared to impurity states in semiconductor materials . because a transition between c and s states can be induced by , _
e.g. , _
temperature @xcite , this can be used as another avenue to program the magnonic response of the square ice . in the remanent state , the unit cell
is composed of two magnetic islands , fig .
[ fig1](c ) .
the band diagrams for a macrospin and stable onion and s configurations are shown in fig .
[ fig4 ] , exhibiting similar features as discussed above .
( color online ) band diagram for the vortex state in a ( a ) onion , ( b ) c , and ( c ) s state .
the bulk ( edge ) modes are depicted in blue ( dashed black ) lines .
the schematic of each static configuration is also shown for each case . ]
we now explore the effect of an applied field @xmath54 on the square ice .
we consider a feasible experimental scenario of an in - plane field along the @xmath55 direction and detection of coherent excitations ( at the @xmath49 point ) by means of resonance measurements ( the effect of field angle is shown in the supplementary material ) .
note that in our framework , the stable magnetization direction of the magnetic nanoislands is set and assumed _ a priori _ , _ i.e. _ , only small amplitude variations are accessible .
in fact , large fields induce imaginary eigenvalues , denoting decaying modes and thus the breakdown of our model .
we study the effect of field magnitudes between @xmath56 oe which maintains real eigenvalues .
the results obtained for both vortex and remanent states under macrospin approximation are shown in fig .
[ fig5](a - b ) . in the case of the vortex state , we observe that the coherent modes , m1 and m4 , have positive and negative tunabilities , respectively , whereas m2 and m3 exhibit only slight tunability . in the case of the remanent state
we observe either a positive or negligible tunability .
the strongly tunable modes can be traced to those magnetic elements parallel to the applied field .
this is also consistent with the landau - lifshitz equation predicting a blue ( red ) shift of frequencies when the _ internal _ field increases ( decreases ) .
the modes with negligible tunability correspond to magnetic elements perpendicular to the field . by considering edge bending , a richer behavior for the tunability of both the vortex and remanent states
is obtained , fig .
[ fig5](c - d ) .
for both the vortex in an onion state and the remanent s state , we observe similar tunabilities for the bulk and low - frequency edge modes . in all cases , the slope of each band is generally different , leading to band crossings , and implying that the bandgaps in square ices can be manipulated by an applied magnetic field .
( color online ) band diagram for the remanent state in a ( a ) macrospin , ( b ) onion , and ( c ) s states .
the bulk ( edge ) modes are depicted in blue ( dashed black ) lines .
the inset shows the magnetization vector configuration of the unit cell for each band .
the schematic of each static configuration is also shown for each case . ]
full - scale micromagnetic simulations were performed for comparison with the semi - analytical model .
we used a computational system containing eight islands and imposing periodic boundary conditions consistent with the geometry described above ( see supplemental material for details ) .
the results are shown as red circles in fig .
[ fig5 ] ( note that the micromagnetic modeling only returns modes that are even in the unit cell because the exciting field is uniform , while the semi - analytical model captures all modes irrespective of symmetry ) . for the vortex state ,
a good agreement for the bulk modes is obtained from the macrospin model .
further comparison with the extended semi - analytical model also shows excellent agreement with the low - frequency edge modes .
for the remanent state , the macrospin model yields a good qualitative agreement with the micromagnetic results .
a three - spin s - state model also yields good agreement with the micromagnetic low - frequency modes , especially in view of the simplistic treatment in the three - spin model of the smooth static equilibrium magnetization in the micromagnetic model .
( color online ) magnon frequencies at the @xmath49 point as a function of an external field applied along the @xmath55 axis for the ( a ) macrospin vortex , ( b ) macrospin remanent , ( c ) onion vortex , and ( d ) remanent s states .
the bulk ( edge ) modes are depicted in blue ( dashed black ) lines .
the red dots are obtained from micromagnetic simulations . ]
we remark that in both real and micromagnetically - modeled nanoislands , there are many higher - order modes , beyond what can be described by the three - spin model considered here , because of the many internal degrees of freedom .
such higher - order modes have many internal nodal lines of the magnetization eigenmodes .
therefore , the magnetostatic fields emanating from such modes decay rather quickly in space .
this results in a weak coupling between different islands so the magnonic bands arising from such modes are non - dispersive with no phase or group velocity , and are not interesting here .
it is also noteworthy that a strong variation of the islands aspect ratio can significantly affect the excited frequencies , _
i.e. _ , in the nanowire and circular dot limits
. moreover , we expect the thickness to play an important role in the ultra - thin film regime , where the anisotropy becomes perpendicular or can favor a vortex state inside each stadium . in summary
, we have calculated the magnon band structure and the mode tunability at the @xmath49 point for a square ice in two equilibrium states , the vortex ( or ground ) state , and the remanent state , using a model that includes internal degrees of freedom of the islands as well as edge bending .
the good quantitative agreement with micromagnetic simulations confirms the accuracy of the small - amplitude semi - analytical model while avoiding the computational limitations intrinsic to fully three - dimensional micromagnetic simulations .
these results show that the magnon spectra , and therefore group and phase velocities as well as band gap , can be manipulated by external fields . in particular ,
the edge modes give rise to separate magnon bands allowing for a larger parameter space in terms of magnon control .
this suggests that square ices can be considered metamaterials for spin waves .
in addition , the square ice is in principle reconfigurable in that the magnetization in individual islands can be changed by the application of external fields ( _ e.g. , _ from vortex to remanent state ) or temperature , or by using more sophisticated techniques such as using spin torque by patterning nanocontacts on the elements or making the elements part of magnetic tunnel junctions .
this opens up the possibility of two - dimensional reprogrammable magnonic crystals comprised of an artificial square spin ice .
e. i. acknowledges support from the swedish research council , reg .
no .
637 - 2014 - 6863 .
the work by o. h. was funded by the department of energy office of science , materials sciences and engineering division .
e. i. was partly supported by the u.s .
department of energy , office of science , materials sciences and engineering division through the materials theory institute .
we gratefully acknowledge the computing resources provided on blues , a high - performance computing cluster operated by the laboratory computing resource center at argonne national laboratory .
the work by r. l. s. was funded by epsrc ep / l002922/1 .
43ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * ( ) and , eds . , @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.107.127204 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.108.015505 [ * * , ( ) ] @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) link:\doibase 10.1103/physrevb.93.100401 [ * * , ( ) ] @noop * * ( ) link:\doibase 10.1002/adfm.201505165 [ ( ) , 10.1002/adfm.201505165 ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) |
throughout this paper , @xmath8 is an arbitrary field and @xmath9 is a fixed non - zero element of @xmath8 . given a positive integer @xmath1 , @xmath6 denotes the symmetric group of degree @xmath1 and @xmath5 denotes the iwahori
hecke algebra of @xmath6 : this is the unital associative @xmath8-algebra with generators @xmath10 and relations @xmath11 the representation theory of @xmath5 closely resembles ( and informs ) the modular representation theory of @xmath6 .
this theory was first explored in detail by dipper and james , whose account @xcite we closely follow . in particular , they define a ` permutation module ' @xmath4 for each composition @xmath2 of @xmath1 , and a specht module @xmath3 for each partition @xmath0 of @xmath1 . understanding the structure of specht modules is an important goal in understanding the representation theory of @xmath5 , and one well - studied aspect of this is the examination of @xmath5-homomorphisms between specht modules .
the aim of this paper is to provide a ` straightening ' result which yields ( as long as @xmath12 ) a fast algorithm to determine @xmath13 .
this result actually concerns the space of homomorphisms from @xmath3 to @xmath4 ; an important family of such homomorphisms is labelled by @xmath0-tableaux of type @xmath2 ; when @xmath12 , this family spans the space @xmath14 ( * ? ? ?
* corollary 8.2 ) .
regardless of @xmath9 , the space spanned by the tableau homomorphisms has a basis consisting of the homomorphisms labelled by _ semistandard _
however , existing proofs of this result involve a loss of information which means that it is not easy to express a given tableau homomorphism explicitly in terms of semistandard homomorphisms . in this paper
we prove a relation between tableau homomorphisms which allows us to do exactly this ; this closely resembles the _ garnir relations _ which are typically used to straighten elements of the specht module . in the next section ,
we very briefly recall the background information that we need , referring to @xcite for most of it . in section [ s : twopart ]
we prove our garnir - like relation for tableaux with only two rows . in section [ s : rowrem ]
we prove the necessary results concerning row removal to show that our garnir relation holds generally .
finally , in section [ s : algo ] we give an algorithm for expressing a tableau homomorphism as a linear combination of semistandard homomorphisms .
we take almost all of our notation from @xcite , with only very minor modifications .
we provide a brief index here of most of the notation , referring the reader to @xcite for definitions .
homomorphisms are considered in more detail below .
note that we follow @xcite by considering right modules and having all functions act on the right .
[ cols= " < , < " , ] in this paper we frequently employ multisets of positive integers . if @xmath15 is a multiset of positive integers , then we write @xmath16 for the number of @xmath17s in @xmath15 . given multisets @xmath18 , we write @xmath19 for the multiset with @xmath20 for each @xmath17 .
if @xmath21 is a tableau , we write @xmath22 for the multiset of entries in row @xmath17 of @xmath21 .
given our fixed @xmath9 , we write @xmath23=1+q+\dots+q^{n-1}$ ] for any non - negative integer @xmath1 , and @xmath23!=\prod_{i=1}^n[i]$ ] . these expressions are useful in the study of @xmath5 , since we have @xmath23!=\sum_{w\in{w_}n}q^{l(w)}$ ] .
furthermore , if @xmath24 , then any @xmath25 can be written uniquely as @xmath26 with @xmath27 , @xmath28 and @xmath29 , so we have @xmath30!}{\prod_i[\la_i]!}.\ ] ] note that the right - hand side of this expression is a polynomial in @xmath9 , so makes sense even when @xmath31!$ ] is zero .
if @xmath9 is an indeterminate and @xmath32 , then we can define the _ quantum binomial coefficient _ @xmath33}{0pt}{}{n}{r}}=\frac{[n]!}{[r]![n - r]!}.\ ] ] in fact , this is a polynomial in @xmath9 , so we can extend the definition to arbitrary @xmath9 by defining @xmath34}{0pt}{}{n}{r}}$ ] to be the specialisation of this polynomial .
we now recall the definitions from @xcite on homomorphisms between permutation modules . throughout this subsection
we fix @xmath35 .
recall that a @xmath0-tableau @xmath21 of type @xmath2 is _ row - standard _ if its entries increase weakly along the rows .
we write @xmath36 for the set of row - standard @xmath0-tableaux of type @xmath2 .
if in addition @xmath37 , then we say that a @xmath0-tableau of type @xmath2 is _ semistandard _ if its entries are weakly increasing along the rows and strictly increasing down the columns , and we write @xmath38 for the set of semistandard @xmath0-tableaux of type @xmath2 .
given a @xmath0-tableau @xmath21 of type @xmath2 , we define the permutation @xmath39 by specifying that @xmath40 is the row - standard @xmath2-tableau ( of type @xmath41 ) in which @xmath17 belongs to row @xmath42 if the position occupied by @xmath17 in @xmath43 is occupied by @xmath42 in @xmath21 . for example , if 1 @xmath44 so that @xmath45 .
the map @xmath46 defines a bijection from @xmath36 to the set @xmath47 of minimal - length ( @xmath48,@xmath49 ) double coset representatives in @xmath6 . for later use
we record the simple fact that if @xmath50 , then @xmath51 now we consider homomorphisms .
if @xmath50 , then we define an @xmath5-homomorphism @xmath52 by @xmath53 we call this a _
tableau homomorphism_. in the case where @xmath0 is a partition , we write @xmath54 for the restriction of @xmath55 to the specht module @xmath3 .
we remark that in more recent literature @xmath55 , @xmath54 are written as @xmath56 , @xmath57 .
now we have the following theorem . [ semi ] 1 . if @xmath58 , then the set @xmath59 is a basis for @xmath60 .
2 . if @xmath37 and @xmath24 , then the set @xmath61 is a basis for the subspace of @xmath14 spanned by @xmath62 .
if @xmath37 , @xmath24 and @xmath12 , then every @xmath5-homomorphism @xmath63 can be extended to a homomorphism @xmath64 .
hence @xmath61 is a basis for @xmath14 .
the focus of this paper is part ( 2 ) of the above theorem ; we provide an explicit algorithm for writing a given homomorphism @xmath55 as a linear combination of semistandard homomorphisms .
of course , the proof of theorem [ semi](2 ) does yield an algorithm for ` semistandardising ' a tableau homomorphism , but it is rather slow , even in the symmetric group case @xmath65 ; one must consider the image of a polytabloid under the different homomorphisms @xmath55 , and keep track of the coefficients of various tabloids in these images .
our algorithm involves just manipulation of row - standard tableaux , forgetting the underlying homomorphisms . in simple computer experiments
, our algorithm seems considerably quicker .
there are several other useful ways to write the image @xmath66 . from @xcite
, we have @xmath67 by considering the row permutations sending @xmath21 to the various @xmath68 , we can also write this ( following ( * ? ? ? * theorem 3.4 ) ) as @xmath69 where @xmath70 is the ` composition defined by reading @xmath21 along rows ' , i.e. the composition @xmath71 we will often use the expression ( h3 ) , and write the term @xmath72 as @xmath73 . symmetrically , we have @xmath74 where @xmath75 is the composition @xmath76 we shall write the factor @xmath77 as @xmath78 .
our main interest is in understanding linear combinations @xmath79 ( where @xmath80 and @xmath81 ) for which @xmath82 .
since @xmath3 is generated by @xmath83 , this condition amounts to saying that @xmath84 is annihilated by @xmath85 .
now we come to our main result .
we begin with tableaux having only two rows . [ main ] suppose @xmath86 is a partition , and @xmath87 are multisets of positive integers with @xmath88 and @xmath89 .
let @xmath90 be the set of all pairs @xmath91 such that @xmath92 and @xmath93 .
for each @xmath94 , let @xmath95 $ ] denote the row - standard @xmath0-tableau with @xmath96 ^ 1=r\sqcup u,\qquad a[u , v]^2=t\sqcup v.\ ] ] then @xmath97}{\bigl[\medmath{\genfrac{}{}{-2pt}{}{t_i+v_i}{t_i}}\bigr]}\prod_{i < j}q^{r_ju_i+t_iv_j}\hat\phi_{a[u , v]}=0.\ ] ] an example of this theorem is given in section [ s : algo ] . the way we prove the theorem will be to express the given homomorphism ( extended to @xmath98 ) as a composition @xmath99 such that @xmath100 . to do this , we need a few simple results concerning composition of tableau homomorphisms .
some of these are probably known , but we prove everything in order to keep our account self - contained .
our first result shows how to use a tableau of the form @xmath101 to merge two rows of a tableau .
[ bigpart ] suppose @xmath102 is a two - part composition of @xmath103 , and @xmath15 is a row - standard @xmath104-tableau of type @xmath105 .
let @xmath106 be the unique row - standard @xmath107-tableau of type @xmath104 , and @xmath21 the unique row - standard @xmath107-tableau of type @xmath108 .
then @xmath109}\prod_{i < j}q^{c^1_jc^2_i}\phi_a.\ ] ] equation ( h1 ) shows that for a row - standard @xmath107-tableau @xmath110 of any type , we have @xmath111 so we just need to show that @xmath112 is the appropriate scalar multiple of @xmath113
. we have @xmath114 by ( h3 ) , while ( h4 ) gives @xmath115 where @xmath75 is the composition @xmath116 .
so @xmath117 to evaluate this expression , we use the fact that @xmath118 for @xmath25 .
the length of @xmath119 is @xmath120 , and since @xmath121 , we have @xmath122!}{\prod_i[\pi_i]!}=\prod_i{\bigl[\medmath{\genfrac{}{}{-2pt}{}{a^1_i}{c^1_i}}\bigr]}.\ ] ] the result follows .
the next result is a more complicated version of the result above .
[ moresoph ] suppose @xmath123 is a composition of @xmath1 .
let @xmath15 be a row - standard @xmath123-tableau of type @xmath24 , and let @xmath106 be the row - standard @xmath124-tableau in which there are @xmath42 @xmath125s and @xmath126 @xmath127s in the first row , and @xmath128 @xmath129s and @xmath130 @xmath131s in the second .
let @xmath21 be the row - standard @xmath124-tableau with @xmath132 and @xmath133 .
then @xmath109}{\bigl[\medmath{\genfrac{}{}{-2pt}{}{a^2_i}{c^3_i}}\bigr]}\prod_{i < j}q^{c^1_jc^2_i+c^3_jc^4_i}\phi_a.\ ] ] since @xmath134 , ( h3 ) gives @xmath135 and ( h4 ) gives @xmath136 where @xmath75 is the composition @xmath137 .
so @xmath138 @xmath139 factorises as @xmath140 , where @xmath141 , so the above expression becomes @xmath142 where @xmath143 is the composition @xmath144 , and @xmath145 .
now @xmath39 conjugates the pair @xmath146 to the pair @xmath147 , where @xmath148 and this gives @xmath149 @xmath150 is contained in @xmath151 , so @xmath152}{\bigl[\medmath{\genfrac{}{}{-2pt}{}{a^2_i}{c^3_i}}\bigr]}x_{(r+u , v+t)}.\ ] ] so we have @xmath153}{\bigl[\medmath{\genfrac{}{}{-2pt}{}{a^2_i}{c^3_i}}\bigr]}\prod_{i < j}q^{c^1_jc^2_i+c^3_jc^4_i}{\settowidth{\adju}{\hbox{\scriptsize $ \scrd^{-1}_\rho\cap w_\la$}}\addtolength\adju{-\bdju}\setlength\adju{0.45\adju}\hspace{-0.8\adju}\sum_{w\in\scrd^{-1}_\rho\cap w_\la}\hspace{-\adju}t_w}t_{1_a}x_{(r+u , v+t)}\\ & = \prod_i{\bigl[\medmath{\genfrac{}{}{-2pt}{}{a^1_i}{c^1_i}}\bigr]}{\bigl[\medmath{\genfrac{}{}{-2pt}{}{a^2_i}{c^3_i}}\bigr]}\prod_{i < j}q^{c^1_jc^2_i+c^3_jc^4_i}x_{(r , u , v , t)}\phi_a.\qhere{\sum_{i < j}}\end{aligned}\ ] ] the next result effectively enables us to split a row of a tableau in two , using a tableau of the form @xmath154 [ easy ] suppose @xmath123 is a composition of @xmath1 , and let @xmath110 be the unique @xmath123-tableau in which the entries in row @xmath125 are equal to @xmath125 , the entries in rows @xmath127 and @xmath129 are all equal to @xmath127 , and the entries in row @xmath131 are all equal to @xmath129 .
let @xmath155 be a row - standard @xmath156-tableau of any type , and let @xmath157 be the set of all row - standard @xmath123-tableaux @xmath15 such that @xmath158 then @xmath159 we have @xmath160 , so by ( h2)@xmath161 where @xmath2 is the type of @xmath155 .
but it is easy to see that @xmath162 so @xmath163 the preceding results allow us to express the homomorphism of theorem [ main ] as a composition of homomorphisms indexed by row - standard tableaux . [ compo ]
suppose @xmath164 are as in theorem [ main ] , and let @xmath165 , @xmath166 , @xmath167 , @xmath168 , @xmath169 .
let @xmath155 be the row - standard @xmath170-tableau which has @xmath171 , let @xmath106 be the @xmath124-tableau from proposition [ moresoph ] , and let @xmath110 be the @xmath123-tableau from lemma [ easy ] . then @xmath172}{\bigl[\medmath{\genfrac{}{}{-2pt}{}{t_i+v_i}{t_i}}\bigr]}\prod_{i < j}q^{r_ju_i+t_iv_j}\phi_{a[u , v]}.\ ] ] from lemma [ easy ] , we have @xmath173 each tableau @xmath174 satisfies @xmath175 for some pair @xmath94 .
when we compose @xmath176 with @xmath177 using proposition [ moresoph ] , the tableau @xmath21 we obtain is precisely @xmath95 $ ] , and the coefficient of @xmath55 is the coefficient of @xmath178}$ ] in the proposition . using proposition [ compo ] , we must show that the composition @xmath179 kills @xmath3 .
in fact , we will show that @xmath180 kills @xmath3 .
the codomain of @xmath180 is the permutation module @xmath181 , while the specht module @xmath3 is generated by @xmath83 , so it suffices to show that @xmath182 .
but this is straightforward : @xmath181 is spanned by elements @xmath183 which correspond to @xmath170-tabloids ( this correspondence is introduced on @xcite ) . if @xmath126 is an @xmath170-tabloid , then because @xmath184 there must be some pair @xmath185 which lie in the same column of @xmath186 and both lie in the second row of @xmath126
@xmath187 may be factorised as @xmath188 , and the fact that @xmath189 both lie in the same row of @xmath126 means that @xmath190 .
hence @xmath191 .
the argument used here to show that @xmath182 is a special case of ( * ? ? ?
* lemma 4.1 ) which shows that @xmath192 whenever @xmath193 .
in this section , we prove a result which allows us to apply our garnir - type relation to tableaux with more than two rows . using the results of this section
, we will be able to prove a @xmath9-analogue of ( * ? ? ?
* lemma 4 ) , which essentially says that given a linear relation between homomorphisms @xmath54 , we can add a fixed combination of rows to each of the tableaux and preserve the relation .
we remark that row - removal for tableau homomorphisms has been considered at length , by the author and lyle @xcite for the symmetric group , and by lyle and mathas @xcite for the hecke algebra .
however , the result we prove here seems not to have appeared before .
we need to introduce some more notation .
given @xmath194 and @xmath195 , we define the homomorphism @xmath196 by @xmath197 .
so @xmath198 is the usual embedding of @xmath199 in @xmath5
. clearly we have @xmath200 for any @xmath201 .
we will also occasionally use @xmath202 to denote the corresponding map @xmath203 .
the main result we want to prove here is the following .
[ topmain ] suppose @xmath37 and @xmath24 , and @xmath204 are row - standard @xmath0-tableaux of type @xmath2 which are identical in row @xmath125 .
write @xmath205 , and let @xmath206 be the @xmath207-tableaux obtained by deleting the first row from each of @xmath204 . if @xmath81 are such that @xmath208 , then @xmath82 .
in order to prove proposition [ topmain ] , we make two comparisons : * we compare @xmath66 with @xmath209 , for a row - standard @xmath0-tableau @xmath21 ; * we compare the term @xmath85 which defines the specht module with @xmath210 .
[ rowtophom ] suppose @xmath21 is a row - standard @xmath0-tableau of type @xmath2 .
write @xmath205 , and let @xmath211 be the @xmath207-tableau obtained by deleting the first row of @xmath21 .
let @xmath212 be the type of @xmath211 .
then we have @xmath213 for some @xmath214 depending only on @xmath215 .
the way we prove proposition [ rowtophom ] is to express @xmath55 as a composition of tableau homomorphisms .
let @xmath106 be the @xmath0-tableau which is obtained from @xmath21 by replacing every entry in the first row with a @xmath125 and adding @xmath125 to every entry below the first row , and let @xmath15 be the tableau whose first row is the same as the first row of @xmath21 , and whose @xmath17th row consists of @xmath216 @xmath217s , for @xmath218 .
note that both the type of @xmath106 and the shape of @xmath15 equal the composition @xmath219 .
1 suppose @xmath220 , @xmath221 and @xmath222 then @xmath223 , and @xmath224 [ comptop ] with the notation above , we have @xmath225 . using ( h3 ) , we get @xmath226 since the entries in the @xmath17th row of @xmath15 are constant for @xmath218 , we have @xmath227 . on the other hand , since the entries in the first row of @xmath106 all equal @xmath125 , we have @xmath228 so @xmath229 and @xmath230 commute .
it is not too hard to see that @xmath231 , and it is immediate from the definition of @xmath73 that @xmath232 , so @xmath233 we can write @xmath234 as @xmath235 , so @xmath236 we have @xmath237 ; as noted in the proof of lemma [ comptop ] , @xmath230 is just @xmath238 , and it is immediate that @xmath239 .
putting this together , we get @xmath240 the tableau @xmath15 depends only on @xmath215 , so the result follows .
the next thing we need in order to prove proposition [ topmain ] is the following .
[ tytop ] suppose @xmath0 is a partition of @xmath1 , and write @xmath205 .
then @xmath241 to prove this , it will be helpful to generalise to _
skew partitions_. for us , a skew partition is an ordered pair of partitions ( written @xmath242 ) where @xmath243 for each @xmath17 .
we define the young diagram @xmath244 $ ] to be the set difference @xmath245\setminus[\la]$ ] .
we identify two skew - partitions which have the same young diagram , so that for example @xmath246 .
if @xmath242 is a skew - partition , then we define its conjugate @xmath247 to be @xmath248 .
we say that a skew partition @xmath207 is obtained from a skew partition @xmath0 _ by translation _ if for some @xmath249 we have @xmath250=\lset{(r+i , c+j)}{(r , c)\in[\mu]}$ ] . for most purposes @xmath207 and @xmath0 can be treated as equal in this case .
if @xmath0 is a skew partition , then the tableaux @xmath43 , @xmath186 , and hence the permutation @xmath251 , can be defined in the obvious way , and so can @xmath187 .
note that if @xmath207 is a skew - partition obtained from @xmath0 by translation , then @xmath252 and @xmath253 .
now proposition [ tytop ] follows by applying the next lemma @xmath254 times .
[ tytopskew ] suppose @xmath0 is a skew partition of @xmath255 , and let @xmath70 be the skew partition obtained by removing the first node from the first non - empty row of @xmath245 $ ]
. then @xmath256 .
suppose the node removed from @xmath245 $ ] to obtain @xmath257 $ ] is @xmath258 .
then @xmath259 is the standard tableau in which the number @xmath125 appears in position @xmath258 , while the numbers @xmath260 appear in order down successive columns .
so , if the entry at the bottom of column @xmath261 is @xmath103 , then @xmath262 , where @xmath263 is the cycle @xmath264 .
moreover , we have @xmath265 , so @xmath266 .
now @xmath187 is the commuting product of factors @xmath267 , where @xmath268 ( with @xmath269 being the group of permutations of the entries in column @xmath270 of @xmath186 ) .
@xmath271 is the product of factors @xmath272 defined similarly .
for @xmath273 we have @xmath274 , and @xmath267 commutes with @xmath275 . for @xmath276 , we can factorise @xmath267 as @xmath277 using right coset representatives , and @xmath278 again commutes with @xmath275 . for @xmath279
we have @xmath280 and @xmath281 . and so @xmath282 let @xmath283 , so that @xmath284
. then @xmath285 , and @xmath286 we have @xmath287 and @xmath288 , where @xmath289 so @xmath290 @xmath3 is generated by @xmath83 , so we need to show that @xmath291 using propositions [ rowtophom ] and [ tytop ] , this equals @xmath292 where @xmath293 . note that @xmath294 is independent of @xmath17 : because @xmath204 all have the same first row , @xmath206 all have the same type .
so we have @xmath295 the term in the middle is @xmath296 , which is zero by assumption .
we now consider a counterpart to proposition [ topmain ] in which we remove a row from the bottom of a tableau .
[ botmain ] suppose @xmath37 and @xmath24 , and let @xmath297 be maximal such that @xmath298 .
suppose @xmath204 are row - standard @xmath0-tableaux of type @xmath2 which are identical in row @xmath297 .
write @xmath299 , and let @xmath206 be the @xmath207-tableaux obtained by deleting the @xmath297th row from each of @xmath204 .
if @xmath81 are such that @xmath208 , then @xmath82 . this is proved in a very similar way to proposition [ topmain ] ; the proof is in fact slightly simpler , because we do not need the functions @xmath300 , and we can more easily avoid skew partitions .
so we just give the two main propositions which are analogous to propositions [ rowtophom ] and [ tytop ] , and leave the reader to fill in the remaining details .
[ rowbothom ] suppose @xmath21 is a row - standard @xmath0-tableau of type @xmath2 , and let @xmath297 be maximal such that @xmath298 . write @xmath299 , and let @xmath211 be the @xmath207-tableau obtained by deleting the last row of @xmath21 .
let @xmath212 be the type of @xmath211 .
then we have @xmath301 where @xmath214 depends only on @xmath215 .
suppose @xmath0 is a partition of @xmath1 , and write @xmath302 , where @xmath297 is maximal such that @xmath298
. then @xmath303
we now explain how to use theorem [ main ] to write a row - standard homomorphism @xmath54 as a linear combination of semistandard homomorphisms .
we begin by considering two - part partitions .
suppose @xmath86 is a two - part partition , and @xmath21 is a row - standard @xmath0-tableau of arbitrary type . if @xmath21 is not semistandard , then there is some @xmath304 such that the @xmath305-entry of @xmath21 is at least as large as the @xmath306-entry .
suppose the @xmath305-entry equals @xmath307 .
now define the following multisets : @xmath308 then @xmath309 and @xmath310 , so @xmath87 satisfy the conditions of theorem [ main ] .
moreover , the tableau @xmath21 arises as @xmath311 $ ] , where @xmath312 note in particular that @xmath313 consists of the @xmath314 largest elements of @xmath315 .
note also that all the elements of @xmath316 are strictly less than all the elements of @xmath313 , and the same is true for @xmath317 and @xmath318 , so the coefficient of @xmath54 in theorem [ main ] is @xmath125 .
so theorem [ main ] allows us to express @xmath54 as a linear combination of homomorphisms @xmath319 , for row - standard tableaux @xmath106 satisfying @xmath320 we can repeat the process for each @xmath106 which is not semistandard , and eventually express @xmath54 as a linear combination of semistandard homomorphisms .
the above inequality guarantees that this process will terminate . for this example
, we abuse notation and write a homomorphism @xmath319 just as @xmath106 . suppose 1 @xmath321 applying theorem [ main ] with @xmath322 we get @xmath323\young(11224,3333)-\young(11234,2333)-q^3\young(11223,3334).\ ] ] applying theorem [ main ] again , we have @xmath324\young(11223,3334)-[2]\young(11224,3333)-\young(11233,2334)\ ] ] so that @xmath325 [ mainrowrem ] suppose @xmath37 and @xmath24 , and @xmath326 .
let @xmath207 be the partition @xmath327 .
suppose @xmath204 are row - standard @xmath0-tableaux of type @xmath2 which are identical outside rows @xmath328 , and let @xmath206 be the @xmath0-tableaux obtained by deleting rows @xmath329 and @xmath330 of @xmath204 . if @xmath81 are such that @xmath208 , then @xmath82 .
now suppose @xmath21 is a row - standard @xmath0-tableau .
if @xmath21 is not semistandard , then there must be some @xmath297 such that the tableau @xmath211 consisting of rows @xmath331 of @xmath21 is not semistandard .
using the algorithm described above , we can express @xmath332 as a linear combination of homomorphisms of the form @xmath333 where each @xmath334 satisfies @xmath335 .
applying theorem [ mainrowrem ] , we can express @xmath54 as a linear combination of homomorphisms @xmath319 where each @xmath106 satisfies @xmath336 so by induction we can express @xmath55 as a linear combination of semistandard homomorphisms .
we now explain briefly how to compute the space of @xmath5-homomorphisms between two specht modules when @xmath12 , which is the main motivation for proving the results in this paper .
a more detailed discussion of this topic can be found in the paper @xcite by lyle .
suppose @xmath337 .
theorem [ semi](2 ) says that the set @xmath338 is a basis for @xmath14 as long as @xmath12 .
since @xmath339 , an @xmath5-homomorphism from @xmath3 to @xmath340 is simply a linear combination of semistandard homomorphisms whose image lies inside the specht module @xmath340 .
fortunately , there is a straightforward way to check this condition : @xmath340 may be characterised ( * ? ? ?
* theorem 7.5 ) as the intersection of the kernels of certain tableau homomorphisms @xmath341 from @xmath4 to other permutation modules @xmath342 . given a tableau homomorphism @xmath52 ,
it is known how to compose @xmath55 with @xmath341 ; this result is ( * ? ? ?
* proposition 2.14 ) , generalising the result ( * ? ? ?
* lemma 5 ) proved for the symmetric group .
however , this composition is given as a linear combination of tableau homomorphisms which are not always semistandard ( even if @xmath21 is semistandard ) .
since the set @xmath62 is linearly dependent in general , this can make it difficult to tell whether a given linear combination @xmath343 is non - zero . but using the results in this paper , one can quickly express this homomorphism as a linear combination of semistandard homomorphisms , and hence discover whether it is non - zero .
so one has a reasonably fast algorithm for computing the space of @xmath5-homomorphisms between two specht modules when @xmath12 .
the author has implemented this in a small package for the gap4 system @xcite .
this package has already yielded new results on homomorphisms between specht modules @xcite . | suppose @xmath0 is a partition of @xmath1 and @xmath2 a composition of @xmath1 , and let @xmath3 , @xmath4 denote the specht module and permutation module defined by dipper and james for the iwahori
hecke algebra @xmath5 of the symmetric group @xmath6 .
we give an explicit fast algorithm for expressing a tableau homomorphism @xmath7 as a linear combination of semistandard homomorphisms . along the way we provide a utility result related to removing rows from tableaux .
= 12.5 cm = 8.839 cm 0 this is the author s version of a work that was accepted for publication in the journal of algebra .
changes resulting from the publishing process , such as peer review , editing , corrections , structural formatting , and other quality control mechanisms may not be reflected in this document .
changes may have been made to this work since it was submitted for publication .
a definitive version was subsequently published in + _ j. algebra _ = 8.27 in = 11.69 in |
many thousands of x - ray sources are discovered using x - ray cameras with modest spectral resolution , such as the ccd detectors on the @xmath1 , @xmath2 , and @xmath3-@xmath4 space observatories . after detecting and locating the source
, the observer often seeks to estimate intrinsic properties of the source , particularly its flux in a broad band of interest and spectral characteristics such as power law index , thermal plasma temperature , or column density of line - of - sight gas absorbing soft x - rays .
low - resolution source and background spectra are extracted from the ccd data , and @xmath5 or maximum likelihood statistical nonlinear regression of these data with multi - parameter spectral models is performed with codes such as xspec , sherpa , or midas .
we have been particularly concerned with the study of large populations of faint x - ray sources associated with pre - main sequence ( pms ) stars residing in or near molecular clouds ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these and other studies of x - ray - bright young stars use a spectral model family involving one- or two - temperature thermal plasmas with a particular pattern of non - solar elemental abundances ( e.g. * ? ? ?
each star is subject to a different absorption due to its chance location in the molecular cloud and , for the youngest systems , its local protostellar envelope .
however , when faint sources are considered , there can be too few counts to obtain reliable spectral fits using regression methods ; a wide range of models can be fit to the same distribution of source photon energies .
while apparent broad - band fluxes can still be estimated , the role of absorption , the intrinsic fluxes , and true plasma temperature(s ) are often poorly constrained from spectral modeling of faint sources . also
, parametric regression analyses are often inhomogeneous in practice ; two - temperature plasma models may be fit to lightly - absorbed , bright ( say , @xmath6200 count ) sources while fainter and/or highly - absorbed sources may be fit with one - temperature plasma models . there is thus a motivation to consider nonparametric approaches to interpreting x - ray ccd energy distributions which can be applied in a uniform fashion to bright and faint sources . in this work ,
we examine the nonparametric inference of intrinsic x - ray luminosities and absorbing column densities which can be efficiently obtained for large populations of both bright and faint x - ray ccd sources .
the concept of the method is similar to the long - standing use of color - magnitude diagrams in optical and infrared astronomy . for pms stars , a photometric magnitude and at least one suitable color index , such as the @xmath7 @xmath8 @xmath9 color magnitude diagram ,
are used in conjunction with distance , age estimates , and stellar interior evolutionary models to estimate stellar masses , bolometric luminosities , and line - of - sight absorption . in our proposed method ,
photometric x - ray quantities such as apparent hard - band background - corrected flux ( analogous to the @xmath7 magnitude ) and background - corrected total - band median energy ( analogous to the @xmath9 color ) are used to estimate absorption - corrected x - ray fluxes , and line - of - sight absorbing column densities .
we apply our nonparametric estimates to a large sample of faint x - ray sources in the m 17 star - forming region ( * ? ? ? * and townsley et al .
, in preparation ) .
intrinsic spectra for this class of stars are calibrated to the high signal - to - noise @xmath2 acis - i spectra from the sensitive @xmath2 orion ultradeep project ( coup ; * ? ? ?
reliability of our nonparametric methods , including both systematic bias and statistical uncertainty , are evaluated using simulated x - ray sources with known properties .
the technical concept of avoiding nonlinear parametric modeling , especially in cases of faint sources , is not new . in _
acis studies , @xcite used the extracted x - ray mean energy and its comparison with the simulated spectra of one - temperature models to infer some stellar properties .
our method differs in the use of more physically realistic spectral models and the use of the median instead of mean energy to be statistically robust against outliers . @xcite
also comment that median energies can give effective spectral estimators in x - ray ccd spectroscopy , but they use the median in conjunction with quartiles which can be statistically less reliable for very faint sources . for example , the 25%-50%-75% energy quartiles are readily estimated for a source with 9 photons ( the energies of the @xmath10 , @xmath11 , and @xmath12 photons ) , but are not readily estimated if 8 or 10 photons are present . this study is closely linked to the _ acis extract _ software package for _ chandra _ acis data analysis which is described by @xcite .
we use _ acis extract _ version 2009 - 01 - 27 in the analysis here .
an initial step requires the definition of the intrinsic spectral shape of the population of x - ray sources under study ; this may be flux - dependent .
this shape may be obtained from previous spectral analysis of brighter sources of the same class , or from astrophysical models . in our application to young stars , we use flux - dependent thermal plasma models . in other applications , such as active galactic nuclei in extragalactic surveys or x - ray binaries in nearby galaxy studies ,
different spectral shapes would be used ( [ other_classes_section ] ) .
we use the coup spectral results as calibrators of acis observations of star populations in more distant star formation regions .
spectral templates derived from bright coup sources will be used to translate apparent photometric properties of weak pms stars into their intrinsic properties . in [ apparent_aefluxes_section ] and [ mede_section ] , we will concentrate on two observable properties of each source : the apparent hard - band x - ray flux , @xmath13 , and the x - ray total - band median energy of background - corrected extracted counts , @xmath14 . here
the hard - band covers the @xmath15 kev energy range and the total - band covers the @xmath16 kev energy range . @xmath14 and a nonparametric estimator of @xmath13
are automatically calculated from the extracted source photon events by the _
acis extract _
package @xcite .
later in the study , we consider total - band fluxes and absorption - corrected fluxes .
we use the coup spectral fits to establish a parametric spectral model family as a basis for estimating broad - band fluxes of pms stars .
@xcite find that two - temperature thermal plasma models provide good fits to high signal - to - noise acis data of @xmath17 coup pms stars .
the temperature of the hot plasma component and emission measure ratio between hot and cool components increase with stellar surface flux , with the temperature of the cool plasma component remaining approximately constant at @xmath18 kev ( @xmath19 mk ) . while it is recognized that the coronae of active stars intrinsically have a distribution of plasma temperatures ( e.g. * ? ? ?
* ) , the two - temperature model is generally adequate for most acis - quality spectra .
coup pms stars follow the general correlation between the temperatures of the hot and cool plasma components seen in main - sequence stars , but with much higher temperatures @xcite .
the 0.8 kev cool plasma component likely describes a persistent coronal structure of pms stars in the form of compact loops with high plasma density , while the hot plasma component reflects the contribution of pms flaring activity ( * ? ? ?
* and references therein ) .
figure [ fig_models ] presents the dependencies of the hot plasma component , @xmath20 , and the ratio of hot to cool component emission measures , @xmath21 , on the absorption - corrected hard - band luminosity , @xmath22 , for coup pms stars .
a similar diagram appears in figure 11 of @xcite .
a few dozen coup super - hot stars with their @xmath20 above 5 kev are omitted @xcite .
the square symbols in the figure and the parameters in table [ tbl1 ] define a set of eleven model families we adopt for the class of pms stars .
these are median values with linear extrapolations at the ends of the @xmath22 range relation @xcite seen in the xmm - newton extended survey of taurus ( xest ; * ? ? ?
* ) , where @xmath23 is the mean electron temperature weighted by the differential emission measure distribution and @xmath24 is the intrinsic x - ray luminosity derived in the @xmath25 kev energy band .
the dashed lines around the median values encompass @xmath26 of coup stars and indicate the uncertainty in choosing an x - ray model among the model families we consider .
this information will be used below to estimate the sensitivity of derived stellar properties on assumed spectral models .
we note that the xest project reports little correlation of plasma temperature with x - ray luminosity in class ii stars ( pms stars with accretion disks ) , and higher plasma temperatures in class ii compared to class iii stars ( pms stars with weak or absent disks ) @xcite .
the coup sample similarly shows a larger fraction of active accretors lying outside of the band of the x - ray model uncertainty ( figure [ fig_models_acc_strata ] ) .
thus the spectral models used here best describe class iii stars , which generally dominate x - ray - selected samples of young rich x - ray stellar populations .
the more general point is that care must be taken to match the spectral model to the underlying population when photometric quantities are examined .
we simulate 11 coup model families from table [ tbl1 ] using the _ fakeit _ command in _ xspec _ @xcite .
each model family is characterized by values of @xmath20 and @xmath21 with @xmath27 kev and is simulated using the wabs @xmath28 ( mekal@xmath29mekal ) _ xspec _ model family with 0.3 solar elemental abundances characteristic of young stars . for each model family
, we generate a large number of high signal - to - noise @xmath2 acis - i spectra over a grid of absorption column densities within the range of @xmath30 @xmath31 .
simulated spectra are then passed through the _ acis extract _
code for photometric analysis including calculation of source net counts , median energies , and apparent fluxes in several energy bands .
the _ acis extract _ ( ae ) package has been used in dozens of @xmath2-acis studies , particularly projects involving analysis of many faint sources @xcite .
it computes several photometric quantities of interest here from the extracted source events . @xmath32 and @xmath33 ( in counts ) are the total - band and hard - band net counts after subtraction of local background . @xmath14 and @xmath34 ( in kev ) are the total - band and hard - band apparent x - ray median energies after background subtraction
. the program estimates photometric photon fluxes ( in photons @xmath31 s@xmath35 ) in the total and hard bands using @xmath36 here @xmath37 is the mean value of a source s auxiliary response file ( arf ) , the effective area averaged over the appropriate band , and @xmath38 is the exposure time at the source location on the detector is designated @xmath39 , while @xmath40 represents another estimator of incident flux where effective areas are measured in narrow energy bands .
the @xmath40-type fluxes are preferred for bright sources but are less stable at low count rates @xcite .
@xmath40-type fluxes are used in the _ chandra _ source catalog where they are called ` aperture source energy fluxes ' ( http://asc.harvard.edu/csc/columns/fluxes.html ) . ] .
we estimate the apparent energy fluxes @xmath41 and @xmath13 ( in ergs @xmath31 s@xmath35 ) as @xmath42 where @xmath43 are in kev and the constant comes from the kev to ergs conversion .
luminosities @xmath44 and @xmath45 are derived from these fluxes by multiplying them by @xmath46 where @xmath47 is the object s distance , in cm . in various analyses below
, we will consider fluxes that have been corrected for soft x - ray absorption scaled to the line - of - sight column density @xmath48 which we derive from @xmath14 .
these intrinsic absorption - corrected flux estimates are labeled @xmath49 and @xmath50 ( in ergs @xmath31 s@xmath35 ) for the total and hard bands , respectively . in oder to assess the fidelity of these flux estimates , figure [ fig_xspec_vs_phot_sim ] compares _ acis extract _
photometric fluxes from equations ( 3)-(4 ) for simulated data with apparent fluxes derived from the flux integration of the input models ( discussed in [ model_section ] ) evaluated on the energy bins set by the _
response matrix .
discrepancies of @xmath51% in flux are seen for typical spectra , and can range up to @xmath52% for extremely soft or hard spectra .
these biases can be explained by inaccuracies in the computation of fluxes using equations ( 1)-(2 ) that are based on effective areas averaged over the broad band assuming a flat incident spectrum .
this approximation is not accurate for realistic spectral shapes . _
acis extract _
@xmath40-type incident fluxes , where effective areas are measured in narrow energy bands , do not exhibit these biases , but they suffer from large poisson errors for weak sources .
for example , a single photon present at energies @xmath53 kev , where the _ chandra _ mirror response is poor , can lead to a large spurious jump in inferred flux .
since our work is oriented towards study of faint sources , we use the more stable @xmath39-type photometric fluxes of equations ( 1)-(2 ) .
the curves in figure [ fig_xspec_vs_phot_sim ] can then be used to reduce the systematic biases in flux extimates for extremely soft and hard spectra ( see also table [ tbl_calibration1 ] ) .
@xmath41 and @xmath13 are not reliable estimators of the intrinsic source flux if absorption is present . for bright sources ,
a typical procedure is to estimate line - of - sight column density @xmath48 via a parametric fit of the apparent photon energy distribution using _ xspec _ or similar code , and calculate an intrinsic unabsorbed flux based on the other spectral model parameters .
we now seek an analogous procedure based on the apparent source median energy , @xmath14 .
figure [ fig_nh_vs_mede_sim ] ( see also table [ tbl_calibration2 ] ) shows the simulated x - ray column density ( @xmath48 ) as a function of the x - ray source s median energy in the total band ( @xmath14 ) .
the results of our simulations confirm the previous coup finding of @xcite that the median energy can be effectively used as a surrogate for absorption column density for @xmath54 kev corresponding to @xmath55 @xmath31 .
the 0.5 kev low energy limit of the acis instrument leads to a degeneracy in absorption values for @xmath56 kev .
the @xmath57 relationship can be used reliably to infer either absorbing column density or source plasma temperature if @xmath48 is known in advance . if both are unknown , the inference has greater error ( [ method_error_section ] ) .
figure [ fig_redcorr_vs_mede_sim ] ( see also table [ tbl_calibration3 ] ) shows the ratio of intrinsic ( absorption - corrected ) flux to apparent flux against apparent median energy in the total band for the simulated spectra . these are flux integrations of the input models ( not re - fitted simulated spectra ) evaluated on the energy bins set by the _
response matrix .
we see that flux ratios for the total band fluxes , @xmath58 , are often factors of @xmath59 , and can reach factors of @xmath60 , for heavily absorbed sources . in the hard band ,
the flux ratios @xmath61 are much smaller , often less than a factor of 2 and always less than a factor of 10 .
@xmath50 is thus a much more stable estimate of source flux than @xmath49 when absorption is present . as described above , the inferred relation between the intrinsic spectral shape and the x - ray luminosity ( [ model_section ] and figures [ fig_models ] and [ fig_models_acc_strata ] ) allowed us to explicitly define and simulate x - ray models of pms stars .
analysis of the simulated data calibrates relationships between spectral models and observable quantities ( [ apparent_aefluxes_section ] and [ mede_section ] , and figures [ fig_xspec_vs_phot_sim ] , [ fig_nh_vs_mede_sim ] , [ fig_redcorr_vs_mede_sim ] ) .
we have shown that the source counts and median energy , which are readily measured for faint sources , often serve as surrogates for source flux and absorption .
similar observable quantities have been used for many years in color - magnitude diagrams for characterizing sources in optical - infrared ( oir ) astronomy .
other classes of sources , different from young stars , may have their own relationships between intrinsic and/or observable quantities that allow simulations of explicit x - ray models ( [ other_classes_section ] ) .
once model simulations are performed , calibration data mapping simulated observable to simulated intrinsic properties can be built . in the case of pms stars we produce four calibration products : tables [ tbl_calibration1 ] ,
[ tbl_calibration2 ] , and [ tbl_calibration3 ] correspond to figures [ fig_xspec_vs_phot_sim]-[fig_redcorr_vs_mede_sim ] discussed above , and table [ tbl_calibration4 ] is used for error analysis ( [ method_error_section ] ) .
once simulated calibration data are built , the measured photometric properties of observed x - ray sources are compared with the calibration data to obtain intrinsic properties and errors on these properties .
the procedure is shown in figures [ fig_flow_chart_part1 ] and [ fig_flow_chart_part2 ] and is presented in detail in this section .
the idl implementation of the procedure can be found at http://www.astro.psu.edu / users / gkosta / xphot/. researchers working with other classes of sources can use the idl code after constructing the four calibration tables similar to tables [ tbl_calibration1]-[tbl_calibration4 ] appropriate for the spectral shapes of their source class . in step 1 of the procedure shown in figure [ fig_flow_chart_part1 ] ,
source photometry properties are obtained by running _
acis extract_. here we provide details on steps 2 - 4 which estimate the source s absorption and intrinsic broad - band fluxes . in step 2 , one of the 11 simulated pms spectral model families listed in table [ tbl1 ]
is assigned to each individual source of interest based on that source s hard - band apparent luminosity and full - band median energy . for lightly obscured stars ( @xmath62 kev ) , the ratio of intrinsic to apparent flux in the hard - band is @xmath63 dex for all spectral models ( figure [ fig_redcorr_vs_mede_sim]b ) , so that apparent hard - band luminosity @xmath45 only slightly underestimates the intrinsic hard - band source luminosity @xmath22 .
the @xmath63 dex luminosity difference is much less than the quantization of 0.5 dex on @xmath22 used to create the coup template models ( figure [ fig_models ] and table [ tbl1 ] ) .
thus , for a lightly obscured pms star , the apparent luminosity @xmath45 can be used directly to assign a specific spectral model family ( `` best model family '' ) from those listed in table [ tbl1 ] . for a heavily obscured pms star ( @xmath64 kev ) , the source s median energy @xmath14 is compared to the calibration of intrinsic to apparent flux ratios ( figure [ fig_redcorr_vs_mede_sim]b and table [ tbl_calibration3 ] ) to choose 11 candidate @xmath65 values corresponding to 11 template model families .
then , the source s apparent hard - band luminosity @xmath45 is used to obtain 11 candidate intrinsic hard - band luminosities , @xmath66 .
the `` best model family '' is the one with the closest match between the @xmath66 and the `` nominal '' intrinsic hard - band luminosity listed in table [ tbl1 ] .
this procedure is analogous to the reddening correction in oir color - magnitude diagrams . in step 3 ,
photometric fluxes are de - biased using calibration curves ( figure [ fig_xspec_vs_phot_sim ] and table [ tbl_calibration1 ] ) for the source s spectral model .
the absorbing column density is then estimated using the source s median energy ( figure [ fig_nh_vs_mede_sim ] and table [ tbl_calibration2 ] ) .
step 4 uses the resulting @xmath48 estimate to obtain absorption - corrected fluxes ( figure [ fig_redcorr_vs_mede_sim ] and table [ tbl_calibration3 ] ) . to visually represent the transformation of apparent to intrinsic stellar properties in our method , figure [ fig_xraycmd ] shows the x - ray `` color - magnitude diagram '' derived from the simulated pms star data . in this diagram , the apparent hard - band luminosity ( @xmath45 ) , apparent total - band median energy ( @xmath14 ) , absorbing column density ( @xmath48 ) , and intrinsic hard - band luminosity ( @xmath22 ) are analogues to oir magnitude , color , extinction , and bolometric luminosity , respectively .
we note again the degeneracy in low values of absorbing column density at @xmath67 kev due to the 0.5 kev limit of the _ chandra_-acis instrument .
we estimate uncertainties in source properties derived using our photometric procedures including both small-@xmath68 statistical errors and systematic uncertainties in our knowledge of the x - ray spectral model .
systematic errors associated with the astronomical source properties , such as distance or x - ray variability , are not considered here .
we also omit systematic errors of instrumental origin .
errors in the arf , which is based on calibration of the _ chandra _ mirrors and acis detector , are anticipated to be small and become comparable to statistical uncertainties only for sources with @xmath69 counts @xcite .
the procedures for the error analysis of photometric quantities described here are outlined in figure [ fig_flow_chart_part2 ] .
first , we estimate statistical errors on the x - ray median energy statistic using simulated data . passing simulated model spectra ( [ model_section ] ) through the marx mirror - detector simulator , we slice the resulting x - ray event lists into thousands of subsamples of different sizes .
distributions of the median energy statistic are then obtained and corresponding standard errors , @xmath70 , encompassing 68% of the values are recorded .
the resulting standard errors for the total - band and hard - band median energies for the 11 pms spectral model families are shown in figure [ fig_mede_errors ] and tabulated in table [ tbl_calibration4 ] .
figure [ fig_mede_errors ] shows that median energy errors range from @xmath71% for sources with 100 counts to @xmath72% for sources with 10 counts .
these values are larger than expected from a standard gaussian distribution , although the errors do scale roughly with @xmath73 . at very low ( @xmath62 kev ) and very high ( @xmath74 kev ) median energy values ,
errors are reduced due to the energy limits of the @xmath2 instrumental response .
the spread in @xmath75 for a fixed source count and @xmath14 value shows that sources with harder intrinsic spectra have larger median energy uncertainties than intrinsically softer sources ) . that low - energy @xmath48 cut - off and
the intrinsic lack of power at high energies give the low - temperature source a narrower observed spectrum , and a correspondingly more accurate estimator for apparent median energy . ] .
again we recall that the results here are specific to the _ chandra _ acis mirror - detector combination and pms stellar spectral models .
researchers working with data from other x - ray telescopes or source classes are advised to produce new error estimates for the median energy statistic .
however , we expect that the qualitative findings will be similar for other systems . to estimate the effects of these median energy uncertainties ( figure [ fig_mede_errors ] or table [ tbl_calibration4 ] ) , we propagate them along with uncertainty on source net counts to errors on apparent fluxes ( equations 1 - 4 ) using standard methods ( e.g. * ? ? ? * ) . this is step 5 of the procedure shown in figure [ fig_flow_chart_part2 ] .
step 6 of the procedure propagates the median energy errors through the relations between median energy and absorbing column density ( figure [ fig_nh_vs_mede_sim ] and table [ tbl_calibration2 ] ) to estimate both statistical and systematic errors in @xmath76 .
this is illustrated in the insert diagram of figure [ fig_nh_vs_mede_sim ] .
statistical errors on @xmath76 are derived by propagating errors on @xmath14 using the curve for the `` best model family '' ( [ translate_section ] ) .
systematic errors on @xmath76 are estimated by propagating the value of the @xmath14 itself to other possible spectral models ; we use the `` second nearest neighbor '' as a plausible range for model uncertainty ( see the dashed lines in figure [ fig_models ] which are discussed in [ model_section ] ) . finally , we estimate statistical and systematic errors on the absorption - corrected hard - band ( @xmath50 ) and full - band ( @xmath49 ) fluxes by propagating errors already obtained for @xmath14 , @xmath13 , and @xmath41 .
this propagation uses the absorption correction calibrations in figure [ fig_redcorr_vs_mede_sim ] and table [ tbl_calibration3 ] . as with @xmath76 , statistical errors on intrinsic source fluxes are estimated using the best spectral model , while systematic errors are estimated using the `` second nearest neighbor '' models .
we now illustrate our analysis methods using the @xmath2 acis - i x - ray data of pms candidate stars in the m17 region .
m17 , one of the brightest hii regions in the sky , is ionized by ob stars in the rich young stellar cluster ngc 6618 which lies on the edge of a massive molecular cloud .
the x - ray stellar population is composed of light to moderately - absorbed pms stars of the central ngc 6618 cluster supplemented by heavily - absorbed stars embedded in or obscured by the surrounding cloud .
a study of the x - ray population from a 40 ks _ chandra _ acis exposure of the region is described by @xcite ; 886 faint x - ray sources were found in this short exposure .
a new @xmath77 ks @xmath2 exposure of this field has been obtained , and the combined dataset yields @xmath0 x - ray sources ( townsley et al . , in preparation ) .
this dataset is chosen for its large sample and wide range of source properties : @xmath78 source counts and @xmath79 kev median energies .
the @xmath2 data have been analyzed following _ acis extract _
procedures described in @xcite .
spectral fits have been performed with _ xspec _ on sources with net counts down to @xmath59 counts ; best - fit models were found by the maximum likelihood method counts . within xspec these moderate - quality backgrounds were modeled with a simple continuous piecewise - linear function with 10 approximately evenly - spaced vertices ( see the discussion of the _ cplinear _ model in the ae manual , http://www.astro.psu.edu/xray/docs/tara/ae_users_guide.pdf ) . in order to explore the range of alternative model fits the tbabs @xmath28 vapec model with thawed parameters was fit to each source five times , using a set of initial parameter values that explore an appropriate region of parameter space .
two additional tbabs @xmath28 vapec models representing `` standard '' stellar spectra ( kt frozen at 0.86 and at 2.6 kev ) were also fit . the best model derived from these seven fits was chosen via a combination of computer algorithm and human review .
the full procedure requires days to be accomplished . ] assuming a one - temperature tbabs @xmath28 vapec model @xcite with the elemental abundances fixed at the values typical for pms or extremely active zero - age main - sequence stars from @xcite .
these are slightly different spectral models to those used in the coup study , but the small change in spectral shape is likely to have negligible effect on the inferred absorption and flux estimates . the m17 data thus give a large sample of pms x - ray sources for which spectral and photometric properties are available .
the absorption and flux estimates derived with the parametric and nonparametric methods can thus be compared .
figure [ fig_nhratio_vs_mede_sim ] compares the nonparametric and parametric estimation of absorption for m17 sources .
panel ( a ) shows the @xmath14 obtained nonparametrically plotted against the column density @xmath76 obtained from _ xspec_. the data accurately follow the relationship shown in figure [ fig_nh_vs_mede_sim ] with increased scatter for fainter sources , as expected .
panel ( b ) shows the ratio of @xmath48 values estimated from @xmath14 as described in [ mede_section ] with those obtained from _
xspec _ fits , as a function of @xmath14 . for the @xmath80 kev range ,
the @xmath76 value inferred nonparametrically is systematically lower than that of the xspec fitting by 0.05 dex .
this is a small effect , and we do not know if it represents a bias in our @xmath57 conversion , a bias in _ xspec _ fitting , or is due to the difference in spectral models used in the simulations and m17 spectral fitting . for sources with @xmath81 kev ,
discrepancies between the nonparametric and parametric absorption estimates show much larger scatter and biases , up to a factor of ten .
neither method can accurately measure absorbing column density for faint soft sources , because @xmath2 acis - i spectra are often insensitive to differences in low column density and there is an ambiguity in spectral model fits of faint soft sources .
we derive statistical and systematic errors on absorption column density following the procedure presented in [ method_error_section ] .
figure [ fig_lognh_errors_vs_me ] shows that @xmath76 values inferred from our method have increasing uncertainties for soft sources with @xmath81 kev .
again , this arises from the weak dependency of apparent median energy on column density for softer sources with light obscuration ( figure [ fig_nh_vs_mede_sim ] ) .
we recommend that tabulations of nonparametric properties of pms stars with @xmath81 kev should report an upper limit of @xmath82 @xmath31 rather than specifying a very uncertain absorbing value . for stars with
@xmath80 kev , figure [ fig_lognh_errors_vs_me]a shows that our method estimates @xmath76 with a statistical accuracy better than @xmath830.3 dex , @xmath830.2 dex , @xmath830.1 dex for the @xmath84 , @xmath85 , @xmath86 count strata , respectively .
panel ( b ) shows that systematic uncertainties due to inadequate knowledge of the precise x - ray spectral model applicable to each source add an additional uncertainty less than @xmath87 dex for all stars , and less than @xmath88 dex for sources with @xmath89 .
these errors are comparable to the scatter between photometric and _ xspec _ estimates of @xmath76 above 22.0 @xmath31 shown in figure [ fig_nhratio_vs_mede_sim]b . in summary , the nonparametric procedure provides reliable estimates of @xmath48 absorption within a factor of @xmath90 for sources with @xmath91 counts , and within a factor of @xmath92 for sources with @xmath59 counts including both statistical and systematic spectral model uncertainties .
figure [ fig_xspec_vs_phot_data]a shows the ratio of the total - band apparent flux obtained using our nonparametric methods in [ apparent_aefluxes_section ] and the flux obtained from _ xspec _ fitting for m17 sources .
the biases shown in figure [ fig_xspec_vs_phot_sim ] have been removed from the nonparametric estimates .
the mean value is @xmath93 indicating that only a small ( @xmath94% ) systematic flux difference between the nonparametric and parametric methods remains , and individual source differences are mostly below @xmath95% . in figure
[ fig_xspec_vs_phot_data]b , similar distributions for hard - band apparent flux ratios are shown , again exhibiting small ( @xmath92% ) bias .
these small systematic differences between nonparametric and _ xspec _ results may become important in the @xmath96 counts regime , and parametric methods are preferred for strong sources .
following procedures outlined in
[ method_error_section ] , we derive statistical errors on apparent fluxes and plot them in figure [ fig_apparentflux_stat_errors_vs_me ] .
panel @xmath97 shows that our method calculates apparent total - band source fluxes with statistical accuracy better than 60% , 50% , 30% , 20% for total - band net count strata of @xmath98 , @xmath51 , @xmath85 , @xmath86 , respectively .
panel @xmath99 shows similar errors for apparent hard - band source fluxes .
these are slightly higher than gaussian @xmath100 errors , and reach @xmath73 levels when @xmath101 counts are present .
figures [ fig_redcorr_simvsdata_vs_mede]-[fig_stat_syst_err_fluxcorr_vs_mede ] are similar to figures [ fig_xspec_vs_phot_data]-[fig_apparentflux_stat_errors_vs_me ] but for broad - band intrinsic fluxes of m17 sources corrected for absorption .
recall that very large corrections have been applied to high-@xmath14 sources ( figure [ fig_redcorr_vs_mede_sim ] ) . here
we see considerably greater bias and scatter compared to the apparent fluxes ; note the expanded vertical scale in figure [ fig_redcorr_simvsdata_vs_mede]b .
the ratio of nonparametric to _ xspec _ intrinsic total - band fluxes is @xmath102 ( @xmath103 ) for the @xmath85 ( @xmath104 ) total - band count strata , and the ratio of hard - band fluxes is @xmath105 ( @xmath106 ) for the @xmath85 ( @xmath104 ) hard - band count strata .
thus , even for high - count rates , the intrinsic photometric total - band ( hard - band ) fluxes are systematically @xmath107% ( @xmath108% ) higher than _ xspec _ fluxes values inferred nonparametrically using two - temperature simulated models are systematically lower than that of the one - temperature xspec fitting ( [ nh_m17_subsection ] ) , it is clear that one - temperature models can not fully recover the soft intrinsic pms x - ray emission subject to absorption .
this is further indicated by a comparison of 2-temperature model fits of the deep coup dataset with 1-temperature model fits of an earlier @xmath109 ks @xmath2 exposure @xcite .
the coup analysis gave @xmath92 times higher fluxes than the earlier analysis .
this result emphasizes the dangers of absorption correction in fluxes that include the soft band . ] . for any broad band flux that includes the soft @xmath110 kev band
, the correction from apparent to intrinsic fluxes can be both very large ( figure [ fig_redcorr_vs_mede_sim]a ) and very uncertain ( figure [ fig_stat_syst_err_fluxcorr_vs_mede]b ) , even for moderate absorption .
weaker sources are particularly vulnerable .
we discourage use of the @xmath49 and @xmath111 quantities derived from photometry for scientific analysis .
although we have not examined it in detail , we suspect that @xmath49 values from parametric _ xspec_-type fitting will also be inaccurate and biased except for very strong sources .
we derive statistical and systematic errors on intrinsic source fluxes following the procedure presented in [ method_error_section ] .
figures [ fig_stat_syst_err_fluxcorr_vs_mede]@xmath97,@xmath99 compare the inferred statistical and systematic errors on total - band intrinsic source flux . here , systematic errors based on x - ray model uncertainty generally exceed statistical errors ; again note the expanded vertical scale in panel @xmath99 . a maximal accuracy of @xmath112 ( or @xmath830.2 dex in log flux ) is achieved only for sources with @xmath86 total - band net counts .
in contrast to total - band absorption - corrected fluxes , the systematic errors of hard - band intrinsic fluxes do not exceed statistical errors for @xmath113 hard - band count sources . here , systematic and statistical errors become comparable for brighter sources .
an accuracy of @xmath114 for hard - band intrinsic fluxes can be achieved for sources with as few as @xmath115 hard - band net counts , unless very high absorptions ( @xmath116 kev ) are present .
we thus find , as expected from qualitative considerations , that hard - band absorption - corrected intrinsic fluxes are much more stable than intrinsic full - band fluxes .
we thus encourage use of @xmath50 when absorption - corrected fluxes and luminosities are sought using nonparametric methods .
recall from [ model_section ] that a bright calibration sample of young stars in the orion nebula indicated a generic spectral model for this class of x - ray sources : a two - temperature thermal plasma with the cooler temperature fixed and the hotter temperature scaled to x - ray luminosity ( figures [ fig_models ] and [ fig_models_acc_strata ] ) .
this property of pms stars allowed us to simulate x - ray models , and calibrate the derivation of astrophysical absorption and fluxes from the observed counts .
other classes of x - ray sources will exhibit different intrinsic spectra shapes with somewhat different dependencies on observable quantities .
spectral model families can be constructed for other classes of x - ray sources , extending nonparametric estimation methods beyond young star populations .
close binary star systems with accreting black holes typically show two spectral components , a thermal disk and a nonthermal powerlaw .
the relative strength of the components scales with the source luminosity and spectral hardness @xcite .
these relationships , based on bright galactic x - ray binary systems , can be applied to x - ray binary systems in nearby spiral ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or elliptical @xcite galaxies .
while the ultraluminous x - ray binaries in these galaxies may be bright enough for detailed parametric spectral modeling , the more typical source is well - adapted to our methods based on broad band counts and median energies .
spectral model families might also be constructed for faint x - ray source populations in the galactic center region @xcite .
the x - ray spectra of faint active galactic nuclei ( agn ) are often parametrized by a power law with index @xmath117 subject to differing amounts of absorption , although more complex emission and absorption components may be present . at low redshift and luminosities , no statistical relationships are seen between @xmath117 , luminosity , and absorption in the x - ray band @xcite .
but at higher redshifts , quasar spectra appear to soften as luminosity increases @xcite .
as there is substantial scatter in @xmath117 at both low and high redshifts , care must be taken to propagate uncertainties in spectral index through our procedures for estimating errors in the inferred broad - band fluxes and absorption .
in this work , we show that important properties of x - ray sources @xmath118 such as line - of - sight absorption , apparent broad - band fluxes , and intrinsic fluxes @xmath118 can be estimated with reasonable accuracy using the easily measured photometric quantities x - ray count rate and median energy .
the translation from observed to intrinsic quantities is achieved without employing often - uncertain and time consuming methods of non - linear parametric spectral modeling .
while parametric modeling is best for strong sources where spectral shape details are well - populated with photons , we believe nonparametric estimates are preferred for faint sources .
the concept of our procedures is similar to long - standing methods based on color - magnitude diagrams in optical and infrared astronomy .
specifically , we first demonstrate that the column density @xmath76 can be estimated directly from the total - band median energy @xmath14 with quantifiable uncertainties . the apparent total - band ( @xmath119 kev ) and hard - band ( @xmath120 kev ) fluxes can be estimated from the source count numbers @xmath32 and @xmath33 , and the absorption - corrected intrinsic fluxes @xmath50 and @xmath49
can then be derived .
the methods are intended for sources with greater than @xmath121 counts but fewer than @xmath122 counts where parametric spectral fitting methods will be superior .
we establish limitations to the method which are probably generally applicable for any reasonable spectral model family .
nonparametric procedures provide poor measures of absorption for sources with @xmath81 kev ( or equivalently , @xmath123 @xmath31 ) .
absorption corrections for total - band fluxes , @xmath49 , are often large and unreliable for sources with @xmath124 total - band counts
. however , absorption corrections for hard band fluxes , @xmath50 , are reliable down to @xmath98 hard - band counts ) .
as discussed in the paper , several net counts in the hard - band are needed in order to derive meaningful intrinsic properties .
this level of hard - band signal may be difficult to detect from an intrinsically soft x - ray source .
researchers interested in non - parametric characterization of intrinsically soft weak x - ray sources might consider modifying the current implementation of the method to use the total - band apparent flux ( instead of the hard - band flux ) as a primary input photometric quantity . ] .
@xmath50 estimation using nonparametric techniques has small biases and uncertainties only somewhat larger than optimal @xmath73 errors .
quantitative results on biases , statistical errors , and systematic errors for each estimator are given in the text and figures .
these detailed results are linked to assumptions of the intrinsic spectral models of the source populations , which we treat here to be absorbed two - temperature thermal plasmas associated with pms stars in star - forming regions .
we calibrated the spectral models to the well - studied coup sample , and successfully applied the methods to a new sample in the m17 star - forming region .
the idl implementation of our nonparametric method , _
xphot.pro_ , is provided .
computing time needed to derive intrinsic properties for @xmath0 x - ray sources is @xmath125 minute .
our method and results can be directly used in statistical x - ray studies of young rich stellar populations ; for example , for constructing column density maps and x - ray luminosity functions of young stellar clusters .
nonparametric results also can be used as initial and/or frozen parameters to use with the parametric _ xspec _ method for fine - tuning of spectral properties .
the procedures described here are developed in two specific contexts , but are more broadly applicable .
first , they use data products generated by the _
acis extract _
data reduction package designed for the _ chandra _ acis detector @xcite .
however , the methods are not restricted to this software package or detector ; the necessary quantities such as median energy and count numbers can be readily obtained using ciao , midas , ftools , or other packages .
the methods could be applied to x - ray ccd observations with the _ asca _ , _ xmm - newton _ , _ suzaku _ , or other space - borne x - ray telescopes if the results are re - calibrated to each mirror - detector combination .
second , the analysis is conducted for specific two - temperature thermal plasma spectral models associated with pms stars , and our quantitative results are restricted to this application .
however , similar analysis , likely resulting in similar results , can be made for other classes of sources , such as active galactic nuclei and x - ray binary star systems in nearby galaxies .
researchers extending these methods to such classes should explicitly define and simulate their own models to construct calibration tables to use with _
we thank michael eracleous ( penn state ) for helpful discussions , and the anonymous referee for helpful comments .
this work is supported by the chandra acis team ( g. garmire , pi ) through the sao grant sv4 - 74018 and nasa astrophysics data program grant nnx09ac74 g ( e. feigelson , pi ) .
broos , p. s. , townsley , l. k. , getman , k. , & bauer , f. e. 2002 , acis extract , an acis point source extraction package ( university park : the pennsylvania state univ . ) http://www.astro.psu.edu/xray/docs/tara/ae_users_guide.html 27.0 & 0.8 & 0.8 & 0.2 + 27.5 & 0.8 & 1.1 & 0.4 + 28.0 & 0.8 & 1.5 & 0.6 + 28.5 & 0.8 & 1.8 & 1.0 + 29.0 & 0.8 & 2.2 & 1.3 + 29.5 & 0.8 & 2.5 & 1.9 + 30.0 & 0.8 & 2.8 & 2.5 + 30.5 & 0.8 & 3.2 & 2.8 + 31.0 & 0.8 & 3.5 & 3.2 + 31.5 & 0.8 & 3.9 & 3.2 + 32.0 & 0.8 & 4.2 & 3.2 + 0.971 & 1.049 & 27.0 & full + 0.986 & 1.072 & 27.0 & full + 1.000 & 1.104 & 27.0 & full + 1.029 & 1.146 & 27.0 & full + 1.059 & 1.187 & 27.0 & full + 1.102 & 1.229 & 27.0 & full + 1.146 & 1.266 & 27.0 & full + 1.190 & 1.298 & 27.0 & full + 1.234 & 1.323 & 27.0 & full + 1.292 & 1.342 & 27.0 & full + 0.971 & 20.000 & 27.0 + 0.986 & 20.658 & 27.0 + 1.000 & 20.977 & 27.0 + 1.029 & 21.301 & 27.0 + 1.059 & 21.477 & 27.0 + 1.102 & 21.602 & 27.0 + 1.146 & 21.699 & 27.0 + 1.190 & 21.778 & 27.0 + 1.234 & 21.845 &
27.0 + 1.292 & 21.903 & 27.0 + 0.971 & 0.01311 & 27.0 & full + 0.986 & 0.06390 & 27.0 & full + 1.000 & 0.11212 & 27.0 & full + 1.029 & 0.23303 & 27.0 & full + 1.059 & 0.33018 & 27.0 & full + 1.102 & 0.41713 & 27.0 & full + 1.146 & 0.49527 & 27.0 & full + 1.190 & 0.56576 & 27.0 & full + 1.234 & 0.62951 & 27.0 & full + 1.292 & 0.68732 & 27.0 & full + 2.431 & 0.2672 & 27.0 & hard & 5 + 2.445 & 0.2716 & 27.0 & hard & 5 + 2.460 & 0.2784 & 27.0 & hard & 5 + 2.475 & 0.2868 & 27.0 & hard & 5 + 2.489 & 0.2923 & 27.0 & hard & 5 + 2.504 & 0.2981 & 27.0 & hard & 5 + 2.518 & 0.3052 & 27.0 & hard & 5 + 2.533 & 0.3131 & 27.0 & hard & 5 + 2.548 & 0.3180 & 27.0 & hard & 5 + 2.562 & 0.3212 &
27.0 & hard & 5 + orion nebula cluster ( onc ) pms stars with available 2-temperature model fits from @xcite . the hot temperature component ( panel a ) and ratio of emission measures between hot and cool components ( panel b )
are plotted against intrinsic hard - band ( @xmath15 kev ) x - ray luminosity .
the dashed curves encompass @xmath126% of the onc stars and are constructed such that at each @xmath127 bin their y - axis value is equal to the y - axis value of the `` second nearest neighbor '' simulated ( square ) model ; that is , the top ( bottom ) dashed curves are versions of the solid curves shifted two @xmath127 bins to the left ( right ) . [ fig_models],width=652 ] but restricted to the coup pms star sub - sample with an available accretion indicator ( e.g. * ? ? ?
coup stars classified as active accretors having the ca ii 8542 @xmath128 line in emission with equivalent width @xmath129 are shown with * @xmath130 * symbols , and stars classified as weakly accreting or non - accreting with absorption equivalent width @xmath131 are shown with @xmath28 symbols .
[ fig_models_acc_strata],width=652 ] kev ) and ( b ) hard band ( @xmath15 kev ) .
results are presented for all 11 simulated spectral model families with the model families for @xmath132 ergs @xmath133 ( @xmath134 ergs @xmath133 ) stars as the top ( bottom ) curves .
[ fig_xspec_vs_phot_sim],width=652 ] .
the top ( blue dashed ) and bottom ( blue solid ) curves correspond to the model families for @xmath135 ergs s@xmath35 and @xmath136 ergs s@xmath35 stars , respectively .
the insert exemplifies estimation of the column density , @xmath137 , and its statistical ( red arrows ) and systematic ( green arrows ) errors . in this example
, the source has 20 counts with median energy @xmath138 kev and median energy uncertainty @xmath139 kev , using the @xmath140 ergs s@xmath35 spectral model family .
[ fig_nh_vs_mede_sim],width=652 ] ergs s@xmath35 and @xmath136 ergs s@xmath35 as the leftmost and rightmost curves , respectively .
model curves are truncated to indicate the locus of @xmath141 onc pms stars from the sensitive coup project .
correction factors are shown for the ( a ) total band ( @xmath16 kev ) and ( b ) hard band ( @xmath15 kev ) .
[ fig_redcorr_vs_mede_sim],width=576 ] kev ) and ( b ) hard band ( @xmath15 kev ) . for each count stratum , shown with a different color , results are shown for the 11 simulated spectral model families from figure [ fig_models ] with @xmath135 ergs s@xmath35 and @xmath136 ergs s@xmath35 as the bottom and top curves , respectively . moving along an individual model family curve from left to right corresponds to increasing the column density parameter of that spectral model family ( see figure [ fig_nh_vs_mede_sim ] ) .
[ fig_mede_errors],width=528 ] m17 sources plotted against x - ray median energy .
two solid curves are calibration lines corresponding to the simulated model families of @xmath142 ergs s@xmath35 ( top ) and @xmath143 ergs s@xmath35 ( bottom ) from figure [ fig_nh_vs_mede_sim ] .
( b ) ratio of x - ray column density obtained from nonparametric estimation to that inferred from _
spectral fits of the m17 data plotted against x - ray median energy .
net count strata are color - coded .
[ fig_nhratio_vs_mede_sim],width=556 ] m17 sources for the total band ( panel a ) , and @xmath144 m17 sources for the hard band ( panel b ) .
net count strata are color - coded .
fluxes are in units of ergs s@xmath35 @xmath31 .
[ fig_xspec_vs_phot_data],width=576 ] m17 sources for the total band flux ( upper panels ) and for @xmath145 m17 sources for the hard band ( lower panels ) .
net count strata are color - coded .
to avoid symbol clutter only @xmath146 and @xmath147 count strata are presented on panels ( b ) and ( d ) .
[ fig_stat_syst_err_fluxcorr_vs_mede],width=576 ] | x - ray sources with very few counts can be identified with low - noise x - ray detectors such as the advanced ccd imaging spectrometer onboard the _ chandra x - ray observatory_. these sources are often too faint for parametric spectral modeling using well - established methods such as fitting with xspec .
we discuss the estimation of apparent and intrinsic broad - band x - ray fluxes and soft x - ray absorption from gas along the line - of - sight to these sources , using nonparametric methods .
apparent flux is estimated from the ratio of the source count rate to the instrumental effective area averaged over the chosen band .
absorption , intrinsic flux , and errors on these quantities are estimated from comparison of source photometric quantities with those of high signal - to - noise spectra that were simulated using spectral models characteristic of the class of astrophysical sources under study .
the concept of this method is similar to the long - standing use of color - magnitude diagrams in optical and infrared astronomy , with x - ray median energy replacing color index and x - ray source counts replacing magnitude .
our nonparametric method is tested against the apparent spectra of @xmath0 faint sources in the _ chandra _ observation of the rich young stellar cluster in the m 17 hii region .
we show that the intrinsic x - ray properties can be determined with little bias and reasonable accuracy using these observable photometric quantities without employing often uncertain and time - consuming methods of non - linear parametric spectral modeling .
our method is calibrated for thermal spectra characteristic of stars in young stellar clusters , but recalibration should be possible for some other classes of faint x - ray sources such as extragalactic active galactic nuclei . |
the development of supermassive black hole accretion theories has always been limited by observations ( pringle 1981 ) .
this has been changing quickly during the past few years , with increasingly higher resolution observations being made over a wide frequency range .
for example , the _ chandra x - ray observatory _ ( _ chandra _ ) and the _ hubble space telescope _ ( _ hst _ ) both have sub - arcsec scale resolutions , while the _ very large baseline interferometer _ ( _ vlbi _ ) can resolve structures on sub - milliarcsec scale . by combining multiwavelength observations from these different instruments ,
we can obtain very strict constraints on the nature of accretion around supermassive black holes . while the innermost structures of agns remain unresolved ( however , cf .
melia & falcke 2001 regarding the galactic center ) , the current data are good enough to constrain the properties of the large - scale accretion disks that are presumably feeding the central engines . a better understanding of these large - scale disks is not only tenable , but will also provide significant implications on the nature of the agns .
most of the relevant theoretical work to date is based on the standard @xmath4-disk prescription ( shakura & sunyaev 1973 ) , in which the accretion disk at large radii is supported by the thermal pressure of the gas ( see , e.g. , hubeny 1990 ; hur et al .
1994 ; collin & hur 1999 )
. this may not be the entire story , however , as some observations suggest that turbulence is prominent in the disks of many radio galaxies ( ferrarese , ford & jaffe 1996 ; ferrarese & ford 1999 ; van langevelde et al .
in addition , the inner disk is irradiated ( maloney , hollenbach & tielens 1996 ) , which should affect , e.g. , its physical properties and line emission characteristics . at a distance of only @xmath5 mpc ,
the fr i radio galaxy ngc 4261 provides us an exceptional opportunity to investigate these new effects . harboring a @xmath6 black hole at its center ( ferrarese et al .
1996 ) , ngc 4261 is well known for its kiloparsec - scale jets ( birkinshaw & davies 1985 ) and its hundred parsec - scale dusty disk imaged by _ hst _ ( jaffe et al .
this disk possesses an atomic surface density of @xmath7 @xmath8 , as inferred from its optical depth assuming a galactic dust / gas ratio . on a smaller scale ,
a thin circumnuclear disk of radius @xmath9 pc was revealed by radio line absorption studies ( jaffe & mcnamara 1994 ; van langevelde et al .
this portion of the disk has an column density of @xmath10 @xmath8 , where @xmath11 k is the spin temperature of the gas . at an even smaller scale , free - free absorption of radio emission from the core by a geometrically - thin disk of radius @xmath12 pc
was detected ( jones et al .
1997 , 2000 , 2001 ; piner , jones & wehrle 2001 ) .
this recent finding adds significant content to the accretion picture because this radius corresponds to only @xmath13 , where the schwarzschild radius @xmath14 cm for the central black hole in ngc 4261 .
although these observations of the accretion disk extend to sub - parsec scales , a fully self - consistent accretion model does not yet exist , particularly for the small free - free absorbing disk ( sudou et al .
jones et al . ( 2000 ) did have success in attributing the observed free - free absorption to a standard @xmath4-disk .
unfortunately , to match the observations , the model requires a dissipative stress that is much greater than that provided by subsonic turbulence . a viable alternative for ( or addition to ) this stress must be found in order for the model to be fully self - consistent .
we suggest here that supersonic turbulence is the responsible agent
. observations of ngc 4261 provide direct evidence that turbulence is prominent in the disk from @xmath15 pc to about @xmath16 pc , with the inner boundary set by observational limitations . at 100 pc ,
the width of narrow lines from the disk suggest a turbulent velocity of @xmath17 km / s ( jaffe et al .
1996 ) , while at 6 pc , the turbulent velocity estimated from the absorption profile is about @xmath18 km / s ( jaffe & mcnamara 1994 ; van langevelde et al .
strong turbulence in the accretion disk is understandable given that the disk is embedded within the core of the galaxy , where it is subject to disruption by strong stellar winds .
the dissipation of turbulence is regulated by the cascade in energy from large scales to small scales , so the total energy associated with the turbulent motion decays roughly as a power - law in time ( lazarian 2003 ) , even in the case of supersonic turbulence ( smith , mac low & zuev 2000a ) .
this slow turbulence decay suggests that the supersonic structure might persist even to smaller radii . in such a supersonic disk ,
the accreting gas is optically - thin and gravitational energy dissipation will overwhelm line cooling at a critical radius @xmath1 , within which the gas is effectively heated to a temperature close to its virial value ( pringle 1981 ) .
thermal pressure of the gas will then dominate the turbulent pressure at @xmath19 , so the supersonic flow appears to terminate at this radius and the gas may make a transition into an advection - dominated accretion flow thereafter ( narayan & yi 1994 ) . on larger scales , the supersonic structure can be inferred directly from observations .
it is interesting to note that with a stress parameter @xmath20 , the value of @xmath1 is roughly @xmath21 pc , which corresponds to the inferred size of the free - free absorbing region .
thus the observed shadowing of the nascent radio jet may be produced by the ionized gas near @xmath1 .
we further note that newly released _
chandra _ data also hint that the nucleus of ngc 4261 may be weakly - active ( peter , zezas & fabbiano 2001 ) , consistent with earlier _ asca _ and _ rosat _ observations ( matsumoto et al . 2001 ; worrall & birkinshaw 1994 ) . the corresponding ionization parameter at @xmath21 pc ( assuming that the radiation is isotropic ) is @xmath22 , where @xmath23 , @xmath24 is the ionizing flux between 1 and @xmath25 ryd , and @xmath26 is the total number density including all species .
note that @xmath27 roughly corresponds to the ionizing - photon - to - gas pressure ratio .
a value of @xmath28 indicates that the gas may be effectively photoionized .
this photoionization will significantly affect the profiles of the lines emitted by the supersonic disk . high - resolution _ hst _ imaging of the central ionized gas has detected broad emission lines from a region smaller than 7 pc in radius ( ferrarese et al .
1996 ) . in this paper , we show that a supersonic disk , irradiated by the agn , can not only account for the observed radio absorption feature , but also reproduce these broad emission lines .
we also show that its large - scale extension is consistent with the radio line absorption observations .
the physical conditions present in the broad line regions ( blrs ) of this class of radio galaxy are likely to be very different than those present in the more - luminous seyfert galaxies and quasars . in both scenarios ,
the line emission is produced by gas having temperature on the order of a few times @xmath2 k , with line broadening caused by bulk motions of the emitting plasma .
however , while the blr in seyferts and quasars is primarily photoionized , our model predicts that heating due to gravitational dissipation plays a significant role in the low - luminosity radio galaxies .
our modeling also indicates a very different blr geometry for these galaxies , with the line emission coming directly from the turbulent accretion disk .
while accretion disks may play an important role in some blr models of seyferts and quasars , the line emission is not believed to originate in the disk itself , instead being associated with , e.g. , magnetohydrodynamic flux tubes ( emmering , blandford , and shlosman 1992 ) or radiation- and gas - pressure driven winds rising from the surface of the disk ( murray et al .
the accretion disk does not even play a direct role in many successful blr models for these objects , e.g. , the bloated star ( alexander & netzer 1994 ) or accretion - shock cooling models ( fromerth & melia 2001 ) .
it is also noted that the density of the blr gas in these galaxies is much lower than in traditional agns . because broad forbidden emission lines are absent in seyfert and quasar spectra , we know that the density of their line - emitting gas is @xmath29
however , the spectra of some low - luminosity radio galaxies contain strong broad forbidden emission lines ( e.g. , [ ] , [ ] ) , indicating gas densities @xmath31 @xmath30 ( see , e.g. , ferrarese & ford 1999 ; barth et al .
the lower density blr gas may be directly tied to their lower mass accretion rates and resulting luminosities .
we first discuss the general properties of a supersonic disk in
[ sec : turbdom ] .
its consistency with the multi - wavelength observations of ngc 4261 is discussed in [ estimate ] .
detailed modeling of the line emission from this supersonic disk , irradiated by its agn , is given in
[ lineemission ] . in
[ discussion ] , we discuss the implication of this work on the nature of agns in radio galaxies .
because it is thought that the large - scale dusty disk feeds the inner accretion of the agn ( jaffe et al .
1996 ) , the detection of a prominent gap in the radio core of ngc 4261 prompted jones et al .
( 2000 , 2001 ) to introduce a free - free absorbing disk to account for the observed features . to produce the radio absorption
, the inner disk must have an emission measure of @xmath32 pc @xmath33 , where @xmath34 is the electron density , @xmath35 is the path length through the absorbing gas , and @xmath36 is the electron temperature in units of @xmath2 k. if dominated by thermal gas pressure , however , the optically thin @xmath4-disk proposed by these authors can not account for this radio shadowing in a self - consistent manner .
in particular , we show below that the implied stress parameter @xmath4 is too large for the flow to be subsonic . for a gas - pressure - dominated accretion disk
, angular momentum conservation requires that @xmath37 at large radii , where @xmath38 is the kinetic viscosity , @xmath4 is the stress parameter ( assumed to be radius - independent and less than one ) , @xmath39 is the radial velocity of the flow , @xmath40 is the scale height of the disk , and @xmath41 is the local sound speed . from the expressions given in jones et al .
( 2000 ) , one can obtain : @xmath42 where @xmath43 is the keplerian velocity , @xmath44 is the black hole mass in units of @xmath45 , @xmath46 gives the accretion rate , and @xmath47 is the distance of the radio absorbing gas from the black hole in units of @xmath48 cm .
in fact , it can be shown that the disk can not be supported by thermal pressure of the gas alone since the corresponding accretion rate is : @xmath49 where quantities with subscript @xmath50 denote the corresponding values inferred from the radio absorption features , and we have approximated the line - of - sight path length through the disk with @xmath51 ( jones et al . 2001 ) , where @xmath52 , @xmath53 is the gas constant and @xmath54 is the molecular weight .
such an accretion rate is so small that the implied emission efficiency for converting accreted rest - mass energy into x - ray emission from the nucleus of ngc 4261 is about @xmath55 ( worrall & birkinshaw 1994 ; matsumoto et al .
2001 ) , which is unacceptable for any @xmath56 in a thermal pressure supported disk .
sudou et al .
( 2000 ) further discussed an optically - thick standard disk and an optically - thin disk cooled by free - free emission .
neither of these alternatives can account for the observed shadowing .
meanwhile , several observations indicate that turbulence dominates the gas motion at large scales ( jaffe et al .
1996 ; van langevelde et al .
2000 ) , where supersonic turbulence may be produced by stellar winds in the core of the galaxy ( jaffe et al .
as the gas falls in toward the black hole , however , the strength of the residual turbulence in the disk is determined by balancing shock dissipation with the turbulence enhancement due to gravitational energy dissipation . in a fully developed turbulent flow , energy
is mostly associated with large - scale gas motion . the decay rate of the total turbulent energy
is then determined by the rate at which energy cascades from large scales to small scales . the turbulent energy density per unit mass , @xmath57 ,
is therefore given by @xmath58 where the dot denotes a derivative with respective to time and @xmath59 gives the characteristic time scale for transporting energy from large to small scales .
the characteristic length scale , @xmath60 , of the system depends on the turbulence generation mechanism .
assuming @xmath60 is independent of time , we have @xmath61 . in their numerical simulations of the decay of supersonic turbulence , smith et al .
( 2000a ) found that the total energy of the turbulence decays roughly with a power law in time , with an index @xmath62 .
we point out that as the large - scale supersonic gas flows inward , the dissipative stress is provided primarily by its turbulent motion .
so the dissipated gravitational energy will go first into producing turbulence .
this turbulence generation mechanism can , in principle , keep the disk supersonic even on smaller scales ( smith , mac low & heitsch 2000b ) .
a complete model addressing the evolution of this supersonic accretion flow , however , would require a detailed numerical simulation beyond the scope of this paper . in the following discussion
, we will assume that the energy associated with the turbulent motion is in sub - equipartition with , and proportional to , the gravitational energy of the gas .
we therefore have @xmath63 where @xmath64 is a `` turbulence parameter '' , which can be fixed by the observed disk properties at large radii , and @xmath65 is the turbulent velocity in the flow .
the supersonic disk model ( in which the stress is dominated by turbulence ) then gives : @xmath66 in eq .
( [ numd ] ) , @xmath67 is the baryon number density averaged in the vertical direction of the accretion disk and @xmath68 is the proton mass . because the h i absorption feature indicates that @xmath69 at @xmath70 pc from the nucleus , we obtain @xmath71 .
the corresponding disk opening angle is @xmath72 . then from equations ( [ height ] ) and ( [ numd ] ) , we have @xmath73 it is interesting to note that the viscous time scale in such a disk ( @xmath74 by definition ) is longer than the corresponding dynamical time scale ( @xmath75 ) by a factor of @xmath76 . adopting the power - law decay rate from the numerical simulations ( smith et al .
2000a ) , one can show that the ratio of the turbulent energy to the dissipated gravitational energy is about @xmath77 .
this is consistent with @xmath78 for @xmath4 less than one .
one should emphasize that only a small fraction ( @xmath79 ) of the dissipated gravitational energy is left to drive the turbulent motion .
most of the dissipated gravitational energy is thermalized via shocks and effectively radiated away . since we are mostly interested in regions where the gas is partially ionized , this cooling is dominated by line emission .
we will see that there is a critical radius @xmath1 in the disk , where the gravitational energy dissipation rate equals the line cooling rate .
the gravitational energy dissipation rate is given by @xmath80 from equation ( [ numd ] ) , we have @xmath81 using this equation to eliminate @xmath82 in equation ( [ heat ] ) , we get @xmath83\ { \rm ergs\ cm^{-3}\ s^{-1}}\ .
\label{eq : addheat}\end{aligned}\ ] ] the line cooling rate of an optically - thin gas is given by @xmath84 \
{ \rm ergs\ cm^{-3}\ s^{-1}},\end{aligned}\ ] ] where @xmath85 is the cooling function ( krolik 1999 ) and @xmath86 . in terms of the critical radius @xmath1 , we have @xmath87 to account for the radio shadowing , the electron number density should be @xmath88 @xmath30 at @xmath12 pc , where we have assumed an inclination angle @xmath89 for the disk ( jones et al .
thus , for a stress parameter @xmath90 , equation ( [ critical ] ) shows that the critical point is located within the radio shadowing region .
it is natural to expect that the observed radio absorption should be produced by the fully ionized gas near @xmath1 , and we will see below that ionizing photons from the central engine play a crucial role in determining the ionization state of the gas at this radius .
_ hst _ observations indicate that the gas closest to the nucleus may be ionized by radiation from the central engine ( ferrarese et al .
further evidence for this can be shown by estimating the ionization parameter at @xmath1 .
recent high - resolution _ chandra _ observations have uncovered a high - energy source at the nucleus of ngc 4261 with a hard band x - ray luminosity of @xmath91 ergs / s ( peter et al .
2001 ) , confirming earlier _ asca _ results ( matsumoto et al .
2001 ) . meanwhile , the soft x - ray luminosity is @xmath92 ergs / s ( worrall & birkinshaw 1994 ) .
so , assuming that the radiation is isotropic and that the gas temperature is @xmath93 k , we can estimate the value of the ionization parameter to be @xmath94 for the radio - absorbing gas at @xmath95 cm ( krolik 1999 ) .
the radio - absorbing gas is therefore effectively photoionized .
although the radiation pressure dominates the thermal pressure of the gas in the radio absorbing region , the turbulent pressure there is @xmath96 ergs @xmath30 , which is much larger than the radiation pressure .
so the disk structure , including the scale height and number density , is still determined by the turbulent motion . however , the gas temperature profile depends on the interaction between the disk and the radiation field .
in the following sections , we will incorporate this photoionization effect using the algorithm cloudy ( ferland 1996 ; cloudy 96-@xmath973 ) and show that the observed broad - line spectrum may also be fitted with this disk model in a self - consistent fashion .
to model line emission from the irradiated disk , we first need to know the intensity and spectral energy distribution ( sed ) of the ionizing radiation from the central source .
these are highly uncertain due to a combination of factors .
first , there are uncertainties associated with intrinsic absorption and galactic contamination which limit the precision to which we can quantify the spectrum in the optical and x - ray bands .
second , the spatial distribution of photons is unknown it is unclear whether the central engine emits isotropically ; the disk at @xmath1 may `` see '' a different spectrum from what we observe .
third , and most importantly , we do not have observations ( or even good estimates ) of the continuum at uv energies .
generally , the shape and intensity of the uv continuum most strongly determine the properties of a photoionized gas .
the correlations between radio and optical spectra have suggested a non - thermal nuclear source for fr i radio galaxies ( chiaberge , capetti & celotti 1999 ) .
further evidence for a power - law nuclear source in ngc 4261 comes from _ rosat _ and _ asca _ observations in the x - ray band ( worrall & birkinshaw 1994 ; matsumoto et al .
2001 ) , which are confirmed by recent high spatial resolution _
chandra _ observations ( rinn & sambruna 2001 ; peter et al .
2001 ) . in the following discussion
, we will adopt a power - law spectrum in the uv - range to bridge the gap between the optical and x - ray data .
the continuum is assumed to be emitted isotropically from a central , point - like source , and the intensity of the ionizing luminosity remains a free parameter in our model to reflect our uncertainties .
we are mostly interested in the region at @xmath12 pc .
however , to determine the spectrum of the incident radiation , we also need to know the gas distribution in the inner region since radiation from the core may be reprocessed by intermediate gases .
fortunately , according to the supersonic disk model , the disk will be heated up very quickly within @xmath1 , which makes the gas there optically - thin to uv and x - ray emission
. one can show this by solving the energy conservation equation in the transition region between @xmath1 and the radius @xmath98 , where the thermal pressure of the gas equals the turbulent pressure and the supersonic accretion flow effectively terminates . for a supersonic accretion disk , we have the energy conservation equation : @xmath99 \ = \
-4\pi\ r\ , h\ , \lambda \ .\ ] ] the first term on the left - hand side corresponds to the gravitational energy dissipation and turbulent advection , while the second term is associated with the thermal energy of the gas .
the right - hand side gives the cooling .
the quantity @xmath100 is about @xmath101 ergs @xmath102 s@xmath103 near the peak ( at @xmath104 k ) of the cooling function ( krolik 1999 ) . assuming that @xmath85 does not depend on radius , we have the following temperature profile in the transition zone : @xmath105 \nonumber \\ & \simeq & t_c\ + \ { 0.15\ ( 1-f)\ g\ , m\over r_c\ r_g}\ \left[{r_c - r\over r}\right]^2 \nonumber \\ &
= & t_c \ + \ 1.2\times 10 ^ 8\ ( 1-f)\ , \left({r_c - r\over r}\right)^2 \nonumber \\ & & \left({r_c\over 10^{18}\ \cm}\right)^{-1 } \left({m\over 4.9\times 10 ^ 8\ \msun}\right)\ { \rm k}\ . \label{temp}\end{aligned}\ ] ] in the second expression above , we have kept only the term proportional to @xmath106 ^ 2 $ ] and have neglected the high order terms in @xmath107 . setting @xmath108 , we obtain the termination radius @xmath98 of the supersonic flow : @xmath109 where @xmath110 so @xmath111 , and @xmath112 k. then the error introduced by neglecting the high order terms in equation ( [ temp ] ) is @xmath113 . in general
, the termination radius @xmath98 is determined by solving the equation @xmath114 so @xmath64 must be smaller than 3/8 to make the thermal pressure of the gas dominate at small radii . a larger value of @xmath64 will cause the turbulent advection to cancel most of the gravitational energy dissipation and keep the disk supersonic .
in addition , a more accurate expression for @xmath85 in the transition zone will make the gas heated even faster because it is on the unstable portion of the cooling curve ( liu , fromerth & melia 2002 ) . within @xmath98 ,
the accretion pattern may transfer into an advection dominated flow ( narayan & yi 1994 ) . within @xmath1 ,
the disk is optically - thin to uv and x - ray radiation , and because of the sharp increase in the gas temperature in this region , free - free absorption is also negligible . due to the effects of photoionization by the agn
, however , the gas will be fully ionized well beyond @xmath1 . to incorporate this effect self - consistently
, we will need to use the algorithm cloudy with additional heating introduced by gravitational dissipation to calculate the temperature and the ionization state of the gas everywhere in the disk . before calculating the line emission from the supersonic disk numerically , it may be illustrative to estimate the physical conditions of the broad - line region
. observations with _
hst _ have identified an ionized gas concentration in a spatially - resolved region with a fwhm of @xmath115 pc .
the broad emission lines are associated with the central @xmath116 aperture position ( farraese et al .
for the h@xmath4 line , the observed flux is @xmath117 ergs @xmath8 s@xmath103 with a fwhm of 2500 km / s . from the electron number density
inferred above and the h@xmath4 volume emissivity of @xmath118 ergs @xmath119 s@xmath103 ( osterbrock 1974 ) , we can estimate the size of the emission region . in our model ,
most of the lines should be produced beyond @xmath120 .
the radius @xmath121 of this region should therefore satisfy the condition @xmath122 implying that @xmath123 pc .
it is interesting to note that the keplerian velocity at such a radius is 1900 km / s , which is consistent with the fwhm of the h@xmath4 line if one takes into account the turbulent motion in the flow and the inclination of the disk . because the contributions to h@xmath4 from the various rings at different radii are comparable to each other between @xmath12 pc and @xmath124 pc , the double peaks that are usually produced by keplerian motion are here filled in , consistent with the observed spectrum .
so a supersonic accretion disk with the properties described above appears to account for the broad emission lines rather naturally .
several observational constraints already limit the parameter space available for fitting the broad emission lines .
images from _ hst _ show that the large - scale dusty disk has an inclination angle of @xmath125 ( jaffe et al .
1993 ) . on the other hand ,
the narrow - line kinematic data from the central region suggests an inclination angle of @xmath126 for the ten - parsec scale disk .
both of these values are consistent with the viewed angle of the radio jet , which is @xmath127 ( piner et al .
we therefore adopt a value of @xmath89 for the inclination angle of the supersonic disk . as stated in section [ sec : turbdom ] , the radio - line absorption observations fix the turbulent parameter at @xmath128 . to account for the radio shadowing near @xmath21 pc
, the electron number density should be around @xmath129 @xmath30 . because we expect hydrogen to be fully ionized in this region , we will choose @xmath130 @xmath30 as the fiducial value for the scale - height - averaged number density of the disk at a radius @xmath131 cm from the supermassive black hole .
the remaining free parameters in the model are the ionizing luminosity @xmath132 , and the stress parameter @xmath4 , which determines the accretion rate and critical radius via equations ( [ mdot ] ) and ( [ critical ] ) .
we use cloudy to calculate the line emissivity as a function of radius for @xmath133 , including the additional heating ( eq . [ eq : addheat ] ) and varying the density with radius ( eq . [ eq : density ] ) as predicted by our model .
the line profiles are then generated by integrating over the appropriate range in radius .
we accomplish this task by partitioning the disk into a large set of concentric tori , modeling the line emission from each torus as a set of @xmath134 discreet clouds , with the understanding that we approach the continuous limit with very large @xmath134 .
the emission line flux from the population of @xmath134 clouds in torus @xmath135 as a function of wavelength is : @xmath136 where the factor @xmath137 is the volume of torus @xmath135 , @xmath138 is the line emissivity including relativistic effects , and @xmath139 is a convolution term relating a cloud s location and velocity to the observed line wavelength . the composite line flux from the entire disk is then obtained by summing the contributions from each torus . taking the line - of - sight velocity to be @xmath140
, the effect of doppler boosting is given by ( @xcite ) : @xmath141 where @xmath142 is the rest - frame emissivity ( generated by cloudy ) and @xmath143 is the line - of - sight velocity including general - relativistic corrections . the observed wavelength @xmath85 of line emission from a given cloud is related to rest - frame line wavelength @xmath144 by @xmath145 where the correction terms on the right are the gravitational , doppler , and cosmological redshifts , respectively .
this relation , appearing as a delta - function @xmath139 in eq .
( [ eq : f_l ] ) , is used to map each cloud s emission into predetermined wavelength bins . the axis of the disk is taken to lie along the @xmath146-axis .
cloud positions and velocities are then determined via monte carlo sampling .
each cloud is given a random azimuthal position @xmath147 in the disk , and the velocity of each cloud is assigned according to @xmath148 where @xmath149 is the keplerian velocity , @xmath150 is the azimuthal direction vector ( note that this depends on the value of @xmath151 ) , and @xmath152 in a randomly - oriented direction .
the line - of - sight velocity is then equal to @xmath153 , where @xmath154 is the direction vector along the line of sight .
we have modeled the continuum - subtracted h@xmath4 @xmath856563 + [ ] @xmath1556548 , 6584 and [ ] @xmath1556717 , 6731 broad emission - line spectrum of ferrarese et al .
the parameters @xmath156 and @xmath4 were varied to determine the best @xmath157 fit to their nuclear pixel data , corresponding to a fwhm spatial resolution of @xmath158 pc .
figure [ fig : best_fit ] shows the resulting best fit ( @xmath159 ; @xmath160 degrees of freedom ) , having parameters @xmath161 erg / s and @xmath162 .
the model nicely reproduces the broad wings and steep core of the h@xmath4 + [ ] composite .
there is an apparent excess in the modeled [ ] emission , which may be remedied by reducing the assumed abundance of sulfur in the model ( we used solar abundances throughout ) .
allowing for this additional free parameter is beyond the scope of this paper , however , as we are more interested in showing a consistency in the line profiles rather than their relative strengths .
figure [ fig : temp_ion ] shows the radial dependence of the gas temperature and the ionization fraction of hydrogen in the best fit model . from this
we see that most of the hydrogen is ionized near @xmath163 pc from the center .
so significant free - free absorption is expected in the region .
thermal equilibrium occurs when the heating and cooling rates balance ; stable equilibria occur only where the slope of the cooling function is positive ( e.g. , shore 1992 ) .
figure [ fig : heat_cool ] illustrates the heating and cooling functions for the gas at radii @xmath164 cm ( top ) and @xmath165 cm ( bottom ) .
the heating rate includes both gravitational dissipation and radiative effects , and the cooling rate must be determined including the dependency on the ionization state of the gas .
cloudy calculates both of these rates in a self - consistent manner . at @xmath164 cm ,
it is clear that only one stable solution exists , @xmath166 k ( compare with fig .
[ fig : temp_ion ] ) , so there is no ambiguity in determining the equilibrium temperature . at large radii , however , it is apparent that two valid physical solutions exist . because our picture has the gas being heated from a low temperature as it accretes , we assign the smaller value in all such cases of uncertainty .
as shown in figure [ fig : temp_ion ] , the gas temperature jumps from @xmath167 k to @xmath93 k near @xmath168 cm when the gas moves from the lower temperature stable branch to the higher temperature one .
this is interesting because neutral atomic hydrogen is concentrated at that radius and this may therefore explain the detected neutral hydrogen radio line absorption .
we have shown that our model can reproduce the observed broad - line spectrum ; we now turn to the question concerning the observed absorption features .
the free - free absorption optical depth at frequency @xmath169 is given by ( walker et al .
2000 ) : @xmath170 \ , \ ] ] where all the quantities are given in cgs units .
the line - center @xmath171 cm optical depth , corrected for stimulated emission , is given by ( @xcite ) : @xmath172 where @xmath173 is the neutral hydrogen column depth and @xmath174 is the doppler line width ( including both turbulent and thermal broadening ) .
note that for @xmath133 , the turbulent velocity of the gas is always much greater than the thermal velocity .
figure [ fig : opt_depth ] shows the radial dependence of both of these optical depths for our best - fit model , under the assumption that the line of sight is perpendicular to the disk .
the free - free absorption is strong enough to account for the observed radio shadowing .
however , the neutral hydrogen absorption seems deficient .
this is not surprising given that we have neglected the density contrast of the supersonic disk in the calculation .
observations of other radio galaxies have revealed that the gas is clumpy at the same spatial scale ( e.g. , harms et al .
1994 ; haschick , baan & peng 1994 ; kartje , knigal & elitzur 1999 ) .
introducing density contrast will not only change the temperature profile of the gas , but also increase the neutral hydrogen column density .
so we would expect significant enhancement of radio line absorption in a detailed modeling that incorporates these additional effects .
figure [ fig : vary_l ] illustrates how the ionizing flux modifies the line spectrum . with an increase in @xmath132 , the heating and ionization parameters of the gas increase as well .
this has the effect of moving the emission zone to larger radii , particularly for the low ionization [ ] and [ ] species . as a result ,
the line profiles become narrower for increasing @xmath132 .
a reduction of @xmath132 from the best fit value does not have such a drastic impact , however .
this is because the inner radius reaches an asymptotic limit in the low luminosity case , determined solely by balancing the gravitational heating and line cooling rates via equation ( [ critical ] ) . in this limit
, the additional heating due to the radiation field is negligible .
it is interesting to note that our best fit lies on the transition to this limit .
the dependence of the line shape on the stress parameter @xmath4 is demonstrated in fig .
[ fig : vary_a ] .
a larger @xmath4 increases the accretion rate ( for a fixed number density at @xmath120 ) , causing additional heating in the gas ( via eq .
[ eq : addheat ] ) .
this moves the emission zone to larger radii , resulting in narrower line profiles .
decreasing @xmath4 has the opposite effect .
we have used the observed radio free - free absorption to fix our value of @xmath175 in the best fit model .
figure [ fig : vary_n ] illustrates the effect of varying @xmath175 .
the density of the gas is important in many aspects , including a determination of the ionization parameter , as well as the gravitational heating and line cooling rates .
it also has a strong impact on the fluxes of the [ sii ] and [ nii ] forbidden lines , which are suppressed relative to h@xmath4 for densities greater than the critical density , at which the radiative and collisional transition rates balance . because the line emission broadening is due mainly to the keplerian motion of the gas , the line width is proportional to @xmath176 , where @xmath177 is the keplerian velocity in the disk .
the projection effect , characterized by the inclination angle @xmath135 , is demonstrated in fig .
[ fig : vary_i ] .
as expected , model spectra associated with small inclination angles have very narrow line profiles . note , however , that even when @xmath178 the lines have non - zero width .
this is the effect of broadening due to the turbulent velocity @xmath179 .
finally , the turbulent parameter @xmath64 has a similar effect on the accretion rate as the stress parameter @xmath4 .
however , by increasing @xmath64 , we not only shift the broad emission line region toward large radii , but also increase the total line flux , as the disk becomes thicker ( thereby increasing the size of the emission region ) .
figure [ fig : vary_f ] shows how the line spectrum changes with @xmath64 .
the accretion flow within @xmath98 is not well constrained .
the fact that the inferred accretion rate from this supersonic disk model gives a radiative efficiency of @xmath180 suggests that advection may dominate in the inner region ( narayan & yi , 1994 ) .
however , if we associate the observed x - ray emission with such an advection dominated accretion flow ( adaf ) at small radii , we find that the observed x - ray spectrum with a spectral index of @xmath181 ( rinn & sambruna 2001 ; matsumoto et al 2001 ) is somewhat softer than that of the model prediction ( narayan et al .
an adaf with strong winds may be able to produce a softer x - ray spectrum to account for the high energy emission from the agn in ngc 4261 ( quataert & narayan 1999 ) . and this outflow may be associated with the jet formation process in the core .
it is also interesting to note that ngc 6251 is quite similar to ngc 4261 in that both of them have a hundred - pc scale dusty disk , large scale symmetric radio jets and an inverted radio spectrum at the core ( melia , liu & fatuzzo 2002 ) .
compared with ngc 4261 , ngc 6251 is about 3 times further away from us , its x - ray luminosity is @xmath5 times stronger , and its broad h@xmath4 line flux is about 100 times bigger .
hst observations suggest that there is a supermassive black hole of @xmath182 , which is very close to the black hole mass in ngc 4261 .
if a similar accretion flow exists in ngc 6251 , equation ( [ line ] ) suggests that the electron number density must increase by a factor of @xmath183 to produce the strong h@xmath4 emission . combining this with the fact that the disk has an inclination angle of @xmath184 ( ferrarese & ford 1999 ) and that the agn is more active
, we would expect that the observed broad emission line may also be fitted with our model .
we also would expect strong radio line absorption in the direction of the counterjet with an h i column density of @xmath185 @xmath8 . in an earlier paper we introduced a hot expanding gas model for the multi - wavelength spectrum of the agn in ngc 6251 ( melia et al .
according to that picture , a standard @xmath4-disk with an accretion rate of @xmath186 g s@xmath103 exists near the black hole .
how such a fossil disk can form after the supersonic disk terminates is still an open question .
however , the implied strong evaporation of the accretion flow below @xmath98 suggests another origin for the hot unbound gas proposed in that model .
further investigation is clearly warranted , and an application of the supersonic disk model to other radio galaxies may also prove to be invaluable in developing a unified theory for supermassive black hole accretion .
this research was partially supported by nasa under grants nag5 - 8239 and nag5 - 9205 , and has made use of nasa s astrophysics data system abstract service .
fm is very grateful to the university of melbourne for its support ( through a miegunyah fellowship ) .
mjf would like to thank gary ferland for his helpful suggestions . | ngc 4261 ( 3c 270 ) is a low - luminosity radio galaxy with two symmetric kiloparsec - scale jets . earlier _ hubble space telescope _
observations indicated the presence of a hundred - parsec scale disk of cool dust and gas surrounding a central , supermassive ( @xmath0 ) black hole .
the recent detection of free - free radio absorption by a small , geometrically - thin disk , combined with earlier studies of the disk s large scale properties , provide the strictest constraints to date on the nature of the accretion process in this system .
we show here that a supersonic disk , illuminated by the active galactic nucleus ( agn ) , can not only account for the observed radio shadowing , but can also produce the optical broad lines emitted from this region . at large radii , the gas is optically - thin because the ram pressure due to turbulence is much larger than the thermal pressure of the gas . at smaller
radii , but beyond a critical radius @xmath1 , line cooling dominates over gravitational dissipation and the gas is effectively cooled down to temperatures below @xmath2 k. within @xmath1 , however , heating due to the release of gravitational energy overwhelms line cooling and the gas falls onto the unstable portion of the cooling curve . because cooling is quite inefficient under these conditions , the plasma is heated very quickly to a temperature close to its virial value as it falls toward the central engine .
thermal pressure of the gas dominates the turbulent ram pressure at a radius @xmath3 , below which the flow probably becomes advection dominated .
the disk is optically - thin to uv and x - ray radiation within @xmath1 , so the ionizing radiation from the agn is preferentially absorbed near @xmath1 , affecting the disk structure significantly . to include the ensuing photoionization effect , we have used the algorithm cloudy with additional heating introduced by gravitational dissipation to calculate the temperature profile and line emission from the disk in a self - consistent manner .
the results of our model calculation are consistent with current multiwavelength observations of the disk in this source .
3gcm^-3 3cm^-3 # 1#1 #1#2#3#1 , a&a , * # 2 * , # 3 .
# 1#2#3#1 , a&a _ ( letters ) _ , * # 2 * , # 3 .
# 1#2#3#1 , a&as , # 2 , # 3 # 1#2#3#1 , aj , # 2 , # 3 # 1#2#3#1 , _ apj _ , * # 2 * , # 3 .
# 1#2#3#1 , _ apj ( letters ) _ , * # 2 * , # 3 .
# 1#2#3#1 , apjs , # 2 , # 3 # 1#2#3#1 , ara&a , # 2 , # 3 # 1#2#3#1 , baas , # 2 , # 3 # 1#2#3#1 , icarus , # 2 , # 3 # 1#2#3#1 , _ mnras _ , * # 2 * , # 3 .
# 1#2#3#1 , _ nature _ , * # 2 * , # 3 .
# 1#2#3#1 , pasj , # 2 , # 3 # 1#2#3#1 , pasp , # 2 , # 3 # 1#2#3#1 , qjras , # 2 , # 3 # 1#2#3#1 , science , # 2 , # 3 # 1#2#3#1 , soviet astr . ,
# 2 , # 3 # 1#2#3#1 , soviet astr .
lett . ,
# 2 , # 3 # 1#2#3#4#1 , # 2 , # 3 , # 4 submitted to the editor of the astrophysical journal 0.5 in |
a black hole , in quantum sense , is not completely black , which can radiate energies continuously .
furthermore , like the black body radiation , it has a temperature @xmath0 , where @xmath1 is the surface gravity of the black hole @xcite .
however , for the thermal radiation spectrum @xcite , the so called information loss paradox " arises , i.e. , if the radiation is thermal , there are no correlations between the emitted field quanta , and then one will lose information about the nature of the matter that originally formed the black hole .
more technically , the complete evaporation of a black hole , whereby a pure quantum state evolves into a thermal state , would violate the quantum mechanical unitarity . in this regard , it is worthy to note that many groups @xcite have attempted addressing this puzzle , but none has been successful . recently , parikh and wilczek @xcite pointed out that hawking radiation is completely non - thermal if one enforces the energy conservation law .
based on this idea , zhang and cai et al .
@xcite proved that for the non - thermal spectrum there are correlations between the radiations , and thus a queue of corrected radiations can transmit encoded information . according to careful calculations of entropy taken out by the emitted particles
, they found that the black hole as tunneling is an entropy conservation process .
thus , they concluded that the black hole evaporation process is unitary .
after their work , a lot of papers @xcite extended this method to other backgrounds , such as quantum corrected black hole @xcite and non - commutated black hole @xcite .
of great interest , all the reexaminations arrived at the same conclusion that information leaks out through the radiation , and the total entropy is conserved .
so , their method may provide a possibility to solve the information loss paradox " .
therefore , it is worthy to check whether the radiation spectrum is non - thermal or not . on the other hand , based on the hamilton - jacobi method beyond semiclassical approximation , r. banerjee and b. r. majhi @xcite computed all quantum corrections in the single particle action revealing that these are proportional to the usual semiclassical contribution .
they gave a quantum modified hawking temperature and entropy .
after their work , many papers appeared to discuss the quantum corrections to temperature and entropy for different backgrounds @xcite . especially , for this quantum corrected case , d. singleton et al .
@xcite used the same method introduced in ref .
@xcite to study the information loss paradox " , and they also found that the total information will be carried away by the correlations of the outgoing radiations when the black hole evaporates completely .
however , some authors @xcite argued that there are no quantum corrections to the hawking temperature and entropy if one chooses the standard definition of the particle energy . and
all the corrections are found to be the contributions to the particle energy .
this choice is reasonable because it keeps the validity of the first law of the thermodynamics .
thus , a dispute , whether the hawking temperature and entropy are modified or not , is caused .
it is worthy to note that zhang et al .
have compared the non - corrected thermal and non - thermal spectra , and they found the thermal spectrum can be distinguished with the non - thermal spectrum from their distinctly different energy covariances @xcite .
however , as introduced above , the puzzles of hawking radiation are not just whether the radiation spectrum is thermal or not , they also contains other cases , such as whether the spectrum is quantum corrected or not , even the quantum corrected spectra , if there are quantum corrections to spectrum , is temperature - corrected or not .
so , it is needed to consider all possible radiation spectra , and give a complete analysis . in this paper , we focus our attentions on these puzzles of hawking radiation introduced above , and try to find out whether the radiation spectrum is thermal or non - thermal with or without high order quantum corrections .
our paper is constructed as follows : in section [ section 1 ] , we simply discuss different radiation spectra coming from taking account of different conditions . in section [ section 2 ]
, we try to analyze the information loss paradox " based on the temperature - corrected and the energy - corrected non - thermal spectra . in section [ section 3 ]
, we compare the average emission energies , average numbers of emissions , and average emission energy fluctuations for the non - corrected thermal spectrum and quantum corrected thermal spectra . in section [ section 4 ] , we compare the average emission energies _ et al .
_ for the temperature - corrected thermal ( non - thermal ) spectrum with that for the energy - corrected thermal ( non - thermal ) spectrum .
in section [ section 5 ] we give an analysis of energy covariances for the thermal spectra and non - thermal spectra .
we finally summarize our conclusions in section [ section 6 ] .
there are six kinds of radiation spectra for black hole , i.e. , non - corrected thermal and non - thermal spectra , temperature - corrected thermal and non - thermal spectra , and energy - corrected thermal and non - thermal spectra .
we now listed them in the following . for the schwarzschild black hole ,
the first obtained spectrum @xcite is @xmath2,\end{aligned}\ ] ] which denotes the possibility of radiating a particle with energy @xmath3 from the black hole with the mass @xmath4 , and it is obviously a thermal spectrum . on the other hand , considering the energy conservation law during the evaporation process , parikh and wilczek @xcite gave a non - thermal spectrum expressed as @xmath5.\end{aligned}\ ] ] obviously , if we omit @xmath6 correction , eq .
( [ non - thermal spectrum ] ) will go back to eq .
( [ thermal spectrum ] ) .
banerjee and majhi @xcite , from the hamilton - jacobi method beyond semiclassical approximation , got an quantum corrected spectrum @xmath7 \\ & = & \exp\left[-\frac{8\pi\varepsilon\mathcal{m}}{\hbar } \big(1-\frac{\alpha\hbar}{\mathcal{m}^{2}}\big)^{-1}\right],\end{aligned}\ ] ] where we have taken @xmath8 as ref .
let s note that this choice is consistent with the result by considering the one loop back reaction effects in the spacetime @xcite .
moreover , the coefficient @xmath9 is related to the trace anomaly .
using conformal field theory techniques , fursaev et al .
@xcite showed that for the schwarzschild black hole @xmath9 is @xmath10 where @xmath11 denotes the number of fields with spin ` @xmath12 ' .
analogously , by using the connection between the tunneling rate ( [ thermal high order quantum corrected spectrum ] ) and the change in entropy given in ref .
@xcite , one can also get the non - thermal spectrum with high order quantum corrections @xmath13.\end{aligned}\ ] ] eqs .
( [ thermal high order quantum corrected spectrum ] ) and ( [ non - thermal high order quantum corrected spectrum ] ) will respectively go back to eqs .
( [ thermal spectrum ] ) and ( [ non - thermal spectrum ] ) when @xmath14 , i.e. , when there are no quantum corrections . in the above discussions , the authors in ref .
@xcite have assumed that the energy of emitted particles is defined by @xmath15 in which @xmath16 is the action of emitted particle without quantum corrections .
therefore , using @xmath17 , they @xcite get a quantum corrected hawking temperature by hamilton - jacobi method beyond semiclassical approximation , which is given by @xmath18 we call this the temperature - corrected case .
it is worthy to note that their method has been discussed and extended widely @xcite .
however , we think that , in these discussions , we should use the standard definition of particle energy in curved spacetime @xcite @xmath19 where @xmath20 is the action of the emitted particle and @xmath21 is a timelike killing vector . by using the hamilton - jacobi method beyond semiclassical approximation
, we find that the particle energy is @xmath22 and then the hawking temperature is @xmath23 obviously , it is the standard hawking temperature .
this implies that , provided we take the standard definition of the particle energy , the hawking temperature is not modified by the quantum tunneling beyond semiclassical approximation . besides , according to @xmath24 , the entropy can be obtained as @xmath25 which is not modified too . because the particle energy @xmath26 in this case is different from @xmath3 ,
we call it the energy - corrected case . for the energy - corrected case
, the possibility of emitted particles for thermal spectrum can be expressed as @xmath27 \\ & = & \exp\left[-\frac{8\pi\mathcal{m}\varepsilon'}{\hbar}\right].\end{aligned}\ ] ] considering the energy conservation law , we can also obtain the possibility of emitted particles for non - thermal spectrum expressed as @xmath28.\end{aligned}\ ] ] compared with the non - corrected thermal and non - thermal spectra ( [ thermal spectrum ] ) and ( [ non - thermal spectrum ] ) , the energy - corrected thermal and non - thermal spectra ( [ energy correction thermal possibility ] ) and ( [ energy correction non - thermal possibility ] ) have the only difference , the radiation energy .
for the former case it is @xmath3 , while for the latter case it is @xmath29 .
above discussions show us that there are several different spectra for the schwarzschild black hole . the non - thermal spectra result from the energy conservation law ,
while quantum corrections come from the hamilton - jacobi method beyond semiclassical approximation ( considering all quantum corrections ) . furthermore , for the quantum corrected cases , different energy definitions of radiation particle can result in different hawking temperature and entropy . in this regard , it is interesting to note that both the temperature and entropy of black hole are not modified by the quantum tunneling beyond semiclassical approximation if we use the standard definition of the particle energy @xmath30 . for the schwarzschild black hole
, so many spectra seem to be possible .
one may ask which spectrum is real . in the following
, we will review how to solve the information loss paradox " based on the temperature - corrected non - thermal spectrum and the energy - corrected one , and then study the features of the six different spectra with the hope of distinguishing them .
an obvious difference between the non - thermal spectra and thermal spectra is that sequential emissions are correlated for the non - thermal spectra , while that for the thermal spectra are not .
what s more , these correlations can reveal where the black hole information goes . through a careful counting of the entropy taken out by the emitted particles , the black hole radiation as tunneling
is proved to be an entropy conservation process .
zhang , cai , you and zhan @xcite have proved that , using the non - corrected non - thermal spectrum , the information is leaking out through the radiations , and the black hole evaporation process is unitary . we will show that , using the temperature - corrected non - thermal spectrum and the energy - corrected one , the information is also leaking out through the radiations , and the black hole evaporation process is still unitary .
these give possible resolutions to the information paradox " under specific conditions .
considering all quantum corrections in the single particle action , together with choosing particle energy @xmath31 , for a black hole with mass @xmath4 one can get a quantum corrected temperature @xmath32 and entropy @xmath33.\end{aligned}\ ] ] as showed in ref .
@xcite , the entropy of the first emission with an energy @xmath34 from a black hole of mass @xmath4 is @xmath35.\end{aligned}\ ] ] the conditional entropy of a second emission with an energy @xmath36 after the @xmath34 emission is @xmath37.\end{aligned}\ ] ] repeating the process , we will find that the entropy of a hawking emission , at an energy @xmath38 , conditional on the earlier emissions labeled by @xmath34 , @xmath36 , ... , and @xmath39 , is of @xmath40.\end{aligned}\ ] ] then , we can calculate the total entropy taken away by the emissions ( @xmath34 , @xmath36 , ... , @xmath41 , ) that exhaust the initial black hole ( @xmath42 ) , which is @xmath43=s'_{\mathrm{bh}}.\end{aligned}\ ] ] eq .
( [ the total entropy ] ) suggests that the entropy of all emitted hawking radiations is equal to the entropy of the black hole , which implies no information is lost in the process of hawking radiation .
however , we should point out that the entropy ( [ the total entropy ] ) is not the entropy of the initial schwarzschild black hole , @xmath44 from eq .
( [ entropy ] ) , we know that the entropy , although by considering all quantum corrections in the single particle action , is not modified if we choose the standard definition of particle energy . for this case , the entropy of a hawking emission , at an energy @xmath45 , conditional on the earlier emissions labeled by @xmath46 , @xmath47 , ... , and @xmath48 , is of @xmath49 thus , after the black hole evaporates completely , the total entropy for a given sequence of emissions ( @xmath46 , @xmath47 , ... , @xmath50 ) with @xmath51 is @xmath52 we can see from eq .
( [ the total entropy1 ] ) that the entropy of all emitted hawking radiation is equal to the entropy of the initial black hole , which implies no information is lost in the process of hawking radiation .
above discussions show us that both the temperature - corrected non - thermal spectrum and the energy - corrected one can provide a possible way to solve the information paradox .
given a queue of emissions , we can calculate the average energies and the covariances of the emitted radiations for the different spectra : thermal and non - thermal with or without quantum corrections ( temperature- or energy - corrected ) .
these are studied below after introducing suitable units for discussing the hawking radiations and their associated properties . to make sure that the leading order correction to entropy is the same with that found earlier in refs .
@xcite by a statistical method , we also take @xmath63 hereafter . in fig .
[ f1 ] we compare the three thermal spectra ( [ three thermal spectra1 ] ) , ( [ three thermal spectra2 ] ) and ( [ three thermal spectra3 ] ) for the schwarzschild black hole at the planck mass scale .
it is shown that there are few difference between them , especial between the temperature - corrected spectrum and the energy - corrected spectrum , they almost stay the same . in units of @xmath60 , we compute the average energy of hawking radiations at any instant . for the thermal spectrum without corrections ,
we have @xmath64 it is easy to find that @xmath65 approaches @xmath66 for @xmath67 and @xmath68 for @xmath69 .
for the temperature - corrected thermal spectrum , we find @xmath70 this quantity also approaches @xmath66 when @xmath67 , and approaches @xmath68 when @xmath69 . for the energy - corrected thermal spectrum ,
the average energy is of @xmath71 it is interesting to note that it is the same with ( [ average energy1 ] ) .
we plot these three average energies of emitted particles in fig .
[ f2 ] as a function of the dimensionless mass @xmath72 .
it is shown that the average energy for the temperature - corrected case is noticeably different from that for the other two cases only near the planck mass scale . for a black hole with mass @xmath72 , the average number of radiations , according to the average energy ( [ average energy1 ] ) , can be obtained for the non - corrected thermal spectrum approximately as @xmath73 we should pointed out that the result given in ref . @xcite may have a typos .
for the temperature - corrected thermal spectrum , we obtain the corresponding average number of radiations according to ( [ average energy2 ] ) , @xmath74 based on ( [ average energy3 ] ) , for the energy - corrected thermal spectrum we have its corresponding average number of radiations @xmath75 it is interesting to note that the non - corrected case and the energy - corrected case give the same average number of radiations , this directly derives from their same average energies of emitted particles ( [ average energy1 ] ) and ( [ average energy3 ] )
. furthermore , eqs .
( [ average number1 ] ) , ( [ average number2 ] ) and ( [ average number3 ] ) give the same limits , @xmath76 for @xmath69 and @xmath77 for @xmath67 . in fig .
[ f3 ] , we show the average number of emissions for the three thermal spectra . above the planck mass scale
, the average number of emissions increases rapidly with the black hole mass .
however , for the small planck mass scale , it remains nearly a constant .
moreover , the average number of emissions for the temperature - corrected case almost has no difference with that of the other two cases .
we now calculate the standard deviations of the emission energies , and we find for the non - corrected thermal spectrum @xmath78 which has two limits , @xmath79 for @xmath67 and @xmath80 for @xmath69 .
analogously , for the temperature - corrected thermal spectrum , we find its corresponding standard deviations of the emission energies @xmath81 whose large and small @xmath72 limits are the same with that of eq .
( [ standard deviation1 ] ) . for the energy - corrected thermal spectrum ,
the standard deviations of the emission energies is of @xmath82 it is worthy to note that the standard deviations ( [ standard deviation3 ] ) is the same with that for the non - corrected thermal spectrum described by eq .
( [ standard deviation1 ] ) .
[ f4 ] shows the standard deviations of the emission energies for the three cases ( [ standard deviation1 ] ) , ( [ standard deviation2 ] ) and ( [ standard deviation3 ] ) .
clearly , these features illustrates the dependence of the variances on the average energy . from our analysis
, we find no obvious differences between the temperature - corrected thermal spectrum and the non - corrected thermal spectrum ( the energy - corrected thermal spectrum ) except for tiny black holes with masses near the planck mass scale .
especially , the three compared quantities for the energy - corrected thermal spectrum are completely the same with that for the non - corrected thermal spectrum .
considering the energy conservation law , the radiation spectra is non - thermal @xcite .
this non - thermal feature shows that the emissions particles is correlated , and it provides a possible way to explain where the information goes during the evaporation process of black hole @xcite , while the thermal spectra can not .
therefore , it is significant to distinguish thermal spectrum and non - thermal spectrum . in the following , we will compare four quantum corrected spectra : the temperature - corrected thermal spectrum , the temperature - corrected non - thermal spectrum , the energy - corrected thermal spectrum and the energy - corrected non - thermal spectrum . by using the dimensionless quantities @xmath72 and @xmath83 , for the temperature - corrected non - thermal spectrum we find @xmath84,\end{aligned}\ ] ] where @xmath85}{\big(\frac{-\alpha+(m - e/8\pi ) } { -\alpha+m^2}\big)^{-4\pi\alpha}}de\right)^{-1}$ ] is the normalization constant . for the energy - corrected non - thermal spectrum , it is of @xmath86 } \exp\left[-e(1-\frac{\alpha}{m^2})^{-1}\left(m-\frac{e}{16\pi}(1-\frac{\alpha}{m^2})^{-1}\right)\right],\end{aligned}\ ] ] where @xmath87 $ ] is dawson function . fig .
[ f5 ] compares the non - thermal spectra ( [ non - thermal spectrum1 ] ) and ( [ non - thermal spectrum2 ] ) with the thermal spectra ( [ three thermal spectra2 ] ) and ( [ three thermal spectra3 ] ) for a black hole at the planck mass scale .
it is shown that the difference between the non - thermal spectra and the thermal spectra concentrated near @xmath88 with the equivalent hawking radiation temperature @xmath89 for a black hole of mass @xmath72 measured in units of @xmath60 . in units of @xmath60 , as for a fixed mass black hole , we calculate the average energy of hawking emissions for the temperature - corrected non - thermal case , which is @xmath90 } { \big(\frac{-\alpha+(m - e/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de.\end{aligned}\ ] ] for the energy - corrected non - thermal case , the corresponding average energy of hawking emissions is @xmath91},\end{aligned}\ ] ] where @xmath92 $ ] is error function .
interestingly , eq .
( [ average energy5 ] ) has the same limits with that of eq .
( [ average energy3 ] ) for @xmath67 and @xmath69 .
we plot eqs .
( [ average energy2 ] ) , ( [ average energy3 ] ) , ( [ average energy4 ] ) and ( [ average energy5 ] ) as a function of dimensionless mass @xmath72 in fig .
[ f6 ] . which illustrates that the distinguishable feature between the non - thermal spectra and the thermal spectra is only near the planck mass scale .
besides , it is also near the planck mass scale where the temperature - corrected spectra and the energy - corrected spectra have the noticeable difference . according to the average energies eqs .
( [ average energy4 ] ) and ( [ average energy5 ] ) , it is easy to obtain the average number of radiations emitted from a black hole with mass @xmath72 .
for the temperature - corrected non - thermal spectrum , it is of @xmath93 } { \big(\frac{-\alpha+(m - e/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de}.\end{aligned}\ ] ] for the energy - corrected non - thermal spectrum , the average number of emitted particles is of @xmath94}{2\pi m\mathrm{er}[2\sqrt{\pi}m]-(e^{4\pi m^2}-1)}.\end{aligned}\ ] ] it approaches @xmath77 and @xmath76 for @xmath67 and @xmath69 , respectively . and these limits are the same with that of eqs .
( [ average number2 ] ) and ( [ average number3 ] ) . in fig .
[ f7 ] , we compare the average number of emissions ( [ average number2 ] ) , ( [ average number3 ] ) , ( [ average number4 ] ) and ( [ average number5 ] ) .
we find that for the small planck mass scale the average number for all the four cases remains nearly a constant . above the planck mass scale
, however , the average number of emissions increases rapidly with the increase of black hole mass .
furthermore , we can seen from fig .
[ f7 ] that average number of emissions for the four thermal spectra has no noticeable difference except for near the planck mass scale .
we now analyse the standard deviations of the emission energies .
for the temperature - corrected non - thermal spectrum , we find @xmath95 } { \big(\frac{-\alpha+(m - e/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de \\ & -&\left(\int^{8\pi m}_0e\frac{\lambda(m)\exp\left[-e(m-\frac{e}{16\pi})\right ] } { \big(\frac{-\alpha+(m - e/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de\right)^2.\end{aligned}\ ] ] and for the energy - corrected non - thermal spectrum , we find @xmath96\big(4me^{4\pi m^2}-\mathrm{er}[2\sqrt{\pi}m]\big)-16\big(e^{4\pi m^2}-1\big)^2}{\mathrm{er}^2[2\sqrt{\pi}m]}.\end{aligned}\ ] ] this quantity has two limits , @xmath97 for @xmath67 and @xmath80 for @xmath69 . fig .
[ f8 ] compares the eqs .
( [ standard deviation2 ] ) , ( [ standard deviation3 ] ) , ( [ standard deviation4 ] ) and ( [ standard deviation5 ] ) . which tells us that the noticeable difference between the thermal and the non - thermal spectra exists only near the planck mass scale .
these features illustrates the dependence of the variance on the average energy . according to our extensive analysis
, we find that there is no drastic difference between the four spectra , the temperature - corrected thermal ( non - thermal ) spectrum , the energy - corrected thermal ( non - thermal ) spectrum , except for tiny black holes with masses near the planck mass scale .
therefore , one would conclude that it is essentially impossible to experimentally distinguish the temperature - corrected thermal ( non - thermal ) spectrum from the energy - corrected thermal ( non - thermal ) one .
and then the puzzles of hawking radiation , whether information can be carried out from a black hole by correlations hidden in the emissions , whether the radiation spectrum has the high order quantum corrections , and the corrections are temperature - corrected or energy - corrected , still perplexes us . nevertheless , we demonstrate below that information stored in the correlations of hawking radiations from the non - thermal spectrum can indeed be observed through a counting of the emission energy covariances
. moreover , the emission energy covariances for the temperature - corrected non - thermal spectrum and the energy - corrected non - thermal spectrum have distinctly different maximums .
it is well known that for the thermal spectrum individual emissions are uncorrelated @xcite , and one thus expects a vanishing covariance . indeed , according to calculations , we obtain @xmath98 when individual emission energies are averaged over an ideal blackbody spectrum .
so , we finally obtain @xmath99 for the non - thermal spectra , we find that the average cross energy therm @xmath100 , due to the existence of correlations between emissions , is nontrivial . and this quantity is strongly correlated with the probability for two emissions , one at an energy @xmath101 and another at an energy @xmath102 . in this regard ,
let s note that the possibility satisfies @xmath103 for an extensive list of black holes as shown in ref .
what s more , a recursive use of this relation allows us to show @xmath104 as long as @xmath105 , or the probability for emissions @xmath106 is the same as the probability for the emission of a single radiation with an energy @xmath107 . obviously , this probability distribution is symmetric with respect to any permutations of the individual emission indices .
thus it allows us to work within one sector and define the normalized probability subjected to the energy conservation constraint @xmath108 $ ] . according to the above analysis , for multiple emissions eqs .
( [ non - thermal spectrum1 ] ) and ( [ non - thermal spectrum2 ] ) thus can be rewritten as @xmath109,\end{aligned}\ ] ] and @xmath110,\end{aligned}\ ] ] which is symmetric with respect to all permutations of indices .
however , unlike eqs .
( [ non - thermal spectrum1 ] ) and ( [ non - thermal spectrum2 ] ) , we must normalize @xmath111 according to @xmath112 .
doing like this , for the energy - corrected non - thermal spectrum we finally give the normalization constant @xmath113)),\end{aligned}\ ] ] for the case of two emissions with energies @xmath114 and @xmath115 , and for the temperature - corrected non - thermal spectrum its normalization constant is of @xmath116 } { \big(\frac{-\alpha+(m-(e_1+e_2)/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de_2de_1\right)^{-1}.\end{aligned}\ ] ] the covariances of successive emissions for the temperature - corrected non - thermal spectrum can be obtained @xmath117 } { \big(\frac{-\alpha+(m-(e_1+e_2)/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de_2de_1 \\ \nonumber & & -\left(\int^{8\pi m}_0e_1\frac{\lambda_(m)\exp\left[-e_1(m-\frac{e_1}{16\pi})\right ] } { \big(\frac{-\alpha+(m - e_1/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de_1\right ) \\ & & \times\left(\int^{8\pi m}_0e_2\frac{\lambda_(m)\exp\left[-e_2(m-\frac{e_2}{16\pi})\right ] } { \big(\frac{-\alpha+(m - e_2/8\pi)}{-\alpha+m^2}\big)^{-4\pi\alpha}}de_2\right).\end{aligned}\ ] ] here , it is needed to point out that we can not get an analytical formula , so we give a numerical integral in the following figure . for the energy - corrected non - thermal spectrum ,
its corresponding energy covariance for two successive emissions is of @xmath118)}{1-e^{4\pi m^2}+2\pi m\mathrm{er}[2\sqrt{\pi}m]}\right ) \\ & & -\frac{\big(4 - 4e^{4\pi m^2}+8\pi m\mathrm{er}[2\sqrt{\pi}m]\big)^2}{\mathrm{er}^2[2\sqrt{\pi}m]},\end{aligned}\ ] ] which has two limits , @xmath119 let s note that eq .
( [ covariance of two emissions ] ) is the same with that for the non - corrected non - thermal spectrum obtained in ref .
@xcite .
we can see from fig .
[ f9 ] that the covariances approaches their maximums also near the planck mass scale , no matter for the temperature - corrected case or the energy - corrected case .
however , they have different maximums , and the maximum covariance for the temperature - corrected non - thermal spectrum is bigger than that of the energy - corrected non - thermal one .
furthermore , it is interesting that both the covariances vanish at small or large masses . as discussed in fig .
[ f7 ] , for both the quantum corrected spectra their average number of emissions become limited ( for instance , two emissions ) when the mass is small .
thus , the covariances vanish at the small mass limit .
which is consistent with the approximate analytical result for the energy - corrected non - thermal spectrum given in eq .
( [ limits1 ] ) when @xmath120 . for large masses ,
the covariances decrease quickly , and finally approach zero for extremely large mass .
this results from the sharp decrease of the average emission energies illustrated in fig .
furthermore , the correlation between the two emissions is proportional to their product . as a result of that
, it is reasonable that the covariances decrease at the large mass limit .
we investigate the long - standing information loss paradox " and the features of the non - corrected thermal ( non - thermal ) spectrum and the temperature - corrected ( energy - corrected ) thermal ( non - thermal ) spectrum .
our analysis show that the largest covariances of successive emissions appear near the planck mass scale , i.e. , for the temperature - corrected non - thermal spectrum @xmath121 for @xmath122 , and for the energy - corrected non - thermal spectrum @xmath123 for @xmath124 dimensional schwarzschild black hole , the fundamental planck scale is reduced depending on the compact space of volume @xmath125 , e.g. , the reduced planck scale @xmath126 with @xmath127 and @xmath128 @xcite ] .
here we give an open problem that the possibility of different radiation spectra discussed above may also exist for other black hole systems , such as the micro black holes model discussed extensively in connection with the experiments of the cern large hadron collider ( lhc ) @xcite . according to research @xcite ,
it is estimated that the minimum black hole mass should be in the range of @xmath129 for @xmath130 collisions at a center - of - mass energy of @xmath131 at lhc .
based on it , a more recent study @xcite showed that the limits on the minimum semiclassical black hole and string - ball masses in the range @xmath132 to @xmath133 for a wide range of model parameters .
so , if the radiation of a micro black hole were observed , then it may be possible to use the energy covariances as a indicator to determine whether the emission spectrum is non - thermal or not , and whether the emission spectrum is temperature - corrected or not .
also note that the energy scale about the production and observation of micro black holes is being debated @xcite .
thus , when and whether the micro black holes could be observed on earth , especially in a lhc experiment , needs a lot of work .
on the other hand , other kinds of manmade black holes , such as those implemented or discussed with optical , acoustic , and cold - atomic systems @xcite , are being discussed , and several experiments @xcite had shown evidence of hawking radiation from the event horizon
. therefore , the same problems we discussed above are also worthy of being studied in these real radiation systems .
we have discussed several radiation spectra of schwarzschild black hole , the non - corrected thermal ( non - thermal ) spectrum , the temperature - corrected thermal ( non - thermal ) spectrum , and the energy - corrected thermal ( non - thermal ) spectrum .
we emphasize that the non - thermal property of radiation spectra comes from taking account of the energy conservation during the evaporation process , and the quantum corrections results from the hamilton - jacobi method beyond semiclassical approximation .
furthermore , we use the temperature - corrected non - thermal spectrum and the energy - corrected one to solve the long - standing information loss paradox " of black hole .
we find that the entropy of the emitted particles , which exhaust the initial black hole , is identical to the original entropy of the black hole , thus entropy is conserved between the initial ( black hole plus no radiation ) and final ( no black hole plus radiation field quanta ) states . which reveals that no information is lost , and the black hole evaporation process in unitary .
these discussions , therefore , may provide a possible way to understand the information loss paradox " . to distinguish different radiation spectra , their corresponding average emission energies ,
average numbers of emissions and average emission energy fluctuations are compared .
it is found that there are no obvious differences between them except for near the planck mass scale .
especially , the energy - corrected spectra have the same corresponding average emission energies , average numbers of emissions and average emission energy fluctuations with that for the non - corrected spectra . of great interest , we find that for all the thermal spectra the energy covariances of hawking radiations completely vanish , while they are nontrivial for all the non - thermal spectra .
especially , the temperature - corrected non - thermal spectrum and the energy - corrected one have distinctly different maximums of energy covariances . as a result of that , these differences provide a way towards experimentally studying the long - standing puzzles of hawking radiation that whether the radiation spectrum of black hole is thermal or non - thermal with or without high order quantum corrections .
this work was supported by the national natural science foundation of china under grant nos .
11175065 , 10935013 ; the national basic research of china under grant no .
2010cb833004 ; the hunan provincial natural science foundation of china under grant no .
11jj7001 ; hunan provincial innovation foundation for postgraduate under grant no cx2012b202 ; the construct program of the national key discipline . | = 0.6 cm * abstract * we investigate the features of the non - corrected thermal ( non - thermal ) spectrum and the quantum corrected thermal ( non - thermal ) spectrum .
we find that : ( i ) using the quantum corrected non - thermal spectra , the black hole radiation as tunneling is an entropy conservation process , and thus black hole evaporation process is unitary ; ( ii ) there are no obvious differences between all spectra except for near the planck mass scale by comparing their average emission energies , average numbers of emissions and average emission energy fluctuations ; ( iii ) the energy covariances of hawking radiations for all the thermal spectra are exactly zero , while they are nontrivial for all the non - thermal spectra .
especially , there are distinctly different maximums of energy covariances for the temperature - corrected and energy - corrected non - thermal spectra .
consequently , these differences provide a possible way towards experimentally analyzing whether the radiation spectrum of black hole is thermal or non - thermal with or without high order quantum corrections .
= 0.8 cm |
several collaborations have reported the observation of the new baryon @xmath0 with positive strangeness and minimal quark contents @xmath8 @xcite .
the existence of such an exotic state with narrow width @xmath9 and @xmath10 was first predicted in the chiral quark soliton model , where the @xmath0 is a member of the baryon antidecuplet @xmath11 @xcite .
the discovery has opened a new field of strong interactions and provides a new opportunity for a deeper understanding of the low energy qcd .
intense theoretical investigations have been motivated to clarify the quantum numbers and to understand the under - structures of the pentaquark state @xmath0 @xcite .
the zero of the third component of isospin @xmath12 and the absence of isospin partners suggest that the baryon @xmath0 is an isosinglet , while the spin and parity have not been experimentally determined yet and no consensus has ever been reached on the theoretical side .
the extremely narrow width below @xmath13 puts forward a serious challenge to all theoretical models , in the conventional uncorrelated quark models the expected width is of the order of several hundred @xmath14 , since the strong decay @xmath15 is okubo - zweig - iizuka ( ozi ) super - allowed . in this article
, we take the point of view that the quantum numbers of the pentaquark state @xmath16 are @xmath17 , @xmath18 , @xmath19 , and study its decay width within the framework of the qcd sum rules approach @xcite .
the article is arranged as follows : we derive the qcd sum rules for the strong coupling constant of the pentaquark state @xmath0 @xmath20 in section ii ; in section iii , numerical results ; section iv is reserved for conclusion .
in the following , we write down the three - point correlation function @xcite , @xmath21 where the @xmath22 , @xmath23 and @xmath24 are the interpolating currents for the neutron , k meson and pentaquark state @xmath0 respectively , @xmath25 for the pentaquark state @xmath0 , we use the diquark - triquark type interpolating current which can give satisfactory mass and stable magnetic moment @xcite . the pseudoscalar mesons @xmath26 and @xmath27 can be taken as both goldstone bosons and quark - antiquark bound states
, we can use the partial conservation of axial current ( pcac ) in constructing the interpolating currents , @xmath28 , @xmath29 if we take the @xmath30 as the interpolating current , more care has to be taken about the possible contaminations from the axial - vector mesons , furthermore , the calculation will be more tedious ( with @xmath28 ) .
the matrix element of the pseudoscalar current between the vacuum and @xmath27 state can be taken as @xmath31 the values of the @xmath32 depend on the masses of the @xmath33 and @xmath34 quarks which have uncertainties , our numerical results indicate small variations of those masses will not lead to large changes about the values of the coupling constant @xmath1 . for the neutron ,
we take the ioffe current @xcite .
the fierz re - ordering of the interpolating currents @xmath35 and @xmath22 can lead to the following sub - structures , @xmath36 a naive result of the fierz re - ordering may be the appearance of the reducible contributions with the sub - structure of @xmath37 ( i.e. @xmath38 ) clusters in the two - point correlation function @xcite , @xmath39 however , in our calculations with the interpolating current @xmath35 in eq.(2 ) , no such factorable @xmath37 terms appear , so there are no reducible @xmath38 contributions to the correlation function @xmath40 .
the re - ordering in the dirac spin space is always accompanied with the color re - arrangement , which involves the underlying dynamics . if we want to factorize out some @xmath38 contributions from the @xmath40 ( with the current @xmath35 in eq.(2 ) ) , tedious manipulations about the re - ordering in the color and dirac spin space must be done .
there are no direct @xmath38 ( or @xmath37 ) components in the interpolating current @xmath35 , which can readily decay to the @xmath41 final state and result in large width .
if the @xmath0 is really a pentaquark state not a @xmath38 molecule , as the @xmath0 lies above the @xmath41 threshold and no need for additional quark - antiquark pairs creation in decay , the decay must be ozi super - allowed and the width is supposed be large , say , about several hundred @xmath14 ; to produce the narrow width , some huge energy barriers are needed to stabilize the @xmath0 in case the kinematical interpretation can not work here .
the appearance of the @xmath38 component in the fierz re - ordering maybe manifest the possibility ( not the probability ) of the evolution from the @xmath0 to the @xmath41 final state without net quark - antiquark pairs creation ( maybe the quark - antiquark pairs created and annihilated subsequently ) , which is significantly in contrast to the conventional baryons , however , we have no knowledge about the detailed process of the evolution .
the @xmath0 may evolve to the @xmath38 final state , and the @xmath38 final state is not presented in the components of the initial pentaquark state @xmath0 , how to implement the evolution with small probability involves complex quark - gluon interactions , whether just re - arrangement in the color space , or creation and annihilation of quark - antiquark pairs . in additive constituent quark models , whether or not additional relative p wave is introduced to changed the ground states from negative parity to positivity parity , special configurations are needed to take into account the narrow decay width as results of small overlaps of the internal and external @xmath38 wave - functions @xcite , the kinematical interpretations based on the color - flavor - spin ( i.e. @xmath42 ) group theory resort to the possible small overlaps , how to realized the small overlaps needs complex re - ordering in the color - flavor - spin space , if there are really no energy barriers to prevent the re - arrangement . while in the cluster quark models , typically , the diquark - diquark - antiquark model @xcite and diquark - triquark model @xcite , extra barriers , for example , relative p waves , are introduced dynamically to prevent the ready decay .
in fact , the re - ordering in the color and dirac spin space involves complex strong interactions , and we know little now about the dynamics which determine the under - structures of the exotic pentaquark states .
the mismatches between the color - flavor - spin states in the initial pentaquark and final baryon - meson color singlet can result in suppression of the decay naturally @xcite . due to the spontaneous breaking of the chiral symmetry and the goldstone nature of the pseudoscalar mesons @xmath26 , @xmath27 and @xmath43 , the quarks may have direct interactions with the pseudoscalar mesons , which lead to the success of some chiral quark models .
the dominating interactions which determine the exotic pentaquark states be color - spin type @xmath44 or flavor - spin type @xmath45 are still in hot debates @xcite , the naive fierz re - ordering can lead to direct @xmath38 component in the @xmath0 will not work here .
if there are really some @xmath38 components in the interpolating current @xmath46 , they should be factorized out , the remainder can not have the correct quantum numbers to interpolate the @xmath0 . in the qcd sum rules
, we construct the interpolating currents with the same quantum number as the corresponding mesons and baryons , that is enough ; the knowledge about the structures of the hadrons can be of much help in the constructing . in ref.@xcite
, the narrow decay width is attributed to the minor breaking of chirality conservation , the color re - arrangement due to the hard gluon - exchange can result in strong suppression of the decay , @xmath47 . in the article , we take quantitative analysis of the decay width within the framework of the qcd sum rules approach .
the diquark - triquark type interpolating current @xmath48 is more likely related to a negative parity pentaquark state , in this work , we make assumption that the parity of the @xmath0 to be negative and study the decay width with the following lagrangian density , @xmath49 according to the basic assumption of current - hadron duality in the qcd sum rules approach @xcite , we insert a complete series of intermediate states satisfying the unitarity principle with the same quantum numbers as the interpolating currents @xmath22 , @xmath23 and @xmath50 into the correlation function in eq.(1 ) to obtain the hadronic representation . after isolating the double - pole and single - pole terms of the lowest ground states , we get the following result @xcite , @xmath51+\cdots \right\ } \nonumber\\ & & \left \{\sigma^{\mu\nu } q_{\mu}p_{\nu } \,\ , + \cdots\right\}+\cdots\ , , \nonumber \\ & = & \left\ { \frac { g_{\theta nk } \lambda_{\theta}\lambda_{n}\lambda_{k } } { m_{\theta}^2- p'^2}\int_0^{\infty}ds \int_0^{\infty}dt \frac{\delta(s - m_{n}^2)\delta(t - m_{k}^2 ) } { ( s- p^2)(t- q^2 ) } + \right.\nonumber\\ & & \left .
\left[\int_0^\infty ds \frac{a(p'^2,q^2)\delta(s - m_{n}^2 ) } { s - p^2}+ \int_0^\infty dt \frac{b(p'^2,p^2)\delta(t - m_{k}^2)}{t- q^2 } \right]+\cdots \right\ } \nonumber\\ & & \left \{\sigma^{\mu\nu } q_{\mu}p_{\nu } \,\ , + \cdots\right\}+\cdots\ , , \end{aligned}\ ] ] with @xmath52 where the following definitions have been used , @xmath53 the coupling constants @xmath54 and @xmath55 can be determined from the two - point qcd sum rules , for the @xmath55 , we use the correlation function @xmath40 , substitute the @xmath22 for the @xmath46 in eq.(8 ) , we can obtain the @xmath54 . in this article
, we choose the dirac tensor structure @xmath56 for analysis , while in ref.@xcite , the authors take the structure @xmath57 . here
the residues of the single - pole terms @xmath58 and @xmath59 have complex dependence on the transitions between the ground states and high resonances ( or continuum states ) .
we have no knowledge about the transitions , even the existence of the @xmath0 is not firmly established .
however , the contributions from the pole - continuum transitions are not exponentially suppressed compared with the double - pole terms , even after double borel transform , furthermore , the contributions can be as large as or larger than the double - pole terms and must be explicitly included in the sum rules .
we only have the fact that the @xmath0 lies a little above the @xmath41 threshold , the contributions from the @xmath0 can be factorized out , so the spectral densities @xmath60 and @xmath61 can be parameterized as @xmath62 the two unknown functions @xmath63 and @xmath64 have complex dependence on the transitions between the ground states and high resonances ( or continuum states ) . from the eqs.(10 - 14 ) , we can obtain @xmath65+\cdots \right\ } \nonumber\\ & & \left \{\sigma^{\mu\nu } q_{\mu}p_{\nu } \,\ , + \cdots\right\}+\cdots\ , , \\ & = & \left\ { \frac { g_{\theta nk } \lambda_{\theta}\lambda_{n}\lambda_{k } } { m_{\theta}^2- p'^2 } \frac{1}{(m_{n}^2- p^2)(m_{k}^2- q^2 ) } + \frac { 1 } { m_{\theta}^2-p'^2}\left[\frac{cc } { m_{n}^2-p^2}+ \frac{dd } { m_{k}^2-q^2 } \right]+\cdots \right\ } \nonumber\\ & & \left \{\sigma^{\mu\nu } q_{\mu}p_{\nu } \,\ , + \cdots\right\}+\cdots\ , , \end{aligned}\ ] ] here we introduce two constants @xmath66 and @xmath67 for convenience , @xmath68 taking the @xmath66 and @xmath67 as some unknown constants has smeared the complex dependence on the energy and high resonance masses ( or continuum states ) , which will certainly impair the prediction power .
we have no knowledge about the transitions between the pentaquark state @xmath0 and the excited states ( or high resonances , or continuum states ) , the @xmath66 and @xmath67 can be taken as free parameters , we choose the suitable values for the @xmath66 and @xmath67 to eliminate the contaminations from the single - pole terms to obtain the reliable sum rules .
the contributions from the single - pole terms may as large as or larger than the double - pole term , in practical manipulations , the @xmath66 and @xmath67 can be fitted to give stable sum rules with respect to variations of the borel parameter @xmath69 in a suitable interval .
if the final numerical results are insensitive to the threshold parameters @xmath70 , @xmath71 and there really exists a platform with the variations of the borel parameters @xmath72 and @xmath73 , the predictions make sense .
the calculation of operator product expansion in the deep euclidean space - time region is straightforward and tedious , here technical details are neglected for simplicity , once the analytical results are obtained , then we can express the correlation functions at the level of quark - gluon degrees of freedom into the following form through dispersion relation , @xmath74}{2 ^ 9 4 ! \pi^6 q^2}\int_0^{s_0}ds \frac{s^2}{s - p^2}-\frac{11\left [ \langle \bar{q}q\rangle^3+\langle\bar{q}q\rangle^2\langle\bar{s}s\rangle \right ] } { 2 ^ 4 3 ^ 2 \pi^4p^2q^2}\right\ } \sigma^{\mu\nu}p_{\mu}q_\nu+\cdots\ , , \end{aligned}\ ] ] we choose @xmath75 and @xmath76 , then take double borel transform with respect to the variables @xmath77 and @xmath78 respectively , match eq.(16 ) with eq.(19 ) , finally we obtain the sum rules for the strong coupling constant @xmath1 , @xmath79m_1 ^ 6e_2(s)}{2 ^ 8 4 ! \pi^6 } -\frac{11\left
[ \langle \bar{q}q\rangle^3+\langle\bar{q}q\rangle^2\langle\bar{s}s\rangle \right]}{2 ^ 4 3 ^2\pi^4 } \right\}\ , , \end{aligned}\ ] ] where @xmath80 here the @xmath81 ( proportional to the @xmath67 , as the @xmath66 terms are eliminated ) denotes the contributions from the transitions between the ground and excited states ( or high resonances , or continuum states ) , we can choose the suitable values for @xmath81 to eliminate the contaminations to obtain the stable sum rules with the variations of the borel parameters @xmath82 and @xmath73 .
the parameters for the condensates are chosen to be the standard values @xcite , @xmath83 , @xmath84 , @xmath85 and @xmath86 .
small variations of those condensates will not lead to large changes about the numerical values .
the coupling constants are taken as @xmath87 @xcite and @xmath88 @xcite from the two - point qcd sum rules , see eq.(8 ) for example .
the threshold parameters @xmath70 and @xmath71 are chosen to vary between @xmath89 and @xmath90 respectively to avoid possible contaminations from high resonances and continuum states .
the borel parameters are taken as @xmath91 and @xmath92 to obtain the stable sum rules .
finally we obtain the values for the coupling constant @xmath93 , @xmath94 @xmath95 \sqrt{\lambda(m_{\theta}^2,m_{n}^2,m_k^2)}\ , , \nonumber\\ & < & 4mev , \\ \lambda(m_{\theta}^2,m_{n}^2,m_k^2)&= & ( m_{\theta}^2 + m_{n}^2 -m_k^2)^2 - 4m_{\theta}^2m_{n}^2\ , . \nonumber\end{aligned}\ ] ] which can explain the narrow width @xmath4 naturally .
the values of the coupling constant @xmath1 with the variations of the threshold parameters ( @xmath70 , @xmath71 ) and borel parameters ( @xmath82 , @xmath73 ) are shown in fig.1 and fig.2 for @xmath96 , @xmath97 , @xmath98 .
with @xmath82 for @xmath99 , @xmath100 . ] with @xmath73 for @xmath101 , @xmath102
in this article , we take the point of view that the pentaquark state @xmath0 has negative parity , and choose the diquark - triquark type interpolating current to calculate the strong coupling constant @xmath1 within the framework of the qcd sum rules approach .
our numerical results indicate that values of the strong coupling constant @xmath1 are very small , @xmath103 , and the width @xmath104 , which can explain the narrow width @xmath4 naturally .
this work is supported by national natural science foundation , grant number 10405009 , and key program foundation of ncepu .
the authors are indebted to dr .
j.he ( ihep ) , dr .
x.b.huang ( pku ) and dr .
l.li ( gscas ) for numerous help , without them , the work would not be finished .
m. a. shifman , a. i. vainshtein and v. i. zakharov , nucl . phys .
* b147 * ( 1979 ) 385 , 448 ; l. j. reinders , h. rubinstein and s. yazaki , phys
* 127 * ( 1985 ) 1 ; s. narison , world sci .
lect.notes phys .
* 26 * ( 1989 ) 1 . c. e. carlson , c. d. carone , h. j. kwee , v. nazaryan , phys
b573 * ( 2003 ) 101 ; f. stancu , d. o. riska , phys . lett . * b575 * ( 2003 ) 242 ; c. e. carlson , c. d. carone , h. j. kwee , v. nazaryan , phys . lett . * b579 * ( 2004 ) 52 ; r. bijker , m. m. giannini , e. santopinto , eur .
j. * a22 * ( 2004 ) 319 . | in this article , we take the point of view that the pentaquark state @xmath0 has negative parity , and choose the diquark - triquark type interpolating current to calculate the strong coupling constant @xmath1 in the qcd sum rules approach .
our numerical results indicate the values of the strong coupling constant @xmath1 are very small , @xmath2 , and the width @xmath3 , which can explain the narrow width @xmath4 naturally . + zhi - gang wang@xmath5 , wei - min yang@xmath6 and shao - long wan@xmath7 + @xmath5 department of physics , north china electric power university , baoding 071003 , p. r. china + @xmath6 department of modern physics , university of science and technology of china , hefei 230026 , p. r.
china + pacs : 12.38.aw , 12.38.lg , 12.39.ba , 12.39.-x
* key words : * qcd sum rules , decay width , pentaquark |
we propose a new model , for seyfert galaxies , involving a point source of relativistic leptons located above the disk ( that could be physically realized by a strong shock terminating an aborted jet ) emitting hard radiation by inverse compton ( ic ) process on soft photons produced by the accretion disk .
the disk itself radiates only through the re - processing of the hard radiation impinging on it , i.e. we do not suppose any internal energy dissipation ( cf fig .
[ main ] ) .
such a geometry is highly anisotropic , which takes a real importance in the computation of ic process ( @xcite , @xcite ) .
we treat both newtonian and general relativistic cases @xcite , deriving a self - consistent solution in the newtonian case .
we have recently added monte - carlo calculations to take into account , in the high energy part of our synthetic spectra , the compton reflection component and the fluorescent iron line @xcite .
we present here the most important results supplied by the model .
it appears that the anisotropy of the soft photon field at the hot source level , leads to an * anisotropic inverse compton process * , with much more radiation being scattered backward than forward .
such an anisotropic re - illumination could naturally explain the apparent x - ray luminosity , usually much lower than the optical - uv continuum emitted in the blue bump @xcite .
it can also explain the equivalent width observed for the iron line , which requires more impinging radiation than what is actually observed @xcite .
we plot in figure [ anisotropytemperature ] the angular distribution of the power emitted by the hot source in newtonian metrics and for different values of the source height in kerr metrics .
it appears that the closer the source to the black hole is , the less anisotropic the high energy photon field is .
this is principally due to the curvature of geodesics making the photons emitted near the black hole arrive at larger angle than in the newtonian case .
the radiative balance between the hot source and the disk allows to compute the temperature profile on the disk surface .
it is , in fact , markedly different from `` standard accretion disk model '' as shown in figure [ anisotropytemperature ] . indeed , even if at large distances , all models give the same asymptotic behavior @xmath1 , in the inner part of the disk , it keeps increasing in `` standard model '' whereas , in our model , for @xmath2 , the * temperature saturates * around a characteristic value @xmath3 .
the differences between newtonian and kerr metrics come only from gravitational and doppler shifts , which are only appreciable for @xmath4 .
the overall uv to x - ray spectra can be deduced from this model .
the bulk of the energy coming from the disk gives the well - known `` blue - bump '' . on the other hand , in order to avoid run - away electrons , the relativistic particle distribution is supposed to be a * power law with an exponential cut - off * about 300 kev , giving the high energy part of the spectrum .
one can see on figures [ specangzovar]a newtonian spectra for different inclination angles for @xmath5 .
for all inclination angles , the kerr spectra are always weaker in uv and brighter in x - ray than the newtonian ones .
these effects are much less pronounced for high @xmath6 values because the emission area is much larger , and thus is less affected by relativistic corrections .
figure [ specangzovar]b shows the overall spectrum , for different values of @xmath6 in kerr metrics for @xmath7 .
the relativistic effects become important for values of @xmath6 smaller than about @xmath8 .
they produce a variation of intensity lowering the blue - bump and increasing the hard x - ray emission .
the change in the uv range is due to the transverse doppler effect between the rotating disk and the observer , producing a net red - shift . in the x - ray range ,
the variation is due to the high energy dependence on @xmath6 ( cf figure [ anisotropytemperature ] ) .
the observed * x / uv ratio * can then be strongly altered by these effects .
quantitatively , the luminosity ratio between the maximum of the blue - bump and the x - ray plateau goes from @xmath9 in the newtonian case , to @xmath10 for @xmath11 .
with some further assumptions , the model predicts scaling laws quite different from the standard accretion models .
if one assumes a constant high energy cut - off ( possibly fixed by the pair production threshold ) and a constant solid angle subtended by the hot source , then the following mass scaling laws apply : @xmath12 the first equality could explain the weak variations of the blue bump component , even though the luminosity l ranges over 6 orders of magnitude from source to source @xcite .
the second one is in agreement with the results of @xcite where they find in a sample of seyfert i and quasars a correlation between the mass and the luminosity under the form @xmath13 with @xmath14 .
we considered an infinite disc of neutral matter with solar abondances ( opacities given by @xcite ) , and hydrogen column density of @xmath15 .
we found that about 10 % of the total energy emitted toward the disc is reflected and does not contribute to the disc heating .
however , the reflection component has important effects on the shape of the observed spectrum . due to the strong anisotropy of the primary source
, the ratio of the reflected to the primary component can be very high .
the spectrum strongly depends on the viewing angle @xmath16 with respect to the disc normal . for nearly edge on inclinations ,
the reflection contribution is weak . as the inclination decrease the reflection component increase @xcite .
conversely , the primary spectrum decreases with decreasing @xmath16 .
so that , for a face on inclination , reflection dominates the hard x - ray spectrum . a rough comparison to ic4329a and ngc 4151 spectra shows that our model is reasonnably close to the data if the orientation is nearly edge - on ( cf . fig [ reflection ] ) .
we can note that the anisotropic illumination model predicts a weak uv component for high inclination angles which is in qualitative agreement with the observed uv to soft x - ray luminosity ratio in these two objects .
+ + * acknowledgment :* we are very grateful to a. zdziarski for providing us the ngc4151 and ic4329a data . | we develop a self - consistent model of seyfert galaxies continuum emission .
the high energy source is assumed to be an optically thin plasma of highly relativistic leptons ( @xmath0 ) , at rest at a given height on the disk axis .
such a geometry is highly anisotropic , which has a strong influence on compton process .
monte - carlo simulations allow the superposition of a reflected component to the uv to x - ray spectrum obtained with our model , leading us to a first comparison with observations by fitting the high energy spectra of ngc4151 and ic4329a . |
nowadays information is one of the most important resources . we can defeat the enemy in a war just manipulating his data .
if we can guess the mechanism of the generation of ( pseudo-)random numbers used by a casino , then we can efficiently cheat in gambling@xcite . however , most of the so - called random number generators has a deterministic algorithm inside .
it is very difficult to develop a reliable pseudo - random number generation ( prng ) method .
although there are tests@xcite that allow to check whether a sequence of numbers conforms to a particular probability distribution , we can never be sure its security without the knowledge how the sequence was generated .
one of the measures of randomness is so - called min - entropy@xcite . in particular , in the context of authentication , min - entropy is the probability of guessing the easiest key in a given distribution of keys@xcite .
if we know the pseudo - random generating algorithm and the initial seed ( or some sequence of generated numbers ) , then the randomness of such a source is equal to zero .
all classical prngs have this drawback . on the other hand
, quantum physics confuses philosophers with randomness on its deepest level .
this randomness is unavoidable . we know
that if certain observables ( _ bell operators _ , which are linear functions of observed probabilities occurring in the experiment ) attain certain thresholds , then the process must be intrinsically random , or we would have to abandon some ideas that are fundamental to all physical theories .
thus , values of these observables guarantee that the results of performed measurements are indeed random , no matter how does the measuring apparatus work .
this way the idea of the quantum randomness certification emerged@xcite .
if we want to be sure that the device we are using does really produce random numbers , we perform _ bell experiment _ , which is a kind of self testing .
such an experiment involves at least two separated parties that perform subsequent measurements with different settings without any communication between them .
after series of such measurements , the collected data is used to estimate the joint probabilities of the outcomes conditioned on the settings used .
the most prominent example of bell operator is so - called clauser - horne - shimony - holt ( chsh)@xcite . because such self testing works independently of the internal workings of the device used ( in particular ,
the exact form of the performed measurements is not important ) , if the bell inequality attains some value , we are sure that the generated results are indeed random , even if the device has been construed by a malevolent party . the amount of the obtained secure randomness is precisely quantified by means of min - entropy@xcite .
this approach , in which we do not trust the vendor of our devices and draw conclusions only from the observed results , is called _ _ device - independent__@xcite , referred further as _
di_. still , bell experiments are very difficult to do .
they require a high degree of precision and extremely high detection efficiencies . so far loophole - free bell experiment has not been successfully performed .
but when we allow to send a state from one part of the device to another , then we do not have any non - locality , which is crucial for that way of certification .
it was shown that , if we can bound the dimension of the communicated system , we still may use this _ prepare and measure scheme _ to certify the randomness@xcite . since we have to know something about the construction of the device ,
this approach is called _ _ semi - device - independent__@xcite ( denoted hereafter _ sdi _ ) .
this offers a good compromise between security and experimental feasibility .
currently all commercial quantum random generators are based on the prepare and measure scheme , _
e.g. _ the _ i d quantique_s@xcite device _ quantis _ , or the _ qstream _ by _ _
quintessence labs__@xcite .
these devices do not perform any self testing , so we are forced to trust their vendors . for this reason ,
methods for certifying randomness in the prepare and measure scheme with the semi - device - independent approach should be investigated . in this framework analogs of bell inequalities , called _ dimension witnesses _ @xcite , are used .
before we proceed we should stress that what we call random number generation is in fact randomness _ expansion _ , the process that starts with some amount of initial randomness and uses it to obtain more of it .
the presented self testing procedure of the device also requires some amount of randomness ( in order to choose the measurement settings in rounds of testing experiments ) .
strictly speaking , all quantum random number generators that use bell inequalities or dimension witnesses to certify the randomness are randomness expanders .
after generation of a string of bits with a certain amount of min - entropy , it is possible to _ extract _ its randomness what means using a certain algorithm to produce a shorter string with a larger min - entropy per bit@xcite .
in our previous paper @xcite we have investigated the relation between random number expansion protocols based on correlations occurring in the scenario where two parties share an entangled state , and on protocols relying on the prepare and measure scheme . in this paper
we develop these ideas .
the organization of this paper is as follows . in the section [ sec : motivation ] we present
scenario in which we are working .
next , in sections [ sec : bi ] and [ sec : dw ] , we give basic information about bell inequalities and dimension witnesses .
then , in the section [ sec : bitodw ] , we recapitulate a heuristic method of obtaining a dimension witness from a bell inequality .
this method was introduced in @xcite . in sections
[ sec : dwtobi ] and [ sec : binarydw ] we precisely state the conditions when the randomness certified by the violation of a bell inequality lower - bounds the randomness certified by a certain value of dimension witness in the semi - device - independent scenario . in the section [ sec : symdw ] we investigate the properties of a certain class of dimension witnesses and introduce a procedure of dimension witness reduction , which can be used to obtain from an existing witness a new one with higher amount of certifiable randomness . in the section
[ sec : explicitexamples ] we give examples of application of the presented methods .
the aims of this paper are as follows .
we clarify the methods from our previous paper @xcite and give a tighter lower bound on randomness . using these methods we obtain better dimension witnesses , in particular the one based on the braunstein - caves bell inequality @xcite .
we also extend the applicability of the methods from @xcite to arbitrary dimensions .
suppose we are a developer of a random number generating device .
since consumers do not trust us , we are interested in finding a way of certification for our device .
common method for the certification of quantum random number generators that are based on measurements on entangled particles is to estimate the value of a certain bell inequality that is attained in this device .
still , it is too difficult to observe a loophole - free violation of bell inequality .
thus we prefer prepare and measure protocols . both for prepare and measure protocols in the semi - device - independent approach , and for correlation protocols in the device - independent scheme
, we would like to define a value that measures how reliable is its particular realization . as this value
we take the expectation value of the relevant dimension witness or bell inequality , respectively , attained in the relevant protocol .
this value is called a _ security parameter_. it is possible to consider several relations .
one may ask whether , having a protocol of one type , we can relate it to some protocol of another type , in such a way that for the same value of their security parameters the min - entropy certified in one of them , is upper or lower bounded by min - entropy certified by the other one . one may start with a protocol based on a bell inequality and construct out of it a prepare and measure protocol certifying a reasonable amount of min - entropy .
this is useful since there are many randomness expansion protocols based on bell inequalities@xcite and it is easy to obtain new ones @xcite .
another situation is when we begin with some sdi protocol and want to lower bound the certified randomness using efficient numerical methods from @xcite , that works in the device - independent approach
. we present a way to obtain a new bell inequality with the property that the di protocol using it certifies at most as much randomness as the sdi protocol .
as mentioned above , sdi protocols are much easier to implement than the protocols based on entanglement .
for this reason it is useful to have a method that allows to develop devices of the first kind with the help of the well established knowledge about the devices of the second type .
we define for a di protocol : [ probdi ] let @xmath0 , @xmath1 , @xmath2 , and @xmath3 be sets .
probability distribution in di scheme is a conditional probability distribution @xmath4 such that @xmath5 where @xmath6 and @xmath7 are sets of povms on a hilbert space @xmath8 , and @xmath9 is a density matrix on @xmath8 , and @xmath10 = 0 \text { if } x \neq y.\ ] ] we denote this probability by @xmath11 .
\nonumber\ ] ] if @xmath12 , then @xmath4 is called binary .
the set of all di probability distributions for given @xmath0 , @xmath1 , @xmath2 and @xmath3 is denoted by @xmath13 .
let us take two sets , @xmath2 and @xmath3 , that label the measurement settings of alice and bob in di scheme , and two sets , @xmath0 and @xmath1 , that label their respective outcomes .
a bell inequality is a linear function defined , in particular , for probability distributions @xmath13 .
it is of the form @xmath14 \equiv \\ & = \sum_{a \in a } \sum_{b \in b } \sum_{x \in x } \sum_{y \in y } \alpha_{a ,
b , x , y } p(a , b|x , y ) + c_i , \end{aligned}\ ] ] where @xmath15 .
we omit @xmath16 if it is obvious which probability distribution is considered .
the constant term @xmath17 in a bell inequality does not change its properties .
still , we retain this general form , both for bell inequalities , and dimension witnesses in the next section . in the following sections this allows to keep the same maximal expected value when performing a transformation leading from one expression to another .
a particular form of bell inequality is the following correlation form @xmath18 \equiv \\ & = \sum_{x \in x } \sum_{y \in y } \hat{\alpha}_{x , y } c(x , y ) + \hat{c}_i , \end{aligned}\ ] ] with @xmath19 , and @xmath20 obviously , the form ( [ bi ] ) conforms the form ( [ bihat ] ) if , and only if @xmath21 , and @xmath4 is binary . for given @xmath0 , @xmath1 , @xmath2 , @xmath3 , @xmath22 , @xmath23 , bell inequality @xmath24 and @xmath25 we define the following terms : @xmath26 @xmath27 @xmath28 \geq s \end{aligned}\ ] ] the expression @xmath29 is called min - entropy , and @xmath30 is the min - entropy certified by the value @xmath31 of @xmath24 .
for a sdi scheme , we have the following definition of the allowed probability distribution [ probsdi ] let @xmath32 , @xmath33 , and @xmath34 be sets , and @xmath8 be a hilbert space of a finite dimension @xmath35 .
a probability distribution in sdi scheme is a conditional probability distribution @xmath36 such that for @xmath37 , @xmath38 and @xmath39 we have @xmath40 , where @xmath41 is a set of density matrices on @xmath8 , and @xmath42 are povms on @xmath8 for all @xmath39 .
we say that @xmath43 is realized by sets @xmath44 and @xmath45 , and denote it @xmath46 .
\nonumber\ ] ] if @xmath47 , then @xmath48 is called a binary probability distribution .
the set of all sdi probability distributions for given @xmath35 , @xmath32 , @xmath33 and @xmath34 is denoted by @xmath49 .
the set of all sdi probability distributions with restrictions that @xmath50 , @xmath47 and @xmath51 is denoted by @xmath52 .
let @xmath33 and @xmath34 be sets labeling the settings of alice and bob , in the sdi scheme , and let @xmath32 be a set of the outcomes that bob can obtain .
dimension witnesses are linear functions of probability distributions of the form @xmath53 \equiv \\ & = \sum_{b \in \bar{b } } \sum_{x \in \bar{x } } \sum_{y \in \bar{y } } \beta_{b , x , y } p(b|x , y ) + c_w , \end{aligned}\ ] ] where @xmath54 , and @xmath55 . if @xmath47 , then the dimension witness is called binary .
if @xmath56 , then the dimension witness is called zero - summing . for given @xmath32 , @xmath33 , @xmath34 , @xmath57 , @xmath58 , dimension witness @xmath59 , @xmath25 and @xmath55 we define the following terms : @xmath60 @xmath61 @xmath62 \geq s \end{aligned}\ ] ] @xmath63 @xmath64 the expression @xmath65
is called min - entropy , and @xmath66 is the min - entropy certified by the value @xmath31 of @xmath59 ( for the dimension @xmath35 ) .
the following lemma summarizes some properties of dimension witnesses .
[ dwlemma ]
let @xmath8 be a hilbert space of a dimension @xmath67 , and let @xmath59 be a binary dimension witness defined by certain @xmath33 , @xmath34 , @xmath68 and @xmath69 .
let @xmath70 be a set of states on @xmath8 , and @xmath71 be a set of binary povm on @xmath8 .
let @xmath72 $ ] .
then , the following implications hold : 1 . if @xmath73 , then there exists a set of binary povm on @xmath8 , @xmath74 , such that @xmath75 , and @xmath76 = s$ ] .
2 . if @xmath77 , and @xmath78 , then for @xmath79 , which is a set of states on @xmath8 , @xmath80 = -s$ ] .
if @xmath78 , then there exist a set of projective measurements , @xmath81 with @xmath82 , such that @xmath76 \geq s$ ] . 1 .
let us take @xmath39 .
let @xmath83 , @xmath84 , and @xmath85 .
obviously @xmath86 + now , we prove that @xmath87 .
there exist an orthonormal basis @xmath88 in that @xmath89 and @xmath90 where @xmath91 $ ] .
we have @xmath92 , and @xmath93 .
thus @xmath94 since @xmath95 and @xmath96 , we have @xmath97 , and @xmath98 .
similarly , we check that @xmath99 and @xmath100 .
+ repeating this construction for all @xmath39 , we obtain a set of povm , @xmath101 .
+ we have @xmath102 and thus @xmath103 = \sum_{b , x , y } \beta_{b , x , y } \operatorname{tr}(\rho_x \tilde{m}^b_y ) \\ & = s + \sum_y c_y \left ( \sum_x \beta_{0,x , y } - \sum_x \beta_{1,x , y } \right ) = s. \nonumber \end{aligned}\ ] ] 2 .
we have @xmath104 = \sum_{b , x , y } \beta_{b , x , y } \operatorname{tr}\left ( ( \openone - \rho_x ) m^b_y \right ) \\ & = \sum_{b , x , y } \beta_{b , x , y } ( 1 - p(b|x , y ) ) \\ & = \sum_{b , x , y } \beta_{b , x , y } - \sum_{b , x , y } \beta_{b , x , y } p(b|x , y ) = -s .
\nonumber \end{aligned}\ ] ] 3 . for any @xmath39 we have @xmath105 and @xmath106 for a certain basis @xmath107 , @xmath108 $ ] .
+ let us define @xmath109 . denote @xmath110 by @xmath111 , and similarly @xmath112 by @xmath113 .
+ we have @xmath114 , and @xmath115 if @xmath116 , then we take @xmath117 and @xmath118 , otherwise we take @xmath119 and @xmath120 .
for @xmath121 it is easy to see that @xmath122 = \sum_y \max(s_{y,0},s_{y,1 } ) \geq \sum_y s_y = s. \nonumber\ ] ] the first statement in this lemma says that in the dimension @xmath67 the condition that all measurement operators have trace @xmath123 is not restrictive with regards to the set of values possible to attain .
the second statement gives sufficient conditions under which an operation of negation of all states gives the same value of a dimension witness but with opposite sign .
the third statement , which may be used to complement the first one , shows that under certain conditions it is not restrictive to use only projective measurements in case when the values possible to be attained are considered .
consider the following bell experiment .
suppose we are given a bell inequality of the form ( [ bi ] ) .
alice and bob share an entangled state .
alice chooses a measurement setting @xmath124 , and obtains an outcome @xmath125 . for each setting @xmath126 and result @xmath127 , we assign a conditional probability @xmath128 .
alice s measurement prepares some state at bob s side .
next , bob chooses a measurement setting @xmath129 , and obtains an outcome @xmath130 . the probability that bob gets @xmath131 , knowing both the setting and the result of alice , is @xmath132 .
we rewrite the joint conditional probability of a given pair of results for a given pair of settings as @xmath133 .
thus , defining @xmath134 , the initial bell inequality is transformed to the _ form _ of a dimension witness ( see the equation ( [ dw ] ) ) , with @xmath135 .
we have @xmath136 , @xmath137 , and @xmath138 .
the fact that it is possible to transform a bell inequality into the form of a dimension witness , leads us to some _
heuristic _ method to achieve an sdi protocol that certifies a reasonable amount of randomness , once we have a di protocol .
we get the sdi protocol if , instead of measuring on alice s side , she gets `` the outcome '' as a part of her input with the probability distribution @xmath128 .
thus , we obtain a pair @xmath139 that we use as an index of the state to be send .
this way , the device on the side of alice prepares one of @xmath140 states @xmath141 .
bob still has @xmath142 measurement settings .
in this section we construct a sequence of devices that shows that the randomness certified by an sdi protocol can be lower bounded by the randomness certified in a certain di protocol minus @xmath143 .
we consider a device @xmath144 that we get from an untrusted vendor , and that consists of two black boxes .
its only parameter that we can verify ( or trust ) , is the dimension of the message send from one part of it , to the another one .
we assume , that the device can not communicate with the world outside the laboratory .
the black box on alice s side has buttons with labels @xmath38 and emits one of the states of the dimension @xmath35 from the set of states @xmath41 .
the states are unknown to us , and are of arbitrary , possibly mixed , form .
the black box on bob s side has buttons with labels @xmath39 and , after receiving the qubit from alice s black box , it performs one of the measurements given by povms from the set@xmath45 .
we do not know , how the measurements are performed .
this description is semi - device independent , since we know only the dimension @xmath35 .
suppose we are given a dimension witness w ( of the form ( [ dw ] ) ) that achieves in the experiments on the device @xmath144 the expected value @xmath145 .
we denote the conditional probability of obtaining the outcome @xmath131 when the chosen settings are @xmath126 and @xmath146 , by @xmath147 . the device @xmath144 is not trusted , but it is possible to consider another device , @xmath148 , that consists of two parts , with buttons labeled by @xmath38 and @xmath39 on the alice s side and on the bob s side , respectively .
the parts are sharing a maximally entangled state of the dimension @xmath35 .
the part on the alice s side performs some measurement , depending on the chosen input @xmath126 .
this measurement projects the alice s part of the singlet on the state @xmath149 that is the same as the relevant state from the device @xmath144 .
if the projection succeeded , which happens with the probability @xmath150 , then the device returns @xmath151 and changes the state on the alice s side into the state @xmath149 , otherwise it returns @xmath152 .
since the shared state is a singlet , this measurement prepares the same @xmath35-dimensional state on the bob s side .
then he performs the same povm @xmath42 as the device @xmath144 , and returns the outcome @xmath37 .
the probability that alice gets the outcome @xmath127 with the setting @xmath126 , and simultaneously bob gets the outcome @xmath131 with the setting @xmath146 is denoted by @xmath153 .
it is easy to see , that @xmath154 .
now let us consider another device , @xmath155 .
it has the same interface like @xmath148 , but the conditions on the internal working are relaxed , _ viz .
_ we do not assume anything about the performed measurements , and alice s and bob s parts are allowed to share any , possibly entangled , state @xmath9 of an arbitrary dimension .
the probability of obtaining the outcomes @xmath127 and @xmath131 with given pair of settings @xmath126 and @xmath146 for alice and bob , respectively , are denoted by @xmath156 .
we apply a constraint @xmath157 , where @xmath158 is the probability of getting the outcome @xmath127 by alice with the setting @xmath126 with the device @xmath155 .
obviously , all the conditional probability distributions that are possible to be obtained by the device @xmath148 ( and thus also by the device @xmath144 ) , are also possible to be obtained by this device .
note that this description is fully device - independent , and that there are semi - definite programs in the npa hierarchy @xcite that efficiently approximate the probability distributions of the device @xmath155 .
since the device @xmath155 is a relaxed version of the initial device @xmath144 , if both of them have the same value of the relevant security parameters , then the certified amount of min - entropy generated by the device @xmath155 gives a lower bound of the min - entropy certified to be generated by the device @xmath144 .
we recapitulate the above results in the following theorem [ dwtobitheorem ] let @xmath159 , @xmath160 and @xmath161 be sets .
let us take @xmath25 , @xmath55 , a bell inequality @xmath24 of the form ( [ bi ] ) , and a dimension witness @xmath59 of the form ( [ dw ] ) , satisfying @xmath162 .
let @xmath163 be a subset of @xmath49 with @xmath55 ( see the definition [ probsdi ] ) that satisfies @xmath164 .
let @xmath165 be a set of all probability distribution defined by @xmath166 , where @xmath4 is a device - independent probability distribution such that @xmath167 = s$ ] , with @xmath168 . then @xmath169 .
this way we obtain a way to get a relation between bell inequalities and dimension witnesses with the property that the amount of randomness certified by a bell inequality lower - bounds the amount of randomness certified by the relevant dimension witness .
one of key features of the set @xmath170 is that it can be efficiently approximated using semi - definite programming with the npa hierarchy . from the definition of @xmath171 , namely using @xmath172
, we get that the certified min - entropy of sdi protocol is lower - bounded by the one of the di protocol minus @xmath143 .
notable property of the method is that we obtain a bound for any dimension of the communicated system changing only a value of the linear bound . ) in section [ sec : explicitexamples ] ) using the methods from the section [ sec : dwtobi ] for different values of the dimension @xmath35.[fig : cglmp],scaledwidth=45.0% ] the figure [ fig : cglmp ] shows an example of application of the theorem [ dwtobitheorem ] .
in this section the properties of binary zero - summing dimension witnesses are investigated .
recall that a dimension witness of the form given by the equation ( [ dw ] ) is called zero - summing if @xmath56 , and binary if @xmath47 .
the reason to examine them is that it is possible to obtain a tighter semi - definite relaxations for this class of dimension witnesses .
let us start with a binary zero - summing dimension witness @xmath173 that is used to certify the randomness generated by measuring the state @xmath57 with the measurement setting @xmath58 .
let @xmath44 and @xmath174 be the states and measurements that maximize the guessing probability ( see the equation ( [ pguess ] ) ) of the generated bits by the untrusted vendor .
first note that the value of the dimension witness does not change if , for arbitrary @xmath39 , the measurement is changed to @xmath175 , where @xmath176 is such that the spectrum of the operators remains in the range @xmath177 $ ] .
thus , since the potential adversary is interested in increasing the probability of a particular outcome of the measurement @xmath178 as much as possible , the form of these measurements that maximizes his guessing probability is the following : @xmath179 by the lemma [ dwlemma].1 and [ dwlemma].3 it is not restrictive for the vendor to use only projectors of trace @xmath123 for the measurements different than @xmath178 .
the strategy of using a measurement of the form ( [ optmeasy0 ] ) for the setting @xmath178 , and projectors of trace @xmath123 for all remaining measurements is equivalent to using the following mixed strategy . in @xmath180 cases , a projective measurement of trace @xmath123
is used for the measurement @xmath178 ( we call this strategy : p ) , and in @xmath181 cases the outcome is deterministic - this is referred hereafter as a deterministic strategy , or simply : d. for the remaining measurements the same projective measurements of trace @xmath123 are used in both cases .
the guessing probability for the strategy d is @xmath123 , and for the strategy p is @xmath182 , thus the average guessing probability is @xmath183 in the case of a zero - summing dimension witness with the deterministic strategy , measurements with the setting @xmath178 give no contribution to the value of the witness .
thus the certification of the randomness with the dimension witness @xmath184 when the vendor of the device uses the mixed strategy is , after applying certain affine transformation ( see equation ( [ deltaaffine ] ) ) , equivalent to the certification with a dimension witness @xmath185 with @xmath186 defined in the equation ( [ deltabeta ] ) , and the strategy p , where the guessing probability of eve is given by the equation ( [ deltaaffine ] ) .
since the vendor may choose any @xmath187 $ ] that allows to observe the required value of the dimension witness @xmath59 when calculating lower - bound on the certified min - entropy , the worst case should be considered for a particular situation .
this way , we have proved the following let @xmath173 be a binary zero - summing dimension witness , @xmath57 , and @xmath58 .
let @xmath188 be a dimension witness , where @xmath189 then @xmath190 } \left ( ( 1 - \delta ) + \delta \cdot p_{guess}^{cert(p)}(w_{(\delta , y_0)},x_0,y_0,s ) \right ) \end{aligned}\ ] ] where @xmath191 and @xmath192 are defined in equations ( [ pguesscert ] ) and ( [ pguesscertp ] ) .
the consequence of restricting the vendor to the dimension @xmath67 and measurements of trace @xmath123 is that the following holds for all @xmath38 and @xmath39 , and for any @xmath193 : @xmath194 where @xmath195 .
this relation allows to refine the relaxation given in the section [ sec : dwtobi ] .
let us consider a device @xmath196 that models the strategy p by sharing the singlet state , projecting on states @xmath44 on the side of alice , and measuring on the side of bob with measurements of trace @xmath123 , @xmath174 .
in contrast to the device @xmath148 , if the projection on a state @xmath197 for any @xmath38 fails , then the prepared state is @xmath198 .
it is easy to see that , by the equation ( [ negrho ] ) , the probabilities obtained in this device are constrained by the following relation : @xmath199 for all @xmath200 , @xmath124 , and @xmath129 .
a further relaxation , analogous to the one leading from the device @xmath148 to the device @xmath155 , allows to obtain a device @xmath201 , satisfying the relation ( [ neganegb ] ) , that can be modeled by a semi - definite program in the device - independent scheme . this way we have proved the following theorem [ zerosumtrm ]
let @xmath160 and @xmath161 be sets .
let us take @xmath25 , a bell inequality @xmath24 of the form ( [ bi ] ) , and a binary zero - summing dimension witness @xmath59 of the form ( [ dw ] ) , satisfying @xmath202 .
let @xmath203 be a subset of @xmath52 ( see the definition [ probsdi ] ) containing those probabilities @xmath204 that satisfies @xmath205 = s$ ] .
let @xmath206 be a set probability distributions defined by @xmath207 , where @xmath4 is a device - independent probability distribution that satisfies @xmath167 = s$ ] and the relation ( [ neganegb ] ) . then @xmath208 .
it is straightforward to check that the following lemma holds : [ probimply ] let @xmath209 , and let us assume that @xmath210 @xmath211 _ i.e. _ the no - signaling principle , and that the outcomes of bob are binary , namely @xmath212 then we have the following implications : 1 .
if @xmath213 holds , then we have @xmath214 and @xmath215 .
if @xmath215 holds , then we have @xmath216 .
if @xmath214 and @xmath216 hold , then we have @xmath217 . from this lemma
we get that the condition @xmath213 is more restrictive than @xmath214 . from this
we conjecture that for any @xmath31 @xmath218 where the sets are defined in theorems [ dwtobitheorem ] and [ zerosumtrm ] .
thus the theorem [ zerosumtrm ] refines the results of the theorem [ dwtobitheorem ] for the case of binary zero - summing dimension witnesses .
figures [ fig : pstrategylower ] and [ fig : pstrategylowerupper ] show examples lower- and upper - bounds for min - entropy certified when the untrusted vendor uses the strategy p. figures [ fig : mixedstrategylowerdelta ] and [ fig : mixedstrategylower ] show lower - bounds for the certified min - entropy in case when the untrusted vendor uses the mixed strategy .
all lower - bounds are calculated via semi - definite programs with the npa hierarchy , using interior point method with the sedumi solver @xcite .
the upper - bounds have been obtained by finding explicit representations of states and measurements .
this optimization has been carried over pure states and projective measurements , and is not guarantied to reach global minima , in contrast to the semi - definite programming method .
interestingly , in all protocols considered in the figure [ fig : mixedstrategylower ] , it is optimal for the adversary to use @xmath219 , _
i.e. _ using the mixed strategy gives no gain comparing to the strategy p.
let us introduce the following definition : [ symdwdef ] a dimension witness @xmath59 of the form ( [ dw ] ) with the set of alice s settings @xmath33 of even size , and @xmath47 , is _ symmetric _ , if there exist an surjective automorphism @xmath220 with @xmath221 and @xmath222 . for a set @xmath223 we define @xmath224 a set @xmath223
satisfying @xmath225 , and @xmath226 is called a half of @xmath33 .
if a set @xmath227 is a half , then @xmath228 is called a dimension witness reduced with respect to @xmath229 .
@xmath230 and @xmath229 may be omitted if it is obvious which automorphism or set is considered .
if a dimension witness is symmetric , then there is a way to reduce the size of @xmath33 , whilst the obtained dimension witness can certify at least the same amount of randomness , as the initial one .
the following theorem is an immediate result of the theorem [ zerosumtrm ] and the lemma [ dwlemma].2 : [ symaddcond ] for a sdi protocol using the strategy p with a symmetric dimension witness that attains the value of the security parameter @xmath31 on a hilbert space of the dimension @xmath67 and certifies the randomness @xmath231 , the same value is still possible to be attained and certifies at least the same randomness , if we impose an additional condition that @xmath232 , which implies @xmath233 . simply speaking this theorem
says that symmetric dimension witnesses posses some kind of degree of freedom that does not increase the range of values possible to be attained , but allows adversary to `` distribute '' the value of the witness among the states in such a way that misleads about the reliability of the device .
the proposed method shows a way to remove this freedom .
this subsection shows how to transform a symmetric dimension witness to a reduced one .
it is possible to use a known bell inequality to obtain a new dimension witness .
examples of such protocols , @xmath234 , @xmath235 , @xmath236 and @xmath237 , are described below in the section [ sec : explicitexamples ] . from a bell inequality of the form @xmath238 ( where @xmath239 ) , using the method from the section [ sec : bitodw ] , we obtain a symmetric dimension witness of the form ( [ dw ] ) , with @xmath240 , and @xmath241 . for the new sdi protocol , we assume that @xmath127 is chosen randomly by alice , with the distribution @xmath242 . note that bell inequalities in the correlation form ( see the equation ( [ bihat ] ) ) , are a special case of the inequalities of the form ( [ symbi ] ) with @xmath243 and @xmath244 , which means that it is always possible to obtain a symmetric dimension witness from a correlation - based bell inequality .
then @xmath245 turns into @xmath246 it is easy to see , that a dimension witness which is a linear combination of expressions ( [ ctowfull ] ) , is symmetric .
we define @xmath247 and @xmath248 .
the condition @xmath249 allows us to take @xmath250 instead of @xmath251 from the equation ( [ ctowfull ] ) , which is an example of the reduction .
note that using the method of reduction of a symmetric dimension witness the number of states used by alice is reduced twice without loss of of ability to certify both the randomness and the dimension . on the other hand every symmetric dimension witness is a linear combination of expressions @xmath252 that refers in di scenario to the expression @xmath253 .
assuming that the dimension of the hilbert space is @xmath67 , and the eavesdropper uses the strategy p , we get from the equation ( [ neganegb ] ) that @xmath254 .
in this section we give five examples of applications of the methods presented above .
four of them , b , c , d , and e , concern bell inequalities in the correlation form and symmetric dimension witnesses .
all figures are plotted with respect to a relative parameter @xmath182 .
the value @xmath255 refers the case when the maximal value of the relevant bell inequality or dimension witness is achieved .
values @xmath256 relate to the situation with noise , when the attained value is equal to the maximum multiplied by @xmath182 . in the first example we start with cglmp inequality introduced in @xcite .
both alice and bob have two measurement settings with three outcomes .
it has the following form @xmath257 using the heuristic method from the section [ sec : bitodw ] we obtain the following dimension witness : @xmath258 applying the method from the section [ sec : dwtobi ] , we get the following expression , which may be used in a semi - definite program : @xmath259 the certified randomness for cglmp is shown in the figure [ fig : cglmp ] .
in the second example we start with a well known braunstein - caves inequality ( denoted below @xmath236 , it is a bell inequality in the form ( [ bihat ] ) ) with three settings for each of the two parties , and convert it to a symmetric dimension witness with six prepared states . after reduction
, we will obtain a dimension witness with three states , and show that the lower bounding bell inequality is identical to the original @xmath236 .
@xmath236 inequality is of the form @xmath260 with @xmath219 .
for @xmath261 we have @xmath262 .
thus we obtain a symmetric dimension witness with six states prepared by alice and three measurements performed by bob .
the explicit form of this symmetric dimension witness is @xmath263 using the method for symmetric dimension witnesses from the section [ sec : symdw ] , this may be transformed into a dimension witness with three states .
we define @xmath247 and @xmath264 .
the explicit form of this reduced dimension witness is @xmath265 now , using the theorem [ zerosumtrm ] , we go from this reduced dimension witness back to the bell inequality that gives a lower - bounding relation . assuming @xmath213 , we get that the lower - bounding bell inequality is exactly the initial braunstein - caves inequality . if we use a full dimension witness , then the bell inequality used in lower - bounding with the theorem [ zerosumtrm ] is @xmath266 where @xmath267 , @xmath268 , @xmath269 , @xmath270 , @xmath271 , and @xmath272 . in the figure [ fig : bc3_npa_p ] , the min - entropies certified with the bell inequality @xmath236 with different additional conditions are plotted .
the figure [ fig : bc3_cert_p ] shows lower - bounds on the min - entropy certified in this sdi protocol , obtained by theorem [ dwtobitheorem ] from the npa hierarchy with additional condition @xmath217 .
these values assume that the untrusted vendor uses the strategy p ( see the section [ sec : binarydw ] ) .
plots relevant to the mixed strategy are shown on figures [ fig : bc3_npa_delta ] and [ fig : bc3_cert ] .
[ [ sub : t3 ] ] t3 ~~ the third example starts with a dimension witness based on @xmath273 to @xmath123 quantum random access code @xcite , and relates it , and its reduced version , to two bell inequalities , where the second one is @xmath235 introduced in @xcite .
in @xmath273 to @xmath123 quantum random access code alice encodes three bits by sending one of the @xmath274 states to bob , who tries to guess one of them , performing one of three measurements .
the average success probability of correctly guessing an arbitrarily chosen bit is directly related to the value of the following dimension witness : @xmath275 where @xmath276 , @xmath277 .
its maximal value attainable with qubits is @xmath278 .
taking @xmath279 ( negation is meant here as bit - wise ) , @xmath280 and @xmath281 , we get the following reduced dimension witness @xmath282 from this dimension witness , using the method from the section [ sec : symdw ] , we get the following bell operator : @xmath283 if we do not reduce the dimension witness and use the formula ( [ t3dwfull ] ) directly , we get the following bell operator : @xmath284 the bell operator defined in the equation ( [ t3bi ] ) is the one used in @xcite .
it is possible to calculate a lower - bound on the certified min - entropy , @xmath285 , with @xmath286 , @xmath287 using the theorem [ zerosumtrm ] , _ i.e. _ via a semi - definite relaxation with a minimization on higher level over @xmath187 $ ] .
the figure [ fig : t3_npa_p ] shows the min - entropies certified with the bell inequality @xmath235 for different additional conditions . in the figure [ fig : t3_cert_p ] lower - bounds on the certified min - entropy obtained by theorem [ dwtobitheorem ] from the npa hierarchy with additional condition @xmath217 are plotted .
these values assume that the untrusted vendor uses the strategy p ( see the section [ sec : binarydw ] ) .
figures [ fig : t3_npa_delta ] and [ fig : t3_cert ] contains the relevant data for the mixed strategy .
[ [ sub : t2 ] ] t2 ~~ a simple bell inequality is obtained from the symmetric dimension witness of the @xmath67 to @xmath123 qrac used in @xcite .
it has the following form @xmath288 where @xmath289 is defined in the equation ( [ ctowfull ] ) and @xmath219 .
the reduced form of this dimension witness is @xmath290 where @xmath59 is defined by the equation ( [ ctow ] ) and @xmath219 .
robustness of the reduced version has been already investigated in @xcite , in the figure @xmath291 .
the randomness certified by these two dimension witnesses is lower - bounded by the values obtained with the following two bell inequalities . for the dimension witness defined in the equation ( [ t2dwfull ] )
, we use a bell inequality @xmath292 and for the dimension witness from the equation ( [ t2dw ] ) , @xmath293 the operator defined in the equation ( [ t2bi ] ) is exactly the chsh bell operator .
lower bounds for this case are shown in figs [ fig : t2_npa_p ] , [ fig : t2_cert_p ] , [ fig : t2_npa_delta ] and [ fig : t2_cert ] .
the reduced witness ( [ t2dw ] ) has recently been experimentally realized @xcite .
the values obtained in this experiment refer to @xmath294 ( 5.51 in the scaling used there ) and @xmath295 ( 5.56 ) , concluded therein to certify @xmath296 and @xmath297 bits of randomness , respectively . if the reduction had not been performed , then only @xmath298 and @xmath299 would have been certified .
in @xcite the following bell operator is investigated : @xmath300 this bell operator is similar in the form to the dimension witness introduced in @xcite .
since the relevant bell inequality is very robust in certifying the randomness , the dimension witness with randomness lower - bounded by it , may also be expected to be robust .
assuming @xmath214 , we turn it into the following dimension witness @xmath301 since this dimension witness is symmetric , we follow the steps which lead from the expression ( [ ctowfull ] ) , to the expression ( [ ctow ] ) , to obtain the following reduced dimension witness @xmath302 if we start with the dimension witness defined in the equation ( [ modchshdwfull ] ) , and do not use the symmetry , we get the following lower - bounding bell inequality @xmath303 the dimension witness from the equation ( [ modchshdwfull ] ) lower - bounds the dimension witness from the equation ( [ modchshdw ] ) , and thus both are lower - bounded ( in the sense of the theorem [ dwtobitheorem ] and the conjecture below it ) by the bell inequality from the equation ( [ modchshbifull ] ) , but only the second dimension witness is proved to be lower - bounded by @xmath237 ( see the equation ( [ modchshbi ] ) ) .
lower - bounds for this set of di and sdi protocols are shown in figs [ fig : modchsh_npa_p ] , [ fig : modchsh_cert_p ] , [ fig : modchsh_npa_delta ] , and [ fig : modchsh_cert ] .
in this paper we explained in more details the ideas from our previous paper @xcite .
in particular all steps of the proof of the theorem [ dwtobitheorem ] were provided .
a tighter bound , using condition @xmath304 in di scheme , has been introduced .
we have presented a new method of dimension witness reduction and a clear distinction between reduced and full dimension witnesses has been made .
reduced dimension witnesses have been shown to be able to certify more randomness .
min - entropies of several protocols , that had not been considered previously in @xcite , were evaluated .
recently a new method that allows to lower - bound the randomness obtained in a sdi scheme directly , using semi - definite programming , has been introduced in @xcite . however , the complexity of the algorithm from @xcite increases significantly with the dimension of hilbert space while in our case the same computation provides a bound for all dimensions .
it remains an open question , what are the conditions on a dimension witness under that the adversary has no gain in using the mixed strategy rather than p.
sdp was implemented in octave using sedumi @xcite toolbox .
this work is supported by ideas plus ( idp2011 000361 ) , ncn grant 2013/08/m / st2/00626 , fnp team and the national natural science foundation of china ( grant no.11304397 ) .
the major part of this work has been written in the forests of sopot .
a. rukhin , j. soto , j. nechvatal , m. smid , e. barker , s. leigh , m. levenson , m. vangel , d. banks , a. heckert , j. dray , s. vo , _ special publication 800 - 22 revision 1a _ , national institute of standards and technology , u.s .
department of commerce , available at http://csrc.nist.gov/publications/pubssps.html r. koenig , r. renner , c. schaffner , ieee trans .
, vol . * 55 * , no . 9
( 2009 ) .
w. e. burr , d. f. dodson , e. m. newton , r. a. perlner , w. t. polk , s. gupta , e. a. nabbu , _ nist special publication 800 - 63 - 2 _ , national institute of standards and technology , u.s .
department of commerce , available at http://csrc.nist.gov/publications/pubssps.html s. pironio , a. acin , s. massar , a. boyer de la giroday , d. n. matsukevich , p. maunz , s. olmschenk , d. hayes , l. luo , t. a. manning , c. monroe , nature * 464 * , 1021 ( 2010 ) .
j. f. clauser , m.a .
horne , a. shimony , r. a. holt , phys .
* 23 * , 880 ( 1969 ) .
r. colbeck , a. kent , j. phys . a : math .
theor . , * 44*(9 ) 095305 ( 2011 ) .
d. mayers and a. yao , in focs 98 : _ proceedings of the 39th annual symposium on foundations of . computer science _ ( ieee computer society , washington , dc , usa ) , 503 ( 1998 ) .
li , z .- q .
yin , y .- c .
wu , x .- b .
zou , s. wang , w. chen , g .- c .
guo , z .- f .
han , phys .
a * 84 * , 034301 ( 2011 ) .
c . liang , t. vertesi , n. brunner , phys .
a * 83 * , 022108 ( 2011 ) .
n. brunner , s. pironio , a. acin , n. gisin , a. a. methot , v. scarani , phys .
lett . * 100 * , 210503 ( 2008 ) .
r. gallego , n. brunner , c. hadley , a. acin , phys .
lett . * 105 * , 230501 ( 2010 ) .
m. pawowski , n. brunner , phys .
a * 84 * , 010302(r ) ( 2011 ) .
m. dallarno , e. passaro , r. gallego , a. acin , phys .
a * 86 * , 042312 ( 2012 ) .
l. trevisan , _ journal of the acm _ * 48 * , 860 ( 2001 ) . a. de , c. portmann , t. vidick , r. renner , siam journal on computing * 41*(4 ) , 915 ( 2012 ) . m. tomamichel , c. schaffner , a. smith , r. renner , ieee trans .
57*(8 ) , ( 2011 ) .
li , p. mironowicz , m. pawowski , z .- q .
yin , y .- c .
wu , s. wang , w. chen , h .-
hu , g .- c .
guo , z .- f .
a * 87 * , 020302(r ) ( 2013 ) .
braunstein , c.m .
caves , phys .
lett . * 61 * , 662 ( 1988 ) .
m. navascues , s. pironio , a. acin , phys .
. lett . * 98 * , 010401 ( 2007 ) .
m. navascues , s. pironio , a. acin , new j. phys .
* 10 * , 073013 ( 2008 ) .
a. ambainis , a. nayak , a. ta - shma , and u. vazirani , _ dense quantum coding and a lower bound for 1-way quantum automata _ , in proceedings of 31st acm symposium on theory of computing , 376 ( 1999 ) .
li , m. pawowski , z .- q .
yin , g .- c .
guo , z .- f .
han , phys .
a * 85 * , 052308 ( 2012 ) .
d. collins , n. gisin , n. linden , s. massar , s. popescu , phys .
lett . * 88 * , 040404 ( 2002 ) .
sturm , optimization methods and software * 11 * , 625 ( 1999 ) . | in this paper we develop a method for investigating semi - device - independent randomness expansion protocols that was introduced in [ li _ et al . _ phys .
rev .
a * 87 * , 020302(r ) ( 2013 ) ] .
this method allows to lower - bound , with semi - definite programming , the randomness obtained from random number generators based on dimension witnesses .
we also investigate the robustness of some randomness expanders using this method .
we show the role of an assumption about the trace of the measurement operators and a way to avoid it .
the method is also generalized to systems of arbitrary dimension , and for a more general form of dimension witnesses , than it the previous paper .
finally , we introduce a procedure of dimension witness reduction , which can be used to obtain from an existing witness a new one with higher amount of certifiable randomness .
the presented methods finds an application for experiments [ ahrens _ et al .
_ phys .
rev . lett . * 112 * , 140401 ( 2014 ) ] . |
classical electrodynamics is fundamentally linked to the concepts of relativity and spacetime symmetry . in particular ,
many high precision tests of relativity are based on electromagnetic phenomena . at present
, there is a growing interest ( both theoretically and experimentally ) in the search for possible _ violations of the lorentz symmetry_. the latter may naturally occur , for example , in the framework of unified theories of physical interactions at the planck length scale , and may then manifest themselves at the scales of usual high - energy particle physics in the form of corrections to the established quantum field - theoretical models .
even though presently the limits of the standard model of elementary particle physics ( sm ) become visible , see sec.10.6 of the new particle data report @xcite , it still remains the basis of our understanding of the strong and the electroweak interactions . at the foundations of the sm
lays special relativity with the poincar symmetry group and , associated with it , the _ light cone _ structure . although this _ rigid _ light cone structure of special relativity is assumed as framework for the sm , we know from general relativity theory ( gr ) and from experiments that the light cone structure is in fact `` flexible '' . in other words , in gr
the metric becomes a space and time dependent field .
this alone demonstrates that the sm can not be strictly correct .
numerous attempts have been made to take care of this state of affairs and to build up a more general framework for the sm and perhaps even to include gr .
one of the promising approaches is the standard model extension ( sme ) of kosteleck and collaborators , see @xcite , in which by means of a set of parameters certain violations of the sm are put in a quantitative form amenable to experimental tests .
the photon sector of the sme has been particularly closely studied in @xcite , where the propagation of electromagnetic waves in vacuum was investigated with lorentz violating terms added in the effective _ constitutive _ tensor . in this paper
, we continue the study of electrodynamical phenomena in the framework of the so - called premetric approach , i.e. , without assuming a special form of the metric structure on a spacetime manifold .
such an approach provides a general technique for an appropriate physical discussion of the experimental tests of possible violations of the lorentz symmetry and of parity , since at the outset no assumptions were allowed about geometric structures , such as the metric or the connection of spacetime . on the contrary ,
the analysis of the dynamics of the electromagnetic field in this approach yields , via the light propagation , the information about the underlying metrical structure of spacetime .
accordingly , this scheme can be viewed as a general test theory of the searches for possible deviations from the lorentz symmetry in the photon sector of sme .
it is worthwhile to note that the multi - parameter approach of kosteleck and mewes @xcite can be naturally embedded into the framework of general premetric electrodynamics .
electromagnetic wave propagation is a very important physical phenomenon in classical field theory . in general , the geometrical structure of spacetime as well as the intrinsic properties and the motion of material media can affect the light propagation .
hence , a theoretical analysis of the latter ultimately results in establishing the properties of the ( genuine or effective optical " ) metric structure on the manifold . in the generally covariant premetric approach to electrodynamics @xcite ,
the axioms of electric charge and of magnetic flux conservation manifest themselves in the maxwell equations for the excitation @xmath0 and the field strength @xmath1 : @xmath2 these equations should be supplemented by a constitutive law @xmath3 .
the latter relation contains the crucial information about the underlying physical continuum ( i.e. , about spacetime and/or about the material medium ) .
mathematically , this constitutive law arises either from a suitable phenomenological theory of a medium or from the electromagnetic field lagrangian
. it can be a nonlinear or even nonlocal relation between the electromagnetic excitation and the field strength .
earlier , we have investigated the propagation of waves in the most general local and linear theory .
one can conveniently split the constitutive relation into the three irreducible pieces which are known as the _ principal _ part , the _ skewon _ part and the _ axion _ part .
it is important to stress that such a decomposition is intrinsically metricfree , and thus it suits nicely for the purpose of the study of the violations of the lorentz symmetry in the photon sector .
more specifically , we investigated the influence of the axion and the skewon fields on the wave propagation by applying a linear response formalism , see @xcite .
subsequently , itin @xcite studied birefringence caused by the axion field in the caroll - field - jackiw model of electrodynamics @xcite .
more recently , we investigated the birefringence of electromagnetic waves and its consequences for the metric of spacetime @xcite .
the possible limits of the lorentz symmetry can be seen in a particularly transparent way , both qualitatively and quantitatively , including also the case of a possible violation of the charge conservation law @xcite .
the axion field and its interference with electrodynamics is a well - studied subject , and the same applies to the influence of the general principal part of the constitutive tensor which was thoroughly analyzed in @xcite .
however , this is not so in the case of the skewon field , a new structure which allows us to exhaust all 36 parameters of a linear spacetime relation ( exact definitions are given below ) .
the only relevant earlier results are those of nieves and pal @xcite which were confined to the case of the spatially isotropic and constant scalar skewon field .
in contrast , we present here some results which hold true for all electrodynamical models with an arbitrary local and linear _ spacetime relation _ that contains a nontrivial skewon part .
we use the expression `` spacetime relation '' , since it applies to spacetime ( `` the vacuum '' ) itself , and in order to distinguish it from the constitutive law of a material medium . from a mathematical point of view
, the general local and linear spacetime relation is similar to the constitutive relation for matter with sufficiently complicated electric and magnetic properties . quite naturally
, the study and the physical interpretation of the wave propagation in a manifold with a general spacetime relation appears to be very close to the analysis of the corresponding wave effects in crystal optics .
the latter subject was extensively studied in the literature , see @xcite , for example . in particular , our analysis reveals the analogy of the effects of the real skewon field to the propagation of waves in absorbing media with circular dichroism , whereas the imaginary skewon field appears to be analogous to the optical activity tensor .
if local coordinates @xmath4 are given , with @xmath5 , we can decompose the excitation and field strength 2-forms into their components according to @xmath6
we confine ourselves to the case in which the electromagnetic excitation and the field strength are related by the local and linear constitutive law , @xmath7 the constitutive tensor @xmath8 has 36 independent components .
one can decompose this object into its irreducible pieces .
obviously , contraction is the only tool for such a decomposition . following post @xcite
, we can define the contracted tensor of type @xmath9 $ ] @xmath10 with 16 independent components .
the second contraction yields the pseudo - scalar function @xmath11 the traceless piece @xmath12 has 15 independent components .
these pieces can now be subtracted out from the original constitutive tensor .
then , @xmath13}^{l ] } + { \frac 1 6}\,\kappa\,\delta_{[i}^k\delta_{j]}^l.\label{kap - dec}\end{aligned}\ ] ] by construction , @xmath14 is the totally traceless part of the constitutive tensor : @xmath15 thus , we split @xmath8 according to @xmath16 , and the @xmath17 $ ] tensor @xmath14 is subject to the 16 constraints ( [ notrace ] ) and carries @xmath18 components .
one may call @xmath14 the principal , or the metric - dilaton part of the constitutive law . without such a term ,
electromagnetic waves are ruled out , see @xcite .
we further identify the two other irreducible parts with a _
skewon _ and an _ axion _ field , respectively .
conventionally , the skewon and the axion fields are introduced by @xmath19 the standard maxwell - lorentz electrodynamics arises when both skewon and axion vanish , whereas @xmath20 here @xmath21 is the vacuum impedance .
a spacetime metric @xmath22 is assumed on the manifold , and with @xmath23 one defines @xmath24 and @xmath25 .
it has been shown recently @xcite that taking the linear spacetime relation for granted , one ends up at a riemannian lightcone provided one forbids birefringence in vacuum , see also @xcite .
along with the original @xmath8-tensor , it is convenient to introduce an alternative representation of the constitutive tensor : @xmath26 substituting ( [ kap - dec ] ) into ( [ chikap ] ) , we find the corresponding decomposition @xmath27 with the principal , skewon , and axion pieces defined by @xmath28},\label{chi2}\\ { } ^{(3)}\chi^{ijkl } & = & { \frac 1 2}\,\epsilon^{ijmn}\,\,{}^{(3 ) } \kappa_{mn}{}^{kl } = { \frac 1 { 12}}\,\epsilon^{ijkl}\,\kappa . \label{chi3}\end{aligned}\ ] ] using the s - identity and the k - identity derived in @xcite , we can verify that @xmath29 is _ skew - symmetric _ under the exchange of the first and the second index pair , whereas @xmath30 is _ symmetric _ : @xmath31 making a @xmath32-decomposition @xcite of covariant electrodynamics , we can write @xmath33 and @xmath34 as column 6-vectors with the components built from the magnetic and electric excitation 3-vectors @xmath35 and electric and magnetic field strengths @xmath36 , respectively
. then the linear spacetime relation ( [ hchif ] ) reads : @xmath37 here the constitutive tensor is conveniently represented by the 6-matrix @xmath38 the constitutive @xmath39 matrices @xmath40 are constructed from the components of the original constitutive tensor as @xmath41 if we resolve with respect to @xmath42 , we find the inverse formulas @xmath43 the contributions of the principal , the skewon , and the axion parts to the above constitutive 3-matrices can be written explicitly as @xmath44 the set of the symmetric matrices @xmath45 and @xmath46 together with the traceless matrix @xmath47 ( i.e. , @xmath48 ) comprise the principal part @xmath49 of the constitutive tensor . usually , @xmath50 is called _ permittivity _ tensor and @xmath51 _ reciprocal permeability _ tensor ( `` impermeability '' tensor ) , since they describe the polarization and the magnetization of a medium , respectively .
the magnetoelectric cross - term @xmath47 is related to the fresnel - fizeau effects .
the skewon contributions in ( [ a ] ) and ( [ b ] ) are responsible for the electric and magnetic faraday effects , respectively , whereas skewon terms in ( [ c ] ) and ( [ d ] ) describe optical activity .
here we briefly summarize the results of previous work @xcite . in the hadamard approach ,
one studies the propagation of a discontinuity in the first derivative of the electromagnetic field .
the basic notions are then the fields of the _ wave _ covector and the _ ray _ vector that encode the information about the propagation of a wave in a spacetime with a general constitutive relation .
the crucial observation about the surface of discontinuity @xmath52 ( defined locally by a function @xmath53 such that @xmath54 on @xmath52 ) is that across @xmath52 the geometric hadamard conditions are satisfied for the components of the electromagnetic field and their derivatives : @xmath55 = 0 , \,[\partial_i f_{jk } ] = q_i\,f_{jk},\,[h_{ij } ] = 0,\,[\partial_i h_{jk } ] = q_i\ , h_{jk}$ ] . here
@xmath56 is the wave covector . then using the maxwell equations ( [ me ] ) and the constitutive law ( [ hchif ] )
, we find a system of algebraic equations for the jump functions : @xmath57 solving the last equation in ( [ 4dwave1 ] ) by means of @xmath58 , we finally reduce ( [ 4dwave1])@xmath59 to @xmath60 .
this algebraic system has a nontrivial solution for @xmath61 only if the determinant of the matrix on the left hand side vanishes .
the latter gives rise to our _ generalized covariant fresnel equation _
@xmath62 with the fourth order tamm - rubilar ( tr )
tensor density of weight @xmath63 defined by @xmath64 it is totally symmetric , @xmath65 , and thus has 35 independent components .
different irreducible parts of the constitutive tensor ( [ kap - dec ] ) contribute differently to the general fresnel equation .
a straightforward analysis @xcite shows that the axion piece drops out completely from the tr - tensor , whereas the two remaining irreducible parts of the constitutive tensor contributes to ( [ g4 ] ) as follows : @xmath66 a ray can be defined by a vector field @xmath67 that is _ dual _ to the wave covector @xmath68 and to the wave front in the following sense @xcite : @xmath69 here the 2-forms @xmath70 and @xmath71 represent the jumps of the first derivatives of the electromagnetic field .
equations ( [ svector ] ) are metric - free .
the system of the first two equations can be analysed in complete analogy to the system ( [ 4dwave1 ] ) . as a result ,
one arrives at the dual " fresnel equation imposed on the components @xmath72 of the ray vector : @xmath73 like the tr - tensor density ( [ g4 ] ) , the totally symmetric tensor density of weight @xmath74 is constructed in terms of the components of the constitutive tensor : @xmath75 as before , a complete symmetrization has to be performed over the four indices @xmath76 , with the vertical lines separating out those indices that are excluded from the symmetrization .
here we introduced a
double dual " of the constitutive tensor density as @xmath77 the contraction of this new object with the original constitutive tensor yields @xmath78 using ( [ chi - dec ] ) , we find the corresponding decomposition @xmath79}{}^m + \hat{\epsilon}_{ijkl}\,\alpha,\label{low - dec}\ ] ] substituting this into ( [ rg ] ) , we arrive at the skewon contribution to the ray surface : @xmath80
the main aim of this paper is to investigate the possible influence of the skewon field @xmath81 on the electromagnetic wave propagation . for simplicity , we assume that the principal part of the constitutive tensor is determined by the maxwell - lorentz law ( [ ml ] ) , that is , @xmath82 however , the skewon structure will be kept as general as possible . as a preliminary step
, we note that direct computations yield @xmath83 the skewon , as a traceless tensor field of type @xmath84 $ ] , has 15 independent components .
given the _
spacetimes metric _
@xmath22 , we can decompose the skewon field into an antisymmetric and a symmetric part .
the latter can be conveniently constructed from a pair of arbitrary ( co)vector fields @xmath85 and @xmath86 , so that we find eventually @xmath87 here @xmath88 is an arbitrary antisymmetric tensor and @xmath89 .
such a parametrization produces almost a general skewon field , since the number of independent components is here @xmath90 . substituting ( [ svec ] ) into ( [ propg8 ] ) and ( [ rayskew ] ) , and taking into account ( [ chi0 ] )
, we find @xmath91,\label{addtr1}\\ \label{addtr2 } { } ^{(1)}\widehat\chi_{\,m(i|n|j}\!\not\!s_k{}^m \!\not\!s_{l)}{}^n & = & -\,{\frac { \lambda_0}{\sqrt{-g}}}\left[\left(a_{m(i}a^m{}_j + 2w_{(i } \widehat{b}_j + v^2\,w_{(i}w_j\right)g_{kl ) } - v_{(i}v_jw_kw_{l)}\right].\end{aligned}\ ] ] here @xmath92 , @xmath93 and @xmath94 .
adding the principal terms ( [ tr1 ] ) to the contributions of the skewon ( [ addtr1 ] ) and ( [ addtr2 ] ) , we see that neither the wave surface nor the ray surface decompose into the product of cones in general . in other words ,
birefringence is absent .
the last terms in ( [ addtr1 ] ) and in ( [ addtr2 ] ) are responsible for that .
let us investigate those particular cases of the skewon field in the ` double vector ' representation ( [ svec ] ) which do admit the birefringence phenomenon .
when the skewon field is represented by its antisymmetric part only , i.e. , @xmath95 @xmath96 ( which is obtained by putting both @xmath97 ) , then the analysis of ( [ propg8 ] ) shows that the light propagation is birefringent with one light cone defined by the spacetime metric @xmath98 and the second optical " metric @xmath99 this result actually holds true for any other parametrizations of the skewon field which will be considered later . at the same time , inspection of ( [ rayskew ] ) yields the decomposition of the ray surface into the light cone of the spacetime metric @xmath22 and a second cone defined by @xmath100 although ( [ opt1 ] ) and ( [ opt2 ] ) look very similar , we notice that they are not mutual inverse , and hence , they do not determine a unique geometric structure on spacetime .
assume now that the skew - symmetric skewon part is absent , which is evidently achieved if @xmath101}$ ] .
then @xmath102 $ ] , @xmath103 $ ] , and we find the tr - tensor in the birefringent form @xmath104 here the two light cones are defined by the two optical metrics @xmath105.\label{ometpm}\ ] ] here @xmath106 . obviously , in order to have the wave propagation along separate light cones , the skewon field must satisfy the condition @xmath107 when this condition is violated , the fresnel wave surface is of 4th order and birefringence is absent . for the equality sign in ( [ u2 ] ) , the birefringence disappears and the two light cones coincide .
however , one can straightforwardly prove that for @xmath108 the optical metric has an euclidean signature .
hence there is no wave propagation in this case .
the type of the signature correlates with the sign of the determinant of the metric : the latter is negative for the lorentzian case and positive for the euclidean one .
a direct computation yields @xmath109\right\}.\label{det}\ ] ] comparing this with ( [ ometpm ] ) , we identify the coefficient as @xmath110/2\lambda_0 ^ 2 $ ] and finally obtain @xmath111.\label{detpm}\ ] ] since the spacetime metric is lorentzian , its determinant is negative , @xmath112 . as a result , we conclude that the optical metrics will be also lorentzian only if @xmath113 note that the two conditions ( [ u3 ] ) and ( [ u2 ] ) look pretty similar .
in fact , we can verify that only one of them is an essential inequality .
indeed , when the two vectors are either _ both _ timelike ( @xmath114 ) or _ both _ spacelike ( @xmath115 ) , the new condition ( [ u3 ] ) makes the constraint ( [ u2 ] ) redundant , whereas when at least one of the vectors is null , ( [ u3 ] ) and ( [ u2 ] ) coincide . finally ,
when one of the vectors is timelike and another spacelike , the only essential condition is ( [ u2 ] ) and ( [ u3 ] ) is redundant . with these conditions , we can classify the effects of the skewon field on wave propagation .
the spatially isotropic skewon field of nieves & pal @xcite , for example , belongs to the subcase when the two vectors @xmath116 and @xmath117 are proportional to each other . indeed ,
let us consider the ( flat ) minkowski spacetime with the lorentzian metric @xmath118 , and let us put @xmath119 with some scalar function @xmath52 of dimension @xmath120 $ ] .
then , from ( [ svec ] ) , we recover the isotropic skewon field @xmath121 as a result , the spacetime relation becomes @xmath122 accordingly , in the special case when skewon and axion become _ constant _ fields , one can speak of the 4 electromagnetic constants for a vacuum spacetime with spatial isotropy : the electric constant @xmath123 , the magnetic constant @xmath124 , the spatially isotropic part @xmath52 of the skewon @xmath81 , and the axion @xmath125 . in this special case
we evidently find @xmath126 .
hence ( [ detpm ] ) yields that both optical metrics have a wrong euclidean signature .
this is in complete agreement with our earlier observations @xcite in which the absence of wave propagation was reported on the background of the isotropic skewon field of nieves & pal .
however , now we obviously obtain a far more general result : whereas nieves & pal have considered only a very special configuration with the spacelike vectors @xmath117 and @xmath116 being proportional to each other , the wave propagation is actually also absent for any @xmath127 and any riemannian spacetime metric @xmath22 . then again @xmath128 and
hence @xmath129 .
it is impossible to construct a general parametrization of the skewon field by using covariant 4-dimensional objects .
instead , we have to switch to the space and time decomposition techniques and to allow for the use of 3-dimensional ( 3d ) tensor objects
. then we can write the required parametrization as follows : @xmath130 here @xmath131 and @xmath132 are 3d ( co)vector fields , whereas @xmath133 is a 3d tensor of type @xmath134 $ ] .
from now on , the indices from the beginning of the latin alphabet denote the spatial 3d components , i.e. , @xmath135 .
the count of the independent components , namely @xmath136 , shows that we indeed have the most general representation of the skewon field .
it is straightforward to establish the dimensions of the different pieces of skewon field : we find that @xmath137 = [ \lambda_0]$ ] , @xmath138 = [ \varepsilon_0]$ ] and @xmath139 = 1/[\mu_0]$ ] . with this general parametrization , the linear spacetime relation ( [ cr ] )
can be written symbolically as follows : @xmath140 here the linear operator on the right - hand side explicitly reads : @xmath141 now we again specialize to the maxwell - lorentz principal part ( [ chi0 ] ) .
moreover , since we are specifically interested in the skewon effects , we will assume for concreteness that the spacetime metric is the _
flat minkowskian _
one with lorentzian signature , @xmath142 .
the euclidean 3-metric tensor @xmath143 should not be confused with the kronecker delta : the components of the latter are always either 0 or 1 , whereas the former has these values in cartesian coordinates only .
then one can further decompose the 3d skewon tensor into its symmetric and skewsymmetric parts : @xmath144 here @xmath145 , and the 3d indices are raised and lowered by means of the euclidean 3d metric , i.e. , @xmath146 and @xmath147 .
the covector field @xmath148 describes the antisymmetric part of the 3d tensor skewon .
obviously , both irreducible parts of the tensor skewon have the same dimension @xmath149 = [ z_c ] = [ \lambda_0]$ ] .
the analysis of the wave propagation in the presence of the general skewon field ( [ sgen ] ) is not possible in the above covariant 4-dimensional framework .
since the space and time components are explicitly separated in the general parametrization ( [ sgen ] ) , we need a corresponding approach to the wave and ray surfaces in which a space and time decomposition has been performed . namely , let us denote the independent components of the tr - tensor ( [ g4 ] ) as follows : @xmath150\nonumber\\ & & \qquad + ( { \cal c}^d{}_c + { \cal d}_c{}^d)({\cal a}^{c(a}{\cal c}^{b)}{}_d + { \cal d}_d{}^{(a}{\cal a}^{b)c } ) - { \cal c}^d{}_d{\cal a}^{c(a}{\cal c}^{b ) } { } _ c\nonumber\\ & & \qquad - { \cal d}_c{}^{(a}{\cal a}^{b)c}{\cal d}_d{}^d - { \cal a}^{dc}{\cal c}^{(a}{}_c { \cal d}_d{}^{b ) } + \left({\cal a}^{(ab)}{\cal a}^{dc}- { \cal a}^{d(a}{\cal a}^{b)c}\right ) { \cal b}_{dc}\,,\label{ma2}\\ m^{abc } & : = & 4\,{\cal g}^{0abc } = \epsilon^{de(c|}\left[{\cal b}_{df } ( { \cal a}^{ab)}\,{\cal d}_e^{\ f } - { \cal d}_e^{\ a}{\cal a}^{b)f}\ , ) \right .
\nonumber \\ & & \qquad\left.+ { \cal b}_{fd}({\cal a}^{ab)}\,{\cal c}_{\ e}^f - { \cal a}^{f|a } { \cal c}_{\ e}^{b ) } ) + { \cal c}^{a}_{\ f } \,{\cal d}_e^{\ b)}\,{\cal d}_d^{\ f } + { \cal d}_f^{\ a } \,{\cal c}^{b)}_{\ e}\,{\cal c}^{f}_{\ d } \right ] \,,\label{ma3}\\ m^{abcd } & : = & { \cal g}^{abcd } = \epsilon^{ef(c}\epsilon^{|gh|d}\,{\cal b}_{hf } \left[\frac{1}{2 } \,{\cal a}^{ab)}\,{\cal b}_{ge } - { \cal c}^{a}_{\ e}\,{\cal d}_g^{\ b)}\right ] \,.\label{ma4}\end{aligned}\ ] ] then , in decomposed form , the generalized fresnel equation ( [ fresnel ] ) reads @xmath151 similarly to ( [ ma0])-([ma4 ] ) , we can decompose the 4-dimensional tensor density ( [ rg ] ) into a set of 3-dimensional objects : @xmath152 \nonumber\\ & & \qquad + ( { \cal c}^c{}_d + { \cal d}_d{}^c)({\cal b}_{c(a}{\cal d}_{b)}{}^d + { \cal c}^d{}_{(a}{\cal b}_{b)c } ) - { \cal d}_d{}^d{\cal b}_{c(a } { \cal d}_{b)}{}^c\nonumber\\ & & \qquad - { \cal c}^c{}_{(a}{\cal b}_{b)c}{\cal c}^d{}_d - { \cal b}_{dc } { \cal d}_{(a}{}^c { \cal c}^d{}_{b ) } + \left({\cal b}_{(ab)}{\cal b}_{dc } - { \cal b}_{d(a}{\cal b}_{b)c}\right){\cal a}^{dc}\,,\label{mh2}\\ \widehat{m}_{abc } & : = & 4\,\widehat{\cal g}_{0abc } = \hat{\epsilon}_{de(c } \left[{\cal a}^{df } ( { \cal b}_{ab)}\,{\cal c}^e{}_f - { \cal c}^e{}_a{\cal b}_{b)f}\ , ) \right .
\nonumber \\ & & \qquad\left.+ { \cal a}^{fd}({\cal b}_{ab)}\,{\cal d}_f{}^e - { \cal b}_{|f|a } { \cal d}_{b)}{}^e ) + { \cal d}_{a}{}^f\,{\cal c}^e{}_{b)}\,{\cal c}^d{}_f + { \cal c}^f{}_a\,{\cal d}_{b)}{}^e\,{\cal d}_f{}^d\right ] \,,\label{mh3}\\ \widehat{m}_{abcd } & : = & \widehat{\cal g}_{abcd } = \hat{\epsilon}_{ef(c } \hat{\epsilon}_{|gh|d}\,{\cal a}^{hf}\left[\frac{1}{2}\,{\cal b}_{ab ) } \,{\cal a}^{ge } - { \cal d}_a{}^e\,{\cal c}^g{}_{b)}\right]\,.\label{mh4}\end{aligned}\ ] ]
then the fresnel type equation for the quartic ray surface ( [ ray0 ] ) in decomposed form reads @xmath153 now we are in a position to study the effects induced by the different irreducible parts of the skewon field .
when only the @xmath154 part of the skewon is nontrivial , whereas @xmath155 and @xmath156 , the electric constitutive relation reads @xmath157 .
we will call this the skewonic electric faraday effect , since it describes the electric faraday effect in optically active media for purely imaginary @xmath154 @xcite . then , from ( [ ma0])-([ma4 ] ) , we find the coefficients of the fresnel equation as @xmath158 , @xmath159 , whereas the nontrivial components read @xmath160.\end{aligned}\ ] ] here , as usual , @xmath161 and @xmath162 . on the other hand , according to ( [ mh0])-([mh4 ] ) , the analysis of the ray surface yields @xmath163 , @xmath164 , and @xmath165 as a result , we obtain the following fresnel surfaces for the wave covectors and the ray vectors , respectively : @xmath166 + ( q^2)^2 & = & 0,\label{fqn}\\ \label{fsn } ( s_0c)^4 - 2(s_0c)^2\,s^2 + s^2\left[s^2 + ( \overline{n}s)^2\right ] & = & 0.\end{aligned}\ ] ] here we denote @xmath167 , @xmath168 , @xmath169 , @xmath170 . moreover , we use the dimensionless skewon vector @xmath171 . with the help of the nontrivial vector @xmath172 , we can split the wave covector @xmath173 into its transversal and longitudinal parts , with @xmath174 and @xmath175 . then the fresnel equation ( [ fqn ] ) is rewritten in the form @xmath176\left[(q_0/c)^2(1 + \overline{n}^2 ) - q_\bot^2\right ] + q_{||}^2\left[-2(q_0/c)^2 + 2q_\bot^2 + q_{||}^2\right ] = 0.\ ] ] accordingly , in the plane transversal to the skewon vector @xmath154 , we find birefringence with the minkowski metric and the second optical metric @xmath177
. however , the fresnel surface does not decompose into a product of the two light cones .
although , for @xmath178 , the wave covector surface is a sphere , it becomes a toroid already for a small positive @xmath179 .
the latter looks like a very thin and narrow belt for small values of @xmath180 .
however , it becomes a thick toroid with the inner radius decreasing to zero for large values of @xmath180 .
a typical fresnel surface for this case is depicted in fig .
[ efarfig1 ] .
physically , the hole of this toroid corresponds to the directions in which waves can not propagate : in spherical coordinates , the wave propagation occurs for @xmath181 , where the spherical angle @xmath182 is counted from the direction of the skewon vector @xmath154 and the limiting angle is determined from @xmath183 . as we see , the hole disappears when @xmath184 .
this qualitative characterization of the wave propagation can be completed by a numerical estimate for the velocity of light .
the phase velocity @xmath117 of the wave propagation is defined by the components of the wave covector according to @xmath185 , where @xmath186 are the components of the 3d unit covector . usually , the mean value of the velocity is of interest that is obtained after averaging over the directions of propagation and the polarizations . for the case under consideration ,
a direct calculation yields @xmath187 thus , since @xmath188 is positive , the velocity of the wave propagation turns out to be less than @xmath189 . making use of the data on the gamma - ray bursts and following @xcite
, we derive as an upper limit for the skewonic electric faraday effect @xmath190 .
the existence of a hole in the wave covector surface confirms our earlier conclusion about the dissipative nature of the skewon . in particular , when we continue to the purely imaginary values of @xmath191 so that the square becomes negative , @xmath192 , the topology of the fresnel surface changes drastically .
the hole disappears and the wave covector surface assumes the form of a spheroid around the origin .
more exactly , there is a single spheroid elongated in the direction of the skewon vector @xmath154 for @xmath193 , and two ( embedded one into another ) spheroids when @xmath194 .
the average radius of the inner spheroid in the latter case is smaller than 1 which , in physical terms , means that the velocity of wave propagation is greater than @xmath189 .
although the negative @xmath192 corresponds to the well known electric faraday effect in crystals , such a superluminal behavior of waves for @xmath194 was never reported in the literature , to the best of our knowledge . in fig .
[ efarfig2 ] and fig .
[ efarfig3 ] , we depict typical fresnel wave covector surfaces for @xmath195 and @xmath196 , respectively .
the wave surface for @xmath197 looks qualitatively like fig .
[ efarfig2 ] with the top and the bottom of the spheroid at @xmath198 . in the opposite case ,
when the @xmath131 part of the skewon is nontrivial , whereas @xmath199 and @xmath156 , the magnetic constitutive relation reads @xmath200 .
we call this case the skewonic magnetic faraday effect , since it corresponds to the magnetic faraday effect when @xmath131 is purely imaginary @xcite . from ( [ ma0])-([ma4 ] ) , in full analogy to the previous subsection , we find that @xmath158 , @xmath159 , whereas the nontrivial components are @xmath201 analogously , from ( [ mh0])-([mh4 ] ) , we find for the ray surface that @xmath163 , @xmath202 , and @xmath203.\end{aligned}\ ] ] here again @xmath204 and @xmath205 .
if , similarly to the previous subsection , we introduce the dimensionless skewon vector by @xmath206 , we end up with the following fresnel wave and ray surfaces , respectively : @xmath207 & = & 0,\\ ( s_0c)^4\,(1 + \overline{m}^2 ) - ( s_0c)^2\left[2s^2 + \overline{m}^2s^2 - ( \overline{m}s)^2\right ] + ( s^2)^2 & = & 0.\label{fsm}\end{aligned}\ ] ] similarly as before , @xmath208 and @xmath209 . repeating the calculation of the velocity of the wave propagation for this case
, we find the mean velocity @xmath210 . as a result , the analysis of the gamma - ray bursts data does not impose any limit on the pure skewon field of the magnetic faraday type . as one can see immediately , the wave covectors are only trivial in the plane transverse to the direction of the skewon vector , i.e. , when @xmath211 .
the waves propagate in that plane without birefringence along the light cone defined by the standard minkowski metric , since then @xmath212 ^ 2 = 0 $ ] is fulfilled .
in contrast , a purely imaginary skewon vector , with @xmath213 , yields more interesting results . in this case
, the mentioned circle represents a particular configuration that turns out to be the intersection curve of the two branches of the fresnel surface .
the typical fresnel wave covector surfaces for the cases @xmath214 and @xmath215 are depicted in fig .
[ mfarfig2 ] and fig .
[ mfarfig3 ] , respectively : both branches are hyperboloids or spheroids , intersecting in the @xmath216 plane .
the wave surface for @xmath217 is qualitatively different : it is represented by paraboloids as shown in fig .
[ mfarfig1 ] .
comparison of the results of the two last subsections demonstrates that the electric and magnetic faraday skewon effects are dual to each other in the following sense : when we interchange the wave and the ray vectors @xmath218 and simultaneously the skewon vectors @xmath219 , then the wave surface ( [ fqn ] ) has the same form as the ray surface ( [ fsm ] ) and vice versa : the ray surface ( [ fsn ] ) has the same form as the wave surface ( [ fqm ] ) . in this sense ,
[ mfarfig2]-[mfarfig1 ] show the ray ( wave ) surfaces dual to the respective wave ( ray ) surfaces depicted in figs .
[ efarfig1 ] and [ efarfig2 ] , with the mentioned replacement of @xmath219 and with the interchange between the dimensionless wave covector variables @xmath220 and the dimensionless ray vector variables @xmath221 .
let us now assume that only the tensor part of the skewon is present in ( [ sgen ] ) , whereas @xmath222 and @xmath199 .
we call this a skewon of the magneto - electric type , since it corresponds to the case of the natural optical activity in matter for the purely imaginary @xmath133 , see @xcite . furthermore
, we consider the effects of the symmetric and skew - symmetric parts ( [ 3tensor ] ) . in the absence of the skew - symmetric tensor part ( @xmath223 )
, we find @xmath158 , @xmath159 and @xmath224,\\ m^{ab } & = & \varepsilon_0\left[2\lambda_0 ^ 2\,\delta^{ab } + \left(2u^{ab } - \delta^{ab}u_d{}^d\right)u_c{}^c - u_e{}^au^{|e|b)}\right].\end{aligned}\ ] ] as a check of the consistency of our formalism , we can verify that the isotropic case @xmath225 reduces to the birefringent case ( [ gpm ] ) and ( [ ometpm ] ) with @xmath119 . as to the ray surface , a direct computation yields @xmath226 , @xmath202 and @xmath227 accordingly
, the form of the fresnel ray and wave surfaces turns out to be exactly the same .
a direct calculation of the mean velocity of the light propagation now yields @xmath228 as usual , we again introduce the dimensionless skewon variable @xmath229 .
then , similarly to the above cases , we find as upper value for the symmetric tensor skewon field from gamma - ray burst data @xcite @xmath230 .
the fresnel wave covector surface has now quite a different form as compared to the two cases above . for concreteness ,
let us take a tensor skewon with only one nontrivial component , @xmath231 .
then , for small values of the skewon @xmath232 , the fresnel wave surface is depicted in fig .
[ u08 ] . as a comment to this figure ,
let us recall that for a vanishing skewon @xmath233 we have a pure vacuum spacetime relation , and the fresnel wave surface is then a _ sphere_. with @xmath234 , this sphere degenerates to a pair of highly deformed intersecting toroids that , for extremely small values of @xmath235 , are nearly covering the sphere . but with growing @xmath235 , the toroids becomes thin and thinner . the intermediate situation is actually depicted in fig . [ u08 ] .
since we do not expect large skewon fields in general , we here limit ourselves to the case of a small @xmath235 . in the complementary situation ,
when @xmath222 , @xmath199 and the symmetric part is absent in ( [ 3tensor ] ) , @xmath236 , we obtain a particular case described in sec .
[ antis ] : birefringence with the minkowski light cone and the second optical metric @xmath237 unlike as with the symmetric tensor skewon , here the optical metric is always lorentzian with the determinant @xmath238 .
this means that the waves propagates along both light cones ( except for a skewon satisfying @xmath239 when the optical metric becomes degenerate ) .
a straightforward computation of the mean speed of the wave propagation now yields @xmath240 and , accordingly , the upper limit for the skewon field from the gamma - ray data is again @xmath241 .
in the previous section we studied the wave propagation on a vacuum spacetime described by the principal part @xmath49 of the constitutive tensor ( [ chi0 ] ) .
we found that the characteristic effect of the skewon field in that case was the emergence of holes in the wave covector surface . in physical terms
this means the complete damping of the wave propagation in certain directions . here
we will briefly study the wave propagation in an anisotropic dielectric medium and demonstrate that a similar effect occurs in the presence of a skewon field .
the constitutive relation for anisotropic matter is most easily formulated in terms of the @xmath39-matrix parametrization ( [ kappachi])-([cd - matrix0 ] ) . in the _ absence _ of a skewon field ,
an arbitrary dielectric medium is described by the spacetime relation @xmath242 the matrices @xmath243 and @xmath244 .
then , from ( [ ma0])-([ma4 ] ) , we find @xmath245 , @xmath246 , @xmath247 , @xmath248 , and @xmath249 as a result , the fresnel equation for the wave surface ( [ decomp ] ) can be recast into the simple form @xmath250 as in the previous section , @xmath167 . as to the ray surface
, we can immediately verify from ( [ mh0])-([mh4 ] ) that @xmath251 and @xmath252 , whereas @xmath253 , @xmath254 , and @xmath255 then , from ( [ decompray ] ) , one can straightforwardly derive the dual equation of the fresnel ray surface for the components of the ray vector @xmath256 : @xmath257 typical wave and ray surfaces are depicted in figs .
[ aniw ] and [ anir ] , respectively .
both surfaces consist of two non - intersecting branches touching each other exactly in 4 points .
when the skewon field is present , the mentioned picture becomes more nontrivial . in particular , for the case of the isotropic skewon ( [ pn ] ) ,
the constitutive relation is modified by adding to ( [ abani ] ) the nonvanishing @xmath39 matrices @xmath258 then , a quick computation shows that all the @xmath259-coefficients remain the same except that ( [ animab ] ) is replaced with @xmath260 here we introduced the dimensionless skewon variable @xmath261 .
analogously we can verify that the dual @xmath262-coefficients are the same as in the skewonless case except that ( [ anihmab ] ) is changed to @xmath263 then the equations of wave and ray surfaces can not be represented any longer in the compact form ( [ wavesurf ] ) and ( [ raysurf ] ) . the skewon affects the form of the surfaces by creating typical holes that correspond to the directions of wave covectors along which no wave propagation occurs . more specifically , the original two branches of the wave and ray surfaces , depicted in figs
. [ aniw],[anir ] , merge and form a single non - simply - connected surface . in the process of such a merging ,
the four points where the original two branches touched each other become the wormholes through which one can move from the outer surface into the inner one .
such a typical skewon effect is shown in fig .
[ holes ] .
the violation of the lorentz symmetry of spacetime results in spoiling " the ordinary light cone and , in particular , in the emergence of birefringence in light propagation .
thus , the investigation of electromagnetic wave phenomena provides an understanding of the lorentz violation in the photon sector of sme .
the birefringence type effects represent the observable consequences of the model and the corresponding measurements impose limits on the lorentz - violating parameters in the general linear spacetime relation .
it was shown previously @xcite by using cosmological and laboratory observations that the birefringence - related parameters of the principal part of the constitutive tensor must be smaller than @xmath264 . in this paper
we investigated the general covariant fresnel equation for the wave covectors as well as for the ray vectors in the case of a linear spacetime relation ( [ hchif ] ) that connects the electromagnetic excitation with the field strength . strictly speaking , the physical information contained in the ray vector surface
is the same as that in the wave covector surface .
the reason why we still analyzed the ray along with the wave surfaces is twofold .
firstly , we developed the four - dimensional covariant approach to the ray surfaces in complete analogy with the earlier results obtained for the fresnel wave surfaces .
such a formalism might be helpful in the study of the various aspects of the wave propagation in media , since the construction of the ray surfaces is quite a common tool in crystal optics , see @xcite .
secondly , we demonstrated explicitly that although the ray surface is dual to the wave surface ( which is also clearly seen from the figures above ) , the optical metric , which is naturally derived in the case of birefringence , is different for the wave and for the ray surfaces and the two metrics derived are _ not _ dual to each other .
the skewon part ( [ salpha])@xmath59 , ( [ chi2 ] ) of the constitutive tensor was in the center of our study .
we introduced its general parametrization ( [ sgen ] ) .
this enabled us to distiguish between three different types of effects : the electric and magnetic faraday effects and the ( magneto - electric ) optical activity .
the influence of a skewon field on the fresnel wave and ray surfaces is qualitatively different in these three case .
this is illustrated in the figures in the corresponding subsections of sec .
[ effects ] .
however , the characteristic sign of the skewon is the emergence of the specific _ holes in the fresnel surfaces _ that correspond to the directions in space along which the wave propagation is damped out completely .
this effect is in complete agreement with our earlier conclusion on the dissipative nature of the skewon field @xcite . besides the qualitative results ,
one may derive numerical limits on the magnitude of the skewon components from the experimental search for the anisotropy of the velocity of light . using the data for the gamma - ray bursts and following @xcite
, we then obtain the estimates @xmath265 and @xmath266 .
this shows that if the skewon indeed spoils the light cone structure ( yielding a violation of the lorentz symmetry and an anisotropy of the propagation of light ) , its influence should be extremely small . as a final remark
we note that the experimental laboratory data @xcite imposes much stronger constraints on the principal part of the constitutive tensor ( see the corresponding analysis in @xcite ) , and we can expect that the same will be true for the skewon too since the theoretical formalisms are pretty much the same in both cases . .
this work was supported by the deutsche forschungsgemeinschaft ( bonn ) with the grant he 528/20 - 1 .
the figures in this paper were constructed with the help of a mathematica program written by sergey tertychniy ( moscow ) .
we thank sergey also for many discussions . c. lmmerzahl , a. macias and h. mller , _ lorentz invariance violation and charge ( non-)conservation : a general theoretical frame for extensions of the maxwell equations _ , _ phys .
rev . _ * d71 * , 025007 ( 2005 ) 15 pages . g.f .
rubilar , yu.n .
obukhov , and f.w .
hehl , _ generally covariant fresnel equation and the emergence of the light cone structure in linear pre - metric electrodynamics _ , _ int
phys . _ * d11 * ( 2002 ) 1227 - 1242 .
hehl , yu.n .
obukhov , and g.f .
rubilar , _ on a possible new type of a t odd skewon field linked to electromagnetism _ , in : _
`` developments in mathematical and experimental physics '' , volume a : cosmology and gravitation _ , eds .
a. macias , f.j .
uribe , and e. diaz ( kluwer academic / plenum publishers : new york , 2002 ) pp . 241 - 256 [ arxiv.org/gr-qc/0203096 ] .
hehl , yu.n .
obukhov , and g.f .
rubilar , _ light propagation in generally covariant electrodynamics and the fresnel equation _ , _ int .
phys . _ * a17 * ( 2002 ) 2695 - 2700 . c. lmmerzahl , _ the geometry of matter fields .
_ in : v. de sabbata and j. audretsch ( eds . ) , _ quantum mechanics in curved space time _ , nato asi series , series b , physics vol
. 230 . plenum press , new york ( 1990 ) p. 23 .
obukhov and g.f .
rubilar , _ fresnel analysis of wave propagation in nonlinear electrodynamics _
* d66 * ( 2002 ) 024042 ( 11 pages ) .
rubilar , _ linear pre - metric electrodynamics and deduction of the lightcone _ , _ ph.d . thesis _
( university of cologne , june 2002 ) ; see _ ann .
( leipzig ) _ * 11 * ( 2002 ) 717 - 782 . | we start from a _ local _ and _ linear _ spacetime relation between the electromagnetic excitation and the field strength
. then we study the generally covariant fresnel surfaces for light rays and light waves .
the metric and the connection of spacetime are left unspecified .
accordingly , our framework is ideally suited for a search of possible violations of the lorentz symmetry in the photon sector of the extended standard model .
we discuss how the skewon part of the constitutive tensor , if suitably parametrized , influences the fresnel surfaces and disturbs the light cones of vacuum electrodynamics .
conditions are specified that yield the reduction of the original _ quartic _ fresnel surface to the double light cone structure ( birefringence ) and to the single light cone .
qualitatively , the effects of the real skewon field can be compared to those in absorbing material media .
in contrast , the imaginary skewon field can be interpreted in terms of non - absorbing media with natural optical activity and faraday effects .
the astrophysical data on gamma - ray bursts are used for deriving an upper limit for the magnitude of the skewon field .
= 10000 |
many of the exact results for effective actions in quantum field theory can be understood as integrated anomalous ward - identities which are protected from higher order corrections .
well known examples are the integration of the chiral ( weyl ) anomaly for massless fermions in two dimensions coupled to a gauge ( graviational ) field or the veneziano - yankielowicz superpotential in @xmath2 yang - mills theory @xcite .
the integration constant in turn reflects the choice of regularization and dynamical scale . in the latter example
this constant is fixed by a one - instanton calculation @xcite .
the purpose of the present paper is to understand the exact results for the low energy effective action of @xmath3 yang - mills theory @xcite as an integrated anomaly equation .
that the low energy effective action for @xmath4 yang - mills theory can be obtained by integrating the superconformal anomaly equation @xcite has been shown a while ago using a rather intricate series of arguments for gauge group @xmath5 for pure gauge theory in @xcite and including fundamental matter in @xcite .
for @xmath6 without matter this can be understood as follows : the superconformal anomaly equation together with the @xmath7 structure of the mass formula implies an ordinary second order differential equation for the first derivative of the prepotential @xmath8 for the massless @xmath3 vector multiplet @xmath9 .
what is more is that the parameters of this equation are completely fixed by the weak coupling asymptotics .
the integration of this equation then determines the prepotential @xmath8 uniquely . for higher rank gauge groups , however , the superconformal anomaly equation of @xmath3 yang - mills is not sufficient simply because it only provides a single equation for a rank @xmath10 number of low energy fields
. on the other hand , recent work pioneered by vafa and collaborators @xcite using results on geometric transitions in string theory has lead , among other results , to a new string theoretic derivation of the low energy effective action in @xmath3 yang - mills theory .
the starting point is a certain class of @xmath11 , @xmath2 theories obtained from @xmath3 yang - mills by adding a superpotential @xmath12 for the chiral multiplet that breaks the gauge symmetry @xmath13 .
this theory can be geometrically engineered via d - branes partially wrapped over certain cycles of a calabi - yau geometry . at low energies ,
the effective theory has a dual formulation where the branes are replaced by fluxes .
furthermore , the dual theory is described in terms of the gluino condensate @xmath14 which , together with the massless @xmath15 vector multiplets @xmath16 in the @xmath17 , form an @xmath3 vector multiplet on the dual calabi - yau geometry @xcite .
the holomorphic part of the effective action for this multiplet is that of an @xmath18 , @xmath3 yang - mills theory spontaneously broken to @xmath19 with a superpotential @xmath20 .
furthermore , this effective superpotential for the glueball superfield can be written as an integral of the holomorphic @xmath21-form over a blown up @xmath22 in the calabi - yau geometry , or equivalently , a period integral over a dual cycle in a genus @xmath23-riemann surface @xmath24 . in this formulation ,
the effective couplings of the @xmath15 vector multiplets appear directly as the period matrix of this riemann surface . upon scaling the classical superpotential @xmath12 to zero
one then recovers the @xmath4 theory at a point in the moduli space given by the minimum of @xmath12 .
although the @xmath14 vanish in the @xmath4 limit the structure of the riemann surface @xmath24 survives and the limit for the coupling @xmath25 for the massless @xmath15 s is smooth . in this paper we investigate how this structure can be seen to arise directly in the field theory formulation by integrating a suitable anomalous ward - identity .
this program has in fact been to a large extent completed in @xcite where a certain generalization of the konishi anomaly was used to establish that the aforementioned riemann surface @xmath24 arises directly in the field theory limit .
we should note that the prepotential for this theory has been understood a while ago via embedding in string theory @xcite and more recently through explicit multi - instanton computations to all orders @xcite . on the other hand
, the observation that the effective action can be understood as an integrated anomaly may put this result in its proper place within the exact results for effective actions and also removes some of the mystery surrounding this model . as we will explain
, it also implies that the multi - instanton calculus in this model is , in fact , equivalent to an anomalous ward - identity together with a one - instanton calculation and thus should elucidate the structure of the multi - instanton calculus .
the plan of this article is the following .
in section 2 we review the chiral ring relations of @xcite relevant for our program and discuss the consequences of these relations for instanton calculus in the @xmath26 theory .
in particular we argue in section 2.3 that a specific one - instanton calculation in the chiral ring together with an anomalous ward - identity completely determines the @xmath27-instanton contributions to the various quantities relevant in the @xmath26 theory .
then , in section 3 we recall first the constructions of the prepotential @xmath28 and the effective couplings @xmath29 for the massless @xmath15 vector multiplets in terms of the hessian of @xmath28 w.r.t .
the key difference with the effective prepotential @xmath8 for the @xmath4 theory on the coulomb branch is that while @xmath8 has an infinite expansion in @xmath30 and thus involves _ a priori _ an infinite number of instanton contributions , @xmath28 is a homogenous function of the `` glueball '' fields @xmath14 and the couplings @xmath31 of the superpotential . in particular , the role of the dynamical scale @xmath30 is reduced to setting the scale for the microscopic gauge coupling which can be evaluated in perturbation theory and a one - instanton calculation .
what remains to be shown is that the hessian of @xmath28 is given by the period matrix of the riemann surface @xmath24 .
we should note that while in the string theory description via geometric transition @xcite this is an immediate consequence of special geometry , the proof in the field theory description is more intricate . in sections 3.2 - 3.4
we will show explicitly that this relation holds in the field theory description .
this is the main technical result of this paper . for @xmath11 broken to @xmath32 and in the limit of vanishing superpotential , it allows us finally to express the low energy effective couplings for the massless @xmath15 vector multiplets in terms of the @xmath3
, @xmath11 casimirs @xmath33 and in this way we derive the low energy effective action @xcite for @xmath34 yang - mills theory .
this is done in section 3.5 .
hence , the low energy effective action has been shown to follow from integrating an anomalous ward - identity .
our starting point is @xmath3 @xmath11 yang - mills broken to an @xmath2 , @xmath32 theory . in @xmath2 superspace
this theory is described by a chiral multiplet @xmath35 and a vector multiplet @xmath36 .
we add to the corresponding action a classical superpotential @xmath37 for the chiral multiplet with @xmath38 as explained in @xcite an important role in the field theory derivation of the glueball potential is played by the ring of @xmath2 gauge - invariant chiral operators . since we will rely on these properties as well we present in this section a short summary of those that are relevant for us and discuss some consequences .
the chiral ring is made up of gauge invariant operators @xmath39 which are chiral , ie .
it follows immediately from this property that correlation functions of operators in the chiral ring in a supersymmetric vacuum ( @xmath41 ) are @xmath42-independent , ie .
@xmath43 an important consequence of ( [ one ] ) is the factorization property @xmath44 .
another property of chiral operators is that in a supersymmetric vacuum the expectation values of @xmath39 and @xmath45 are identical provided @xmath46 is some gauge invariant operator .
thus , we consider the chiral ring as being generated by the equivalence classes @xmath47 one can represent ( not necessarily gauge invariant ) chiral operators as the lowest components of chiral superfields @xmath48 ( since @xmath49 ) . then , using @xmath50 where @xmath51 is the super- and gauge covariant derivative , one finds that @xmath52 ( anti ) commutes with @xmath48 in the chiral ring .
taking this into account one sees that the @xmath19 chiral ring is generated by the elements @xmath53 in the rest of this paper certain relations between elements of the chiral ring will play a crucial role . in order to describe these relations for the complete set of elements simultaneously
it is useful to introduce a generating function for all elements of the chiral ring .
such a generating function is given by @xmath54 where @xmath55 and @xmath56 is a grassman - valued parameter .
the various elements of the chiral ring are then given by the coefficients in the expansion of @xmath57 in powers of @xmath56 and @xmath58 .
as explained in the introduction the basic idea we will employ is to integrate an anomalous ward identity in order to obtain the effective action for the massless degrees of freedom in the coulomb branch of @xmath4 yang - mills theory .
concretely this anomaly is manifested as a quantum correction to a classical relation in the chiral ring .
the relation we consider is simply a consequence of the equation of motion , @xmath59 , where @xmath12 is the classical superpotential .
since classically @xmath60 commutes with any chiral field we can multiply this equation by @xmath61 to get @xmath62 upon quantization , normal ordering effects have to be taken into account which lead to an anomaly in the above relation @xcite .
this anomaly can be determined , for instance , using pauli - villars regularisation ( e.g. @xcite ) . in perturbation
theory this one loop calculation is exact .
it was shown in @xcite that for the types of superpotentials considered here there are no non - perturbative corrections to the anomaly .
the anomalous relation is then given by @xcite @xmath63 \right]_{lk}\,,\ ] ] where @xmath64 are the entries of the matrix @xmath35 .
note that the above equation still holds if we replace @xmath52 by @xmath65 since @xmath56 ( anti ) commutes with everything . still following @xcite and using the identity @xmath66\right]_{lk}= \tr\left(\frac{\chi_1\chi_2}{z-\phi}\right ) \tr\left(\frac{\chi_1\chi_2}{z-\phi}\right)\,,\ ] ] valid in the chiral ring for anti commuting operators @xmath67 and @xmath68 and taking expectation values , we have @xmath69 finally , recalling the factorization properties of the chiral ring this leads to @xmath70 where @xmath71 is a polynomial of order @xmath72 in @xmath73 defined by @xmath74 the coefficients @xmath75 are thus given by @xmath76 where we have defined @xmath77 eqn .
( [ rq ] ) is then solved for @xmath78 as @xmath79 in the above expectation values it is understood that the high energy degrees of freedom have been integrated out while the low energy degrees of freedom play the role of background fields which are assumed to take values compatible with unbroken supersymmetry , so that the chiral ring properties are realized .
it has been shown in @xcite that each solution of the chiral ring relations corresponds to a supersymmetric vacuum of the gauge theory . in what follows
we consider the case where the @xmath11 gauge symmetry is broken down to @xmath32 .
in analogy with the case where the breaking is to a product of @xmath17 groups , we take the low energy fields to be the massless @xmath80 vector multiplets @xmath81 together with the glueball superfields @xmath14 with the condensate of the massless gluinos as their lowest components .
in particular , the chiral multiplet which acquires a mass due to the classical superpotential @xmath12 is integrated out .
the light degrees of freedom are then conveniently combined with the help of the auxiliary variable @xmath56 as @xmath82 where the contour @xmath83 is around the cut extending from the @xmath84-th minimum of the function @xmath85 which are assumed to be non - degenerate .
the details concerning the projection onto @xmath14 in terms of the contour integrals can be found in @xcite .
it follows from ( [ m1 ] ) that the expectation value @xmath78 depends on @xmath56 only implicitly through the @xmath86 .
furthermore , defining @xmath87 and using eq.([rl ] ) gives @xmath88 note also that @xmath86 can be considered as a function of the variables @xmath31 and @xmath89 and conversely @xmath90 .
the result ( [ rl ] ) , obtained in @xcite shows the appearance of a genus @xmath91 riemann surface @xmath24 defined by eq.([reim1 ] ) ( evaluated at @xmath92 ) as a consequence of the anomalous relation ( [ ar ] ) in the chiral ring .
the period matrix @xmath93 of @xmath24 will be identified with the low energy effective couplings @xmath94 of the massless fields in the coulomb branch of the @xmath4 theory . in order to establish this result
we first need to understand how @xmath95 enters in the effective action for the massless vector multiplets @xmath96 .
this is done in section 3 . as explained above the expectation values in eqn .
( [ rq ] ) stand for integrating out the high energy degrees of freedom in the presence of the `` external '' fields @xmath14 and @xmath96 .
however , another interpretation is possible in which all degrees of freedom are integrated out . in this case
the coefficients @xmath97 defined by eq.([fi ] ) are functions of @xmath31 only .
if the superpotential is such that the gauge symmetry is broken to @xmath32 , then there is no strong infrared dynamics such that the coefficients @xmath97 can be evaluated by treating the superpotential perturbatively in the @xmath26 theory .
in addition @xmath97 , or rather @xmath98 ( see eqn . ( [ tkdef ] ) ) is saturated by classical and instanton contributions .
this is because @xmath99 is annihilated by @xmath100 implying that the path integral localizes on just these configurations .
concretely we have for the @xmath101-instanton contribution @xmath102 with @xmath103 and subject to the constraints @xmath104 and @xmath105 .
these selection rules can be understood from symmetries @xcite or alternatively by looking at the details of the @xmath101-instanton calculation : since we are expanding about the theory with vanishing superpotential ( @xmath106 ) only non - negative powers of @xmath31 appear in the expansion yang - mills theory ( see @xcite and references therein ) where gauge and dilatation symmetry is spontaneously broken from the outset . ] . in order to get a non - vanishing contribution from an instanton with topological charge
@xmath101 we need to saturate the @xmath107 fermionic zero modes : two gluino zero modes appear in @xmath108 .
further zero modes appear by expanding in terms of the yukawa coulings @xmath109 .
it is then not hard to see that in order to produce @xmath110 gluino @xmath111 and `` quark ''
@xmath112 zero modes we need to have exactly one contribution @xmath113 from the potential @xmath114 .
this explains the first selection rule given above .
furthermore , the total number of scalars contained in @xmath99 and @xmath115 must match the @xmath116 scalars coming from the expansion in the yukawa couplings .
this then leads to the second selection rule stated above which now has the interpretation of a tree level amplitude in the presence of @xmath107 fermionic zero modes .
a consequence of these selection rules is that the @xmath99 s vanish for @xmath117 and @xmath118 is determined by a 1-instanton contribution . for @xmath119
the connected part of @xmath120 is still subject to the selection rules above and similarly for the connected part of @xmath121 , that is @xmath122 here the connected part of the expectation value is the part that can not be written as a product of expectation values with less than @xmath123 fields . in particular , for @xmath124 , @xmath33 is fully determined by its classical expression in terms of the @xmath31 s .
combining the selection rules for ( [ fkeval ] ) and ( [ ukeval ] ) with the chiral ring relation ( [ rl ] ) one then concludes that the expansions of @xmath99 and @xmath33 in the dynamical scale @xmath30 are completely determined by a single @xmath125-instanton calculation and the classical expressions for the @xmath33 , @xmath126 .
indeed , expanding both sides of eq.([rq ] ) in @xmath127 we obtain the recurrence relation @xmath128 where @xmath129 .
the recursion relation is then obtained by comparing equal powers in @xmath73 ( see also @xcite for a related discussion ) .
an important application of this discussion is that , of the coefficients @xmath97 , only @xmath130 is non - vanishing @xcite and is then equal to @xmath131 .
the factor @xmath132 can in principle be determined by evaluating @xmath118 in the explicit one - instanton computation .
instead we will determine @xmath132 below by comparison with the @xmath125-instanton contribution to the superconformal anomaly in the @xmath3 theory .
this , with the help of ( [ rl ] ) , fixes @xmath78 and , in particular , @xmath14 completely in terms of the @xmath31 and the one - instanton coefficient @xmath130 .
this suggests that the chiral ring relations , or equivalently the konishi anomaly imply a relation between instanton contributions of different topological charges ( see @xcite for other approaches to instanton recursion relations ) .
in order to determine the couplings of the massless u(1 ) vector multiplets @xmath133 , we will need to determine the holomorphic part of the effective action for these fields . in the @xmath26 theory this action is usually expressed in terms of a prepotential @xmath8 for the @xmath26 vector multiplet @xmath9 .
however , as explained in the introduction this prepotential involves an infinite number of instanton contributions .
the key observation @xcite allowing to circumvent this problem is that the low energy effective couplings can equally be obtained from an effective superpotential @xmath134 which in contrast to @xmath8 does not receive higher instantons contributions . in analogy with the usual effective action in field theory the holomorphic part of the effective action is given by the sum over 1pi graphs with @xmath14 and @xmath96 insertions .
in fact we can say more .
a `` non - renormalisation theorem '' given in @xcite shows that in perturbation theory the only contributions , compatible with the expected symmetries of the effective action come from planar graphs .
furthermore these planar graphs have either exactly one @xmath96 insertion at two of the index loops and one @xmath14 insertion at each of the remaining index loops or one @xmath14 insertion in all but one index loop . the index loop without @xmath14 insertion being proportional to @xmath135 , where @xmath136 counts the degeneracy of the vacuum corresponding to the @xmath137 minimum of the potential .
this observation then implies that the holomorphic part of the effective action for the low energy fields can be expressed in terms of a single function @xmath138 .
indeed let @xmath138 be the sum over all 1pi graphs with exactely one @xmath14 insertion at each of the index loops . then , since the effective potential is obtained by either replacing one @xmath14 by @xmath135 or two @xmath14 insertions by @xmath96 s , we have @xmath139 this latter relation can be written equivalently as @xmath140 where @xmath141 is obtained from @xmath138 simply by the substitution @xmath142 . in order to evaluate the coupling for the massless @xmath15 fields @xmath96 we then set the @xmath14 `` on shell '' .
there are two equivalent ways of implementing this . in the last section
we have seen that on shell the @xmath97 s are expressed in terms of @xmath31 . on the other hand , from the point of view of @xmath143
this means that @xmath144 so that the coupling matrix is @xmath145 however , to compare ( [ mat ] ) with the couplings for the massless @xmath15 vector multiplets in @xmath146 @xmath4 yang - mills we need to change variables from the @xmath147 to usual cartan - weyl basis for @xmath146 .
we will call these variables @xmath148 , @xmath149 .
they are defined by @xmath150 while @xmath151 . in particular
, @xmath152 vanishes on shell .
the couplings for the @xmath153 are then given by @xmath154 . in principle @xmath138
can now be computed directly order by order in perturbation theory is a homogenous function in @xmath31 and @xmath14 .
in particular the instanton contributions are summarized in the @xmath31-independent integration constant which is fixed by a one - instanton calculation .
see also @xcite for a discussion of the ambiguities in the field theory computation of @xmath28 . ]
@xcite ( or in an associated matrix model @xcite ) .
however , for the couplings , we only need its second derivative @xmath138 with respect to @xmath14 .
we will now show explicitely that the hessian of @xmath155 can be expressed in terms of the period matrix of the riemann surface described by ( [ reim1 ] ) .
this is the main technical result of the present paper .
first we recall a result in @xcite relating the derivative of @xmath138 with respect to @xmath31 to the expectation value of @xmath156 introduced in section [ scr ] . to start with we have from eq.([wphi ] ) @xmath157 on the other hand @xmath143 is expressed in terms of @xmath138 via eq.([fwpsi ] ) .
the claim is that this implies @xmath158 the identity ( [ peffphi ] ) follows from the observation that the right hand side depends on the background fields @xmath159 only in the combination @xmath86 . on the other hand
( [ fwpsi ] ) and ( [ wgk ] ) imply that the @xmath160 component of both sides of the equation ( [ peffphi ] ) agree . finally , by construction
both sides vanish at @xmath161 .
consequently the two functions are the same .
we will now show that @xmath162 is expressed in terms of the ( dual ) @xmath163-periods of the riemann surface @xmath24 defined by eq .
( [ reim1 ] ) . in the string theory description of this model
@xcite this property follows directly from the special geometry of cy spaces . in field theory
we are not aware of a previous derivation of this property ( see also comment at the end of section 4 in @xcite ) .
the strategy we use is to first compute @xmath164 and then to show that the result obtained can be integrated to obtain @xmath165 up to an integration `` constant '' independent of @xmath31 .
we start with eq .
( [ peffphi ] ) which gives for @xmath166 @xmath167\ , , \label{fresy}\end{aligned}\ ] ] where we have used eq.([eqlast ] ) , ( [ rl ] ) and @xmath168 . the next step is then to characterize completely @xmath169 .
as it is a mereomorphic @xmath125-form with a singular point of order @xmath170 at infinity , it can be written as @xcite @xmath171 with the following definitions : @xmath172 are canonically normalized holomorphic @xmath125-forms such that @xmath173 . @xmath174
behave at infinity like @xmath175 and satisfy - forms . ]
@xmath176 for all @xmath84 .
we have similarly @xmath177 and @xmath178 for @xmath179 .
let us now determine the coefficients @xmath180 , @xmath181 and @xmath182 .
the coefficient @xmath182 is obtained by integrating around the contour @xmath83 .
it follows from ( [ si ] ) , the previous definitions and from ( [ sal ] ) that @xmath183 . the coefficient @xmath180 is obtained by considering the behavior of @xmath184 at infinity .
more precisely , as @xmath185 we have @xmath186 for @xmath187 which leads to @xmath188 .
finally , we have @xmath189 as @xmath190 thus we have shown that [ sectiony ] @xmath191 we prove then in appendix @xmath192 that @xmath184 is homogenous of degree one in @xmath153 , @xmath193 ( or @xmath14 ) and @xmath31 . in particular
, we have @xmath194 after these preparations we consider @xmath195 .
we get then from equations ( [ fresy ] ) and ( [ deriofy ] ) @xmath196 in the next subsection we integrate this relation .
we will now show that the expression ( [ eq1 ] ) can be written as a derivative with respect to @xmath31 of an integral over the cycle @xmath163 dual to @xmath83 .
for this , we use the riemann bilinear relations associated with one form of first kind ( i.e. holomorphic ) , @xmath172 , and the other of second kind ( i.e. meromorphic with no residues ) , @xmath174 : @xmath197= 2 \pi i \left [ \res_p(\omega_k \xi_i ) + \res_{\tilde{p}}(\omega_k \xi_i ) \right ] .
\label{rb}\ ] ] let us recall briefly that such relations are obtained by computing in two different ways one particular integral @xcite , @xmath198 , where the riemann surface @xmath24 is thought here as a polygon with some identifications ( see figure [ figrblr ] ) .
then , the l.h.s .
corresponds to the computation of this integral on the contour shown on figure [ figrblr ] while the r.h.s . of eq.([rb ] ) is simply obtained by use of cauchy s formula where @xmath199 and @xmath200 have coordinates @xmath201 with @xmath202 . , height=113 ] however , for the l.h.s .
we use the definitions of @xmath172 and @xmath174 , which give respectively @xmath203 and @xmath204 . for the r.h.s .
, we have @xmath205 because both @xmath206 and @xmath172 change sign when going from @xmath199 to @xmath200 .
furthermore these residues are equal to @xmath207 as @xmath208 and @xmath209 . using the result ( [ deriofy ] )
we finally get @xmath210 this enables us to write eqn.([eq1 ] ) as @xmath211 which can be integrated w.r.t .
@xmath31 to end up with @xmath212 where @xmath213 is an integration constant .
the purpose of the next subsection is to fix this function _ via _ the superconformal anomaly . in order to determine
@xmath213 we need to know the @xmath31-independent part of @xmath214 .
this can be done with the help of the superconformal anomaly which can be computed in the microscopic @xmath11 theory using standard methods such as pauli - villars regularization .
this leads to @xmath215 where @xmath216 is the @xmath19 supercurrent and @xmath217 is the @xmath146 glueball field . more precisely , we have @xmath218 this anomaly ought to be reproduced by the low energy effective potential @xmath219
. the charges of @xmath220 and @xmath221 are @xmath21 , @xmath222 , and @xmath223 respectively .
we thus conclude that @xmath224 satisfies the equation @xmath225 where we have assumed @xmath226 and @xmath227 in the last equality to show that @xmath228 i.e. @xmath229 is a homogenous function of degree @xmath125 . ] . in order to see how this anomaly is reproduced in the low energy effective theory
we need to relate the dynamical scale @xmath230 of the low energy theory to the scale @xmath30 of the @xmath11 theory @xcite .
we have @xmath231 where @xmath232 are the roots of @xmath233 .
taking into account the dimensions of @xmath31 and @xmath232 it is then not hard to see that the anomaly ( [ w3 ] ) is reproduced by @xmath234 where @xmath235 is a homogeneous polynomial in @xmath31 and @xmath236 transforming with weight @xmath21 .
on the other hand , from ( [ fw ] ) we have @xmath237 let @xmath238 be the @xmath31-independent part of @xmath28 .
then , eq.([eqqq ] ) and eq.([apreqqq ] ) imply that @xmath239 where @xmath240 is an undetermined constant .
alternatively this gives for @xmath241 @xmath242 where we have used the symmetry under the permutation of the @xmath14 .
let us then compare this result with the @xmath31-independent term in @xmath243 .
this can be done using a scaling argument ( see also @xcite ) .
suppose we rescale all couplings @xmath31 by @xmath111 , so that @xmath244 , @xmath245 and consider @xmath246 which is thus given by @xmath247 if we then let @xmath111 go to infinity , @xmath248 goes to zero .
geometrically this limit corresponds to vanishing @xmath83 cycles since @xmath249 .
therefore we can rotate the two endpoints of the cut along one vanishing cycle @xmath83 .
under such a transformation , @xmath14 picks up a phase @xmath250 .
the transformation of the cycles can be worked out from figure [ mono ] : under a monodromy transformation @xmath251 all the @xmath163 change to @xmath252 while under a monodromy transformation @xmath253 , with @xmath254 , only @xmath163 is changed in @xmath255 . for @xmath256 [ mono],height=113 ] this implies the asymptotic form of the period integrals @xmath257 where @xmath258 is an invariant polynomial of @xmath259 and @xmath260 with no constant term as the period integrals vanish in the limit @xmath261 .
therefore , we get for large @xmath262 ( @xmath263 ) @xmath264 where @xmath265 is a @xmath31-independent linear combination in @xmath266 and @xmath267 since only the linear terms of @xmath268 contribute in the limit @xmath269 . on the other hand
we see from ( [ scales ] ) that rescaling of @xmath31 by @xmath111 corresponds simply to rescaling of @xmath230 so that we get @xmath270 where the dots refer to terms which depend on @xmath31 and where we have used the property that any choice of the basis of cycles @xmath83 is physically equivalent .
thus , the comparison of eq.([fbhi ] ) on one hand with eq.([gkindependent ] ) and eq.([in110 ] ) on the other hand give that the `` integration constant '' @xmath271 is linear or more precisely that @xmath272 the result ( [ eq53 ] ) enables us then by use of eq.([deriofy ] ) to obtain the low energy couplings @xmath273 for the vector multiplets in terms of the period matrix of the riemann surface @xmath24 defined by anomalous chiral ring relations ( [ reim1 ] ) , that is @xmath274 where @xmath275 denotes the value on shell .
in particular @xmath276 vanishes .
since the terms proportional to @xmath277 do not match the structure of the @xmath125-loop correction , we have @xmath278 .
furthermore , using the result @xmath279 on shell , @xmath24 is given by @xmath280 the constant @xmath132 will be fixed in the next subsection . to complete our derivation of the low energy couplings for the massless u(1 ) vector multiplets
@xmath133 we first need to express the period matrix in eq.([tauf ] ) in terms of the casimir variables @xmath281 , @xmath282 ) , of the @xmath3 theory @xcite and then fix the constant @xmath132 appearing in the equation for the riemann surface @xmath24 . for this purpose ,
if @xmath283 is the classical value of @xmath284 obtained by extremizing the tree - level superpotential , we write @xmath285 , where the coefficients of the polynomial @xmath286 depend on @xmath287 through the relation @xmath288 we know however from section 2.3 .
that there are no quantum corrections to the casimir variables i.e. that @xmath289 .
thus , @xmath290 and the equation ( [ tauf2 ] ) for @xmath24 becomes @xmath291 or equivalently @xmath292 the low energy couplings @xmath94 are thus given by the period matrix of the riemann surface @xmath24 defined by eq.([y ] ) .
furthermore , since the expectation values of the massless scalars are given by @xmath293 we get from eq.([eqlast ] ) together with eq.([y ] ) @xmath294 the two equations ( [ tauf ] ) and ( [ a ] ) then determine the prepotential @xmath295 via @xmath296 up to physically irrelevant integration constants @xcite .
finally let us fix the constant @xmath132 .
this can be done in several ways . here
we will fix @xmath132 by comparing the prediction for the observable @xmath297 with the explicit one - instanton calculation .
indeed , we have on one hand from the superconformal ward identity @xcite @xmath298 upon substitution of the asymptotic instanton expansion for @xmath299 this then leads to @xcite @xmath300 with @xcite @xmath301 . here
@xmath302 stands for the diagonal entries of @xmath303 .
eqn ( [ 58 ] ) is then a direct consequence of ( [ tauf ] ) and ( [ tauf2 ] ) . on the other hand , comparing ( [ 58 ] ) with the explicit one - instanton calculation @xcite implies then @xmath304 .
thus we have shown that the low energy couplings are indeed given by the period matrix of @xmath24 with @xmath304 .
in this article we have completed the field theory proof of the claim that the holomorphic effective action of @xmath3 @xmath146 yang - mills action can be obtained by integrating a suitable anomaly . in particular
, we have field the gap that arises in field theory ( not present in string theory ) due to the absence of special geometry relation between the periods of the riemann surface @xmath24 .
in addition we have pointed out that the anomalous chiral ring relations have important consequences for the multi - instanton calculus .
indeed , the instanton contributions to the expectation values of chiral ring elements are completely fixed by the @xmath125-instanton observable @xmath305 .
this also applies in a similar way to the expectation value @xmath306 . on the other hand ,
expanding the contour integral in @xmath30 one reads off the contribution of the @xmath27-instanton contribution to the function @xmath307 ( or equivalently @xmath308 ) .
this seems to suggest a recursive structure that would be interesting to investigate within the multi - instanton calculus .
* acknowledgments : * + it is a pleasure to thank andreas brandhuber , franois delduc , dominique lnge , steffen metzger , guiseppe policastro , christian rmelsberger , samson shatashvili and especially peter mayr for many helpful discussions .
the work of i.s . is supported by rtn under contract 005104 forcesuniverse , by dfg 1096 , string theory and by sfb 375 of the deutsche forschungsgesellschaft .
m.m . acknowledges support from the european network euclid - hprnc - ct-2002 - 00325 .
in this appendix , we prove that @xmath309 is homogeneous of degree 1 in @xmath14 and @xmath31 . for that purpose , we compute the various derivatives of @xmath184 and compare the results obtained with the expression ( [ whatisy ] ) of @xmath184 found in section [ sectiony ] , @xmath310 the general strategy in the computations that will be done is to identify the holomorphic contributions .
let us recall at this stage that the space of holomorphic forms on the riemann surface @xmath24 is @xmath311 dimensional and two special basis are @xmath312 and @xmath313 with @xmath179 and @xmath314 .
we change then the variables from @xmath315 to @xmath316 where @xmath317 is the coefficient of order @xmath84 of @xmath318 .
indeed , @xmath319 is holomorphic except for @xmath320 .
however , it is easy to prove that @xmath321 by computing the residue of @xmath184 at infinity as it follows from the definition of @xmath184 and the result ( [ newwhatisy ] ) that it is respectively equal to @xmath322 and to @xmath323 .
we first compute the derivative w.r.t .
we have successively : @xmath324 the first contribution in the r.h.s . of ( [ r3 ] )
is a linear combination of holomorphic terms while the second term is a meromorphic form that has residue @xmath125 at infinity .
consider then the case @xmath325 . in that case
, we also have from the definition ( [ si ] ) of @xmath14 @xmath326 for @xmath327 .
thus @xmath328 is a 1-form whose @xmath83 integrals for @xmath327 vanish and which has residue @xmath125 at infinity .
however , there is an unique form satisfying these properties and by definition it is @xmath329 .
thus , we have shown that @xmath330 take then @xmath331 . for @xmath332 ,
we get @xmath333 . thus , if we define the @xmath125-form @xmath334 we have on one hand @xmath335 for @xmath336 and on the other hand @xmath337 .
thus , for the same reason as above we have @xmath338 and so @xmath339 it is now easy to compute the derivatives w.r.t . @xmath153 and @xmath193 to get @xmath340
let us compute now the derivative w.r.t .
we have : @xmath341 suppose now that @xmath342 .
then , in eq.([derygk ] ) , the first term is equal to @xmath343 , the second term is holomorphic and the third term vanishes by use of eq.([derfn1 ] ) .
thus , we have @xmath344 . if @xmath345 , it is easy to see that we come to the same conclusion thanks again to eq.([derfn1 ] ) .
furthermore , @xmath346 and thus @xmath347 we conclude that @xmath184 is homogeneous of degree one in @xmath348 ( and thus in @xmath349 ) by comparing eq.([newwhatisy ] ) with the results ( [ derai ] ) , ( [ derplus ] ) and ( [ dergk ] ) .
novikov , m.a .
shifman , a.i .
vainshtein and v.i .
zakharov , nucl .
b * 260 * ( 1985 ) , 157 ( yad . fiz .
42 , 1499 ) ; m.a .
shifman and a.i .
vainshtein , nucl .
b * 296 * , ( 1988 ) , 445 ( sov . phys .
jetp 66 , 1100 ) ; d. amati , g.c . rossi and g. veneziano , nucl . phys .
b * 249 * ( 1985 ) , 1 ; j. fuchs and m.g .
schmidt , z. phys .
* c30 * ( 1986 ) , 161 ; d. amati , k. konishi , y. meurice , g.c . rossi and g. veneziano , phys . rep . * 162 * ( 1988 ) 169 .
n. seiberg and e. witten , nucl .
b * 426 * ( 1994 ) , 19 [ arxiv : hep - th/9407087 ] ; erratum - ibid , b * 430 * ( 1994 ) 485 ; n. seiberg and e. witten , nucl .
b * 431 * ( 1994 ) , 484 [ arxiv : hep - th/9408099 ] ; a. klemm , w. lerche , s. theisen and s. yankielowicz , phys . lett .
b * 344 * ( 1995 ) 169 [ arxiv : hep - th/9411048 ] ; a. klemm , w. lerche and s.theisen , int .
a11 * ( 1996 ) , 1929 [ arxiv : hep - th/9505150 ] ; p.
c. argyres , a. e. faraggi , phys .
74 ( 1995 ) , 3931 [ arxiv : hep - th/9411057 ] .
g. bonelli , m. matone and m. tonin , phys .
d * 55 * ( 1997 ) 6466 [ arxiv : hep - th/9610026 ] ; r. flume , m. magro , l.oraifeartaigh , i. sachs and o. schnetz , nucl .
b * 494 * ( 1997 ) , 331 [ arxiv : hep - th/9611123 ] .
f. cachazo , m.r .
douglas , n. seiberg and e. witten , jhep * 0212 * ( 2002 ) , 071 [ arxiv : hep - th/0211170 ] .
s. kachru , a. klemm , w. lerche , p. mayr and c. vafa , nucl . phys .
* b459 * ( 1996 ) , 537 [ arxiv : hep - th/9508155 ] .
s. katz , a. klemm and c. vafa , nucl .
b * 497 * ( 1997 ) 173 [ arxiv : hep - th/9609239 ] .
a. klemm , w. lerche , p. mayr , c. vafa and n. p. warner , nucl . phys .
* b477 * ( 1996 ) , 746 [ arxiv : hep - th/9604034 ] ; e. witten , nucl . phys .
b * 500 * ( 1997 ) 3 [ arxiv : hep - th/9703166 ] . n. nekrasov , adv .
* 7 * ( 2004 ) 831 [ arxiv : hep - th/0206161 ] ; n. nekrasov and s. shadchin , commun .
* 252 * ( 2004 ) 359 ; [ arxiv : hep - th/0404225 ] ; w. moore , n. nekrasov and s. shatashvili , commun .
* 209 * ( 2000 ) 77 [ arxiv : hep - th/9803265 ] .
k. intriligator , p. kraus , a. v. ryzhov , m. shigemori and c. vafa , nucl .
b * 682 * ( 2004 ) 45 [ arxiv : hep - th/0311181 ] .
n. dorey , v.v .
khoze and m.p .
mattis , phys .
* d54 * ( 1996 ) , 2921 [ arxiv : hep - th/9603136 ] ; n. dorey , t.j .
hollowood , v.v .
khoze and m.p .
mattis , phys .
* 371 * ( 2002 ) , 231 [ arxiv : hep - th/0206063 ] .
m. matone , phys .
b * 357 * ( 1995 ) 342 [ arxiv : hep - th/9506102 ] .
j. d. edelstein , m. marino and j. mas , nucl .
b * 541 * ( 1999 ) 671 [ arxiv : hep - th/9805172 ] .
g. bertoldi , s. bolognesi , m. matone , l. mazzucato and y. nakayama , jhep * 0405 * ( 2004 ) 075 [ arxiv : hep - th/0405117 ] .
r. dijkgraaf , m.t .
grisaru , c.s .
lam , c. vafa and d. zanon , phys .
b * 573 * ( 2003 ) , 138 [ arxiv : hep - th/0211017 ] .
j. sonnenschein , s. theisen and s. yankielowicz , phys .
b * 367 * ( 1996 ) 145 [ arxiv : hep - th/9510129 ] .
k. ito and n. sasakura , mod .
a * 12 * ( 1997 ) 205 [ arxiv : hep - th/9609104 ] . k. ito and n. sasakura , phys .
b * 382 * ( 1996 ) 95 [ arxiv : hep - th/9602073 ] ; e. dhoker , i. m. krichever and d. h. phong , nucl .
b * 489 * ( 1997 ) 179 [ arxiv : hep - th/9609041 ] .
f. fucito and g. travaglini , phys.rev .
* d55 * ( 1997 ) , 1099 [ arxiv : hep - th/9605215 ] . | based on chiral ring relations and anomalies , as described by cachazo , douglas , seiberg and witten , we argue that the holomorphic effective action in n=2 yang - mills theory can be understood as an integrated u(1 ) anomaly from a purely field theory point of view .
in particular , we show that the periods of the riemann surface arising from the generalized konishi anomaly can be given a physical interpretation without referring to special geometry .
we also discuss consequences for the multi - instanton calculus in n=2 yang - mills theory .
lmu - asc 25/05 + * low energy effective action in n=2 yang - mills as an integrated anomaly * 0.5truecm @xmath0 and i. sachs@xmath1 + 0.5 cm @xmath0 laboratoire de physique , + cole normale suprieure de lyon , + 46 alle ditalie , 69364 lyon cedex 07 , france + [email protected] + 0.2 cm @xmath1 arnold sommerfeld center , department fr physik + ludwig - maximilians universitt , + theresienstrasse 37 , d-80333 , mnchen , germany + [email protected] |
in recent years , @xmath2 dimensional , banados - teitelboim - zanelli ( btz ) @xcite black hole has attracted much attention .
one of the basic reasons for this attraction is that the btz black hole has a relatively simple tractable mathematical structure so that it provides a better understanding of investigating the general aspects of black hole physics since the btz black hole carries all the characteristic features such as the event horizon and hawking radiation , it can be treated as a real black hole . another motivation to study
btz black hole is the ads / cft correspondence which relates thermal properties of black holes in the ads space to a dual cft . in view of these points , the unresolved black hole properties belonging to @xmath3 or higher dimensional black holes at the quantum level ,
make the btz black hole an excellent background for exploring the black hole physics .
another interesting subject is the study of naked singularities that can be considered as a threat to the cosmic censorship hypothesis . compared to the black holes ,
the naked singularities are less understood . today , there is no common consensus either on the structure or the existence of the naked singularities .
recently , pitelli and letelier ( pl ) @xcite have analysed the occurrence of naked singularities for the btz spacetime from quantum mechanical point of view . in their analysis ,
the criteria proposed by horowitz and marolf ( hm ) @xcite is used .
the classical naked singularity is studied with the quantum test particles that obey klein - gordon and dirac equations .
they confirmed that the naked singularity is `` healed '' when tested by massless scalar particles or fermions without introducing extra boundary conditions .
however , for massive scalar particles additional information is needed .
despite the recent developments on the concept of quantum singularities hk , our understanding of naked singularities as far as quantum gravity concerned is still far from being complete .
the purpose of this paper is to analyse the naked singularities within the context of the quantum mechanics that form in the matter coupled @xmath4 dimensional black hole spacetimes .
our motivation here is to investigate the effect of the matter fields on the quantum singularity structure of the btz spacetime because the surface at @xmath1 for the btz black hole is not a curvature singularity , but is a singularity in the causal structure .
this situation changes when a matter field is coupled .
this is precisely the case that we shall elaborate on in this article . for this purpose
we consider the charged btz spacetime both in linear and non - linear electrodynamics .
this is analogous to a kind of einstein - maxwell extension of the work presented in @xcite .
furthermore , we extend the analysis to cover @xmath4 dimensional einstein - maxwell - dilaton coupled black hole spacetime .
the presence of charge both in linear and non - linear case and also the dilaton field modify the resulting spacetime geometry significantly . near the origin , the spacetime is not conic and true curvature singularity develops at @xmath1 .
consequently , the spacetime geometry that we have investigated in this study differs when compared with the case considered in @xcite .
the plan of the paper is as follows . in section
ii , we first review the definition of quantum singularities for general static spacetimes . in section iii ,
we consider the charged btz black hole in non - linear electrodynamics .
klein - gordon and dirac fields are used to test the quantum singularity .
we also discuss the sobolev norm which is used first time in this context by ishibashi and hosoya @xcite . in sections
iv and v , we consider charged btz in linear electrodynamics and dilaton coupled @xmath0 black hole spacetime in the einstein - maxwell and einstein - maxwell - dilaton theory , respectively .
dirac and scalar fields are used to judge the quantum singularity .
the paper ends with a conclusion in section vi .
in classical general relativity , the spacetime is said to be singular if the evolution of timelike or null geodesics is not defined after a proper time . horowitz and marolf , based on the pioneering work of wald @xcite ,
have proposed the criteria to test the classical singularities with quantum test particles that obey the klein - gordon equation for static spacetime having timelike singularities . according to this criteria ,
the singular character of the spacetime is defined as the ambiguity in the evolution of the wave functions .
that is to say , the singular character is determined in terms of the ambiguity when attempting to find self - adjoint extension of the operator to the entire space .
if the extension is unique , it is said that the space is quantum mechanically regular .
the brief review is as follows : consider a static spacetime @xmath5 with a timelike killing vector field @xmath6 .
let @xmath7 denote the killing parameter and @xmath8 denote a static slice.the klein - gordon equation on this space is @xmath9 this equation can be written in the form of @xmath10 in which @xmath11 and @xmath12 is the spatial covariant derivative on @xmath13 .
the hilbert space @xmath14 is the space of square integrable functions on @xmath13 .
the domain of the operator @xmath15 @xmath16 is taken in such a way that it does not enclose the spacetime singularities .
an appropriate set is @xmath17 , the set of smooth functions with compact support on @xmath13 .
operator @xmath18 is real , positive and symmetric therefore its self - adjoint extensions always exist .
if it has a unique extension @xmath19 then @xmath18 is called essentially self - adjoint @xcite . accordingly , the klein - gordon equation for a free particle satisfies @xmath20 with the solution @xmath21 \psi\left ( 0\right ) .\ ] ] if @xmath18 is not essentially self - adjoint , the future time evolution of the wave function ( eq . ( 4 ) ) is ambiguous .
then , horowitz and marolf define the spacetime quantum mechanically singular .
however , if there is only one self - adjoint extension , the operator @xmath18 is said to be essentially self - adjoint and the quantum evolution described by eq.(4 ) is uniquely determined by the initial conditions . according to the horowitz and marolf criterion
, this spacetime is said to be quantum mechanically nonsingular . in order to determine the number of self - adjoint extensions , the concept of deficiency indices is used .
the deficiency subspaces @xmath22 are defined by ( see ref .
@xcite for a detailed mathematical background ) , @xmath23 the dimensions @xmath24 are the deficiency indices of the operator @xmath18
. the indices @xmath25 are completely independent of the choice of @xmath26 depending only on whether @xmath27 lies in the upper ( lower ) half complex plane .
generally one takes @xmath28 and @xmath29 , where @xmath30 is an arbitrary positive constant necessary for dimensional reasons .
the determination of deficiency indices then reduces to counting the number of solutions of @xmath31 ; ( for @xmath32 ) , @xmath33 that belong to the hilbert space @xmath34 . if there is no square integrable solutions ( i.e. @xmath35 , the operator a possesses a unique self - adjoint extension and it is essentially self - adjoint .
consequently , a sufficient condition for the operator @xmath18 to be essentially self - adjoint is to investigate the solutions satisfying eq .
( 6 ) that do not belong to the hilbert space .
the action describing @xmath36 - dimensional einstein theory coupled with non - linear electrodynamics is given by @xcite , @xmath37 the field equations via variational principle read as , @xmath38 @xmath39 @xmath40 in which @xmath41 stands for the derivative of @xmath42 with respect to @xmath43 .
the non - linear field is chosen so that the energy momentum tensor ( 9 ) has a vanishing trace .
the trace of the tensor gives , @xmath44 hence , to have a vanishing trace , the electromagnetic lagrangian is obtained as @xmath45 where @xmath46 is an integration constant . with reference to the paper @xcite , the complete solution to the above action is given by the metric , @xmath47 where the metric function @xmath48 is given by , @xmath49 here @xmath50 is the mass , @xmath51 the case @xmath52 @xmath53 , that corresponds with an asymptotically de - sitter ( anti de - sitter ) spacetime , and @xmath54 is the electric charge .
this metric represents the btz spacetime in non - linear electrodynamics .
if @xmath55 , we have an asymptotically flat solution coupled with coulomb - like field . the kretschmann scalar which indicates the occurrence of curvature singularity is given by , @xmath56 in which @xmath57
it is clear that @xmath1 is a typical central curvature singularity . according to the values of @xmath58 , @xmath59 and @xmath54
, this singularity may be clothed by single or double horizons .
( see the reference @xcite for details ) .
however , for specific values of @xmath58 , @xmath59 and @xmath54 the central curvature singularity becomes naked and it deserves to be investigated within the framework of quantum mechanics . to find the condition for naked singularities the metric function is written in the following form , @xmath60 where @xmath61 and @xmath62 since the range of coordinate @xmath63 varies from 0 to infinity , the negative root will indicate the condition for a naked singularity . in order to find the roots
, we set @xmath64 which yields @xmath65 the standard procedure is followed for a solution via a new variable defined by @xmath66 that transforms the equation to @xmath67 this equation can be solved easily and the final answer is @xmath68 in which @xmath69 with a constraint condition @xmath70 after some algebra , we end up with the following equation @xmath71 where @xmath72 and @xmath73 it can be verified easily that the expression inside the curly bracket in eq . ( 18 ) is always positive .
hence , the only possibility for a negative root is @xmath74 .
this implies @xmath75 therefore , the condition @xmath76 is imposed from the constraint condition . as a result , for a naked singularity , @xmath77 or @xmath78 should be satisfied .
our aim now is to investigate the quantum singularity structure of the naked singularity that may arise if the constant coefficients satisfy @xmath79 using seperation of variables , @xmath80 , we obtain the radial portion of eq .
( 6 ) as @xmath81 where a prime denotes derivative with respect to @xmath63 .
the coulomb - like field in metric function ( 14 ) becomes negligibly small and hence , the metric takes the form @xmath83 this particular case overlaps with the results already reported in @xcite .
hence , no new result arises for this particular case .
this is expected because the effect of source term vanishes for large values of @xmath63 .
the case near origin is topologically different compared to the analysis reported in @xcite . here
, the spacetime is not conic .
the approximate metric near origin is given by , @xmath85 this metric can also be interpreted as the @xmath4 dimensional topological schwarzchild - like black hole geometry . for the solution of the radial equation ( 19 ) , we assume that a massless case ( i.e. @xmath86 ) , and ignore the term @xmath87 ( since it is negligible near the origin )
. then it takes the form , @xmath88 whose solution is @xmath89 where @xmath90 and @xmath91 are the first and second kind modified bessel functions and @xmath92 the behaviour of the modified bessel functions for real @xmath93 as @xmath94 are given by ; @xmath95 , \text { \ \ \ \ } \nu=0}{\frac{\gamma\left ( \nu\right ) } { 2}% \left ( \frac{2}{x}\right ) ^{\nu},\text { \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ } % \nu\text{\ } \neq0 } , \notag\end{aligned}\ ] ] thus @xmath96 and @xmath97 @xmath98 checking for the square integrability of the solution ( 23 ) requires the behaviour of the integral for @xmath90 @xmath99 and @xmath100 which are both convergent as @xmath94 .
any linear combination is also square integrable .
it follows the solution ( 23 ) belonging to the hilbert space @xmath34 and therefore the operator @xmath18 described in eq.(6 ) is not essentially self - adjoint .
so , the naked singularity at @xmath1 is quantum mechanically singular if it is probed with quantum particles .
another approach to remove the quantum singularity is to choose the function space to be the sobolev space @xmath101 which is used first time in this context by ishibashi and hosoya @xcite . here
, the function space is defined by @xmath102 where the norm defined in @xmath4 dimensional geometry as , @xmath103 which involves both the wave function and its derivative to be square integrable .
the failure in the square integrability indicates that the operator @xmath18 is essentially self - adjoint and thus , the spacetime is `` wave regular '' . according to this norm , the first integral is square integrable while the second integral behaves for the functions @xmath90 as @xmath104 and @xmath91 integral vanishes . as a result ,
the wave functions are square integrable and thus the spacetime is quantum mechanically wave singular .
it should be noted that the sobolev space is not the natural quantum mechanical hilbert space .
we apply the same methodology as in @xcite for finding a solution to dirac equation .
since the fermions have only one spin polarization in @xmath4 dimensions @xcite , dirac matrices are reduced to pauli matrices @xcite so that , @xmath105 where latin indices represent internal ( local ) indices . in this way , @xmath106 where @xmath107 is the minkowski metric in @xmath4dimensions and @xmath108 is the identity matrix .
the coordinate dependent metric tensor @xmath109 and matrices @xmath110 are related to the triads @xmath111 by @xmath112 where @xmath113 and @xmath114 are the external ( global ) indices .
the dirac equation in @xmath4 dimensional curved spacetime for a free particle with mass @xmath115 becomes @xmath116 \psi\left ( x\right ) = m\psi\left ( x\right ) , \ ] ] where @xmath117 is the spinorial affine connection and is given by @xmath118 s^{\lambda\nu}(x),\ ] ] @xmath119 .\ ] ] the causal structure of the spacetime indicates that there are two singular cases to be investigated . the asymptotic case , @xmath120 has already been analysed by pl .
the case of @xmath94 is not conical so there is a topological difference in the spacetime near @xmath1 .
hence , the suitable triads for the metric ( 21 ) are given by , @xmath121 the coordinate dependent gamma matrices and the spinorial affine connection are given by @xmath122 now , for the spinor @xmath123 the dirac equation can be written as @xmath124 the following ansatz will be employed for the positive frequency solutions : @xmath125 the radial parts of the dirac equation for investigating the behaviour as @xmath126 , are @xmath127 where @xmath128 @xmath129 , @xmath130 and @xmath131 then , the solutions are given by ;
@xmath132 where @xmath133 , @xmath134and @xmath135 .
when we look for the square integrability of the above solutions , we obtained that both functions @xmath136 and @xmath137 are square integrable near @xmath138 ( or @xmath1 ) for both @xmath139 and @xmath140 .
one has,@xmath141 ^{2}d\rho<\infty,\ ] ] and @xmath142 ^{2}d\rho<\infty.\ ] ] we note that these results are verified first by expanding the whittaker functions in series form up to the order of @xmath143 and then by integrating term by term in the limit as @xmath94 .
@xmath144the set of solutions for the dirac equation for the spacetime ( 21 ) is given by @xmath145 and an arbitrary wave packet can be written as @xmath146 where @xmath147 is an arbitrary constant . hence , initial condition @xmath148 is sufficient to determine the future time evolution of the particle .
the spacetime is then quantum regular when tested by fermions .
the metric for the charged btz spacetime in linear electrodynamics is given by @xcite , @xmath47 with the metric function @xmath149 where @xmath54 is the electric charge and @xmath50 is the mass and @xmath150 .
the kretschmann scalar is given by , @xmath151 which displays a power - law central curvature singularity at @xmath1 .
according to the values of @xmath59 , @xmath152 and @xmath153 this central singularity is clothed by horizons or it remains naked .
our interest here is to investigate the quantum mechanical behaviour of the naked singularity . in order to find the condition for naked singularity , we set @xmath154 and the solution for @xmath155 is @xmath156 in which represents the lambert function @xcite.fig.1 displays ( unmarked region ) the possible values of @xmath59 and @xmath157 that result in naked singularity .
the causal structure is similar to the case considered in the previous section .
there are two singular cases to be investigated .
the case for @xmath158 is approximately the same case considered in @xcite .
therefore , the results reported by pl are valid for this case as well . for small @xmath63 values
, the approximate metric can be written in the following form @xmath159 in which @xmath160 the radial equation for the metric ( 44 ) is obtained for the massless case as , @xmath161 since @xmath162 , the solution can be written in terms of zeroth order first and second kind modified bessel functions , @xmath163 where @xmath164 .
as @xmath165 , @xmath166 the behaviour of the modified bessel functions for @xmath167 are @xmath168 and @xmath169 these functions are always square integrable for @xmath170 that is @xmath171 these results indicate that charged btz black hole in linear electrodynamics is quantum mechanically singular when probed with quantum test particles that obey klein - gordon equation .
if we use the sobolev norm ( 25 ) , the second integral which involves the derivative of the wave function @xmath172 becomes @xmath173 numerical integration has revealed that as @xmath170 @xmath174 on the other hand for the wave function @xmath175 , the second integral in the sobolev norm is solved numerically as @xmath176 @xmath177 @xmath178 which is square integrable . as a result , charged coupled btz black hole in linear electrodynamics is quantum mechanically wave regular if and only if the arbitrary constant parameter is @xmath179 in eq.(46 ) .
consequently , if the naked singularity both in linear and non - linear electrodynamics is probed with quantum test particles , the following results are obtained : \1 ) in classical point of view , the kretschmann scalar in non - linear case diverges faster than in the linear case .
\2 ) in quantum mechanical point of view , if the chosen function space is sobolev space , spacetime remains singular for non - linear case , but the spacetime can be made wave regular for linear case . from these results
we may conclude that the structure of the naked singularity in the non - linear electrodynamics is deeper rooted than the singularity in the linear case . the effect of the charge when @xmath120 does not contribute as much as the term that contains the cosmological constant .
therefore , we ignore the mass and the charged terms in the metric function ( 42 ) .
this particular case has already been analysed in @xcite .
the contribution of the charge is dominant when @xmath180 the dirac equation for the metric ( 44 ) is solved by using the same method demonstrated in the previous section .
we obtain the radial equation in the limit @xmath94 as @xmath181 whose solution is given by @xmath182 where @xmath183 and @xmath184 are arbitrary constants .
the solution given in eq.(48 ) is square integrable .
the arbitrary wave packet can be written as , @xmath185 thus , the spacetime is quantum mechanically regular when probed with fermions . in this section
, we consider @xmath0 black holes described by the einstein - maxwell - dilaton action , @xmath186 where @xmath187 is the ricci scalar , @xmath188 is the dilaton field , @xmath189 is the maxwell field and @xmath58 , @xmath190 , @xmath191 , and @xmath192 are arbitrary couplings .
the general solution to this action is given by @xcite , @xmath193 where @xmath194 here , @xmath18 is a constant of integration which is proportional to the quasilocal mass @xmath195 , @xmath196 is an integration constant and @xmath197 is the charge .
the dilaton field is given by @xmath198 in which @xmath199 is a @xmath196 related constant parameter .
note that , the above solution for @xmath200 contains both the vacuum btz metric if one takes @xmath201 and the btz black hole if @xmath202 .
however , if the constant parameters are chosen appropriately , the resulting metric represents black hole solutions with prescribed properties .
for example , when @xmath203 , @xmath204 the metric function given in equation ( 52 ) becomes @xmath205 and therefore the corresponding metric is @xmath206 where @xmath207 is a constant parameter .
the kretschmann scalar for this solution is given by @xmath208 + 40r^{\frac{1}{3}}q^{2}\left [ 2\left ( 5q^{2}-mr^{\frac{2}{3}}\right ) -5\lambda r^{2}\right ] \right\ } } { % 81\alpha^{2}r^{7}},\ ] ] which indicates a central curvature singularity at @xmath1 that is clothed by the event horizon . to find the location of horizons , @xmath209
is set to zero and we have @xmath210 there are three possible cases to be considered .
case 1 : if @xmath211 , the equation admits two positive roots indicating inner and outer horizons of the black hole
. case 2 : if @xmath212 , this is an extreme case and the equation ( 57 ) has one real positive root .
this means that there is only one horizon .
case 3 : if @xmath213 there is no real positive root and the solution does not admit black hole so that the singularity at @xmath1 is naked . with reference to the detailed analysis given in @xcite ,
the penrose diagram of the solution illustrates the timelike character of the singularity at @xmath1 .
our aim in this section is to investigate the behaviour of this naked singularity when probed with klein - gordon and dirac fields in the framework of quantum mechanics .
the radial equation for the metric ( 55 ) is obtained for the massles case@xmath214 as , @xmath215 the behaviour of the radial equation as @xmath94 is @xmath216 where @xmath217 the solution is given by @xmath218 both solutions are square integrable in hilbert space , that is , @xmath219 therefore , the spacetime is quantum mechanically singular when probed with quantum particles obeying klein - gordon equation .
if we use the sobolev norm , @xmath220 although the first integral of the solution is square integrable , the second integral for @xmath221 fails to be square integrable and the spacetime is quantum mechanically wave regular . the dirac equation can be written as , @xmath222 where @xmath223 is given in ( 54 ) . by using the same anzats as in ( 36 ) ,
the radial part of the dirac equation becomes , @xmath224 in which @xmath225 , @xmath226 the solution becomes , @xmath227 which is square integrable .
this is verified first by expanding the whittaker functions in series and then by integrating term by term in the limit as @xmath180 consequently , the spacetime is quantum mechanically regular when probed with dirac fields .
matter coupled @xmath4 dimensional black hole spacetimes are shown to share similar quantum mechanical singularity structure as in the case of pure btz black hole .
the inclusion of matter fields changes the topology and creates true curvature singularity at @xmath1 .
the effect of the matter fields allows only specific frequency modes in the solution of klein - gordon and dirac fields . if the quantum singularity analysis is based on the natural hilbert space of quantum mechanics which is the linear function space with square integrability @xmath228 , the singularity at @xmath1 turns out to be quantum mechanically singular for particles obeying the klein- gordon equation and regular for fermions obeying the dirac equation .
we have proved that the quantum singularity structure of @xmath4 dimensional black hole spacetimes are generic for dirac particles and the character of the singularity in quantum mechanical point of view is irrespective whether the matter field is coupled or not .
this result suggests that the dirac fields preserve the cosmic censorhip hypothesis in the considered spacetimes that exhibit timelike naked singularities . instead of horizons ( that clothes the singularity in the black hole cases ) the repulsive barrier is replaced against the propagation of dirac fields .
however , for particles obeying klein - gordon fields , the singularity becomes worse when a matter field is coupled .
however , we have also shown that in the charged btz ( in linear electrodynamics ) and dilaton coupled black hole spacetimes specific choice of waves exhibit quantum mechanical wave regularity when probed with waves obeying klein - gordon equation , if the function space is sobolev with the norm defined in ( 25 ) .
the singularity at @xmath1 is stronger in the non - linear electrodynamic case .
it should be reminded that , one may not feel comfortable to use sobolev norm in place of natural linear function space of quantum mechanics . | quantum singularities considered in the @xmath0 btz spacetime by pitelli and letelier ( phys . rev .
d77 : 124030 , 2008 ) is extended to charged btz and 3d einstein - maxwell - dilaton gravity spacetimes .
the occurence of naked singularities in the einstein - maxwell extension of the btz spacetime both in linear and non - linear electrodynamics as well as in the einstein - maxwell - dilaton gravity spacetimes are analysed with the quantum test fields obeying the klein - gordon and dirac equations .
we show that with the inclusion of the matter fields ; the conical geometry near @xmath1 is removed and restricted classes of solutions are admitted for the klein - gordon and dirac equations .
hence , the classical central singularity at @xmath1 turns out to be quantum mechanically singular for quantum particles obeying klein - gordon equation but nonsingular for fermions obeying dirac equation .
explicit calculations reveal that the occurrence of the timelike naked singularities in the considered spacetimes do not violate the cosmic censorship hypothesis as far as the dirac fields are concerned .
the role of horizons that clothes the singularity in the black hole cases is replaced by repulsive potential barrier against the propagation of dirac fields . |
exact solutions of field equations are of fundamental importance as in theoretical as experimental physics
. such solutions of the pauli and dirac equation are described in a variety of books ( see , for example , @xcite ) . however , as far as the author knows , there are no publications on two - dimensional ( 2d ) square integrable solutions of these equations in rotating electromagnetic fields with a constant magnetic field along the rotation axis . in this paper
we present such solutions for the pauli equation in a rotating magnetic field and for the dirac equation in the field of a traveling circularly polarized electromagnetic wave . for finding the solution the translation of spinor into a rotating frame
is necessary .
this translation must be accompanied by the transformation of coordinates and time .
these manipulations turn the non - stationary into stationary problem . however what must be the transformation ? in the appendix
we discuss some physical arguments for the choice of the reasonable transformation .
the galilean or lorentz transformation is the term pertaining to mechanics .
we use here this term keeping in mind that the role of velocity is played by frequency .
we start from the pauli equation@xmath0 in standard notations .
@xmath1 is the potential : @xmath2 @xmath3 @xmath4 $ ] , @xmath5 is the constant component of the magnetic field along the @xmath6-axis , @xmath7 is the amplitude of the rotating transverse component@xmath8 _ _ _ _ @xmath9 _ _ and @xmath10 is the plank constant , mass , charge , speed of light , magnetic moment and pauli s matrix respectively , @xmath11 is the frequency of rotation . @xmath11 may take as positive as negative values .
negative @xmath11 corresponds to the reverse rotation .
we consider solutions with the definite momentum @xmath12 along the @xmath6-axis .
solutions must be continuous and square integrable over all the cross - section transversely to this axis.the potential correspond to the magnetic field@xmath13 the search for solutions consists of the wave function translation into the rotating frame @xmath14and a transformation of coordinates@xmath15where _
_ _ _ @xmath16 _ _ _ _ @xmath17__. _ _ above two operations reduce the pauli equation to the stationary case .
further this equation is putted to the diagonal form @xmath18 where @xmath19after that we use new coordinates @xmath20and investigate stationary states @xmath21 , where @xmath22 is the `` energy '' in the rotating frame .
finally denote @xmath23then the pauli equation may be written in the following stationary form@xmath24 in this equation the comma means the differentiation . in eq .
( [ eq0 ] ) two components of the wave function are defined by independent equations .
exact solutions for every component of the wave function exist in the form @xmath25 where @xmath26 corresponds to the sign of diagonal components of @xmath27 , @xmath28@xmath29 is a polynomial in @xmath30 and @xmath31 , the quantum number @xmath32 is defined as maximal number @xmath33 in the product @xmath34 @xmath35 are integers .
the complete two - component wave function has to be defined as a combination of two solutions @xmath29 . in the initial frame
the wave function is@xmath36@xmath37where @xmath38 are constants .
the wave function ( [ sol ] ) describes a state with the spin rotation .
every @xmath39 is connected with @xmath40 energy levels @xmath41 with condition @xmath42 .
the levels are determined from an algebraic equation of the power @xmath43 .
in particular , first three energy levels are @xmath44@xmath45 , \label{eps } \\ \tau & = & \frac{\hbar ^{2}}{2m}\sqrt{(d_{11}-d_{22})^{2}+d_{12}^{2}+f^{2}}. \label{tau}\end{aligned}\]]the subscript of @xmath22 equals @xmath46 energy levels are numbered by superscripts @xmath47 .
two values of both superscripts @xmath48 and @xmath49 corresponds two sign of @xmath27 and two sign before @xmath49 in ( e1 ) , ( [ e2d ] ) .
parameters @xmath50 are defined by equations@xmath51it may be shown with help these relations that if @xmath52 then @xmath53 . in this
limit solutions turn out into that of 2d axially symmetric harmonic oscillator .
the wave function normalization@xmath54where the integration is over all the cross - section , imposes one condition on two coefficients @xmath38 @xmath55the second one follows application scenarios . in particular , at the magnetic resonance we have following expression for the average value of spin @xmath56the first condition of the magnetic resonance @xmath57 is defined by making the constant part of @xmath58 equal zero .
it is achieved by the initial polarization .
the second condition @xmath59 is defined by making the amplitude of the variable part equal maximum .
it is achieved by the adjustment of @xmath11 or @xmath5 .
the same procedure for the magnetic resonance definition is used in the next section for solutions of dirac s equation .
desirable solutions does not exist if @xmath60 @xmath61 since in this case the integral ( [ norm ] ) tends to infinity and if @xmath62since in this interval energy has complex values . at boundaries of ( [ gf ] ) energy
is real but at the lower and upper boundary @xmath63 and @xmath64 respectively , i.e. , solutions are not square integrable . with help of ( [ gf ] )
we may evaluate the forbidden zone for values of the magnetic moment - or rather its @xmath65 factor - using the condition of the magnetic resonance @xmath66 or@xmath67(for electron @xmath68 and @xmath69 has the same sign ) . from ( [ mr ] ) and ( [ ds3 ] ) we obtain @xmath70 @xmath71 then from ( [ gf ] ) we find the forbidden zone for the @xmath65 factor@xmath72usually @xmath73 and the zone represents a narrow vicinity of two points @xmath74 and @xmath75 .
the second inequality in eq . ( [ zone ] ) establishes an important result .
namely , the magnetic moment must be always anomaluos .
note that the pauli principle @xcite in the given case may be reformulated as follows .
any state in a rotating magnetic field may have two electrons with the opposite sense of the spin rotation . in conclusion of this
section note that similar exact solutions are not found for the dirac equation . on the other hand solutions in the field of a traveling circularly polarized electromagnetic wave
are not found for the pauli equation but they exist for the dirac equation .
it is connected with the unusual character of solutions .
solutions of the dirac equation considered below do not consist of a small and large two - component spinor .
therefore they can not be presented in the first approximation as solutions of the pauli equation
consider exact solutions of the dirac equation@xmath76 in the field of a powerful traveling circularly polarized wave and a constant magnetic field directed along the axis of rotation ( propagation ) .
such a field corresponds to the potential @xmath77 @xmath78 @xmath79 _ _ _ _ is the propagation constant , where values @xmath80 and @xmath81 correspond to propagation of fermions and wave in the same and opposite direction respectively .
make manipulations analogously to previous section .
however we must take into account the coordinate @xmath6 in the translation of the wave function @xmath82\tilde{\psi}$ ] as well as in the transformation of coordinates@xmath83 further we consider the stationary case with the definite momentum @xmath84 @xmath85 then in coordinates @xmath86 @xmath87 exact solutions , somewhat more symmetric than that of the pauli equation , are @xmath88 where @xmath89@xmath90 is the spinor polynomial in @xmath30 and @xmath31 .
such solutions exist as for @xmath91 as for @xmath92 . in the first case
@xmath93 the parameter @xmath94 is always positive since solutions must decrease at infinity , @xmath95 @xmath96 as an example , we present here the normalized wave function of the ground state for the first case @xmath97 where @xmath98 .
a distinguish feature of the solution ( [ wf ] ) is that bispinor does not consist of a small and large two - component spinor .
it means that this solution can not be reduced to that of the pauli equation .
solution in the form of _ _ _ _ an expansion in infinite series of terms like ( [ sol ] ) is possible . but
such series evidently do not converge to a finite value .
another distinguish feature is that energy of the ground state is defined by an algebraic equation of the third order@xmath99where @xmath100 for @xmath101 eq . ( [ eqe ] )
has two positive and one negative roots @xmath102 .
therefore a combination of two solutions with positive roots may be used as a solution . in the second case @xmath103
@xmath104 @xmath105and above expressions ( [ wf]-[n ] ) are still valid by substituting @xmath106 in the initial frame the wave function is @xmath107\psi _ { n}. \label{wfd}\]]according to the procedure of the previous section consider the average value of spin @xmath108and its temporal evolution .
the wave function consists of two solutions with two positive roots of eq .
( [ eqe ] ) : @xmath109 the normalization integral ( [ norm ] ) and the equality to zero of the constant part of @xmath58 results in following values of normalization constants @xmath110 @xmath111@xmath112 } , \label{teta}\]]where @xmath113for the average spin we obtain @xmath114 \sin 2\theta \cos \phi , \label{s}\]]where @xmath115 and @xmath116 is the normalization parameter ( [ n ] ) and the parameter ( [ dh ] ) for @xmath117 respectively , @xmath118 defines the frequency of the spin oscillation@xmath119 it may be shown that for small @xmath120 in the first approximation the amplitude maximum of @xmath58 is realized at @xmath121 .
it is easily to see that this equality is the condition of the magnetic resonance ( [ mr ] ) for @xmath122 .
however next approximations lead to conflicting results .
once again emphasize the solution ( [ sol ] ) is not an approximation of ( [ wf ] ) .
the solution ( [ wf ] ) represents a new class of solutions . in the general case
the amplitude of the average spin as well as the frequency of the spin oscillation is a complicated function of all parameters .
however both variables can be calculated in accordance with ( s ) , ( [ ft ] ) , adjusted and tested experimentally .
the concept of the point rotation frame arises in electrooptics , when light propagates through electrooptical crystal with the rotating optical indicatrix ( index ellipsoid ) @xcite .
the optical indicatrix is an example of such a frame however we attach a more fundamental meaning to this concept .
the concept is applicable to any rotating field , in particular to rotating magnetic field in quantum mechanics .
a distinguish feature of the point rotation frame is the existence of the rotation axis at every point .
all points in the plane perpendicular to the rotation axis are equivalent but time in frames rotating at different frequencies may be different .
centrifugal forces are absent in this frame .
coordinates of the frame are an angle and time ; the frequency of rotation is a parameter @xcite .
such non - rotating frames are used in quantum field theory for a long time . above it
is shown that the first step in studying solutions of the pauli and dirac equation in the rotating electromagnetic field is the translation of the wave function into the rotating frame .
this translation must accompanied by the transformation of coordinates and time . for finding of above solutions
the galilean transformation ( [ d2 ] ) or ( [ d3 ] ) is used .
however there is possible a phase mismatch if the real transformation is non - galilean and time in the initial and rotating frame is different .
the problem in a sense is similar to that in mechanics where the consideration in the system of the mass center is preferable .
the correct result in mechanics can be obtained only if the lorentz transformation is used for the transition into this system . in the opposite case
some corrections of the result are necessary . in this sense
the frame with the resting magnetic field is of particular importance as an analog of the mass center in mechanics .
the general linear 2d transformation for the point rotating frames was considered in @xcite .
it was assumed that symmetry exists by the exchange angle @xmath123 time and shown that in this condition a non - trivial transformation together with the lorentz transformation is possible .
this transformation possesses unusual properties and enables the existence of allowable frequency regions with lower and upper boundaries ( see an example in @xcite ) . from the general physical viewpoint it seems more preferable in comparison with the galilean ( with infinite frequencies ) or lorentz ( with the sole limiting frequency ) transformation
. however the transformation can not be uniquely determined with help of the assumption on the symmetry only .
in @xcite the 2d transformation was used for a phenomenological description of a circularly polarized light wave in the electrooptical single - sideband modulator @xcite , @xcite .
this description is similar to an elastic collision in mechanics , however such a consideration is not applicable to quantum mechanics . the 3d galilean transformation ( [ d3 ] ) is used for finding solutions of the dirac equation in the field of traveling circularly polarized wave .
some aspects of the general 3d transformation is considered in @xcite .
this transformation also can not be uniquely determined . in both cases
there is lacking a principle which would allow determining the transformation explicitly .
we have presented the new class of exact solutions of the pauli and dirac equation in rotating electromagnetic fields .
their properties differ significantly from known presently exact solutions . that especially is related to the dirac equation . in the standard approach
the pauli equation is not the first approximation of dirac s equation for such solutions .
the obtained solutions may be a basis for experiments at large values of parameters , in particular , for relativistic electron .
we have discussed non - galilean transformation .
main problem is the absence of a principle for determining the transformation .
main advantage of such a transformation is the possibility of different lower and upper frequency boundaries for different fields .
this is important from the general physical viewpoint and calls for detail investigations of this issue . | we describe a new class of exact square integrable solutions of the pauli and dirac equation in rotating electromagnetic fields .
solutions obtained by putting equations in the stationary form with help of a coordinate transformation corresponding to the transition into a rotating frame .
the transformation is assumed to be galilean one however a non - galilean transformation is of particular interest for such solutions .
obtained solutions , especially of dirac s equation , are valid for arbitrary values of parameters and may be tested experimentally . |
the time - resolved gamma - ray burst ( grb ) spectra have the mean observed photon spectral index @xmath10 close to @xmath11 ( i.e. @xmath12 ) and some spectra can be as hard as @xmath13 @xcite .
synchrotron shock models ( see , e.g. , @xcite ) assume that a fraction of the available energy is given instantaneously to the electrons which are injected at lorentz factor @xmath14 .
the typical cooling time - scale is smaller that the light crossing time of the source by a factor @xmath15 , where the compactness @xmath16 is about @xmath17 for typical grb parameters @xcite . because the cooling time is small
, one observes the time - averaged spectrum of cooling electrons @xcite . for relativistic electrons ,
the photon emission frequency is @xmath18 , where @xmath19 for synchrotron or synchrotron - self - compton ( ssc ) radiation .
thus , the rate of the frequency change @xmath20 , where @xmath21 is the emitted power . the time - averaged flux is then @xmath22 and the `` cooling '' spectrum is @xmath0 ( i.e. @xmath23 ) for synchrotron as well as for ssc @xcite , which is much softer that those observed from grbs .
the immediate conclusion from this simple exercise is that _ all _ the models involving injection of electrons at large energies should be rejected .
the efficient cooling can be prevented if the energy is supplied to the electrons in small steps . in this paper , we discuss consequences of the assumption that the available energy is supplied continuously to the particles over the life - time of a source .
we study how the resulting spectrum depends on main parameters and determine parameter range where model spectra are consistent with the majority of the observed ones .
a number of physical models can produce a source of continuously heated particles . these can be plasma instabilities behind ultrarelativistic shocks @xcite or dissipation of the magnetic energy in a poynting flux dominated outflow ( e.g. @xcite ) .
therefore , we consider a toy model where energy is injected to the emission region with the constant rate during comoving time @xmath24 . we adopt a slab geometry with the thickness @xmath25 ( in units of @xmath26 ) which may depend on the specific scenario .
the energy is injected uniformly over the slab volume in a form of acceleration of electrons ( and pairs ) which obtain equal amount of energy per unit time .
the model is fully described by four parameters : ( i ) the initial thomson optical depth across the slab , @xmath27 ; ( ii ) the comoving size @xmath26 ; ( iii ) the dissipation compactness , @xmath28 related to the observed isotropic energy release @xmath29 and the ejecta bulk lorentz factor @xmath30 , which determines the rate of pair production ; and ( iv ) the magnetic compactness @xmath31 determining the role of synchrotron radiation .
see @xcite for details .
let us first consider how particles ( electron and positrons ) of lorentz factor @xmath32 are heated and how do they cool .
the energy gain rate of a particle is simply given by the heating rate @xmath33 divided by the number of particles ( which is proportional to the total thomson optical depth , including pairs , across the slab , @xmath1 ) .
particles cool by emitting synchrotron radiation and by scattering this radiation ( ssc ) .
the energy balance equation can be written as : @xmath34 where @xmath35 accounts for the reduced synchrotron cooling due to synchrotron self - absorption and @xmath36 is the compactness corresponding to the energy density of soft photons in the thomson regime .
it can be expressed as a sum of the synchrotron @xmath37 and first compton scattering @xmath38 energy densities , where @xmath39 is the compton parameter . if mean lorentz factor @xmath40 , then one need to account for further scattering orders .
the typical cooling time , @xmath41 $ ] , is orders of magnitude smaller than the light - crossing time @xmath24 for grb conditions .
the balance between heating and cooling is achieved at @xmath42 and thus @xmath43 .
particles with @xmath44 lose energy faster than they gain it , while at @xmath45 the situation is opposite resulting in a narrow distribution peaked at @xmath46
. main radiative process is determined by the instantaneous optical depth @xmath1 , which can be significantly larger than the initial @xmath47 because of pair production at @xmath48 . if @xmath49 , then @xmath50 and the main emission mechanism is synchrotron .
a typical energy of a synchrotron photon in the frame comoving with the ejecta ( in units @xmath51 ) is @xmath52 , for @xmath53 ( where @xmath54 and @xmath55 ) and compton scattering is in klein - nishina limit . if @xmath56 , @xmath57 and @xmath58 .
this time , first compton scattering peak is at @xmath59 , and one can neglect further scatterings . at @xmath60 , second compton scattering will dominate the energy output . for high compactness source @xmath1 can reach 1 , then @xmath3 and
multiply compton scatterings ( quasi - thermal comptonization ) have to be accounted for .
the simulations were performed using a large particle monte carlo code described in @xcite .
it treats compton scattering , photon - photon pair production and pair annihilation , synchrotron radiation and synchrotron self - absorption .
the electron / pair and photon distributions are computed self - consistently . as an example
, we consider @xmath61 , @xmath62 , @xmath63 ( corresponding to @xmath64 and @xmath65 erg ) , and @xmath66 , which are typical for grb ejecta @xcite .
the resulting evolution of broad - band photon spectra and electron distributions is shown in fig .
[ fig : spectra ] . at the start of simulations @xmath67 and @xmath68 ( see dashed curves ) .
partially self - absorbed synchrotron peaks at @xmath69 ( uv in the observer frame ) with low - energy tail @xmath70 may be responsible for the prompt optical emission seen in grb 990123 @xcite .
when @xmath1 starts to grow due to pair production , @xmath71 drops as @xmath72 , and the first compton peak moves to softer energies crossing the ` batse window ' .
this spectral evolution is consistent with the observed in time - resolved pulses @xcite .
the low - energy slope of this component , corresponding to compton scattering by isotropic monoenergetic electrons , can as hard as @xmath73 ( i.e. @xmath74 ) , reproducing most of the observed grb spectra .
the second compton component peaking at @xmath75@xmath76 mev rises later and decays on a longer time - scale .
it can be responsible for the delayed emission observed by egret from grb 941017 @xcite .
thus , at @xmath77 , the model spectra reproduce well the observed ones from grbs , including their hardness , spectral evolution , as well as optical and 100 mev emission .
small @xmath1 required for synchrotron to be responsible for the batse grbs are difficult to get in any physical scenario .
additionally , the synchrotron low energy spectrum @xmath78 is softer that the observed ones @xcite . at @xmath79
, the resulting quasi - thermal comptonization spectrum may be too soft for most of the grb spectra .
the emission peaks at 1050 kev in the comoving frame of the ejecta and , for @xmath80 , this peak shifts to an uncomfortably high energy @xcite . in conclusion , ssc from continuously heated electrons ( pairs ) seems to be the most promising model to explain the observed properties of grbs . | we discuss here constraints on the particle acceleration models from the observed gamma - ray bursts spectra .
the standard synchrotron shock model assumes that some fraction of available energy is given instantaneously to the electrons which are injected at high lorentz factor .
the emitted spectrum in that case corresponds to the spectrum of cooling electrons , @xmath0 , is much too soft to account for the majority of the observed spectral slopes .
we show that continuous heating of electrons over the life - time of a source is needed to produce hard observed spectra . in this model
, a prominent peak develops in the electron distribution at energy which is a strong function of thomson optical depth @xmath1 of heated electrons ( pairs ) . at @xmath2 , a typical electron lorentz factor @xmath3 and quasi - thermal comptonization operates .
it produces spectrum peaking at a too high energy .
optical depths below @xmath4 would be difficult to imagine in any physical scenario . at @xmath5@xmath6 , @xmath7 and
synchrotron self - compton radiation is the main emission mechanism .
the synchrotron peak should be observed at 10100 ev , while the self - absorbed low - energy tail with @xmath8 can produce the prompt optical emission ( like in the case of grb 990123 ) .
the first compton scattering radiation by nearly monoenergetic electrons peaks in the batse energy band and can be as hard as @xmath9 reproducing the hardness of most of the observed grb spectra .
the second compton peak should be observed in the high - energy gamma - ray band , possibly being responsible for the 10100 mev emission detected in grb 941017 .
a significant electron - positron pair production reduces the available energy per particle , moving spectral peaks to lower energies as the burst progresses .
[ 1999/12/01 v1.4c il nuovo cimento ] |
in recent years , the field of nonequilibrium surface growth has been investigated using various discrete models and continuous equations@xcite .
the comprehension of nonequilibrium surface growth plays an important role in understanding and controlling many interesting interface processes , such as vapor deposition@xcite , crystal growth@xcite , molecular beam epitaxy ( mbe)@xcite , and so on . during the mbe growth process
, the conserved growth condition is applied without defects , such as overhangs and vacancies .
various discrete conserved models for mbe , describing the kinetic properties of this type of surface growth , have been proposed and studied by intensive numerical simulations .
the main purpose of studying discrete models is to measure scaling exponents for the kinetic roughening , which determines the asymptotic behavior of surface growth on a large length scale in a long time limit .
the important result of the kinetic roughening studies is that a large variety of different discrete growth models can be divided into only a few universality classes .
the surface width @xmath2 , which measures the root - mean - square fluctuation of the surface height , scales as w(l , t ) ~l^f(t / l^z ) , where the asymptotic behavior of the scaling function @xmath3 is constant for @xmath4 and @xmath5 for @xmath6 with @xmath7 .
the scaling behavior of the growth is characterized by three exponents : the roughness exponent @xmath8 , the growth exponent @xmath9 , and the dynamical exponent @xmath10 .
these exponents determine the universality class . in the coarse - grained picture , evolution of the growing surface
is usually described by a stochastic differential equation ( sde ) for the height variable @xmath11 as a function of the surface coordinate and time . for discrete models of mbe growth ,
several sdes were suggested and it was generally believed that there is a correspondence between discrete growth models and continuum sdes . the common way of establishing the link between discrete models and continuous equations is a simple comparison of critical exponents determined from computer simulations of the discrete model with exponents for the continuous equation .
there have also been attempts to establish the correspondence in an explicit way .
the systematic procedure for establishing a continuous equation corresponding to discrete models , starting from the master equation in discrete space was proposed by vvedensky _
et al._@xcite and has been successfully applied to the derivation of growth equations for some discrete models , including the solid - on - solid ( sos ) model , the restricted solid - on - solid ( rsos ) model , as well as the wolf - villain and das sarma - tamborenea models@xcite .
however , there are several difficulties with this procedure ; in particular , in converting from the equation system for a discrete set of heights to an equation for a continuous function @xmath11 , the procedure requires the regularization step , in which the nonanalytic quantities are expanded and replaced with analytic quantities , i.e. , the step function is approximated by an analytic shifted hyperbolic tangent function expanded in a taylor series .
however , the form of the regularized function is uncertain , and different choices of this function lead to different results .
some forms of the regularized function have been suggested , but the problem of a proper choice of a regularization scheme has not been discussed . to overcome this kind of uncertainty , the authors proposed a new schematic formalism@xcite , deriving the continuous equations , such as the edwards - wlikinson ( ew)@xcite and kardar - parisi - zhang ( kpz)@xcite equations , from the body centered solid - on - solid model and the rsos model . in this paper , we apply our formalism to a new kind of mbe growth model proposed by kim , park , and kim@xcite .
this model allows the deposited particle to relax to the nearest site where the rsos condition on neighboring heights is satisfied and has the conserved growth condition constraint , which means the deposited particles are possible to hop for an infinite distance until they eventually find the site with the rsos condition satisfied .
applying our formalism to the above conserved rsos ( crsos ) model , we not only derive the villain - lai - das sarma ( vld ) equation@xcite for the model which belongs to a different universality class from the ew and kpz equations , but also we are able to predict the coefficients in the vld equation , which was not possible by other methods . observing that the rsos model belongs to the kpz class and
the crsos model belongs to the vld class , we went one step further to study the rsos model with the finite range hopping ( rsos / h ) . in this rsos / h model
, it is possible for the deposited particles to hop a finite distance @xmath12 until they find the site with the rsos condition satisfied . if they fail to find the site with the rsos condition satisfied within the distance @xmath12 in both directions , the deposition process is rejected .
the rsos model corresponds to the rsos / h model with @xmath13 and to the crsos model with @xmath14 .
we apply our formalism to the rsos / h model with @xmath12 finite and find that this model belongs to the kpz class , contrary to the previous report by kim and yook@xcite , who concluded that there is a phase transition between the kpz class to the vld class along the parameter @xmath12 . in sec .
[ an ] our formalism to derive the continuous equation from the discrete model is briefly explained and the procedure of derivation is described .
the detailed calculations are attached in the appendices .
extensive numerical simulations are presented in sec .
[ num ] and the summary and discussion are given in sec .
in this section , we derive continuous equations for the one - dimensional rsos / h model with a hopping distance @xmath12 and for the crsos model corresponding to @xmath14 . we restrict ourselves to the case wherein the height difference between two nearest neighbors is not larger than 1 . for a succinct description of the dynamics
, we introduce the nomenclature that if a site @xmath15 satisfies the condition @xmath16 , this site is called `` stable . ''
following this nomenclature , the growth algorithm of the one - dimensional rsos / h model is as follows : ( i ) we choose a site @xmath15 randomly .
( ii ) sites from @xmath17 to @xmath18 are examined to determine if they are stable sites .
( iii ) if a stable site is found within the interval from @xmath17 to @xmath18 , a new particle is deposited to the nearest stable site from @xmath15 ( @xmath15 itself can be a candidate for deposition ) .
however , if stable sites are nonexistent in the examined interval , the particle drop is rejected and the system remains unchanged . after this try , the time is increased by @xmath19 , where @xmath20 is the system size .
we assume periodic boundary conditions . since the height difference between two nearest neighbors
is restricted not to be larger than 1 , the rsos / h model is mapped onto the reaction - diffusion system of hard - core particles with two species .
the step - up ( -down ) is mapped to an @xmath21 ( @xmath22 ) particle .
if two nearest neighbor sites have equal height , a particle vacuum is located between these two sites .
the site where the particles reside is labeled by an integer , and the site for height by a half - integer .
this mapping is depicted in fig.[config ] . according to this mapping
, the dynamics of the rsos / h model can be described by the ( imaginary time ) schrdinger equation @xmath23 for the state vector @xmath24 , where @xmath25 is the probability with which the system is in state @xmath26 at time @xmath27 , and @xmath28 , called a hamiltonian , is an evolution operator = _ n ( i - _ n - _ n ) _n , [ hamiltonian ] where _n = i + 12 _
j=1 ^ 2 l_0 ( _ k=1^j _ n+k+ _ k=1^j _ n - k ) , and ccl _ n & = & _ n _
n + _ n+1 _ n+1 - _ n
_ n+1 , + + _ n & = & ( _
n + _ n ) ( _ n+1 + _
the role of the rejection operator @xmath29 is to check the stability of site @xmath30 , that is , if a configuration @xmath31 has a stable site at @xmath30 , @xmath32 and otherwise @xmath33 .
the adsorption operator @xmath34 describes the configuration change after a successful deposition .
@xmath35 ( @xmath36 ) is the annihilation operator of an @xmath21 ( @xmath22 ) particle at site @xmath37 and @xmath38 ( @xmath39 ) is the corresponding creation operator , satisfying the mixed commutation relations presented in ref.@xcite , _
n _ n = _ n _
n = i - _ n _
n - _ n _ n,_n _ n = _
n _ n = 0 .
all operators at different sites commute with each other . to find the sde for the rsos / h model , we apply the method recently introduced by the authors@xcite .
first we assume the existence of the lattice version of the sdes in terms of density .
those equations are supposed to take the forms _
t a_n&= & c_n^a ( \ { a},\{b } ) + _ m , [ sdea ] + _ t b_n&= & c_n^b(\{a},\{b } ) + _ m , [ sdeb ] where @xmath40 and @xmath41 , @xmath42 are white noises with correlation & _
m^x(t ) = _ nm _ x , x ( t - t ) & + & ( x , x = either a or b ) , & where @xmath43 and @xmath44 are kronecker deltas and @xmath45 is the dirac delta function .
the matrix @xmath46 is related to the kramers - moyal coefficient in such a way that & _ r , x
= c_nm^xx ( \{a},\{b } ) & + & ( x , x,x = either a or b ) .
& here we are using the it interpretation . the field @xmath47 ( @xmath48 ) in the curly bracket represents the density of species @xmath21 ( @xmath22 ) at all sites . from here on , without a hat above itself a mathematical symbol is a pure number as opposed to an operator .
@xmath49 should not be confused with @xmath35 .
the former is a density at site @xmath37 that runs over real numbers , while the latter is an annihilation operator . by requiring that the noise average of observables in eqs .
( [ sdea ] ) and ( [ sdeb ] ) has the same behavior with the ensemble average of the number operator , we find@xcite c_n^a _ t & = & _ t , + c_n^b _ t & = & _ t , + c_nm^aa _ t & = & _ t , [ commutation ] + c_nm^ab _ t & = & c_mn^ba _ t = _ t , + c_nm^bb _ t & = & _ t , where the @xmath50 on the left hand side represents the average over noise at time @xmath27 and that on the right hand side stands for the ensemble average .
the arguments of the kramers - moyal coefficients are dropped for brevity .
as presented in ref.@xcite , the ensemble average of any operator can be interpreted as an average of number operators due to the property of the projection state @xmath51 , which is defined as a sum over all possible microscopic configurations and is itself a left eigenstate of @xmath28 with eigenvalue @xmath52 , and in turn , the ensemble average of a number operator is mapped to the noise average of density .
this procedure leads us to find the kramers - moyal coefficients @xmath53 , @xmath54 , in terms of the density fields .
we call this procedure _ figurization _ , which means `` expression in number . ''
to represent the _ figurization _ , we use the symbol @xmath55 to the left of which is an operator ( or a product of operators ) and to the right of which is the corresponding density representation . to complete the derivation
, we perform the commutation relations between the hamiltonian and the density operators and so forth : & = & _ n ( _ n+1 + _
n+1 ) _n - _ n ( _ n-1 + _
n-1 ) _n-1 , [ den_commute_a ] + & = & -_n ( _ n+1 + _
n+1 ) _n + _ n ( _ n-1 + _
n-1 ) _n-1 , [ den_commute_b ] + & = & _ ln ( _ n ( _ n+1 + _
n+1 ) _n + _ n ( _ n-1 + _
n-1 ) _n-1 ) + & & -_l , n+1 _ n_n+1_n - _ l , n-1 _ n _ n-1 _n-1 , + & = & _ ln ( _ n ( _ n+1 + _
n+1 ) _n + _ n ( _ n-1 + _ n-1 ) _n-1 ) + & & - _ l , n+1 _ n _ n+1 _n - _ l , n-1 _ n _ n-1 _n-1 , + & = & _ l , n+1 _ n _ n+1 _n + _ l , n-1 _ n _ n-1 _n-1 , where @xmath56 is a kronecker delta . following the _ figurization _ , we find the kramers - moyal coefficients . for a later purpose , we give some examples of the _ figurization_. the _ figurization _ of @xmath29 is @xmath57 , and the symbolic representation of the _ figurization _ of the product of @xmath58 s is _ k=1^j _
n+k r_n^j , where the superscript @xmath59 should not be confused with the power .
when @xmath60 , @xmath61 is denoted as @xmath62 .
@xmath63 are given by the following recursion relations : r_n^j & = & a_n+1 r_n+1^j-1 + ( 1 - a_n+1 ) _
k=2^j+1 b_n+k , [ recursive1 ] + r_n - j-2^j & = & b_n-1 r_n - j-2^j-1 + ( 1 - b_n-1 ) _ k=2^j+1 a_n - k , [ recursive2 ] where @xmath64 and we define @xmath65 . the physical meaning of eqs .
( [ recursive1 ] ) and ( [ recursive2 ] ) is as follows : we divide the situation that prohibits the height increase at sites @xmath66 by the condition at site @xmath67 .
if there is an @xmath21 particle at site @xmath67 , the height increase is suppressed at site @xmath68 irrespective of the condition at site @xmath69 .
hence the first term of eq .
( [ recursive1 ] ) follows .
if there is no @xmath21 particle at site @xmath67 , to suppress the deposition at site @xmath68 , there must exist a @xmath22 particle at site @xmath69 and this should be continued until the site @xmath70 , because at site @xmath71 ( @xmath72 ) no @xmath21 particle is present ; this condition is represented by the second term . to comprehend eq .
( [ recursive2 ] ) , we only have to perform the mirror transformation relative to site @xmath37 . by the mirror transformation relative to a site @xmath37 , we mean the exchange of @xmath47 and @xmath48 ( @xmath73 ) , followed by @xmath74 . under this transformation ,
@xmath75 changes into @xmath76 for an arbitrary @xmath77 .
the mirror transformation of eq .
( [ recursive1 ] ) is eq .
( [ recursive2 ] ) . with these notions , we will find the sde of the rsos / h model . at first , the deterministic part of sde is found . since the main concern is not the respective dynamics of the @xmath21 and @xmath22 particles , but it is @xmath78 ( the local slope ) and @xmath79 ( the slope density ) , we will write the sde for @xmath80 and @xmath81 rather than for the @xmath21 , @xmath22 particles . the kramers - moyal coefficient @xmath82
is obtained by subtracting eq .
( [ den_commute_b ] ) from eq .
( [ den_commute_a ] ) and @xmath83 by adding eq .
( [ den_commute_b ] ) and ( [ den_commute_a ] ) followed by the _ figurization_. c_n^d & = & 12_j = 1 ^ 2 l_0 + & + & 12 ( r_n-2 ^ 2l_0 + 1 - r_n-1 ^ 2l_0 + 1 + r_n-2l_0 - 2 ^ 2l_0 + 1 - r_n-2l_0 - 1 ^ 2l_0 + 1 ) , [ km_of_height ] + c_n^s & = & 2 - 3 ( a_n + b_n ) - ( b_n a_n+1 + a_n b_n-1 - 3 b_n b_n+1 - 3 a_n a_n-1 ) + & + & 12 _ j=1 ^ 2l_0 ( r_n^j - r_n-2^j - r_n - j^j + r_n - j-2^j ) - b_n ( 1 - s_n+1 ) b_n+2 - a_n ( 1 - s_n-1 ) a_n-2 + & - & 12 ( r_n-1 ^ 2 l_0 + 1 + r_n-2 l_0 - 1 ^ 2 l_0 + 1+r_n-2 ^ 2 l_0 + 1+r_n-2 l_0 -2 ^ 2 l_0 + 1 ) + & -&_j=1 ^ 2l_0 - 1 + & -&_j=1 ^ 2l_0 - 1 .
[ km_of_step ] as pointed out in ref.@xcite , the mass term in eq .
( [ km_of_step ] ) makes the step density saturate fast . as a result , the step density @xmath81 becomes a slave field of the slope @xmath80 and takes the form s = 2 ( l_0 ) + _ 0 ( l_0 ) d + _ 0 ( l_0 ) d^2 + , [ step_density ] where @xmath84 is the ( mean - field ) stationary step density and @xmath85 and @xmath86 may depend on @xmath12 . here
since this system has the mirror symmetry whose continuum version is invariant under the transformations @xmath88 and @xmath89 , we do not expect the occurrence of @xmath80 in eq .
( [ step_density ] ) .
the parameters @xmath90 , @xmath91 , and @xmath92 are determined by the stationarity of eq .
( [ km_of_step ] ) . in appendix [ rj ] , we obtain @xmath93 , @xmath94 , and @xmath85 as functions of @xmath12 and show the approximate solutions of these parameters
. we can now rewrite eq .
( [ km_of_height ] ) in terms of the height field @xmath95 using eq .
( [ step_density ] ) and d = h. [ slope ] the last task is to determine the noise strength .
this is accomplished using eqs .
( [ sdea ] ) and ( [ sdeb ] ) . after eliminating @xmath81 in favor of @xmath80 ,
there is only one kind of noise with strength c_nm
= c_nm^aa + c_nm^bb - c_nm^ab - c_nm^ba .
[ noise_strength ] using _ figurization _ of eq .
( [ commutation ] ) and keeping the most relevant term , we find @xmath96 where @xmath97 for finite @xmath12 we found _
t h = v_+ ^2 h + 2 ( h)^2 + ( x , t),[kpzeq ] + ( x , t ) ( x,t ) = d _ ( x - x ) ( t - t ) , where v_&=&1-r(2 l_0 + 1 ) 1-_^2 l_0 + o(_^4 l_0 ) , + & = & -(2 l_0 + 1)_^2 l_0 ( l_0 + 1 ) ( 2 l_0 + 1 ) ( 6l_0 + 3 - 4 ) + o(_^4 l_0 ) , + & = & -2 ( 2 l_0 + 1)- _ ^2 l_0 ( 1 + l_0 ) ( 1 + 2 l_0 ) ( 1 + + 1 3 l_0 ) + o(_^4 l_0 ) , + d _ & & 2 - 12 + o(_^4 l_0 ) , with @xmath98 .
the numerical values of @xmath99 , @xmath100 , and @xmath101 are given for several values of @xmath12 in table i. for infinite @xmath12 _ t h = v _ - ^4 h + ^2 ( h)^2 + ( x , t),[vldeq ] + ( x , t ) ( x,t ) = d _ ( x - x ) ( t - t),where v_&= & 1 , + & = & 21 - 12 2 , + & = & 10 - 3 2 , + d _ & = & 2 - 1 2 .
equation ( [ kpzeq ] ) directly shows that the rsos / h model , for any finite @xmath12 , belongs to the kpz class , and eq .
( [ vldeq ] ) suggests that the crsos model is described by the vld equation .
however , the first line of eq .
( [ km_of_height ] ) has the form _
j = 1 ^ 2 l_0 [ r_n-2^j - 2 r_n-1^j + r_n^j - \{mirror terms of ( r_n-2^j - 2 r_n-1^j + r_n^j ) } ] .
[ firstline ] which has the following implications : if the continuum version of @xmath102 has a nonvanishing coefficient of @xmath103 with a non - negative integer @xmath104 , it is not certain that eq .
( [ firstline ] ) is a lattice version of @xmath105 ; consider the mirror transformation in the continuum limit .
the occurrence of @xmath103 in eq .
( [ firstline ] ) is directly related to the appearance of @xmath106 in eq .
( [ vldeq ] ) , which generates an ew term by the dynamic renormalization group@xcite .
appendix [ proof_no_kpz ] shows that this is not the case .
the vanishing of @xmath107 in the continuum limit of @xmath75 guarantees the vanishing of @xmath103 in @xmath102 . in view of this
, we conclude that the continuum equation of the crsos model is the vld equation . to confirm this conclusion , we have performed monte carlo simulations as outlined in sec .
although the rsos / h model was studied numerically by kim and yook@xcite , their results are contradictory to our derivation . as a result , we need to perform extensive numerical simulations to verify our results . in the derivation
, we found that the coefficients of the ew terms and the kpz terms are vanishingly small , though finite , for large @xmath12 .
hence , we may find a crossover of roughness exponents from vld ( @xmath108 ) to kpz ( @xmath109 ) .
we are preoccupied with the numerical observation of this crossover . in fig .
[ alpha ] , we draw the saturated width @xmath110 as function of the system size @xmath20 for some @xmath12 s .
the system sizes are 64 , 90 , 128 , 180 , 256 , 360 , 512 , 720 , and 1024 .
for relatively small system sizes , the roughness exponents are near to the values reported in ref . @xcite .
as expected from our derivation , we see a crossover for large system sizes . to clarify the crossover behavior , a scaling plot is given in fig .
[ 1dcross ] .
the anticipated scaling form of the saturated width is w_sat ( l_0 , l ) = l^_vld g(l_0^/ l ) , where @xmath111 is the crossover exponent . the asymptotic behavior of the scaling function @xmath46 is expected to be g(x ) ~ \ { lccccc const & & when & x & & + x^_vld - _ kpz & & when & x & & 0 . .
the best fit for the data set shown in fig .
[ alpha ] corresponds to @xmath112 and @xmath113 .
the fitting parameter @xmath114 obtained is somewhat smaller than the known value of the roughness exponent of the crsos model .
this is most likely due to the smallness of @xmath12 .
for example , the data for @xmath115 in fig .
[ alpha ] yield 0.86 rather than 0.95 .
the meaning of @xmath111 is as follows : when the system size is sufficiently large , we expect the width of the characteristic mountain to be @xmath116 . due to the rsos condition , the linear size of the mountain is also expected to be @xmath116 .
the smaller the system size , the less the rejection events occur due to the shrinking of the characteristic mountain . if @xmath12 is comparable with the linear size of the characteristic mountain ( @xmath117 ) , the system starts to behave differently .
eventually , the system with small size , @xmath118 , can not feel the existence of @xmath12 .
thus the crossover length @xmath119 is expected to be @xmath120 , that is , @xmath121 .
we studied the rsos / h model using both analytical and numerical methods .
we derived the continuum equation for the microscopic discrete model analytically , and found coefficients of the ew term @xmath122 , the kpz term @xmath123 , as well as the vld term @xmath124 .
we observed that the coefficients of the ew and kpz terms behave as @xmath125 for sufficiently large @xmath12 , which is consistent with the previous numerical study@xcite .
accordingly , we concluded that the rsos / h model for any finite @xmath12 eventually belongs to the kpz class and the crsos model belongs to the vld class .
numerically , we reported the crossover from the vld class to the kpz one , which confirms our derivation .
moreover , we found a crossover exponent @xmath111 , which is argued to be 2 . besides these studies , we can offer an ( nonrigorous plausible ) argument to anticipate the universality class of the rsos / h model by employing the block spin concept of kadanoff s .
consider a system with linear size @xmath20 and hopping distance @xmath12 .
similar to the block spin in the ising model , we introduce a coarse - graining parameter @xmath48 which blocks the @xmath48 sites by one .
although the exact transformation of coarse - graining can not be determined , we expect that if it exists , @xmath12 may be renormalized as @xmath126 .
hence , we expect that the stable fixed point corresponds to @xmath13 , which is the kk model@xcite and the unstable fixed point corresponds to @xmath14 , which is the crsos model@xcite .
we are grateful to byungnam kahng for helpful discussions .
this work was supported by grant no.2000 - 2 - 11200 - 002 - 3 from the brp program of the kosef .
in this appendix , we derive the continuum limit of @xmath127 where @xmath37 is the reference point .
one may obtain the continuum limit directly from eq .
( [ recursive1 ] ) , but we do not follow this route . rather , we make recursion relation about the parameters that appear in front of the field ( see below ) . the use of eq .
( [ step_density ] ) makes it possible to represent the continuum limit in terms of @xmath80 , which is directly related to the height by eq .
( [ slope ] ) .
we can find the correct coefficient that appears in eq .
( [ step_density ] ) after we find the continuum limit of @xmath128 .
the continuum limit of @xmath128 takes the form r_n+k^l r(l ) + ( l , k ) d + ( l , k ) d^2 + ( l , k ) d + , [ cont_lim ] where the implicit dependence on @xmath12 is assumed and the argument of d is dropped for simplicity .
the symbol `` @xmath129 '' represents the continuum limit of a quantity .
equation ( [ recursive1 ] ) allows us to find the recursion relations . before going further
, the explicit form of @xmath130 in the case of no tilt boundary condition is found .
the recursion relation becomes r(l ) & = & r(l-1 ) + ( 1 - ) ^l , + ^-l r ( l ) & = & ^-(l-1 ) r(l-1 ) + 1 - with an ( sort of ) initial condition @xmath131 .
@xmath93 is the abbreviation of @xmath99 in eq .
( [ step_density ] ) .
it is trivial to find the solution that reads r(l ) = ^l + ^l ( 1 - ) l. to find the recursion relations , we need a continuum limit of @xmath132 , which is defined as p_l^k ( 1 - a_n+k+1 ) _ m=2^l+1 b_n + k + m = p_l-1^k b_n+k+l+1 .
[ recursive_of_p ] one may directly calculate the continuum limit of @xmath132 , but we follow another path .
let the continuum limit of @xmath132 be p_l^k ( 1-)^l + _ 1 ( l , k ) d + _ 1 ( l , k ) d^2 + _ 1 ( l , k ) d + .
[ cont_p ] from eq .
( [ recursive_of_p ] ) , we find the recursion relations ( @xmath133 ) & = & _ 1(l-1,k)^l-1 - 1 - 2 , + _ 1 ( l , k)^l & = & _ 1(l-1,k)^l-1 - _ 1(l-1,k ) 2 ^l + _ 0 1 - 2 , [ relation_p ] + _ 1 ( l , k)^l & = & _ 1 ( l-1,k)^l-1 + ( _ 0 - k - l-1 ) 1 - 2 , with initial conditions _ 1(1,k ) & = & - 1 2 , + _ 1 ( 1,k ) & = & 1 4 ( 1 + 2 _ 0 - 4 _ 0 ) , [ initial_p ] + _ 1 ( 1,k ) & = & 12 ( _ 0 -k-2- 2_0 + ) , @xmath86 and @xmath85 are functions of @xmath12 as shown in eq . ( [ step_density ] ) , which are to be determined by eq .
( [ km_of_step ] ) .
the solutions of eq .
( [ relation_p ] ) under the condition ( [ initial_p ] ) are ( @xmath134 ) _ 1(l ,
k ) & = & ^l-1 , + _ 1(l , k ) & = & ^l-1 , [ solution_p ] + _ 1 ( l , k ) & = & ^l-1 .
now we can obtain the continuum limit of @xmath128 .
using eqs .
( [ recursive1 ] ) , ( [ cont_p ] ) , and ( [ solution_p ] ) , we find the recursion relations of parameters given in eq .
( [ cont_lim ] ) ( l , k ) & = & ( l-1,k+1 ) , + ( l , k ) & = & ( l-1,k+1 ) + _ 1(l , k ) + _ 0 2^l-1 , [ relation_r ] + ( l , k ) & = & ( l-1,k+1 ) + _ 1 ( l , k ) + ^l-1 2 ( 1 + k + _ 0 ) , with conditions ( 1,k ) & = & 0 , + ( 1,k ) & = & 14 + _ 0 ( 1- ) , + ( 1,k ) & = & ( 1 - ) ( _ 0 - 12 ) .we solve eq .
( [ relation_r ] ) step by step .
it is trivial to find that @xmath135 .
as shown in appendix [ proof_no_kpz ] , the vanishing of @xmath111 is not a coincidence .
since @xmath136 has no @xmath77 dependence , we expect that @xmath137 also has no @xmath77 dependence .
we find ( l ) = l ^l-1 , [ kpzterm ] where @xmath77 is removed from the argument of @xmath138 due to the independence .
@xmath139 seems to have an explicit @xmath77 dependence , but by inserting @xmath140 directly , we find the independence of @xmath139 on @xmath77 .
the result reads ( we drop @xmath77 by the same reason as @xmath138 ) ( l ) = ( 1- ) ^l-1 l ( l+1 ) 4 ( 2 _ 0 - l3 - 23 ) , [ ewterm ] by requiring @xmath141 at the stationary state , we can determine @xmath93 , @xmath86 , and @xmath85 as a function of @xmath12 . to find @xmath93 , @xmath86 , and @xmath85 , we need the continuum limit of @xmath142 we determine @xmath99 from eq .
( [ km_of_step ] ) .
@xmath93 is the solution of the equation 2 x^2 - 4 x + 1 - x^2 l_0 + 1 = 0 , [ mf_steady ] whose approximate solution is ( l_0 ) _ ( 1 + 14 _ ^ 2 l_0 + 1 ( 2 l_0 + 1 ) ) + o(_^4l_0 ) , with @xmath143 , which is the solution for infinite @xmath12 . by the same reasoning
, we can find @xmath86 and @xmath85 , whose approximate solutions read _ 0(l_0 ) & & 4 + o(_^4l_0 ) , + _ 0 ( l_0 ) & & 4 - 3 4 + o(_^4l_0 ) . for finite @xmath12 , the most relevant terms arise from the second line of eq .
( [ km_of_height ] ) and the resulting equation is the kpz one , with coefficients @xmath144 we give the numerical values of @xmath93 , @xmath145 , and @xmath146 in the table .
these numbers are determined from the direct calculation of @xmath93 , @xmath86 and @xmath85 using eq .
( [ km_of_height ] ) .
these @xmath146 s should be compared with the previous numerical results@xcite .
note that @xmath146 in ref.@xcite is one half of @xmath146 here . for infinite @xmath12 , the second line of eq .
( [ km_of_height ] ) vanishes , of which the physical meaning is that a dropped particle eventually find a stable site .
hence the continuum equation becomes the vld equation with coefficients @xmath147
this appendix proves that the first line of eq .
( [ km_of_height ] ) does not generate terms @xmath149 , where @xmath150 is a non - negative integer . to this end , consider a taylor expansion of @xmath128 and set the lattice constant to be @xmath52 .
we call this quantity @xmath151 , and the corresponding quantity for @xmath81 @xmath152 .
one should not confuse this procedure with the continuum limit .
it is enough to prove that @xmath153 .
it is clear that @xmath151 has no @xmath77 dependence because the lattice constant is set to @xmath52 .
this fact enables us to write the recursion relation of @xmath151 in the symmetric form 2 t_l^l_0 ( d ) = s_l_0(d ) t_l-1^l_0 ( d ) & & + ( 1 - s_l_0(d)+d 2 ) ( s_l_0(d ) - d 2 ) ^ l + & & + ( 1 - s_l_0(d)-d 2 ) ( s_l_0(d ) + d 2 ) ^ l. hence we see that 2 ( t_l^l_0 ( d ) - t_l^l_0 ( -d ) ) = s_l_0(d ) ( t_l-1^l_0 ( d ) - t_l-1^l_0 ( -d ) ) the mirror symmetry of this system restricts the form of @xmath154 to be even in @xmath80 , so @xmath155 implies @xmath156 .
indeed @xmath157 and the logic of the mathematical induction proves the disappearance of the odd powers of @xmath80 in @xmath158 .
the vanishing @xmath159 in appendix [ rj ] is a consequence of this property . as a result
, we can safely affirm that crsos model is described by the vld equation . | we study the restricted solid - on - solid ( rsos ) model with finite hopping distance @xmath0 , using both analytical and numerical methods .
analytically , we use the hard - core bosonic field theory developed by the authors [ phys .
rev .
e * 62 * , 7642 ( 2000 ) ] and derive the villain - lai - das sarma ( vld ) equation for the @xmath1 case which corresponds to the conserved rsos ( crsos ) model and the kardar - parisi - zhang ( kpz ) equation for all finite values of @xmath0 .
consequently , we find that the crsos model belongs to the vld universality class and the rsos models with any finite hopping distance belong to the kpz universality class .
there is no phase transition at a certain finite hopping distance contrary to the previous result .
we confirm the analytic results using the monte carlo simulations for several values of the finite hopping distance . |
the non - trivial vacuum structure of su(2 ) yang - mills - higgs ( ymh ) theory allows for the existence of regular non - perturbative finite energy solutions , such as monopoles , multimonopoles and monopole - antimonopole systems .
while spherically symmetric monopoles carry unit topological charge @xcite , monopoles with charge @xmath7 are axially symmetric @xcite or possess no rotational symmetry at all @xcite . in the bogomolnyi - prasad - sommerfield ( bps ) limit of vanishing higgs potential , the monopole and
multimonopole solutions satisfy a set of first order equations , the bogomolnyi equations @xcite . the spherically symmetric and axially symmetric bps ( multi)monopole solutions are known analytically @xcite . in these solutions
all nodes of the higgs field are superimposed at a single point . in bps
multimonopole solutions with only discrete symmetries , recently constructed numerically , the nodes of the higgs field can be located at several isolated points @xcite .
the energy of bps solutions satisfies exactly the lower energy bound given by the topological charge .
in particular , in the bps limit the repulsive and attractive forces between monopoles exactly compensate , thus bps monopoles experience no net interaction @xcite .
the configuration space of ymh theory consists of topological sectors characterized by the topological charge of the higgs field .
as shown by taubes @xcite , each topological sector contains besides the ( multi)monopole solutions further regular , finite energy solutions , which do not satisfy the first order bogomolnyi equations , but only the set of second order field equations .
these solutions form saddlepoints of the energy functional @xcite . the simplest such solution , the monopole
antimonopole pair solution @xcite , is topologically trivial .
it possesses axial symmetry , and the two nodes of its higgs field are located symmetrically on the positive and negative @xmath8-axis .
recently we have constructed axially symmetric saddlepoint solutions , where the higgs field vanishes at @xmath9 isolated points on the symmetry axis .
these represent chains of single monopoles and antimonopoles located on the symmetry axis in alternating order @xcite . for an equal number of monopoles and antimonopoles ,
i.e. for even @xmath1 , the chains reside in the topologically trivial sector .
when the number of monopoles exceeds the number of antimonopoles by one , i.e. for odd @xmath1 , the chains reside in the sector with topological charge one . while such chains can also be formed from doubly charged monopoles and antimonopoles ( @xmath4 ) , chains with higher charged monopoles ( @xmath10 )
can not be formed .
instead completely different solutions appear , where the higgs field vanishes on one or more rings centered around the symmetry axis @xcite .
we therefore refer to these solutions as vortex rings .
when gravity is coupled to ymh theory , this has significant effect on the monopole , multimonopole and monopole - antimonopole pair solutions .
when the coupling constant is increased from zero @xcite , a branch of gravitating monopole solutions emerges smoothly from the flat space t hooft - polyakov monopole solution .
the coupling constant @xmath11 , entering the einstein - yang - mills - higgs ( eymh ) equations , is proportional to the higgs vacuum expectation value @xmath12 and the square root of the gravitational constant @xmath13 and to the square of the higgs vacuum expectation value @xmath12 .
with increasing coupling constant @xmath11 the mass of the gravitating monopole solutions decreases .
the branch of gravitating monopole solutions extends up to a maximal value @xmath14 , beyond which gravity becomes too strong for regular monopole solutions to persist @xcite . for vanishing higgs self - coupling constant ,
this first gravitating monopole branch merges with a second branch at @xmath14 , which extends slightly backwards , until at a critical value @xmath15 of the coupling constant a degenerate horizon develops @xcite .
the exterior space time of the solution then corresponds to the one of an extremal reissner - nordstrm ( rn ) black hole with unit magnetic charge @xcite . at @xmath15 this second branch of gravitating monopole solutions
thus bifurcates with the branch of extremal reissner - nordstrm ( rn ) solutions .
remarkably , the additional attraction in the ymh system due to the coupling to gravity also allows for bound monopoles , not present in flat space @xcite . for the monopole - antimonopole pair solution ,
on the other hand , we observe a different coupling constant dependence for the gravitating solutions @xcite . again
a branch of gravitating monopole - antimonopole pair solutions emerges smoothly from the flat space solution , and merges at a maximal value of the coupling constant @xmath14 with a second branch of gravitating monopole - antimonopole pair solutions .
this second branch , however , extends all the way back to vanishing coupling constant . along the first branch
the mass of the solutions decreases with increasing @xmath11 , since with increasing gravitational strength the attraction in the system increases . along the second branch
the mass increases strongly with decreasing coupling constant and diverges in the limit of vanishing coupling constant .
this upper branch may be considered as being obtained by decreasing the higgs expectation value . along the upper branch
the solutions shrink to zero size in the limit @xmath16 . scaling the coordinates , the mass and the higgs field by @xmath11 shows , that the scaled solutions approach the lowest mass bartnik - mckinnon ( bm ) solution @xcite of einstein - yang - mills ( eym ) theory in the limit @xmath16 . here
we consider the effect of gravity on the monopole - antimonopole chains and vortex rings @xcite .
characterizing these solutions by the integers @xmath1 and @xmath2 , related to the polar angle @xmath17 and the azimuthal angle @xmath18 , respectively , we first consider chains , where @xmath19 , and then vortex rings , where @xmath20 .
when @xmath1 is even , these solutions reside in the topologically trivial sector , thus we expect a similar dependence on the coupling constant as for the monopole - antimonopole pair . when @xmath1 is odd ,
the solutions reside in the sector with charge @xmath2 , where the corresponding multimonopoles reside
. therefore at first sight a connection with rn solutions appears to be possible , similar as observed for gravitating ( multi)monopoles .
we find , however , that all these unstable gravitating solutions , chains and vortex rings alike , show the same general coupling constant dependence as observed for the monopole - antimonopole pair .
two branches of solutions arise , a lower branch connected to the flat space solution and an upper branch connected to an eym solution . for @xmath21 these eym solutions
correspond to the spherically symmetric bm solutions ( @xmath3 ) or their axially symmetric generalizations ( @xmath7 ) @xcite . for @xmath22 new eym solutions
arise in the limit @xcite , implying further branches of gravitating eymh solutions .
in this paper we present gravitating monopole - antimonopole chains and vortex rings for vanishing higgs self - coupling . in section
ii we present the action , the axially symmetric ansatz and the boundary conditions . in section
iii we discuss the properties of these solutions with particular emphasis on the limit @xmath16 , and we present our conclusions in section iv .
we consider static axially symmetric monopole - antimonopole chains and vortex rings in su(2 ) einstein - yang - mills - higgs theory with action @xmath23 with curvature scalar @xmath24 , @xmath25 field strength tensor @xmath26 \ , \ ] ] gauge potential @xmath27 , and covariant derivative of the higgs field @xmath28 in the adjoint representation @xmath29 \ .\ ] ] here @xmath13 and @xmath30 denote the gravitational and gauge coupling constants , respectively , @xmath12 denotes the vacuum expectation value of the higgs field , and @xmath31 the strength of the higgs self - coupling . under su(2 )
gauge transformations @xmath32 , the gauge potentials transform as @xmath33 and the higgs field transforms as @xmath34 the nonzero vacuum expectation value of the higgs field breaks the non - abelian su(2 ) gauge symmetry to the abelian u(1 ) symmetry .
the particle spectrum of the theory then consists of a massless photon , two massive vector bosons of mass @xmath35 , and a massive scalar field @xmath36 . in the bps limit
the scalar field also becomes massless , since @xmath37 , i.e. the higgs potential vanishes .
variation of the action ( [ action ] ) with respect to the metric @xmath38 leads to the einstein equations @xmath39 with stress - energy tensor @xmath40 variation with respect to the gauge field @xmath41 and the higgs field @xmath42 leads to the matter field equations , @xmath43 = 0 \ , \label{feqa } \\ & & \frac{1}{\sqrt{-g } } d_\mu(\sqrt{-g } d^\mu \phi ) + \lambda ( \phi^2 -\eta^2 ) \phi = 0 \ , \label{feqphi}\end{aligned}\ ] ] respectively . to obtain gravitating static axially symmetric solutions , we employ isotropic coordinates @xcite . in terms of the spherical
coordinates @xmath44 , @xmath17 and @xmath45 the isotropic metric reads @xmath46 where the metric functions @xmath47 , @xmath1 and @xmath48 are functions of the coordinates @xmath44 and @xmath17 , only .
the @xmath8-axis ( @xmath49 ) represents the symmetry axis .
regularity on this axis requires @xcite @xmath50 we parametrize the gauge potential and the higgs field by the ansatz @xcite @xmath51 where the @xmath25 matrices @xmath52 , @xmath53 , and @xmath54 are defined as products of the spatial unit vectors @xmath55 with the pauli matrices @xmath56 , i.e. @xmath57 with @xmath58 . for @xmath59 , @xmath3
the ansatz corresponds to the one for the monopole - antimonopole pair solutions @xcite , while for @xmath60 , @xmath7 it corresponds to the ansatz for axially symmetric multimonopoles @xcite .
the four gauge field functions @xmath61 and two higgs field functions @xmath62 depend on the coordinates @xmath44 and @xmath17 , only .
all functions are even or odd w.r.t . reflection symmetry , @xmath63 .
the gauge transformation @xmath64 leaves the ansatz form - invariant @xcite . to construct regular solutions
we have to fix the gauge @xcite . here
we impose the gauge condition @xcite @xmath65 with this ansatz the equations of motion reduce to a set of 9 coupled partial differential equations , to be solved numerically subject to the set of boundary conditions , discussed below . to obtain globally regular asymptotically flat solutions with the proper symmetries
, we must impose appropriate boundary conditions @xcite .
_ boundary conditions at the origin _
regularity of the solutions at the origin ( @xmath66 ) requires for the metric functions the boundary conditions @xmath67 whereas the gauge field functions @xmath61 satisfy @xmath68 and the higgs field functions @xmath62 satisfy @xmath69 @xmath70 \right|_{r=0 } = 0 \ , \ ] ] i.e. @xmath71 , @xmath72 .
_ boundary conditions at infinity _
asymptotic flatness imposes on the metric functions of the solutions at infinity ( @xmath73 ) the boundary conditions @xmath74 considering the gauge field at infinity , we require that solutions in the vacuum sector @xmath75 , where @xmath76 , tend to a gauge transformed trivial solution , @xmath77 and that solutions in the sector with topological charge @xmath2 , where @xmath78 , tend to @xmath79 where @xmath80 is the asymptotic solution of a charge @xmath2 multimonopole , and @xmath81 , both for even and odd @xmath1 . in terms of the functions
@xmath82 , @xmath83 , @xmath84 these boundary conditions read @xmath85 @xmath86 @xmath87 @xmath88 _ boundary conditions along the symmetry axis _
the boundary conditions along the @xmath8-axis ( @xmath89 and @xmath90 ) are determined by the symmetries .
the metric functions satisfy along the axis @xmath91 whereas the matter field functions satisfy @xmath92
we have constructed numerically gravitating chains and vortex rings , subject to the above boundary conditions .
we first briefly address the numerical procedure , and then present our results for the chains ( @xmath3 , 2 ) and vortex rings ( @xmath20 ) . in particular , we consider the dependence of the solutions on the value of the coupling constant @xmath11 , and restrict to vanishing higgs self - coupling @xmath93 . to construct solutions subject to the above boundary conditions , we map the infinite interval of the variable @xmath44 onto the unit interval of the compactified radial variable @xmath94 $ ] , @xmath95 i.e.
, the partial derivative with respect to the radial coordinate changes according to @xmath96 the numerical calculations are performed with the help of the fidisol package based on the newton - raphson iterative procedure @xcite .
it is therefore essential for the numerical procedure to have a reasonably good initial configuration .
( for details see description and related documentation @xcite . )
the equations are discretized on a non - equidistant grid in @xmath97 and @xmath17 with typical grids sizes of @xmath98 .
the estimates of the relative error for the functions are of the order of @xmath99 , @xmath100 and @xmath101 for solutions with @xmath59 , @xmath102 and @xmath103 , respectively .
let us introduce the dimensionless coordinate @xmath97 , @xmath104 the equations then depend only on the dimensionless coupling constant @xmath11 , @xmath105 the mass @xmath106 of the solutions is related to the dimensionless mass @xmath107 , via @xmath108 where @xmath107 is determined by the derivative of the metric function @xmath47 at infinity @xcite @xmath109 let us consider first monopole - antimonopole chains composed of poles of charge @xmath3 .
the topological charge of these chains is either zero ( for even @xmath1 ) or unity ( for odd @xmath1 ) .
the monopole - antimonopole chains possess @xmath1 nodes of the higgs field on the @xmath8-axis . because of reflection symmetry , to each node on the positive @xmath8-axis
there corresponds a node on the negative @xmath8-axis . for even @xmath1 ( @xmath76 ) the higgs field does not have a node at the origin , while for odd @xmath1 ( @xmath78 ) it does .
associating with each pole the location of a single magnetic charge , these chains then possess a total of @xmath1 magnetic monopoles and antimonopoles , located in alternating order on the symmetry axis @xcite .
the dependence of the @xmath60 solution , i.e. of the t hooft - polyakov monopole , on gravity has been studied before .
when gravity is coupled , a branch of gravitating monopoles emerges from the flat space solution and extends up to a maximal value @xmath14 , where it merges with a second branch .
this second branch extends slightly backwards , and bifurcates at a critical value @xmath15 with the branch of extremal rn solutions of unit charge .
the @xmath59 chain , i.e. the monopole - antimonopole pair , shows a different dependence on the coupling constant @xcite .
again , when gravity is coupled , a branch of gravitating monopole - antimonopole pair solutions emerges from the flat space solution , and merges with a second branch of monopole - antimonopole pair solutions at a maximal value of the coupling constant @xmath14 .
but since the monopole - antimonopole pair solutions reside in the topologically trivial sector and thus carry no magnetic charge , this upper branch of solutions can not bifurcate with a branch of extremal rn solutions at a critical value @xmath15 of the coupling constant .
instead the upper branch extends all the way back to @xmath110 . along the first branch
the mass of the solutions decreases with increasing @xmath11 , since with increasing gravitational strength the attraction in the system increases . along the second branch , in contrast , the mass increases strongly with decreasing @xmath11 , and the solutions shrink correspondingly . in the limit of vanishing coupling constant the mass then diverges and the solutions shrink to zero size . scaling the coordinates and the higgs field ,
however , via @xmath111 leads to a limiting solution with finite size and finite scaled mass @xmath112 @xcite , @xmath113 indeed , after the scaling the field equations do not depend on @xmath11 .
instead @xmath11 appears in the asymptotic boundary conditions of the higgs field , @xmath114 .
the higgs field then becomes trivial on the upper branch as @xmath16 .
consequently , the scaled monopole - antimonopole solutions approach an eym solution as @xmath16 .
this limiting solution is the lowest mass bm solution @xcite .
the mass and the scaled mass of the monopole - antimonopole pair solutions are exhibited in fig . 1 .
let us now consider chains with even @xmath1 , @xmath9 , which , like the @xmath59 monopole - antimonopole pair , also reside in the vacuum sector .
as one may expect , these chains exhibit an analogous dependence on the coupling constant @xmath11 as the monopole - antimonopole pair . in the limit @xmath16 , a lower branch of gravitating solutions emerges from the flat space solution , and merges at a maximal value @xmath14 with an upper branch of solutions , which extends all the way back to @xmath16 .
the mass of the chains with @xmath115 and @xmath116 is exhibited in fig .
here , for @xmath16 , the emergence of the lower branches from the ymh limit is seen as well as the divergence along the upper branches . in fig .
1b the scaled mass is exhibited . clearly , the scaled mass of these solutions approaches the mass of the lowest bm solution in the limit @xmath16 along the upper branch .
the values of the metric functions @xmath47 and @xmath48 at the origin are shown in fig . 2 for both branches of the chains with @xmath115 and @xmath116 .
they evolve from unity in the flat space limit and reach the corresponding values of the lowest mass bm solution , when @xmath16 on the upper branch .
let us now consider the nodes of the higgs field for these even @xmath1 chains . in fig .
3a we recall the dependence of the single positive node of the monopole - antimonopole pair on the coupling constant .
starting from the flat space limit , the location of the node continuously moves inwards along both branches , and reaches the origin in the @xmath16 limit on the upper branch .
( we do not consider excited monopole - antimonopole pair solutions with several nodes here , which are related to excited bartnik - mckinnon solutions @xcite . )
the flat space @xmath115 solution has two positive nodes .
the evolution of these nodes along both branches is exhibited in fig .
the location of the inner node again moves continuously inwards along both branches , and reaches the origin in the limit @xmath16 on the upper branch .
in contrast , the location of the outer node reaches a finite limiting value in the limit @xmath16 on the upper branch , which , interestingly , agrees with the location of the single node of the flat space monopole - antimonopole pair solution .
similarly , as seen in fig . 3c , of the 3 positive nodes of the @xmath116 solution the location of the innermost node reaches the origin in the limit @xmath16 on the upper branch , whereas the locations of the other two nodes reach finite limiting values , which agree with the locations of the two nodes of the flat space @xmath115 chain . to obtain a better understanding of the limit @xmath16 on the upper branch for these chains ,
let us consider the matter and metric functions for small values of the coupling constant . in fig . 4 ,
for instance , we exhibit the gauge field function @xmath117 , the modulus of the higgs field @xmath118 and the metric function @xmath47 for the @xmath115 chain for a small value of the coupling constant , @xmath119 , on the upper branch .
we observe , that we have to distinguish two regions for the matter function @xmath117 , an outer region and an inner region . in the outer region ,
the gauge field function @xmath117 of the @xmath115 chain agrees well with the ( by 2 shifted ) corresponding gauge field function @xmath117 of the flat space ( @xmath59 ) monopole - antimonopole pair solution , while in the inner region the gauge field function @xmath117 of the @xmath115 chain agrees well with the corresponding function of the bm solution ( in by @xmath11 scaled coordinates ) .
likewise , in the outer region the modulus of the higgs field @xmath118 of the @xmath115 chain agrees well with the modulus of the higgs field of the flat space ( @xmath59 ) monopole - antimonopole pair solution . in the bm solution , of course
, no higgs field is present .
( an analogous pattern holds for the functions @xmath120 , @xmath121 , @xmath122 , @xmath83 and @xmath84 after a gauge transformation . ) the metric function @xmath47 , on the other hand , agrees well with the metric function of the bm solution ( in by @xmath11 scaled coordinates ) everywhere .
( the same holds for the functions @xmath1 and @xmath48 . )
the metric of the flat space ( @xmath59 ) monopole - antimonopole pair solution is , of course , trivial .
thus , for small @xmath11 on the upper branch , the @xmath115 chain may be thought of as composed of a scaled bm solution in the inner region and a flat space ( @xmath59 ) monopole - antimonopole pair solution in the outer region .
taking the limit @xmath16 then , the inner region shrinks to zero size , yielding a solution which is singular at the origin and which has infinite mass , while elsewhere it corresponds to the flat space ( @xmath59 ) monopole - antimonopole pair solution .
this explains the above observation concerning the location of the higgs field nodes , made in fig .
3 . on the other hand , taking the limit @xmath16 with scaled coordinates , the inner region covers all of space and contains the bm solution with its finite scaled mass , while the flat space ( @xmath59 ) monopole - antimonopole pair solution is shifted all the way out to infinity . for the @xmath116 chain
, we observe an analogous pattern for the functions in the limit @xmath16 on the upper branch , as exhibited in fig . 5 .
in the outer region , the gauge field function @xmath117 of the @xmath116 chain agrees well with the ( by 2 shifted ) corresponding gauge field function @xmath117 of the flat space @xmath115 chain , while in the inner region it agrees well with the corresponding bm function ( in by @xmath11 scaled coordinates ) .
likewise , in the outer region the modulus of the higgs field @xmath118 of the @xmath116 chain agrees well with the modulus of the higgs field of the flat space @xmath115 chain , while the metric function @xmath47 of the @xmath116 chain agrees well with the metric function of the bm solution ( in by @xmath11 scaled coordinates ) everywhere .
turning now to monopole - antimonopole chains in the topological sector with unit charge , one might expect a completely different pattern for the coupling constant dependence of the solutions , namely a pattern which would be similar to the coupling constant dependence of the gravitating monopoles @xcite , where the upper branch of gravitating monopoles bifurcates with the branch of extremal rn solutions @xcite .
such an expectation does not prove true , however .
instead , the odd @xmath1 chains show a similar pattern as the even @xmath1 chains . two branches of gravitating @xmath1 chain solutions merge at a maximal value @xmath14 .
the lower branch of solutions emerges from the flat space @xmath1 chain , and the upper branch extends all the way back to @xmath123 . for small values of @xmath11 on the upper branch , the solutions may be thought of as composed of a scaled bm solution in the inner region and a flat space @xmath124 solution in the outer region . for the @xmath125 and @xmath126 chains , we exhibit the mass and scaled mass in fig . 1 , the values of the metric functions @xmath47 and @xmath48 at the origin in fig . 2 , and the location of the positive node(s ) , in fig .
note , that in the limit @xmath16 on the upper branch , the location of the outer positive node of the @xmath126 chain agrees with the location of the positive node of the flat space @xmath125 chain .
the composition of the solutions in terms of a scaled bm solution in the inner region and a flat space @xmath124 solution in the outer region is illustrated in fig .
6 , where we exihibit the gauge field function @xmath117 for the @xmath125 and @xmath126 chains . for the @xmath125 chain
clearly the gauge field function corresponds in the outer region to the function of a flat space ( @xmath60 ) monopole , while for the @xmath126 chain it corresponds to a flat space @xmath125 chain .
concluding , we observe the same general pattern for all gravitating chain solutions with @xmath3 . from the flat space @xmath1 chain a lower branch of gravitating @xmath1 chains emerges , which merges at a maximal value @xmath14 with an upper branch .
the value of @xmath14 decreases with increasing @xmath1 . for small @xmath11 the solutions on the upper branch
may be thought of as composed of a scaled bm solution in the inner region and a flat space @xmath124 solution in the outer region .
consequently , the mass diverges in the limit @xmath16 , while the scaled mass approaches the mass of the lowest bm solution of eym theory . for the @xmath1 chains composed of monopoles and antimonopoles with charge @xmath4 @xcite
, we observe a completely analogous pattern as for the chains composed of monopoles and antimonopoles with charge @xmath3 . for small @xmath11 the solutions on the upper branch may be thought of as composed of a scaled generalized bm solution with @xmath4 in the inner region @xcite , and a flat space @xmath124 solution in the outer region .
consequently , the mass again diverges in the limit @xmath16 , while the scaled mass approaches the mass of the generalized bm solution with @xmath4 .
we illustrate this behaviour in fig . 7 for the @xmath115 chain with @xmath4
clearly , the upper branch of the scaled mass approaches the mass of the generalized bm solution with @xmath4 , and the values of the metric functions @xmath47 and @xmath48 at the origin approach those of the generalized bm solution with @xmath4 as well .
the composition of the solutions on the upper branch for small values of @xmath11 in terms of an inner generalized bm solution with @xmath4 and an outer flat space @xmath59 solution is exhibited in fig .
8 for the gauge field function @xmath117 , the modulus of the higgs field @xmath42 and the metric function @xmath47 . for larger values of @xmath2
the flat space solutions completely change character , since the structure of the nodes of the higgs field in these solutions is totally different @xcite . whereas @xmath1 chains ( with @xmath19 ) possess only isolated nodes on the symmetry axis , for the solutions with @xmath20 the modulus of the higgs field vanishes on rings in the @xmath128-plane or in planes parallel to the @xmath128-plane , centered around the @xmath8 axis .
for even @xmath1 the solutions possess only such vortex rings and no nodes on the symmetry axis .
for odd @xmath1 they possess vortex rings as well as a node at the origin , where a charge @xmath2-monopole is located .
let us now consider the effect of gravity on vortex solutions with @xmath127 . here
the same general pattern is seen for the coupling constant dependence as for the gravitating chains . from the flat space vortex solution
a lower branch of gravitating vortex solutions emerges , and merges at a maximal value @xmath14 with an upper branch . on the upper branch , for small values of @xmath11
the solutions may be thought of as composed of a scaled generalized bm solution with @xmath127 in the inner region and a flat space @xmath124 solution in the outer region .
thus the mass diverges in the limit @xmath16 on the upper branch , while the scaled mass approaches the mass of the generalized bm solution with @xmath127 .
it is now tempting to conclude , that for arbitrary values of @xmath1 and @xmath2 the same general pattern holds : from the flat space ymh solution a lower branch of gravitating eymh solutions emerges , which merges at a maximal value @xmath14 with an upper branch . on the upper branch , for small values of @xmath11 the solutions may be thought of as composed of a scaled generalized bm solution in the inner region and a flat space @xmath124 solution in the outer region .
this conjectured pattern is indeed observed for the @xmath59 and @xmath125 eymh solutions with @xmath129 . for @xmath129 and @xmath115 , however , a surprise in encountered . as illustrated in fig . 7 , on the upper branch the scaled mass of the @xmath129 and @xmath115 solutions does not approach the mass of the generalized bm solution with @xmath129 @xcite , but instead it approaches the mass of a new axially symmetric @xmath129 eym solution @xcite .
moreover , a second new axially symmetric @xmath129 eym solution exists , which is slightly higher in mass @xcite .
this second eym solution constitutes the endpoint of a second upper branch of @xmath129 and @xmath115 eymh solutions , which merges at a second maximal value of the coupling constant @xmath11 with a second lower branch , and it is along this second lower branch , that for small values of @xmath11 the solutions may be thought of as composed of a scaled generalized bm solution with @xmath129 in the inner region and a flat space @xmath59 solution in the outer region .
thus for @xmath129 and @xmath115 there are four branches of solutions instead of two , and consequently there are four limiting solutions when @xmath16 .
the masses of the generalized bm solutions and of the new eym solutions are exhibited in fig .
note , that the boundary conditions of the new eym solutions differ from those of the generalized bm solutions at infinity @xcite .
the eym solutions can thus be classified by these boundary conditions . denoting @xmath76 ,
the new solutions have @xmath130 , whereas the generalized bm solutions have @xmath131 @xcite . in fig .
10 we illustrate the dependence of the location of the vortex of the higgs field on the coupling constant for these four branches of @xmath129 and @xmath115 solutions .
the flat space solution has two vortex rings , located symmetrically in planes parallel to the @xmath128 plane .
starting from the flat space location , the vortex rings move continuously inwards and towards the @xmath128 plane along the first lower branch and the first upper branch , until at a critical value of the coupling constant the two rings merge in the @xmath128 plane .
when @xmath11 is decreased further along the first upper branch , a bifurcation takes place and two distinct rings appear in the @xmath128 plane . in the limit
@xmath16 they shrink to zero size . in scaled coordinates , in contrast
, they reach a finite limiting size . likewise , in scaled coordinates there are two distinct rings with finite size on the second upper branch ( located not too far from the two rings of the first upper branch ) . following the second upper branch and then the second lower branch
, the inner ring shrinks to zero size in the limit @xmath16 ( in ordinary coordinates ) , while the outer ring approaches the location of the flat space vortex ring of the @xmath129 and @xmath59 solution .
summarizing , we find the following pattern : from the flat space @xmath115 and @xmath129 vortex solution the first lower branch of gravitating vortex solutions emerges , which merges at the first maximal value @xmath14 with the first upper branch .
along the first upper branch , for small values of @xmath11 the solutions approach the scaled @xmath130 and @xmath129 eym solution with lower mass .
close to the first upper branch , for small values of @xmath11 the second upper branch arises from the scaled @xmath130 and @xmath129 eym solution with higher mass . this second upper branch merges at the second smaller maximal value @xmath14 with the second lower branch of solutions . along the second lower branch , for small values of @xmath11 the solutions
may be thought of as composed of a scaled @xmath131 and @xmath129 eym solution in the inner region and a flat space @xmath59 and @xmath129 solution in the outer region .
as seen in fig .
9 , no eym solutions with @xmath132 and @xmath129 appear .
we therefore conjecture , that for solutions with @xmath5 and @xmath129 an analogous but generalized pattern holds .
there are four branches of solutions . from the flat space @xmath5 and @xmath129 vortex solution
the first lower branch emerges , and merges at a first maximal value @xmath14 with the first upper branch .
along the first upper branch , the solutions approach a composite solution for small values of @xmath11 , consisting of the scaled @xmath130 and @xmath129 eym solution with lower mass in the inner region and an @xmath133 and @xmath129 ymh solution in the outer region .
the second upper branch arises from a composite solution consisting of the scaled @xmath130 and @xmath129 eym solution with higher mass in the inner region and an @xmath133 and @xmath129 ymh solution in the outer region .
the second upper branch merges at a second maximal value @xmath14 with the second lower branch of solutions . along the second lower branch , for small values of @xmath11 the solutions
may be thought of as composed of a scaled @xmath131 and @xmath129 eym solution in the inner region and a flat space @xmath124 and @xmath129 solution in the outer region . for vortex solutions with @xmath134
we expect an analogous pattern as for the @xmath129 vortex solutions .
in particular , there should be four branches of solutions : a first lower branch emerging from the flat space solution , a first upper branch approaching a composite solution consisting of the lower scaled @xmath130 eym solution and a @xmath133 ymh solution , a second upper branch arising from a composite solution consisting of the higher scaled @xmath130 eym solution and a @xmath133 ymh solution , and a second lower branch approaching a composite solution consisting of the scaled @xmath131 eym solution and a @xmath124 ymh solution .
we illustrate this pattern for the @xmath126 and @xmath134 vortex solutions in fig . 11 , where we show the gauge field function @xmath117 for solutions on all four branches , for small values of @xmath11 .
the scaled mass along the four branches is shown in fig .
11c . considering higher values of @xmath2
, we observe in fig .
9 , that for @xmath135 and @xmath136 at least five eym solutions exist : one @xmath131 , two @xmath130 and two @xmath137 solutions .
this suggests , that for @xmath135 and @xmath116 six branches of gravitating vortex solutions exist : a first lower branch emerging from the @xmath135 flat space solution , a first upper branch approaching the lower mass scaled @xmath137 and @xmath135 eym solution , a second upper branch arising from the higher mass scaled @xmath137 and @xmath135 eym solution , a second lower branch approaching a composite solution consisting of the lower mass scaled @xmath130 and @xmath135 eym solution and a @xmath133 and @xmath135 ymh solution , a third upper branch arising from a composite solution consisting of the higher mass scaled @xmath130 and @xmath135 eym solution and a @xmath133 and @xmath135 ymh solution , and a third lower branch approaching a composite solution consisting of the scaled @xmath131 and @xmath135 eym solution and a @xmath124 and @xmath135 ymh solution .
we indeed obtain these six branches numerically . surprisingly , there are two more branches for a small range of @xmath11 .
we exhibit the scaled mass along these branches in fig .
the value of the metric function @xmath48 at the origin is exhibited in fig .
we have considered the effect of gravity on the monopole - antimonopole chains and vortex rings of yang - mills - higgs theory @xcite .
the resulting solutions are regular , static , axially symmetric and asymptotically flat .
they are characterized by two integers @xmath1 and @xmath2 . solutions with @xmath3 and @xmath4
correspond to gravitating chains of @xmath1 monopoles and antimonopoles , each of charge @xmath138 , where the higgs field vanishes at @xmath1 isolated points along the symmetry axis .
larger values of @xmath2 give rise to gravitating vortex solutions , where the higgs field vanishes on one or more rings , centered around the symmetry axis .
concerning their dependence on the coupling constant @xmath11 , we observe the same general pattern for all gravitating chain solutions . from the flat space @xmath1 chain a lower branch of gravitating @xmath1 chains
emerges , which merges at a maximal value @xmath14 with an upper branch . for small @xmath11 on the upper branch , the solutions may be thought of as composed of a scaled ( generalized ) bm solution in the inner region and a flat space @xmath124 solution in the outer region .
consequently , the mass diverges in the limit @xmath16 , while the scaled mass approaches the mass of the lowest ( generalized ) bm solution of eym theory .
the same pattern holds for gravitating vortex solutions with @xmath127 , and for @xmath139 when @xmath21 . for @xmath129 and @xmath22
, however , a new phenomenon occurs . instead of two branches of solutions four branches of solutions arise .
the reason for these is the presence of two additional eym solutions @xcite . in this case
we find ( and conjecture for higher @xmath1 ) the following pattern for the coupling constant dependence : a first lower branch emerges from the flat space solution , a first upper branch approaches a composite solution consisting of the lower scaled @xmath130 eym solution and a @xmath133 ymh solution , a second upper branch arises from a composite solution consisting of the higher scaled @xmath130 eym solution and a @xmath133 ymh solution , a second lower branch approaches a composite solution consisting of the scaled @xmath131 eym solution and a @xmath124 ymh solution
. for @xmath140 and @xmath136 not only two but four additional eym solutions arise @xcite .
we consequently find ( and conjecture for higher @xmath1 ) six branches of gravitating vortex solutions , and for a small range of @xmath11 even eight branches . for larger values of @xmath2
we expect this trend to continue .
we observe that for all the gravitating monopole - antimonpole pair solutions , chains of monopoles and antimonopoles and vortex rings we considered in this paper there is no formation of a degenerate horizon , in contrast to the gravitating monopole solution @xcite .
we note that for gravitating skyrmions @xcite the dependence of the solutions on the gravitational coupling parameter follows the same pattern as for the monopole - antimonpole pair .
again , two branches of solutions merge at the maximal value of the coupling parameter , and the upper branch is connected to the ( scaled ) lowest mass bm solution . in this paper
we have only considered gravitating chain and vortex solutions in the limit of vanishing higgs self - coupling . for finite values of the self - coupling constant
the structure of the flat space solutions can become more involved , yielding solutions with vortex rings and several isolated nodes of the higgs field @xcite .
their continuation to curved space remains to be studied .
furthermore , it would be interesting to consider dyonic gravitating chain and vortex solutions @xcite , as well as black holes with monopole or dipole hair , consisting of chain - like or vortex - like structures @xcite . finally , recent work on gravitating monopoles and monopole - antimonopole pairs in the presence of a cosmological constant might be generalized to the study of gravitating chain and vortex solutions @xcite . e.j .
weinberg , and a.h .
guth , phys .
* d14 * ( 1976 ) 1660 . c. rebbi , and p. rossi , phys
* d22 * ( 1980 ) 2010 . r.s .
ward , comm .
* 79 * ( 1981 ) 317 ; + p. forgacs , z. horvath , and l. palla , phys . lett . * 99b * ( 1981 ) 232 ; + m.k .
prasad , comm .
phys . * 80 * ( 1981 ) 137 ; + m.k .
prasad , and p. rossi , phys .
* d24 * ( 1981 ) 2182 .
b. kleihaus , j. kunz , and ya .
shnir , phys . lett . *
b570 * , ( 2003 ) 237 ; + b. kleihaus , j. kunz , and ya .
shnir , phys . rev .
* d68 * ( 2003 ) 101701 ; + b. kleihaus , j. kunz , and ya .
shnir , phys . rev . * d70 * ( 2004 ) 065010 .
k. lee , v.p .
nair , and e.j .
weinberg , phys .
* d45 * ( 1992 ) 2751 ; + p. breitenlohner , p. forgacs , and d. maison , nucl . phys . *
b383 * ( 1992 ) 357 ; + p. breitenlohner , p. forgacs , and d. maison , nucl . phys . *
b442 * ( 1995 ) 126 .
b. kleihaus , and j. kunz , phys .
* 78 * ( 1997 ) 2527 ; + b. kleihaus , and j. kunz , phys .
* 79 * ( 1997 ) 1595 ; + b. kleihaus , and j. kunz , phys . rev .
* d57 * ( 1998 ) 834 ; + b. kleihaus , and j. kunz , phys . rev .
* d57 * ( 1998 ) 6138 .
b. hartmann , b. kleihaus , and j. kunz , mod .
a * 15 * ( 2000 ) 1003 ; + y. brihaye , b. hartmann , and j. kunz , phys .
* b441 * ( 1998 ) 77 ; + y. brihaye , b. hartmann , j. kunz , and n. tell , phys
* d60 * ( 1999 ) 104016 .
e. radu , and d.h .
tchrakian , hep - th/0411084 .
h. luckock , and i. moss , phys . lett .
* b176 * ( 1986 ) 341 ; + p. bizon and t. chmaj , phys.lett . *
b297 * ( 1992 ) 55 ; + s. droz , m. heusler and n. straumann , phys
* b268 * ( 1991 ) 371 ; + b. kleihaus , j. kunz and a. sood , phys . lett .
* b352 * ( 1995 ) 247 ; + t. ioannidou , b. kleihaus and w. zakrzewski , phys . lett . *
b600 * ( 2004 ) 116 . | we construct monopole - antimonopole chain and vortex solutions in yang - mills - higgs theory coupled to einstein gravity .
the solutions are static , axially symmetric and asymptotically flat .
they are characterized by two integers @xmath0 where @xmath1 is related to the polar angle and @xmath2 to the azimuthal angle .
solutions with @xmath3 and @xmath4 correspond to chains of @xmath1 monopoles and antimonopoles . here
the higgs field vanishes at @xmath1 isolated points along the symmetry axis .
larger values of @xmath2 give rise to vortex solutions , where the higgs field vanishes on one or more rings , centered around the symmetry axis . when gravity is coupled to the flat space solutions , a branch of gravitating monopole - antimonopole chain or vortex solutions arises , and merges at a maximal value of the coupling constant with a second branch of solutions .
this upper branch has no flat space limit .
instead in the limit of vanishing coupling constant it either connects to a bartnik - mckinnon or generalized bartnik - mckinnon solution , or , for @xmath5 , @xmath6 , it connects to a new einstein - yang - mills solution . in this latter case
further branches of solutions appear . for small values of the coupling constant on the upper branches ,
the solutions correspond to composite systems , consisting of a scaled inner einstein - yang - mills solution and an outer yang - mills - higgs solution . |
according to inflationary cosmology @xcite , the universe expanded exponentially quickly for a fraction of a second very early in its history growing from a patch as small as @xmath0 m , one hundred billion times smaller than a proton , to macroscopic scales on the order of a meter , all within about @xmath1 s before slowing down to the more stately rate of expansion that has characterized the universe s behavior ever since .
the driving force behind this dramatic growth , strangely enough , was gravity .
[ for technical introductions to inflationary cosmology , see @xcite ; a more popular description may be found in @xcite . ]
although this might sound like hopeless ( or , depending on one s inclinations , interesting ) speculation , in fact inflationary cosmology leads to several quantitative predictions about the present behavior of our universe predictions that are being tested to unprecedented accuracy by a new generation of observational techniques .
so far the agreement has been excellent .
how could gravity drive the universal repulsion during inflation ?
the key to this rapid expansion is that in einstein s general relativity ( physicists reigning description of gravity ) , the gravitational field couples both to mass - energy ( where mass and energy are interchangeable thanks to einstein s @xmath2 ) and to pressure , rather than to mass alone . in the simplest scenario , in which at least the observable portion of our universe can be approximated as being homogeneous and isotropic
that is , having no preferred locations or directions einstein s gravitational equations give a particularly simple result .
the expansion of the universe may be described by introducing a time - dependent `` scale factor , '' @xmath3 , with the separation between any two objects in the universe being proportional to @xmath3 .
einstein s equations prescribe how this scale factor will evolve over time , @xmath4 .
the rate of acceleration is proportional to the density of mass - energy in the universe , @xmath5 , plus three times its pressure , @xmath6 : @xmath7 , where @xmath8 is newton s gravitational constant ( and we use units for which the speed of light @xmath9 ) .
the minus sign is important : ordinary matter under ordinary circumstances has both positive mass - energy density and positive ( or zero ) pressure , so that @xmath10 . in this case
, gravity acts as we would expect it to : all of the matter in the universe tends to attract all of the other matter , causing the expansion of the universe as a whole to slow down .
the crucial idea behind inflation is that matter can behave rather differently from this familiar pattern .
ideas from particle physics suggest that the universe is permeated by scalar fields , such as the higgs field of the standard model of particle physics , or its more exotic generalizations .
( a scalar field takes exactly one value at every point in space and time .
for example , one could measure the temperature at every position in a room and repeat the measurements over time , and represent the measurements by a scalar field , @xmath11 , of temperature .
electric and magnetic fields are vector fields , which carry three distinct components at every point in space and time : the field in the @xmath12 direction , in the @xmath13 direction , and in the @xmath14 direction .
scalar fields are introduced in particle physics to describe certain kinds of particles , just as photons are described in quantum field theories in terms of electromagnetic fields . )
these scalar fields can exist in a special state , having a high energy density that can not be rapidly lowered , such as the arrangement labeled ( a ) in fig . 1 .
such a state is called a `` false vacuum . ''
particle physicists use the word `` vacuum '' to denote the state of lowest energy .
`` false vacua '' are only metastable , not the true states of lowest possible energy . in the early universe , a scalar field in such a false vacuum state
can dominate all the contributions to the total mass - energy density , @xmath5 . during this period , @xmath5 remains nearly constant , even as the volume of the universe expands rapidly : @xmath15 .
this is quite different from the density of ordinary matter , which decreases when the volume of its container increases . moreover
, the first law of thermodynamics , in the context of general relativity , implies that if @xmath16 while the universe expands , then the equation of state for this special state of matter must be @xmath17 , a negative pressure .
this yields @xmath18 : rather than slowing down , the cosmic expansion rate will grow rapidly , driven by the negative pressure created by this special state of matter . under these circumstances ,
the scale factor grows as @xmath19 , where the hubble parameter , @xmath20 , which measures the universe s rate of expansion , assumes the constant value , @xmath21 .
the universe expands exponentially until the scalar field rolls to near the bottom of the hill in the potential energy diagram .
what supplies the energy for this gigantic expansion ?
the answer , surprisingly , is that no energy is needed @xcite .
physicists have known since the 1930s @xcite that the gravitational field carries _ negative _ potential energy density . as vast quantities of matter
are produced during inflation , a vast amount of negative potential energy materializes in the gravitational field that fills the ever - enlarging region of space .
the total energy remains constant , and very small , and possibly exactly equal to zero .
there are now dozens of models that lead to this generic inflationary behavior , featuring an equation of state , @xmath22 , during the early universe @xcite .
this entire family of models , moreover , leads to several main predictions about today s universe .
first , our observable universe should be spatially flat .
einstein s general relativity allows for all kinds of curved ( or `` non - euclidean '' ) spacetimes .
homogeneous and isotropic spacetimes fall into three classes ( fig .
2 ) , depending on the value of the mass - energy density , @xmath5 . if @xmath23 , where @xmath24 , then einstein s equations imply that the spacetime will be positively curved , or closed ( akin to the two - dimensional surface of a sphere ) ; parallel lines will intersect , and the interior angles of a triangle will always add up to more than @xmath25 . if @xmath26 , the spacetime will be negatively curved , or open ( akin to the two - dimensional surface of a saddle ) ; parallel lines will diverge and triangles will sum to less than @xmath25 . only if @xmath27 will spacetime be spatially flat ( akin to an ordinary two - dimensional flat surface ) ; in this case , all of the usual rules of euclidean geometry apply .
cosmologists use the letter @xmath28 to designate the ratio of the actual mass - energy density in the universe to this critical value : @xmath29 .
although general relativity allows any value for this ratio , inflation predicts that @xmath30 within our observable universe to extremely high accuracy . until recently , uncertainties in the measurement of @xmath28 allowed any value in the wide range , @xmath31 , with many observations seeming to favor @xmath32 . a new generation of detectors ,
however , has dramatically changed the situation .
the latest observations , combining data from the wilkinson microwave anisotropy probe ( wmap ) , the sloan digital sky survey ( sdss ) , and observations of type ia supernovae , have measured @xmath33 @xcite an amazing match between prediction and observation .
in fact , inflation offers a simple explanation for why the universe should be so flat today . in the standard big bang cosmology ( without inflation ) , @xmath30 is an _ unstable _ solution : if @xmath28 were ever - so - slightly less than 1 at an early time , then it would rapidly slide toward 0 . for example , if @xmath28 were @xmath34 at 1 s after the big bang , it would be only @xmath35 today . if @xmath28 were 1.1 at @xmath36 s , then it would have grown so quickly that the universe would have recollapsed just 45 s later . in inflationary models ,
on the other hand , any original curvature of the early universe would have been stretched out to near - flatness as the universe underwent its rapid expansion ( fig .
quantitatively , @xmath37 , so that while @xmath38 and @xmath19 during the inflationary epoch , @xmath28 gets driven rapidly to 1 .
the second main prediction of inflation is that the presently observed universe should be remarkably smooth and homogeneous on the largest astronomical scales .
this , too , has been measured to extraordinary accuracy during the past decades .
starting in the 1960s , earth - bound , balloon - borne , and now satellite detectors have measured the cosmic microwave background ( cmb ) radiation , a thermal bath of photons that fills the sky .
the photons today have a frequency that corresponds to a temperature of 2.728 k @xcite .
these photons were released @xmath39400,000 years after the inflationary epoch , when the universe had cooled to a low enough temperature that would allow stable ( and electrically neutral ) atoms to form . before that time , the ambient temperature of matter in the universe was so high that would - be atoms were broken up by high - energy photons as soon as they formed , so that the photons were effectively trapped , constantly colliding into electrically charged matter . since stable atoms formed , however , the cmb photons have been streaming freely .
their temperature today is terrifically uniform : after adjusting for the earth s motion , cmb photons measured from any direction in the sky have the same temperature to one part in @xmath40 @xcite . without inflation , this large - scale smoothness appears quite puzzling . according to ordinary ( noninflationary ) big bang cosmology , these photons should never have had a chance to come to thermal equilibrium : the regions in the sky from which they were released would have been about 100 times farther apart than even light could have traveled between the time of the big bang and the time of the photons release @xcite .
much like the flatness problem , inflation provides a simple and generic reason for the observed homogeneity of the cmb : today s observable universe originated from a _ much _ smaller region than that in the noninflationary scenarios .
this much - smaller patch could easily have become smooth before inflation began .
inflation would then stretch this small homogeneous region to encompass the entire observable universe .
a third major prediction of inflationary cosmology is that there should be tiny departures from this strict large - scale smoothness , and that these ripples ( or `` perturbations '' ) should have a characteristic spectrum .
today these ripples can be seen directly as fluctuations in the cmb . although the ripples are believed to be responsible for the grandest structures of the universe
galaxies , superclusters , and giant voids in inflationary models they arise from quantum fluctuations , usually important only on atomic scales or smaller .
the field @xmath41 that drives inflation , like all quantum fields , undergoes quantum fluctuations in accord with the heisenberg uncertainty principle
. during inflation these quantum fluctuations are stretched proportionally to @xmath3 , rapidly growing to macroscopic scales .
the result : a set of nearly scale - invariant perturbations extending over a huge range of wavelengths @xcite .
cosmologists parameterize the spectrum of primordial perturbations by a spectral index , @xmath42 .
a scale - invariant spectrum would have @xmath43 ; inflationary models generically predict @xmath44 to within @xmath45 .
the latest measurements of these perturbations by wmap and sdss reveal @xmath46 @xcite . until recently ,
astronomers were aware of several cosmological models that were consistent with the known data : an open universe , with @xmath47 ; an inflationary universe with considerable dark energy ( @xmath48 ) ; an inflationary universe without @xmath48 ; and a universe in which the primordial perturbations arose from topological defects such as cosmic strings .
dark energy @xcite is a form of matter with negative pressure that is currently believed to contribute more than @xmath49 of the total energy of the observed universe .
cosmic strings are long , narrow filaments hypothesized to be scattered throughout space , remnants of a symmetry - breaking phase transition in the early universe @xcite .
[ cosmic strings are topologically nontrivial configurations of fields which should not be confused with the fundamental strings of superstring theory .
the latter are usually believed to have lengths on the order of @xmath1 m , although for some compactifications these strings might also have astronomical lengths @xcite . ]
each of these models leads to a distinctive pattern of resonant oscillations in the early universe , which can be probed today through its imprint on the cmb . as can be seen in fig .
4 @xcite , three of the models are now definitively ruled out .
the full class of inflationary models can make a variety of predictions , but the prediction of the simplest inflationary models with large @xmath48 , shown on the graph , fits the data beautifully .
research in recent years has included investigations of what might have preceded inflation , and how an inflationary epoch might have ended .
soon after the first inflationary models were introduced , several physicists @xcite realized that once inflation began , it would in all likelihood never stop .
regions of space would stop inflating , forming what can be called `` pocket universes , '' one of which would contain the observed universe . nonetheless , at any given moment some portion of the universe would still be undergoing exponential expansion , in a process called `` eternal inflation . '' in the model depicted in fig . 1
, for example , quantum - mechanical effects compete with the classical motion to produce eternal inflation .
consider a region of size @xmath50 , in which the average value of @xmath41 is near ( b ) or ( d ) in the diagram .
call the average energy density @xmath51 . whereas the classical tendency of @xmath41 is to roll slowly downward ( red arrow ) toward the minimum of its potential
, the field will also be subject to quantum fluctuations ( green arrows ) similar to those described above
. the quantum fluctuations will give the field a certain likelihood of hopping up the wall of potential energy rather than down it . over a time period @xmath50
, this region will grow @xmath52 times its original size .
if the probability that the field will roll up the potential hill during this period is greater than @xmath53 , then on average the volume of space in which @xmath54 increases with time @xcite . the probability of upward fluctuations tends to become large when the initial value of @xmath41 is near the peak at ( a ) or high on the hill near ( d ) , so for most potential energy functions the condition for eternal inflation is attainable . in that case
the volume of the inflating region grows exponentially , and forever : inflation would produce an infinity of pocket universes .
an interesting question is whether or not eternal inflation makes the big bang unnecessary : might eternal inflation have been truly `` eternal , '' existing more or less the same way for all time , or is it only `` eternal '' to the future once it gets started ?
borde and vilenkin have analyzed this question ( most recently , with guth ) , and have concluded that eternal inflation could not have been past - eternal : using kinematic arguments , they showed @xcite that the inflating region must have had a past boundary , before which some alternative description must have applied .
one possibility would be the creation of the universe by some kind of quantum process .
another major area of research centers on the mechanisms by which inflation might have ended within our observable universe .
the means by which inflation ends have major consequences for the subsequent history of our universe . for one thing ,
the colossal expansion during inflation causes the temperature of the universe to plummet nearly to zero , and dilutes the density of ordinary matter to negligible quantities . some mechanism must therefore convert the energy of the scalar field , @xmath41 , into a hot soup of garden - variety matter . in most models , inflation ends when @xmath41 oscillates around the minimum of its potential , as in region ( c ) of fig . 1 .
quantum - mechanically , these field oscillations correspond to a collection of @xmath41 particles approximately at rest . early studies of post - inflation `` reheating '' assumed that individual @xmath41 particles would decay during these oscillations like radioactive nuclei .
more recently , it has been discovered @xcite that these oscillations would drive resonances in @xmath41 s interactions with other quantum fields . instead of individual @xmath41 particles decaying independently
, these resonances would set up collective behavior @xmath41 would release its energy more like a laser than an ordinary light bulb , pouring it extremely rapidly into a sea of newly created particles .
large numbers of particles would be created very quickly within specific energy - bands , corresponding to the frequency of @xmath41 s oscillations and its higher harmonics .
this dramatic burst of particle creation would affect spacetime itself , as it responded to changes in the arrangement of matter and energy .
the rapid transfer of energy would excite gravitational perturbations , of which the most strongly amplified would be those with frequencies within the resonance bands of the decaying @xmath41 field . in some extreme cases ,
very long - wavelength perturbations can be amplified during reheating , which could in principle even leave an imprint on the cmb @xcite .
although superstring theory promises to synthesize general relativity with the other fundamental forces of nature , it introduces a number of surprising features such as the existence of microscopic strings , rather than particles , as the fundamental units of matter , along with the existence of several extra spatial dimensions in the universe .
could our observable universe really be built from such a bizarre collection of ingredients ?
navely , one might expect the extra dimensions to conflict with the observed behavior of gravity . to be successful , string theory , like general relativity , must reduce to newton s law of gravity in the appropriate limit . in newton s formulation , gravity can be described by force lines that always begin and end on masses . if the force lines could spread in @xmath55 spatial dimensions , then at a radius @xmath56 from the center , they would intersect a hypersphere with surface area proportional to @xmath57 . an equal number of force lines would cross the hypersphere at each radius , which means that the density of force lines would be proportional to 1/@xmath57 . for @xmath58 ,
this reproduces the familiar newtonian force law , @xmath59 , which has been tested ( along with its einsteinian generalization ) to remarkable accuracy over a huge range of distances , from astronomical scales down to less than a millimeter @xcite .
an early response to this difficulty was to assume that the extra spatial dimensions are curled up into tiny closed circles rather than extending to macroscopic distances .
because gravity has a natural scale , known as the planck length , @xmath60 m [ where @xmath61 is planck s constant divided by @xmath62 , physicists assumed that @xmath63 sets the scale for these extra dimensions . just as the surface of a soda straw would appear one - dimensional when viewed from a large distance
even though it is really two - dimensional our space would appear three - dimensional if the extra dimensions were `` compactified '' in this way . on scales
much larger than the radii of the extra dimensions , @xmath64 , we would fail to notice them : the strength of gravity would fall off in its usual @xmath65 manner for distances @xmath66 , but would fall off as @xmath67 for scales @xmath68 @xcite .
the question remained , however , what caused this compactification , and why this special behavior affected only some but not all dimensions .
recently arkani - hamed , dimopoulos , and dvali @xcite realized that there is no necessary relation between @xmath63 and @xmath64 , and that experiments only require @xmath69 mm . shortly afterward
, randall and sundrum @xcite discovered that the extra dimensions could even be infinite in extent ! in the randall - sundrum model , our observable universe lies on a membrane , or `` brane '' for short , of three spatial dimensions , embedded within some larger multidimensional space .
the key insight is that the energy carried by the brane will sharply affect the way the gravitational field behaves . for certain spacetime configurations ,
the behavior of gravity along the brane can appear four - dimensional ( three space and one time ) , even in the presence of extra dimensions .
gravitational force lines would tend to `` hug '' the brane , rather than spill out into the `` bulk '' the spatial volume in which our brane is embedded . along the brane , therefore , the dominant behavior of the gravitational force would still be @xmath65 . in simple models ,
in which the spacetime geometry along our brane is highly symmetric , such as the minkowski spacetime of special relativity , the effective gravitational field along our brane is found to mimic the usual einsteinian results to high accuracy @xcite . at very short distances there are calculable ( and testable ) deviations from standard gravity , and there may also be deviations for very strong gravitational fields , such as those near black holes .
there are also modifications to the cosmological predictions of gravity . in the usual case ,
when einstein s equations are applied to a homogeneous and isotropic spacetime , one finds @xmath70 , where @xmath71 is a constant connected to the curvature of the universe .
if instead we lived on a brane embedded within one large extra dimension , then @xmath72 , where @xmath73 is a constant @xcite . under ordinary conditions
, @xmath5 decreases as the universe expands , and so the new term in the effective einstein equations should have minimal effects at late times in our observable universe .
but we saw above that during an inflationary epoch , @xmath74 , and in these early moments the departures from the ordinary einsteinian case can be dramatic . in particular ,
the @xmath75 term would allow inflation to occur at lower energies than are usually assumed in ordinary ( nonembedded ) models , with potential energy functions that are less flat than are ordinarily needed to sustain inflation .
moreover , the spectrum of primordial perturbations would get driven even closer to the scale - invariant shape , with @xmath76 @xcite .
brane cosmology thus leads to some interesting effects during the early universe , making inflation even more robust than in ordinary scenarios .
although theories of extra dimensions establish a connection between string theory and cosmology , the developments of the past few years have pushed the connection much further . [ for reviews , see @xcite . ]
the union of string theory and cosmology is barely past its honeymoon , but so far the marriage appears to be a happy one .
inflation , from its inception , was a phenomenologically very successful idea that has been in need of a fundamental theory to constrain its many variations .
string theory , from its inception , has been a very well - constrained mathematical theory in need of a phenomenology to provide contact with observation .
the match seems perfect , but time will be needed before we know for sure whether either marriage partner can fulfill the needs of the other . in the meantime
, ideas are stirring that have the potential to radically alter our ideas about fundamental laws of physics .
for many years the possibility of describing inflation in terms of string theory seemed completely intractable , because the only string vacua that were understood were highly supersymmetric ones , with many massless scalar fields , called moduli , which have potential energy functions that vanish identically to all orders of perturbation theory .
when the effects of gravity are included , the energy density of such supersymmetric states is never positive .
inflation , on the other hand , requires a positive energy density , and it requires a hill in the potential energy function .
inflation , therefore , could only be contemplated in the context of nonperturbative supersymmetry - breaking effects , of which there was very little understanding .
the situation changed dramatically with the realization that string theory contains not only strings , but also branes , and fluxes , which can be thought of as higher - dimensional generalizations of magnetic fields .
the combination of these two ingredients makes it possible to construct string theory states that break supersymmetry , and that give nontrivial potential energy functions to all the scalar fields .
one very attractive idea for incorporating inflation into string theory is to use the positions of branes to play the role of the scalar field that drives inflation .
the earliest version of this theory was proposed in 1998 by dvali and tye @xcite , shortly after the possibility of large extra dimensions was proposed in @xcite . in the dvali - tye model ,
the observed universe is described not by a single three - dimensional brane , but instead by a number of three - dimensional branes which in the vacuum state would sit on top of each other .
if some of the branes were displaced , however , in a fourth spatial direction , then the energy would be increased .
the brane separation would be a function of time and the three spatial coordinates along the branes , and so from the point of view of an observer on the brane , it would act like a scalar field that could drive inflation . at this stage , however , the authors needed to invoke unknown mechanisms to break supersymmetry and to give the moduli fields nonzero potential energy functions . in 2003 ,
kachru , kallosh , linde , and trivedi @xcite showed how to construct complicated string theory states for which all the moduli have nontrivial potentials , for which the energy density is positive , and for which the approximations that were used in the calculations appeared justifiable .
these states are only metastable , but their lifetimes can be vastly longer than the 10 billion years that have elapsed since the big bang .
there was nothing elegant about this construction the six extra dimensions implied by string theory are curled not into circles , but into complicated manifolds with a number of internal loops that can be threaded by several different types of flux , and populated by a hodgepodge of branes .
joined by maldacena and mcallister , this group @xcite went on to construct states that can describe inflation , in which a parameter corresponding to a brane position can roll down a hill in its potential energy diagram . generically the potential energy function is not flat enough for successful inflation , but the authors argued that the number of possible constructions was so large that there may well be a large class of states for which sufficient inflation is achieved .
iizuka and trivedi @xcite showed that successful inflation can be attained by curling the extra dimensions into a manifold that has a special kind of symmetry .
a tantalizing feature of these models is that at the end of inflation , a network of strings would be produced @xcite
. these could be fundamental strings , or branes with one spatial dimension .
the cmb data of fig .
4 rule out the possibility that these strings are major sources of density fluctuations , but they are still allowed if they are light enough so that they do nt disturb the density fluctuations from inflation .
string theorists are hoping that such strings may be able to provide an observational window on string physics .
a key feature of the constructions of inflating states or vacuumlike states in string theory is that they are far from unique .
the number might be something like @xmath77 @xcite , forming what susskind has dubbed the `` landscape of string theory . ''
although the rules of string theory are unique , the low - energy laws that describe the physics that we can in practice observe would depend strongly on which vacuum state our universe was built upon .
other vacuum states could give rise to different values of `` fundamental '' constants , or even to altogether different types of `` elementary '' particles , and even different numbers of large spatial dimensions ! furthermore , because inflation is generically eternal , one would expect that the resulting eternally inflating spacetime would sample every one of these states , each an infinite number of times . because all of these states are possible , the important problem is to learn which states are probable .
this problem involves comparison of one infinity with another , which is in general not a well - defined question @xcite .
proposals have been made and arguments have been given to justify them @xcite , but no conclusive solution to this problem has been found .
what , then , determined the vacuum state for our observable universe ?
although many physicists ( including the authors ) hope that some principle can be found to understand how this choice was determined , there are no persuasive ideas about what form such a principle might take .
it is possible that inflation helps to control the choice of state , because perhaps one state or a subset of states expands faster than any others .
because inflation is generically eternal , the state that inflates the fastest , along with the states that it decays into , might dominate over any others by an infinite amount .
progress in implementing this idea , however , has so far been nil , in part because we can not identify the state that inflates the fastest , and in part because we can not calculate probabilities in any case .
if we could calculate the decay chain of the most rapidly inflating state , we would have no guarantee that the number of states with significant probability would be much smaller than the total number of possible states .
another possibility , now widely discussed , is that _ nothing _ determines the choice of vacuum for our universe ; instead , the observable universe is viewed as a tiny speck within a multiverse that contains _ every _ possible type of vacuum .
if this point of view is right , then a quantity such as the electron - to - proton mass ratio would be on the same footing as the distance between our planet and the sun .
neither is fixed by the fundamental laws , but instead both are determined by historical accidents , restricted only by the fact that if these quantities did not lie within a suitable range , we would not be here to make the observations .
this idea that the laws of physics that we observe are determined not by fundamental principles , but instead by the requirement that intelligent life can exist to observe them is often called the anthropic principle .
although in some contexts this principle might sound patently religious , the combination of inflationary cosmology and the landscape of string theory gives the anthropic principle a scientifically viable framework .
a key reason why the anthropic approach is gaining attention is the observed fact that the expansion of the universe today is accelerating , rather than slowing down under the influence of normal gravity . in the context of general relativity
, this requires that the energy of the observable universe is dominated by dark energy .
the identity of the dark energy is unknown , but the simplest possibility is that it is the energy density of the vacuum , which is equivalent to what einstein called the cosmological constant . to particle physicists it is not surprising that the vacuum has nonzero energy density , because the vacuum is known to be a very complicated state , in which particle - antiparticle pairs are constantly materializing and disappearing , and fields such as the electromagnetic field are constantly undergoing wild fluctuations . from the standpoint of the particle physicist , the shocking thing is that the energy density of the vacuum is so low .
no one really knows how to calculate the energy density of the vacuum , but nave estimates lead to numbers that are about @xmath78 times larger than the observational upper limit .
there are both positive and negative contributions , but physicists have been trying for decades to find some reason why the positive and negative contributions should cancel , but so far to no avail .
it seems even more hopeless to find a reason why the net energy density should be nonzero , but 120 orders of magnitude smaller than its expected value .
however , if one adopts the anthropic point of view , it was argued as early as 1987 by weinberg @xcite that an explanation is at hand : if the multiverse contained regions with all conceivable values of the cosmological constant , galaxies and hence life could appear only in those very rare regions where the value is small , because otherwise the huge gravitational repulsion would blow matter apart without allowing it to collect into galaxies .
the landscape of string theory and the evolution of the universe through the landscape are of course still not well understood , and some have argued @xcite that the landscape might not even exist .
it seems too early to draw any firm conclusions , but clearly the question of whether the laws of physics are uniquely determined , or whether they are environmental accidents , is an issue too fundamental to ignore .
during the past decade , cosmology has unquestionably entered the domain of high - precision science .
just a few years ago several basic cosmological quantities , such as the expansion parameter , @xmath79 , and the flatness parameter , @xmath28 , were known only to within a factor of 2 .
now new observations using wmap , sdss , and the high - redshift type ia supernovae measure these and other crucial quantities with percent - level accuracy .
several of inflation s most basic quantitative predictions , including @xmath30 and @xmath80 , may now be compared with data that are discriminating enough to distinguish inflation from many of its theoretical rivals .
so far , every measure has been favorable to inflation . even with the evidence in favor of inflation now stronger than ever , much work remains .
inflationary cosmology has always been a framework for studying the interconnections between particle physics and gravitation a collection of models rather than a unique theory .
the next generation of astronomical detectors should be able to distinguish between competing inflationary models , whittling down the large number of options to a preferred few .
one important goal is the high - precision measurement of polarization effects in the cmb , which allows the possibility of uncovering the traces of gravity waves originating from inflation .
gravity waves of the right pattern would be a striking test of inflation , and would allow us to determine the energy density of the `` false vacuum '' state that drove inflation .
the new cosmological observations also offer physicists one of the best resources for evaluating the latest developments in idea - rich ( but data - poor ) particle theory , where much of the current research has been aimed at the high - energy frontier , well beyond the range of existing accelerators .
perhaps the interface between string theory and cosmology will lead to new predictions for the astronomers to test .
whether such tests are successful or not , physicists are certain to learn important lessons about the nature of space , time , and matter .
meanwhile , several major puzzles persist . now that physicists and astronomers are confident that @xmath30 to high accuracy
, the question remains of just what type of matter and energy is filling the universe .
ordinary matter , such as the protons , neutrons , and electrons that make up atoms , contributes just @xmath81 to this cosmic balance .
nearly one - quarter of the universe s mass - energy is some form of `` dark matter '' different in kind from the garden - variety matter we see around us , and yet exerting a measurable gravitational tug that shapes the way galaxies behave .
particle physicists have offered many candidates for this exotic dark matter , but to date no single contender has proved fully convincing @xcite . even more bizarre is the dark energy now known to contribute about @xmath49 to @xmath28 .
this dark energy is driving a mini - inflationary epoch today , billions of years after the initial round of inflation .
today s accelerated expansion is far less fast than the earlier inflationary rate had been , but the question remains why it is happening at all .
could the dark energy be an example of einstein s cosmological constant ?
or maybe it is a variation on an inflationary theme : perhaps some scalar field has been sliding down its potential energy hill on a time scale of billions of years rather than fractions of a second @xcite .
whatever its origin , dark energy , much like dark matter , presents a fascinating puzzle that will keep cosmologists busy for years to come .
the authors thank gia dvali , hong liu , max tegmark , sandip trivedi , and henry tye for very helpful comments on the manuscript .
this work was supported in part by funds provided by the u.s .
department of energy ( d.o.e . ) under cooperative research agreement # df - fc02 - 94er40818 .
d. n. spergel _
et al . _ , `` first year wilkinson microwave anisotropy probe ( wmap ) observations : determination of cosmological parameters , '' _ astrophys .
j. suppl . _
* 148 * , 175 ( 2003 ) [ arxiv : astro - ph/0302209 ] .
we thank max tegmark for providing this graph , which shows the most precise data points for each range of @xmath82 from recent observations , as summarized in @xcite .
the cosmic string prediction is taken from @xcite .
the other curves were all calculated for @xmath83 , @xmath84 , and @xmath79 = 70 km s@xmath85 mpc@xmath85 , with the remaining parameters fixed as follows . ``
inflation with @xmath48 '' : @xmath86 , @xmath87 , and optical depth parameter @xmath88 ; `` inflation without @xmath48 '' : @xmath89 , @xmath90 , @xmath91 ; `` open universe '' : @xmath92 , @xmath90 , @xmath91 .
the 1-sd error bars include both observational uncertainty and `` cosmic variance , '' the intrinsic quantum uncertainty in the predictions , as calculated from the `` inflation with @xmath48 '' model . with our current ignorance of the underlying physics
, none of these theories predicts the overall amplitude of the fluctuations ; the `` inflation with @xmath48 '' curve was normalized for a best fit , and the others were normalized arbitrarily . | understanding the behavior of the universe at large depends critically on insights about the smallest units of matter and their fundamental interactions .
inflationary cosmology is a highly successful framework for exploring these interconnections between particle physics and gravitation .
inflation makes several predictions about the present state of the universe such as its overall shape , large - scale smoothness , and smaller - scale structure which are being tested to unprecedented accuracy by a new generation of astronomical measurements .
the agreement between these predictions and the latest observations is extremely promising .
meanwhile , physicists are busy trying to understand inflation s ultimate implications for the nature of matter , energy , and spacetime .
the scientific community is celebrating the international year of physics in 2005 , honoring the centennial of albert einstein s most important year of scientific innovation . in the span of just a few months during 1905
einstein introduced key notions that would dramatically change our understanding of matter and energy as well as the nature of space and time .
the centennial of these seminal developments offers an enticing opportunity to take stock of how scientists think about these issues today .
we focus in particular on recent developments in the field of inflationary cosmology , which draws on a blend of concepts from particle physics and gravitation .
the last few years have been a remarkably exciting time for cosmology , with new observations of unprecedented accuracy yielding many surprises .
einstein s legacy is flourishing in the early 21st century .
inflation was invented a quarter of a century ago , and has become a central ingredient of current cosmological research .
describing dramatic events in the earliest history of our universe , inflationary models generically predict that our universe today should have several distinct features features that are currently being tested by the new generation of high - precision astronomical measurements .
even as inflation passes more and more stringent empirical tests , theorists continue to explore broader features and implications , such as what might have come before an inflationary epoch , how inflation might have ended within our observable universe , and how inflation might arise in the context of our latest understanding of the structure of space , time , and matter .
particle theory has been changing rapidly , and these theoretical developments have provided just as important a spur to inflationary cosmology as have the new observations . during the 1960s and 70s , particle physicists discovered that if they neglected gravity , they could construct highly successful descriptions of three out of the four basic forces in the universe : electromagnetism and the strong and weak nuclear forces .
the `` standard model of particle physics , '' describing these three forces , was formulated within the framework of quantum field theory , the physicist s quantum - mechanical description of subatomic matter .
inflationary cosmology was likewise first formulated in terms of quantum field theory .
now , however , despite ( or perhaps because of ) the spectacular experimental success of the standard model , the major thrust of particle physics research is aimed at moving beyond it . for all its successes , the standard model says nothing at all about the fourth force : gravity . for more than 50 years
physicists have sought ways to incorporate gravity within a quantum - mechanical framework , initially with no success .
but for the past 25 or more years , an ever - growing group of theoretical physicists has been pursuing superstring theory as the bright hope for solving this problem . to accomplish this task ,
however , string theorists have been forced to introduce many novel departures from conventional ideas about fundamental forces and the nature of the universe .
for one thing , string theory stipulates that the basic units of matter are not pointlike particles ( as treated by quantum field theory ) , but rather one - dimensional extended objects , or strings .
moreover , in order to be mathematically self - consistent , string theories require the existence of several additional spatial dimensions . whereas our observable universe seems to contain one timelike dimension and three spatial dimensions
height , width , and depth string theory postulates that our universe actually contains at least six additional spatial dimensions , each at right angles to the others and yet somehow hidden from view . for measurements at low energies , string theory should behave effectively like a quantum field theory , reproducing the successes of the standard model of particle physics . yet
the interface between cosmology and string theory has been a lively frontier . for example
, some theorists have been constructing inflationary models for our universe that make use of the extra dimensions that string theory introduces .
others have been studying the string theory underpinnings for inflationary models , exploring such topics as the nature of vacuum states and the question of their uniqueness .
as we will see , inflation continues to occupy a central place in cosmological research , even as its relation to fundamental particle physics continues to evolve . |
computers have become an integral part of modern life , and are essential for most academic research @xcite . since the middle of last century , researchers have invented new techniques to boost their calculation rate , e.g. by engineering superior hardware , designing more effective algorithms , and introducing increased parallelism .
due to a range of physical limitations which constrain the performance of single processing units @xcite , recent computer science research is frequently geared towards enabling increased parallelism for existing applications . by definition
, parallelism is obtained by concurrently using the calculation power of multiple processing units . from small to large spatial scales , this is respectively done by : facilitating concurrent operation of instruction threads within a single core , of cores within a single processor , of processors within a node , of nodes within a cluster or supercomputer , and of supercomputers within a distributed supercomputing environment .
the vision of aggregating existing computers to form a global unified computing platform , and to focus that power for a single purpose , has been very popular both in popular fiction ( e.g. , the borg collective mind in star trek or big brother in orwell s 1984 ) and in scientific research ( e.g. , amazon ec2 , projects such as teragrid / xsede and egi , and numerous distributed computing projects @xcite ) .
although many have tried , none have yet succeeded to link up more than a handful of major computers in the world to solve a major high - performance computing problem .
very few research endeavors aim to do distributed computing at such a scale to obtain more performance .
although it requires a rather large labour investment across several time zones , accompanied with political complexities , it is technically possible to combine supercomputers to form an intercontinental grid .
we consider that combining supercomputers in such a way is probably worth the effort if many machines are involved , rather than a few .
combining a small number of machines is hardly worth the effort of doubling the performance of a single machine , but combining hundreds or maybe even thousands of computers together could increase performance by orders of magnitude @xcite . here
we share our experiences , and lessons learned , in performing a large cosmological simulation using an intercontinental infrastructure of multiple supercomputers .
our work was part of the cosmogrid project @xcite , an effort that was eventually successful but which suffered from a range of difficulties and set - backs .
the issues we faced have impacted on our personal research ambitions , and have led to insights which could benefit researchers in any large - scale computing community .
we provide a short overview of the cosmogrid project , and describe our initial assumptions in section [ sec : vision ] .
we summarize the challenges we faced , ascending the hierarchy from thread to transcontinental computer , in section [ sec : parallel ] and we summarize how our insights affected our ensuing research agenda in section [ sec : after ] .
we discuss the long - term implications of cosmogrid in section [ sec : future ] and conclude the paper with some reflections in section [ sec : discuss ] .
the aim of cosmogrid was to interconnect four supercomputers ( one in japan , and three across europe ) using light paths and 10 gigabit wide area networks , and to use them concurrently to run a very large cosmological simulation .
we performed the project in two stages : first by running simulations across two supercomputers , and then by extending our implementation to use four supercomputers concurrently .
the project started as a collaboration between researchers in the netherlands , japan and the united states in october 2007 , and received support from several major supercomputing centres ( sara in amsterdam , epcc in edinburgh , csc in espoo and naoj in tokyo ) .
cosmogrid mainly served a two - fold purpose : to predict the statistical properties of small dark matter halos from an astrophysics perspective , and to enable production simulations using an intercontinental network of supercomputers from a computer science perspective . for cosmogrid , we required a code to model the formation of dark matter structures ( using @xmath1 particles in total ) over a period of over 13 billion years .
we adopted a hybrid tree / particle - mesh ( treepm ) n - body code named _ greem _
@xcite , which is highly scalable and straightforward to install on supercomputers .
greem uses a barnes - hut tree algorithm @xcite to calculate force interactions between dark matter particles over short distances , and a particle - mesh algorithm to calculate force interactions over long distances @xcite . later in the project
, we realized that further code changes were required to enable execution across supercomputers .
as a result , we created a separate version of greem solely for this purpose .
this modified code is named _ sushi _ , which stands for simulating universe structure formation on heterogeneous infrastructures @xcite .
our case for a distributed computing approach was focused on a classic argument used to justify parallelism : multiple resources can do more work than a single one .
even the world s largest supercomputer is about an order of magnitude less powerful than the top 500 supercomputers in the world combined @xcite . in terms of interconnectivity
the case was also clear .
our performance models predicted that a 1 gbps wide area network would already result in good simulation performance ( we had 10 gbps links at our disposal ) , and that the round - trip time of about 0.27 s between the netherlands and japan would only impose a limited overhead on a simulation that would require approximately 100 s per integration time step .
world - leading performance of our cosmological n - body integrator was essential to make our simulations possible , and our japanese colleagues optimized the code for single - machine performance as part of the project .
snapshots / checkpoints would then be written distributed across sites , and gathered at run - time . at the start of cosmogrid , we anticipated to run across two supercomputers by the summer of 2008 , and across four supercomputers by the summer of 2009 .
we assumed a number of political benefits : the simulation we proposed required a large number of core hours and produce an exceptionally large amount of data .
these requirements would have been a very heavy burden for a single machine , and by executing a distributed setup we could mitigate the computational , storage and data i / o load imposed on individual machines .
we also were aware of the varying loads of machines at different times , and could accommodate for that by rebalancing the core distribution whenever we would restart the distributed simulation from a checkpoint .
overall , we mainly expected technical problems , particularly in establishing a parallelization platform which works across supercomputers . installing
homogeneous software across heterogeneous ( and frequently evolving ) supercomputer platforms appeared difficult to accomplish , particularly since we did not possess administrative rights on any of the machines .
in addition , the greem code had not been tested in a distributed environment prior to the project .
although we finalized the production simulations about a year later than anticipated , cosmogrid was successful in a number of fundamental areas .
we managed to successfully execute cosmological test simulations across up to four supercomputers , and full - size production simulations across up to three supercomputers @xcite .
in addition , our astrophysical results have led to new insights on the mass distribution of satellite halos around milky - way sized galaxies @xcite , on the existence of small groups of galaxies in dark - matter deprived voids @xcite , the structure of voids @xcite and the evolution of barionic - dominated star clusters in a dark matter halo @xcite . however , these results , though valuable in their own right , do not capture some of the most important and disturbing lessons we have learned from cosmogrid about distributed supercomputing . here
we summarize our experiences on engineering a code from the level of threads to that of a transcontinental machine , establishing a linked infrastructure to deploy the code , reserving the required resources to execute the code , and the software engineering and sustainability aspects surrounding distributed supercomputing codes .
we are not aware of previous publications of practical experiences on the subject , and for that reason this paper may help achieve a more successful outcome for existing research efforts in distributed hpc .
the greem code was greatly re - engineered during cosmogrid .
this was necessary to achieve a complete production simulation within the core hour allocations that we obtained from ncf and deisa .
our infrastructure consisted of three cray xt4 machines with little endian intel chips and one ibm machine with big endian power7 chips . at the start of cosmogrid ,
the power7 architecture was not yet in place .
greem had been optimized for the use of sse and avx instruction sets , executing 10 times faster when these instructions are supported .
sse was available in intel chips and avx was expected to be available in power7 chips .
however , the support for avx in power7 never materialized , forcing us to find alternative optimization approaches .
an initial 2-month effort by ibm engineers , leading to a 10% performance improvement , did not speed up the code sufficiently .
we then resorted to manual optimization without the use specialized instruction sets , e.g. , by reordering data structures and unrolling loops .
this effort resulted in a @xmath2 performance increase , which was within a factor of 3 of our original target .
much of the parellelization work on greem was highly successful , as evidenced by the gordon bell prize awarded in 2012 to ishiyama et al .
however , one unanticipated problem arose while scaling up the simulation to larger problem sizes .
greem applies a particle - mesh ( pm ) algorithm to resolve the interactions beyond a preconfigured cutoff limit .
the implementation of is algorithm was initially serial , as the overhead was a negligible component ( @xmath3% ) of the total execution time for smaller problems .
however , the overhead became much larger we scaled up to mesh sizes beyond @xmath4 mesh cells , forcing us to move from a serial implementation to a parallel implementation .
initially we considered executing greem as - is , and using a middleware solution to enable execution across supercomputers . in the years leading up to cosmogrid , a large number of libraries emerged for distributed hpc ( e.g. , mpich - g2 @xcite , openmpi @xcite , mpig @xcite and pacx - mpi @xcite ) .
many of these were strongly recommended by colleagues in the field , and provided mpi layers that allowed applications to pass messages across different computational resources .
although these were well - suited for distributed hpc across self - administered clusters , we quickly found that a distributed supercomputing environment was substantially different .
first , supercomputers are both more expensive and less common than clusters , and the centres managing them are reluctant to install libraries that require administrative privileges , due to risks of security and ease of maintenance ( mpi distributions tend to require such privileges ) .
second , the networks interconnecting the supercomputers are managed by separate organizational entities , and the default configurations at each network endpoint are almost always different and frequently conflicting .
this is not the case in more traditional ( national ) grid infrastructures , where uniform configurations can be imposed by the overarching project team .
the heterogeneity in network configurations resulted in severe performance and reliability penalties when using standard tcp - based middleware ( such as mpi and scp , unless we could find some way to either ( a ) customize the network configuration for individual paths or ( b ) adopt a different protocol ( e.g. , udp ) which ignores these preset configurations . in either case , we realized that using standard tcp - based mpi libraries for the communication between supercomputers was no longer a viable option .
using any other library had the inevitable consequence of modifying the main code , and eventually we chose to customize greem ( the customized version is named sushi @xcite ) and establish a seperate communication library ( mpwide @xcite ) for distributed supercomputing . having a code to run , and computer time to run it on is insufficient to do distributed concurrent supercomputing .
the amount of data to be transported is @xmath5 per integration time step and should not become the limiting factor in measuring performance .
each integration step would take about 100s . when allowing a 10% overhead we would have to require a network speed of @xmath6gb / s .
our collaboration with cees de laat ( university of amsterdam ) and kei hiraki ( tokyo university ) enables us to have two network and data transport specialists at each side of the light path . at the time
, russia had planned to make their military 10gbps dark fiber available for scientific experiments , but due to their enhanced military use we were unable to secure access to this cable .
the eventual route of the optical cable is presented in fig.[fig : cgnetworktopology ] .
we had a backup network between the ntt - com router at jgn2plus and the starlight interconnect to guarantee that our data stream remained stable throughout our calculations .
one of the interesting final quirks in our network topology was the absence of an optical network interface in the edinburgh machine ( which was installed later ) , and the fact that the optical cable at the japanese side reached the computer science building on the mitaka campus in tokyo next to where the supercomputer at naoj was located .
a person had to go physically to dig a hole and install a connecting cable between the two buildings . from a software perspective , we present our design considerations on mpwide fully in groen et al .
@xcite . here
we will summarize the main experiences and lessons that we learned from cosmogrid , as well as our experiences in readying the network in terms of software configuration .
we initially attempted to homogenize the network configuration settings between the different supercomputers .
this effort failed , as it was complicated by the presence of over a dozen stakeholder organizations , and further undermined by the lack of diagnostic information available to us .
for example , it was not always possible for us to pinpoint specific configuration errors to individual routers , and to their respective owners .
we also assessed the performance of udp - based solutions such as quanta @xcite and udt @xcite , which operate outside of the tcp - specific configuration settings of network devices .
however , we were not able to universally adopt such a solution , as some types of routers filter or restrict udp traffic .
we eventually converged on basing mpwide on multi - stream tcp @xcite , and combine this with mechanisms to customize tcp settings on each of the communication nodes @xcite .
our initial tests were marred with network glitches , particularly on the path between amsterdam and tokyo ( see fig .
[ fig : cgglitch ] for an example where packets were periodically stalled ) .
however , later runs resulted in more stable performance once we used a different path and adjusted the mpwide configuration to use very small tcp buffer sizes per stream @xcite . particles , run in parallel across the supercomputers in amsterdam and tokyo ( 32 cores per site ) .
we present the total time spent ( blue dotted line ) , time spent on calculation ( green dashed line ) , and the time spent on communication with mpwide ( red solid line ) .
stalls in the communication due to dropped packets resulted in visible peaks in the communication performance measurements .
reproduced from groen et al .
note : the communication time was relatively high ( @xmath7 of total runtime ) ) in this test run due to the small problem size , in @xcite we also present results from a test run with @xmath1 particles , which had a communication overhead of @xmath8 of total runtime.,width=288 ] the ability to have concurrent access to multiple supercomputers is an essential requirement for distributed supercomputing . within cosmogrid
, we initially agreed that the four institutions involved would provide so - called `` phone - based '' advance reservation for the purpose of this project .
however , due to delays in the commisioning of the light path , and due to political resistance regarding advance reservation within some of the supercomputing centres ( in part caused due to increasing demand of the machines ) , it was no longer possible to use this means of advance reservation on all sites .
eventually , we ended up with a different `` reservation '' mechanism for each of the supercomputers .
the original approach of calling up was still supported for the huygens machine in the netherlands . for the louhi machine in finland ,
calling up was also possible , but the reservation was established through an exclusively dedicated queue , as no direct reservation system was in place .
this approach had the side effect of locking other users out , and therefore we only opted to use it as a last resort .
for the hector machine in the uk , it was possible to request higher priority , but not to actually reserve resources in advance .
support for advance reservation was provided there shortly after cosmogrid concluded , but at the time the best method to `` reserve '' the machine was to submit a very large high priority job right before the machine maintenance time slot .
we then need to align all other reservations to the end of that maintenance time slot , presumed usually to be 6 hours after the start of the maintenance .
for the cray machine in japan , reservation was no longer possible due to the high work load .
however , some mechanisms of augmented priority could be established indirectly , e.g. by chaining jobs , which allowed for a job to be kept running at all times .
the combination of these strategies made it impossible to perform a large production run using all four sites .
we did perform smaller tests using the full infrastructure ( using regular scheduling queues and hoping for the best ) , but we were only able to do the largest runs using either the three european machines , or huygens combined with the cray in tokyo .
the task of engineering a code for distributed supercomputing is accompanied with a number of unusual challenges , particularly in the areas of software development and testing @xcite . at the time
, we had to ensure that greem and sushi remained fully compatible with all four supercomputer platforms . with no infrastructure in place to do continuous integration testing on supercomputer nodes ( even today this is still a rarity ) , we performed this testing periodically by hand . in general , testing on a single site is straightforward , but testing across sites less so .
we were able to arrange proof - of - concept tests across 4 sites using very small jobs ( e.g. , 16 cores per site ) by putting these jobs with long runtimes in the queue on each machine and waiting with starting the run until the jobs are running simultaneously . for slightly larger jobs ( e.g. , 64 cores per site )
this became difficult during peak usage hours as the queuing times of each job became longer and less predictable .
however , we have been able to perform a number of tests during more quiet periods of the week ( e.g. , at 2 am ) , without using advance reservation . for yet larger test runs
we were required to use advance reservation , reduce the number of supercomputers involved , or both .
software testing is instrumental to overall development , particularly when developing software in a challenging environment like an intercontental network of supercomputers . here
, the lack of facilities for advance reservation and continuous integration made testing large systems prohibitively difficult , and had an adverse effect on the development progress of sushi .
we eventually managed to get an efficient production calculation running for 12 hours across three sites and with the full system size of @xmath1 particles @xcite , but with better facilities for testing across sites we could well have tackled larger problems using higher core counts .
more emphasis and investment in testing facilities at the supercomputer centres would have boosted the cosmogrid project , and such support would arguably be of great benefit to increase the user uptake of supercomputers in general .
our experience with cosmogrid changed how we approached our computational research in two important ways .
first , due to the political hardships and the lack of facilities for advance reservation and testing , we changed our emphasis from distributed concurrent supercomputing towards improving code performance on single sites @xcite .
second , our expertise enabled us the enter the relatively young field of distributed multiscale computing with a head start .
multiscale computing involves the combination of multiple solvers to solve a problem that encompasses several different length and time scales .
each of these solvers may have different resource requirements , and as a result such simulations are frequently well - suited to be run across multiple resources .
drost et al .
applied some of our expertise , combining it with their experience using ibis , to enable the amuse environment to run different coupled solvers on different resources concurrently @xcite .
our experiences with cosmogrid were an important argument towards redeveloping the muscle coupling environment for the mapper project . in this eu - funded consortium ,
borgdorff et al . developed a successor ( muscle 2 ) , which is optimized for easier installation on supercomputers , and automates the startup of solvers that run concurrently on different sites ( a task that was done manually in cosmogrid ) .
in addition , we integrated mpwide in muscle2 @xcite and used mpwide directly to enable concurrently running coupled simulations across sites @xcite .
distributed high - performance supercomputing is not for the faint at heart .
it requires excellent programming and pluralization skills , stamina , determination , politics and hard labour .
one can wonder if it is worth the effort , but the answer depends on the available resources and the success of proposal writing . about 2030% of the proposals submitted to incite ( http://www.doeleadershipcomputing.org/faqs/ ) or prace ( http://www.prace-ri.eu/prace-kpi/ ) are successful , but
these success rates are generally lower for the very largest infrastructures ( e.g. , ornl titan ) . in addition , some of the largest supercomputers ( such as tianhe-2 and the k computer ) provide access only to closely connected research groups or to projects that have a native principal investigator .
acquiring compute time on these largest architectures can be sufficiently challenging that running your calculations on a number of less powerful but earlier accessible machines may be easier to accomplish . accessing several of such machines through one project is even harder , and probably not very realistic .
similarly , for the different architectures , it would be very curious to develop a code that works optimal on k computer and ornl titan concurrently .
achieving 25 pflops on titan alone is already a major undertaking @xcite , and combining such an optimized code with a tofu - type network architecture ( which is present on the k computer ) would make optimization a challenge of a different order .
we therefore do not think that distributed architectures will be used to beef - up the world s fastest computers , nor to connect a number of top-10 to top-100 supercomputers to out - compute the number 1 .
the type of distributed hpc as discussed in this article is probably best applied to large industrial or small academic computer clusters .
these @xmath9petaflop architectures are found in many academic settings or small countries , and are relatively easily accessible , by peer review proposals or via academic license agreements . in this context
, we think it is more feasible to connect 10 to 100 of such machines to outperform a top 1 to 10 computer .
we have presented our experiences from the cosmogrid project in high - performance distributed supercomputing .
plainly put , distributed high - performance supercomputing is a tough undertaking , and most of our initial assumptions were proven wrong .
much of the hardware and software infrastructure was constructed with very specific use - cases in mind , and was simply not fit for purpose to do distributed high - performance supercomputing .
a major reason why we have been able to establish distributed simulations at all is due to the tremendous effort of all the people involved , from research groups , networking departments and supercomputer centres .
it was due to their efforts to navigate the project around the numerous technical and political obstacles that distributed supercomputing became even possible .
cosmogrid was unsuccessful in establishing high - performance distributed supercomputing as the future paradigm for using very large supercomputers .
however , the project did provide a substantial knowledge boost to our subsequent research efforts , which is reflected by the success of projects such as mapper @xcite and amuse @xcite . the somewhat different approach taken in these projects ( aiming for more local resource infrastructures , and with a focus on coupling different solvers instead of parallelizing a single one ) resulted in tens of publications which relied on distributed ( super-)computing .
the hpc community has recently received criticism for its conservative approaches and resistance to change ( e.g. , @xcite ) . through cosmogrid
, it became obvious to us that resource providers can be subject to tricky dilemmas , where the benefits of supporting one research project need to be weighed against the possible reduced service ( or support ) incurred by other users . in light of that
, we do understand the conservative approaches followed in hpc to some extent . in cosmogrid
, we tried to work around that by ensuring that our software was installable without any administrative privileges , and we recommend that new researchers who wish to do distributed supercomputing do so as well ( or , perhaps , adopt very robust , flexible and well - performing tools for virtualization ) .
in addition , we believe that cosmogrid would have been greatly helped if innovations such as automated advance reservation systems for resources and network links , facilities for systematic software testing and continuous integration , and streamlined procedures for obtaining access to multiple sites had been in place . even today , such facilities make hpc infrastructures more convenient for new types of users and applications , and strengthen the position of the hpc community in an increasingly cloud - dominated computing landscape .
both spz and dg contributed equally to this work .
we are grateful to tomoaki ishiyama , keigo nitadori , jun makino , steven rieder , stefan harfst , cees de laat , paola grosso , steve macmillan , mary inaba , hans blom , jeroen bdorf , juha fagerholm , tomoaki ishiyama , esko kernen , walter lioen , jun makino , petri nikunen , gavin pringle and joni virtanen for their contributions to this work .
this research is supported by the netherlands organization for scientific research ( nwo ) grant # 639.073.803 , # 643.200.503 and # 643.000.803 and the stichting nationale computerfaciliteiten ( project # sh-095 - 08 ) .
we thank the deisa consortium ( eu fp6 project ri-031513 and fp7 project ri-222919 ) for support within the deisa extreme computing initiative ( gbbp project ) .
this paper has been made possible with funding from the uk engineering and physical sciences research council under grant number ep / i017909/1 ( http://www.science.net ) .
s. hettrick , m. antonioletti , l. carr , n. chue hong , s. crouch , d. de roure , i. emsley , c. goble , a. hay , d. inupakutika , m. jackson , a. nenadic , t. parkinson , m. i. parsons , a. pawlik , g. peru , a. proeme , j. robinson , and s. sufi , `` uk research software survey 2014 , '' dec . 2014 .
[ online ] .
available : http://dx.doi.org/10.5281/zenodo.14809 a. gualandris , s. portegies zwart , and a. tirado - ramos , `` performance analysis of direct n - body algorithms for astrophysical simulations on distributed systems . ''
_ parallel computing _ , vol .
33 , no . 3 , pp .
159173 , 2007 .
d. groen , s. portegies zwart , t. ishiyama , and j. makino , `` high performance gravitational n - body simulations on a planet - wide distributed supercomputer , '' _ computational science and discovery _
, vol . 4 , no .
015001 , jan .
e. agullo , c. coti , t. herault , j. langou , s. peyronnet , a. rezmerita , f. cappello , and j. dongarra , `` qcg - ompi : \{mpi } applications on grids , '' _ future generation computer systems _ , vol .
27 , no . 4 , pp . 357
369 , 2011 .
[ online ] .
available : http://www.sciencedirect.com/science/article/pii/s0167739x10002359 f. j. seinstra , j. maassen , r. v. van nieuwpoort , n. drost , t. van kessel , b. van werkhoven , j. urbani , c. jacobs , t. kielmann , and h. e. bal , `` , '' in _ _ , ser .
computer communications and networks , m. cafaro and g. aloisio , eds .
1em plus 0.5em minus 0.4emspringer london , 2011 , pp .
167197 .
m. ben belgacem , b. chopard , j. borgdorff , m. mamonski , k. rycerz , and d. harezlak , `` distributed multiscale computations using the mapper framework , '' _ procedia computer science _ ,
18 , no . 0 , pp . 1106 1115 , 2013 , 2013 international conference on computational science .
[ online ] .
available : http://www.sciencedirect.com/science/article/pii/s1877050913004195 j. borgdorff , m. ben belgacem , c. bona - casas , l. fazendeiro , d. groen , o. hoenen , a. mizeranschi , j. l. suter , d. coster , p. v. coveney , w. dubitzky , a. g. hoekstra , p. strand , and b. chopard , `` performance of distributed multiscale simulations , '' _ philosophical transactions of the royal society a : mathematical , physical and engineering sciences _ , vol .
372 , no . 2021 , 2014 .
g. hoekstra , a. , s. portegies zwart , m. bubak , and p. sloot , _ towards distributed petascale computing_.1em plus 0.5em minus 0.4empetascale computing : algorithms and applications , by david a. bader ( ed . ) .
chapman & hall / crc computational science series 565pp .
( isbn : 9781584889090 , isbn 10 : 1584889098 ) , 2008 .
s. portegies zwart , t. ishiyama , d. groen , k. nitadori , j. makino , c. de laat , s. mcmillan , k. hiraki , s. harfst , and p. grosso , `` simulating the universe on an intercontinental grid , '' _ computer _ , vol .
43 , pp . 6370 , 2010 .
t. ishiyama , t. fukushige , and j. makino , `` greem : massively parallel treepm code for large cosmological n - body simulations , '' _ publications of the astronomical society of japan _ ,
61 , no . 6 , pp .
13191330 , 2009 .
d. groen , s. rieder , and s. portegies zwart , `` high performance cosmological simulations on a grid of supercomputers , '' in _ proceedings of infocomp 2011_.1em plus 0.5em minus 0.4emthinkmind.org , sep . 2011 .
t. ishiyama , s. rieder , j. makino , s. portegies zwart , d. groen , k. nitadori , c. de laat , s. mcmillan , k. hiraki , and s. harfst , `` the cosmogrid simulation : statistical properties of small dark matter halos , '' _ the astrophysical journal _ , vol .
767 , no . 2 , p. 146
[ online ] .
available : http://stacks.iop.org/0004-637x/767/i=2/a=146 t. ishiyama , k. nitadori , and j. makino , `` 4.45 pflops astrophysical n - body simulation on k computer : the gravitational trillion - body problem , '' in _ proceedings of the international conference on high performance computing , networking , storage and analysis _
sc 12.1em plus 0.5em minus 0.4emlos alamitos , ca , usa : ieee computer society press , 2012 , pp .
5:15:10 .
n. karonis , b. toonen , and i. foster , `` mpich - g2 : a grid - enabled implementation of the message passing interface , '' _ journal of parallel and distributed computing _ , vol .
63 , no . 5 , pp . 551 563 , 2003 , special issue on computational grids . c. coti , t. herault , and f. cappello ,
`` , '' in _ _ , ser .
lecture notes in computer science , h. sips , d. epema , and h .- x .
lin , eds.1em plus 0.5em minus 0.4emspringer berlin heidelberg , 2009 , vol .
5704 , pp .
466477 .
s. manos , m. mazzeo , o. kenway , p. v. coveney , n. t. karonis , and b. toonen , `` distributed mpi cross - site run performance using mpig , '' in _ proceedings of the 17th international symposium on high performance distributed computing _
hpdc 08.1em plus 0.5em minus 0.4em new york , ny , usa : acm , 2008 , pp .
229230 .
m. muller , m. hess , and e. gabriel , `` grid enabled mpi solutions for clusters , '' in _
cluster computing and the grid , 2003 .
ccgrid 2003 .
3rd ieee / acm international symposium on_.1em plus 0.5em minus 0.4emieee , 2003 , pp .
d. groen , s. rieder , p. grosso , c. de laat , and s. portegies zwart , `` a light - weight communication library for distributed computing , '' _ computational science and discovery _ , vol . 3 , no .
015002 , aug . 2010 .
e. he , j. alimohideen , j. eliason , n. krishnaprasad , j. leigh , o. yu , and t. defanti , `` quanta : a toolkit for high performance data delivery over photonic networks , '' _ future generation computer systems _ , vol .
19 , no . 6 , pp . 919 933 , 2003
. t. j. hacker , b. d. athey , and b. noble , `` the end - to - end performance effects of parallel tcp sockets on a lossy wide - area network , '' in _ vehicle navigation and information systems conference , 1993 . ,
proceedings of the ieee - iee _ , oct 1993 .
j. bdorf , e. gaburov , m. s. fujii , k. nitadori , t. ishiyama , and s. portegies zwart , `` 24.77 pflops on a gravitational tree - code to simulate the milky way galaxy with 18600 gpus , '' in _ proceedings of the international conference for high performance computing , networking , storage and analysis _ , ser .
sc 14.1em plus 0.5em minus 0.4em piscataway , nj , usa : ieee press , 2014 , pp .
[ online ] .
available : http://dx.doi.org/10.1109/sc.2014.10 n. drost , j. maassen , m. van meersbergen , h. bal , f. pelupessy , s. portegies zwart , m. kliphuis , h. dijkstra , and f. seinstra , `` high - performance distributed multi - model / multi - kernel simulations : a case - study in jungle computing , '' in _ parallel and distributed processing symposium workshops phd forum ( ipdpsw ) , 2012 ieee 26th international _ , may 2012 , pp .
150162 .
j. borgdorff , m. mamonski , b. bosak , k. kurowski , m. ben belgacem , b. chopard , d. groen , p. v. coveney , and a. g. hoekstra , `` distributed multiscale computing with \{muscle } 2 , the multiscale coupling library and environment , '' _ journal of computational science _
, vol . 5 , no . 5 , pp .
719 731 , 2014 .
[ online ] .
available : http://www.sciencedirect.com/science/article/pii/s1877750314000465 d. groen , j. borgdorff , c. bona - casas , j. hetherington , r. nash , s. zasada , i. saverchenko , m. mamonski , k. kurowski , m. bernabeu , a. hoekstra , and p. coveney , `` flexible composition and execution of high performance , high fidelity multiscale biomedical simulations , '' _ interface focus _
, vol . 3 , no . 2 , p. 20120087 , 2013 .
s. f. portegies zwart , s. l. w. mcmillan , a. van elteren , f. i. pelupessy , and n. de vries , `` multi - physics simulations using a hierarchical interchangeable software interface , '' _ computer physics communications _
184 , no . 3 , pp .
456 468 , 2013 . | we describe the political and technical complications encountered during the astronomical cosmogrid project .
cosmogrid is a numerical study on the formation of large scale structure in the universe .
the simulations are challenging due to the enormous dynamic range in spatial and temporal coordinates , as well as the enormous computer resources required .
in cosmogrid we dealt with the computational requirements by connecting up to four supercomputers via an optical network and make them operate as a single machine .
this was challenging , if only for the fact that the supercomputers of our choice are separated by half the planet , as three of them are located scattered across europe and fourth one is in tokyo .
the co - scheduling of multiple computers and the gridification of the code enabled us to achieve an efficiency of up to @xmath0 for this distributed intercontinental supercomputer . in this work , we find that high - performance computing on a grid can be done much more effectively if the sites involved are willing to be flexible about their user policies , and that having facilities to provide such flexibility could be key to strengthening the position of the hpc community in an increasingly cloud - dominated computing landscape . given that smaller computer clusters owned by research groups or university departments usually have flexible user policies
, we argue that it could be easier to instead realize distributed supercomputing by combining tens , hundreds or even thousands of these resources . |
x - ray bursts observed in accreting neutron stars ( ns ) in low - mass x - ray binaries ( lmxbs ) come in two types , sometimes similar in appearance ( @xmath0 min time scales , @xmath1 erg s@xmath2 peak luminosities ) but thought to be due to very different mechanisms , because of the large difference ( by factors @xmath3 ) in average power .
type i bursts arise from the heating and cooling of the ns photosphere after a thermonuclear shell flash of accreted material ( see reviews by @xcite , @xcite , @xcite ) .
this is consistent with thermal spectra of varying temperature and roughly constant emitting area . despite some remaining open questions ( e.g. , , , )
a general understanding of the physics underlying type i bursts has been reached , also thanks to the wide range of circumstances in which they occur ( type of lmxb , accretion rates , etc ) provided in particular by the relatively large amount of bursters ( over a hundred ) .
type ii bursts remain largely a puzzle .
soon after their discovery , @xcite showed that their short recurrence times rule out the possibility that they are powered by nuclear burning , ascribing the source instead to gravitational potential energy .
little more is understood about their origin , also because they are only confidently detected in two sources : the rapid burster ( mxb 1730335 , hereafter rb ) , a recurring transient which is the object of the present study and which also exhibits type i bursts , and the bursting pulsar ( bp ) , a slowly rotating ( 2 hz ) accretion - powered pulsar @xcite .
one of the first properties of type ii bursts discovered in the rb is the relaxation - oscillator behaviour : the fluence in a type ii burst is proportional to the waiting time to the next burst @xcite .
this is not observed in the bp @xcite .
the relaxation - oscillator behaviour seems to point to the presence of a mass storage from which matter can be released and then replenished at a constant rate , so that the amount of depleted material determines the time to refill the buffer @xcite .
the accretion disc is the most obvious candidate for this buffer and all proposed models involve one ( e.g. , @xcite ) .
the various time scales at different positions in the disc can help to explain the range of observed type ii burst time scales , while a key role in acting as a gating mechanism might be played by a magnetic field ( e.g. , recent simulations by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . notwithstanding that models involving some role for the ns magnetic field seem promising ,
the strength of the magnetic field is unfortunately not known in the rb due to the lack of pulsations . for the bp , estimates of the surface strength of the dipolar component of the magnetic field @xmath4 put it in the range @xmath5 g @xcite .
while this is larger than in typical ns lmxbs @xcite , it is not unique among bursters ( estimates put the magnetic field of the ns in igr j17480 - 2446 at @xmath6 g ; e.g. , * ? ? ?
therefore , at least another parameter besides the magnetic field strength ( e.g. , the stellar spin , or the inclination of @xmath4 with respect to the stellar spin axis ) must determine the type ii burst occurrence .
* hereafter b13 ) have shown that thermonuclear ( type i ) x - ray bursts in the rb occur over an unusually large range of accretion luminosities , from about 45 per cent of the eddington luminosity @xmath7 ( assuming it is located in the globular cluster liller 1 at a distance of @xmath8 kpc , @xcite ) down to almost quiescence . however , while the type i bursts span the entire range of luminosities , type ii bursts only appear below a critical luminosity of about 10 per cent @xmath7 ( * ? ?
hence , there are a type - i - only phase and a mixed phase .
the type - i - only state is a soft state because the persistent emission spectrum is dominated by a disc black - body component , and it spans the range 1045 per cent of @xmath7 .
this is almost unique among type i burst sources : all other bursters show nearly no type i bursts above @xmath9 , except for the slow ( 11 hz ) pulsar igr j174802446 @xcite .
the abundance of type i bursts in the rb above this threshold prompted b13 to propose that it is a slow rotator too , and to speculate that this might actually be one of the necessary ingredients for producing type ii bursts .
as the rb exits the soft state , the first type ii bursts appear .
these first type ii bursts are the most energetic , long and infrequent .
a brief ( typically a few days ) intermediate state exists in which these are accompanied by double - peaked type i bursts and a qpo at 0.25 hz ( * ? ? ?
* hereafter b14 ) .
the unusual shape of the type i bursts found in this short - lived intermediate outburst phase seems to be due to a temporary obscuration of the ns surface during the burst decay , producing the double peaks .
the ensuing type ii bursts , the obscuration phenomenon and the occurrence of a qpo are probably related to the change in the accretion geometry that is widely thought to take place between the soft and the hard state .
having studied the soft and intermediate states in our aforementioned papers , we now complete our study of the entire _ rossi x - ray timing explorer _
( _ rxte _ , 19952012 ) rb database by focussing on the type - ii - burst dominated hard state , so called because of the power - law component that dominates the persistent emission spectrum .
we aim at constructing a population study of type ii bursts to compare it with the predictions of the few available theoretical models , and as a future reference . in sec .
[ sec : obs ] we introduce the dataset , and in sec .
[ sec : ana ] we illustrate our methodology . in sec .
[ sec : res ] we present the type ii burst properties and their relationship with the persistent emission and the type i bursts , taking advantage of the unprecedented detail provided by the _
we summarize our findings , some of which were previously unknown , in sec .
[ sec : sum ] , and discuss their physical implications in sec .
[ sec : dis ] . finally , in sec .
[ sec : mod ] we review proposed theoretical models , checking their predictions against the properties of our sample .
+ this study is about measurements with the proportional counter array ( pca ) onboard _ rxte_. it consists of five co - aligned proportional counter units ( pcus ) that combine to a total photon collecting area of 8000 @xmath10 over a 2 to 60 kev bandpass @xcite .
nearly all _ rxte _ observations of the rb ( exposure time 2.4 ms ) were carried out in one of three configurations . for 553 ks
, the pca was directly pointed at the source .
these observations are contaminated by the presence of the persistently bright 4u 1728 - 34 at an angular distance of 0.56@xmath11 . for 612 ks ,
the pointing was offset by just this amount , thus avoiding the contamination at the expense of roughly half the effective area . finally , for 1.3 ms the pca was pointed at 4u 1728 - 34 , serendipitously encompassing the rb .
we have investigated all of the 2.4 ms of data . we extracted a lightcurve from the standard-1 data for all the observations containing the rb in the field of view ( fov ) .
these data have 0.125 s time resolution and no photon energy resolution .
the telescope orientation was determined with the ftool pcaclrsp , and deemed stable if the jitter was below 20 arcsec .
we performed a visual inspection of all the light curves .
for the details of source ( rb vs 4u 1728 - 34 ) and burst ( type i vs type ii ) identification , we refer the reader to appendix [ sec : appendix_search ] . in total
, we identified 123 type i and 8458 type ii bursts from the rb .
as already mentioned , type ii bursts are only emitted by the rb at persistent luminosities below 10 per cent of the eddington limit ( b13 , * ? ? ?
* ) , when the rb is in the hard state ( b14 ) .
actually the rb almost never reaches higher luminosities during outbursts after 1999 (; b14 ) .
three modes of type ii bursting behaviour follow one another in a smooth transition , differing in lightcurves , energetics and bursting patterns @xcite . to illustrate the evolution of the type ii burst behaviour , we plot the light curve from _ rxte _ data of the outburst in june - july 2006 ( fig . [ fig : outburst92026 ] ) .
the first type ii bursts are hundreds of seconds long , often with flat peaks , and have the longest recurrence times ( sometimes longer than a few ks observation ) .
the persistent emission drops right before and after them .
these are the so - called mode-0 type ii bursts ( fig . [ fig : outburst92026 ] , top - left inset ) .
we identified a record long mode 0 burst with a duration of at least 1100 s. it is visible in fig .
[ fig : outburst92026 ] as the long plateau on june 26 . at lower persistent fluxes ,
the large bursts are increasingly preceded by sequences of 8 to 40 short bursts .
after the final longer burst , a longer gap is measured , during which the persistent emission rises to a level higher than that of the intra - burst emission , in a `` hump '' shape . this bursting pattern is called mode 1 ( fig .
[ fig : outburst92026 ] , top - right inset ) .
as the outburst decays further , the sequences of mode 1 type ii bursts become increasingly long , the final bursts less energetic , and the humps shorter and less pronounced .
eventually all bursts have similar energetics and durations : this is the so - called mode 2 ( fig .
[ fig : outburst92026 ] , bottom - middle inset ) .
finally , even these bursts become increasingly sparse , as the source moves to quiescence ( fig .
[ fig : outburst92026 ] , bottom - right inset ) .
we caution that the sequence `` no type ii bursts @xmath12 mode 0 @xmath12 mode 1 @xmath12 mode 2 '' is only a _
general _ trend .
for instance , the flat peak in fig .
[ fig : outburst92026 ] starting on june 27 is not a very long mode 0 type ii burst , but an actual switch of the source to the soft state , during which a type i burst can be seen , but no type ii bursts .
two observations were taken the day before and the day after showing long mode 0 type ii bursts , meaning that the soft state lasted briefly .
also , the rb can sometimes be seen switching back and forth between modes .
the switch is sudden rather than gradual ( as shown in fig .
[ fig : mode_switching ] , already noted by * ? ? ?
* ) , although the rb can be seen to transit back and forth between modes for a few ks .
it is clear from the picture that the minima before and after a mode 0 burst are not `` dips '' : rather , the humps in between are what is peculiar and are a feature common to all very energetic type ii bursts , independent of bursting mode .
the sampling along an outburst is often patchy , meaning that it is not easy to establish an average duration for each mode .
the outburst in fig .
[ fig : outburst92026 ] is probably the best sampled one .
it features the longest mode 0 phase in the entire dataset , lasting @xmath13 days , albeit including a brief return into the soft state . the longest uninterrupted mode 0 bursting phase occurred in an earlier outburst , and lasted @xmath14 days .
mode 1 bursts are visible in fig . [ fig : outburst92026 ] for @xmath15 days . in a later outburst ,
they were observed to last for twice that time , @xmath16 days .
finally , mode 2 burst were continuously observed for up to 24 days , the longest stretch in the dataset .
our search routine found 8458 type ii bursts in the entire dataset ( see sec .
[ sec : sample ] ) .
the results are summarized in table [ table : bursts ] .
we determined the instrumental and cosmic diffuse background with the ftool pcabackest .
no bursts were analyzed that are outside so - called good time intervals ( gti ) , are affected by data gaps are overlapping with type i bursts ( from either the rb or 4u 1728 - 34 ) , or are in obsids for which the collimator response was not constant because of a slew .
this left us with 7601 bursts .
of these , 3662 type ii bursts are present in 389 ks of direct pointings , while 3183 were observed during 341 ks of offset pointings , and 756 during 92 ks of 4u 1728 - 34 observations .
the time resolution of the spectral analysis varied across the different observations and within observations is chosen to follow the variability in flux . in total 374 spectra
were extracted from these 11 observations , involving 8 mode-0 bursts ( including the longest one detected ) , 25 mode-1 bursts and 23 mode-2 bursts .
the spectral analysis involves the following generalities .
spectra were extracted from event - mode data with 64 spectral channels .
all active pcus were employed .
the bandpass was limited to 2.5 - 20 kev which is where the pca is well calibrated @xcite and contains most of the signal .
a systematic error of 0.5 per cent per channel was included ( following * ? ? ?
all spectra were corrected for particle background and cosmic diffuse background as predicted through the pcabackest tool ( version 3.8 ) .
the pca response matrix was calculated with version 11.7.1 of tool pcarsp .
this includes a correction for the collimator response which is always non - standard because the rb is always considerably off - axis in the selected obsids .
the collimator throughput for the rb was usually about 0.4 .
spectral bins were combined to obtain at least 20 counts per bin to ensure applicability of the @xmath18 statistic .
all spectral modelling was carried out with xspec version 12.8.2 .
a constant cold interstellar absorption component was assumed to apply , equivalent to @xmath19 @xmath20 @xcite .
the absorption model of @xcite was followed .
a significant contribution may be expected from the galactic ridge . to determine this , off - axis observation i d 92026 - 01 - 04 - 00
was chosen which represents the observation with the lowest flux measured for the rb field .
it was taken on june 9 , 2006 , 11 days before the onset of the outburst presented in fig . 1 .
the background - subtracted data was modelled with the combination of a raymond - smith component @xcite and a power law , following the prescription of @xcite as also determined from pca ridge data .
the fit was satisfactory with @xmath21 ( @xmath22 ) .
this constant spectral component was assumed to be present in all studied rb spectra .
we tested various models on the data : single component models ( power law , planck function , thermal bremsstrahlung , comptonization ) and models with combinations of two components .
it was found that models with three free parameters were sufficient to describe all spectra .
we chose the comptonization model according to @xcite as baseline since this is the simplest model fitting all data .
furthermore , we find that this model performs best to the highest quality spectrum ( 20093 - 01 - 07 - 01r whose spectrum contains 10@xmath23 counts ) , with @xmath24 for @xmath22 . as an example , for the same data @xmath25 for a model consisting of a disk black body and a power law , and 6.88 for a cut - off power law .
we note that it was necessary to complement the continuum model with a narrow line at about 6.5 kev , in the spectral region where the fe - k resonance line resides ( 6.4 - 6.9 kev , depending on the ionization state of fe ) .
possibly it is related to instrumentally scattered emission from the bright source 4u 1728 - 34 just outside the field of view , or to an inadequate modelling of the galactic ridge at this location .
figure [ spec_examplek ] shows an example of a time history for observation 92026 - 01 - 06 - 00 which shows a sequence of 5 bright and relatively short mode-0 bursts .
what is clearly visible here , and that applies in general to all our spectral results , is that plasma temperature and optical depth show no significant variability , while photon seed temperature does .
the plasma temperature varies in general between 2.5 and 3.5 kev while at the same time the plasma optical depth varies between 7 and 4 , both typical 1-sigma errors of 1 in value .
a more detailed analysis of the spectral data is subject for a future paper .
we here confine the analysis to inferring an average translation factor between photon fluxes and fluences and energy fluxes and fluences . for the most accurately determined 112 spectral shapes , we find that a photon flux of 1 c s@xmath2pcu@xmath2 for the whole bandpass translates on average to a 2.5 - 20 kev energy flux of @xmath26 erg @xmath20s@xmath2 .
all 112 conversion values are contained within 7% from this average .
thus , the maximum systematic error by assuming a constant conversion factor is 7% . converting from 2.5 - 20 kev energy flux to unabsorbed bolometric flux involves a factor of @xmath27 .
all 112 bolometric correction factors range between -7% and + 2% from the mean .
therefore , the maximum systematic error in the bolometric correction factor is 7% . in fig .
[ spec_spectra ] we show spectra of the five most diverse states of the rb . the conversion factor from c s@xmath2pcu@xmath2 to unabsorbed bolometric erg @xmath20s@xmath2 varies between @xmath28 and @xmath29 .
spectra for five widely different states of rb . from bottom to top : non - burst emission between slow mode-2 bursts ( black curve , obsid 30424 - 01 - 06 - 01r ) , non - burst emission in long intermission between mode-1 bursts ( light blue , 30424 - 01 - 02 - 02r ) , peak of longest mode-0 burst ever measured ( green ; 92026 - 01 - 06 - 03 ) , soft state prior to mode-0 ( red ; 20418 - 01 - 01 - 01r ) , peak of bright short mode-0 burst ( dark blue ; 30424 - 01 - 01 - 02r ) .
the lowest spectrum shows the iron line clearly whose origin is undetermined .
[ spec_spectra ] ] | type ii bursts are thought to arise from instabilities in the accretion flow onto a neutron star in an x - ray binary .
despite having been known for almost 40 years , no model can yet satisfactorily account for all their properties . to shed light on the nature of this phenomenon and provide a reference for future theoretical work , we study the entire sample of _ rossi x - ray timing explorer _ data of type ii bursts from the rapid burster ( mxb 1730335 ) .
we find that type ii bursts are eddington - limited in flux , that a larger amount of energy goes in the bursts than in the persistent emission , that type ii bursts can be as short as 0.130 s , and that the distribution of recurrence times drops abruptly below 1518 s. we highlight the complicated feedback between type ii bursts and the ns surface thermonuclear explosions known as type i bursts , and between type ii bursts and the persistent emission .
we review a number of models for type ii bursts . while no model can reproduce
all the observed burst properties and explain the source uniqueness , models involving a gating role for the magnetic field come closest to matching the properties of our sample .
the uniqueness of the source may be explained by a special combination of magnetic field strength , stellar spin period and alignment between the magnetic field and the spin axis .
[ firstpage ] stars : neutron x - rays : binaries x - rays : bursts x - rays : individual : mxb 1730 - 335 |
many scientists have contributed to our knowledge and understanding of the accelerating universe and the properties of the dark energy . in keeping with the request of the conference organizers
, this paper only addresses my contributions to this field .
the years 1997 and 1998 were exciting times for cosmology !
i had been studying powerful classical double radio galaxies since the early 1990s @xcite , and had proposed a new method of using radio galaxies as a modified standard yardstick in 1994 @xcite .
this work continued with princeton university phd thesis students eddie guerra , lin wan , and greg wellman , and some of this work included cosmological studies ( @xcite , @xcite , @xcite , @xcite , @xcite ) , and studies of outflows from the supermassive black holes that power the radio sources ( @xcite , @xcite , @xcite , @xcite , @xcite ) .
the cosmological studies were done in the context of two cosmological world models : one that included non - relativistic matter , a cosmological constant , and space curvature , and another that included `` quintessence '' with constant equation of state , non - relativistic matter , and zero space curvature .
later , radio galaxies were studied in the context of a cosmological model that included a rolling scalar field , non - relativistic matter , and zero space curvature @xcite , @xcite .
the cosmological results eventually published by @xcite were presented on january 9 , 1998 at the aas meeting in washington , d. c. ; the sample of twenty radio galaxies studied is briefly mentioned in aas bulletin abstract 95.04 .
these results indicated that a cosmological constant provided a good fit to the radio galaxy data . in late 1997 , steve maran from the aas press office invited me to prepare a press release on cosmological studies with radio galaxies to be presented on january 8 , 1998 , and i accepted this invitation .
the press release explains how distant radio galaxies can be used to study the expansion history of the universe . for
a given observed angular size of the radio source , a large intrinsic size meant that the coordinate distance to the source was large , and that the universe was accelerating in its expansion , or , in the words of the release , `` the expanding universe will continue to expand forever , and will expand more and more rapidly as time goes by . ''
this is explained again later in the release where it states `` the universe will continue to expand forever and will expand at a faster and faster rate as time goes by . ''
the press release session included supernova results presented by adam riess and saul perlmutter , and i had an opportunity to discuss my conclusions with adam and saul in detail .
they were surprised to hear that a cosmological constant provided a good fit to the radio galaxy data , which implied the universe would expand at an ever increasing rate .
each expressed a similar concern , this concern being the dependence of the result on the cosmological model or world view under consideration .
the result that i was reporting on was obtained in the context of a cosmological model that allowed for non - relativistic matter , a cosmological constant , and space curvature .
if different components were present in the universe , would the data still imply that the universe was accelerating ?
in fact , addressing this concern was part of the motivation for developing a completely model - independent approach to the analysis and interpretation of radio galaxy and supernova data ( described below ) .
two supernova groups showed that a cosmological constant provides a good description of the supernova data ( e.g. @xcite , @xcite ) .
as time passed and more data were analyzed , it became clear that these two completely independent methods , friib radio galaxies and type ia supernovae , based on totally different types of sources and source physics , yield very similar results @xcite
, @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
this was important because it suggested that neither method was plagued by unknown systematic errors .
one of the reasons the radio galaxy method provides interesting results with a relatively small number of sources is that many of the radio sources are at relatively high redshift .
for example , the highest redshift source in either the radio galaxy or supernovae samples is the radio galaxy 3c 239 at a redshift of 1.79 , which has been included in the radio galaxy studies since 1998 @xcite .
differences between predictions of various cosmological models become large at high redshift , so high redshift data points can have a strong impact on results .
the methods of using type ia supernova and type iib radio galaxies for cosmological studies are empirically based .
it could be empirically demonstrated that the methods worked well , but the underlying physical processes were not understood well enough to explain why the methods worked so well .
this changed in 2002 for the radio galaxies , when the reason that the radio galaxy method works so well began to become clear @xcite .
the radio galaxy method is applied to very powerful classical double radio galaxies , such as the radio source cygnus a ( 3c 405 ) .
these friib radio galaxies are powered by very energetic , highly collimated outflows from regions very close to a supermassive black hole located at the center of a galaxy .
when the collimated outflow impacts the ambient gas , a strong shock wave forms , and a shock front separates the radio emitting material from the ambient gas @xcite .
the physics of strong shocks is fairly simple and straight - forward , and makes these systems ideal for cosmological studies .
large - scale outflows from supermassive black holes are thought to be powered by the spin energy of the hole ( e.g. @xcite , @xcite , @xcite ) .
when @xcite cast the radio galaxy method in the language of the blandford - znajek model to extract the spin energy from a rotating black hole @xcite , it became clear that the outflow from the hole occurs when the strength of the magnetic field near the hole reaches a maximum or limiting value .
this value can be written as a function of the black hole mass , spin , and the radio galaxy model parameter , @xmath0 . when the radio galaxy model parameter has one particular value , @xmath1 , the relationship between the magnetic field strength and the properties of the rotating hole is greatly simplified , and the field strength depends only upon the black hole spin .
empirical studies by @xcite and @xcite found that the value of @xmath0 is very close to 1.5 , @xmath2 .
thus , the reason the radio galaxy model works so well is that the outflow from the supermassive black hole is triggered when the magnetic field strength reaches a maximum or limiting value that depends only upon the black hole spin .
interestingly , other models , such as that by @xcite , have the same functional form as the blandford - znajek model @xcite but with a different constant of proportionality , and the results of @xcite apply to any model with the same functional form as the blandford - znajek model .
from 1998 to 2002 the study of the acceleration of the universe was done in the context of particular cosmological world models , and the question of whether the acceleration of the universe could be studied independent of a particular cosmological model and independent of a theory of gravity captivated my interest . to address this question
, i worked to develop an assumption - free , or model - independent , method of analyzing supernova , radio galaxy , or other data sets that provide coordinate distances .
the method was proposed in 2002 @xcite , and , in collaboration with george djorgovski , was developed and applied to supernova and radio galaxy data sets @xcite , @xcite , @xcite , @xcite , @xcite . assuming only that the friedmann - lematre - robertson - walker ( flrw ) line element is valid , coordinate distance measurements can be used to obtain the expansion and acceleration rates of the universe as functions of redshift .
coordinate distance measurements are easily obtained from luminosity distances or angular size distances to any type of source ( e.g. supernovae or radio galaxies ) .
the flrw line element is the most general metric describing a homogeneous and isotropic four - dimensional space - time .
these determinations of the expansion and acceleration rates of the universe are independent of a theory of gravity , and independent of the contents of the universe ( @xcite , @xcite ) .
it was shown by @xcite that the zero redshift value of the dimensionless acceleration rate of the universe is independent of space curvature , and that very similar results are obtained for the dimensionless acceleration rate @xmath3 and the expansion rate of the universe @xmath4 for zero and reasonable non - zero values of space curvature .
thus , the model - independent method can be applied without requiring that space curvature be set equal to zero .
it was shown by @xcite , @xcite , @xcite , @xcite , and @xcite that the universe is accelerating today and was most likely decelerating in the recent past , and this result is independent of a theory of gravity , of the contents of the universe , and of whether space curvature is non - zero ( for reasonable non - zero values ) .
recent determinations of @xmath4 and @xmath3 obtained using the model - independent method are compared with predictions in a standard lambda cold dark matter ( lcdm ) model in fig .
1 ( the thin solid line shows the lcdm prediction ) .
these results indicate that the lcdm model provides a good description of the data to a redshift of about one ( e.g. @xcite , @xcite , @xcite , @xcite , @xcite ) .
the lcdm model assumes that general relativity ( gr ) is the correct theory of gravity , space curvature is equal to zero , and two components contribute to the current mass - energy density of the universe , a cosmological constant and non - relativistic matter with 70 % and 30 % , respectively , of the normalized mean mass - energy density of the universe at the current epoch . as discussed by @xcite , a comparison of model - independent determinations of @xmath4 and @xmath3 with predictions in the lcdm and other models provides a large - scale test of gr .
current observations suggest that gr provides an accurate description of the data over look back times of about ten billion years @xcite .
there is a hint of a deviation of the data from predictions in the lcdm model at redshifts of about one ( @xcite , @xcite ) .
the model - independent approach can be extended to solve for the properties of the dark energy as a function of redshift @xcite , where the `` dark energy '' is the name given to whatever is causing the universe to accelerate .
assuming that gr is valid on very large length scales , and that space curvature is zero , supernova and radio galaxy data can be used to solve for the pressure , energy density , equation of state , and potential and kinetic energy densities of the dark energy as functions of redshift ( @xcite , @xcite ) , as shown in fig .
2 . results obtained using the model - independent approach can provide valuable information to theorists developing new ideas to explain the acceleration history of the universe and the properties of the dark energy .
this is complementary to the commonly adopted approach of assuming a particular dark energy model and cosmological world model and solving for best fit model parameters ( e.g. @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . in studies of the properties of the dark energy ,
the equation of state of the dark energy has surfaced as an important parameter . a cosmological
constant has an equation of state that is always equal to @xmath5 . to study the equation of state of the dark energy in a model - independent manner
, @xcite defined a new model - independent function , called the dark energy indicator .
the dark energy indicator provides a measure of deviations of the equation of state from @xmath5 as a function of redshift .
current data suggest that a value of @xmath6 provides a good description of data at redshift less than 1 ( see fig .
the radio galaxies described above are powered by large - scale outflows from the vicinity of supermassive black holes .
studies of the properties of a radio galaxy allow the energy per unit time , known as the `` beam power , '' that is being channeled from the vicinity of the supermassive black hole to the large - scale outflow to be determined .
studies of the beam power and other source properties provide important insights and information on these black hole systems ( e.g. @xcite , @xcite , @xcite ) .
for example , the beam power can be combined with the radio galaxy model parameter @xmath0 to solve for the total energy that will be channeled away from the vicinity of the supermassive black hole over the full lifetime of the outflow ( e.g. @xcite , @xcite , @xcite ) .
the total energy of the outflow can be combined with the black hole mass to obtain a lower bound on the spin of the supermassive black hole @xcite , assuming only that the highly collimated outflow is powered by the spin energy of the supermassive black hole .
this is one of the very few direct indications of the spin of supermassive black holes known at present .
the ratio of the total outflow energy to the black hole mass appears to be constant for these black hole systems .
this ratio provides an important diagnostic of the physical state of the black hole system at the time the outflow is generated , and the results of @xcite indicate that each system is in a similar physical state when the outflow is triggered . thus , these studies provide insights into the physical conditions of supermassive black holes systems and their state at the time powerful outflows are generated
. dimensionless coordinate distances ( filled symbols indicate radio galaxies and open symbols indicate supernovae ) , @xmath4 for supernovae ( middle left panel ) and radio galaxies ( bottom left panel ) , and @xmath3 for supernovae ( top right panel ) and radio galaxies ( bottom right panel ) ; from daly et al .
, width=292 ] dimensionless coordinate distances ( filled symbols indicate radio galaxies and open symbols indicate supernovae ) , @xmath4 for supernovae ( middle left panel ) and radio galaxies ( bottom left panel ) , and @xmath3 for supernovae ( top right panel ) and radio galaxies ( bottom right panel ) ; from daly et al .
( 2008 ) . , width=292 ] dark energy pressure ( top left panel ) , energy density ( middle left panel ) , and equation of state ( bottom left panel ) , and the dark energy indicator ( right panel ) as functions of redshift for a combined sample of supernovae and radio galaxies ; from daly et al .
( 2008 ) . , width=292 ] dark energy pressure ( top left panel ) , energy density ( middle left panel ) , and equation of state ( bottom left panel ) , and the dark energy indicator ( right panel ) as functions of redshift for a combined sample of supernovae and radio galaxies ; from daly et al .
, width=292 ] it is a pleasure to thank david cline for encouraging me to give this presentation , and my collaborators in these endeavors , especially george djorgovski , chris odea , preeti kharb , and stefi baum .
this work was supported in part by u.s .
nsf grants ast-0096077 , ast-0206002 , and ast-0507465 . | the years 1998 to 2008 were very exciting years for cosmology .
it was a pleasure to accept this invitation to describe my contributions to the development of our knowledge and understanding of the universe over the course of the past decade .
here , i begin by describing some of my work on radio galaxies as a modified standard yardstick and go on to describe model - independent studies of the accelerating universe and the properties of the dark energy .
during the course of these studies , i came upon interesting ways to study the spin and other properties of supermassive black holes , some of which are briefly mentioned .
address = department of physics , penn state university , berks campus , p. o. box 7009 , reading , pa 19610 |
the welded pipelines subjected to high pressure and temperature are widely used in different branches of industry . under such conditions , the creep and damage effects should be taken into account for accurate assurance of long - term reliability @xcite .
application of computational continuum creep damage mechanics ( see , for example @xcite , @xcite ) coupled with increasing power of computers can accomplish this task . in recent years the finite element method has become the widely accepted tool for the structural analysis in the creep range @xcite .
a user defined creep material subroutine with appropriate constitutive and evolution equations can be developed and incorporated into the commercial finite element code to perform a numerical time step solution of creep and long term strength problems .
on the other hand , in addition to more and more sophisticated numerical analysis , simplified models of creep response are required .
these models should provide a better intuitive insight into the problem and give a quantitative description of the solution .
the assessment of reliability of user - defined creep material subroutines and the choice of suitable numerical parameters like the element type , the mesh density , and time step control are complicated problems , particulary if studying creep of multi - material structures .
therefore , it is important to have reference solutions of benchmark problems .
such solutions should be obtained by use of alternative analytical or semi - analytical methods which do not require the spatial discretization techniques and allow for studying the behavior of stress and deformation gradients .
the objective of this paper is to develop an alternative semi - analytical solutions to creep problems for multi - material pipe structures .
particularly we address the analysis of stress gradients in the local zones of material connections . to obtain a semi - analytical solution we shall make the following simplifications .
we assume the idealized material behavior having the secondary creep stage only . in this case
the steady state solution of creep in the pipe exists , for which the stresses do not depend on time .
we assume that the difference between the material properties of constituents is not great .
particularly the difference between the minimum creep rates for the same stress level should not exceed the value of 2 .
the lifetime of a welded pipe under creep conditions is less than that of homogeneous one .
the effect of reliability reduction is of big interest , therefore large numbers of model problems were proposed .
the most commonly used approach simulates weldment as a region with non - uniformly distributed material properties ( @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . within the framework of this approach
it is often necessary to consider a number of parameter distribution cases .
the amount of problems to be solved increases with the number of changing material parameters and parametric analysis becomes very complicated .
drawing an analogy with some simple systems , for which an analytical solution is available ( @xcite , @xcite ) is useful for understanding how the parameter change can affect the solution .
however this is not enough for a proper estimation of stress distribution .
two main types of constitutive equations are often used in weld modelling , namely , norton s steady - state creep law and continuum damage mechanics equations for tertiary creep @xcite . below only the constitutive equations of norton s law
are considered . even by this assumption
the structural response may be captured very well .
some special techniques are used to predict the failure life more precisely using steady - state solution ( @xcite , @xcite , @xcite , @xcite ) . as shown in @xcite , @xcite for the particular pipe welds investigated ,
the steady - state analysis underestimates the failure time by about 20 - 40 percent , but predicts failure position quite well . to simplify the parametric analysis
, we study a family of boundary - value problems for multi - component pipe creep depending on a small parameter @xmath0 . when @xmath1 , the problem is reduced to the case of homogeneous pipe creep .
the corresponding solution @xmath2 for the steady - state stress distribution is well known ( @xcite , @xcite , @xcite ) and considered to be the basic solution . in order to get the common solution @xmath3 from this basic one
, correction terms should be added .
the equations for the correction terms are formally obtained as perturbations of a boundary - value problem with respect to @xmath0 .
these equations formulate a problem of linear elasticity with respect to linear - elastic solid with anisotropic elastic properties .
the utility of used technique is guaranteed especially because the theory of linear elasticity is in a very satisfactory state of completion ; every complicated case of parameter distribution can be treated in a routine manner as a combination of simple ones .
correction terms are obtained numerically for some problems with the help of the ritz method . in some cases
the simplicity of geometry enables us to construct an approximate analytical solution .
an exact expression for stress jumps at the interface is obtained .
numerical solutions of the nonlinear problem are obtained with the help of ansys finite element code for comparison . throughout light - face letters
we denote scalars , the bald - face letters stand for tensors .
the notation @xmath4 is used in vector - matrix form of constitutive equations to designate vectors @xmath5 and @xmath6 of stress and strain components .
in this section we consider a two - material model only . it will be shown later that the solution for some multi - material models can be reduced to this case .
the configuration analyzed is shown in fig .
[ fig1 ] .
[ m][][1][0]parent [ m][][1][0]material [ m][][1][0]@xmath7 [ m][][1][0]weldment [ m][][1][0]@xmath8 [ m][][1][0](a ) [ m][][1][0](b ) [ m][][1][0]@xmath9 [ m][][1][0]@xmath10 [ m][][1][0]@xmath11 [ m][][1][0]@xmath12 [ m][][1][0]@xmath13 [ m][][1][0]@xmath14 [ m][][1][0]@xmath15 [ m][][1][0]@xmath16 [ m][][1][0]@xmath14 [ m][][1][0]@xmath17 assume that @xmath18 , @xmath19 , @xmath20 describe the volume occupied by the solid , by the weld metal , and by the parent material , respectively .
here @xmath21\times[0,h)$ ] , @xmath22\times[h , h]$ ] , @xmath23 .
the basic equations of the problem are given below .
equations of equilibrium @xmath24 where @xmath25 is the stress tensor . in the volumetric forces are ignored .
strain - displacement relations @xmath26 where @xmath27 is the linearized strain tensor and @xmath28 is the displacement vector .
the governing equations can be summarized as norton s creep law ( see , for example , @xcite ) @xmath29 where @xmath30 is the time derivative , @xmath31 is the stress deviator , @xmath32 is the von mises equivalent stress , @xmath33 is the second rank unit tensor , and @xmath34 , @xmath35 are material constants .
note that norton s creep law is often written as @xmath36 .
it means @xmath37 in our notation . if we take into account that the problem is axisymmetric , we have in cylindrical coordinates the following equilibrium equations @xmath38 strain - displacement relations @xmath39 and boundary conditions @xmath40 @xmath41 finally we define the distribution of the parameter @xmath42 in norton s creep law as a piecewise constant function in @xmath43 @xmath44 the stress field @xmath25 , defined by , does not change if instead of @xmath42 we use @xmath45 , where @xmath46 is any nonzero constant .
consequently , without loss of generality it can be assumed that is not necessary a small parameter . in practice ,
@xmath47 and @xmath48 might differ essentially . ] @xmath49 if we eliminate displacements from the strain - displacement relations , we obtain compatibility equations @xmath50
@xmath51 we consider the weak form of compatibility equations expressed by the equation of complementary virtual power principle @xcite @xmath52 @xmath53 here we use the brackets @xmath54 to enclose the argument of a linear operator ; @xmath55 is the strain rate defined by and ; @xmath56 ; @xmath57 is a virtual stress field that satisfy the equations of equilibrium and homogeneous boundary conditions @xmath58 in what follows we search function @xmath59 , such that @xmath60
now suppose that @xmath61 , i.e. @xmath1 . the problem is reduced to a one - dimensional , and the solution @xmath62 of this problem is well known ( @xcite , @xcite , @xcite ) @xmath63 @xmath64 let @xmath0 be a small parameter .
assume that @xmath65 here @xmath66 is the unknown derivative which must satisfy the equations of equilibrium and the homogeneous boundary conditions ; @xmath67 is the little - o landau symbol .
from it follows that @xmath68 with @xmath69 , @xmath70 . substituting and in , we get @xmath71 since @xmath72 and @xmath73 as @xmath74 , we have an equation for @xmath75 @xmath76 we will analyze more closely the linear operator @xmath77 in the next section .
in the previous section it was shown that the correction term @xmath75 can be found from the linear equation .
it is clear that @xmath78 where @xmath79 let us introduce a vector notation @xmath4 as @xmath80 substituting in and differentiating we obtain @xmath81 here we have introduced the compliance matrix as follows @xmath82 @xmath83 \,.\ ] ] thus , the left - hand side of can be treated as an internal complementary virtual work with respect to linear elastic solid with constitutive law @xmath84 the eigenvalues of the compliance operator @xmath85 are @xmath86 the problem of steady - state creep is reduced to the elasticity problem for orthotropic , incompressible , and inhomogeneous solid .
we now seek to convert the right - hand side of into a surface integral through gauss theorem .
it can be proved that @xmath87 let us show that prescribes a jump of displacements at the interface @xmath88 ( see fig .
[ fig2 ] ) .
[ m][][1][0]@xmath89 [ m][][1][0]@xmath13 [ m][][1][0]@xmath12 [ m][][1][0]@xmath90 [ m][][1][0]@xmath91 [ m][][1][0]@xmath92 [ m][][1][0]@xmath7 [ m][][1][0]@xmath8 consider the principle of virtual complementary work for linear elastic solids @xmath93 and @xmath94 .
one gets @xmath95 @xmath96 @xmath97 @xmath98 suppose @xmath99 latexmath:[\[\label{dispjump } u^+_z - u^-_z=0 , \quad u^+_r - u^-_r=3^{\frac{n-1}{2}}a_r that the solution @xmath75 of has the form @xmath101 we define a notation for jumps of field variables at the interface @xmath102 @xmath103=v^+ - v^-.\ ] ] substituting for @xmath104 in , we get @xmath105=-c / r^2 , \quad [ \varepsilon_\theta]=c / r^2 , \quad c=3^{\frac{n-1}{2}}a_r |a_r|^{n-1}\frac{1}{n^n}.\ ] ] if we combine this with , , and , we obtain @xmath106=-\frac{\sqrt{3 } a_r |a_r|}{n^3 } r^{-4/n } , \quad [ \sigma^1_{\theta}]=\frac{\sqrt{3 } a_r |a_r|}{n^3 } r^{-4/n}.\ ] ] in this subsection we consider @xmath107-material structure ( @xmath108 ) @xmath109 @xmath110 \times [ z_j , z_{j+1 } ] , \
z_1=0 , \ z_{m+1}=h,\ ] ] @xmath111 there exist unique @xmath0 and @xmath112 such that ( fig .
[ fig3 ] ) [ m][][1][0]@xmath113 [ m][][1][0]@xmath114 [ m][][1][0]@xmath115 [ m][][1][0]@xmath116 [ m][][1][0]@xmath13 [ m][][1][0]@xmath117 [ m][][1][0]@xmath118 [ m][][1][0]@xmath119 @xmath120 here @xmath121 is the heaviside function .
if we use the small parameter method to solve this problem , then we get an approximation in the form ; @xmath62 is given by and @xmath75 is obtained from . since is valid , it follows that the right - hand side of has the form @xmath122 owing to the fact that is linear , the perturbation @xmath75 is a linear combination of @xmath123 solutions of type .
in this section we construct an analytical solution for the first perturbation term @xmath124 . consider a pair of stress functions @xmath125 , such that condition is satisfied @xmath126 these functions were constructed from compatibility equations , using the standard technique @xcite .
this choice of stress functions simplifies the application of variational methods .
we use the kantorovich method ( see , for example , @xcite ) to reduce the 2-d variational problem to 1-d variational problem .
suppose that for @xmath125 , that define the solution @xmath124 of , the following static hypothesis is valid @xmath127 where @xmath128 are given kantorovich trial functions , such that @xmath129 we use the kantorovich method to obtain a system of differential equations and boundary conditions for @xmath130 .
this method gives a projection of @xmath124 on the subspace , defined by .
it is more convenient to solve in the form . for the sake of brevity
we describe only the application of the kantorovich method for equation . for equation
this procedure can be arranged in a similar manner .
since @xmath131 and @xmath132 are fixed , the variation of gives @xmath133 let us rewrite , in terms of @xmath134 @xmath135 @xmath136 @xmath137 here @xmath138 and @xmath139 are defined by , , and .
using gauss theorem and the fundamental lemma of the calculus of variation @xcite we obtain compatibility equations and boundary conditions @xmath140 @xmath141 @xmath142 here @xmath143 and @xmath144 are defined by , .
after integration , system has the form @xmath145 elimination of @xmath146 from gives @xmath147 after integration in , , we obtain @xmath148 @xmath149 @xmath150 dealing with , in a similar fashion we obtain compatibility equations and boundary conditions for @xmath151 in @xmath152 . finally , taking into account , , we have eight boundary conditions @xmath153 @xmath154=0 , \quad [ \frac{d \psi_2}{dz}]=0,\ ] ] @xmath148 @xmath155 @xmath156=0,\ ] ] @xmath157=\int^{r_o}_{r_i } \frac{c}{r^2 } \psi_1(r ) dr.\ ] ] solution @xmath151 is uniquely defined by these equations combined with two ordinary differential equations ( ode ) of fourth order ( one equation in @xmath158 and another in @xmath152 ) . in this section
we solve problem numerically with the help of the ritz method .
let @xmath159 be a system of stresses , satisfying and .
suppose @xmath160 then the unknown constants @xmath161 are defined from a system of @xmath162 linear algebraic equations @xmath163 we use the system @xmath164 , produced by means of .
the complete system of @xmath165 is given by combinations of trigonometric functions @xmath166
@xmath167 @xmath168 @xmath169 in order to approximate the stress - jump at the interface @xmath88 it could be useful to consider additionally discontinuous functions @xmath170 @xmath171 here @xmath121 is the heaviside function .
if we put @xmath172 , then the solution tends to the approximate analytical solution from the previous section .
the specimen dimensions , applied loads , and material constants are given by @xmath173 , @xmath174 , @xmath175 , @xmath176 , @xmath177 , @xmath178 .
first we compare the approximate analytical solution of ( section 5.1 ) with the numerical solution which was obtained using the ritz method ( section 5.2 ) .
we used @xmath179 in to construct the analytical solution .
values of constants @xmath180 are given in appendix .
the numerical solution by the ritz method is performed witn @xmath181 in .
we use @xmath182 to illustrate on fig .
[ fig4 ] the distribution of stress components along the @xmath13 axis . along @xmath182
the @xmath183 stress is negligibly small for both solutions , therefore we do not plot this component .
hypothesis imposes essential restrictions on the class of solutions , nevertheless the value of shear stress @xmath184 is captured very well .
the error for the hoop stress @xmath185 and the radial @xmath186 stress is significant in the vicinity of the interface @xmath88 .
[ m][][1][0]@xmath89 [ m][][1][0]@xmath187 [ m][][1][0]@xmath188 [ m][][1][0]@xmath189 [ m][][1][0]@xmath190 [ m][][1][0]@xmath191 [ m][][1][0]@xmath192 [ m][][1][0]@xmath193 [ m][][1][0]@xmath194 [ m][][1][0]@xmath13 [ m][][1][0]@xmath184 [ m][][1][0]@xmath195 [ m][][1][0]@xmath196 [ m][][1][0]@xmath197 [ m][][1][0]@xmath32 [ m][][1][0]@xmath198 [ m][][1][0]@xmath199 [ m][][1][0]@xmath200 [ m][][1][0]@xmath201 [ m][][1][0]@xmath185 [ m][][1][0]@xmath202 [ m][][1][0]@xmath200 [ m][][1][0]@xmath203 [ m][][1][0]@xmath204 the approximate analytical solution gives only a smooth part of the stress field .
this solution can be more useful in case of smooth changing material properties .
we investigate the error of the perturbation method @xmath205 .
numerical solutions @xmath59 of the system are obtained with the help of ansys finite element code for a set of @xmath0 .
the geometry of the half of the pipe was represented by 3200 axisymmetric plane183 finite elements ( fig .
[ fig5 ] ) .
we used a uniform mesh with 160 elements along the axis direction z and 20 elements along the radial direction r. [ m][][1][0]pressure @xmath206 this type of element can model creep behavior but the norton s constitutive law is available only as a secondary creep equation .
that is , the total strain is a sum of elastic strain and creep strain @xmath207 if the applied load remains the same with time @xmath208 , then we have @xmath209 and we obtain the steady - creep solution as @xmath210 . in calculations we used @xmath211 for young modulus and @xmath212 for poisson ratio to simulate the elastic material .
we consider the solution to be close enough to the asymptotic solution at the moment @xmath213 of time @xmath208 .
the solution @xmath75 of the auxiliary problem is given by the ritz method ( see section 5.2 ) .
here we used series with @xmath181 .
the leading term in the asymptotic series @xmath214 is a good approximation even if the `` small '' parameter @xmath0 equals 0.5 ( fig .
[ fig6 ] ) . with subsequent increase of @xmath0
the error grows dramatically .
[ m][][1][0]@xmath89 [ m][][1][0]@xmath187 [ m][][1][0]@xmath188 [ m][][1][0]@xmath189 [ m][][1][0]@xmath190 [ m][][1][0]@xmath191 [ m][][1][0]@xmath192 [ m][][1][0]@xmath193 [ m][][1][0]@xmath194 [ m][][1][0]@xmath13 [ m][][1][0]@xmath184 [ m][][1][0]@xmath185 [ m][][1][0]@xmath215 [ m][][1][0]@xmath216 [ m][][1][0]@xmath217 [ m][][1][0]@xmath218 [ m][][1][0]@xmath219 [ m][][1][0]@xmath220 [ m][][1][0]@xmath221 [ m][][1][0]@xmath222 [ m][][1][0]@xmath223 [ m][][1][0]@xmath224 [ m][][1][0]@xmath195 [ m][][1][0]@xmath196 [ m][][1][0]@xmath197 [ m][][1][0]@xmath225 [ m][][1][0]@xmath200 [ m][][1][0]@xmath226 [ m][][1][0]@xmath227 [ m][][1][0]@xmath228 [ m][][1][0]@xmath229 [ m][][1][0]@xmath230 [ m][][1][0]@xmath187 [ m][][1][0]@xmath231 [ m][][1][0]@xmath232 [ m][][1][0]@xmath233 [ m][][1][0]@xmath234 [ m][][1][0]@xmath235 [ m][][1][0]@xmath236 [ m][][1][0]@xmath237 [ m][][1][0]@xmath238 [ m][][1][0]@xmath239 [ m][][1][0]@xmath236 [ m][][1][0]@xmath240 [ m][][1][0]@xmath241 [ m][][1][0]@xmath198 [ m][][1][0]@xmath226 [ m][][1][0]@xmath200
the application of the perturbation method to the steady - state creep problem was investigated .
high performance of this method in prediction of creep response was validated .
the perturbation method allow one to reduce an initial nonlinear problem to the sequence of simpler ones .
this technique is especially attractive if the unperturbed solution is given in closed form as it was in this paper .
another example is the creep response of the thick - walled homogeneous pipe under plane stress conditions @xcite . such solution could be used for perturbation analysis of creep in open - ended pipes .
the error of the perturbation method becomes substantial when the creep properties differ from one another by one order of magnitude .
nevertheless we note that asymptotic expansion gives a good simplified model of structure response .
this model treats changes in parameter distribution as jumps of displacements in linear elastic material . in that way
, the solution for every complicated case of parameter distribution is represented as a combination of simple solutions .
a.v . shutov is grateful for the support provided by the german academic exchange service .
assume @xmath179 , @xmath175 , @xmath176 .
after integration in , , we have the following values of constants @xmath242 @xmath243 @xmath244 @xmath245 the ode has four linearly independent solutions @xmath246 @xmath247 where @xmath46 is one of solutions of characteristic equation @xmath248 to be definite , we use @xmath249 h. altenbach , v. kushnevsky , k. naumenko . 2001 . on the use of solid- and shell - type finite elements in creep - damage predictions of thin - walled structures , archive of applied mechanics , vol .
71 , 164 - 181 .
browne , b. cane , j.d .
parker and d. walters .
creep failure analysis of butt welded tubes . in proceedings of the conference on creep and fracture of engeneering materials and structures , swansea , 645 - 649 .
hayhurst . 2001a .
computational continuum damage mechanics : its use in the prediction of creep in structures - past , present and future .
iutam symposium on creep in structures , kluwer academic publishers , 175 - 188 .
hayhurst , m.t .
wong , f. vakili - tahami .
2001b . the use of cdm analysis techniques in high temperature creep failure of welded structures , proceedings of creep 7 , japan society of mechanical engineering ( tsukuba , june 3 - 8 , 2001 ) , sa-10 - 6(106 ) 519 - 526 .
hyde , a. yaghi , m. proctor .
use of the reference stress method in estimating the life of pipe bends under creep conditions , international journal of pressure vessels and piping , vol .
75 , 161 - 169 .
nikitenko , v.a .
design calculation of structural elements with vulnerability of the material in creep taken into account , strength of materials ( historical archive ) , vol .
11(4 ) , 354 - 360 . i.j .
perrin , d.r .
continuum damage mechanic analyses of type iv crepp failure in ferritic steel crossweld specimens , international journal of pressure vessels and piping , vol .
76 , 599 - 617 .
roche , c.h.a .
townley , v. regis , h. hubel .
1992 . structural analysis and available knowledge , in : larson lh(ed . ) , high temperature structural design .
mechanical engineering publications , london , 161 - 180 . | the stress analysis of pressurized circumferential pipe weldments under steady state creep is considered .
the creep response of the material is governed by norton s law .
numerical and analytical solutions are obtained by means of perturbation method , the unperturbed solution corresponds to the stress field in a homogeneous pipe .
the correction terms are treated as stresses defined with the help of an auxiliary linear elastic problem .
exact expressions for jumps of hoop and radial stresses at the interface are obtained .
the proposed technique essentially simplifies parametric analysis of multi - material components .
key words : creep , circumferential pipe weldments , stress analysis , parametric analysis , perturbation method _ ams subject classification _ : 74g10 , 74d10 , 74g70 , 74s05 . |
the high spatial resolution of the _ chandra x - ray observatory _ allows us to study the intracluster medium ( icm ) in clusters of galaxies in unprecedented detail .
one interesting result is the interplay between the x - ray emitting gas and non - thermal radio emission .
observations of several galaxy clusters have found x - ray cavities in the icm coincident with bright radio emission ( e.g. , perseus ( a426 ) ; * ? ? ?
* ; * ? ? ?
the central radio sources produce the bubbles of non - thermal plasma responsible for the radio emission .
these bubbles then rise , possibly by buoyancy , displacing the x - ray emitting gas and causing cavities in the latter .
a number of galaxy clusters contain large - scale diffuse radio sources not associated with active galactic nuclei .
these sources are classified into two groups : radio halos which are centrally located , and radio relics .
radio relics are extended diffuse radio sources often located at the periphery of clusters of galaxies .
@xcite present a classification scheme for radio relics based on their observational properties ( but see also * ? ? ?
radio relics typically have steep radio spectra ( @xmath0 assuming @xmath1 ) and are linearly polarized @xcite .
the x - ray surface brightness around most radio relics is low , making it difficult to study the relationship between the x - ray gas and radio emission .
lccc cxou j001316.6@xmath2193201 & 00:13:16.67 & @xmath219:32:01.2 & [email protected] + cxou j001319.0@xmath2193018 & 00:13:19.09 & @xmath219:30:18.5 & [email protected] + cxou j001320.0@xmath2192748 & 00:13:20.06 & @xmath219:27:48.6 & [email protected] + cxou j001321.9@xmath2192750 & 00:13:21.98 & @xmath219:27:50.7 & [email protected] + cxou j001326.0@xmath2193111 & 00:13:26.02 & @xmath219:31:11.4 & [email protected] + cxou j001335.7@xmath2192805 & 00:13:35.74 & @xmath219:28:05.5 & [email protected] + cxou j001339.9@xmath2192935 & 00:13:39.94 & @xmath219:29:35.8 & [email protected] + cxou j001341.1@xmath2192809 & 00:13:41.15 & @xmath219:28:09.5 & [email protected] + cxou j001342.5@xmath2193209 & 00:13:42.58 & @xmath219:32:09.0 & [email protected] + cxou j001344.6@xmath2192914 & 00:13:44.60 & @xmath219:29:14.1 & [email protected] + cxou j001344.7@xmath2193051 & 00:13:44.70 & @xmath219:30:51.0 & [email protected] [ tab : srcprop ] many radio relics , including some of the best known and studied ( e.g. , the coma cluster relic 1253@xmath4275 and abell 3667 ; * ? ? ?
* ; * ? ? ?
* ) , are located far from the cluster center , on the order of mpcs .
however , there exists a subclass of radio relics which are located significantly closer to the cluster center , @xmath550 to 350 kpc . @xcite
classify these as relic sources near the first ranked galaxy .
these radio sources share many of the same properties of classical radio relics including steep spectra , filamentary structure , and polarization @xcite .
abell 133 is one of the class of radio relics near the first ranked galaxy .
observation revealed a tongue - like feature in the x - ray emission comprised of cool gas ( @xmath6 kev ) which extended from the cd galaxy to the radio relic @xcite .
a number of possible origins for the tongue were explored including the suggestion that the cool gas in the tongue was uplifted from the cluster core by the buoyantly rising radio relic .
a followup observation with _ xmm - newton _ is consistent with the uplifted bubble scenario @xcite .
however the detection of a cold front southeast of the cluster core and a weak shock in the core suggests that an unequal mass merger is responsible for the tongue feature . in either model
, the radio plasma appears to have originated from the active galactic nucleus ( agn ) in the central cd , rather than as a direct result of merger shock acceleration . in this paper , we present x - ray data obtained with the _ chandra x - ray observatory _ for the cluster abell 13 , another unusual system with a radio relic near the first ranked galaxy .
abell 13 is an x - ray luminous ( @xmath7 erg s@xmath8 in the 0.510 kev range ) cluster at @xmath9 @xcite .
the relic has an extremely steep ( @xmath10 ) radio spectrum , a linear polarization of @xmath1112% at 1425 mhz , and spans up to 2 @xcite .
no identification of a host galaxy associated with the relic was possible , although the two brightest cluster members are candidates .
interestingly , these two galaxies are separated by 2621 km s@xmath8 in recessional velocity @xcite .
similarly , from redshift data of the eso nearby abell clusters survey ( enacs ) , @xcite concluded that the radial velocity distribution of abell 13 may also be interpreted as a bimodal one , with peaks at @xmath12 ( 16 redshifts , vel.disp .
@xmath13 km s@xmath8 ) and @xmath14 ( 21 redshifts vel.disp .
@xmath15 km s@xmath8 ) , or @xmath16 and 29160 km s@xmath8 .
we assume @xmath17 km s@xmath8 mpc@xmath8 , @xmath18 , and @xmath19 unless otherwise mentioned . at the distance of abell 13
, 1 corresponds to 1.75 kpc .
we observed abell 13 on 2004 august 25 for 55 ks with _
chandra _ using the advanced ccd imaging spectrometer ( acis ; * ? ? ?
the observation was performed so that both the cluster center and the radio relic were located on the back - illuminated s3 ccd .
the `` level 1 '' event file was processed using the ciao v3.2 data analysis package following the standard reduction procedure .
we removed the afterglow correction and created a new observation specific bad pixel file .
the level 1 event file was reprocessed to include the new bad pixel file and utilize the improved background event detection in vfaint mode .
we then filtered the data to remove hot pixels , bad columns , and events with grades 1 , 3 , 5 , and 7 .
data from the back - illuminated s1 ccd were used to determine time intervals with high background using the tool lc_clean .
we filtered the event file on good times , yielding a final exposure time of 50.7 ks .
background data were extracted from the blank sky observations compiled by m. markevitch . our analysis is restricted to data from the s3 ccd . in figure
[ fig : x ] , we show both the raw _ chandra _ x - ray image of the 0.310 kev energy range , and an adaptively smoothed image of the x - ray data produced with the ciao tool csmooth .
the smoothed image has a minimum signal - to - noise ratio of 3 per smoothing beam , and has been corrected for exposure and background . the x - ray emission from abell 13 shows complex structure in the core region .
a knot of x - ray emission is also clearly detected 86 ( 150 kpc ) to the northwest of the cluster s x - ray peak , along with a more diffuse extension of the emission in the same direction .
we identify 11 point sources in our full band image ( 0.310.0 kev ; see table [ tab : srcprop ] ) .
this is consistent with the expected number of background sources from the _ chandra _ deep fields @xcite .
no point source was found at the center of the x - ray emission .
since the cluster emission will reduce our point source detection sensitivity , we also searched for point sources in just the hard band ( 4.010.0 kev ) data . in the hard x - ray band ,
the diffuse emission from the cluster is weaker while the emission from agn is stronger than at softer bands .
therefore , any agns should be more easily detected .
we found no point sources within 05 of the cluster center , defined to be the position of galaxy h , the brightest cluster galaxy . in figure
[ fig : op ] , we show the optical digital sky survey ( dss ) image of the cluster with the x - ray contours overlaid . the x - ray contours are based on the image in figure [ fig : x ] _
right panel_. we have labeled the brightest galaxies following @xcite . the x - ray emission is centered near the bright elliptical h ( also 2masx j00133853@xmath21930007 ) , the brightest cluster galaxy . the knot of emission to the nw is coincident with the elliptical galaxy f ( also 2masx j00133401@xmath21929017 ) , the second brightest cluster galaxy .
the steepness of the x - ray gradient to the north and east is an artifact of the ccd edges .
we are also interested in comparing the x - ray data with the radio emission from abell 13 .
figure [ fig : xr ] shows the x - ray data in red and the radio data in green .
we find no enhancement of the x - ray emission near the radio relic , which is different from what was seen for abell 133 @xcite .
given the low surface brightness of the x - ray emission , it is difficult to determine if there is a deficit of x - ray emission at the position of the radio relic , commonly referred to as an x - ray cavity or radio bubble .
the absolute astrometric accuracy of the radio and x - ray maps is roughly 05 for each .
the radio point source near the cluster center is not associated with galaxy h. instead , it seems coincident with a barely resolved pair of galaxies in the dss and two micron all sky survey ( 2mass ) images , listed in ned as 2masx j00133712@xmath21930067 .
no velocity information is available for 2masx j00133712@xmath21930067 , making association with the cluster uncertain . to enhance the structures in the cluster core , we fitted the observed cluster emission with a smooth elliptical isophotal model , using the iraf / stsdas task ellipse .
we allowed the ellipticity , position angle , and intensities of each isophote to vary while keeping the centroids fixed to the position of galaxy h. the model was then subtracted from the data after multiplication of the model counts by 0.5 to avoid oversubtraction . in figure
[ fig : comp ] , we show the subtracted image of the cluster in the region around the radio relic with the radio contours overlaid . no excess emission is associated with the radio relic .
the cluster emission is faint near the radio relic , and continues to weaken at projected cluster - centric distances greater than that of the relic in the same direction .
more interesting structure is seen in the cluster center .
figure [ fig : compcen ] shows the same image as figure [ fig : comp ] with a different color scale to enhance the variations in the inner region of the cluster .
the knot of emission in the upper right is the second - brightest elliptical galaxy , f. the bright yellow knot of emission is coincident with galaxy h. there is a more diffuse bright extension of emission to the west of galaxy h. in addition , there is other structure in the cluster center perpendicular to the bright extension .
interestingly , the emission extension to the west of galaxy h seems to point in the same direction as one might expect the radio relic to extend if it were projected back to galaxy h. enhanced x - ray emission has been seen in the large elliptical galaxies in the coma cluster @xcite .
we compared the enhanced emission of galaxy f with the results from coma .
first , we determined the radial profile of the enhancement to determine its size and the appropriate source region for galaxy f. the emission was clearly more extended than the _
chandra _ point spread function .
we found that the enhancement measured @xmath11152 in radius , or roughly 3 kpc at the distance of abell 13 .
@xcite found that the compact cores associated with the brightest galaxies in the coma cluster had a physical size of 3 kpc in radius .
we attempted to extract a spectrum of the enhanced emission associated with galaxy f from a circular region with radius of 2 . to account for both the cosmic x - ray background and the cluster emission
, we used a local background extracted from an annulus centered on galaxy f with radii of 3 and 5 .
there are only 31 counts in a 2 region around galaxy f , and we expect 17.5 of these to be background .
this is too few counts for spectral fitting or even to determine a useful color for the source .
even a basic color comparison can not yield any constraint on the temperature of the galaxy gas .
we used pimms to determine the flux from galaxy f. we measured a 0.58.0 kev count rate of @xmath20 cts s@xmath8 and assumed a gas temperature of 1 kev for a raymond - smith plasma model .
the largest elliptical galaxies have x - ray gas temperatures of @xmath111 kev ( e.g. , * ? ? ?
* ) . from pimms
we find a 0.52.0 kev unabsorbed flux of @xmath21 erg @xmath22 s@xmath8 , or a luminosity of @xmath23 erg s@xmath8 .
the galaxies in coma had similar , although slightly larger ( 7.69.1@xmath24 erg s@xmath8 ) luminosities .
we also calculated the mass of gas necessary to produce the excess emission from galaxy f. to do this , we first calculated the model normalization of a 1 kev apec plasma model that would produce the flux measured from galaxy f. we then calculated the average electron number density in the gas from the model normalization and assuming a spherical gas distribution with 2 radius .
translating the electron number density into a gas mass , we find that the required gas mass is @xmath25 m@xmath26 .
this is comparable to the gas masses found in the coma galaxies and other galaxies in nearby clusters @xcite .
given the similarity in size and luminosity , we may expect that the cause of the x - ray enhancement in galaxy f is the same as that found for the coma galaxies @xcite .
the gas in the coma galaxies is of low temperature ( 12 kev ) and is likely the remains of the galactic x - ray halos .
@xcite suggest that the cool gas is in pressure equilibrium with the hotter , intercluster medium .
spectra were extracted from the level 2 event file , excluding point sources within the regions of interest .
background spectra were extracted from the blank - sky background file . for each spectrum ,
weighted response files ( rmfs and arfs ) were generated using the ciao tools mkacisrmf and mkwarf , respectively .
finally , spectra were grouped to obtain a minimum of 25 counts per bin .
spectral fitting was primarily performed in xspec v11.0 , except for the temperature map spectral fits ( see [ sec : tmap ] below ) .
we first determined the integrated spectrum of the cluster using an elliptical region with semi - major and semi - minor axes of 25 and 19 , respectively , and position angle of 121@xmath27 measured from north to east ( see figure [ fig : specshape ] ) .
this region was the largest elliptical region whose shape agreed with the outer cluster x - ray isophotes and fit entirely on the s3 ccd .
we fit the data with an absorbed plasma model , using the apec model in xspec .
we fixed the redshift of the cluster to the mean redshift of abell 13 found in the enacs study ( @xmath9 ; * ? ? ? * ) and the equivalent hydrogen column density to the average galactic value in the direction of abell 13 , @xmath28 @xmath22 @xcite .
the best - fit model had a temperature @xmath29 kev and a metal abundance of @xmath30 solar ( see table [ tab : spec ] and figure [ fig : clusspec ] ) .
we also performed a spectral study to determine if a cooling flow is present in the center of abell 13 , i.e. the peak in the x - ray emission which is also coincident with galaxy h. we extracted spectra in circular annuli out to 100 from the cluster center .
each spectrum was required to have at least 1000 counts .
the width of the annuli ranged from 15 in the cluster center to 5 at the periphery of the cluster .
we fitted each spectrum with the same apec model used to describe the total cluster emission , allowing only the temperature and normalization to vary ( the metallicity was fixed at the best fit value found above ) .
no statistically significant variations in temperature were found .
this result is consistent with the temperature map determined in [ sec : tmap ] . in order to determine the spectral distribution in the cluster emission
, we created an adaptively binned temperature map of abell 13 , shown in figure [ fig : tmap ] .
the temperature map was produced using isis @xcite . a spectrum at each map pixel ( 1 map pixel @xmath31 16.3 acis pixels @xmath31 80 ) was extracted from the event data ( excluding point sources ) in a square region centered on the pixel .
the region size was determined by requiring at least 800 net ( i.e. background subtracted ) counts in the 0.47.0 kev energy range , but with a maximum box size of 183@xmath32183 acis pixels , or 90@xmath3290 .
the maximum box size was only reached in the outskirts of the cluster .
the total counts in each region were between 1.5 and 2 times the 800 count minimum to account for background subtraction . for each spectrum ,
a background spectrum was extracted from the blank sky background event file in the same region .
rmf and arf files were constructed for specific locations on the ccd , with a spacing of 32 and 50 pixels , respectively .
this was done to account for variations in the response across the ccd , without having to create individual rmf and arf files for each region which would be computationally demanding . for a given region ,
the rmf and arf used to fit the spectrum were chosen by selecting the response files with positions closest to the count - weighted average position of the events in each region .
the spectra were grouped to have 25 counts per bin and were fitted in isis using the apec thermal model .
the galactic hydrogen column density , redshift , and abundance were fixed to the values found for the full cluster spectrum ( [ sec : full ] above ) . only the temperature and normalization of the apec model were allowed to vary . in general , the errors on the temperature map are about 20% in the temperature .
although , for pixels with higher best - fit temperatures or those near the edge of the map , the errors are larger , in the range of 5070% .
this is because of the low high - energy sensitivity of _ chandra _ and the lower statistics of the outer regions .
the temperature map shown in figure [ fig : tmap ] is colored such that the coolest regions ( @xmath33 kev ) are red , and the hottest regions ( @xmath34 kev ) purple .
the intensity contours for the x - ray and radio emission from abell 13 are overlaid .
we note that small differences in temperature from pixel to pixel are not statistically meaningful given the errors in the fits ( an average error of 1.2 kev at 6.0 kev ) .
since the fitted region size is larger than a single map pixel , there is overlap of the fitted regions . of particular interest
are the lack of very cool gas in the cluster center and the presence of cool gas associated with the radio relic region .
the temperature map shows a large concentration of lower temperature gas associated with the radio relic .
below ( [ sec : rad ] ) we perform a detailed spectral study of this gas to better constrain its properties .
there may also be some hotter gas ( purple / blue region ) associated with the steep gradient in the x - ray surface brightness contours on the southern edge of the cluster ( see e.g. , the x - ray contours in figure [ fig : tmap ] ) but the temperature errors are large in this region . to try and confirm this result , we extracted spectra on the northern and southern sides of the x - ray gradient .
no significant difference in the best - fit temperatures was found , although both regions had large errors .
higher count spectra are required to verify the increased gas temperature along the cluster s southern edge .
from the temperature map , we see that the x - ray emission associated with the radio relic is colder than the average cluster temperature
. this immediately rules out a shock origin as we would expect the emission to be hotter in that case . to determine the nature of this cooler emission
, we performed a more rigorous x - ray spectral study of the radio relic region .
we fit the spectra from a 20@xmath3214 ( major / minor axis ) elliptical region surrounding the radio relic ( figure [ fig : specshape ] ) .
this region was chosen to encompass the outer portion of the radio relic that is associated with the cooler gas in the temperature map .
this region has the advantage that it is a region of lower surface brightness for the hot gas component .
if we add in the radio relic region to the east ( the tail leading back to galaxy h ) , the spectrum has a much larger hot gas component and is less sensitive to a cool gas component .
lc @xmath35 ( @xmath22 ) & @xmath36 + @xmath37 ( kev ) & [email protected] + abundance ( solar ) & [email protected] + norm & ( [email protected])@xmath38 + @xmath39/dof & 316.45/297 + [ tab : spec ] the data were fit with different spectral models , including a single temperature gas model , a two - temperature gas model , and a gas @xmath4 powerlaw emission model .
the results are given in table [ tab : rrspec ] .
in all fits we have frozen the metallicity of the thermal components to that of the general cluster emission .
if allowed to vary , the metallicity is unconstrained . the single temperature gas model , while a decent fit to the data , does not reproduce the excess emission in the data near 0.8 kev ( see figure [ fig : rrspec ] ) .
the two - temperature gas model is the best fit to the data .
one of the components is consistent with the overall cluster temperature , while the other component has a temperature of @xmath110.9 kev and accounts for the soft excess seen in the data .
the gas @xmath4 powerlaw model is a worse fit , with the powerlaw component replacing the hotter gas component .
the two - temperature gas model is the most appropriate , as it would include both a lower temperature gas associated with the relic as well as hotter cluster emission seen in projection .
we have presented the _ chandra _ data of the galaxy cluster abell 13 .
the x - ray emission is centered on galaxy h of @xcite , the brightest cluster galaxy .
the spatial structure near the core is complicated , with a bright tail of excess emission to the west of galaxy h , and a fainter excess to the north .
no point source is associated with galaxy h and no cooling flow is found .
we also found excess x - ray emission associated with the second - brightest elliptical galaxy , f , in the cluster .
the mass of gas associated with galaxy f is consistent with other bright elliptical galaxies in cluster environments @xcite . in general
, there is an extension of the x - ray emission to the northwest with respect to the x - ray peak coincident with galaxy h. finally , no excess or significant deficit in emission is found coincident with the unusual radio relic in abell 13 , although the spectral properties of the gas in this region are different from those in other areas of the cluster .
we found a soft excess in the spectrum of gas coincident with the radio relic , which can be well fit by a @xmath110.9 kev thermal model , suggesting the presence of cooler gas there .
the redshift distribution of the galaxies in abell 13 can be described by a bimodal distribution @xcite .
the subclumps are not distinguishable in their location on the sky .
interestingly , the two brightest galaxies , h and f , are associated with different peaks in the redshift distribution and rather than being associated with the peak velocity , either of the cluster as a whole or of either subclump , the velocities of galaxies h and f are on the very edges of the velocity distribution .
galaxy f is on the lower edge for the low - velocity subcluster and galaxy h on the higher edge of the high - velocity subcluster .
this , combined with the complicated structure of the x - ray emission , points to an ongoing merger within the cluster and that galaxies h and f have a significant velocity relative to the general cluster distribution .
the spatial distribution of the galaxies in abell 13 shows a northwest - southeast extension , paralleling the x - ray distribution , but extending to larger scales . in the following discussion
, we assume that the x - ray gas associated with the radio relic has two components , a hot ( 6 kev ) gas component due to the general cluster emission , and a cooler ( 0.9 kev ) gas component local to the relic .
this model is based on our best - fit spectrum found in [ sec : rad ] .
we assume that the two components are in pressure equilibrium . from @xcite ,
the minimum pressure of the radio relic in abell 13 is @xmath40 dyne @xmath22 .
their estimate of the thermal pressure of the hot x - ray gas was very rough and based solely on scaling the properties of abell 85 since no imaging observation of abell 13 was available .
our _ chandra _ data allow for a more accurate estimate of the thermal pressure at the location of the relic . however , due to an unfortunate placement of the cluster on the ccd , the only regions which extended out to a distance larger than the relic were in the direction of the relic and to the south . since the southern edge of the cluster shows a steep gradient in the x - ray emission , we did not consider this region in determining the surface brightness distribution .
instead , we determined the surface brightness of the cluster in a conical region extending from the cluster center out to 41 and encompassing position angles of 240312@xmath27 measured from the north to the east .
this region includes the radio relic and associated cool gas which will affect our results ; however , while not a perfect measure of the properties of the x - ray gas , it represents a significant improvement on the estimates of @xcite .
a larger field of view observation would be better suited to address the surface brightness and electron density distributions of the full cluster emission .
lccccc & 5.9@xmath41 & ( [email protected])@xmath42 & & & 89.4/69 + apec @xmath4 apec & 0.86@xmath43 & ( [email protected])@xmath44 & 10@xmath45 & ( [email protected])@xmath42 & 74.3/67 + apec @xmath4 pl & 0.89@xmath43 & ( [email protected])@xmath44 & [email protected] & ( [email protected])@xmath44 & 81.4/67 + [ tab : rrspec ] using the surface brightness distribution , we estimate the electron density distribution of the hot component following the spherical deprojection method of @xcite .
we assumed a single temperature of the hot gas of 6.0 kev and used the apec thermal model to convert from counts to electron density . from this , we find an electron density at the relic of @xmath46 @xmath47 .
given this and the average cluster temperature ( 6.0 kev ) determined in
[ sec : full ] , we find that the thermal pressure of the surrounding hot gas at the location of the radio relic is @xmath48 dyne @xmath22 . this value is larger than the estimate for the pressure of the radio plasma , as has been seen in other x - ray cluster radio bubbles @xcite .
we note that if the relic were at a larger radius than the projected radius , the electron density and thermal pressure would be lower , bringing it closer to the radio plasma pressure estimate .
in addition , in the case of abell 13 , if the values assumed by @xcite for ( 1 ) the ratio of masses in heavy particles and electrons ; ( 2 ) the plasma filling factor ; ( 3 ) the low frequency spectral index , and ( 4 ) the angle between the line of sight and field direction are given more realistic values then we may increase the radio plasma pressure to that of the confining x - ray gas .
it is unlikely that the radio relic in abell 13 is a peripheral cluster relic associated with a cluster merger shock .
first , there is no evidence in the _ chandra _ image or spectra for any shock feature near the relic .
second , the radio image shows a long filament which extends almost all the way back to the brightest elliptical galaxy h ( figure [ fig : xr ] ) .
thus , we assume that the radio emitting plasma originated in an agn at the center of galaxy h. based on the classification system of @xcite , which looks at the physical properties and proposed origins of extended cluster radio sources , we would classify abell 13 as an agn relic system .
we propose two explanations for the multi - wavelength properties of abell 13 .
the first is the buoyant bubble scenario suggested by @xcite for abell 133 .
the radio source may have uplifted the cool gas from the core of abell 13 . in this scenario ,
galaxy h would have been an active radio emitter in the past . at some later time , the radio plasma would have risen by buoyancy , and entrained the cool gas in the center , uplifting it to a larger distance from the cluster center .
this might also remove the source of fuel for the central agn .
this would explain the lack of cool gas in the cluster center and the fact that currently , galaxy h is neither a radio ( @xmath49 mjy at 1.4 ghz ; * ? ? ?
* ) nor a hard x - ray source .
to estimate the time scale for the cool gas and radio source to rise by buoyancy from the position of galaxy h , we assume that the cool gas and radio source move at a velocity roughly 1/3 the sound speed , @xmath50 , of the hot gas component . for a 6.0 kev gas , @xmath50 is 980 km s@xmath8 .
the projected separation between the relic and galaxy h is 2231 .
this yields a buoyancy time scale of 7@xmath51 yr , roughly consistent with the relic age determined through spectral fitting @xcite . to estimate the mass of cool gas uplifted , we take the volume to be that of a prolate spheroid with major and minor axes equal to that of the region used to extract the spectrum of the radio relic . from the two - temperature spectral model ,
we find that the enclosed mass of cool gas is @xmath52 m@xmath26 .
this is a factor of 10 greater than the mass of gas contained in the tongue of x - ray gas found in abell 133 @xcite .
however , we note that the electron density for the cool gas is @xmath53
@xmath47 , lower than the hotter cluster gas ( particularly the gas towards the center of the cluster ) making buoyant uplift a possibility .
a cool gas mass of @xmath52 m@xmath26 implies that previously abell 13 would have had a mass deposition rate of @xmath54 m@xmath26 yr@xmath8 assuming a cooling time of @xmath55 yr .
this is on the low end of the distribution of known cooling flow mass deposition rates ( see e.g. , * ? ? ?
this is only an estimate of the mass of cool gas to determine if the value is realistic .
the temperature map shows additional cool regions to the north and east of the radio relic which if taken into account would increase the mass of cool gas .
however , the radio relic is filamentary in structure and it is possible that the cool gas also resides in filaments rather than a single big blob , which would decrease the gas mass estimate . since it is difficult to determine the extent of either of these situations , we merely suggest that the estimated amount of cool gas in abell 13 seems reasonable for a galaxy cluster
. the lower temperature of the gas near the radio source might be due both to the lower entropy of the gas in the cool core of the cluster , plus the effects of adiabatic expansion as this gas was moved out to a lower - pressure region . from our above estimate of the electron density distribution of the cluster gas
, we find that the electron density at the cluster center is @xmath116 times that at the radio relic .
the pressure difference between the two locations should also be 6 times assuming a constant hot gas temperature .
we assume that the cool gas component is always in pressure equilibrium with the hot gas component that dominates the cluster emission . as a result of adiabatic expansion , the temperature of the cool gas in its current location , @xmath56 is related to its initial temperature , @xmath57 , by @xmath58 where @xmath59 is the pressure in the cluster core and @xmath60 is the pressure at the location of the radio relic , and the adiabatic index is @xmath61 .
observationally , we found @xmath62 , giving @xmath63 for adiabatic expansion .
this difference alone can not account for the cooling of the gas from the average cluster temperature of @xmath116 kev to its current temperature of @xmath64 kev .
therefore , the cool gas associated with the radio relic must have been cooler than the average cluster temperature to begin with .
we can estimate the initial cool gas temperature assuming that adiabatic expansion was the only heating or cooling mechanism as the cool gas was uplifted .
this requires that the initial temperature of the cool gas was about @xmath65 kev when this gas was located at the cluster core .
this is a reasonable temperature for a cooling core .
if the relic were at a larger radius than the projected radius , then we would expect the electron density to be lower and therefore the adiabatic cooling to be stronger .
this would make the initial temperature of the cool gas higher , possibly more in line with the average cluster temperature .
our second proposed explanation is that a dynamical event is the origin of the unusual radio relic and disturbed x - ray morphology .
both the x - ray and optical properties ( radial velocity distribution ) suggest that abell 13 is undergoing a merger event . the x - ray extension from galaxy h ( to the west )
points back to the radio relic .
this may imply that galaxy h is moving to the east and that the radio relic is at the former position of the galaxy .
( we note that the extension may also favor the buoyant uplift scenario . )
we propose that before the merger event , galaxy h had a cool core and showed agn activity .
as the merger progressed , the cold core and radio lobes were displaced ( or stripped ) from the galaxy as a result of ram pressure from the gas in the other subcluster , which did not affect the stars and dark matter in galaxy h. @xcite found that the travel time from galaxy h was comparable to the estimated relic age .
the long filament of radio emission leading back almost to galaxy h ( figure [ fig : xr ] ) suggests that the agn remained active during the initial phases of this stripping process .
the concentration of x - ray emission around galaxy h ( figure [ fig : compcen ] ) suggests that the stripping process is not complete ; there is still gas around galaxy h , but it is relatively hot , perhaps due to heating processes associated with the merger .
the _ chandra _
data do not conclusively determine which of these two models is correct , or strongly reject either .
, j. c. , blanton , e. l. , clarke , t. e. , enlin , t. a. , johnston - hollitt , m. , & rudnick , l. 2004 , in the riddle of cooling flows in galaxies and clusters of galaxies , ed . t. reiprich , j. kempner , & n. soker ( charlottesville : univ .
virginia ) , 335 | we present results from the _ chandra _ x - ray observation of abell 13 , a galaxy cluster that contains an unusual noncentral radio source , also known as a radio relic .
this is the first pointed x - ray observation of abell 13 , providing a more sensitive study of the properties of the x - ray gas .
the x - ray emission from abell 13 is extended to the northwest of the x - ray peak and shows substructure indicative of a recent merger event .
the cluster x - ray emission is centered on the bright galaxy h of @xcite .
we find no evidence for a cooling flow in the cluster .
a knot of excess x - ray emission is coincident with the other bright elliptical galaxy f. this knot of emission has properties similar to the enhanced emission associated with the large galaxies in the coma cluster . with these _ chandra _ data we are able to compare the properties of the hot x - ray gas with those of the radio relic from vla data , to study the interaction of the x - ray gas with the radio emitting electrons .
our results suggest that the radio relic is associated with cooler gas in the cluster .
we suggest two explanations for the coincidence of the cooler gas and radio source .
first , the gas may have been uplifted by the radio relic from the cluster core .
alternatively , the relic and cool gas may have been displaced from the central galaxy during the cluster merger event . |
massive early - type galaxies ( etgs ) reveal halos of hot gas .
this x - ray emitting diffuse gas varies in luminosity by a factor of @xmath3100 - 1000 for a given optical luminosity , or equivalently , stellar mass ( osullivan , forbes & ponman 2001 ; fabbiano 2006 ; boroson , kim & fabbiano 2011 ) .
it has been suggested that stellar mass loss dominates the source of this gas ( sun et al .
2007 ; sarzi et al . 2013 ) . however , in a large x - ray study of etgs , goulding et al .
( 2016 ) concluded that the data are not consistent with a simple stellar mass loss picture .
an additional source of gas is indicated by the cold dark matter ( cdm ) paradigm for galaxy formation . here
infalling pristine gas in massive halos is shock - heated to x - ray emitting temperatures as these halos collapse .
the hot gas slowly cools while emitting x - rays , with the x - ray luminosity directly related to the depth of the potential well ( white & frenk 1991 ) . in the gimic simulations of crain
et al . ( 2010 ) , the density of the hot gas is less concentrated than that of the canonical dark matter density profile .
this leads to much lower x - ray luminosities than predicted by white & frenk ( 1991 ) , which is then closer to the x - ray luminosity as observed from late - type galaxy halos .
however , for early - type galaxies , agn heating of hot gas is also potentially important .
recently , choi et al . (
2015 ) included agn in cosmological models of massive etgs .
they found better agreement with the x - ray luminosities of etgs than when using models without agn feedback , although some discrepancies remained .
these cosmological simulations all suggest that the key parameter determining the luminosity of the diffuse x - ray gas ( l@xmath0 ) is the galaxy mass ( which determines the depth of the potential well ) with agn perhaps playing a secondary role in etgs . combining new _
chandra _ data with results from the literature , kim & fabbiano ( 2013 ; kf13 ) have shown that the dynamical ( i.e. baryonic plus dark matter ) mass may indeed be the key parameter in determining l@xmath0 surrounding an etg ( see also mathews et al .
in particular , they showed that measurements of l@xmath0 correlated strongly with the dynamical mass within 5 r@xmath1 ( where r@xmath1 is the effective radius of the galaxy optical light ) for a sample of 14 early - type galaxies .
the total mass measurements within 5 r@xmath1 came from the work of deason et al . ( 2012 ) who used a combination of planetary nebulae ( pne ) and globular cluster ( gc ) kinematics from the literature , and the assumption of power - law mass and tracer density profiles . here
, we use a homogeneous analysis of dynamical masses from gc kinematics within 5 r@xmath1 for 29 etgs ( alabi et al .
this analysis is based on data from the sluggs survey ( brodie et al .
2014 ) and supplemented by data from the literature .
x - ray gas luminosities mostly come from the compilation of kim & fabbiano ( 2015 ; hereafter kf15 ) and are extracted from radii which vary from 30 to 240 arcsec ( @xmath3 2 - 5 r@xmath1 . ) we investigate the trends of the x - ray gas luminosity for our sample with stellar mass , dynamical mass within 5 r@xmath1 and with extrapolated virial mass .
we compare these new observations with simulated galaxy halos from crain et al .
( 2010 ) and choi et al .
( 2015 ) , and discuss whether the hot gas halos around etgs are consistent with an origin , and heating mechanisms , as described by the @xmath2cdm paradigm .
globular cluster kinematics for galaxies from the sluggs survey are supplemented by literature data for ngc 1316 , 1399 , 4472 and 4636 to give a total sample of 29 galaxies , to derive dynamical masses ( alabi et al . 2017 ) .
the sluggs ( brodie et al . 2014 ) survey targets nearby etgs with stellar masses around 10@xmath4 m@xmath5 .
the survey has accumulated over 4000 high precision gc velocities ( see e.g. pota et al .
2013 ) .
dynamical masses for the sample galaxies have been derived from the gc kinematics by alabi et al .
they used the tracer mass estimator of watkins et al .
( 2010 ) with the assumption of isotropic orbits for the gcs ( mildly radial or tangential orbits give masses within @xmath310% of the isotropic case ) .
small corrections have been made for flattening , rotation or substructures within the gc systems . in fig .
1 we show a comparison of the total dynamical mass within 5r@xmath1 used in this work with that used by kf13 , i.e. from deason et al .
although there is a reasonable match between the two studies , fig.1 shows that alabi et al .
( 2017 ) find systematically higher masses for low mass galaxies than deason et al .
the masses for most of these low mass galaxies were derived using pne , which alabi et al .
( 2016 ) showed may underestimate the true galaxy mass compared to other tracers of mass ( i.e. gcs and x - rays ) .
here we use the alabi et al .
( 2017 ) masses which are derived only from gcs . assuming an nfw - like halo , alabi et al .
( 2017 ) also estimated the total virial mass for each galaxy by extrapolating from the 5r@xmath1 mass .
the diffuse hot gas x - ray luminosities in the 0.3 - 8 kev band are taken from the work of kf15 .
they have carefully removed the contribution from discrete sources such as low mass x - ray binaries ( fabbiano 2006 ) to the total x - ray luminosity , leaving the diffuse gas contribution l@xmath0 . a correction to bolometric would increase the x - ray luminosities by 0.08 dex on average .
most of the x - ray data come from _ chandra _ observations .
however , for some high mass galaxies the x - ray emission is particularly extended ( e.g. ngc 4374 , 4486 , 4649 , 5846 ) and in those cases _ rosat _ data from osullivan , forbes & ponman ( 2001 ) , corrected to the _ chandra _ energy band , are used .
although the contribution from discrete sources in the rosat data can not be subtracted as accurately as it can for chandra data , their contribution is only about 1% of the diffuse gas luminosity for these high mass galaxies ( see osullivan , forbes & ponman 2001 ) . for further details
here we make a very small correction to the kf15 l@xmath0 luminosities for the distances used in the sluggs survey ( brodie et al .
the kf15 compilation did not include several galaxies that appear in the alabi et al .
( 2017 ) study .
here we also include the x - ray luminosities for ngc 720 , ngc 1316 and ngc 3115 from boroson et al .
( 2011 ) , and for ngc 5128 from kf13 .
( 2014 ) conducted a detailed xmm and _ chandra _ study of ngc 1400 . as well as
the x - ray emission centred on ngc 1400 , they detected an enhanced region of x - rays to the ne of the galaxy that they associated with stripped gas .
here we use the x - ray luminosity centred on ngc 1400 with a small adjustment to our x - ray band and distance , and assume an uncertainty of 20% .
we note that the x - ray luminosity would double if the enhanced region were also included .
two galaxies in alabi et al .
( 2017 ) but not included here are ngc 2974 ( not observed by _
chandra _ ) and ngc 4474 ( the _ chandra _ observation was only 5 ksec ) . in table 1
we list the properties of our sample galaxies .
these include the distance - adjusted l@xmath0 , the presence of a core or a cusp in the central optical light profile , the local environment of the galaxy , the stellar mass calculated from the @xmath6-band luminosity after the correction described by scott et al .
( 2013 ) and assuming m / l@xmath7 = 1 , the dynamical mass within 5 r@xmath1 and an estimate of the total virial mass . [ cols="<,<,<,<,<,<,<,^,^ " , ]
notes : columns are ( 1 ) galaxy name , ( 2 ) distance , ( 3 ) effective radius , ( 4 ) 1=core , 2=intermediate , 3=cusp central light profile from krajnovic et al .
( 2013 ) and lauer et al .
( 2007 ) , ( 5 ) environment : f = field , g = group , c = cluster , @xmath8 = central dominant galaxy , ( 6 ) log x - ray gas luminosity in the 0.3 - 8 kev band from kf15 ( except @xmath9 = boroson et al .
( 2011 ) , @xmath10 = kf13 , @xmath11 = su et al .
( 2014 ) , ( 7 ) log total stellar mass from the @xmath6-band luminosity ( an uncertainty of @xmath120.2 dex allows for reasonable variations in imf , age and metallicity ) , ( 8) log dynamical mass within 5 r@xmath1 and uncertainty , ( 9 ) estimated total virial mass and uncertainty ( data for columns 7 - 9 come from alabi et al .
in fig . 2 we show the x - ray luminosity against stellar mass , dynamical mass within 5 r@xmath1 ( i.e. on a scale that is comparable to the x - ray extraction radii used by kf15 and one in which our sample galaxies are dominated by dark matter ) and virial mass .
these measurements and their uncertainties are given in table 1 .
we also show in fig .
2 model predictions from the cosmological simulations of crain et al .
( 2010 ) and choi et al .
crain et al .
( 2010 ) used the gimic hydro simulations to predict the x - ray gas properties of 458 isolated late - type galaxy halos .
a key finding of their work is that the gas is in quasi - hydrostatic equilibrium and that the large scatter in the optical and ir luminosity
l@xmath13 relation is largely due to the variation of stellar mass at a given halo mass .
in contrast , the halo virial mass l@xmath13 relation is much tighter , since it directly traces the depth of the potential well .
we note that the observations are in the 0.3 - 8 kev energy range , while the crain et al .
simulations predict the 0.5 - 2 kev range .
we expect a small ( @xmath140.1 dex ) increase in their predicted x - ray luminosities to account for the wider band in the observations .
crain et al .
did not include agn in their simulations .
agn , and the x - ray cavities that they form , are now known to have an important effect on the hot gas halos around early - type galaxies ( mcnamara & nulsen 2007 ; diehl et al .
if an agn imparts sufficient mechanical energy to the halo gas it can result in an outflow which reduces the overall gas density ( @xmath15 ) leading to a significant reduction in the x - ray luminosity ( as l@xmath16 ) .
the choi et al .
( 2015 ) simulations incorporate agn mechanical heating from a wind for a subsample of 20 etgs from the cosmological zoom simulations of oser et al .
compared to their simulations without agn , and agn with pure thermal heating , agn with mechanical heating produce x - ray luminosities that are 100 - 1000@xmath17 lower for a given galaxy .
the simulated galaxies are well matched in stellar mass to that of the sluggs sample , and they predict l@xmath0 in the same x - ray energy band as the observed data . in both the choi et al . and
crain et al .
cosmological simulations the hot gas is initially heated to the virial temperature by shocks as it collapses in the forming dark matter halos . in the case of the choi et al .
simulation , feedback from agn provides additional heating ( which reduces the overall x - ray luminosity since the gas becomes more diffuse due to an outflow ) .
the left panel of fig .
2 shows the total stellar mass vs x - ray luminosity ( we note that using the stellar mass within 5r@xmath1 would give similar results as it is @xmath390% , or 0.05 dex , of the total stellar mass for a typical galaxy ) .
the model predictions of crain et al .
( 2010 ) for late - type galaxies show a large range in x - ray luminosity for a given stellar mass .
the observations for our etgs reveal a factor of 100 - 1000@xmath17 range in l@xmath0 at a given stellar mass and are generally located on the right hand edge of the crain et al . distribution . on the other hand ,
the observations are well matched by the predictions from choi et al .
( 2015 ) , albeit over the limited mass range of their simulations .
we suggest that crain et al . overpredicted the x - ray luminosities appropriate for etgs at a given stellar mass due to the absence of agn in their model .
we return to this point below . in the middle panel
we show the dynamical mass within 5 r@xmath1 vs x - ray luminosity . over this radial range ,
the galaxies are dark matter dominated ( alabi et al .
a similar trend was seen by kf13 from the x - ray emission of 14 etgs . to determine the best fit we have applied a bisector linear regression with errors determined from bootstrap resampling .
this fit is shown in the middle panel of fig .
this fit is consistent with that found by kf13 ( who used masses from deason et al .
2012 ) , but shifted to slightly higher masses for low mass galaxies as indicated in fig .
1 . the best fit has a scatter of 0.5 dex about the relation : + log l@xmath0 = 3.13(@xmath120.32 ) log mass ( @xmath145r@xmath1 ) + 3.64(@xmath123.67 ) we also show the simulated galaxies of choi et al .
( 2015 ) in the middle panel , and although they only cover a limited range they are well matched to the observations ( crain et al .
did not predict the mass within 5 r@xmath1 ) .
the choi et al .
simulations support the interpretation that the data are well represented by a single linear relation .
we note that choi et al . compared their predictions with the best fit relation of kf13 ( which used masses from deason et al .
2013 ) in their figure 4 and showed that their l@xmath0 values are systematically too low for a given mass .
( when using velocity dispersion as a mass proxy the agreement was very good . with our larger data set and the use of ( systematically higher ) masses from alabi et al .
( 2017 ) their predictions now agree very well with the observations of l@xmath0 and mass within 5 r@xmath1 . in the right panel of fig .
2 we show the total virial mass vs x - ray luminosity . the virial mass of the observed etgs has been estimated from the dynamical mass measured at 5 r@xmath1 and extrapolated to the virial radius assuming a navarro , frenk & white ( 1997 ) dark matter profile ( see alabi et al .
2017 for details ) . because the ratio of virial mass to dynamical mass within 5 r@xmath1 varies ( due to a non - linear stellar mass to halo mass relation )
, we expect l@xmath0 to reveal a non - linear relation with virial mass .
a non - linear relation can be seen in the observations shown in the right panel of fig .
we also show the 20 model galaxies from choi et al .
( 2015 ) and the crain et al .
( 2010 ) predicted trend , and approximate scatter , for their 458 model galaxies .
crain et al .
showed that the scatter in the x - ray luminosity relation using the virial mass compared to the stellar mass is much reduced , and argued that the virial mass is the primary driver of the x - ray luminosity from galaxy halos . compared to our sample of etgs , the crain et al .
model predictions appear to provide an upper envelope ; their l@xmath0 values need to be reduced by at least an order of magnitude to match the observations of etgs .
choi et al .
compared their predicted l@xmath0 values with virial masses for galaxies and groups from mathews et al .
once again , the choi et al . ( 2015 ) simulations predicted l@xmath0 values that are too low at a given mass ( their figure 4 ) .
however , with our updated measurements their predictions now reveal a good match to the observations of l@xmath0 and virial mass .
we note that several of the galaxies in our sample show ` hot cores ' ( i.e. central gas temperatures that are hotter than the surrounding gas ) when spatially resolved temperture profiles are available , e.g. ngc 1316 ( kim & fabbiano 2013 ) , ngc 3115 ( wong et al .
2011 ) , ngc 4278 ( pellegrini et al .
2012 ) and ngc 4649 ( paggi et al . 2014 ) .
these hot cores are thought to be due to recent heating by the agn and/or associated radio jets ( e.g. paggi et al .
the influence of agn may extend down to our lowest stellar masses of 10@xmath18 m@xmath5 ( cheung et al .
2016 ) .
kf13 found that the most x - ray luminous ( log l@xmath0 @xmath19 40 ) galaxies ( i.e. 7 galaxies after excluding ngc 4374 which deviated significantly from their trend ) revealed significantly less scatter than the 6 low x - ray luminosity galaxies . with our sample of 29 galaxies , and a new 5 r@xmath1 dynamical mass for ngc 4374 , we find it lies well within the distribution of the x - ray luminous galaxies .
we note that the @xmath33 gyr old merger remnant galaxy ngc 1316 ( fornax a ) also lies within the scatter of the general trend .
using a sample of 61 galaxies , kf15 were able to explore the l@xmath0t@xmath13 relation subdivided by the shape of their central optical profile , i.e. core or cusp like .
they found that core galaxies , which tend to be x - ray luminous , have a very tight relation .
in contrast , low x - ray luminosity cusp galaxies revealed no clear correlation between l@xmath0 and t@xmath13 . in fig .
2 we have colour - coded our sample by their central optical profile ( see table 1 ) .
the galaxies are classified as core , cusp , intermediate or unknown surface brightness profile from krajnovic et al .
( 2013 ) or lauer et al .
we note that there is some disagreement in the literature about the central profile in some galaxies , e.g. dullo & graham ( 2013 ) suggested that ngc 4473 does not contain a central core but rather a disk . in all three panels
, the sample divides clearly between core ( high x - ray luminosity ) and cusp ( low x - ray luminosity ) galaxies , with 5 additional galaxies ( 3 unknown and 2 intermediate profiles ) .
the sample divides more clearly in terms of l@xmath0 than mass ( see also pellegrini 2005 ) .
we note that the central profile shape correlates very strongly with central rotation , i.e. core galaxies @xmath20 slow rotators , and cusp galaxies @xmath20 fast rotators ( krajnovic et al . 2013 ) .
focusing on the middle panel , the 17 core galaxies reveal a somewhat tighter relation than the 9 cuspy galaxies .
we find a p - value for the spearman rank correlation test of 0.0007 for the core galaxies , whereas the cusp galaxies have a value of 0.26 indicating almost no relation .
a bisector fit to only the cuspy galaxies also suggests no statistically significant relation .
this supports the findings of kf15 and suggests that sn heating may play an increased role in cuspy ( low mass ) etgs ( david et al .
so although the x - ray luminosity of both core ( high x - ray luminosity ) and cuspy ( low x - ray luminosity ) etgs are primarily determined by total mass , and hence the depth of the potential well , other factors may play a role for the lower mass galaxies .
using new dynamical masses we find a strong linear relationship between the diffuse gas x - ray luminosity for a sample of 29 nearby massive early - type galaxies and their total dynamical mass within 5 r@xmath1 .
this result supports the earlier analysis by kf13 based on 14 galaxies .
we also show that the cosmological simulations of choi et al .
( 2015 ) , which incorporate agn mechanical heating , now agree with observations of the x - ray luminosity and mass within 5 r@xmath1 .
this good agreement with model galaxies in a cosmological framework supports the idea that the diffuse x - ray luminosity is primarily driven by the depth of a galaxy s potential well , with a significant contribution from agn mechanical heating .
supernova heating may contribute in lower mass galaxies .
we thank r. crain for useful discussion and e. choi for her model data .
we also thank the referee for useful feedback ( of the non - agn kind ) .
daf thanks the arc for financial support via dp130100388 .
this work was supported by nsf grant ast-1211995 and nasa contract nas8 - 03060 ( cxc ) .
gf thanks the aspen center for physics , funded by nsf grant 1066293 , for their hospitality while this paper was completed .
dwk acknowledges support by the chandra go grant ar4 - 15005x and by 2014 smithsonian competitive grant program for science .
alabi a. , et al .
2016 , mnras , in press + alabi a. , et al .
2017 , in prep + boroson b. , kim d .- w . , fabbiano g. , 2011 , apj , 729 , 12 + brodie j. p. , et al . , 2014 ,
apj , 796 , 52 + cheung e. , et al . , 2016 ,
natur , 533 , 504 + choi e. , et al .
2015 , mnras , 449 , 4105 + crain r. a. , et al . 2010 , mnras , 407 , 1403 + diehl s. , li h. , fryer c. l. , rafferty d. , 2008 , apj , 687 , 173 - 192 + dullo b. t. , graham a. w. , 2013 , apj , 768 , 36 + david l. p. , et al . 2006 , apj , 653 , 207 + deason a. j. , et al . 2012 , apj , 748 , 2 + fabbiano g. , 2006 , ara&a , 44 , 323 + goulding a. d. , et al .
, 2016 , arxiv:1604.01764 + kim , d .- w . ,
fabbiano , g. 2003 , apj , 586 , 826 + kim d .- w . , fabbiano g. , 2013 , apj , 776 , 116 ( kf13 ) + kim d .- w . , fabbiano g. , 2015 , apj , 812 , 127 ( kf15 ) + krajnovi d. , et al . , 2013 , mnras , 433 , 2812 + lauer t. r. , et al . , 2007 ,
apj , 664 , 226 + mathews w. g. , et al .
2006 , apj , 652 , l17 + mcnamara b. r. , nulsen p. e. j. , 2007 , ara&a , 45 , 117 + navarro j. f. , frenk c. s. , white s. d. m. , 1997 , apj , 490 , 493 + oser l. , et al . 2010 , apj , 725 , 2312 + osullivan e. , forbes d. a. , ponman t. j. , 2001 , mnras , 328 , 461 + paggi a. , et al .
2014 , apj , 787 , 134 + pellegrini s. , 2005 , apj , 624 , 155 + pellegrini s. , et al .
2012 , apj , 758 , 94 + pota v. , et al . , 2013 , mnras , 428 , 389 + sarzi m. , et al . , 2013 , mnras , 432 , 1845 + scott n. , graham a. w. , schombert j. , 2013 , apj , 768 , 76 + su y. , gu l. , white r. e. , iii , irwin j. , 2014 , apj , 786 , 152 + sun m. , donahue m. , voit g. m. , 2007 , apj , 671 , 190 + watkins l. l. , evans n. w. , an j. h. , 2010 , mnras , 406 , 264 + white s. d. m. , frenk c. s. , 1991 , apj , 379 , 52 + wong k .- w . , et al . 2011 , apj , 736 , l23 | here we utilise recent measures of galaxy total dynamical mass and x - ray gas luminosities ( l@xmath0 ) for a sample of 29 massive early - type galaxies from the sluggs survey to probe l@xmath0mass scaling relations .
in particular , we investigate scalings with stellar mass , dynamical mass within 5 effective radii ( r@xmath1 ) and total virial mass .
we also compare these relations with predictions from @xmath2cdm simulations .
we find a strong linear relationship between l@xmath0 and galaxy dynamical mass within 5 r@xmath1 , which is consistent with the recent cosmological simulations of choi et al .
that incorporate mechanical heating from agn .
we conclude that the gas surrounding massive early - type galaxies was shock heated as it fell into collapsing dark matter halos so that l@xmath0 is primarily driven by the depth of a galaxy s potential well .
heating by an agn plays an important secondary role in determining l@xmath0 .
[ firstpage ] galaxies : star clusters galaxies : evolution galaxies : xrays |
the recent proliferation of smartphones and tablets is seen as a key enabler for anywhere , anytime wireless communications .
the rise of websites , such as facebook and youtube , significantly increases the frequency of users online activities . due to this continuously increasing demand for wireless access
, a tremendous amount of data is circulating over today s wireless networks .
this increase in demand is straining current cellular systems , thus requiring novel approaches for network design . in order to cope with this wireless capacity crunch ,
device - to - device ( d2d ) communications underlaid on cellular systems has recently emerged as a promising technique that can significantly boost the performance of wireless networks . in d2d communication , user equipments ( ues )
transmit data signals to each other over a direct link instead of through the wireless infrastructure , i.e. , the cellular network s evolved node bs ( enbs ) .
the key idea is to allow direct d2d communications over the licensed band and under the control of the cellular system s operator @xcite .
furthermore , as d2d communication often occurs over shorter distances , it is expected to yield higher data rates for the ues than infrastructure - based communications .
d2d communication is regarded as a promising technology for improving the spectral utilization of wireless systems .
recent studies have shown that the majority of the traffic in cellular systems consists of the download of content such as videos or mobile applications .
if we can cut off the traffic of this part , a large amount of capacity can be freed out .
usually , popular contents , such as certain youtube videos , are requested by much more frequently than others . as a result ,
the enbs often end up serving different mobile users with the same contents for multiple times .
as the enb has already sent the contents to mobile users , the contents are now locally accessible to other users in the same area , if cellular ues resource blocks ( rbs ) can be shared with others .
upcoming users who are within the transmission distance can request the contents from those users through d2d communication . in this case , the enb will have to only serve those users who request new " content , which has never been downloaded before . through this d2d communication , we can reduce considerable redundant requests to the enb , so that the traffic burden of enb can be released .
the main contribution of this paper is to propose a novel approach to d2d communications that allows to exploit the social network characteristics so as to improve the performance and reduce the load on the wireless cellular system . to achieve this goal , there are some key points we are going to discuss .
the first of all is to establish a stable d2d subnetwork to maintain the data transmission successfully .
a stable connection should guarantee that once the d2d communication is set up , the link should not be dropped easily . as a d2d subnetwork
is formulated by individual users , the connectivity among users sometimes is intermittent .
if the connection is too sensitive to users movement and easy to interrupt , it can neither offload the traffic of enb nor meet users satisfaction .
it is difficult to employ such dynamic information to make reasonable decisions .
however , the social relations in real world tend to be stable over time .
such social ties can be utilized to achieve efficient data transmission in the d2d subnetwork .
we name this social relation assisted data transmission network by offline social network ( offsn ) .
second , we assess the amount of traffic that can be offloaded to d2d communication , i.e. , what probability the requested contents can be served locally . to analyze this problem
, we study the probability that a certain content is selected .
this probability is affected by two different aspects : the external ( external influence from media , friends , etc . ) , and internal ( user s own interests ) aspects . while users interests are hard to know , the external influence is more easier to estimate .
the choices of the users are mutually dependent .
consequently , the network operator ( e.g. , via the enb ) can generate an online social network ( onsn ) to keep track of users access to online websites , and maintain the distribution over offsn users selection .
if we can estimate people s selection based on external influences , we can get the probability of contents being served locally . in this paper
, we will adopt practical metrics to establish offsn , and involve the novel learning process - indian buffet process to model the influence in onsn .
then we can get solutions for the previous two problems .
latter we will integrate offsn and onsn to carry out the traffic offloading algorithm for the cellular network .
our simulations also proved our analysis for the traffic offloading performance .
consider a cellular network with one enb and @xmath0 users .
the ues can receive signals from the enb through cellular network , or from other ues through d2d pairs using licensed spectrum resources . in the system ,
two network layers exist over which information is disseminated .
the first layer is onsn .
the links of contents spread out on popular websites , users access the links to contents . hence , the onsn is the platform over which users acquire the links of contents .
once a link is accessed , the data package of contents must be transmitted to the ues through the actual physical network .
taking advantage of the social ties , offsn is the physical layer network for contents behind the links spread out .
information dissemination in both onsn and offsn.,scaledwidth=28.0% ] an illustration of this proposed model is shown in fig .
[ fig : onsn and offsn ] .
each user active in the onsn corresponds to an ue in the offsn .
users access the link of a content in an increasing order of their labels . in onsn
, the link of a content is spread out according their popularity from frequent users to regular users .
in particular , a group of users , which we refer to as @xmath1 , have a high online activity , and , thus , are the main source of influence and information dissemination . in this respect ,
the choices of the @xmath2 , who access the onsn less frequently , are usually influenced by the frequent users . in the offsn , the first request of the content
is served by the enb .
subsequent coming users can thus be served by previous users who hold the content , if they are within the d2d communication distance . in this section
, we will study the properties of offsn and onsn , and quantify the relation between these two networks , before developing the proposed offloading algorithm . in the area covered by an enb , the distribution of the users can often be properly modeled .
for instance , in public areas such as office buildings and commercial sites , the density of users is much higher than in other locations such as sideways and open fields .
in addition , users are less likely to browse the web when they are walking or driving .
indeed , the majority of the data transmissions occurs in those fixed places . in such high density locations , forming d2d network as offsn becomes a natural process .
thus , we can distinguish two types of areas : highly dense areas such as office buildings , and white " areas such as open fields . in the former
, we assume that d2d networks are formed based on users social relations . while in the latter , due to the low density , the users are served directly by the enb .
d2d communications carry out in those subnetworks can effectively improve the cellular system throughput . if proper resource management is adopted , the interference among each subnetwork can be restricted @xcite .
thus , each offsn can be mutually independent from others to some extent .
optimizing the performance of each offsn can improve the overall performance of the cellular network .
the offsn is a reflection of local users social ties .
proper metrics need to be adopted to depict the degree of the connections among users in the offsn .
for example , users who are within the d2d communication range and more regularly meet can be seen as more robust connection . due
to mobility , grouping users by using only their last known location , can lead to dropped connections .
indeed , in public areas such as airports and train stations , most of the users have high mobility patterns and it is difficult to predict their future locations .
however , if we define users connection degree in offsn according to their encounter history or daily routes , stable d2d connections can be formed .
for example , officemates , classmates and family members always meet one another more regularly and frequently than others .
thus , those social ties lead to higher probabilities to transmit data among users
. if any two ues are within the d2d communication distance , the enb can detect and mark them as encountered .
but the specific way to collect data is not our main focus in this paper .
the contact duration distribution between two users is assumed to be a continuous distribution , which has a positive value for all real values greater than zero . for users who regularly meet ,
their encounter duration usually centers around a mean value .
so we can adopt a @xmath3 distribution to model the encounter duration between two users .
this distribution is widely used in modeling the call durations and has been shown to have a high accuracy @xcite .
to find the value for the two parameter @xmath4 and @xmath5 , we need to derive the mean and variance of the contact duration . as shown in fig .
[ fig : encounter history ] , given the contact duration @xmath6 and the number of encounters @xmath7 between user @xmath8 and user @xmath9 , an estimate of the expected contact duration length @xmath10 : @xmath11 encounter history between user i and j.,scaledwidth=13.0% ] the variance represents the fluctuation in the contact period .
if two cases have the same average contact period , the one with larger fluctuations would be less preferable since the contact length would be more uncertain .
thus , we measure the variance of the contact period distribution to reflect the fluctuation using an irregularity metric @xmath12 defined by @xcite : @xmath13 given the mean and variance of the encounter period , we can derive the encounter duration distribution : @xmath14 .
the probability density distribution ( pdf ) of encounter duration is : @xmath15 where @xmath16 .
then , we can calculate the probability of the contact durations that are qualified for data transmission .
if the contact duration is not enough to complete a data package transmission , the communication session can not be carried out successfully .
we adopt a closeness metric , @xmath17 to represent the probability of establishing a successful communication period between two ues @xmath8 and @xmath9 , which ranges from @xmath18 to @xmath19 .
the qualified contact duration is the complementary of the disqualified communication duration probability , so @xmath17 can be represented as : @xmath20 where @xmath21 is the minimal contact duration required to successfully transmit one content data package .
@xmath21 is a random variable that depends on the channel conditions between the two ues ( e.g. , the higher the signal strength between the ues , the smaller @xmath21 ) .
moreover , @xmath21 should depend on the content size .
@xmath22 is the lower incomplete gamma function .
larger closeness @xmath17 indicates a better future contact opportunity between user @xmath8 and user @xmath9 .
an offsn can be seen as a d2d network that is constructed by a group of users with stable connections .
hence , we can use the closeness metric @xmath17 to describe the communication probability between two users , which can also be seen as the weight of the link between user @xmath8 and user @xmath9 .
then , a threshold @xmath23 can be defined to filter the boundary between different offsns and white " areas .
if we have lower @xmath23 , more users will be added in , and the covered area of the offsn will increase . even though the probability that a user will be served locally will increase , it is likely that the group members who own the content are located far away . in this case , the prospective gain would not compensate the cost on associated power consumption for relaying and transmitting . also , offsns may overlap with each other , and cause inter - offsn interference which can further complicate the analysis . on the other hand ,
if @xmath23 is too high , only a small number of users can be grouped which makes it difficult to perform d2d communication .
therefore , one important problem is to find a proper @xmath23 that can balance the tradeoff between the cost and gain , and thus to get the best performance for the system .
first , we define the number of users in onsn as @xmath0 which is corresponding to @xmath0 ues in offsn .
the total number of available content in onsn is denoted by @xmath24 . given the large volume of data available online , we can assume that @xmath25 , @xmath26 .
@xmath27 represents the set of contents that have viewing histories and @xmath28 is the set of contents that do not have any viewing history . in onsn
, users select contents partially based on external influence .
it is possible to predict current users selections by analysing the onsn activities of previous users within the same offsn . to draw the probability of current user s selection , we adopt the indian buffet process ( ibp ) @xcite which serves as a powerful analytical tool for predicting users behaviors . in an ibp
, there are infinite dishes for customers to choose . the first customer will select its preferred dishes according to a @xmath29 distribution with parameter @xmath30 .
since all dishes are new to this customer , no reference or external information exists so as to influence this customer s selection .
however , once the first customer completed the selection , the following customers will have some prior information about those dishes based on the first customer s feedback .
therefore , the decisions of subsequent customers are influenced by the previous customers feedback .
customers learn from the previous selections to update their beliefs on the dishes and the probabilities with which they will choose the dishes .
the behavior of content selection in onsn is analogous to the behavior of dish selection in an ibp .
if we view onsn as an indian buffet , the online content as the infinite number of dishes , and the users as customers , we can interpret the contents spreading process online by an ibp .
users enter onsn sequentially to download their desired content .
when a user downloads its content , the recorded downloading times of contents changed .
this action will affect the probability that this content to be requested .
popular contents will be requested more frequently . while those contents that are only favored by a few number of people , or those new produced content will be requested less frequently .
so the probability distribution can be implemented from the ibp directly .
indian buffet process . ] in fig .
[ fig : ibp ] , we show one realization of ibp .
customers are labeled by ascending numbers in a sequence .
the shaded block represent the @xmath31th user selected dish @xmath4 . in ibp
, the first customer selects each dishes with equal probability of @xmath32 , and ends up with the number of dishes follows @xmath33 distribution . for subsequent customers @xmath34 , the probability of also having dish @xmath4 already belonging to previous customers is@xmath35 , where @xmath36 is the number of customers prior to @xmath31 with dish @xmath4 @xcite . repeating the same argument as the first customer
, customer @xmath31 will also have @xmath37 new dishes not tasted by the previous customers follows a @xmath38 distribution .
the probabilities of selecting certain dishes act as the prior information @xmath39 . for old " dishes which have been tasted before , @xmath40 . for
new" dishes which have not been sampled before , @xmath41 .
after user @xmath31 completes its selection , the prior will be updated to @xmath42 .
this learning process is also illustrated in fig .
[ fig : ibp ] .
@xmath43 is the number of dishes that have not been sampled before user @xmath31 s selection session .
we can see the selection probability for dish @xmath4 updated every time after each customer s selection .
as the inter - offsn interference of d2d communication can be restricted by methods such as power control @xcite@xcite . for simplicity , we place an emphasis on the intra - offsn interference due to resource sharing between d2d and cellular communication . in the offsn that acts as the subnetwork of cellular network
, the d2d transmissions and the enb transmissions will interfere . during the downlink period of d2d communication
, ues will experience interference from other cellular and d2d communications as they share the same subchannels . in this respect ,
the received power of link @xmath8 to @xmath9 can be expressed as @xcite : @xmath44 where @xmath45 is the transmit power , @xmath46 is path loss exponent which ranges from @xmath47 , @xmath48 is the channel response of the link @xmath8 to @xmath9 , @xmath49 is the complex gaussian channel coefficient that follows the complex normal distribution @xmath50 .
thus , we can define the transmission rate of users served by the enb and by d2d communication with co - channel interference @xcite : @xmath51 @xmath52 where @xmath53 , @xmath54 and @xmath55 are the transmit power of enb , d2d transmitter @xmath56 and @xmath57 , respectively , @xmath58 is the additive white gaussian noise ( awgn ) at the receivers , and @xmath59 represents the presence of interference from d2d to cellular communication , satisfying @xmath60 , otherwise @xmath61 . here
, @xmath62 , so @xmath63 represents the interference from the other d2d pairs that share spectrum resources with pair @xmath56 .
the transmission rate of the users that are only serviced by the enb and that do not experience co - channel interference in the offsn is given by : @xmath64 in the studied model , the enb maintains the contents distribution of each onsn .
when the @xmath31th user starts to surf online , this user will sample the content based on the prior information @xmath39 . to this end , this user will access old content with probability @xmath65 , and access new content with a @xmath38 distribution .
after content selection , @xmath39 is updated to the posterior probability @xmath66 .
the total amount of content each user selected can be draw from a @xmath29 distribution @xmath67 in the proposed model , the users aim at maximizing their data rate while minimizing the cost . without loss of generality , we assume that every content has the same size . consequently , we propose the following utility function for user @xmath31 : @xmath68 ( r_d - c_t)+m_n^0 ( r_c- c_m),\ ] ] where @xmath69 is the cost related to the d2d transmission power @xmath45 in ( 6 ) .
@xmath70 is the cost paid by the user to the enb for the data flow .
the utility function consists of two parts .
the first part is the utility of receiving old content via d2d communication .
the second part is the utility of downloading new content from the enb . from the enb perspective
, the goal is to maximize the overall data rate while also offloading as much traffic as possible . even though enb can offload traffic by d2d communication , controlling the switching over cellular and d2d communication causes extra data transmission . there
thus exists some cost such as control signals transmission and information feedback during the access process @xcite .
therefore , for the enb that is servicing a certain user @xmath31 , we propose the following utility function : @xmath71r_d+m_n^0 r_c - m_n c_c,\ ] ] where @xmath72 is the cost for controlling the resource allocation process .
the total traffic enb offloaded by d2d communication is @xmath73 . *
offsn generation * enb collects encounter information in cellular network locate the frequent users in high user density areas find the closeness @xmath17 between two users * 2 .
user activity detection * * 3 . service based on onsn activities * we propose a novel algorithm form which consists of multiple stages . in the first stage ,
the enb focuses on high user density areas , and collects the encounter history between users .
the enb locates one frequent user and its neighboring users via well - known algorithms in order to compute their closeness @xmath17 . by checking if @xmath74 , i.e. , if user @xmath8 and user @xmath9 satisfy the predefined closeness threshold , the enb can decide on whether to add this user into the offsn or not . by choosing a proper @xmath23 and power control ,
the interference among different offsns can be avoided .
this process will continue until no more user can be further added to the enb s list .
then , the users in the established offsn can construct a communication session with only intra - offsn interference . for websites that provide a portal to access content , such as facebook and youtube , the enb will assign a special tag .
once a user visits such tagged websites , the enb will inspect whether the user is located in an offsn or a white " area .
if the user is in white " area , user s any requests will be served by the enb directly .
if the user is located in an offsn , the enb will wait until the user requests contents . by serving previous user s requests ,
the enb has already built up a history file including the prior information @xmath39 of the content distribution in the onsn .
as soon as receiving user requests data , the enb detects if there are any resources in the offsn , and then choose to set up a d2d communication or not based on the feedback . for old content , the enb will send control signal to the @xmath8th ue with the highest closeness @xmath75 with user @xmath31 .
then ue @xmath31 and ue @xmath8 establish the d2d communication .
even if the d2d communication is setup successfully , the enb still wait until the process finishes successfully .
if the d2d communication fails , the enb will revert back to serving the user directly . for new content
, the enb serves the user directly .
after the selection is complete , the prior information updates to the posterior probability @xmath66 .
the proposed d2d communication algorithm is summarized in algorithm @xmath19 .
in this section , we give the simulation results to show traffic offloading performances of the system from different aspects and provide numerical results of how different parameters affect the system s performance .
consider @xmath76 active users than are randomly distributed in an offsn .
the size of the content library is unbounded .
we assume that the content selection process has been carried on for a period of times .
so the enb can obtain the prior information of the content distribution . in our simulation
, we have proved that , once the parameters are specified , the order in which the users perform their selection does not affect the performance of the system .
the main physical layer parameters are listed as follows .
the radius of an offsn is set up as @xmath77 m .
noise spectral density is @xmath78dbm / hz . noise figure at device is @xmath79db .
antenna gains and transmit power of enb is @xmath80dbi and @xmath81dbm . for device
, it is @xmath18dbi and @xmath82dbm , respectively . the impact of the parameter @xmath30 on offloading traffic.,scaledwidth=27.5% ] in fig .
[ fig : traffic and alpha ] , we investigate whether the user online activity degree will affect the amount of traffic that can be offloaded . as we can see from fig .
[ fig : traffic and alpha ] , with the increase of parameter @xmath30 , the amount of traffic that is offloaded from the enb decreases , but traffic still can be released compared to the enb serving only system .
the data rate is nearly @xmath18 when @xmath30 is small .
this result is due to the fact that , when @xmath30 is low , users are more likely to choose old content . here
, the requests are served by d2d communication in most of the cases , and cause nearly no traffic on the enb . as @xmath30 increases
, the data rate begins increase . indeed ,
when the users tend to make more selections , they may choose more new content .
the offloaded traffic amount generally decreases with the increase of online activities , which is coincidence with the common sense that , more contents downloading will cause more traffic to enb .
the traffic includes not only the contents data , but also the control signals the enb needs to send for the d2d communication arrangement .
the relationship between the offloading traffic and maximum d2d communication distance.,scaledwidth=27.5% ] as the d2d communication distance increases , the enb will have more possibilities for detecting available contents providers . as a result
, the performance of traffic offloading will be better with larger maximum distance .
this assumption is shown in fig .
[ fig : traffic and distance ] . in this figure
, we can see that , with @xmath30 setting to @xmath83 , increasing the maximum communication distance , yields a decrease in the enb s data rate and an increase in the amount of offloaded traffic . however , we note that , with the increase of the transmission distance , the associated ue costs ( e.g. , power consumption ) will also increase .
thus , the increase of d2d communication distance will provides additional benefits to the enb , but not for users .
average sum - rate at the enb , as the cost for control signaling varies.,scaledwidth=27.5% ] in fig .
[ fig : traffic and enb cost ] , we show the variation of the sum - rate at the enb as the cost for control signal varies . as the enb has to arrange the inter change process between cellular and d2d communication , necessary control information are needed . moreover ,
additional feedback signals are required for monitoring the d2d communication and checking its status .
those costs will affect the traffic offloading performance of the system . in our simulation
, we define the cost as the counteract to the gain in data rate from 5% to 50% . as we can see from fig .
[ fig : traffic and enb cost ] , increasing the cost on control signal , the offload traffic amount is decreased .
in this paper , we have proposed a novel approach for improving the performance of d2d communication underlaid over a cellular system , by exploiting the social ties and influence among individuals .
we formed the offsn to divide the cellular network into several subnetworks for carrying out d2d communication with only intra - offsn interference .
also we established the onsn to analyse the offsn users online activities . by modeling the influence among users on contents selection online by indian buffet process
, we obtain the distribution of contents requests , and thus can get the probabilities of each contents to be requested . with the algorithm we proposed ,
the traffic of enb has been released .
simulation results have shown that different parameters for enb and users will lead to different traffic offloading performances .
users with more online activities will increase and fluctuate the data rate of the enb .
enable larger maximum d2d communication distance will release more traffic burden of the enb . while if the cost on d2d communication arrangement is high , the traffic offloading performance will decrease . c. xu , l. song , z. han , d. li , and b. jiao , resource allocation using a reverse iterative combinatorial auction for device - to - device underlay cellular networks , " _ ieee globe communication conference ( globecom ) _ , anaheim , ca , dec .
3 - 7 , 2012 .
n. golrezaei , a. f. molisch , and a. g. dimakis , base - station assisted device - to - device communication for high - throughput wireless video networks , " _ ieee international conference on communications ( icc ) _ , pp . 7077 - 7081 , ottawa , canada , jun .
10 - 15 , 2012 .
j. guo , f. liu , and z. zhu estimate the call duration distribution parameters in gsm system based on k - l divergence method , " _ ieee international conference on wireless communications , networking and mobile computing ( wicom ) _ , pp . 2988 - 2991 , shanghai , china , sep .
21 - 25 , 2007 .
f. li and j. wu ,
localcom : a community - based epidemic forwarding scheme in disruption - tolerant networks , " _ proc .
ieee conference sensor , mesh and ad hoc communications and networks ( secon ) _ , pp .
574 - 582 , rome , italy , jun .
22 - 26 , 2009 . | device - to - device ( d2d ) communications is seen as a major technology to overcome the imminent wireless capacity crunch and to enable novel application services . in this paper
, we propose a novel , social - aware approach for optimizing d2d communications by exploiting two network layers : the social network and the physical , wireless network .
first we formulate the physical layer d2d network according to users encounter histories .
subsequently , we propose a novel approach , based on the so - called indian buffet process , so as to model the distribution of contents in users online social networks . given the online and offline social relations collected by the evolved node b , we jointly optimize the traffic offload process in d2d communication .
simulation results show that the proposed approach offload the traffic of evolved node b successfully . |
observing @xmath8 decay would tell us that the total lepton number is not a conserved quantity , and that neutrinos are massive majorana fermions .
answering these questions is obviously a crucial part of the search for the `` physics beyond the standard model '' .
consequently , experimental searches for the @xmath8 decay , of ever increasing sensitivity , are pursued worldwide ( for a recent review of the field , see e.g. @xcite ) .
however , interpreting existing results as a determination of the neutrino effective mass , and planning new experiments , is impossible without the knowledge of the corresponding nuclear matrix elements .
their determination , and a realistic estimate of their uncertainty , are therefore an integral part of the problem .
the nuclear matrix elements @xmath9 of the @xmath8 decay must be evaluated using tools of nuclear structure theory .
unfortunately , there are no observables that could be simply and directly linked to the magnitude of @xmath8 nuclear matrix elements and that could be used to determine them in an essentially model independent way . in the past , knowledge of the @xmath10-decay rate , and therefore of the corresponding matrix elements @xmath6 , and of the ordinary @xmath11 decay @xmath12 values and the corresponding beta strength distributions , were used to constrain the nuclear model parameters , in particular when the quasiparticle random phase approximation ( qrpa ) was employed @xcite . in the present paper we discuss a novel relation between these nuclear matrix elements .
very early , primakoff and rosen @xcite speculated that since the operators governing @xmath8 and @xmath10 decays differ by a relatively gentle radial dependence , approximately of the form @xmath13 , the corresponding matrix elements might be proportional to each other with the proportionality constant @xmath14 , where @xmath15fm is the nuclear radius .
at that time the authors also believed that the @xmath10 decay can be treated in closure , thus that the corresponding matrix element is dimensionless , while in fact the realistic @xmath10 matrix element has dimension energy@xmath16 . also , the @xmath8 matrix elements are now , by convention , made dimensionless by including the nuclear radius @xmath17 as a multiplicative factor , which is compensated by the factor @xmath18 in the corresponding phase space function @xmath19 . modern nuclear structure evaluations of these matrix elements do not support the conjecture of proportionality between @xmath9 and @xmath6 .
the rate of the @xmath10 decay has been determined experimentally in many nuclei , and hence the @xmath10 matrix elements @xmath6 are known .
they exhibit pronounced shell effects and vary rather abruptly between nuclei with different @xmath5 and @xmath4 . at the same time
, the calculated @xmath8 nuclear matrix elements , whether based on the qrpa @xcite , nuclear shell model @xcite , or the interacting boson model @xcite , do not show such a variability ; instead they vary relatively smoothly between nuclei with different @xmath20 .
the reason for the difference is , presumably , the very different momentum transfer @xmath21 involved in these matrix elements , even though they involve the same initial and final nuclear states . in the @xmath10 decay the momentum transfer @xmath21 is restricted to @xmath22 , where @xmath23 is the energy difference of the initial and final atomic masses .
hence , the allowed approximation is valid , @xmath24 , and only the gamow - teller operator @xmath25 and only the @xmath26 virtual intermediate states , contribute . on the other hand , in the @xmath8 decay
the momentum transfer is of the order of the nucleon fermi momentum @xmath27 mev , @xmath28 , and all @xmath29 virtual intermediate states can contribute significantly .
our discussion here sheds more light on the different behavior of the @xmath8 and @xmath10 matrix elements .
it is worthwhile to remember another type of relation , explored in the classic paper by pontecorvo @xcite . at that time
the available information on @xmath1 decay was based on the geochemical determination of the total decay rate , @xmath30 . since these two modes scale very differently with @xmath23 ( @xmath31 for @xmath32 and @xmath33 for @xmath34 ) pontecorvo suggested that comparing the total lifetimes of two isotopes , @xmath35te and @xmath36te , which have very different @xmath23 values , might reveal the presence of the lepton number violating @xmath32 decay , provided the nuclear matrix elements of these two isotopes are identical . while the matrix elements of these two isotopes are indeed rather close , they are not quite the same .
moreover , we know today that the @xmath32 decay rate is very much smaller , if it is indeed nonvanishing , than the @xmath34 decay rate . the present paper is structured as follows . in the next section
ii we describe the formalism that leads to the relation between the gamow - teller part of the @xmath8 matrix element and the @xmath10 matrix element evaluated in the closure approximation .
we also discuss the validity of the closure approximation in the @xmath8 case . in the following section iii we discuss this novel relation in more detail and show numerous examples . in section
iv we briefly discuss the issue of quenching of the axial current matrix elements . while closure is a rather poor approximation in the @xmath10 case , we argue in section v that combining the known lifetimes with the often measured distribution of the @xmath37 and @xmath38 strengths constrains the @xmath2 values substantially .
we believe that the relation found here allows one to better understand the different behavior of these matrix elements .
we conclude in the last section .
assuming that the @xmath8 decay is caused by the exchange of the light majorana neutrinos , the halflife and the nuclear matrix element are related through @xmath39 where @xmath40 is the easily calculable phase space factor , @xmath41 is the effective neutrino majorana mass whose determination is the ultimate goal of the experiments , and @xmath42 is the nuclear matrix element consisting of gamow - teller , fermi and tensor parts , @xmath43 where @xmath44 and @xmath45 are the matrix element ratios that are smaller than unity and , presumably , less dependent on the details of the applied nuclear model . in the following we concetrate on the gt part , @xmath46 ,
which can be somewhat symbolically written as @xmath47 where @xmath48 is the neutrino potential described in detail below and @xmath49 is the relative distance between the two neutrons that are trasformed in the decay into the two protons .
as function of relative distance of two @xmath11-decaying neutrons in the @xmath50 nucleus .
calculation performed for @xmath51ge with 23 single particle levels model space .
positive parity multipoles are shown in the upper panel and the negative parity ones in the lower panel .
( color online ) ] as a function of the relative distance of the decaying neutron pair for different nuclei .
( color online ) ] in ref .
@xcite , based on the qrpa , as well as in ref .
@xcite based on the nuclear shell model , the function @xmath52 that describes the dependence of the @xmath9 on the distance @xmath49 was introduced .
formally , this function can be defined as @xcite @xmath53 where @xmath54 is the dirac delta function . obviously , this function is normalized by @xmath55 and has the dimension lenght@xmath16 .
the shape of @xmath56 is very similar in both qrpa and nsm and in all cases consists of a peak with maximum at @xmath57 fm ending near @xmath58 fm , and of very little contributions for larger values of @xmath3 .
now lets turn to the case of the @xmath34 decay mode .
the matrix element @xmath6 governing the @xmath10 decay mode is of the form @xmath59 where the sumation extends over all @xmath26 virtual intermediate states .
we can introduce also the closure analog of @xmath6 , denoted by @xmath2 , by replacing the energies @xmath60 by a properly defined average value @xmath61 .
thus , @xmath62 in analogy with eq .
( [ eq : c(r ) ] ) we can define the new function @xmath63 while the matrix elements @xmath6 and @xmath2 get contributions only from the @xmath26 intermediate states , the function @xmath64 gets contributions from all intermediate multipoles .
this is the consequence of the @xmath65 function in the definition of @xmath66 . when expanded , all multipoles
naturally , when integrated over @xmath3 only the contributions from the @xmath26 are nonvanishing .
an example of the multipole decomposition of @xmath66 is shown in fig .
[ fig : c2nu_mult ] , and in fig .
[ fig : c2nu_nucl ] we show the functions @xmath66 for a variety of @xmath1 decaying nuclei . for completeness
we show here the qrpa formula used for the evaluation of the function @xmath66 and its multipole decomposition depicted in fig .
[ fig : c2nu_mult ] .
first , the function @xmath67 is introduced where @xmath3 is the relative distance between the neutrons in the states @xmath68 and @xmath69 , respectively protons in @xmath70 and @xmath71 .
then , the part of @xmath66 with the multipolarity @xmath29 is given by @xmath72_j || j^{\pi } k_f \rangle \langle j^{\pi } k_f | j^{\pi } k_i \rangle \langle j^{\pi } k_i || [ c_p^+ \tilde{c}_n]_j || 0_i^+ \rangle .
\nonumber\end{aligned}\ ] ] here @xmath73 and @xmath74 are the labels of the excited states with the multipolarity @xmath29 in the intermediate nucleus built on the initial and final nuclear ground states , and @xmath75_j || j^{\pi } k_f \rangle$ ] and @xmath76_j || 0_i^+ \rangle$ ] are the corresponding qrpa amplitudes .
it is now clear that , by construction , @xmath77 which is valid for any shape of the neutrino potential @xmath78 .
thus , if @xmath66 is known , @xmath56 and therefore also @xmath0 can be easily determined .
the equation ( [ eq : basic ] ) represents the basic relation between the @xmath32 and @xmath34 @xmath1-decay modes that we will explore further . note
that while the function @xmath66 has a substantial negative tail past @xmath79 fm , these distances contribute very little to @xmath56 .
this is a consequence of the shape of the neutrino potential @xmath80 that decreases fast with increasing values of the distance @xmath3 .
the neutrino potential @xmath81 governing the gamow - teller part of the matrix element @xmath9 is defined as @xmath82 where @xmath83 takes into account the finite size of the nucleon and is usually approximated using the above dipole type form factor with @xmath84 gev @xcite(varying @xmath85 between 1.0 - 1.2 gev makes little difference ) .
the function @xmath86 includes the terms from higher order hadron currents , namely induced pseudoscalar and weak - magnetism @xcite . the short range correlations are included using the method of ref .
the jastrow - like two - body function derived there is applied when the radial integrals in both functions @xmath87 and @xmath64 are evaluated ; they do not appear explicitly in eq .
( [ eq : pot ] ) .
we show in fig .
[ fig : pot ] the shape of the potential . when the finite nucleon size , higher order terms are neglected , and @xmath88 is assumed , the potential has coulomb - like shape @xmath89 .
the full potential , eq .
( [ eq : pot ] ) , however , is finite at @xmath90 , @xmath91 .
including the higher order currents and finite @xmath92 in eq .
( [ eq : pot ] ) increases the value of @xmath93 by @xmath9430% . .
different approximate forms , as well as the exact one , are shown ( color online ) . ] for the indicated nuclei evaluated in the closure approximation as a function of the assumed average excitation energy ( color online ) .
the values of @xmath9 obtained without the closure approximation are 5.24(@xmath51ge ) , 2.62 ( @xmath95zr ) , 4.99 ( @xmath96mo ) , and 4.07 ( @xmath35te ) . ]
the closure approximation , i.e. the replacement of the summation over the virtual intermediate states by matrix element of a two - body operator , is typically used in the evaluation of the @xmath9 .
it is worthwhile to test the validity of this approximation .
such test can be conveniently performed within the qrpa , where the sum over the intermediate states can be easily carried out .
in fact , the calculations performed in refs .
@xcite do not use closure . in this context
one can ask two questions : how good is the closure approximation ? and what is the value of the corresponding average energy ? in fig . [
fig : cl ] we illustrate the answers to these questions .
the exact qrpa matrix elements shown in the caption can be compared with the curves obtained by replacing all intermediate energies with a constant @xmath92 , which is varied there between 0 and 12 mev .
one can see , first of all , that the @xmath9 changes modestly , by less than 10% when @xmath92 is varied and , at the same time , that the exact results are quite close , but somewhat larger , than the closure ones .
thus , using the closure approximation is appropriate for the evaluation of @xmath9 even though it slightly underestimates the @xmath9 values .
however , the corresponding uncertainty is not more than the other uncertainties involved .
we compare the qrpa exact and closure @xmath9 for all nuclei of interest in the next section .
we evaluated the nuclear matrix elements ( nme ) @xmath0 and @xmath9 with and without closure approximation using the quasiparticle random phase approximation ( qrpa ) .
for all nuclear systems the single - particle model space consisted of @xmath97 oscillator shells plus @xmath98 and @xmath99 levels both for protons and neutrons ( 23 single particle states ) .
the single particle energies were obtained from the coulomb corrected woods
saxon potential .
two - body interaction g - matrix elements were derived from the argonne v18 one - boson exchange potential within the brueckner theory .
the pairing interaction was adjusted to fit the empirical pairing gaps @xcite .
the particle - particle and particle - hole channels of the g - matrix interaction of the nuclear hamiltonian @xmath100 were renormalized by introducing the parameters @xmath101 and @xmath102 , respectively .
while @xmath103 was used throughout , the particle - particle strength parameter @xmath101 was fixed by the data on the two - neutrino double beta decay rates @xcite for each nucleus separately . in the calculation of the @xmath8-decay nmes
the two - nucleon short - range correlations derived from same potential as residual interactions , namely from the argonne potential @xcite , were applied .
the unquenched value of the axial current coupling constant , @xmath104 was used here .
the modifications caused by the quenching of the weak axial current are discussed in the following two sections . on the other hand ,
the absolute values of @xmath105 were deduced from the averaged values of @xmath10-decay half - lives of ref .
@xcite . in table [ tab:1 ] we show both the calculated @xmath8 nmes evaluated with and without the closure approximation , as well as only the gt parts of their values .
also shown are the experimental @xmath10-decay nmes . using the qrpa method the closure matrix elements @xmath106 were also evaluated .
one can see that the spread among the candidate nuclei of the @xmath10 nmes is significatly larger when compared with the spread of the calculated @xmath8-decay nmes .
the table also demonstrates that using the closure approximation for evaluation of @xmath9 makes relatively little difference and that the gt part of @xmath9 is dominant in all considered nuclei . [ cols="<,^,^,^,^,^,^,^,^,^",options="header " , ] + since our goal is the determination of the gt part of the @xmath8 matrix element @xmath0 , a priori the knowledge of the @xmath2 , which depends only on the @xmath26 intermediate states , is insufficient . according to the eq .
( [ eq : basic ] ) we need for that purpose the function @xmath66 that depends , in principle , on all intermediate multipoles .
however , is we could use the expansion of the spherical bessel function @xmath107 in eq.([eq : pot ] ) in powers of @xmath108 and keep just the first term , the neutrino potential @xmath80 would be represented by a constant and the eq.([eq : basic ] ) would predict a simple proportionality between @xmath2 and @xmath0 .
however , such an expansion does not work .
in reality in eq .
( [ eq : pot ] ) @xmath109 and we can not approximate the neutrino potential @xmath80 by its value at @xmath90 . hence , we do not expect a proportionality between @xmath2 and @xmath0 and the qrpa evaluation supports this conclusion .
since the nuclear matrix elements @xmath9 must be determined theoretically , it is of obvious interest to search for any relation between their numerical values and other quantities that are either known from experiments or at least constrained by them .
here we describe such a relation between the dominant gamow - teller part @xmath0 of @xmath9 and the matrix element @xmath2 of the observed @xmath10-decay evaluated in the closure approximation .
the relation is based on the evaluation of the auxiliary functions @xmath56 and @xmath66 that describe the dependence of the corresponding nuclear matrix elements on the distance @xmath3 between the pair of neutrons that is transformed in the @xmath1 decay into a pair of protons .
thus ( see eqs .
( [ eq : c(r)int ] ) , ( [ eq:2nucl ] ) and ( [ eq : basic ] ) ) @xmath110 represents the required relation . however ,
while the matrix elements @xmath6 and @xmath2 depend only on the transition strengths and energies of the @xmath26 virtual intermediate states ( they are pure @xmath111 quantities ) , the function @xmath66 gets contribution from all multipoles .
thus , the relation that we found is an indirect one ; even if @xmath106 would be precisely known , the evaluation of the function @xmath66 requires additional nuclear theory input .
nevertheless , the relation in eq . ( [ eq : basic ] ) allows us to obtain a better insight into the problem of the @xmath4 and @xmath5 dependence of the matrix elements @xmath6 and @xmath9 .
while the known @xmath6 have a strong shell dependence , the calculated @xmath9 vary much less .
analysis of the functions @xmath66 and @xmath112 makes it possible to better understand where this fundamental difference comes from .
we show that , so far , the qrpa values of closure approximation @xmath2 matrix elements do not agree well with the same quantities based on the measured @xmath37 and @xmath38 strength functions and on the assumption of coherence ( i.e. same sign ) of contributions of individual states . until this discrepancy
is resolved , it is difficult to employ @xmath2 in order to constrain the magnitude of the @xmath8 matrix elements @xmath0 .
useful discussions with kazuo muto are appreciated .
the work of p.v . was partially supported by the us department of energy under contract no.de-fg02-88er40397 .
a.f . , r.h . and
f. acknowledge the support in part by the dfg project 436 slk 17/298 , the transregio project tr27 `` neutrinos and beyond '' and by the vega grant agency under the contract no .
1/0249/03 .
v. a. rodin , a. faessler , f. imkovic and p. vogel , phys .
c*68 * , 044302(2003 ) .
v. a. rodin , a. faessler , f. imkovic and p. vogel , nucl .
a766 * , 107 ( 2006 ) , and erratum * a793 * , 213 ( 2007 ) .
f. imkovic , a. faessler , v. a. rodin , p. vogel , and j. engel , phys .
cbf 77 , 045503(2008 ) . | we show that the dominant gamow - teller part , @xmath0 , of the nuclear matrix element governing the neutrinoless @xmath1 decay is related to the matrix element @xmath2 governing the allowed two - neutrino @xmath1 decay . that relation is revealed when these matrix elements are expressed as functions of the relative distance @xmath3 between the pair of neutrons that are transformed into a pair of protons in the @xmath1 decay .
analyzing this relation allows us to understand the contrasting behavior of these matrix elements when @xmath4 and @xmath5 is changed ; while @xmath0 changes slowly and smoothly , @xmath6 has pronounced shell effects .
we also discuss the possibility of phenomenological determination of the @xmath2 and from them of the @xmath0 values from the experimental study of the @xmath7 strength functions . |
the nuclear shell model has been very successful in microscopic descriptions of nuclear structure , and in this approach the shell - model effective interaction @xmath12 has played an important role and its determination has been extensively studied @xcite . as discussed in these references , @xmath12 may be determined using either an empirical approach where it is required to reproduce selected experimantal data or a microscopic one where @xmath12 is derived from realistic nucleon - nucleon interactions using many - body methods .
the interactions used in such microscopic calculations have been mostly the two - nucleon ( nn ) interaction @xmath1 @xcite . should @xmath12 have also contributions from the three - nucleon ( nnn ) force @xmath3 in addition to the two - nucleon one ?
in fact the need of @xmath3 in nuclear many - body problems has long been recognized .
the use of @xmath1 alone has been inadequate in reproducing the empirical nuclear matter saturation properties ( see e.g. @xcite and references quoted therein ) .
the inclusion of @xmath3 has been of essential importance in describing the binding energies and low - lying spectra of light nuclie @xcite and in explaining the long half - life of the @xmath13 @xmath14-decay @xcite .
@xcite have shown that the inclusion of @xmath3 plays a crucial role in describing the oxygen isotopes near the drip line . in the present work
, we shall calculate the shell - model effective interactions for the @xmath5 and @xmath6 shells using the chiral @xmath15 two - nucleon potential @xmath1 @xcite with and without the inclusion of the in - medium nnn force @xmath2 , which is a density - dependent two - nucleon potential recently derived from the leading - order chiral nnn force @xmath3 by holt , kaiser and weise @xcite .
we shall apply these effective interactions to shell - model calculations for nuclei @xmath7 , @xmath8 , @xmath9 and @xmath10 , to study the effects of @xmath2 in these nuclei .
we shall first calculate the low - momentum @xmath4 interactions @xcite from @xmath1 and @xmath2 .
our shell model effective interactions @xmath12 will then be calculated from these @xmath4 interactions using a folded - diagram formalism @xcite . in this formalism , @xmath12 is given as a folded - diagram expansion .
for the degenerate @xmath5 case , this expansion can be summed up using the commonly employed lee - suzuki ( ls ) iteration method @xcite . for the non - degenerate @xmath6 case we shall sum up the expansion using the extended krenciglowa - kuo iteration method recently developed by okamoto , suzuki , kumagai and fujii ( ekko ) @xcite .
the ekko method is efficient for deriving the effective interactions of non - degenerate model spaces such as the @xmath6 two - shell space , and it can also be conveniently applied to degenerate model spaces such as the @xmath5 one - shell one .
@xcite as mentioned earlier , the in - medium nnn potential @xmath2 is dependent on the nuclear medium density @xmath16 .
we shall study this @xmath16 dependence for both low and moderately high densities . clearly , at zero density @xmath2 vanishes .
nuclear matter calculations using @xmath1 alone have not been able to satisfactorily reproduce the empirical nuclear matter saturation properties @xcite .
it may be useful to study if the @xmath16 dependence of @xmath2 near the nuclear matter saturation density @xmath17 ( 0.16 @xmath18 ) may play an important role for nuclear matter saturation .
we shall do so by carrying out nuclear matter calculations with and without the inclusion of @xmath2 .
similarly this dependence may also be important for closed - shell nuclei , such as @xmath11 , whose nucleons are embedded in a nuclear medium of densities near @xmath17 .
we shall calculate closed - shell nucleus @xmath11 with and without the inclusion of @xmath2 , as a further study of its effect near @xmath17 .
the valence nucleons , such as those of @xmath7 , are in a medium of densities much less than @xmath19 .
thus our shell - model calculations for valence nuclei mentioned earlier are a study of @xmath2 at low densities .
the organization of the present paper is as follows .
we shall describe some details about the derivation of @xmath2 from @xmath3 in section ii .
it may be noted that @xmath2 is an effective density - dependent two - nucleon interaction , which is more convenient than its underlying three - nucleon potential for nuclear many - body calculations .
the methods we shall employ for the derivation of the shell - model effective interactions @xmath12 from @xmath1 and @xmath2 will be outlined there . in section iii
we shall present first our results of nuclear matter calculations using @xmath1 with and without the inclusion of @xmath2 .
the ring - diagram formalism @xcite employed for our nuclear matter calculations will also be outlined .
next we shall report the results of a similar ring - diagram calculation for @xmath11 .
results of our shell - model calculations of @xmath7 , @xmath8 , @xmath9 and @xmath10 using @xmath1 with and without the inclusion of @xmath2 will be presented and discussed in this section .
a summary and conclusion is presented in section iv .
we first describe how we include the effects of the leading - order chiral three - nucleon interaction , @xmath3 , in our calculations .
we consider the nuclear interaction @xmath20 as given by @xmath21 , where @xmath1 is the n@xmath0lo idaho two - nucleon potential @xcite and @xmath22 is a density - dependent two - body interaction obtained from the chiral three - nucleon force by closing one pair of external lines and summing over the filled fermi sea of nucleons .
the leading contribution to @xmath3 occurs at n@xmath23lo in the chiral power counting and is composed of a long - range two - pion exchange component @xmath24 , a medium - range one - pion exchange term @xmath25 , and a pure contact interaction @xmath26 : @xmath27 @xmath28 @xmath29 where @xmath30 is the average pion mass , @xmath31 is the difference between the final and initial momentum of nucleon _
i _ and @xmath32 the low - energy constants @xmath33 , @xmath34 , and @xmath35 appear already in the n@xmath23lo two - nucleon potential and are therefore constrained by low - energy nn phase shifts @xcite .
the low - energy constants @xmath36 and @xmath37 are typically fit to reproduce the properties of light nuclei @xcite .
a general three - body force may be written in second quantization as @xmath38,[456 ] ) b^{\dagger}_1 b^{\dagger}_2 b^{\dagger}_3 b_6 b_5 b_4,\ ] ] where the antisymmetrized matrix element is @xmath39,[456])\equiv \langle 123|v_{3n}|456 + 645 + 564 - 654 - 546 - 465 \rangle .
\label{a3n}\ ] ] here @xmath40 is a simple product matrix element , and @xmath41 and @xmath42 are creation and destruction operators defined with respect to the particle - hole vacuum @xmath43 with @xmath44 = 0 for all _ k_. from eqs .
( [ 3n1]-[3n4 ] ) we can write @xmath45 , where @xmath46 is the component of @xmath3 that is symmetric with respect to the interchange @xmath47
. now we contract one pair of the @xmath41 and _ b _ operators of the above @xmath48 ( both operators must be holes ) , and this leads to an effective two - body force @xmath49,[45 ] ) b^{\dagger}_1 b^{\dagger}_2 b_5 b_4,\ ] ] with @xmath50,[45])&=&\sum_{i\leq k_f}\left [ \langle i12|v^{(2)}_{3n}|i45 \rangle + \langle 1i2|v^{(2)}_{3n}|4i5 \rangle \right .\\
\nonumber & + & \langle 12i|v^{(2)}_{3n}|45i \rangle - \langle 1i2|v^{(2)}_{3n}|i45 \rangle \\ \nonumber & - & \langle i12|v^{(2)}_{3n}|4i5 \rangle - \langle 12i|v^{(2)}_{3n}|4i5 \rangle \\ \nonumber & - & \langle 1i2|v^{(2)}_{3n}|45i \rangle + \langle i12|v^{(2)}_{3n}|45i \rangle \\ & + & \left .
\langle 12i|v^{(2)}_{3n}|i45 \rangle - ( 4 \leftrightarrow 5)\right ] , \end{aligned}\ ] ] where @xmath51 denotes the nine exchange terms .
the above result is unchanged when @xmath52 is replaced by either @xmath53 or @xmath54 . in our calculations
we consider a background medium of symmetric nuclear matter at constant density characterized by a fermi momentum @xmath55 . in this way
analytic expressions can be obtained for @xmath22 , as shown in refs .
the above is a density dependent effective ` two - nucleon ' interaction which , unlike its underlying three - nucleon force , can be readily used in many - body problems .
as detailed in @xcite , the partial - wave potentials of the above @xmath2 have been derived from the lowest - order three - nucleon force .
we shall use them in our calculations , namely we shall consider the nucleon interaction as given by ( @xmath56 ) .
we use the folded - diagram theory @xcite to calculate the effective interaction . briefly speaking , in this theory @xmath12
is given by a folded - diagram series v_eff= - ^ + ^ - ^ , where @xmath57 represents a so - called @xmath57-box consisted of irreducible diagrams , as illustrated by the 1st- and 2nd - order diagrams of fig .
1 . ( the @xmath58-box is the same as the @xmath57-box except that @xmath58 does not have diagrams 1st - order in the interaction . )
suppose we include only @xmath1 .
then each vertex in the diagrams shown represents a @xmath4 low - momentum interaction @xcite derived from @xmath1 .
when we include also @xmath2 in the calculation of @xmath12 , the @xmath57-box diagrams for effective interactions will have a specific type of @xmath2 vertices as illustrated in fig .
2 . recall that diagram d4 of fig .
1 represents the interaction between two valence nucleons via @xmath1 .
this diagram becomes diagram ( a ) of fig . 2 if its @xmath1 is replaced by @xmath2 . here
the two valence nucleons have to have the participation of a sea - nucleon ( below fermi sea ) in order to activate @xmath3 , as indicated by the hole - line loop in the diagram .
( note each @xmath2 vertex of fig .
2 represents a @xmath4 interaction derived from @xmath2 . )
similarly diagram d1 of fig .
1 represents the interaction of a valence nucleon with a sea - nucleon via the two - nucleon interaction @xmath1 .
if this interaction is replaced by @xmath2 , this diagram becomes diagram ( b ) of fig .
@xmath3 must involve three nucleons , and hence here the valence nucleon interacts with ` two ' sea - nucleons as indicated by the two hole - line loops attached to @xmath3 . diagram ( c ) is the @xmath3 core polarization diagram corresponding to diagram d7 of fig .
again here one hole - line loop is needed for each vertex to activate @xmath3 . as seen from eqs .
( 31 ) and ( 32 ) , @xmath2 requires the involvement of at least one sea - nucleon , and it is this requirement which is reflected by the hole - line loops attached to the @xmath3 vertices in fig . 2 . to summarize , the @xmath57-box diagrams of fig .
1 are used for calculating @xmath12 when we use only @xmath1 , while additional diagrams as illustrated by fig . 2
are also included when the interaction is consisted of both @xmath1 and @xmath3 . after obtaining the @xmath57-box ,
we calcualte the effective interaction of eq . ( 9 ) using iteration methods .
the ls method @xcite is convenient for degenerate model spaces , and it is used to calculate the _ sd _ one - shell effective interactions . for the non - degenerate _
sdpf _ two - shell effective interactions , one can use either the krenciglowa - kuo iteration method ( kk ) @xcite or the recent extended kk method ( ekko ) of okamoto _ et al .
_ @xcite . both the ekko and kk methods are convenient for calculating the effective interaction for non - degenerate model spaces , while the ekko method being more efficient ( faster converging rate ) .
the main difference between the ekko and kk methods is the following .
the kk method employs the irreducible vertex function @xmath57-box which may be written as q()=[pvp+pvqqvp]_l where v represents the nn interaction , p denotes the model - space projection operator , and @xmath59 . the energy variable @xmath60 is determined self - consistently @xcite .
note that @xmath57 contains only valence - linked diagrams as indicated by the subscript _
l_. in the ekko method , the above @xmath57-box is replaced by the @xmath61-box defined by @xcite @xmath62,\ ] ] where @xmath63 is the first derivative @xmath64 .
when the p- and q - space are not sufficiently separated , the above @xmath57-box may have singularities .
an important advantage of the ekko method is that the @xmath61-box is well behaved even when the corresponding @xmath57-box is singular , overcoming the above singularity difficulty and making the ekko method more efficient @xcite .
we shall calculate the _ sdpf _ two - shell effective interactions using both the ekko and kk iteration methods , to cross check their convergences .
as described in section ii , we shall calculate the shell - model effective interactions with the inclusion of the medium - dependent three - nucleon force @xmath2 which is obtained from a chiral nnn force by integrating one participating sea - nucleon over the fermi sea .
the @xmath2 interaction is a density dependent effective interaction , and we shall first carry out nuclear matter calculations with the inclusion of @xmath2 , to study the properties of this interaction at various densities near the nuclear matter saturation density @xmath17 . before doing so , we need to choose or decide the low - energy constant of @xmath3 ( see eqs .
( 1 - 3 ) ) to be used in our calculations .
the low - energy constants @xmath65 , @xmath66 and @xmath67 of @xmath68 are well determined ; they are constrained by low - energy nn scattering data @xcite .
their values so determined ( section ii ) will be used in the present work .
but the low - energy constants @xmath36 and @xmath37 , of @xmath69 and @xmath70 respectively , are less well known ; their values determined from properties of light nuclei exhibit considerable variations @xcite .
the @xmath2 interaction depends explicitly on the fermi momentum @xmath55 which is well defined for nuclear matter . but
@xmath55 is not well defined for finite nuclei , causing uncertainty in determining the low - energy constants of @xmath2 from properties of finite nuclei .
but for nuclear matter , @xmath55 is well defined ; we have thus chosen to study these constants by way of nuclear matter calculations with @xmath2 included .
we calculate the equation of state ( eos ) for symmetric nuclear matter using a ring - digram formalism @xcite . a brief description of this formalism is presented below , to outline how we include @xmath2 in our calculations .
using familiar renormlization procedures @xcite , we first calculate the low - momentum interactions @xmath71 and @xmath72 respectively from @xmath1 @xcite and @xmath2
. a common decimation scale of @xmath73 is employed .
this value is chosen because at this scale the low - momentum interactions derived from different nn potentials @xcite are remarkably close to each other @xcite , leading to a nearly unique low - momentum interacton . using the above ring - diagram framework , the ground - state energy shift @xmath74 is given by the all - order sum of the _ pphh _ ring diagrams as illustrated in fig .
( @xmath75 is defined as ( @xmath76 ) where @xmath77 is the true ground state energy and @xmath78 that for the non - interacting system . ) as shown in fig.3 , diagram ( 1 ) is a 1st - order ring diagram with its vertex @xmath79 calculated from the n@xmath0lo chiral nn potential @xcite .
similarly diagram ( 2 ) is a 1st - order ring diagram with @xmath80 obtained from the lowest order chiral @xmath3 as described in section ii .
diagram ( 3 ) has vertices from both @xmath79 and @xmath80 .
similar to fig . 2 , to have the pair of nucleons in the ring diagrams interact with @xmath3 there must be the participation of a third nucleon .
thus the @xmath3 vertices in fig .
3 all have a one - hole - line loop attached to them . with these ring diagram summed to all orders ,
the ground - state energy shift for nuclear matter is given as @xmath81|kl \rangle.\end{aligned}\ ] ] the transition amplitudes above are @xmath82 the _ _ m__th true eigenstate of the ( _ a_-2 ) system , and @xmath83 if @xmath84 and @xmath85 if @xmath86 .
note that @xmath87 is a strength parameter , integrated from 0 to 1 .
the amplitudes _ y _ are calculated from a rpa equation @xcite based on the ( @xmath1+@xmath2 ) interactions .
note that our calculation is the same as the usual hartree - fock ( hf ) one if we include only the 1st - order ring diagrams ( 1 ) and ( 2 ) of fig .
4 , and in this case the above _ y _ amplitude becomes @xmath88 where @xmath89=1 for @xmath90 and = 0 otherwise .
our ring - diagram calculations include particle - hole fluctuations of the fermi sea , while not so for the hf calculations .
we have carried out ring - diagram calculations for symmetric nuclear matter using a wide range of values for @xmath36 and @xmath37 .
our results with @xmath91 and @xmath92 are displayed in fig .
4 ; they give @xmath93 and @xmath94 , both in satisfactory agreement with the empirical values .
two curves are shown in the figure , one with @xmath1 alone and the other with the addition of @xmath2 . comparing with the @xmath1 curve , it is of interest that the effect of @xmath2 is slightly attractive for low densities ( @xmath95 ) while becomes strongly repulsive at high densities .
recall that we have used a common decimation scale of @xmath96 @xcite for both @xmath79 and @xmath80 .
it is of interest that in the ring - diagram calculations with brown - rho scaling @xcite a large @xmath97 of @xmath98 is needed for obtaining satisfactory nuclear matter saturation properties , while in the present calculation with @xmath2 satisfactory results can be obtained with the use of a smaller @xmath97 ( 2.1 @xmath99 ) .
we have calculated the hf s.p .
spectrum @xmath100 in nuclear matter with the inclusion of both @xmath1 and three - nucleon force @xmath2 .
the inclusion of the latter has been found to have significant effect to the spectrum .
@xmath100 can be well fitted by the quadratic expression ( @xmath101 ) where @xmath102 is the effective mass and @xmath103 is a well - depth parameter representing the s.p .
energy at zero momentum . in fig .
5 we present our results for @xmath103 and @xmath104 for various densities .
two curves are shown : the lower one without and the top one with the inclusion of @xmath2 .
the densities for the 7 data points of each curve are , from left to right , ( 0.25 , 0.5 , 1.0 , 1.5 , 2.0 , 2.5 , 3.0 @xmath17 ) respectively .
as shown by the lower curve of the figure , @xmath103 and @xmath104 both vary monotonically with the density when only @xmath1 is employed . with increasing density , @xmath103 becomes incresingly more negative and @xmath104 increasingly larger , exhibiting no saturation ( up to 3.0 @xmath17 ) .
the trend is quite different for the upper curve where the three - nucleon force is included .
here @xmath104 still monotonically increases with density , but @xmath103 arises after saturation , approaching zero at some higher density .
it is of interest that @xmath2 has a large effect in raising the chemical potential of nucleons in high density nuclear matter ; this raise of the chemical potential may play an important role in enabling such nucleons decaying into other baryons such as hyperons . as indicated by fig .
4 , our ring - diagram calculations with the inclusion of the nnn force @xmath2 have described well the nuclear matter saturation properties . it would be useful and of interest to check if this approach is also satisfactory for describing closed - shell nuclei whose nucleons are mainly embedded in a similar nuclear medium of densities near @xmath19 , the nuclear matter satturation density . to study this possiblity ,
we have extended the above ring - diagram nuclear matter calculation to the closed shell nucleus @xmath105 , using the same low - energy parameters @xmath106 and @xmath107 ( -2.70 and 0.70 ) determined earlier . note that these parametrs will also be used later in our @xmath5 and @xmath6 shell - model calculations .
our ring - diagram calculation for @xmath11 is quite similar to the nuclear matter one .
the ground - state energy shift for @xmath105 is also given by the all - order sum of the ring diagrams illustrated in fig .
3 , except for the following differences . for nuclear matter , the s.p . states are plane - wave states and each particle line in the figure denotes a plane - wave particle with momentum @xmath108 and each hole line with @xmath109 . for @xmath105
we use harmonic oscillator s.p .
wave functions ( @xmath110=14 mev ) , and each particle line in the figure denotes a nucleon in the particle shells ( _ 1s0d1p0f _ ) , and each hole line a nucleon in the _ 0s0p _ shells .
for nuclear matter , we use momemtum model space where all particles have momentum @xmath111 . for @xmath105
, we use an oscillator model space composed of the s.p .
orbits of the @xmath112 shells .
indeed the ring - diagram calculations for nuclear matter and @xmath105 are quite similar .
the ground - state energy shift for @xmath105 is also given by eq.(12 ) , except that the orbits ( _ ijkl _ ) now all refer to shell - model orbits and the summation restriction of @xmath113 is replaced by that of _ ijkl _ belonging to the above oscillator model space .
with the same @xmath1 and @xmath2 interactions used for our nuclear matter calculations , we have carried out ring - diagram calculations for the ground - state energy of @xmath105 . since @xmath2 is dependent on the nuclear density @xmath16 , we need to have the value of @xmath16 to carry on the calculation . for nuclear matter , this @xmath16 is well defined , but not so for finite nuclei .
a preliminary estimate for the density for @xmath114 can be made by considering it as a uniform sphere @xcite ; in this way the average density for @xmath11 is estimated as @xmath115 . a more realistic estimate for
this density should be lower , as physically the nuclear surface is diffused ( not sharp ) .
in table i we present results for the ground - state energy per nucleon of @xmath105 for sevearl densities , @xmath116 0 , 0.5 , 0.70 and 0.75 .
as seen , the binding energy calculated with @xmath117 ( i.e. no @xmath2 ) is clearly not adequate .
the inclusion of @xmath2 lowers the ground - state energy per nucleon significantly ; for example the inclusion of @xmath2 at @xmath118=0.75 has increased the ring - diagram ground - state energy from -6.97 ( only @xmath1 ) to -8.01 mev ( with @xmath2 ) per nucleon , in good agreement with experiment .
since we have not included the coulomb interaction in our calculations , an empirical coulomb energy of 1.14 mev per nucleon @xcite has been included in the above calculation for the ground - state energy of @xmath105 .
it may be noted that our results for @xmath118= 0.70 and 0.75 are rather close to each other .
we have calculated the potential enengy(pe ) from the @xmath119 interaction .
results given by the 1st- and 2nd- and all - order ring diagrams , denoted by pe-1st , pe-2nd and pe - ring respectively , are listed in the table .
it may be noted that pe-2nd is much smaller than pe-1st . and
the difference between ( pe-1st+pe-2nd ) and pe - ring is rather small , about 5@xmath120 .
_ @xcite have found that @xmath4 is suitable for perturbative nuclear matter calculations .
the good convergence shown in table i suggests that this interaction is also suitable for perturbative calculations of finite nuclei .
.ring - diagram calculations of the ground - state energy per nucleon @xmath77 for @xmath105 .
the interaction ( @xmath56 ) is used . the potential energy per nucleon given by the 1st- , 2nd- and all - order ring diagrams
are denoted respectively by pe-1st , pe-2nd and pe - ring . @xmath121 and @xmath122
are given by the 1st- and all - order ring diagrams respectively .
the experimental result is from @xcite .
all energies are in mev . [
cols="^,^,^,^,^,^",options="header " , ] having seen that the medium - dependent three - nucleon force @xmath2 has given desirable results for nuclear matter and closed - shell nucleus @xmath11 , we shall now study its effects on the nuclear effective interactions for the degenerate _
one - shell and the non - degenerate _
sdpf _ two - shell model spaces .
we use the same methods as described in @xcite for the calculation and application of these effective interactions , except for the following difference : in @xcite only the two - nucleon interaction @xmath1 was employed . in the present work
we include both @xmath1 and @xmath2 .
we first calculate the low - momentum interactions from these interactions , and using them to obtain the @xmath57-box diagrams as shown in figs . 1 and 2 .
the same @xmath1 and @xmath2 interactions we used earlier in our calculations for nuclear matter and @xmath11 are employed .
the degenerate @xmath5 and non - degenerate @xmath6 effective interactions are obtained from the @xmath57-box using respectively the ls iteration method @xcite and the ekko iteration method @xcite .
the nondegenerate _
sdpf _ effective interactions can also be calculated using the kk iteration method @xcite .
we first apply the above effective interactions with @xmath2 to nuclei @xmath7 and @xmath8 .
since @xmath2 is density dependent , we need to know the ` local density ' @xmath123 felt by the valence nucleons in these nuclei .
it should be small , but its precise value is rather uncertain . in the present work , we shall estimate @xmath123 by comparing the density profile of @xmath11 and that for the valence nucleons . in fig .
6 we plot the density profile @xmath124 of @xmath11 obtained with its wave function assumed to be a closed shell - model @xmath125 core .
we also plot the distribution @xmath126 where @xmath127 is the radial harmonic oscillator wave function for @xmath128 , with the scaling parameter @xmath129=0.1 . as seen from the plots ,
the pair of @xmath128 nucleons reside primarily in the low - density region of @xmath130 . depending on the averaging procedure employed ,
we have estimated @xmath123 from @xmath131 and @xmath132 , obtaining values ranging from @xmath133 to @xmath134 .
we have considered another scheme to estimate @xmath123 .
the rms radius @xmath135 for the shell - model @xmath128 orbit is indicated by an arrow in the figure . a simple scheme to estimate @xmath123 is to let it equal to @xmath136 , the value obtained in this way being @xmath137 .
this value may be reduced if realistic s.p .
wave functions are employed .
the @xmath138 orbit is nearly unbound , and its rms radius should be considerably larger than that for @xmath128 .
assuming @xmath110 = 10 mev for the @xmath138 orbit , its @xmath135 would be @xmath139 giving @xmath140 for this orbit .
we believe that a suitable range for @xmath123 is from @xmath141 to @xmath142 ( @xmath143 ) . in the present work ,
we adopt @xmath144 , to illustrae the effects of @xmath2 . in fig .
7 the matrix elements of the @xmath5-shell effective interactions calculated from @xmath1 with and without @xmath2 , denoted respectively as nn+nnn and nn , are compared .
as seen , the two sets of matrix elements are rather similar to each other , with the magnitudes of those including @xmath2 being slightly larger .
the low - lying spectrum of @xmath7 and @xmath8 calculated with the above interactions are presented in figs . 8 and 9 .
the main effect of the ` nnn ' force is a small downward shift for the ground states of @xmath7 and @xmath8 , while leaving the other states largely unchanged .
this trend is consistent with what we have observed for nuclear matter calculations ( see fig .
4 ) , namely the effect of @xmath2 is slightly attractive at low densities . with the inclusion of thre nnn force ,
the ground - state energy of @xmath8 is in good agreement with experiment while that for @xmath7 is slightly overbound .
the agreements of the other states with experiments for @xmath7 are better than @xmath8 . using the method outlined in section ii and @xcite
, we have also calculated the effective interaction for a @xmath6 two - shell space using @xmath1 with and without the inclusion of @xmath2 . in figs . 8 and 9 ,
the spectra of @xmath7 and @xmath8 obtained with the former effective interaction are also presented ( two - shell , nn+nnn ) .
as seen , the two - shell and one - shell results are rather similar to each other . to further study the effects of the three - nucleon force @xmath2 , we have extended the calculations of figs . 8 and 9 to nuclei @xmath9 and @xmath10 , with results presented in figs .
10 and 11 . here ,
similar to @xmath7 and @xmath8 , the ground - state energies of @xmath9 and @xmath10 obtained with @xmath1 only are both too high compared with experiments ( see the nn columns ) .
as shown by the nn+nnn columns , the inclusion of @xmath2 slightly lowers these energies . the one - shell and
two - shell results are rather close to each other for @xmath9 .
but they are slightly different for @xmath10 , with the nn+nnn ground - state energy noticebly lower than the nn one . as shown in figs .
8 - 11 , although there are qualitative agreements between the calculated and experimental spectra of @xmath7 , @xmath8 , @xmath9 and @xmath10 , there are significant disagreements between them .
for example , the orderings of the calculated low - lying states of @xmath9 are in fair agreement with the experimental ones , but their relative spacings are not .
as shown by the figures , the effects from @xmath2 are generally small and attractive , mainly slightly lowering their ground - state energies while having nearly no influence on the relative spacings of the calculated spectrum
. it should be useful to further study how to reduce the above disagreements .
we have employed a low - order approximation for the @xmath57-box , including only 1st- and 2nd - order diagrams .
the inclusion of higher - order diagrams may alter our results . as carried out by by holt et al .
@xcite , certain classes of planar diagrams of the @xmath57-box can be summed up to all orders using the kirson - babu - brown induced interaction method .
we plan to extend our present calculations by including these planar diagrams , and study their effects to the shell - model description of these nuclei . ?
we have carried out shell - model calculations for nuclei @xmath7 , @xmath8 , @xmath9 , @xmath10 and @xmath11 using effective interactions derived from the chiral two- and three - nucleon potentials .
we start from the @xmath15 two - nucleon potential @xmath1 together with an in - medium three - nucleon potential @xmath145 which is a density - dependent effective interaction derived from the lowest - order chiral three - nucleon force @xmath3 by holt , kaiser and weise @xcite .
the @xmath4 low - momentum interactions are then derived from these potentials , and are used in a folded - diagram expansion for calculating the @xmath12 interactions for both the degenerate @xmath5 and non - degenerate @xmath6 model soaces . the well - known lee - suzuki iteration method @xcite is used for the derivation of the @xmath5 effective interactions , while the recently developed iteration method of okamoto _ et al . _
@xcite is employed for the @xmath6 ones .
we have studied the effects of @xmath2 , which is an effective two - body interaction depending on the medium density @xmath16 , at various densities near and below the nuclear matter saturation density @xmath19 .
we first apply the @xmath2 interaction to nuclear matter calculations for which the medium density is uniform and well defined .
a ring - diagram method for nuclear matter has been employed .
the low - energy constants @xmath106 and @xmath107 , of respectively the one - pion and contact terms of the leading - order chiral three - nucleon potential @xmath3 , are not well known .
we have chosen these constants by requiring that their use in our nuclear matter calculations with @xmath1 and @xmath2 , which is derived from @xmath3 , satisfactorily reproduces the saturation properties of nuclear matter . as indicated in fig .
4 , the inclusion of @xmath2 has largely improved our nuclear matter results comparing with those given by @xmath1 alone , and the resulting nuclear matter saturation propertiesa are in good agreement with the empirical values .
we then calculate the @xmath5 and @xmath6 effective interactions with the inclusion of the above @xmath2 , and apply them to nuclei @xmath7 , @xmath8 , @xmath9 and @xmath10 .
an ambiguity here is the determination of the local density @xmath123 felt by the valence nucleons of these nuclei .
this density should be small compared with the saturation density @xmath19 of nuclear matter , but it is difficult to determine its value precisely .
we have estimated this value based on the density profile of @xmath11 given by a simple @xmath125 shell - model wave function .
this is an approximation which should be further investigated and improved upon . in our calculations for these nuclei , we have adopted the estimation of @xmath146=0.15 . as indicated by figs .
8 - 11 , the main effect of @xmath2 is to lower the ground - state energies of these nuclei slightly , making them in better agreements with experiments than the energies given by @xmath1 alone .
the effects of @xmath2 to these nuclei are general small , mainly because their valence nucleons are in a low - density medium . the above @xmath2
has also been applied to a shell - model calculation of closed - shell nucleus @xmath11 .
the nucleons in closed - shell nuclei are embedded in nuclear medium of much larger density than that for the valence nucleons mentioned above . for them
the effect of @xmath2 is expected be more important .
indeed this is confirmed by our calculations .
as illustrated in table i , the inclusion of this interaction has been essential for satisfactorily reproducing the empirical binding energy of @xmath11
. it should be useful and of interest to further study this in - medium three - nucleon potential by applying it to closed - shell nuclei and those with a few valence holes .
we are very grateful to l. coraggio , a. covello , a. gargano , n. itaco , m. machleidt and r. okamoto for many helpful discussions .
partial supports from the us department of energy under contracts de - fg02 - 88er40388 and the dfg ( deutsche forschungsgemeinschaft ) cluster of excellence : origin and structure of the universe are gratefully acknowledged .
brown and b.h .
wildenthal , ann .
* 38 * , 29(1988 ) .
b.a . brown , w.a .
richter , phys .
c74 , 0343150 ( 2006 ) .
m. hjorth - jensen , t. t. s. kuo , and e. osnes , phys .
rep . * 261 * , 126 ( 1995 ) , and references therein .
l. coraggio , a. covello , a. gargano , n. itaco and t.t.s .
kuo , prog .
phys . , 62 ( 2009 ) 135 , and references quoted therein .
siu , j.w .
holt , t.t.s . kuo and g.e .
brown , phys . rev . *
c79 * , 054004 ( 2009 ) .
h. dong , t.t.s .
kuo and r. machleidt , phys .
rev . * c80 * , 065803 ( 2009 ) .
a. nogga , h. kamada and w. glockle , phys .
lett . * 94 * , 944(2000 ) .
peiper , k. varga and r. b. waringa , phys .
rev . * c66 * , 044310 ( 2002 ) .
p. navratil , v.g .
gueorguiev , j.p .
vary , w.e .
ormand and a. nogga , phys .
lett . * 99 * , 042501(2007 ) . j. w. holt , n. kaiser and w. weise , phys . rev . * c79 * , 054331 ( 2009 ) .
j. w. holt , n. kaiser and w. weise , phys . rev .
* c81 * , 024002 ( 2010 ) .
t. otsuka , t. sizuki , j.d .
holt , a. schwenk and y. akaishi , phys .
lett . * 105 * , 032501 ( 2010 ) .
d. r. entem , r. machleidt , and h. witala , phys .
c * 65 * , 064005 ( 2002 ) .
s. k. bogner , t. t. s. kuo and l. coraggio , nucl .
* a684 * , ( 2001 ) 432 .
bogner , t.t.s .
kuo , l. coraggio a. covello , and n. itaco , phys .
c * 65 * , 051301(r ) ( 2002 ) .
s. k. bogner , t. t. s. kuo , and a. schwenk , phys .
* 386 * , 1 ( 2003 ) .
t. t. s. kuo and e. osnes , _ lecture notes in physics _
( springer - verlag , new york , 1990 ) , vol . 364 .
kuo , s.y .
lee and k.f .
ratcliff , _ nucl .
phys . _ * a176 * ( 1971 ) 172 .
lee and k. suzuki , _ phys .
_ * 91b * , 173 ( 1980 ) . k. suzuki and s.y .
lee , _ prog .
phys . _ * 64 * , 2091 ( 1980 ) .
r. okamoto , k. suzuki , h. kumagai and s. fujii , to be published in ` proceedings of the 10th international spring seminar on nuclear physics ( may 21 - 25 , sur mare vietri , italy , ed .
by a. covello ) ' ; [ nucl - th ] arxiv:1011.1994v1 .
k. suzuki , r. okamoto , p.j .
ellis and t.t.s .
kuo , _ nucl .
phys . _ * a567 * ( 1994 ) 576 .
krenciglowa and t.t.s .
kuo , _ nucl .
* a235 * , 171 ( 1974 ) .
kuo , f. krmpotic , k. suzuki and r. okamoto , _ nucl .
phys . _ * a582 * , 205(1995 ) .
h. dong , t.t.s .
kuo and j.w .
holt , preprint ( may 2011 , to be submitted to arxiv and prc ) . h. q. song , s. d. yang and t.t.s .
kuo , nucl .
* a462 * , 491 ( 1987 ) . m. c. m. rentmeester , r. g. e. timmermans , and j. j. de swart , phys . rev .
c * 67 * ( 2003 ) 044001 .
e. epelbaum , prog .
57 * , 654(2006 ) .
r. machleidt , phys .
c * 63 * , 024001 ( 2001 ) . p. ring and p. schuck , _ the nuclear many - body problem _ ( springer - verlag , new york , 1980 )
. s. k. bogner , a. schwenk , r. j. furnstahl and a. nogga , nucl .
a763 * ( 2005 ) 59 .
j. d. holt , j. w. holt , t. t. s. kuo , g. e. brown and s. k. bogner , phys . rev . * 72 * , 041304r ( 2005 ) . | shell - model calculations for several mass 16 - 19 nuclei are performed using the n@xmath0lo two - nucleon potential @xmath1 with and without the addition of an in - medium three - nucleon potential @xmath2 , which is a density - dependent effective two - nucleon potential recently derived from the leading - order chiral three - nucleon force @xmath3 by holt , kaiser , and weise .
we first calculate the @xmath4 low - momentum interactions from @xmath1 and @xmath2 .
the shell - model effective interactions for both the @xmath5 one - shell and @xmath6 two - shell model spaces are then obtained from these low - momentum interactions using respectively the lee - suzuki and the recently developed okamoto and suzuki iteration methods .
the effects of @xmath2 to the low - lying states of @xmath7 , @xmath8 , @xmath9 and @xmath10 are generally small and attractive , mainly lowering the ground - state energies of these nuclei and making them in better agreements with experiments than those calculated with @xmath1 alone .
the excitation spectra of these nuclei are not significantly affected by @xmath2 .
the low - lying spectra of these nuclei calculated with the @xmath5 and @xmath6 model spaces are closely similar to each other .
our shell - model calculations for @xmath11 indicate that the @xmath2 interaction is important and desirable for the binding energy of this nucleus . |
strongly correlated classical and quantum two - component systems are of high interest in many fields , for a recent overview see @xcite .
examples include electron - hole plasmas in semiconductors , e.g. @xcite , dense electron - ion plasmas ( warm dense matter ) and complex or dusty plasmas @xcite .
while the simple one - component plasma ( or jellium ) model of a coulomb system has been studied in detail it disregards many fundamental properties , in particular the formation and phases of bound states .
much less is known about `` real '' two - component plasmas .
although a number of studies have been performed in recent years , e.g. @xcite , many questions remain . in particular , the phase diagram and its dependence on the mass ratio of the two components is known only qualitatively . here , we analyze these questions by considering the simplest realization of a neutral two - component plasma : a bilayer system with spatially separated light negative and heavy positive charges , in the following referred to as electrons ( `` @xmath0 '' ) and holes ( `` @xmath1 '' ) . at low temperature and large mass anisotropy , @xmath2 , the ( classical ) heavy component has a tendency to form a wigner crystal whereas the light component forms a neutralizing background .
thus , a strong ( weak ) coupling situation is realized for the heavy ( light ) component , which can be quantified in terms of the one - component coupling parameter , @xmath3/[\hbar^2/m_\alpha \bar{a}^2]=\bar{a}/a_b$ ] , where @xmath4 are the effective particle masses @xmath5 , @xmath6 is the elementary charge , @xmath7 the dielectric constant , and @xmath8 denotes the reduced planck constant . for each species
, @xmath9 depends on @xmath4 ( via the bohr radius @xmath10 ) and the particle number density @xmath11 ( @xmath12 is the average inter - particle distance and @xmath13 the dimensionality of the system ) .
when @xmath0-@xmath1 bound state ( exciton ) formation takes place , the residual ( screened ) interaction between these almost neutral complexes weakens , i.e. @xmath14 is reduced , up to the point where a crystallization into a wigner solid is prevented .
generally the formation of few - particle bound states is limited to low temperatures .
their correct theoretical description calls for a full quantum - mechanical treatment , in particular if temperature and density effects , as well as their stability against quantum fluctuations is addressed for low - dimensional systems .
this provides the motivation of the present work : after giving a brief overview on existing results on e - h bilayers , we perform a first - principle numerical analysis of thermodynamically stable phases of a two - component coulomb plasma in the bilayer geometry with spatially separated electrons and holes , and map out major regions of the phase diagram .
thereby the focus is on the inter - layer and intra - layer particle correlations , leading to bound state formation and charge ordering , respectively .
the main advantage of a bilayer system is the easy control of the coupling between the layers via the layer separation @xmath15 .
actual realizations range from electron - ion quantum plasmas @xcite to electron - hole systems in semiconductors @xcite . in the latter spatially separated electrons and holes
can be realized in double quantum well heterostructures @xcite .
recent advances in fabrication technology ( lithography techniques and molecular beam epitaxy ) allow for a good control of the electron / hole density and the spatial inter - well distance .
both classical and quantum electron - hole bilayers have been a subject of active research during the last decades @xcite .
_ classical bilayers _ ( corresponding to high temperature and low density ) have been investigated by molecular dynamics simulations to reveal their static and dynamical properties @xcite .
the phase diagram has been explored as a function of the layer separation @xmath16 , with the principle result that for @xmath17 the inter - layer correlations stabilize the wigner crystals in both layers ( the melting temperature is increased ) compared to a single layer wigner crystal . for _ quantum bilayers _
, at zero temperature , the static and dynamic dielectric response has been analyzed in refs .
the presence of a second ( oppositely charged ) layer was found to enlarge the region where a crystalline phase exists in the @xmath18 plane to higher densities , @xmath19 , or @xmath20 ( @xmath21 denotes the brueckner parameter ) .
the critical value of the parameter @xmath22 , characterizing the critical particle density , changes from @xmath23 for independent layers @xcite to @xmath24 for a bilayer .
quantum melting of a coulomb crystal is caused by the zero - point motion of the particles around their lattice sites .
similar to the classical case , these fluctuations are suppressed by attraction to the particles residing in the second layer , in particluar if they also form a lattice . in this way
, a stack of two wigner lattices appears to be more stable then a single layer crystal .
the second component plays also a prominant role when electron - hole bound states are formed , e.g. for systems of spatially indirect excitons .
then the quantum bilayer is similar to a 2d system of dipoles .
the transition from coulomb ( dominant at @xmath25 ) to dipolar ( @xmath26 ) correlations shifts the critical density of the solid - gas phase transition to a lower value @xcite , which scales with the dipole moment ( or the layer separation ) as @xmath27 . in the dipole approximation ( @xmath28 )
, the bilayer phase diagram should be similar to that for bosonic dipoles in 2d and includes superfluid and normal gases @xcite as well as a dipolar solid , if the composite particles obey bose statistics .
a more accurate treatment that accounts for the internal structure of bound states and allows to release the restriction @xmath28 , was recently performed , see ref .
there it was shown that the a realistic finite layer width restores the @xmath29-coulomb scaling at small distances which makes an exciton solid unstable at high densities similar to the case of a 2d wigner crystal .
besides macroscopic systems , _ mesoscopic bilayers _ with full account of fermi statistics have been studied recently @xcite .
such systems are of particular relevance for experiments with carriers confined by an external trap potential , e.g. electron - hole systems in coupled quantum dots , where both the distance between the dots and the in - plane carrier density are tunable parameters . in contrast to bulk systems , in a mesoscopic bilayer
, a change of @xmath15 leads to a sequence of structural transitions and shell structure formation @xcite similar to complex plasmas @xcite . another important aspect , which has not been studied in detail so far , concerns the influence of the mass - asymmetry @xmath30 between the bilayer charge carriers . a mass ratio @xmath31 is expected to have a significant effect on the phase diagram , as the zero - point energies of light electrons and heavy holes differ , which in turn will influence inter - particle correlations @xcite . depending on the particle density , the light component ( electrons ) can be in a homogeneous fermi gas phase , while the heavier component ( holes ) can form a wigner lattice when the mass ratio exceeds a critical value , @xmath32 ( with @xmath33 for typical electron densities in semiconductors which is about a factor @xmath34 lower than in a @xmath35 plasma @xcite ) .
the critical value @xmath36 can be tuned by the layer separation @xmath15 @xcite .
this is in favor of wigner crystallization ( at least for a heavy component ) in a two - component coulomb system .
in contrast , in a 2d one - component system ( electron gas ) crystalliztation ( which is expected to appear for @xmath37 @xcite ) could not yet be observed since the typical particle densities are too high ( @xmath38 ) , and quantum fluctuations destroy any spatial ordering . in the following , we present a more systematic analysis of _ macroscopic bilayers _ , by considering a wider range of densities / inter - layer separations and mass ratios .
we identify different structural phases by analyzing the pair distribution functions and reconstruct the phase diagram .
for this purpose we adopt the path integral monte carlo ( pimc ) technique for simulations of spatially separated electron - hole systems @xcite .
we note that quantum effects between electrons and holes are fully taken into account by the pimc formalism based on the stochastic sampling of the @xmath39-particle density matrix .
a wide class of electron - hole bilayers can be described by the hamiltonian @xmath40 the last term in eq .
( [ eq1 ] ) mimics the coupling between the electrons and holes residing in different layers .
obviously , the ground - state and thermodynamic properties of the system depend on the following experimentally accessible and managable parameters : ( i ) the hole - to - electron mass ratio @xmath41 , ( ii ) the inter - layer separation @xmath42 , ( iii ) the in - plane particle density @xmath43 , and ( iv ) the temperature @xmath44 ( measured in hartree : @xmath45 ) . in the simulations we considered a bilayer system with @xmath46 electrons and holes .
the particle species are confined to their respective two - dimensional layer of zero width , separated by @xmath15 . to simulate an infinite system we use a rectangular simulation cell with periodic boundary conditions .
the dimensions of the cell with @xmath47 and the number of particles were chosen to fit the symmetry of a hexagonal lattice , once a wigner crystal is formed .
the temperature was kept fixed to @xmath48 .
the other parameters ( @xmath49 ) were varied , each in the range @xmath50 .
more details regarding the setup of the pimc simulations can be found in ref .
@xcite . in order to identify different structural phases we exploit the pair distribution function between different species ( @xmath51 ) @xmath52 in particular
, the formation of a wigner solid in the electron ( hole ) layer is ascribed to a jump in the ratio of the first maximum and the next minimum in @xmath53 , i.e. @xmath54\,\ , / \min\limits_{r < 2r_c}[g_{\alpha\alpha}(r ) ] \ , , \label{delta}\end{aligned}\ ] ] where @xmath55 $ ] is the radius of the first correlation shell .
this criterion has been found sensitive to a solid - liquid / gas phase transition @xcite .
furthermore , to identify phases composed of excitons we use the fraction of bound electron - hole states , which can be estimated by the inter - layer pair distribution function @xmath56 ( for more details see @xcite ) , @xmath57 \,\text{d}r } { \displaystyle \int_0^{r_b } r g_\text{eh } \,\text{d}r } \ , .
\label{bsf}\end{aligned}\ ] ] here @xmath58 is the second zero of @xmath59 $ ] on the right - hand side of the exciton peak located near @xmath60 .
[ fig : snapshots ] presents our simulation data for the pair distribution functions ( left panels ) and corresponding density `` snapshots '' ( right panels ) for characteristic configurations of the electron - hole bilayer system .
, black lines ) , electron - electron ( @xmath61 , red lines ) and electron - hole ( @xmath62 , green lines ) pair distribution functions ( averaged over @xmath63 monte carlo steps ) . _ middle / right columns _ : snapshots of the instantaneous particle density ( averaged over @xmath64 mc steps ) in the electron and hole layer , respectively .
varied parameters are mass ratio @xmath30 , inter - layer distance @xmath15 , and the brueckner parameter @xmath22 .
the three upper rows ( from top to bottom ) typify the excitonic phase : a ) a classical exciton gas , b ) a degenerate exciton gas and c ) an exciton crystal .
the last two rows ( no e - h bound states ) mark d ) a hole wigner crystal and e ) an electron - hole plasma .
all pimc simulations were performed with @xmath65 particles in each layer .
the temperature was kept fixed at @xmath66 . ] , black lines ) , electron - electron ( @xmath61 , red lines ) and electron - hole ( @xmath62 , green lines ) pair distribution functions ( averaged over @xmath63 monte carlo steps ) . _ middle / right columns _ : snapshots of the instantaneous particle density ( averaged over @xmath64 mc steps ) in the electron and hole layer , respectively .
varied parameters are mass ratio @xmath30 , inter - layer distance @xmath15 , and the brueckner parameter @xmath22 .
the three upper rows ( from top to bottom ) typify the excitonic phase : a ) a classical exciton gas , b ) a degenerate exciton gas and c ) an exciton crystal .
the last two rows ( no e - h bound states ) mark d ) a hole wigner crystal and e ) an electron - hole plasma .
all pimc simulations were performed with @xmath65 particles in each layer .
the temperature was kept fixed at @xmath66 . ] , black lines ) , electron - electron ( @xmath61 , red lines ) and electron - hole ( @xmath62 , green lines ) pair distribution functions ( averaged over @xmath63 monte carlo steps ) . _ middle / right columns _ : snapshots of the instantaneous particle density ( averaged over @xmath64 mc steps ) in the electron and hole layer , respectively .
varied parameters are mass ratio @xmath30 , inter - layer distance @xmath15 , and the brueckner parameter @xmath22 .
the three upper rows ( from top to bottom ) typify the excitonic phase : a ) a classical exciton gas , b ) a degenerate exciton gas and c ) an exciton crystal .
the last two rows ( no e - h bound states ) mark d ) a hole wigner crystal and e ) an electron - hole plasma .
all pimc simulations were performed with @xmath65 particles in each layer .
the temperature was kept fixed at @xmath66 . ] , black lines ) , electron - electron ( @xmath61 , red lines ) and electron - hole ( @xmath62 , green lines ) pair distribution functions ( averaged over @xmath63 monte carlo steps ) . _ middle / right columns _ : snapshots of the instantaneous particle density ( averaged over @xmath64 mc steps ) in the electron and hole layer , respectively .
varied parameters are mass ratio @xmath30 , inter - layer distance @xmath15 , and the brueckner parameter @xmath22 . the three upper rows ( from top to bottom ) typify the excitonic phase : a ) a classical exciton gas , b ) a degenerate exciton gas and c ) an exciton crystal .
the last two rows ( no e - h bound states ) mark d ) a hole wigner crystal and e ) an electron - hole plasma .
all pimc simulations were performed with @xmath65 particles in each layer .
the temperature was kept fixed at @xmath66 . ] , black lines ) , electron - electron ( @xmath61 , red lines ) and electron - hole ( @xmath62 , green lines ) pair distribution functions ( averaged over @xmath63 monte carlo steps ) . _ middle / right columns _ : snapshots of the instantaneous particle density ( averaged over @xmath64 mc steps ) in the electron and hole layer , respectively .
varied parameters are mass ratio @xmath30 , inter - layer distance @xmath15 , and the brueckner parameter @xmath22 .
the three upper rows ( from top to bottom ) typify the excitonic phase : a ) a classical exciton gas , b ) a degenerate exciton gas and c ) an exciton crystal .
the last two rows ( no e - h bound states ) mark d ) a hole wigner crystal and e ) an electron - hole plasma .
all pimc simulations were performed with @xmath65 particles in each layer .
the temperature was kept fixed at @xmath66 . ]
_ classical exciton gas . _ fig .
[ fig : snapshots]a exemplifies the excitonic gas phase as realized for a mass - asymmetric bilayer with @xmath67 and an inter - layer distance much smaller compared to the intra - layer particle distance ( @xmath68 ) .
therefore we are in a regime where the dipole approximation is valid .
indeed , @xmath69 is sharply peaked at @xmath60 reflecting the formation of bound states ( indirect excitons ) .
a comparison of the snapshots of the ( instantaneous ) particle density in each layer shows that all particles are bound pairwise ( cf .
the uppermost two right - hand panels ) .
the structureless form of @xmath53 shows that the excitons interact only weakly , i.e. , they move almost freely and form a gaseous phase . expressed in terms of the dipole coupling parameter @xmath70 ( see ref .
@xcite ) , this regime is characterized by @xmath71 , and we expect a dipolar gas with moderate correlations and a superfluid transition temperature @xmath72 \equiv k_b t_c / e_0\sim 1.25 $ ] ( cf . the @xmath70-dependence of @xmath73 in ref .
@xcite scaled in units @xmath74 ) . since @xmath73 is by factor @xmath75 smaller than the temperature simulated in fig .
[ fig : snapshots]a , where @xmath76\equiv
k_b t / e_0=(k_b t/{\rm ha } ) ( m+1 ) \pi r_s^2\sim 115 $ ] , the exciton gas is essentailly _ classical _ , and effects of bose / fermi statistics are irrelevant .
the situation does not change much if @xmath77 , because the degeneracy of the exciton gas remains low . _
degenerate exciton gas / liquid .
_ we now consider the mass - symmetric case and increase the particle density ( reduce @xmath22 ) , see fig .
[ fig : snapshots]b .
since for @xmath78 and @xmath79 the condition @xmath80 still holds , the excitons still behave as dipoles .
this is supported by an analysis of the microscopic effective exciton - exciton interaction potential @xcite .
for the parameters used in fig .
[ fig : snapshots]b , we have @xmath81 and @xmath72\sim 1.3 $ ] .
the simulated temperature , @xmath76\sim 0.84 $ ] , is now below @xmath73 .
accordingly the exciton gas is superfluid . in the quasi-2d bilayer excitonic
system the superfluid transition will be of the berezinskii - kosterlitz - thouless type @xcite and is associated with the emergence of topological order , i.e. the pairing of vortices with opposite circulation . the correlations of the order parameter ( for example , the one - particle density matrix ) decay algebraically in space . a true long - range order is recovered only at @xmath82 .
this scenario is applicable to a wide variety of quasi two - dimensional systems .
recent experimental realizations of degenerate dipolar atomic gases @xcite , bosonic molecules @xcite , and indirect excitons in quantum wells @xcite have stimulated theoretical analyses of their quantum many - body properties .
the present bilayer system provides an interesting test case . in our system , the onset of degeneracy and spatial coherence in both layers ( which leads to the superfluid response ) is reflected in the behavior of the pair distribution functions .
the incipient overlap of the excitonic wavefunctions ( around @xmath83 ) can be noted from the particle density plots ( right - hand panels of fig .
[ fig : snapshots]b ) .
simultaneously , we observe the build - up of short - range correlations which is indicated by the few ( weak ) oscillations in @xmath53 .
the density snapshots point to a local hexagonal ordering : though noticeable spatial correlations exist between nearest neighbor and even next - nearest neighbor excitons , true long - range order does not appear .
such behavior is typical for correlated liquids with repulsive interactions . finally , we note , that the correlations not only influence the static structural properties , but they significantly modify also the quasi - particle excitation spectrum ( leading to the formation of a phonon - maxon - roton branch ) .
the latter has a strong effect on the superfluid transition temperature ( and stability of the superfluid phase to thermal fluctuations ) , in particular , in 2d dipolar systems @xcite . _ excitonic crystal .
_ an interesting situation appears when the inter - layer and intra - layer correlations are of the same order .
if @xmath84 , the bilayer system does not behave as a pure dipolar system .
for the parameters used in fig .
[ fig : snapshots]c we get @xmath85 and @xmath76=4.6 $ ] , i.e. , we are in the classical domain of the phase diagram determined in ref .
@xcite , and far away from the quantum crystallization transition observed at @xmath86 @xcite . in this case , a crude estimate of the melting temperature of a dipolar crystal follows from the classical dipole coupling parameter @xmath87
@xcite , yielding @xmath88 .
this means the simulated temperature ( @xmath89 ) is @xmath90 times below the melting temperature and , indeed , we observe an almost perfect hexagonal lattice in both layers .
obviously the bilayer geometry plays a crucial role , and we directly observe the stabilizing effect of the second layer mentioned in the introduction : a single electron layer would form a wigner crystal only at @xmath91 @xcite , whereas wigner crystal formation in a bilayer sets in at significantly higher densities , e.g. at about @xmath92 for the mass - symmetric case @xmath77 , see refs . @xcite . in qualitative agreement ,
our data for a bilayer with heavy holes yield an exciton crystal already for @xmath93 .
this state is characterized by strongly oscillating pair distribution functions in both layers . the observed difference in the peak heights of @xmath94 and @xmath61 results from the mass - asymmetry .
since the lighter electrons have a larger zero - point energy , they are not as localized as the holes around the hexagonal lattice sites
. furthermore we can test the appearance of the present bilayer exciton solid against the predictions of the exciton approximation @xcite ( here the analysis has been performed for a mass ratio @xmath95 , which corresponds to the paramaters of znse - based quantum wells ) .
the excitonic phase diagram derived in ref .
@xcite predicts that such kind of exciton solid can exist only in a finite density interval , which depends on the exciton dipole moment ( and on the inter - layer separation , in case of a bilayer ) . for @xmath96 , the density simulated in fig .
[ fig : snapshots]c ( @xmath79 ) fits in the exciton solid density window , @xmath97 , estimated from fig . 7 in @xcite . _ hole crystal .
_ we now turn to the case where electrons and holes are weakly correlated .
this happens at high in - plane particle densities ( and large layer distances ) when the coulomb attraction between electrons and holes is screened to a large extent .
then the correlation energy within the layer , @xmath98 , overcomes the inter - layer binding energy , @xmath99 . as a result
the electron - hole pairs dissociate and the spatially separated electron and hole subsystems undergo structural phase transitions essentially independently . for the strongly mass - asymmetric case ,
zero - point fluctuations destroy any ordering in the layer containing the light electrons , while the heavy holes can form a wigner lattice , see the results for @xmath100 in fig .
[ fig : snapshots]d . in the hole layer ,
the one - component quantum parameter scales with the particle density as @xmath101 , leading to @xmath102 well above @xmath23
formation of the hole crystal does not necessarily go along with a significant modulation of the electron density .
this can be seen from the density snapshot but also from the behavior of the e - h pair distribution function , which stays practically constant ( @xmath103 ) for all distances .
_ electron - hole plasma_. the hole crystal melts , of course , if we lower the mass of the holes . for the mass - symmetric case shown in fig .
[ fig : snapshots]e , a structureless electron ( hole ) plasma emerges in both layers .
plane for mass ratios @xmath100 ( left column ) , @xmath67 ( middle ) , and @xmath77 ( right ) .
the upper ( lower ) row gives data for the electron ( hole ) layer .
symbols denote different thermodynamic phases : crystal ( squares ) , degenerate gas ( triangles ) and classical gas ( circles ) . bound states ( indirect excitons )
are marked by filled symbols .
the superfluid phase for 2d dipole systems is located just below the blue curve ( superfluid - normal phase boundary ) and left of the green line ( superfluid - dipole crystal boundary , corresponding to @xmath104 ) as taken from ref .
the degenerate exciton gas , shown as filled triangles , is in the superfluid state .
the melting transition boundary of the classical dipole phase is mareked by the red line .
the dotted line is a guide for the eye showing the @xmath18 diagonal . ]
after having qualitatively characterized the various phases found in the electron - hole bilayer system , we now construct the phase diagram in the @xmath18 plane , see fig .
[ fig : phasediagram ] .
the identification of the different states and configurations is based on the max - min ratio @xmath105 and the bound state fraction @xmath106 given by eq .
( [ delta ] ) and ( [ bsf ] ) , respectively .
the wigner crystal phase is found above a critical value @xmath107 and the excitonic phase was specified for @xmath108 .
of course , these specified critical values yield only an estimate of the exact phase boundary in the thermodynamic limit .
the first general observation from the phase diagram is that , independent of the mass ratio @xmath30 , excitonic bound states exist only for @xmath109 and @xmath110 .
these states are marked in fig .
[ fig : phasediagram ] by filled symbols .
a similar observation was reported in ref .
. _ large mass - asymmetry . _ for @xmath100 and @xmath111
we observe some of the exciton phases discussed in the previous section ( cf .
the left most column in fig .
[ fig : phasediagram ] ) . in the neighborhood of @xmath112 a bilayer exciton crystal
is formed ( filled squares ) . by lowering the density ( @xmath113 ) the exciton crystal vanishes , as predicted by the dipole approximation valid in this regime .
once the inter - particle spacing @xmath12 is increased , the dipolar potential energy @xmath114 will decay faster than the zero - point kinetic energy @xmath115 .
hence , at some point quantum fluctuations will destroy any spatial ordering and initiate a phase transition to a non - degenerate ( classical ) exciton gas ( filled circles ) , far away from the diagonal @xmath112 ( recall that large @xmath22 implies a dilute , weakly interacting system ) . below the diagonal ( @xmath116 ) , the excitons dissociate . as a result
the two layers are weakly coupled and exhibit essentially independent structural phases . for the scanned range of @xmath22-values , @xmath117 ,
the wigner crystal in the hole layer is the thermodynamically stable phase , due to the large hole mass .
by contrast , the electron layer at high densities ( small @xmath22 ) forms a degenerate gaseous phase up to @xmath118 . here an electron wigner crystal is not observed , because the two points we have simulated ( @xmath79 and @xmath113 ) are either below or well above the critical density @xmath23 @xcite . for @xmath113 ,
the simulated temperature @xmath89 is not low enough to see coulomb crystallization of a classical electron gas .
the corresponding classical coulomb coupling parameter @xmath119 is well below the critical value @xmath120 , required for crystallization of a classical electron gas .
_ moderate mass - asymmetry .
_ if we decrease the mass ratio to @xmath67 the stability of the excitonic phase is not much affected ( cf .
the middle panels of fig .
[ fig : phasediagram ] ) . for @xmath121 and @xmath122
their phase - space domain remains almost the same . at @xmath123 a bilayer exciton crystal forms
( see also fig .
[ fig : snapshots]c ) . here
we note the stabilizing effect of the lattice in the hole layer . due to the coupling of the electrons to the holes arranged in a regular pattern ,
the electron wigner crystal is observed significantly below the corresponding critical value @xmath23 ( above the density ) of a single layer .
obviously , the hole lattice produces the same pinning effect as impurities studied in ref .
@xcite . in the latter case , in - plane impurities
enhance the electron wigner crystallization , i.e. , they shift the transition density to @xmath124 .
the same scenario applies to bilayers with large inter - layer separation , provided the screening effect produced by the electrons on the hole - hole interaction can be neglected .
our simulation show that at higher electron densities ( @xmath125 ) zero - point fluctuations may overcome the pinning effect of the hole lattice and the electron wigner crystal melts on the diagonal @xmath126 ( see middle column of fig .
[ fig : phasediagram ] , @xmath127 ) . at an inter - layer separation @xmath128 the melting of the hole lattice
is accompanied by the formation of excitonic bound states .
that is to say we observe a destabilization of the hole wigner lattice by the electron layer due to screening effects . _
mass - symmetric bilayer .
_ in the case @xmath77 , obviously , all phase transitions are identical in both layers . for @xmath129 and @xmath130
we find an electron - hole plasma , otherwise a ( non - degenerate or degenerate ) excitonic gas phase , see right panels of fig . [ fig : phasediagram ] .
note , that for the layer separation @xmath131 , exciton bound states are stable for @xmath132 , in agreement with zero - temperature diffusion monte carlo results @xcite .
we can not confirm , however , the two - component plasma phase boundary reported in ref .
@xcite , simply because the corresponding region @xmath133 was not accessible by our pimc simulations so far . as noted before , the simulated temperature is not sufficiently low to observe a single - layer wigner crystal at @xmath134 and @xmath135 .
a similar argument applies to the exciton crystal , which should exist for layer separations above a critical value @xmath136 , where @xmath137 for @xmath95 @xcite . in the vicinity of the excitonic mott transition more refined simulations , including an accurate analysis of the fraction of ionized and bound states and a full account of the fermi statistics @xcite of electrons and holes would be required .
in this report , we analyzed the structure and stability of possible phases of a neutral electron - hole bilayer in a wide range of the system s control parameters : layer separation , particle density , and mass ratio at a fixed low temperature . to this end
we performed large - scale unbiased path integral monte carlo simulations .
the properties of the observed exciton gas , exciton crystal , hole crystal and electron - hole plasma phases are characterized by pair distribution functions and snapshots of the instantaneous particle densities . for a representative set of mass ratios , @xmath77 , @xmath138 , and @xmath139 ,
the phase diagram was mapped out .
it shows the complex interplay between the intra - layer and inter - layer correlations as well as stabilization and destabilization effects of the second layer on the wigner crystal in a 2d coulomb system
. currently , extensive simulations are underway that use a denser grid of simulation parameters that will allow to draw the phase boundaries with higher accuracy .
our results are important for strongly correlated two - component charged particle systems in general @xcite and for complex plasmas in particular @xcite .
common to all these systems are collective phenomena resulting in a rich variety of phases
. of particular interest is the possibility to control the interparticle interaction in e - h bilayers by changing external parameters ( inter - layer separation , density ) .
these concepts are also of interest for complex plasmas and for charged particles in contact with surfaces @xcite where similar pinning and crystal stabilization effects occur due to the presence of surface defects @xcite .
this work was funded by the deutsche forschungsgemeinschaft , sfb tr-24 , projects a5 and a7 .
numerical simulations were performed on the tera - flop pc - cluster at the institute of physics , university greifswald .
j. schleede acknowledges the hospitality of the institute for theoretical physics and astrophysics , christian - albrechts - universitt zu kiel .
l. butov _ et al .
_ , nature * 417 * , 47 ( 2002 ) ; hammack _ et al .
* 96 * , 227402 ( 2006 ) d. snoke _ et al .
_ , nature * 418 * , 754 ( 2002 ) ; r. rapaport _ et al .
* 92 * , 117405 ( 2004 ) ; d. snoke _ et al .
_ , solid state comm . * 134 * , 37 ( 2005 ) .
a.v . gorbunov and v.b .
timofeev , jetp lett . * 84 * , 329 ( 2006 ) ; v.b .
timofeev and a.v .
gorbunov , j. appl . phys .
* 101 * , 081708 ( 2007 ) ; v.b .
timofeev , a.v .
gorbunov , and a.v .
larionov , j. phys . : condens .
matter * 19 * , 295209 ( 2007 ) .
v. s. filinov , h. fehske , m. bonitz , v. e. fortov and p. r. levashov , proton crystallization in a dense hydrogen plasma contrib . plasma phys .
* 52 * , 224 ( 2012 ) ; v. s. filinov , m. bonitz , h. fehske , p. r. levashov and v. e. fortov , structures of quantum 2d electron - hole plasmas j. phys . a : math .
general * 42 * , 214014 ( 2009 ) ; v. s. filinov , h. fehske , m. bonitz , v. e. fortov and p. levashov , ordered structure formation in 2d mass asymmetric electron - hole plasmas phys .
a * 372 * , 5208 ( 2008 ) .
v. s. filinov , m. bonitz , v. e. fortov , h. fehske , and p. r. levashov , coulomb crystal and quantum melting in electron - hole plasmas of semiconductors under high pressure phys .
( b ) * 244 * , 474 ( 2007 ) ; m. bonitz , v. s. filinov , v. e. fortov , p. r. levashov , and h. fehske , hole crystallization in semiconductors j. phys .
a * 39 * , 4717 ( 2006 ) .
p. ludwig , a. filinov , y.e .
lozovik , h. stolz , and m. bonitz , contrib .
plasma phys .
* 47*(4 - 5 ) , 335 ( 2007 ) .
p. hartmann , z. donko , and g. j. kalman , europhys .
72 * , 396 ( 2005 ) .
s. ranganathan , r.e .
johnson , phys .
b * 75 * , 155314 ( 2007 ) . j. szymanski , l. swierkowski and d. neilson , phys . rev .
b * 50 * , 11002 ( 1994 ) .
l. liu and l. swierkowski , d. neilson , j. szymanski , phys . rev .
b * 53 * , 7923 ( 1996 ) . r.k .
moudgil , gaetano senatore , l.k .
saini , phys .
b * 66 * , 205316 ( 2002 ) .
b. tanatar and d.m .
ceperley , physical review b * 39 * , 5005 ( 1989 ) . c. mora , o. parcollet , and x. waintal , phys .
b * 76 * , 064511 ( 2007 ) ; g.e .
astrakharchik , j. boronat , i.l .
kurbakov , and yu.e .
lozovik , phys .
lett . * 98 * , 060405 ( 2007 ) ; h.p .
bchler , e. demler , m. lukin , a. micheli , n. prokofev , g. pupillo , p. zoller , phys .
* 98 * , 060404 ( 2007 ) a. filinov , n. prokofev , and m. bonitz , phys .
lett . * 105 * , 070401 ( 2010 ) .
j. bning , a. filinov , and m. bonitz , phys .
b * 84 * , 075130 ( 2011 ) .
a. filinov , m. bonitz , p. ludwig , and yu.e .
lozovik , phys .
( c ) * 3 * issue 7 , 2457 ( 2006 ) ; p. ludwig , k. balzer , a. filinov , h. stolz , m. bonitz , new journal of physics * 10 * , 083031 ( 2008 ) .
j. werner , a. griesmaier , s. hensler , j. stuhler , t. pfau , a. simoni , and e. tiesinga , phys .
lett . * 94 * , 183201 ( 2005 ) ; j. stuhler , a. griesmaier , t. koch , m. fattori , t. pfau , s. giovanazzi , p. pedri , and l. santos , phys .
lett . * 95 * , 150406 ( 2005 ) ; t. lahaye , c. menotti , l. santos , m. lewenstein and t. pfau , rep .
. phys . * 72 * , 126401 ( 2009 ) .
sage at al .
, phys . rev . *
94 * , 203001 ( 2005 ) ; d. wang at al .
, phys . rev . *
93 * , 243005 ( 2004 ) ; j.g .
danzl _ et al .
_ , science * 321 * , 1062 ( 2008 ) ; k. aikawa _ et al .
_ , new journal of physics * 11 * , 055035 ( 2009 ) . | we investigate exciton bound - state formation and crystallization effects in two - dimensional electron - hole bilayers . performing unbiased path integral monte carlo simulations all quantum and coulomb correlation effects
are treated on first principles .
we analyze diverse pair distribution functions in dependence on the layer separation , particle density and hole - to - electron mass ratio and derive a schematic phase diagram for the neutral mass - asymmetric bilayer system .
our simulations reveal a great variety of possible phases namely an exciton gas , an exciton crystal , an electron - hole liquid and a hole crystal embedded in an electron gas . |
the interplay between thermodynamic properties and quantum behavior has been studied through several models of quantum heat engines @xcite .
even though quantum systems show non - classical features like entanglement , coherence and so on , these models are often found to be consistent with thermodynamic interpretations @xcite . on the other hand ,
some models of these thermal machines have been reported to show an unexpected behavior such as extraction of work from a single heat bath @xcite , exceeding carnot efficiency @xcite and cooling to absolute zero @xcite . in this paper , we focus on the interesting phenomenon of intrinsic friction in quantum engines @xcite .
this effect arises due to non - commutativity of the internal and the external part of the hamiltonian leading to non - commutativity of the hamiltonians at different times .
further to reduce friction , an effect called quantum lubrication has also been proposed @xcite . in order to better understand intrinsic friction and its relevance for analysis of dissipation in quantum systems
, it seems interesting to look for this effect in other similar models . in the quantum heat engine that we discuss below , the working medium ( system )
consists of two spin - half particles with heisenberg interaction , kept in an external magnetic field .
the system is driven by selectively changing the external field in finite time , such that the field on either spin is different ( inhomogenous ) . in this case , we observe the frictional effect .
but if the field on both spins remains homogeneous , then friction is absent . as expected , we find that if the driving that creates inhomogeneity of fields on the spins , is performed very slowly , the friction effect is again absent .
an analogous system under an inhomogeneous magnetic field plays important role in quantum computing @xcite
. thermal entanglement of such spin system @xcite and similar models @xcite have also been studied .
we consider a heat cycle analogous to otto cycle , consisting of two adiabatic branches and two thermalization branches . on the thermalization branches ,
the magnetic fields on both spins are kept constant . to simplify our model , we consider the latter branches to take sufficiently long time so that the system attains equilibrium with the bath at the end of the process .
then the system is decoupled from the bath and a thermodynamically adiabatic process is carried out on the system .
so the initial state of the system before the adiabatic process is a thermal state , diagonal in the eigenbasis of the hamiltonian .
this will help us to understand the coherence developed in the system during the adiabatic process @xcite .
the development of coherence leads to an increase in the entropy of the system which is a signature of the friction observed in our model .
the paper is organised as follows . in section
ii , we introduce the model of the quantum heat engine . entropy production in the adiabatic branches of the cycle is discussed in section iii .
section iv is devoted for understanding the work extraction in our model . here
we discuss lower and upper bound of the work that can be extracted .
section v is devoted to discussion .
we analyse the cycle using numerical simulations by alloting finite time to the adiabatic branches and conclude with a summary and future directions .
we consider two spin - half particles with isotropic exchange interaction , as the working substance for a quantum otto cycle . in general , the hamiltonian is written as @xmath0 , where @xmath1 is the external hamiltonian which can be controlled and @xmath2 is the internal hamiltonian . in our model , we control in time , the magnetic field applied to particle labeled 2 .
so we have @xmath3 @xmath4 where @xmath5 are the pauli matrices , @xmath6 is the isotropic exchange constant and @xmath7 , @xmath8 are the magnetic fields applied along @xmath9-axis to the first and the second spin respectively .
so the magnetic field applied to the individual spins are not always equal during the adiabatic branch which results in @xmath10\neq0 $ ] .
this non - commutativity of the external and the internal hamiltonian when leading to non - commutativity of the hamiltonian at different times , is the cause of internal friction in our model @xcite . as a special case ,
we show in section v that the non - commutative property of external and the internal hamiltonian by itself is not a sufficent condition for friction .
now we analyse the system with inhomogeneous magnetic field in more detail . in this case
the eigenbasis of the hamiltonian is @xmath11 \{@xmath12 , @xmath13 , @xmath14 , @xmath15 } , where @xmath12 and @xmath14 are given by @xmath16 and @xmath17 respectively and \{@xmath13 , @xmath18 , @xmath19 , @xmath15 } forms the computational basis . here
@xmath20 and @xmath21 , where @xmath22 and @xmath23 .
the corresponding eigenvalues are \{@xmath24 , @xmath25 , @xmath26 , @xmath27 } , where @xmath28 .
the equilibrium density matrix when the system is attached to a bath at temperature @xmath29 , is given by @xmath30 , where @xmath31 is partition function of the system , and we have set boltzmann s constant to unity .
the eigenvalues of @xmath32 , or the occupation probabilities of the energy levels , are given by @xmath33 now we are ready to discuss the quantum heat cycle , which consists of the following four stages : _ stage 1 _ : the coupled - spins system is attached to a cold bath with temperature @xmath34 .
the system attains equilibrium with the bath .
the magnetic field applied to the first and second spins are identical ( @xmath35 ) .
the density matrix is diagonal in the hamiltonian s eigenbasis . because of the homogeneous magnetic field , the eigenstates @xmath12 and @xmath14 are maximally entangled bell states , with @xmath36 .
the occupation probability @xmath37 for the state with energy eigenvalue @xmath38 is calculated from eq .
( [ probability ] ) by setting @xmath39 and @xmath40 .
so the mean energy at the end of the first stage is tr@xmath41 .
_ stage 2 _ : in this stage , the system is isolated from the bath and it can exchange only work with the surroundings .
the magnetic field applied to the second spin is changed from @xmath42 to @xmath43 in finite time and the system may undergo a non - adiabatic evolution . by non - adiabatic evolution ,
we mean that the system may be driven fast enough so that the quantum adiabatic theorem does not hold @xcite .
the density matrix undergoes a unitary evolution .
the eigenstates of @xmath44 are also time dependent . in general , the eigenstates of @xmath45 are not the same as @xmath44 . in the infinitely slow limit ( @xmath46 ) , the adiabatic theorem holds and eigenstates of the density matrix are identical to the eigenvectors of the instantaneous hamiltonian .
so in case of fast driving , the final state of the system may not be diagonal in the eigenbasis of the final hamiltonian .
when we project the final density matrix onto the eigenbasis of the hamiltonian , the corresponding occupation probability of the eigenstate of the hamiltonian with eigenvalue @xmath47 is given as @xmath48 , where @xmath49 is the eigenvector of the final hamiltonian .
a pictorial representation is shown in fig .
[ a1 ] . at the end of the second stage , the mean
energy can be written as tr@xmath50 .
the difference of the initial and the final mean energy during the adiabatic process is equal to the work performed : @xmath51 .
_ stage 3 _ : the system under inhomogeneous magnetic field is attached to a hot bath with temperature @xmath52 and it attains equilibrium by absorbing heat from the bath .
the occupation probabilities ( @xmath53 ) are calculated from eq .
( [ probability ] ) by putting @xmath54 and @xmath55 . at the end of the third stage ,
the system is in a thermal state with mean energy @xmath56 .
_ stage 4 _ : the system again undergoes a unitary evolution by a change of the magnetic field of the second spin from @xmath57 to @xmath7 , whereby the energy levels change from @xmath47 back to @xmath58 .
the occupation probabilities @xmath59 in the eigenstates of the hamiltonian are calculated by projecting the density matrix onto the eigenbasis of the hamiltonian .
so the mean energy at the end of the process is @xmath60 .
the difference in the mean energy due to this process is @xmath61 . to close the cycle ,
the system is again brought in contact with cold bath . the system releases on average an amount of heat to cold bath .
as we show below , @xmath62 and @xmath63 are the work done _ by _ and _ on _ the system , respectively .
represent the populations in the energy eigenbasis @xmath64 . in the infinite time limit
, we get @xmath65 and the eigenstates of the density matrix are same as that of the hamiltonian . ]
now we analyse the irreversibility associated with the adiabatic branch by quantifying the entropy production .
the adiabatic process is represented by a unitary process so that after time @xmath66 , the system - state evolves to @xmath67 , where @xmath68 .
the von neumann entropy @xmath69 remains constant throughout the process .
but energy - entropy @xmath70 , defined with the occupational probabilities of the energy levels , changes .
@xmath70 in the initial state is given by @xmath71 , where @xmath72 .
since the initial state is a thermal state , we have @xmath73 .
but after the finite - time adiabatic step , @xmath70 increases where as @xmath69 remains unchanged .
initially , we have @xmath74=0 $ ] .
two of the eigenvectors @xmath75 and @xmath15 of the hamiltonian are not functions of the applied magnetic field and hence are independent of time .
so if the system is in any of these @xmath76 eigenstates , it will remain there during the process .
thus the initial population in theses states remains constant throughout the adiabatic process .
but the eigenvectors @xmath77 and @xmath14 of hamiltonian depend on the magnetic field and hence are time dependent .
so if the system is initially in one of these states , then changing the hamiltonian with a finite rate results in a non - adiabatic evolution . in in other words ,
the final state of the system is then not an eigenstate of the final hamiltonian .
let the eigenvectors of the final hamiltonian be given as @xmath78 , @xmath13 , @xmath79 , @xmath80 and the set of the eigenvectors of the final density matrix is \{@xmath81 , @xmath13 , @xmath82 , @xmath15}. since @xmath81 and @xmath82 are orthogonal to each other as well as to @xmath13 and @xmath15 , we can express the kets @xmath81 and @xmath82 as linear combinations of @xmath83 and @xmath79 as @xmath84 where @xmath85 . now consider a projection of the system - state on the eigenbasis of hamiltonian .
two of the populations remain unchanged such that @xmath86 and @xmath87 .
the occupation probabilities for the eigenstates @xmath88 and @xmath89 are @xmath90 and @xmath91 respectively . now project the density matrix onto the eigenbasis @xmath64 of the final hamiltonian . from eq .
( [ phipsi ] ) , we get the occupation probabilities corresponding to @xmath92 and @xmath93 as @xmath94 due to @xmath95 , we can write @xmath96 as the difference between @xmath97 and @xmath98 gets reduced as compared to the one between @xmath90 and @xmath91 , and recalling that @xmath86 and @xmath87 , the distribution @xmath99 is more uniform than @xmath100 , we have @xmath101 which signifies that the energy - entropy @xmath70 increases in the finite - time adiabatic process . in the infinite time process ( @xmath102 )
, the system undergoes quantum adiabatic evolution and in this limit @xmath70 remains unchanged .
the total entropy production versus the total time allocated to adiabatic branch will be discussed in section v.
the work is performed by or on the system only during the adiabatic branches i.e. in stages 2 and 4 , when the evolution of the system is governed by liouville - von neumann equation ( with @xmath103 ) @xmath104 . \label{liouville}\ ] ] the instantaneous mean energy of the system is given by @xmath105 . differentiating with respect to time we get @xmath106 in general , comparing with the first law of thermodynamics , we identify @xcite the first term on the right hand side as the rate of heat flow ( @xmath107 ) and the second term as the power ( @xmath108 ) . for an adiabatic process , the first term above on the right hand side vanishes due to eq .
( [ liouville ] ) . upon integrating the power
, we get the expression for work as @xmath109 thus the work performed during the adiabatic process lasting for a time interval @xmath66 , is equal to the change in the mean energy of the system , upto time @xmath66 . furthermore , it can be shown that the work done in a infinitely slow process is always higher than the work done in a finite - time process .
thus the lower bound for work extracted is obtained for an extremely fast process ( @xmath110 ) . to evaluate the lower bound , we assume that the density matrix of the system remains unchanged . in case of equilibrium with the cold bath ,
the initial density matrix is given as @xmath111 where @xmath112 and @xmath113 .
since the system is in thermal state , the initial hamiltonian commutes with the density matrix and both have the same set of eigenvectors . in the sudden limit ( @xmath114 )
, the density matrix remains the same as the initial , because @xmath115 .
but the hamiltonian is changed to @xmath116 where @xmath117 and @xmath118 .
now we find the population of the corresponding eigenstates of the hamiltonian by projecting the density matrix onto the eigenbasis of hamiltonian as @xmath119 while @xmath86 and @xmath87 .
similarly for the second adiabatic process where the hamiltonian is returned to its initial stage with eigenbasis @xmath120 , we obtain upon projecting the density matrix @xmath121 for this process , as @xmath122 and @xmath123 and @xmath124 .
now the work extracted in complete cycle ( @xmath125 ) with fast adiabatic processes is given by @xmath126 using the probabilities calculated above for extremely fast ( sudden ) processes , we get the lower bound of work @xmath127 the upper bound for work is obtained for the slow process ( @xmath46 ) . according to quantum adiabatic theorem the system remains in the instantaneous eigenstate of the hamiltonian .
the work expression is in general written as @xmath128 yields the upper bound for the extractable work , @xmath129 these bounds are compared with the finite - time work in fig .
( [ figurework ] ) .
analytic expressions for work can be derived both in the case of a very slow driving and a sudden one . to estimate the finite - time evolution of the system on the adiabatic branch
, we have to integrate the liouville - von neumann equation , eq .
( [ liouville ] ) .
we accomplish this using the fourth - order runge - kutta method @xcite . in the first adiabatic process
, @xmath130 changes from @xmath131 to @xmath57 .
this is modeled by applying a pulse @xmath132 for a time @xmath133 , where @xmath134 is half of the time period .
similarly the second adiabatic process is done by applying a pulse @xmath135 for the same time interval .
thus we allot equal time intervals to both the adiabatic branches .
the total work performed and the total entropy production due to the finite time process are plotted in fig .
( [ figurework ] ) . as discussed in the previous section , the work extracted decreases monotonically with a finite rate of driving .
this is also reflected in the corresponding increase of entropy production in the finite - time case . )
allocated for both the adiabatic processes .
here we use @xmath136 , @xmath137 , @xmath138 , @xmath139 and @xmath140 .
the work is bounded above by @xmath141 ( eq . ( [ wub ] ) , dashed line ) .
the thick horizontal line depicts the lower bound @xmath142 , eq . ( [ wlb ] ) , obtained for a sudden adiabatic process ( @xmath143 ) .
the inset shows total entropy production on the adiabatic branches versus total time ( @xmath134 ) .
as @xmath134 is increased , the total entropy production reduces monotonically to zero and the frictional effect vanishes.,width=294,height=188 ] let us consider a cycle in which the magnetic fields applied to the first and second spins in stage 1 have different values , @xmath144 and @xmath131 respectively . in this case
the internal and external part of the hamiltonian do not commute with each other .
now suppose that during the first adiabatic process , @xmath7 and @xmath145 vary at equal rates so that the difference ( @xmath146 ) keeps constant during the process . as we have seen in section ii ,
the parameters @xmath147 and @xmath148 appearing in the eigenbasis of the hamiltonian , are functions of @xmath6 and @xmath149 .
since @xmath149 remains constant during the adiabatic process , the energy eigenstates of the hamiltonian become time independent which implies that hamiltonians at different times commute with each other and so friction is absent in this case .
using the similar argument , no friction is expected on the second adiabatic process , when the magnetic fields are restored to their initial values ( @xmath150 and @xmath131 ) .
this serves as an example to appreciate that the non - commutative property of the internal and external hamiltonian caused by the inhomogeneous magnetic fields may not always lead to non commutativity of hamiltonian at different times to cause friction .
rather the inhomogeneous driving in which @xmath149 changes with time leads to the non commutative property of the hamiltonian at different times and thereby to frictional effect . to conclude , we have studied a model of quantum heat engine where the inhomogeneous driving at a finite rate , of the components of the quantum working medium leads to a frictional effect .
this effect is characterized by increase in the entropy of the system .
as expected of a thermodynamic system , the entropy production leads to decrease in the work obtained from a cycle . the work is plotted versus the time allotted for the adiabatic branches .
the amount of work that can be obtained from our model is bounded from both above and below .
the upper bound is obtained for a slow process where the frictional effect vanishes and quantum adiabatic theorem holds , while the lower bound is obtained for a sudden process .
some interesting future problems include the study of frictional effect on models with anisotropic interactions and with systems using higher number of spins .
the possibility of quantum lubrication @xcite to reduce the intrinsic friction can also be studied .
gt acknowledges financial support from iiser mohali .
rsj gratefully acknowledges reearch grant from the department of science and technology , india , under project no .
sr / s2/cmp-0047/2010(g ) titled , _ quantum heat engines : work , entropy and information at nanoscale_. 99 h.e.d .
scovil and e.o .
schulz - dubois , phys .
* 2 * , 262 ( 1959 ) ; j.e .
geusic , e.o .
schulz - dubois , and h.e.d .
scovil , phys . rev . * 156 * , 343 ( 1967 ) .
r alicki , j. phys . a : math . gen . * 12 * , l103 ( 1979 ) .
e. geva and r. kosloff , j. chem . phys . *
96 * , 3054 ( 1992 ) .
r. kosloff and t. feldmann , phys .
e * 65 * , 055102 ( 2002 ) .
t. feldmann and r. kosloff , phys .
e * 68 * , 016101 ( 2003 ) .
scully , m.s .
zubairy , g.s .
agarwal , and h. walther , science * 299 * , 862 ( 2003 ) .
j. gemmer , m. michel , and g. mahler , _ quantum thermodynamics _ , springer , berlin ( 2004 ) .
t. feldmann and r. kosloff , phys .
e * 73 * , 025107 ( 2006 ) .
kieu , phys .
lett . * 93 * , 140403 ( 2004 ) ; eur .
j. d * 39 * , 115 ( 2006 ) .
t. zhang , w .- t .
liu , p .- x .
chen , and c .- z .
li , phys . rev .
a * 75 * , 062102 ( 2007 ) .
a. e. allahverdyan , r. s. johal , and g. mahler , phys .
e * 77 * , 041118 ( 2008 ) .
r. dillenschneider and e. lutz , europhys.lett .
* 88 * , 5003 ( 2009 ) .
t. feldmann and r. kosloff , phys .
e * 85 * , 051114 ( 2012 ) .
x. l. huang , l. c. wang , and x. x. yi , phys .
e 87 , 012144 ( 2013 ) .
, edited by h.s .
leff and a.f .
rex ( institute of physics , bristol , 2003 ) .
m. kolar , d. gelbwaser - klimovsky , r. alicki , and g. kurizki , phys .
lett . * 109 * , 090601 ( 2012 ) .
r. kosloff , entropy * 15 * , 2100 ( 2013 ) .
y. rezek , entropy * 12 * , 1885 ( 2010 ) .
x. hu , r. de sousa and s. das sarma , phys .
lett * 86 * , 918 ( 2001 ) .
m. asoudeh and v. karimipour , phys .
a * 71 * , 022308 ( 2005 ) .
e. albayrak , the european physical journal b , * 72 * , 491 ( 2009 ) .
m. born and v. fock , z. phys .
* 51 * , 165 ( 1928 ) .
t. kato , j. phys .
japan * 5 * , 435 ( 1950 ) .
m. j. henrich , m. michel , and g. mahler , europhys .
76 * , 1057 ( 2006 ) . | we consider two spin-1/2 particles with isotropic heisenberg interaction , as the working substance of a quantum heat engine .
we observe a frictional effect on the adiabatic branches of the heat cycle , which arises due to an inhomogeneous driving at a finite rate of the external magnetic field .
the frictional effect is characterized by entropy production in the system and reduction in the work extracted .
corresponding to a sudden and a very slow driving , we find expressions for the lower and upper bounds of work that can be extracted on the adiabatic branches .
these bounds are also confirmed with numerical simulations of the corresponding liouville - von neumann equation . |
in everyday experience , even when we do not realize it , we use our intuition to draw inferences . for example
when we hear doorbell we immediately ask ourselves `` who has come ? '' . on the way to the door we consider many different possibilities .
maybe it is our friend , who said some days before that he visit us .
maybe it is our neighbor , who came to say that we should lower the music or maybe somebody came to inform us that we have won in the lottery and so on and so on .
all of these possibilities are more or less probable , with some outcomes which we can not figure out _
ex ante_. but every bit of information influences our expectations and we stick on outcome which seems to be most probable ( subjectively ) to us .
so we look through the window and see a car on the street , that looks like our friend s , we are almost sure that it is he in fact . or if we really play music very loudly we can expect that sooner or later somebody will be angry . when we do not expect somebody we know
, we assume that it may be the postman or someone who got the wrong address .
in fact we always choose the simplest case .
it means that we never assume that superman is standing behind the door , even when there is a reason why he should visit us .
this intuitive feeling to choose the simplest solution is called occam s razor @xcite and it states formally _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` accept the simplest explanation that fits the data '' . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in our example the data means in fact what we already know .
this example is very easy and our faculties manage very well to solve problems of this kind .
however there is a vast variety of problems for which there are uncountable possibilities and such problems can be treated only in an approximated manner . on the other hand
there are problems , which are too tedious to be solved by humans in reasonable time and we should employ to do job .
so , can we enclose this intuitive knowledge in the form of mathematically defined theory and use it instead of mind ?
the answer is yes , this exciting idea is embodied in the form of bayesian inference @xcite .
below we introduce bayesian inference and show how it works in practice .
we start from general considerations which lead us to the connection with thermodynamics .
subsequently we show how to write down problems on computer with use of the monte - carlo approach ( the metropolis algorithm ) .
we describe possible applications in the modern cosmology .
there is a huge number of places in cosmology where bayesian inference can often be applied .
we have residual observations and a lot of theories .
some of them are easy , some are pure models , some are brilliant new ideas .
this is as with our example with the door bell .
we hear the bell and we must predict who is at the door . without any information we can not predict who is ringing , because the sound of the bell is always the same
. however some people ring only once and some of them more times .
therefore we must listen very carefully when something is ringing in cosmology .
in the last section we try to put bayesian inference into a larger perspective : containing epistemological aspects of the method as well as suggested limitations .
our goal it to present a way to choose among alternative theories taking into account their conformity with data .
of course these theories can have different basis .
they can be connected with everyday experience , data analysis , biology , physics etc .
because we want to apply finally bayesian inference in physics we can restrict ourselves now , without loss of generality , to physical theories .
so let us consider some unknown physical phenomenon and let us say we possess @xmath0 theories @xmath1 that can potentially describe it .
all these theories differ from one another .
information about phenomenon investigated is contained in the collection of experimental data @xmath2 . both theories and experimental data we may consider as elements of the same set space @xmath3 so @xmath4 .
the set @xmath3 together with measure @xmath5 and @xmath6-algebra @xmath7 build probability space @xmath8 .
in such a well defined theory of probability , a natural concept of conditional probability occurs .
so the probability of a given theory @xmath9 , when we have data @xmath2 , is defined as @xmath10 this probability tells us which theory describes experimental data @xmath2 better and is called posterior probability . on the other hand we can ask about probability of outcomes @xmath2 when theory @xmath9 is the true one @xmath11 this probability tells us about different predictions @xmath2 from the theory @xmath9 and is a marginal likelihood ,
commonly called evidence . because @xmath12 we can combine equations ( [ posterior ] ) and ( [ evidence ] ) what give us @xmath13 this equation is the famous bayes theorem . the probability @xmath14 in this equation is called prior probability and it is in fact hard to describe this number .
it describes our initial beliefs about a given theory .
it is a human factor to choose this number and can be non - objective .
sometimes one theory is chosen because of its mathematical beauty although a more probable alternative exists .
the other factor is that some theory can well describe variety of others similar phenomena .
however , if we do not have strong motivation to introduce some initial selection of the theories , then the most natural choice is to assume a homogeneous distribution of the prior probability @xmath15 then none of theories is favored .
the probability @xmath16 is a simple normalization constant , what we can calculate thanks to the normalization condition @xmath17 which together with the bayes theorem ( [ bayes ] ) give us @xmath18 each of the theories @xmath1 contains a number of parameters described by the vector @xmath19 for a particular theory .
the simple theories ( simple mathematically ) contain in general a small number of parameters .
the main increase of number of parameters enlarges the complexity of theories .
this complexity can be in some cases accepted due to intrinsic beauty of a mathematical structure of theory .
nevertheless the theory which has one factor to explain a phenomenon is preferable over the theory which employ many factors for description of it .
effective theories belong to the type of simple theories .
@xmath20 parameters @xmath21 of the model @xmath22 are elements of the set space @xmath3 .
the values of these parameters may be fixed or significantly bounded by a theory .
but when no limits are put on these parameters ( there is no prior knowledge ) the evidence is calculated as the marginal probability integrated over all the allowed range of values of the parameters of the model .
now we can go back to the bayes theorem and explain the idea of bayesian inference . considering the bayes theorems for two models @xmath23 and @xmath24 and dividing the respective equations ( [ bayes ] ) by sides we obtain @xmath25 which is called the bayes factor .
if the priors @xmath14 for all @xmath23 are equal then the bayes factor reduces to the ratio of evidences ( @xmath26 ) .
the values of @xmath27 can be interpreted as follows : if @xmath28 then inference is inconclusive , if @xmath29 we have weak , if @xmath30 we have moderate and if @xmath31 we have strong evidence in favor of a model indexed by @xmath23 over the model indexed by @xmath24 @xcite .
so the main problem is now to calculate the evidence .
the direct calculation is generally impossible .
that is the reason is to use the monte carlo methods to do it .
first we introduce the parameter @xmath32 and redefine evidence to the form @xmath33 where @xmath34 the @xmath35 is in fact likelihood and we denote it by @xmath36 .
so @xmath37 and @xmath38 = \exp \int_0 ^ 1 d\lambda \langle \log l
\rangle_{\lambda}. \label{eq12}\ ] ] now we can show a connection between our approach and thermodynamics . introducing @xmath39
we obtain @xmath40 and equation ( [ eq11 ] ) takes a known form @xmath41 this is the energy of the system in the temperature @xmath42 . when we calculate it for different temperatures we can directly evaluate the integral in the expression ( [ eq12 ] ) and hence the evidence . as we see , to perform the bayesian inference we have to calculate the thermodynamical integral ( [ int ] ) .
this kind of integrals can be solved analytically only in case of very simple systems .
numerical methods to solve this kind of problems are known as the monte carlo .
it is not the subject of this paper to describe how they work in detail .
however to make this paper self - contained we add a short appendix a introducing basics of the monte carlo methods .
we also present experimental demonstration of property of ergodicity which is important in the context of monte carlo simulation ( see appendix b ) .
an interested reader can find more on monte carlo simulations e.g. in @xcite .
now we have all theoretical equipment to show this approach in action . in this example
we show how to perform the bayesian inference in a very simple case .
we consider a very simple kind of theories and a small sample of data - points to make computer computation short .
we also design it for clarity and better understanding .
however generalizations to more advanced problems are straightforward . in the next section
we will mention how to apply bayesian methods to more complicated problems .
let us consider some experiment in which we perform measurements of some physical variable @xmath43 for six different values of parameter @xmath44 . in the experiment we also measure standard error of the outcomes @xmath43 .
in fact we one can repeat many times measurements of @xmath43 for a given @xmath44 value .
then one can obtain the mean values of parameter @xmath43 together with its dispersion .
these data points we present in table [ tab:1 ] .
.in the table we collect the exemplary pairs @xmath45 together with the uncertainty of @xmath43 .
the uncertainty @xmath46 can be the result of the instrumental resolution . [
cols="^,^,^",options="header " , ] the phenomenon which we instigate is still not undetermined , but we possess three polynomial models to describe them .
we list these models below @xmath47 models 1 and 3 look more simple because each is described by two parameters when model 2 contains three unknown parameters .
the first step of bayesian inference is to fit these models to experimental data .
we can use for example method of least squares .
we obtain @xmath48 the important ingredient in the bayesian inference is a choice of priors .
it is the choice of the intervals and probability distribution for parameters .
the parameter intervals should be specified , because we must perform the integration ( look for the solution ) in a finite parameter space .
standard errors of the parameters give us intervals necessary for bayesian inference .
of course , the different choices of the parameter intervals can lead to the different values of the posterior probabilities . in our case , we choose the intervals as the @xmath49-@xmath6 confidence interval for the parameters .
namely , @xmath50 $ ] , where @xmath51 is the best fit value and @xmath6 is the standard error .
next we assume that outcomes are from the gaussian distribution and the likelihood function has then a form @xmath52.\ ] ] now we apply theory described in the previous section . with use of the metropolis algorithm
we calculate energies @xmath53 for models considered .
we perform calculations for the values of @xmath32 from the range @xmath54 as it is necessary to calculate the integral in equation ( [ eq12 ] ) .
we show these results in fig . [ 1 ] , [ 2 ] , [ 3 ] .
dependence for the first model.,width=264 ] dependence for the second model.,width=264 ] dependence for the third model.,width=264 ] now with use of this data we can perform integrals in the form @xmath55 from equation ( [ eq12 ] ) we see that @xmath56 obtained values for three models considered are @xmath57 now we can directly calculate bayes factors @xmath58 based on the obtained values of @xmath59 one can conclude : * because @xmath60 , we have weak evidence that the linear model is favored over the three parametric quadratic model . * because @xmath61 , the two parametric quadratic model is moderately preferred over the the three parametric quadratic model . * because @xmath62 , there is no preferences between the two parametric models , linear and quadratic .
we can also directly calculate posterior probabilities @xmath63
what we see is that the posterior probability suggest that the last model explains the experimental data in the best way . even that it may look more complicated than linear model number one .
the second model can be discarded .
this model contains remaining models inside and naturally fits better to the experimental data because has more degrees of freedom . but comparing to the other models he is too complicated and not necessarily properly explains experimental data .
we see that however first and last model possess the same number of degrees of freedom the last function seems to explain in the better way the nature of the investigated physical process . on the other hand , from the bayes factor criterion
, one can not seen such a preference .
it is because the borders in the bayes factor criterion were established empirically in the very careful manner .
the example presented here illustrates that the different criteria give us stronger or weaker conclusions . in particular ,
note that posterior probability criterion favors the model @xmath64 over the model @xmath65 . on the other hand
the bayes factor criterion indicates that there is no preference between the models @xmath64 and @xmath65 .
therefore , this simple example gives us here the first lesson to be very careful interpreting statistical inference results .
the second lesson is to be careful about results of estimations .
the results presented above , base on the @xmath49-@xmath6 interval choice of the parameters intervals . in case of the @xmath66-@xmath6 intervals the resulting bayes factors are the following @xmath67 therefore , there the first model is weakly favored with respect to the third one . also the second model is weakly favored with respect to the third one .
the first and the second models are indistinguishable .
the @xmath66-@xmath6 results differ from these performed previously .
the comparison between the @xmath49-@xmath6 and @xmath66-@xmath6 examples was performed to show explicitly the sensitivity of the bayesian inference on the parameter intervals .
this dependence on the priors becomes significant in case of the _ weak data _
, e.g. for the small sample size or significant error .
for the _ strong data _ the impact from the priors becomes irrelevant . in the case
considered here , the data sample is small and the significant errors are present . because of this , the inference is sensitive on the choice of the priors as observed .
in this section we present some applications of bayesian inference in modern cosmology .
the bayesian methods have been introduced to cosmology relatively late .
the proliferation of dark energy models forced using the formal methods of model selection .
starting from the pioneering work of m. v. john and j. v. narlikar @xcite the bayesian methods have become popular @xcite .
the reason is that cosmology really needs these methods , because our knowledge about the universe is still very limited .
we have a lot of theories about different stages of the universe but only a modest number of observations to verify them .
so it is the right place for the bayesian methods .
the first case that we would like to talk about is connected with very mysterious behavior of the universe , discovered in the end of the last decade .
namely , observations of the distant type ia supernovae ( snia ) indicated that the universe expansion accelerates @xcite .
this discovery was apparently in the conflict with other observations and beliefs that universe in filled with normal matter like dust , stars , planets etc . given this assumption the universe always decelerates .
so what is happening in the universe ? why did it start to accelerate after a previous phase of deceleration ? what is the mysterious component of the universe that we call dark energy ?
there is presently an enormous number of possible answers for this question .
the presence of the cosmological constant @xmath68 , phantom fluid , modified gravity , field of quintessence , quantum gravitational effects , brane world models , vacuum energy and so on and so on .
the number of possible solutions is really impressive .
so , which one of them is the real solution ? which one describes nature in the right way ? or maybe none of them , maybe we still must be looking for new models .
let us neglect the possibility of building new models to explain acceleration of the universe and limit ourselves only to some already proposed solutions .
this is precisely what we can do thanks to bayesian inference .
due to bayesian methods one can obtain a ranking of accelerating models @xcite .
the analysis show that standard cosmological model , the so - called @xmath68cdm model , is on the top off all theoretical propositions @xcite .
the assumed model of the universe is the friedmann - robertson - walker model filled with dust matter , dark and baryonic , and dark energy . in bayesian estimation of cosmological model parameters for these components
we start from the cosmography which determines the luminosity function as the function of density parameters .
the density parameters define the fraction of energy in the total energy budget of the universe .
they are dimensionless and their sum is equal one .
the additional parameter is the hubble constant which describes the rate of expansion of the current universe . while the kinematic part of the cosmology is controlled by cosmography basing on analysis of trajectories of photons , the observations of cmb control the dependence of perturbation on time . among many theoretical propositions
the simplest candidate for dark energy is the time - independent cosmological constant .
the other propositions : the chaplygin gas , models with varying coefficient of the equation of state , quintessence , phantoms , etc .
these models are fitted using the astronomical data : the measurement of absolute magnitudes of high redshift type ia supernovae , cosmic microwave background ( cmb ) radiation , the measurement of baryon acoustic peak ( bao ) in galaxies correlation function , gas mass fraction value in galaxy clusters , gamma - ray bursts ( grb ) which are at higher redshift than snia . in selection of cosmological models with different forms of dark energy different criteria
the simplest criteria are the information criteria aic , bic and bayes factor , posterior probability .
these methods allow us pick out the best model in the light of data at our disposal .
this intriguing question corresponds to the presence of possible observational phenomena of quantum gravity .
the quantum gravitational effects are predicted to be very small and unreachable by the present and any future generations of accelerators .
however the quantum gravitational effect can survive as the relict from the very early universe in which quantum gravitational effects have been dominant .
these effects can influence the spectrum of inflationary perturbations @xcite .
these primordial fluctuations then lead to the fluctuations of matter and finally to the structures formation in the universe .
so can we deduce some information about quantum gravity from observations of microwave background radiation and large scale structures ? at first glance it can sound strange because quantum gravity describes microscopic property of gravitational field at a very deep level .
however the same effects were very important in the early universe and could influence its global properties .
is here a place for bayesian inference ?
the answer is _
yes_. we have now a lot of predictions from quantum theories of gravity like loop quantum gravity @xcite and still a very limited number of observations in the region where they can be important , e.g. non - gaussianity in primordial fluctuation in cmb or the primordial gravitational waves spectrum .
the two examples presented in this section are very important but they are not the only ones . there is a lot of other places in cosmology where bayesian inference is and should be applied . for example , we still do not know what the dark matter ( the second dominant component of the universe ) is .
it may be an axion , higisno , gluino other super particles @xcite or just neutrinos etc .
it is an ideal place for the bayesian inference to point out the best candidate .
unfortunately still too little is known from observations .
it is an obvious fact that bayes theorem is a theorem .
nevertheless , it should be explicitly explained , that to be a bayesian is more than _ to know and use the theorem_. the bayesian theory of confirmation achieves a great success in such areas of human activity as physics , biology , medicine , cognitive science ( decision theory ) on the one hand , and encounters serious limitations of the analyzed method on the other hand .
a radical bayesian would probably say that it can always be applied : we always have to do with a joint distribution and a numerical representation of belief is always possible .
however , a bayesian epistemology could be treated not only as competing with other methods of confirmation , but as a way of making a specific ( not exclusive ) model of beliefs and related evidences , which can be easily elaborated and understood . in the bayesian approach probability
is attributed to hypotheses that are being confirmed .
this confirmation can be interpreted both qualitatively and quantitatively , since inference is based on empirical data and there are relations between hypotheses , theories and observations to be explicated .
it is indeed a crucial point in understanding bayesian inference the meaning which is ascribed to probability . using probabilistic methods one can measure two things : how often a specific event occurs and how strong evidence ( confirming our beliefs ) is .
let us generally state that bayesianism can be treated as an epistemic theory which examines the relation between beliefs and empirical evidence as to measure the strength of the beliefs . as it has been shown the most important concept used to gain that goal
is the notion of conditional probability @xcite . using the bayesian inference we not only measure the strength of beliefs but also propose the method for rational estimating a change of the beliefs under the influence of a new evidence .
sometimes , among scientists and philosophers of science , there is a bit of hesitation about exclusiveness of such an approach .
s. okasha in his study on van frassen s conception of induction wrote @xcite : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ he accepts the bayesian representation of opinion in terms of degrees - of - belief , and he agrees that synchronic probabilistic coherence is a necessary condition of rationality .
however , he does not accept the bayesian thesis that conditionalization is the only rational way to respond to new evidence ; though he allows that it is a rational way .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ it can be said in that sense that bayesianism offers a solution to old problems of induction .
we have got an approximately coherent and reasonable model for probability corrections .
of course , it is possible if having initial probability and evidence ( priors ) .
the proposed solution has its weakness : its method often tells nothing how to estimate these probabilities .
the crucial point in bayesian inference lies in the fact that it is able to deal with the problem , only if we manage the input of some probabilities .
it makes sense , since we never start reasoning with absolutely no knowledge .
the result we achieve
@xmath69 are always `` conditional '' : it reveals a property of @xmath70 which is not objective , but related to @xmath71 and certain _ knowledge _ , called _
background knowledge_. a more subtle epistemic analysis can be carried out with reference to the types of background knowledge .
p. wang studied the problem of underlying knowledge and discerned two types of conditions which influence the evaluation of probability distribution function @xcite in the formulas , as follows : * explicit conditions @xmath72 * * @xmath71 is a binary proposition , * * belongs to proposition space ( then we can define @xmath73 ) , * * @xmath74 ; * implicit conditions @xmath75 * * non - binary propositions allowed , * * there may be statements outside the proposition space , * * `` even if a proposition is assigned to a prior probability of zero according to one knowledge source , it is possible for the proposition to be assigned a non - zero probability according to another knowledge source '' .
all these discernments are not trivial , since we have to answer the question : _ whether all the background knowledge can be probabilistic - valued?_. this is one of the most important epistemic questions of bayesian theory of confirmation , beside the others : * _ are there degrees of belief ? _ the answer , maybe , lies in an attempt to distinguish ` rational ' degrees of belief from belief in general .
are the corrections in probability , gained in the bayesian procedure , just new probabilistic information or do they deliver _ new reason to believe _ that the proposition considered is true ? * when are we actually updating our belief and when there is just a revision of probability ( known problem of old evidence ) ? while elaborating empirical data in cosmology , one can use classical or bayesian statistics . in a classical approach we rely on the classical definition of likelihood , but doing the same _ bayesian way _ we are dealing with _ a priori _ and _ a posteriori _ probability , respectively .
there is an opinion among some cosmologists that the standard approach ( _ the standard maximum - likelihood technique _ ) is satisfactory indeed , which is reasonable if there is a single model under consideration .
when we have several competitive hypotheses , bayesian statistics would be a better choice .
there are two groups among bayesians who differ from each other with respect to criteria used in choosing of priors : objective bayesians : e. t. jaynes @xcite , h. jeffreys @xcite , r. d. rosenkrantz @xcite ) and subjective bayesians : b.de finetti @xcite , c. howson and p. urbach @xcite .
it is not only the problem ( or problems ) of induction , that bayesian inference tries to deal with , but also the problem of finding a justification of induction inference itself , which can be explicated in several schemas : * inductive generalization + nobody denies that a finite number of experimental data can not deliver an exhaustive proof to a universal statement but , according to induction , empirical evidence confirms generalization @xcite .
an example of that is enumerative induction which principle explicates , as follow : + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ several crows are black . + therefore , all crows are black . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ + _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ every load added so far has not damaged this truck . + therefore , the next piece of load will not damage this truck . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * hypothetical induction + it occurs when some hypothesis deductively entails the evidence .
+ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ an evidence confirms hypothesis , if the evidence is a logical consequence of the hypothesis . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ + in the case of existing multiple competing hypotheses one can try to show that the falsity of the hypothesis entails the falsity of the evidence or use additional criteria of hypotheses selection , like simplicity or inference to the best explanation . however , these proposals create _ market of hypotheses _
( for example cosmological models ) with rules for successful selection but in fact they can not give any rational explanation to the evidence .
it may be a truism , but the difference between explanation and confirmation should be treated with special care .
the more so because there is not a unity among bayesians concerning representation of the degree to which evidence supports an hypothesis .
the most popular are three options : * * a difference measure : @xmath76 ; * * a normalized difference measure : @xmath77 ; * * a likelihood measure : @xmath78}{[1-p(h|e ) ] p(h)}$ ] the inductive generalization , which has the simple pattern : extrapolation from particular data do general conclusions , suffers several problems called paradoxes of confirmation .
goodman s paradox , know in the literature as problem of `` grue '' , is particularly interesting @xcite .
especially the question of its counterpart in the field of cosmology . in a traditional version : * we have two hypotheses : ( 1 ) all emeralds are green and ( 2 ) all emeralds are grue ( green if examined until some time @xmath79 and blue otherwise ) . * evidence : _ found emerald is green _ confirms both : ( 1 ) and ( 2 ) .
any satisfying resolutions to the paradox propose additional assumptions ; for example pointing out on `` green '' as a natural kind term instead of `` grue '' . in a search for possible cosmological version of the paradox we can compare for example two related models the cold dark matter cosmological model ( cdm model ) and the lambda cold dark matter cosmological model with the positive cosmological constant term ( lcdm model ) .
the latter seems to be the simplest candidate for the dark energy description .
the bayesian method of confirmation dedicated to select between these two models reveals a quite opposite verdict while used in the 90s and currently . using the sample of perlmutter et al .
@xcite there is not enough information to distinguish these models .
the extended sample with additional 42 high @xmath80 snia @xcite gives a weak evidence to favor the lcdm model over the cdm one .
however , in our opinion it is a misunderstanding to treat this study case as a paradox in goodman s sense .
it becomes obvious , because when new observational data confirm better the lcdm model in comparison with the cdm model , the latter simply disappears out the stage .
the paradox of confirmation would occur when related to a certain family of models there will be the same degree of confirmation ( the same time and evidence ) assigned to hypotheses differing from each other for example with regard to foreseeable future scenarios of universe evolution .
to illustrate this situation let us consider two hypotheses 1 .
the universe decelerates .
the universe decelerates until some time @xmath79 and accelerates afterward .
the cdm model is valid with the first hypothesis and the lcdm model is in agreement with the second one . from the 60 s
it was known that the universe is expanding with the decelerating rate .
so we have a paradox here .
however the evidences of accelerating universe due to snia data falsified the first model . and
the paradox is naturally solved .
this example teaches us that paradoxes of goodman s type ( in the logic of induction ) are common in evolutionary sciences but they are not dangers because we hope that new evidences ( which appear due to science development ) we discriminate between two hypotheses . in goodman s paradox
there is only one kind of evidence ; we need to draw an emerald and check its color . in the case of cosmological hypotheses
we are not left with only one evidence .
a new evidence appears and resolves the paradox in favor of one of the hypothesis .
it comes from new observations .
we know that this evidence will appear eventually because we , scientists look for it .
the reason that there is no paradox after a new evidence appears , is that one hypotheses is falsified ( the cdm model ) and only one hypothesis ( the lcdm model ) becomes in agreement with this new evidence .
it is often said that a scientific theoretical research means achieving two specific goals : ( 1 ) finding a model which approximates a phenomenon best and ( 2 ) constructing a hypothesis that offers best prediction .
it is a good example to show how in this context two criteria of model selection are being compared : the akaike information criterion ( aic ) and bayesian information criterion ( bic ) @xcite . although these model comparison methods are put together as competitors , they in fact try to ask different questions @xcite .
the aic estimates predictive power of an elaborated hypothesis , while the bic goodness - of - fitting @xcite .
m. forster and e. sober have explained this nuance with respect to the fitting problem @xcite : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ even though a hypothesis with more adjustable parameters would fit the data better , scientists seem to be willing to sacrifice goodness - of - fit if there is a compensating gain in simplicity .
( ) + since we assume that observation is subject to error , it is overwhelmingly probable that the data we obtain will not fall exactly on that true curve .
( ) since the data points do not fall exactly on the true curve , such a best - fitting curve will be _
false_. if we think of the true curve as the ` signal ' and the deviation from the true curve generated by errors of observation as ` noise ' , then fitting the data perfectly involves confusing the noise with the signal .
it is overwhelmingly probable that any curve that fits the data perfectly is false .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the general comments of this section can be summed up by a statement that bayesian inference is a method dedicated to specific goals in scientific practice @xcite . with respect to cosmology , the mentioned lcdm cdm models comparison reveals in bayesian inference context another problem .
it strictly concerns currently changing concept of the model in physics @xcite . at present
there is a special emphasis placed on effectiveness and mediating function of models in physics .
this status of scientific models is determined by the way they are designed : they are not simply derived from the underlying theory , nor fixed by the evidence only .
their `` nature '' is determined by a mediating role ( between a theory and phenomena ) .
morrison states , as follows @xcite : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ although they are designed for a specific purpose these models have an autonomous role to play in supplying information , information that goes beyond what we are able to derive from the data / theory combination alone . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in cosmology built on general relativity , the solutions of einstein equations can be treated as the geometrical models of the universe .
a construction of a model starts from assuming specific idealizations ( symmetries , etc ) .
it means in a practice that we reduce degrees of freedom ( all apart gravitational ones are neglected ) .
for example assumption of spacial homogeneity means that the einstein equations which constitute the system of non - linear partial differential equation system reduce to the ordinary differential equation system in the cosmological time .
it is said that those formulations of scientific laws are certain approximations of the investigated phenomena .
there has been recently quite an important and interesting discussion about validity of application the bayesian inference to idealization itself @xcite .
the problem concerns idealized hypotheses and a question of assignment probability to them , since they can be treated as counterfactuals .
what is a posterior probability of the ideal gas law or the law of motion for simple pendulum ?
jones showed that solution lies exactly in the understanding of the procedure of elaborating a model .
if we treat the model idealizations not as a result of abstraction but as a distortion , the methodological consequences may exclude for bayesian inference@xcite : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ given that most scientific hypotheses are idealized in some way , bayesianism seems to entail that most scientific hypotheses can not be confirmed .
+ bayesians thus confront an apparent trilemma : either develop a coherent proposal for to assign prior probabilities to counterfactuals ; or embrace the counterintuitive result that idealized hypotheses can not be confirmed ; or reject bayesianism . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the general bayesian conception of empirical evidence can be put into three main statements / consequences : * less probable evidence delivers best confirmation to hypothesis ; * evidence confirms better those hypothesis in context of which it is more probable . *
if the hypothesis probability is very little , it can be confirmed only by very strong evidence .
in this paper we have presented basics of bayesian inference and showed how to use it in practice . we have introduced some mathematical background and formulated a problem in the similarity with thermodynamics . as a case study
we choose three simple models .
then the known monte carlo methods and metropolis algorithm were used to select the best model in the light of data .
the general remark which can be derived from these considerations is to be careful in evaluation of the models in the light of the data and in using the complementary indicators .
the bayesian methods started to be popular due to new discoveries in cosmology at the beginning of xxi century .
we presented the areas of cosmology where bayesian inference has been applied , namely problem of dark energy , dark matter , and testing quantum effects by astronomical data .
subsequently we have studied epistemological aspects of the bayesian confirmation theory in the context of problems of modern cosmology where the bayesian approach offers not only the estimation of model parameters from the observational data but also methods of the comparison of models ( selection ) .
we have demonstrated that the bayesian inference is based on some assumptions of philosophical character .
the philosophical issues of inference in context of cosmological models on the example of models without and with the dark energy component ( the cosmological constant ) are discussed .
we pointed out that goodman s famous paradox does not appear in the cosmology reconstructed using bayesian methodology .
the reason for this we are looking for new evidences which falsify one hypothesis such that only one hypothesis becomes in agreement with observational data .
note that the bayesian framework enable us to test and select between competing hypotheses so one can construct the ranking of cosmological models explaining acceleration of the current universe .
therefore , we obtain the best model favored by data .
in this appendix we show how to compute thermodynamical integral ( [ int ] ) with use of the monte carlo simulations .
our short introduction to this subject is based partially on this made in @xcite .
let us consider state of the system labeled by @xmath81 and corresponding energy @xmath82 .
our task is to compute integral @xmath83 where integration is performed over all available states @xmath81 . since in numerical computations
we always discretize the system , integration @xmath84 is replaced by the summation .
our task now is to write a program which generates states @xmath81 from the canonical ensemble given with the probability @xmath85 .
the crucial observation is that we do not have to generate all possible states to calculate ( [ intapp ] ) .
the main contribution to their value comes from the equilibrium states .
therefore the idea is to find these equilibrium states and average over them .
starting from some arbitrary initial state we create a sequence of states @xmath86 finally finding ensemble of equilibrium states
. then one can calculate @xmath87 in order to find equilibrium states the markov chain method can be applied .
we consider sequence of transitions @xmath88 with probability @xmath89 .
moreover we assume @xmath90 practical realization of the above conditions is given by _
metropolis algorithm_. namely it states : * take initial state @xmath91 . *
make some move to neighboring state @xmath92 . *
if @xmath93 , accept the change . *
if @xmath94 , accept the change conditionally with the probability @xmath95/t}$ ] .
all computations performed in sec . [ simpleexample ] have been done applying this simple set of rules . as an example
we show here the markov chains in the parameter space for the models considered there . in fig .
[ fig:1l ] we show sequence of moves for the first model considered in sec . [ simpleexample ] .
we assume values of the parameter @xmath96 . since @xmath97 the higher value of @xmath32 corresponds to lower temperatures .
we investigate here a broad range in @xmath32 , however for the calculations of the evidence only values of @xmath98 $ ] are required .
@xmath99{1l01.eps } & \includegraphics[width=6cm]{1l1.eps } \\
\includegraphics[width=6cm]{1l10.eps } & \includegraphics[width=6cm]{1l100.eps } \end{array } $ ]
the very important question related to monte carlo simulations is the `` ergodicity '' of the algorithm .
it means that in the finite number of steps ( finite time ) the system must be freely close to any point in the phase space .
this prevents the system to being trapped in a subset of states . in the monte carlo simulations it causes that we can always find a proper energy minimum , even for very low temperatures when fluctuations are small . to check it we performed markov chains in the low temperature system . in such a system ,
when algorithm is not ergodic , a markov chain can not always lead to the proper minimum . in fig .
[ erg ] we show markov chain in the parameters space for the third model considered in sec . [ simpleexample ] .
we show that starting from the different points in the parameter space system always go to the same region where the proper minimum is placed .
this is a visual proof of ergodicity for a kind of function considered .
it is possible that it is not true for more complicated kind of functions . | we discuss epistemological and methodological aspects of the bayesian approach in astrophysics and cosmology .
the introduction to the bayesian framework is given for a further discussion concerning the bayesian inference in physics .
the interplay between the modern cosmology , bayesian statistics , and philosophy of science is presented .
we consider paradoxes of confirmation , like goodman s paradox , appearing in the bayesian theory of confirmation . as in goodman
s paradox the bayesian inference is susceptible to some epistemic limitations in the logic of induction .
however goodman s paradox applied to cosmological hypotheses seems to be resolved due to the evolutionary character of cosmology and accumulation new empirical evidences .
we argue that the bayesian framework is useful in the context of falsificability of quantum cosmological models , as well as contemporary dark energy and dark matter problem . |
-5 mm a lot of fascinating mathematics has been created when successful tools are transferred from one research area to another .
we shall describe two instances of such transfers , both into knot theory .
the first transfer realizes commutator calculus of group theory by _ embedded versions _ in 3- and 4-space , and produces many interesting geometric equivalence relations on knots , called _ grope cobordism _ in 3-space and _ grope concordance _ in 4-space .
it turns out that in 3-space these new equivalence relations give a geometric interpretation ( theorem [ thm:3d ] ) of vassiliev s finite type invariants @xcite and that the kontsevich integral @xcite calculates the new theory over @xmath0 ( theorem [ thm : kontsevich ] ) . in 4-space the new equivalence relations factor naturally through knot concordance , and
in fact they organize all known concordance invariants in a wonderful manner ( theorem [ thm:4d ] ) .
they also point the way to new concordance invariants ( theorem [ thm : cot2 ] ) and these are constructed using a second transfer , from the spectral theory of self - adjoint operators and von neumann s continuous dimension @xcite .
to explain the first transfer into knot theory , recall that every knot bounds a seifert surface ( embedded in 3-space ) , but only the trivial knot bounds an embedded disk .
thus all of knot theory is created by the difference between a surface and a disk .
the new idea is to filter this difference by introducing a concept into knot theory which is the analogue of iterated commutators in group theory .
commutators arise because a continuous map @xmath1 extends to a map of a surface if and only if @xmath2 represents a commutator in the fundamental group @xmath3 .
iterated commutators can similarly be expressed by gluing together several surfaces .
namely , there are certain finite 2-complexes ( built out of iterated surface stages ) called _ gropes _ by cannon @xcite , with the following defining property : @xmath1 represents an element in the @xmath4-th term of the _ lower central series _ of @xmath3 if and only if it extends to a continuous map of a _ grope of class @xmath4_. similarly , there are _ symmetric gropes _ which geometrically implement the derived series of @xmath3 , see figures [ fig : gropes ] and [ fig : symmetric gropes ] gropes , therefore , are not quite manifolds but the singularities that arise are of a very simple type , so that these 2-complexes are in some sense the next easiest thing after surfaces . two sentences on the history of the use of gropes in mathematics
are in place , compare ( * ? ? ?
* sec.2.11 ) .
their inventor stanko worked in high - dimensional topology , and so did edwards and cannon who developed gropes further .
bob edwards suggested their relevance for topological 4-manifolds , where they were used extensively , see @xcite or @xcite .
it is this application that seems to have created a certain angst of studying gropes , so we should point out that the only really difficult part in that application is the use of _ infinite constructions _ ,
i.e. when the class of the grope goes to infinity . one purpose of this note is to convince the reader that ( finite ) gropes are a very simple , and extremely powerful tool in low - dimensional topology .
the point is that once one can describe iterated commutators in @xmath3 by maps of gropes , one might as well study _ embedded _
gropes in 3-space ( respectively 4-space ) in order to organize knots up to isotopy ( respectively up to concordance ) . in section [ sec:3d
] we shall explain joint work with jim conant on how gropes embedded in 3-space lead to a geometric interpretation of vassiliev s knot invariants @xcite and of the kontsevich integral @xcite . in section [ sec:4d ] we study symmetric gropes embedded in 4-space , and explain how they lead to a geometric framework for
all known knot concordance invariants and beyond .
more precisely , we explain our joint work with tim cochran and kent orr @xcite , where we define new concordance invariants by inductively constructing representations into certain solvable groups @xmath5 , and associating a hermitian form over the group ring @xmath6 to the knot @xmath7 , which is derived from the intersection form of a certain @xmath8-manifold with fundamental group @xmath5 and whose boundary is obtained by @xmath9-surgery on @xmath7 .
this intersection form represents an element in the cappell - shaneson @xmath10-group @xcite of @xmath6 and we detect it via the second transfer from a different area of mathematics : the standard way to detect elements in witt groups like the @xmath10-group above is to construct unitary representations of @xmath5 , and then consider the corresponding ( twisted ) signature of the resulting hermitian form over @xmath11 .
it turns out that the solvable groups @xmath5 we construct do not have any interesting finite dimensional representations , basically because they are `` too big '' ( e.g. not finitely generated ) , a property that is intrinsic to the groups @xmath5 in question because they are `` universal solvable '' in the sense that many @xmath8-manifold groups ( with the right boundary ) must map to @xmath5 , extending the given map of the knot group .
however , every group @xmath5 has a fundamental unitary representation given by @xmath12 , the hilbert space of square summable sequences of group elements with complex coefficients .
the resulting ( weak ) completion of @xmath13 is the _ group von neumann algebra _ @xmath14 .
it is of type @xmath15 because the map @xmath16 extends from @xmath13 to give a finite faithful trace on @xmath14 .
the punchline is that hermitian forms over the completion @xmath14 are much easier to understand than over @xmath13 because they are diagonalizable ( by functional calculus of self - adjoint operators ) . here
one really uses the von neumann algebra , rather than the @xmath17-algebra completion of @xmath13 because the functional calculus must be applied to the characteristic functions of the positive ( respectively negative ) real numbers , which are bounded but _ not _ continuous .
the subspace on which the hermitian form is positive ( respectively negative ) definite has a _ continuous dimension _ , which is the positive real number given by the trace of the projection onto that subspace . as a consequence
, one can associate to every hermitian form over @xmath14 a real valued invariant , the _ von neumann signature_. in @xcite we use this invariant to construct our new knot concordance invariants , and a survey of this work can be found in section [ sec:4d ] it is not only related to embedded gropes in 4-space but also to the existence of towers of _ whitney disks _ in 4-space .
unfortunately , we wo nt be able to explain this aspect of the theory , but see ( * ? ? ? * thm.8.12 ) . in section [ sec:4d ]
we shall hint at how the interesting representations to our solvable groups are obtained .
but it is well worth pointing out that the methods developed for studying knot concordance have much simpler counterparts in 3-space , i.e. if one is only interested in isotopy invariants .
a typical list of knot invariants that might find its way into a text book or survey talk on _ classical _ knot theory , would contain the alexander polynomial , ( twisted ) signatures , ( twisted ) arf invariants , and maybe knot determinants .
it turns out that all of these invariants can be computed from the homology of the infinite cyclic covering of the knot complement , and are in this sense `` commutative '' invariants . instead of the maximal abelian quotient one can use other solvable quotient groups of the knot group to obtain `` noncommutative '' knot invariants .
the canonical candidates are the quotient groups @xmath18 of the derived series @xmath19 of the knot group ( compare section [ sec:4d ] for the definition ) .
one can thus define the higher order _ alexander modules _ of a knot @xmath7 to be the @xmath20-modules @xmath21 the indexing is chosen so that @xmath22 is the classical alexander module .
for @xmath23 these modules are best studied by introducing further algebraic tools as follows . by a result of strebel
the groups @xmath18 are torsionfree .
therefore , the group ring @xmath24 satisfies the ore condition and has a well defined ( skew ) quotient field .
this field is in fact the quotient field of a ( skew ) polynomial ring @xmath25 $ ] , with @xmath26 the quotient field of @xmath27 $ ] and @xmath28 .
thus one is exactly in the context of ( * ? ? ?
* sec.2 ) and one can define explicit noncommutative isotopy invariants of knots . for example , let @xmath29 be the dimension ( over the field @xmath30 ) of the _ rational _ alexander module @xmath31.\ ] ] it is shown in ( * ? ? ?
* prop.2.11 ) that these dimensions are finite with the degree of the usual alexander polynomial being @xmath32 .
moreover , cochran @xcite has proven the following non - triviality result for these dimensions .
if @xmath7 is a nontrivial knot then for @xmath23 one has @xmath33 moreover , there are examples where these numbers are strictly increasing up to any given @xmath34 .
if one of the inequalities in the above theorem is strict then @xmath7 is not fibered . furthermore ,
@xmath9-surgery on @xmath7 cross the circle is not a symplectic 4-manifold .
the first statement is clear : for fibered knots the degree of the alexander polynomial @xmath32 equals twice the genus of the knot @xmath7 .
the second statement follows from a result of kronheimer @xcite who showed that this equality also holds if the above @xmath8-manifold is symplectic .
recently , harvey @xcite has studied similar invariants for arbitrary @xmath35-manifolds and has proven generalizations of the above results : there are lower bounds for the thurston norm of a homology class , analogous to @xmath36 , that are better than mcmullen s lower bound , which is the analogy of @xmath32 . as a consequence , she gets new algebraic obstructions to a @xmath8-manifold of the form @xmath37 admitting a symplectic structure .
just like in the classical case @xmath38 , there is more structure on the rational alexander modules . by (
* thm.2.13 ) there are higher order _ blanchfield forms _ which are hermitian and non - singular in an appropriate sense , compare ( * ? ? ?
* prop.12.2 )
. it would be very interesting to know whether the @xmath34-th order blanchfield form determines the von neumann @xmath39-invariant associated to the @xmath40-cover .
so far , these @xmath39-invariants are very mysterious real numbers canonically associated to a knot
. only in the bottom case @xmath38 do we understand this @xmath39-invariant well : the @xmath41-index theorem implies that the von neumann @xmath39-invariant corresponding to the @xmath42-cover is the von neumann signature of a certain @xmath8-manifold with fundamental group @xmath42 .
moreover , this signature is the integral , over the circle , of all ( levine - tristram ) twisted signatures of the knot ( * ? ? ?
* prop.5.1 ) ( and is thus a concordance invariant ) .
for @xmath23 there is in general no such @xmath8-manifold available and the corresponding @xmath39-invariants are not concordance invariants .
-5 mm we first give a more precise treatment of the first transfer from group theory to knot theory hinted at in the introduction . recall that the fundamental group consists of continuous maps of the circle @xmath43 into some target space @xmath44 , modulo homotopy ( i.e.@xmath45-parameter families of continuous maps ) .
quite analogously , classical knot theory studies smooth _ embeddings _ of a circle into @xmath46 , modulo isotopy ( i.e. @xmath45-parameter families of embeddings ) . to explain the transfer
, we recall that a continuous map @xmath1 represents the trivial element in the fundamental group @xmath3 if and only if it extends to a map of the disk , @xmath47 .
moreover , @xmath2 represents a commutator in @xmath3 if and only if it extends to a map of a surface ( i.e. of a compact oriented 2-manifold with boundary @xmath43 ) .
the first statement has a straightforward analogy in knot theory : @xmath48 is trivial if and only if it extends to an embedding of the disk into @xmath46 . however , every knot `` is a commutator '' in the sense that it bounds a _ seifert surface _ , i.e. an embedded surface in @xmath46 .
[ fig : gropes ] recall from the introduction that gropes are finite 2-complexes defined by the following property : @xmath1 represents an element in the @xmath4-th term @xmath49 of the lower central series of @xmath3 if and only if it extends to a continuous map of a grope of class @xmath4 . here
@xmath50 is defined inductively for a group @xmath5 by the iterated commutators @xmath51 \text { and } g_k:=[g , g_{k-1 } ] \text { for } k>2.\ ] ] accordingly , a grope of class 2 is just a surface , and one can obtain a grope of class @xmath4 by attaching gropes of class @xmath52 to @xmath53 disjointly embedded curves in the _ bottom _ surface . here
@xmath53 is the genus of the bottom surface and the curves are assumed to span one half of its homology .
this gives gropes of class @xmath4 with one boundary circle as on the left of figure [ fig : gropes ] it s not the most general way to get gropes because of re - bracketing issues , and we refer to ( * ? ? ? *
sec.2.1 ) for details .
the boundary of a grope is by definition just the boundary of the bottom surface , compare figure [ fig : gropes ] [ def:3d ] two ( smooth oriented ) knots in @xmath46 are _ grope cobordant _ of class @xmath4 , if there is an embedded grope of class @xmath4 in @xmath46 ( the _ grope cobordism _ ) such that its boundary consists exactly of the given knots .
an _ embedding _ of a grope is best defined via the obvious @xmath35-dimensional local model .
since every grope has a 1-dimensional spine , embedded gropes can then be isotoped into the neighborhood of a 1-complex . as a consequence ,
embedded gropes abound in 3-space !
it is important to point out that if two knots @xmath54 cobounds a grope then @xmath55 and @xmath56 might very well be linked in a nontrivial way .
thus it is a much stronger condition on @xmath7 to assume that it is the boundary of an embedded grope than to say that it cobounds a grope with the unknot .
for example , if @xmath7 bounds an embedded grope of class @xmath35 in @xmath46 then the alexander polynomial vanishes .
together with stavros garoufalidis , we recently showed @xcite that the 2-loop term of the kontsevich integral detects many counterexamples to the converse of this statement . in joint work with jim conant @xcite , we show that grope cobordism defines equivalence relations on knots , one for every class @xmath57 . moreover , theorem [ thm:3d ] below implies that the resulting quotients are in fact finitely generated abelian groups ( under the connected sum operation ) .
for the smallest values @xmath58 and 5 , these groups are isomorphic to @xmath59 and they are detected by the first two vassiliev invariants ( * ? ? ?
* thm.4.2 ) .
the following theorem is formulated in terms of _ clasper surgery _ which was introduced independently by habiro @xcite and goussarov @xcite , as a geometric answer to finite type invariants a l vassiliev @xcite .
we can not explain the definitions here but see ( * ? ? ?
* thm.1 and 3 ) .
we should say that the notion of a _ capped _
grope is well known in 4 dimensions , see ( * ? ? ?
* sec.2 ) . in our context , it means that all circles at the `` tips '' of the grope bound disjointly embedded disks in 3-space which are only allowed to intersect the bottom surface of the grope .
[ thm:3d ] two knots @xmath60 and @xmath55 are grope cobordant of class @xmath4 if and only if @xmath55 can be obtained from @xmath60 by a finite sequence of clasper surgeries of _ grope degree @xmath4 _ ( as defined below ) .
moreover , two knots are _ capped _ grope cobordant of class @xmath4 if and only if they have the same finite type invariants of vassiliev degree @xmath61 . as a consequence of this result ,
the invariants associated to grope cobordism are highly nontrivial as well as manageable .
for example , we prove the following result in ( * ? ? ?
* thm.1.1 ) : [ thm : kontsevich ] the ( logarithm of the ) kontsevich integral ( with values in @xmath62 ) , graded by the new _ grope degree @xmath4 _ , is an obstruction to finding a grope cobordism of class @xmath4 between two knots .
moreover , this invariant is rationally faithful and surjective .
here @xmath62 is one of the usual algebras of feynman diagrams known from the theory of finite type invariants , but graded by the _ grope degree_. more precisely , @xmath62 is the @xmath0-vector space generated by connected uni - trivalent graphs of grope degree @xmath63 , with at least one univalent vertex and a cyclic ordering at each trivalent vertex .
the relations are the usual ihx and as relations .
the grope degree is the vassiliev degree ( i.e. half the number of vertices ) _ plus _ the first betti number of the graph .
observe that both relations preserve this new degree .
read backwards , our results give an interpretation of the kontsevich integral in terms of the geometrically defined equivalence relations of grope cobordism .
-5 mm we now turn to the 4-dimensional aspects of the theory .
it may look like the end of the story to realize that any knot with trivial arf invariant bounds a grope of arbitrary big class embedded in @xmath64 , ( * ? ? ?
* prop.3.8 ) .
however , group theory has more to offer than the lower central series . recall that the _ derived series _ of a group @xmath5 is defined inductively by the iterated commutators @xmath65 \text { and } g^{(h)}:=[g^{(h-1)},g^{(h-1 ) } ]
\text { for } h>1.\ ] ] accordingly , we may define _ symmetric gropes _ with their complexity now measured by _
, satisfying the following defining property : a continuous map @xmath1 represents an element in @xmath66 if and only if it extends to a continuous map of a symmetric grope of height @xmath67 .
thus a symmetric grope of height 1 is just a surface , and a symmetric grope of height @xmath67 is obtained from a bottom surface by attaching symmetric gropes of height @xmath68 to a _ full _ symplectic basis of curves .
this defines symmetric gropes of height @xmath67 with one boundary circle as in figure [ fig : symmetric gropes ] note that a symmetric grope of height @xmath67 is also a grope of class @xmath70 , just like in group theory .
but conversely , not every grope is symmetric .
it should also be clear from figure [ fig : symmetric gropes ] how one defines symmetric gropes with half - integer height ( even though there is no group theoretic analogue ) . in the following definition we attempt to distinguish the terms `` cobordant '' and `` concordant '' in the sense that the latter refers to 4 dimensions , whereas the former was used in dimension 3 , see definition [ def:3d ] .
historically , these terms were used interchangeably , but we hope not to create any confusion with our new distinction . [ def:4d ] two oriented knots in @xmath46 are _ grope concordant _ of height @xmath71 , if there is an embedded symmetric grope of height @xmath67 in @xmath72 $ ] such that its boundary consists exactly of the given knots @xmath73 .
observe that since an annulus is a symmetric grope of arbitrary height we indeed get a filtration of the _ knot concordance group_. this group is defined by identifying two knots which cobound an embedded annulus in @xmath72 $ ] , where there are two theories depending on whether the embedding is smooth or just topological ( and locally flat ) . for grope concordance
the smaller topological knot concordance group is the more natural setting : a locally flat topological embedding of a grope ( defined by the obvious local model at the singular points ) can be perturbed to become smooth .
this perturbation might introduce many self - intersection points in the surface stages of the grope .
however , these new singularities are arbitrarily small and thus they can be removed at the expense of increasing the genus of the surface stage in question but without changing the height of the grope . in joint work with tim cochran and kent orr @xcite , we showed that all known knot concordance invariants fit beautifully into the scheme of grope concordance !
in particular , all known invariants turned out to already be invariants of grope concordance of height @xmath74 : 1 .
@xmath54 have the same arf invariant if and only if they are grope concordant of height @xmath75 ( or class 3 ) .
@xmath54 are _ algebraically concordant _ in the sense of levine @xcite ( i.e. all twisted signatures and twisted arf invariants agree ) if and only if they are grope concordant of height @xmath76 .
if @xmath54 are grope concordant of height @xmath74 then they have the same casson - gordon invariants @xcite .
the third statement includes the generalizations of casson - gordon invariants by gilmer , kirk - livingston , and letsche . in @xcite
we prove an even stronger version of the third part of theorem [ thm:4d ] .
namely , we give a weaker condition for a knot @xmath7 to have vanishing casson - gordon invariants : it suffices that @xmath7 is _ @xmath77-solvable_. all the obstruction theory in @xcite is based on the definition of _
@xmath78-solvable _ knots , @xmath71 , which we shall not give here .
suffice it to say that this definition is closer to the algebraic topology of @xmath8-manifolds than grope concordance . in ( * ? ? ?
* thm.8.11 ) we show that a knot which bounds an embedded grope of height @xmath79 in @xmath64 is @xmath78-solvable .
it should come as no surprise that the invariants which detect grope concordance have to do with solvable quotients of the knot group .
in fact , the above invariants are all obtained by studying the witt class of the intersection form of a certain @xmath8-manifold @xmath80 whose boundary is obtained by @xmath9-surgery on the knot .
the different cases are distinguished by the fundamental group @xmath81 , namely so the previously known concordance invariants stopped at solvable groups which are extensions of abelian by abelian groups . to proceed further in the understanding of grope ( and knot ) concordance , one must be able to handle more complicated solvable groups .
a program for that purpose was developed in @xcite by giving an elaborate boot strap argument to construct inductively representations of knot groups into certain _
universal solvable _ groups . on the way , we introduced blanchfield duality pairings in solvable covers of the knot complement by using noncommutative localizations of the group rings in question .
the main idea of the boot strap is that a particular choice of `` vanishing '' of the previous invariant _ defines _ the map into the next solvable group ( and hence the next invariant ) . in terms of gropes
this can be expressed quite nicely as follows : pick a grope concordance of height @xmath82 and use it to construct a certain @xmath8-manifold whose intersection form gives an obstruction to being able to extend that grope to height @xmath83 .
there is an obvious technical problem in such an approach , already present in @xcite : to show that there is no grope concordance of height @xmath83 , one needs to prove non - triviality of the obstruction _ for all _ possible gropes of height @xmath67 .
one way around this problem is to construct examples where the grope concordances of small height are in some sense unique .
this was done successfully in @xcite for the level above casson - gordon invariants , and in @xcite we even obtain the following infinite generation result .
let @xmath84 be the graded quotient groups of knots , grope concordant of height @xmath67 to the unknot , modulo grope concordance of height @xmath83 .
then the results of levine and casson - gordon show that @xmath85 and @xmath86 are not finitely generated .
the easiest example of a non - slice knot with vanishing casson - gordon invariants is given in ( * ? ? ? * fig.6.5 ) .
as explained in the introduction , the last step in the proof of theorem [ thm : cot2 ] is to show that the intersection form of the @xmath8-manifold in question is nontrivial in a certain witt group .
our new tool is the von neumann signature which has the additional bonus that it takes values in @xmath88 , which is not finitely generated as an abelian group .
this fact makes the above result tractable .
we can not review any aspect of the von neumann signature here , but see ( * ? ? ?
* sec.5 ) .
last but not least , it should be mentioned that we now know that for every @xmath82 the groups @xmath84 are nontrivial . this work in progress @xcite uses as the main additional input the cheeger - gromov estimate for von neumann @xmath39-invariants @xcite in order to get around the technical problem mentioned above .
it is very likely that non of the groups @xmath89 are finitely generated . | we explain new developments in classical knot theory in 3 and 4 dimensions , i.e. we study knots in 3-space , up to isotopy as well as up to concordance . in dimension
3 we give a geometric interpretation of the kontsevich integral ( joint with jim conant ) , and in dimension 4 we introduce new concordance invariants using von neumann signatures ( joint with tim cochran and kent orr ) .
the common geometric feature of our results is the notion of a grope cobordism .
4.5 mm * 2000 mathematics subject classification : * 57m25 , 57n70 , 46l89 . * keywords and phrases : * knot , signature , von neumann algebra , concordance , kontsevich integral , grope . |
molecular outflows are one of the first manifestations of the formation of a new star .
they are thought to play an essential role in the removal of angular momentum from collapsing dense cores , which eventually enable them to contract to stellar sizes .
although the details remain unclear , the general consensus is that outflows are driven by rotating magnetic fields anchored to the disk - star system @xcite .
accretion disks are known to rotate , so the material ejected from them should inherit a toroidal angular momentum component @xcite . indeed , observational evidence for this component have been presented for several outflows driven by young stellar objects @xcite . arguably the most direct evidence has been presented by launhardt et al .
( 2009 ) who reported millimeter interferometric observations of the isolated bok globule cb 26 .
their @xmath0co(2 - 1 ) line observations revealed the presence of a systematic velocity gradient perpendicular to the flow axis along its entire length .
zapata et al ( 2010 ) , on the other hand , reported three independent @xmath0co(2 - 1 ) and so(6 - 5 ) observations toward the ori - s6 molecular outflow .
all three observations revealed velocity asymmetries about the outflow axis which are suggestive of rotation at different size scales .
ic 348 is a young star cluster ( @xmath2 2 myr ) located near the eastern edge of the perseus dark cloud complex at about 300 pc from the sun ( muench et al .
2007 ; luhman et al . 2003 ) .
it contains more than a dozen protostars , many of them driving outflows .
the ic 348-mms / smm2 source is a dense condensation within that region which drives a large collimated north - south co ouflow @xcite .
the optical and infrared counterpart of this molecular outflow is the herbig - haro system hh 797 @xcite .
the low radial velocities displayed by the outflow ( blue to the north and red to the south ) and its large extension ( @xmath2 0.5 pc ) suggest that it is nearly in the plane of the sky . in this study , we present millimeter observations , made with the submillimeter array ( sma ) , of the ic 348-mms / smm2 object and its associated hh 797 outflow .
these interferometric observations reveal velocity asymmetries about the flow axis over the entire length of the flow , that will be interpreted as jet rotation .
m spitzer - irac image ( color scale ) , and the 1.4 mm continuum emission from the object ic 348-mms / smm2 ( blue contours ) .
the white contours range from 15% to 80% of the emission peak in steps of 5% .
the emission peak for the co(2 - 1 ) observations is 31 jy beam@xmath1 km s@xmath1 .
the blue contours range from 40% to 90% of the peak emission in steps of 9% .
the emission peak for the millimeter continuum observations is 160 mjy beam@xmath1 .
the synthesized beam of the line and continuum observations is shown in the lower right corner . ]
the observations were obtained from the sma archive , and were collected on november 2005 , when the array was in its compact configuration .
the 21 independent baselines in the compact configuration ranged in projected length from 10 to 55 k@xmath3 .
the phase reference center for the observations was at @xmath4 = @xmath5 , @xmath6 = @xmath7 .
two frequency bands , centered at 220.538 ghz ( lower sideband ) and 230.538 ghz ( upper sideband ) were observed simultaneously .
the observations were made in mosaicing mode using the half - power point spacing between field centers and thus covering the entire hh 797 outflow .
the primary beam of the sma at 230 ghz has a fwhm @xmath8 .
the sma digital correlator was configured in 24 spectral windows ( `` chunks '' ) of 104 mhz and 128 channels each .
this provides a spectral resolution of 0.812 mhz ( @xmath2 1.1 km s@xmath1 ) per channel .
the zenith opacity ( @xmath9 ) , measured with the nrao tipping radiometer located at the nearby caltech submillimeter observatory , fluctuated between 0.12 and 0.19 , indicating reasonable weather conditions during the observations .
observations of uranus provided the absolute scale for the flux density calibration .
the gain calibrators were the quasars 3c 111 and 3c 84 , while 3c 273 was used for bandpass calibration .
the uncertainty in the flux scale is estimated to be between 15 and 20@xmath10 , based on the sma monitoring of quasars .
further technical descriptions of the sma and its calibration schemes can be found in @xcite .
the data were calibrated using the idl superset mir , originally developed for the owens valley radio observatory ( ovro , * ? ? ?
* ) and adapted for the sma.cqi/mircook.html . ]
the calibrated data were imaged and analyzed in the standard manner using the miriad and karma @xcite softwares .
a 1.4 mm continuum image was obtained by averaging line - free channels in the upper sideband .
we set the robust parameter of the task invert to 0 to obtain an optimal compromise between resolution and sensitivity .
the resulting r.m.s .
noise for the continuum image was about 7 mjy beam@xmath1 at an angular resolution of @xmath11 @xmath12 @xmath13 with a p.a .
= @xmath14 . the r.m.s .
noise in each channel of the spectral line data was about 70 mjy beam@xmath1 at the same angular resolution . .
the color - scale bars on the right indicate the lsr velocities in km s@xmath1 .
the e - w dashed lines mark the positions of the position - velocity cuts shown in figure 3 .
the n - s continuous black line represents approximately the position of the outflow axis .
the synthesized beam of the line and continuum observations is shown in the lower right corner . ]
in figure 1 , we show the 1.4 mm continuum map and the integrated intensity @xmath0co(2 - 1 ) line emission from hh 797 overlaid on the spitzer - irac 4.5 @xmath15 m image .
this irac band contains several h@xmath16 transitions which trace shock - excited material @xcite associated ( in the present case ) with the supersonic flow in hh 797 .
this image reveals the highly collimated outflow ( with an opening angle of about 6@xmath17 at a position angle of @xmath1830@xmath17 ) already mapped in co lines by @xcite at a lower spatial resolution . from our observations ,
it is evident that the @xmath0co(2 - 1 ) emission traces the innermost portion ( @xmath2 0.1 pc ) of the outflow .
the blueshifted radial velocities run from @xmath182 to @xmath198 km s@xmath1 and the redshifted ones from @xmath1910 to @xmath1920 km s@xmath1 ; the systemic lsr radial velocity of the ambient molecular cloud is at about @xmath19 9 km s@xmath1 .
these velocity ranges are in very good agreement with those reported by @xcite .
most of the receding radial velocities ( redshifted ) of the outflow are located toward the south , while the approaching radial velocities ( blueshifted ) are toward the north .
we note that the optical and infrared components of hh 797 have been extensively discussed by @xcite , @xcite , and @xcite .
the outflow emanates from ic 348-mms / smm2 , a class 0 object @xcite .
our observations resolve the dust emission from this source , with a deconvolved size of ( 3.0@xmath20 ) @xmath12 ( 1.5@xmath21 ) , at p.a .
@xmath22 @xmath23 .
this implies a physical size of about 900 @xmath12 450 au ( assuming a distance of 300 pc ) and an orientation almost exactly perpendicular to the outflow axis .
the source is centered at the @xmath4 = @xmath24 , @xmath6 = @xmath25 and has a flux density at 1.4 mm of 220 @xmath26 15 mjy .
assuming optically thin isothermal dust emission , a gas - to - dust ratio of 100 , a dust temperature of 40 k , a dust mass opacity @xmath27 = 1 @xmath28 g@xmath1 , and an emissivity index @xmath29 , we estimate a total mass for the source of 0.09 m@xmath30 .
this combination of properties suggest that the emission seen at 1.4 mm is dominated by the accretion disk , although a contribution from the inner envelope might also be present . in objects where the masses of the disks and central stars have been estimated , they usually are in a ratio of order m@xmath31/m@xmath32 @xmath2 10 @xcite .
this suggests that the central protostar in ic 348-mms / smm2 is roughly of solar mass .
we searched for a centimeter counterpart using the very large array archive data from project s9056 , taken on 2008 march 13 , 14 , 18 and 19 .
we did not detect a source , setting 4-@xmath33 upper limits of 0.05 and 0.06 mjy at 4.86 and 8.46 ghz , respectively . in figure 2 , we present the intensity - weighted velocity map of the @xmath0co(2 - 1 ) emission , which confirms that most of the redshifted velocities are found to the south , while the blueshifted velocities are to the north .
interestingly , this figure also shows a clear gradient _ across _ the outflow , detected along the entire length of the flow .
this gradient is such that , at a given distance from the protostar , the velocity of the gas near the eastern edge of the jet is systematically more red - shifted than that of the gas on the western edge .
this asymmetry is confirmed by the position - velocity ( pv ) diagrams for four cuts across the outflow ( figure 3 ) , which reveal the spatial structure of the gas across the flow as a function of the radial velocities .
one can observe that in all cases , a `` v''-shape ( or , perhaps , more accurately , a triangular shape ) velocity pattern is apparent .
near the axis of the jet , high and low velocities coexist , extending from the systemic velocity up to about @xmath19 2 km s@xmath1 on the north ( blue ) side and about @xmath19 16 km s@xmath1 on the south ( red ) side . in contrast , near the edges of the flow , only comparatively low velocities are detected .
if the velocity of the gas were parallel to the direction of the outflow , one would expect these v shapes to be symmetric with respect to the axis of the flow . in the present case , however , there is a clear asymmetry .
consider , for instance , the cuts at positions b1 and b2 . near the western edge ( at angular offset @xmath34 ) , the bulk gas is at a velocity of about @xmath19 8 km s@xmath1 ( i.e. slightly below the systemic velocity , as appropriate for the blue side ) . but near the eastern edge ( angular offset @xmath35 ) , the gas reaches redshifted velocity of about @xmath19 11 km s@xmath1 .
similar patterns are seen at positions r1 and r2 and have been reported in two other collimated outflows ( orion - s6 and hh 212 ; * ? ? ?
* ; * ? ? ?
* ) where they have been interpreted as evidence for outflow rotation . in the present case ,
the amplitude of the velocity asymmetry is about 2 km s@xmath1 .
we should mention , finally , that the structure of the emission at the systemic velocity ( where the outflow emission is blended with the extended ambient emission ) is poorly reconstructed by our interferometric data .
the apparent lack of continuity in the pv diagrams at velocity @xmath19 9 km s@xmath1 is a result of this effect .
the most natural interpretation of the velocity asymmetries observed here is jet rotation .
let us , however , consider and discuss alternative explanations . the most obvious alternate interpretation would be that the flow driven by ic 348-mms / smm2 is in fact the superposition of two nearly parallel outflows , not spatially resolved by our observations , and powered by a central tight binary system . in a recent study , @xcite present a model of such a situation , and show that the parallel outflows eventually merge , resulting in a persistent kink in the final structure .
moreover , the precession induced by the binarity quickly lead to bending jet trajectories . although they can not be entirely ruled out ,
neither of these effects are apparent in our sma data .
in addition , there is currently no independent direct evidence supporting the binarity of ic 348-mms / smm2 .
a second possibility to consider is that of a precessing ( non - rotating ) jet / outflow system .
this possibility would naturally fit with the `` wavy '' morphology of the outflow at large distances from the source , as revealed at infrared wavelengths ( see figure 1 ) .
this situation might plausibly mimic rotation because the jet would be the superposition of successive events where gas was ejected in different directions at different times .
such a scenario , however , does not easily explain the symmetric v - shape morphology seen in the pv diagrams of hh 797 ( which are consistent , instead , with a high velocity jet along the outflow axis ) .
in addition , precession should produce a point - symmetric situation : if the asymmetries were east to west along the blue outflow lobe , they should be west to east along the red lobe . in the present case , however , the asymmetry is consistently east to west along the entire length of the flow . since alternative interpretations fail to reproduce the observed kinematics of the hh 797 flow , we interpret the observed velocity asymmetries as rotation .
considering the situation where a jet is launched from a rotating protostellar disk and then accelerated and collimated by mhd forces , @xcite provide a formula relating the jet properties measured at large distances from the disk to the position ( the `` footpoint '' ) on the disk where the jet is anchored .
@xmath36 here @xmath37 is the observed radial distance of the jet shell from the flow axis , @xmath38 the radius on the disk from where that shell s material leaves , v@xmath39 and v@xmath40 are the toroidal and poloidal velocities observed for the shell at @xmath37 , and @xmath41 the mass of the ( proto)star at the center of the disk . in this case , we assume from observation @xmath37 @xmath2 1000 au ( about 4@xmath42 ) , v@xmath39 @xmath2 2 km s@xmath1*cos @xmath43 , v@xmath40 @xmath2 10/sin @xmath43 km s@xmath1 and @xmath41 @xmath2 1 m@xmath30 , with @xmath43 the unknown angle between flow direction and plane of the sky .
this results in a footpoint radius of 100*(sin @xmath44*(cos @xmath45 au .
since the apparent poloidal flow velocity of @xmath2 10 km s@xmath1 relative to ambient is relatively low , it seems likely that @xmath43 is small . for @xmath46 20@xmath17 the footpoint radius would drop to below 20 au , a reasonable value that fits with the innermost part of the above - mentioned disk dimensions .
the value @xmath37 = 20 au is very similar to the one found by @xcite for the rotating outflow ori - s6 energized by the young class 0 protostar 139 - 409 and the one estimated by @xcite for the the outflow cb 26 powered by a more evolved low - mass t - tauri star . on the other hand ,
@xcite have found much smaller values for @xmath37 in outflows powered for both class 0/1 protostars .
thus , at this point , we can not role out any conclusion about the variation of @xmath37 with time or some similarities on these values .
in this paper , we have reported @xmath0co(2 - 1 ) line observation of hh 797 , and discovered velocity asymmetries about the flow axis with an amplitude roughly on the order of 2 km s@xmath1 over distances of about 1000 au .
the same velocity asymmetries are found on both sides of the outflow , and we interpret them as evidence for jet rotation , thus , hh 797 appears to be promising laboratory for future studies of magneto - centrifugal models of jet acceleration .
g. p. , l.a.z , l. l. and l. f. r. acknowledge the financial support from dgapa , unam , and conacyt , mxico .
l. l. is indebted to the alexander von humboldt stiftung for financial support .
we are grateful with the anonymous referee whose recomendations helped us to improve the paper .
anderson , j. m. , li , z .- y . , krasnopolsky , r. , & blandford , r. d. 2003 , , 590 , l107 choi , m. , kang , m. , & tatematsu , k. 2011 , , 728 , l34 eislffel , j. , froebrich , d. , stanke , t. , & mccaughrean , m. j. 2003 , , 595 , 259 gooch , r. 1996 , astronomical data analysis software and systems v , 101 , 80 guilloteau , s. , & dutrey , a. 1998 , , 339 , 467 hatchell , j. , & dunham , m. m. 2009 , , 502 , 139 fendt , c. 2011 , , 737 , 43 ho , p. t. p. , moran , j. m. , & lo , k. y. 2004 , , 616 , l1 launhardt , r. , pavlyuchenkov , y. , gueth , f. , et al .
2009 , , 494 , 147 lee , c .- f .
, ho , p. t. p. , bourke , t. l. , et al .
2008 , , 685 , 1026 lee , c .- f . ,
hirano , n. , palau , a. , et al .
2009 , , 699 , 1584 murphy , g. c. , lery , t. , osullivan , s. , et al .
2008 , , 478 , 453 mccaughrean , m. j. , rayner , j. t. , & zinnecker , h. 1994 , , 436 , l189 pudritz , r. e. , ouyed , r. , fendt , c. , & brandenburg , a. 2007 , protostars and planets v , 277 shang , h. , li , z .- y . , & hirano , n. 2007 , protostars and planets v , 261 scoville , n. z. , carlstrom , j. e. , chandler , c. j. , et al .
1993 , , 105 , 1482 smith , m. d. , & rosen , a. 2005 , , 357 , 1370 schreyer , k. , semenov , d. , henning , t. , & forbrich , j. 2006 , , 637 , l129 tafalla , m. , kumar , m. s. n. , & bachiller , r. 2006 , , 456 , 179 walawender , j. , bally , j. , & reipurth , b. 2005 , , 129 , 2308 walawender , j. , bally , j. , kirk , h. , et al .
2006 , , 132 , 467 rodrguez , l. f. , dalessio , p. , wilner , d. j. , et al .
1998 , , 395 , 355 zapata , l. a. , schmid - burgk , j. , muders , d. , et al . 2010 , , 510 , a2 | we present @xmath0co(2 - 1 ) line and 1.4 mm continuum archival observations , made with the submillimeter array , of the outflow hh 797 located in the ic 348 cluster in perseus .
the continuum emission is associated with a circumstellar disk surrounding the class 0 object ic 348-mms / smm2 , a very young solar analog .
the line emission , on the other hand , delineates a collimated outflow , and reveals velocity asymmetries about the flow axis over the entire length of the flow .
the amplitude of velocity differences is of order 2 km s@xmath1 over distances of about 1000 au , and we interpret them as evidence for jet rotation although we also discuss alternative possibilities .
a comparison with theoretical models suggests that the magnetic field lines threading the protostellar jet might be anchored to the disk of a radius of about 20 au . |
consider a finite chain , say @xmath4 under the natural ordering and let @xmath5 and @xmath6 be the full transformation semigroup and the semigroup of all partial transformations on @xmath7 , under the usual composition , respectively .
we shall call a partial transformation @xmath8 , _ order - decreasing ( order - increasing ) _ or simply _ decreasing ( increasing ) _ if @xmath9 for all @xmath10 in @xmath11 , and @xmath12 is _ order - preserving _ if @xmath13 implies @xmath14 for @xmath15 in @xmath11 .
this paper establishes bijections between certain types of lattice paths and certain subsemigroups of @xmath0 , the semigroup of all order - preserving partial transformations of @xmath7 .
thus combinatorial problems of these lattice paths can be translated into combinatorial problems of these subsemigroups of @xmath0 , and vice - versa .
+ various enumerative problems of an essentially combinatorial nature have been considered for certain classes of semigroups of transformations .
for example , it is well known and indeed obvious that @xmath5 and @xmath6 have orders @xmath16 and @xmath17 , respectively . only slightly less obvious
are their number of idempotents given by @xmath18 the first usually attributed to tainiter @xcite is actually ex 2.2.2(a ) in clifford and preston @xcite .
the second can be deduced easily via vagner s method of representing a partial transformation by a full transformation @xcite , which has been used to good effect by garba @xcite . the following list ( which is by no means exhaustive ) of papers and books @xcite each contains some interesting combinatorial results pertaining to semigroups of transformations .
combinatorial properties of @xmath19 , the semigroup of all decreasing and order - preserving partial transformations of @xmath7 have been investigated recently by laradji and umar @xcite .
one surprising outcome from @xcite is the discovery of many integer sequences that are not yet recorded or have just recently been recorded in sloane s encyclopaedia of integer sequences @xcite .
the advantage of the approach in this paper is that it establishes a connection between the theories of lattice paths and partial transformations .
+ in section 2 , we give the necessary definitions that we need in the paper as well as show that the semigroup @xmath19 of all decreasing and order - preserving partial transformations of @xmath7 , is a disjoint union of two subsemigroups of the same cardinality . in section 3
, we establish a bijection between a set of certain lattice paths and @xmath19 .
hence we obtain the order of @xmath19 as the large or double schrder number @xcite , and the order of the two subsemigroups as the small schrder number .
we also discuss several consequences of this bijection . in section 4 ,
we gather some remarks concerning the results of the paper and we also give an alternative proof of the result that set of all idempotents of @xmath19 is of cardinality @xmath20 .
for standard terms and concepts in transformation semigroup theory see @xcite or @xcite .
we now recall some definitions and notations to be used in the paper .
consider @xmath4 and let @xmath21 be a partial transformation .
we shall denote by @xmath11 and @xmath22 , the _ domain _ and _ image set _ of @xmath12 , respectively .
the semigroup @xmath6 , of all partial transformations contains two important subsemigroups which have been studied recently .
they are @xmath23 and @xmath0 the semigroups of all order - decreasing and respectively order - preserving partial transformations , ( see @xcite and @xcite ) .
now let @xmath24 be the semigroup of all decreasing and order - preserving partial transformations of @xmath7 .
next let @xmath25 be the set of all maps in @xmath19 all of whose domain does contain the element 1 .
then evidently we have the following result .
both @xmath26 and its set complement @xmath27 are subsemigroups of @xmath19 .
moreover , @xmath28 .
less evidently , we have @xmath29 .
define a map @xmath30 from @xmath26 into @xmath27 by @xmath31 where @xmath32 it is clear that @xmath30 is a bijection since @xmath33 for all @xmath34 in @xmath27 , and @xmath35 for all @xmath12 in @xmath26 .
informally , we may argue as follows : to get @xmath27 from @xmath26 we delete 1 from @xmath11 ( for each @xmath36 ) , if @xmath37 ; otherwise we delete 1 from @xmath38 as well .
conversely , to get @xmath26 from @xmath27 we extend each @xmath39 by defining @xmath35 and @xmath40 for all @xmath41 .
our main objective in this section is to give an alternative proof for @xmath42 .
we begin our investigation by considering lattice paths in the cartesian plane that start at @xmath1 , end at @xmath43 , contain no points above the line @xmath44 , and composed only of steps @xmath45 and @xmath46 , i.e. , @xmath47 and @xmath48 .
we shall call such lattice paths _ type i _ , in this paper .
the diagrams in figure 1 illustrate all such paths in the @xmath49 and @xmath50 squares , respectively .
the total number of such paths is known to be the large or double schrder number @xmath51 , @xcite . to establish a bijection between the set of type i paths ( in an @xmath52 square ) and the set of all decreasing and order - preserving partial transformations of @xmath7 , we first make the following observation : 1 . in each lattice path from @xmath1
to @xmath2 there are exactly same number of horizontal steps as there are vertical steps .
to see this , we note that a horizontal step corresponds to the map @xmath53 , a vertical step corresponds to the map @xmath54 , and a diagonal step corresponds to the map @xmath55 .
it is easy to check that @xmath56 and @xmath57 . now
if a path starts from @xmath1 and ends at @xmath43 using only the above steps then each such path can be written as a product : @xmath58 , and so we have @xmath59 hence @xmath60 , which implies @xmath61 .
next , we note that it is convenient ( in this section ) to express @xmath12 in @xmath62 ( with base set @xmath63 ) as @xmath64 where the @xmath65 s are distinct , but the @xmath66 s are not necessarily distinct .
moreover , we may also assume that @xmath67 and @xmath68 .
thus @xmath69 however , we refer to the sequence @xmath70 as the _ simage _ of @xmath12 , denoted by @xmath71 ( note that if all the @xmath66 s are distinct then @xmath72 coincides with @xmath73 , otherwise @xmath74 ) now to each vertical step from @xmath75 to @xmath76 in an arbitrary type i path ( in an @xmath52 square ) we put @xmath77 in @xmath78 , and to each horizontal step from @xmath75 to @xmath79 we put @xmath77 in @xmath80 .
the domain is then arranged in a strictly increasing order while the simage is arranged in a nondecreasing order , and by virtue of ( ob1 ) this gives rise to a unique order - preserving map .
two examples should make these ideas more clear .
the path given in figure 2 implies @xmath81 and @xmath82 .
thus the associated order - preserving map is @xmath83 and the path given in figure 3 implies @xmath84 and @xmath85 .
thus the associated order - preserving map is @xmath86 moreover , the condition that type i paths never cross above the diagonal line @xmath44 , forces these associated order - preserving maps to be decreasing as well .
in fact , at any level , a horizontal step will take us 1-unit away from the diagonal line @xmath44 , and a vertical step will take us 1-unit closer to the diagonal line @xmath44 .
thus , since our target is the point @xmath2 on this diagonal line we must take vertical steps at higher levels to compensate for each horizontal step , and so a decreasing map results . finally , note that the unique all diagonal path corresponds to the empty map .
hence we have shown that to every type i lattice path in an @xmath52 square there corresponds a unique partial order - preserving map .
conversely , we show that every @xmath87 corresponds to a type i path in an @xmath52 square . for a given @xmath12 in @xmath62
we first express it as in ( 3.1 ) and write @xmath88 for some @xmath89 in @xmath90 . to construct an associated type i path , first note
that since , in general , there may be repetitions in @xmath80 , we shall consider it as consisting of blocks of subsequences , where an @xmath10-block consists only of @xmath10 in @xmath91 , repeated say @xmath92 times @xmath93 .
thus we may write @xmath94 where @xmath95 .
( note that @xmath96 next we label the horizontal rows of the @xmath97 square from the bottom to the top , starting from 0 to @xmath98 . now starting from @xmath99
if @xmath100 take as many horizontal steps as the length of the @xmath101-block , otherwise take a diagonal step to the next level ( level 1 ) .
note that the order - decreasing property guarantees that @xmath102 , however there is no guarantee that @xmath103 ( for @xmath104 , but this will not be of any disadvantage . in general , at level @xmath92 @xmath105 , check 1 . if @xmath106 and @xmath107 , take as many horizontal steps as the length of the @xmath92-block followed by a vertical step to the next level ; 2 . if @xmath106 and @xmath108 , take as many horizontal steps as the length of the @xmath92-block followed by a diagonal step to the next level ; 3 . if @xmath109 , take a vertical or diagonal step to the next level depending on whether @xmath110 or @xmath111 , respectively .
now since in ( i ) and ( ii ) we give priority to horizontal step(s ) over vertical and diagonal steps , it follows by the order - decreasing property that at level @xmath92 we must have had at least as many horizontal steps as there are vertical steps ( up to level @xmath112 .
this in turn guarantees that our paths never overshoot to cross the diagonal line @xmath44 .
moreover the bijection between @xmath11 and @xmath80 guarantees that our paths always end up at @xmath2 as required .
an example would be quite appropriate .
consider the map @xmath113 then @xmath114 now start from @xmath1 and take 3 horizontal steps , since the first block of @xmath80 is a 0-block of length 3 .
we are still at level 0 , however , we take a vertical step to level 1 since @xmath115 . now since @xmath116 , we take a diagonal step to level 2 .
next , since @xmath117 we take a vertical step to level 3 . then since @xmath118 we take a vertical step to level 4 .
now we take 2 horizontal steps and then a diagonal step since @xmath119 .
we are now at level 5 , from where we take a vertical step to level 6 since @xmath120 . finally , we take another vertical step to level 7 , since @xmath121 .
thus we have the path indicated in figure 4 .
we have now established the main result of this section : [ thrm31 ] there is a one - one correspondence between the set of all decreasing and order - preserving partial transformations of an @xmath122-element chain and the set of all type i lattice paths in an @xmath52 square .
an immediate consequence of theorem [ thrm31 ] is that it furnishes us with the order of @xmath42 .
however , before we formally state this result we first deduce from pergola and sulanke @xcite and stanley @xcite that the large ( or double ) schrder number denoted by @xmath51 could be defined as @xmath123 moreover , @xmath51 satisfies the recurrence @xmath124 for @xmath125 , with initial conditions @xmath126 and @xmath127 .
the ( small ) schrder number is usually denoted by @xmath128 and defined as @xmath129 and so it satisfies the same recurrence as @xmath51 .
* theorem 2.10)[thrm32 ] let @xmath19 be as defined in ( 2.1 ) .
then @xmath130 , the double schrder number .
several consequences of this result will now be exhibited .
for the semigroup @xmath26 ( defined in ( 2.2 ) ) and its set complement @xmath27 , we now have @xmath131 , the ( small ) schrder number . it is also known ( see for example , pergola and sulanke@xcite ) that the number of type i lattice paths ( in an @xmath52 square ) without a diagonal step is the @xmath122-@xmath132 catalan number : @xmath133 . however , since clearly such paths correspond to full transformations , higgins ( * ? ? ?
* theorem 3.1 ) follows immediately : [ thrm34 ] let @xmath134 be the semigroup of all decreasing and order - preserving full transformations of @xmath7
. then @xmath135 , the @xmath122-@xmath132 catalan number .
further , we give a sequence of results which arise naturally in the semigroup context which ( with the exception of theorem [ thrm38 ] ) may not have been asked yet in lattice path theory .
first , as in @xcite we define the numbers @xmath136 , @xmath137 and @xmath138 , respectively by @xmath139 @xmath140 @xmath141 where @xmath142 is an arbitrary semigroup of transformations . then from ( 3.3 ) and
* proposition 2.8 ) we deduce [ thrm35 ] let @xmath136 be as defined in ( 3.3 ) , where @xmath143 .
then the number of type i lattice paths ( in an @xmath97 square ) with exactly r vertical steps is @xmath144 from ( 3.4 ) and ( * ? ? ?
* proposition 2.7 ) we deduce [ thrm36 ] let @xmath137 be as defined in ( 3.4 ) , where @xmath143 .
then the number of type i lattice paths ( in an @xmath97 square ) in which the last horizontal segment is at level @xmath145 is @xmath146 , where @xmath147 , @xmath148 and for @xmath149 , @xmath150 from ( 3.4 ) and ( * ? ? ?
* proposition 3.10 ) we deduce [ thrm37 ] let @xmath137 be as defined in ( 3.4 ) , where @xmath151 .
then the number of type i lattice paths without diagonal steps ( in an @xmath97 square ) in which the last horizontal segment is at level @xmath145 is @xmath146 , where for @xmath152 , @xmath153 from ( 3.5 ) and ( * ? ? ?
* proposition 3.6 ) we deduce [ thrm38 ] let @xmath154 be as defined in ( 3.5 ) , where @xmath151 .
then the number of type i lattice paths without diagonal steps ( in an @xmath97 square ) with exactly @xmath89 horizontal segments is @xmath138 , where for @xmath155 , @xmath156 define a subsemigroup of @xmath0 as @xmath157 now if we consider the ( unrestricted ) lattice paths in the sense that our lattice paths consist of the same steps as the type i lattice paths but they can now cross the diagonal line many times if necessary then the total number of such paths is known to be the central delannoy number , @xmath158 @xcite , where the ( arbitrary ) delannoy numbers @xmath159 are given by @xmath160 and hence @xmath161 where @xmath162 is a hypergeometric function .
moreover , @xmath158 satisfies the recurrence @xmath163 the proof that there is a bijection between the set of ( unrestricted ) lattice paths in an @xmath97 square and the semigroup @xmath164 is for most part similar to the one given earlier between the set of restricted ( or type i ) lattice paths and the semigroup @xmath19 , since @xmath165 applies to all lattice paths .
however , note that below the diagonal of the @xmath97 square a horizontal step takes us 1-unit away from the diagonal while a vertical step takes us 1-unit closer to the diagonal , but above the diagonal the reverse is the case . in either case ,
the diagonal step is neutral in the sense that it neither takes us closer to nor away from the diagonal .
moreover , in this case we need not worry about our lattice paths crossing the diagonal of the @xmath97 square , as our paths can now cross over several times if necessary . as in the above
we exemplify how to construct a lattice path from a partial transformation , and vice - versa .
the path in figure 5 implies @xmath166 and @xmath167 .
thus the associated partial order - preserving map is @xmath168 similarly , consider the partial order - preserving map @xmath169 it is now routine to verify that it corresponds to the path given in figure 6 .
thus we have established there is a one - one correspondence between the set of all order - preserving partial transformations in @xmath164 and the set of all ( unrestricted ) lattice paths in an @xmath52 square .
immediately we deduce let @xmath164 be as defined in ( 3.6 ) .
then @xmath170 , the n - th central delannoy number . from (
* proposition 2.11 ) we deduce the total number of lattice paths ( in an @xmath97 square ) whose last step is never a horizontal step is @xmath171 , where @xmath172 and for all @xmath173 @xmath174 from ( 3.3 ) and ( * ? ? ?
* proposition 2.7 ) we deduce [ thrm312 ] let @xmath136 be as defined in ( 3.3 ) , where @xmath175 .
then the number of lattice paths ( in an @xmath176 square ) whose last step is not a horizontal step and with exactly r vertical steps is @xmath177 from ( 3.4 ) and ( * ? ? ?
* lemma 2.5 ) we deduce [ thrm313 ] let @xmath137 be as defined in ( 3.4 ) , where @xmath175 .
then the number of lattice paths ( in an @xmath176 square ) without diagonal steps and in which last step is not a horizontal step while the last horizontal segment is at level @xmath145 is @xmath146 , where @xmath147 , @xmath178 and for @xmath149 , @xmath150 from ( 3.5 ) and ( * ? ? ?
* lemma 4.5 ) we deduce [ thrm314 ] let @xmath154 be as defined in ( 3.5 ) , where @xmath175 .
then the number of lattice paths without diagonal steps ( in an @xmath179 square ) , whose last step is not a horizontal step and with exactly @xmath89 horizontal segments is @xmath138 , where @xmath180 , and for @xmath181 , @xmath182 from ( 3.4 ) and ( * ? ? ?
* corollary 3.11 ) we deduce [ thrm315 ] let @xmath137 be as defined in ( 3.4 ) , where @xmath183 .
then the number of lattice paths ( in an @xmath97 square ) without diagonal steps and in which the last horizontal segment is at level @xmath145 is @xmath146 , where for @xmath152 , @xmath184 from ( 3.5 ) and ( * ? ? ?
* corollary 3.10 ) we deduce [ thrm316 ] let @xmath154 be as defined in ( 3.5 ) , where @xmath183 .
then the number of lattice paths without diagonal steps ( in an @xmath185 square ) with exactly @xmath89 horizontal segments is @xmath138 , where for @xmath155 , @xmath186
as far as we know the subsemigroups @xmath26 , @xmath27 and @xmath164 have not been studied .
we can now compose lattice paths via the bijection with certain subsemigroups of @xmath0 , thereby making these lattice paths into algebraic objects .
a more important consequence is we believe , the link of combinatorial questions between lattice paths and partial transformations is now firmly established . as stated in the introduction
the number of idempotents of various classes of semigroups of transformations has been computed . for further results
see @xcite .
the number of idempotents @xmath187 , has been shown in @xcite to be @xmath20 .
it is not very difficult to see that lattice paths in which a diagonal step never succeeds a horizontal segment , and where every length @xmath188 horizontal segment is followed by exactly @xmath188 vertical steps plus some ( may be none ) diagonal steps before another horizontal segment , correspond to idempotents in @xmath0 .
however , while idempotents are natural elements to study in a semigroup , their corresponding paths do not seem to have a natural description .
we would like to take this opportunity to give an alternative easier proof of the formula for @xmath187 , though quite admittedly not via lattice paths this time .
first , we consider @xmath189 then clearly we have @xmath190 where the former corresponds to the empty map and the latter corresponds to the identity map .
more generally , we have if @xmath193 then @xmath194 , by idempotency and so there are @xmath195 idempotents of this type .
if on the other hand @xmath196 then either @xmath197 , for some @xmath198 , of which there are again @xmath195 idempotents of this type ; or @xmath199 , of which there are @xmath200 idempotents of this type , by the order - decreasing property .
hence the result follows . by using lemma [ 4.1 ]
, we have @xmath206+[2e(n-1,1 ) + e(n-1,0)]+[2e(n-1,2 ) + e(n-1,1 ) ] \\ & & + [ 2e(n-1,3 ) + e(n-1,2)]+ ... + [ 2e(n-1,n ) + e(n-1,n-1 ) ] \\ & = & 3e(n-1,0)+3e(n-1,1)+3e(n-1,2)+ ... + 3e(n-1,n-1)-1 \\ & = & 3\sum^{n-1}_{r=0}e(n-1,r ) -1 = 3e_{n-1}-1 .
\end{aligned}\ ] ] | let @xmath0 be the semigroup of all order - preserving partial transformations of a finite chain .
it is shown that there exist bijections between the set of certain lattice paths in the cartesian plane that start at @xmath1 , end at @xmath2 , and certain subsemigroups of @xmath0 .
several consequences of these bijections were discussed . * lattice paths and order - preserving partial transformations * + * a. laradji * + @xmath3 + department of mathematics and statistics + king fahd university of petroleum & minerals + dhahran 31261 + saudi arabia + and + * a. umar * + @xmath3 + department of mathematics and statistics + sultan qaboos university + al - khod pc 123 + sultanate of oman + email : ` alaradjikfupm.edu.sa ` and ` aumarhsqu.edu.om ` keywords : semigroup , subsemigroup , partial order - preserving transformation , partial order - decreasing transformation , lattice path , schrder number , catalan number , central delannoy number .
msc2010 : 20m18 , 20m20 , 05a10 , 05a15 . |
the cradles of planet formation protoplanetary disks are ubiquitous around young stars , at least of low and intermediate masses ( @xmath2 ) .
it is imperative to study disk structure and evolution in detail if one aims to understand initial conditions for planet formation .
these objects are best studied at ( sub-)millimeter wavelengths in thermal dust continuum and various molecular lines tracing their physical and chemical structure . however , being small ( @xmath3 au ) and of relatively low mass ( @xmath4 ) , such disks are hard to study observationally as high spatial resolution and sensitivity are required .
to date only several nearby disks around t tauri and herbig ae / be stars have been spatially resolved in a handful of molecular lines @xcite .
these observational data coupled to advanced modeling allowed us to constrain basic disk parameters such as size , mass , temperature , density distribution , kinematics , and ionization structure ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
an interesting result concerns the vertical temperature gradient that seems to be present in some disks ( e.g. , , ) while absent in others ( e.g. , ) , see @xcite . the situation will change dramatically when the atacama large millimeter array ( alma ) , equipped with 50 12-m antennas , will come into operation in 2012 .
alma will be capable of imaging protoplanetary disks at spatial resolutions up to @xmath5 in a frequency range between 100 and 950 ghz .
this will allow direct detection and characterization of disk instabilities , resulting in clumpy structures ( vorticities , `` spiral arms '' , etc . ) as well as inner gaps and holes induced by forming giant planets @xcite .
the large - scale chemical structure of protoplanetary disks will become accessible in rotational lines of many abundant molecules @xcite . as demonstrated by @xcite , excitation of molecular lines in disks with strong gradients of physical conditions and chemical structure
can be a complicated process and may be hard to interpret .
the main aim of this work is to unravel the potential of alma to distinguish between various temperature and chemical effects in protoplanetary disks , with an emphasis on discerning spatial resolution and integration time needed for this goal .
such observations will allow us to fully explore the validity of state - of - the - art disk chemical models on a global scale _ for the first time _ , and advance our understanding of the disk physics enormously .
in contrast to previous studies of @xcite and @xcite , we take into account both the non - lte excitation and realistic chemical structure .
we adopt a steady - state flared disk model with a radius of 800 au , an accretion rate @xmath6yr@xmath7 , a viscosity parameter @xmath8 , and a mass of @xmath9 @xcite .
the disk has a microturbulent velocity of 0.1 kms@xmath7 and a keplerian rotation profile , @xmath10 .
this disk model resembles the disk around the young t tauri star dm tau located at the distance of @xmath11 pc @xcite .
its thermal and density structure is shown in fig .
[ disk_struc ] ( left and middle panels ) . to mimic the case when the vertical temperature gradient is weak or absent
, we also consider the same disk model , but with a vertical temperature distribution fixed to the value at one pressure scale height .
using both disk models , 5 myr of disk chemical evolution is modeled with a gas - grain chemical model from @xcite . the gas - phase reaction data are adopted from the rate06 database @xcite .
ionization and dissociation of the gas - phase species as well as desorption of surface species due to cosmic rays , stellar x - rays , stellar and interstellar uv radiation are taken into account .
the disk is illuminated by uv radiation from the central star with an intensity @xmath12 at @xmath13 au and by interstellar uv radiation with intensity @xmath14 @xcite .
the x - ray ionization rate in a given disk region is computed according to @xcite .
the adopted cosmic - ray ionization rate is 1.3@xmath1510@xmath16 s@xmath7 @xcite .
molecules are assumed to stick to grains with 100@xmath17 efficiency .
we also consider dissociative recombination of ions on charged dust grains and grain re - charging , with a dust - to - gas mass ratio of @xmath18 . to model surface chemistry on the @xmath19 m olivine grains , a set of surface reactions from @xcite and @xcite is used . in what follows
, we will concentrate on hco@xmath0 a representative abundant molecule that traces the ionization fraction and possesses strong rotational transitions , and which is readily observed in disks @xcite .
the modeled distribution of the hco@xmath0 absolute abundances ( molecules@xmath20 ) in the dm tau disk at 5 myr is shown in fig .
[ disk_struc ] ( right panel ) .
a layered chemical structure with the maximal hco@xmath0 concentration reached at intermediate disk heights is clearly visible .
many observed species have a similar stratification , e.g. , co , cs , hcn , etc . @xcite .
strong turbulent mixing and/or global advection flows can smooth these abundance gradients , leading to more uniformly distributed abundances @xcite .
therefore , we also consider a model with an uniform hco@xmath0 abundance of @xmath21 relative to h@xmath22 .
thus , our study is based on three disk models : ( 1 ) the model with chemical stratification and vertical temperature gradient , ( 2 ) the same model but with uniform abundances , and ( 3 ) the model with uniform abundances and no vertical temperature gradient .
using these three dm tau - like disk models , and the 2d non - lte line radiative transfer code of @xcite with thermal dust continuum included , we synthesize `` ideal '' ( beam - unconvolved ) channel maps .
the collisional rates are taken from the leiden atomic and molecular database @xcite , while dust opacities for silicate grains are from @xcite . at the @xmath23 inclination
the integrated hco@xmath0(4 - 3 ) line width is about 3 kms@xmath7 , with a peak intensity reached at about 0.8 kms@xmath7 ( see fig . 10 in * ? ? ?
the corresponding @xmath24 kms@xmath7 wide velocity channel at @xmath25 kms@xmath7 of the continuum - subtracted hco@xmath0(4 - 3 ) map is shown in fig .
[ chan_maps_3 ] .
this close to edge - on orientation is particularly favorable as the layered and uniform disk chemical structures can be clearly distinguished in the channel maps , if sufficiently small channel widths of @xmath26 kms@xmath7 are used .
all synthetic channel maps reveal a complex pattern and are asymmetric .
the model with both non - zero temperature and abundance gradients has a remarkable `` omega '' shape , where the cold midplane with low hco@xmath0 concentration appears as two intensity `` holes '' ( fig .
[ chan_maps_3 ] , left panel ) .
a steep temperature gradient in radial direction toward the inner disk region is clearly visible , while the vertical temperature gradient can not be fully traced with the hco@xmath0 lines due to strong chemical stratification of this species .
the hco@xmath0(4 - 3 ) emission is thermalized and optically thin in this case .
the two models with uniform abundances have relatively high hco@xmath0 column densities so that self - absorption becomes important . in the model with fixed vertical temperature
this effect leads to a layer with low intensity in the upper part of the map ( fig .
[ chan_maps_3 ] , right panel ) . however , the overall intensity of the hco@xmath0(4 - 3 ) emission is higher compared to the model with layered abundances .
the radial temperature gradient is also prominent in this case .
the model without chemical stratification but with vertical temperature gradient also reveals the self - absorption layer in the upper part of the channel map .
the excitation temperature of the hco@xmath0(4 - 3 ) transition changes strongly with the disk height and has the lowest value in the midplane .
this zone of low intensity appears as a fake `` spiral arm '' in the lower part of the map ( marked as `` midplane '' in fig .
[ chan_maps_3 ] , middle panel ) .
note that such a pattern somewhat resembles the `` omega '' structure for the model with chemical stratification of hco@xmath0 .
while we show only one representative channel , one should bear in mind that analysis of the interferometric data is more reliable when all channels and dust continuum emission are taken into account .
despite the fact that all the models considered can be distinguished in the synthetic channel maps , it is important to study whether alma is able to disentangle the effects of thermal and chemical structure in disks . to simulate the observations with alma
, we use the synthetic hco@xmath0(4 - 3 ) channel maps as an input to the gildas simulation software of @xcite ( version from october 2007 ) . a position close to dm tau
is assumed : @xmath27h 33 m 48s , @xmath28 @xmath29 @xmath30 @xcite .
we focus on the @xmath31 kms@xmath7 channel ( with 120 khz bandwidth or @xmath24 kms@xmath7 velocity width ) .
array configurations with 64 antennas built in gildas are considered .
note that in reality alma will consist of 50 antennas with different configurations , so our analysis will tend to underestimate required observing time by a factor of 1.5 but the simulated spectra will not be significantly affected .
typical weather conditions at the chajnantor plateau are assumed ( see * ? ? ?
* ) , with the following main types of errors leading to noise : 1 ) receiver temperature is 80 k , system temperature is 230 k ( at @xmath32 ghz , see * ? ? ?
* ) , 2 ) random pointing errors during the observation are @xmath33 , 3 ) relative amplitude errors are 3% with a 6% hour@xmath7 drift , 4 ) residual phase noise after calibration is @xmath34 , 5 ) anomalous refraction .
the object is assumed to pass the meridian in the middle of a single observational run .
we use 30 minutes of integration time or increase it such that the deconvolved maps look similar to the input model with chemical gradients .
in contrast to modern interferometric facilities there will be no problem to achieve a good @xmath35 coverage with alma in a fraction of an hour .
the simulated alma channel maps of hco@xmath0(4 - 3 ) at various spatial resolutions for the three adopted models and the inclination angle of @xmath23 are presented in fig .
[ alma ] ( upper 3 rows ) .
all significant features that are present in the `` ideal '' molecular emission spectra appear clearly at the resolution of @xmath36 ( `` zoom - c '' array configuration ) and some are still visible at half the resolution ( `` zoom - d '' configuration ) .
the use of even lower resolution ( `` zoom - e '' ) makes it difficult to disentangle the chemical and thermal features without extensive modeling . the dm tau disk model with layered hco@xmath0 abundances and 2d temperature gradient has the lowest line intensities and hence requires the longest observation of 2 hours , while for the models with uniform hco@xmath0 abundances it can be as short as 30 minutes or less .
in addition , we perform a similar analysis for a face - on orientation with @xmath37 . at such low inclination channel maps for distinct disk models look similar and multi - line , multi - molecule analysis become a must .
the corresponding @xmath38 kms@xmath7 channel of the hco@xmath0(4 - 3 ) channel map for the disk model layered abundances is presented in fig .
[ alma ] ( bottom row ) .
this channel map has a ring - like shape with two emission peaks located around the low - intensity midplane .
the intensity in this channel is by a factor of 2 stronger than for the disk model with @xmath39 and thus the observational time required to detect all these features is only 1 hour with the @xmath36 beam size and less than 30 minutes at lower resolutions .
all major features in the @xmath37 channel map are resolved even with the compact `` zoom - e '' array configuration .
given the importance of multi - line observations and ability of alma to simultaneously observe several transitions of several species in various frequency bands , we continue our study for other observationally important molecular tracers ( cs , hcn , co isotopes ) as well as a smaller disk with a radius of @xmath40 au .
all results are summarized in table [ obs_conditions ] .
the channel maps for other transitions at millimeter wavelengths and for molecules located in the intermediate layer are similar to that of hco@xmath0 , though their intensities are vastly different .
we find that smaller disks are better studied at high frequencies of 400700 ghz with moderately extended array configurations leading to smaller beam sizes than at millimeter wavelengths , albeit longer integration times are required for that .
the contamination of the high - lying emission lines by the optically thick dust continuum from the disk inner regions and non - lte excitation can be an issue for the analysis of these high - frequency data .
the molecular lines also comprise kinematic information .
high - quality alma spectra will allow to derive accurate dynamical masses of the central stars and thus calibrate pre - main - sequence evolutionary track models @xcite , discern gas motions along spiral arms in gravitationally unstable disks @xcite , and trace flows of matter around a forming planet @xcite .
we do not pretend to analyze the alma requirements necessary for such studies as it would deserve a separate publication .
we showed that the alma interferometer will allow us to distinguish the effects of temperature gradients and chemical stratification in disks through molecular line observations , in particular for highly - inclined objects .
we found that moderately extended array configurations ( with baselines of @xmath41 km ) and 0.510 hours of integration time will be necessary to pursue such observations .
the authors are grateful to an anonymous referee for valuable comments and suggestions .
this research has made use of nasa s astrophysics data system .
van dishoeck , e. f. 1988 , rate coefficients in astrochemistry .
proceedings of a conference held in umist , manchester , united kingdom , september 21 - 24 , 1987 .
editors , t.j .
millar , d.a .
williams ; publisher , kluwer academic publishers , dordrecht , boston , 1988 .
isbn # 90 - 277 - 2752-x .
lc # qb450 .r38 1988 .
, 1988 , 49 lccllll hco@xmath0(1 - 0 ) & 89 & 30 & zoom - c ( @xmath42h ) & zoom - c ( @xmath43h ) & zoom - a / b ( @xmath44h ) & zoom - a / b ( @xmath44h ) + c@xmath45o(2 - 1 ) & 220 & 75 & zoom - d ( @xmath46h ) & zoom - c ( @xmath47h ) & zoom - c ( @xmath48h ) & zoom - c ( @xmath43h ) + @xmath49co(2 - 1 ) & 220 & 75 & zoom - d ( @xmath47h ) & zoom - d ( @xmath47h ) & zoom - c ( @xmath50h ) & zoom - c ( @xmath51h ) + cs(5 - 4 ) & 245 & 80 & zoom - e ( @xmath52h ) & zoom - d ( @xmath53h ) & zoom - b ( @xmath44h ) & zoom - b ( @xmath44h ) + hcn(3 - 2 ) & 266 & 90 & zoom - e ( @xmath47h ) & zoom - d ( @xmath46h ) & zoom - c ( @xmath48h ) & zoom - b ( @xmath44h ) + hco@xmath0(4 - 3 ) & 357 & 120 & zoom - d ( @xmath47h ) & zoom - e ( @xmath47h ) & zoom - c ( @xmath50h ) & zoom - c ( @xmath52h ) + hco@xmath0(7 - 6 ) & 624 & 210 & zoom - e ( @xmath47h ) & zoom - e ( @xmath54h ) & zoom - c ( @xmath53h ) & zoom - d ( @xmath44h ) + @xmath49co(6 - 5 ) & 661 & 220 & zoom - e ( @xmath47h ) & zoom - e ( @xmath46h ) & zoom - d ( @xmath46h ) & zoom - c ( @xmath55h ) | we predict how protoplanetary disks around low - mass young stars would appear in molecular lines observed with the alma interferometer .
our goal is to identify those molecules and transitions that can be used to probe and distinguish between chemical and physical disk structure and to define necessary requirements for alma observations .
disk models with and without vertical temperature gradient as well as with uniform abundances and those from a chemical network are considered . as an example
, we show the channel maps of hco@xmath0(4 - 3 ) synthesized with a non - lte line radiative transfer code and used as an input to the gildas alma simulator to produce noise - added realistic images .
the channel maps reveal complex asymmetric patterns even for the model with uniform abundances and no vertical thermal gradient .
we find that a spatial resolution of @xmath1 and 0.510 hours of integration time will be needed to disentangle large - scale temperature gradients and the chemical stratification in disks in lines of abundant molecules . |
the interpretation of the higher order terms in the @xmath0 expansion of the nekrasov partition function as gravitational corrections for the gauge theory , has been central in many works since the first conjecture in @xcite . in the limit @xmath2
it is now well understood that the underlying microscopic theory is the geometrically engineered topological string ; open question is the general case .
this hypothetical one parameter refinement of topological strings can be seen either as a deformation of the background in which the theory lives or of the worldsheet 2-d theory itself ( or some combination of the two ) .
we will concentrate on the latter possibility for which , in the past , various suggestions have already been proposed , @xcite and @xcite , @xcite ( and @xcite working with heterotic strings ) , even if none of them , to my understanding , fully successful .
the possibility that such a refined theory actually exists and it is some kind of string theory is highly probable essentially because , depending on the a or b model version of the refinement , it shares different and inherently `` stringy '' features of the topological models , with only a few modifications . a list would include its relationship with the graviphoton supergravity correction @xcite , the connection with the counting of bps d0-d2 brane bound states @xcite , the existence of a refinement of a well established computational method for toric calabi yau ( again @xcite ) and , central for this paper , the fact that it satisfies a slight deformation of the holomorphic anomaly equation @xcite , @xcite , @xcite , @xcite ( and @xcite for application to ale space ) .
this equation has been originally derived from the very worldsheet definition of topological string theory and so it is plausible that its refined version not only suggests the existence of a refined purely worldsheet - type description but it also gives hints for its actual form . following this idea
we have found an explicit two parameter expansion for the nekrasov partition function where the novel deformation parameter that arises away from the @xmath2 limit behaves as if being associated to an additional genus expansion , these new handles different from the purely topological ones ( meaning those associated to the parameter that survives the @xmath2 limit ) . in section [ sez1 ]
we will give a brief ( and much incomplete ) review for the derivation of the holomorphic anomaly equation emphasizing its relationship with the definition of the topological string amplitudes .
in section [ sez2 ] it will be ( re)explained in what sense the nekrasov partition function satisfies the holomorphic anomaly equation , what shape this equation assumes depending on the parametrization chosen and what are the consequences .
the main result will be the construction of a general ansatz for the proposed worldsheet definition of the refined string amplitudes satisfying two conditions : * it should reduce to the topological string amplitudes in the limit @xmath2 .
* it has to be consistent with the refined holomorphic anomaly equation .
section [ sez3 ] tackles the problem of the relationship between the nekrasov partition function and f - terms in the four dimensional effective field theory , generalizing the well known results of @xcite . while until this point we have worked on the b model side , in this section we will switch to the a model perspective
. the discussion here will be somehow naive , nonetheless the goal is twofold : to provide physical reason for the two coupling constants expansion of our refined partition function and to explicitly identify inside the nekrasov partition function what is related to the unconstrained part of our worldsheet ansatz .
finally in section [ sez4 ] we analyse more quantitatively an easy example using the gopakumar - vafa interpretation of topological string .
let us start the discussion with the holomorphic anomaly equation @xcite ( @xcite for the open case with purely closed moduli ) .
this is a recursion relation for topological strings that equates an antiholomorphic target space moduli derivative of a genus @xmath3 amplitude to covariant derivatives of lower genus amplitudes .
it can be derived starting from the definition of a topological amplitude as an integral over the moduli space of the riemann surface weighted by a jacobian given by the contraction of the spin two supercurrents and @xmath4 to be the two supercurrents that have been twisted to spin two , while @xmath5 and @xmath6 to spin one .
this is the typical b - model notation , for the a - model case the sign of the @xmath7 current flips in the right moving sector ] @xmath8 and @xmath4 with the appropriate beltrami differentials @xmath9 , @xmath10 : @xmath11 being @xmath12 and @xmath13 the coordinates for the complexified target moduli space of the theory , deriving with respect to @xmath13 corresponds to insert in the amplitude the operator associated to the @xmath13-deformation of the action .
this turns out to be exact with respect to the topological charge , @xmath14 , and so naively trivial .
however , before annihilating the vacuum , @xmath15 and @xmath16 should commute with the @xmath17 and @xmath18 insertions from ( [ a ] ) ; the result is , because of the supersymmetry algebra of the original worldsheet non linear sigma model , the contraction of the energy momentum tensor with a beltrami differential which is the same as an external derivative with respect to the riemann surface moduli : @xmath19 @xmath20 this gives a nonvanishing contribution only from the ( real codimension two ) boundary of the moduli space of @xmath21 which amounts to a degenerate riemann surface with either one dividing or non dividing handle shrunk to form a node . with the additional property that the boundary contribution is vanishing unless the @xmath22 sits on the node , and replacing twice
the long thin tube on the right and the left of the @xmath22 position , by complete sets of marginal operators plus the appropriate metric , brings the result @xmath23 the marginal operator insertions are represented by the covariant holomorphic derivatives whose connection has both a mathematical interpretation on a certain bundle over the moduli space , and a physical one as regularization for divergent contact terms ; the first term on the right of ( [ b ] ) corresponds to a non dividing handle degeneration ( two operators on the same @xmath24 riemann surface ) while the second to a dividing one ( one operator on each genus @xmath25 and @xmath26 surface ) .
finally the sphere three points function @xmath27 is what remains of the portion of the long thin tube on which the @xmath22 and two marginal operators were dwelling .
the story can be generalized for amplitudes with @xmath28 boundaries ( and crosscaps ) and @xmath29 operator insertions , @xmath30 , still the moral remains the same : the antiholomorphic derivative forces the commutation of the topological currents with part of the measure of the amplitude , thus bringing contributions only from the boundary of the moduli space of riemann surfaces . in the most general case with purely closed moduli
the boundary component can be either of real codimension one or two and , beside the straightforward generalization of ( [ b ] ) , it contains only the additional terms : * @xmath31 corresponding to a boundary shrinking to zero size or , conformally equivalent , moving far away from the rest of the riemann surface , giving rise to an amplitude with one less hole and the usual covariant derivative ( @xmath32 ) , contracted with a disk containing @xmath22 and the other marginal operator ( @xmath33 ) . * @xmath34 coming from contact terms arising when @xmath22 approaches one of the operator insertions on a riemann surface of non vanishing worldsheet curvature .
in this section we consider the generalization of the holomorphic anomaly equation for the refined topological string amplitudes .
first we explain how the antiholomorphic dependence by the target space moduli arises in the nekrasov partition function .
then we write what is the equation obeyed by these anthiholomorphic moduli using different parameterizations in @xmath35 and selecting one ( parameterization ) for the construction of our worldsheet ansatz .
long time will be spent dealing with the analysis of all the constrains on our ansatz coming from the refined holomorphic anomaly equation .
the nekrasov partition function @xmath36 represents a regularized partition function for supersymmetric @xmath37 four dimensional gauge theories , @xcite and @xcite , and depends on the regularization parameters @xmath35 , on the vacuum expectation values for the diagonal ( usually @xmath38 ) gauge field @xmath39 , later on briefly indicated as `` @xmath40 '' , end eventually on various masses ( the instanton counting parameter is kept implicit ) . we know from the celebrated work of seiberg and witten @xcite that @xmath36 is expected to be holomorphic in the @xmath40 but for a finite number of singularities .
moreover @xmath40 is valid only locally as a parameter on the moduli space and objects depending on it will in general transform under the modular group contained in @xmath41 , with physical quantities required to be invariant . in order to achieve this latter condition for the coefficients of a generic expansion in the @xmath35 parameters of @xmath42 , it is in general not possible to maintain the local holomorphic dependence by @xmath40 but it is necessary to add an antiholomorphic completion , @xcite,@xcite , @xcite and @xcite
this can be made explicit writing the amplitudes in terms of modular holomorphic functions plus quasi - modular holomorphic generators that become modular after adding an antiholomorphic piece .
for example for @xmath43 we have the amplitudes depending on the eisenstein series @xmath44 @xcite,@xcite . under the redefinition @xmath45 , with @xmath46 and @xmath47 the prepotential ,
they become modular invariant objects and all the antiholomorphic dependence is carried by @xmath48 .
so we should consider the holomorphicity as arising only as a local limit of well defined modular non holomorphic objects .
this is indeed the same story as for the holomorphic anomaly equation .
chosen a point in the moduli space by fixing @xmath49 and @xmath50 , the topological amplitudes are holomorphic only with respect to local coordinates , usually called @xmath51 , parameterizing marginal operator deformations of the action , @xcite .
this @xmath51 is also the parameter with respect to which the f terms @xmath52 , in the four dimensional effective field theory from type ii superstrings compactification , are holomorphic .
these f terms come from the superspace integration of @xmath53 , where @xmath54 is a function of the chiral superfields @xmath55 , @xmath56 is the gravitational superfield and @xmath57 and @xmath58 were respectively the self dual part of the field strength of the graviton and the graviphoton .
they `` correspond '' to the topological amplitudes in the sense that the coefficient @xmath59 is the genus @xmath3 amplitude for the topological theory with target space the same calabi yau used for the superstring compactification @xcite .
being an f term we expect it to be holomorphic in the parameters ( @xmath51 ) with the holomorphic anomaly equation introducing the correct global @xmath60 dependence . in @xcite,@xcite , @xcite and @xcite
it has been shown that the coefficients of a two parameters expansion of the nekrasov partition function satisfy an equation that is but a slight extension of the holomorphic anomaly equation for topological string amplitudes .
this can be checked assuming the form of the equation and then solving it order by order to match the nekrasov results in the holomorphic limit for a variety of examples .
it was also shown how the holomorphic ambiguity that arises when integrating equations like ( [ b ] ) , can be completely fixed relying on the behaviour of the nekrasov partition function in selected points of the moduli space .
an obvious but important thing to keep in mind is that the explicit form of the equation depends on the chosen expansion .
the authors of @xcite define from @xmath61 to @xmath3 is the one used in all the ansatz inside @xcite and it is always enough for a full parameterization ; in fact it is even overabundant . ] .
@xmath62 instead @xcite expands as but this brings an `` @xmath63 '' in the identification with the topological string coupling constant @xmath64 in the limit @xmath65 which is not present in @xcite .
the addition of the minus sign simplifies the comparison and redefines @xmath66 .
this naively would bring a minus sign on one of the two sides of ( [ g ] ) but , being the three point function @xmath67 derived from the prepotential , it also carries a minus sign and the equation remains untouched .
it is easy to check that this agrees completely with ( [ f ] ) ] @xmath68 in the expansion ( [ d ] ) the holomorphic anomaly equation for @xmath69 reads @xmath70 where @xmath71 and the power of @xmath72 has to be matched on the two sides of the equation . instead using ( [ e ] )
we have @xmath73 for both ( [ f ] ) and ( [ g ] ) the first term on the right is zero for @xmath74 and the one loop case @xmath75 , that usually has an additional term , is here not different ( due to local calabi yau properties ) than the direct @xmath75 reduction of ( [ f ] ) and ( [ g ] ) .
moreover the covariant derivatives appearing are the same as the ones in the usual holomorphic anomaly equation ( [ b ] ) .
correctly ( [ f ] ) and ( [ g ] ) are equivalent as can be checked algebraically .
the partition function can depend on hypermultiplet masses as well . in this case
it was found that the holomorphic anomaly equation develops , at least in some examples , an additional term , @xcite .
this has the same form of the degenerating disk that appears in the open holomorphic anomaly equation of @xcite , and it requires the presence of terms with odd power in @xmath76 inside ( [ d ] ) .
however it was later understood in @xcite that an appropriate shift for the mass parameters defines a different partition function whose anholomorphic completion obeys the same equation ( [ g ] ) of the massless case,@xmath77 compatible to the expansion ( [ d ] ) ] . for this reason from now on we will use ( [ f ] ) and ( [ g ] ) as the most general cases leaving implicit , when having non zero masses , that the correct shift for them has been performed .
let us consider instead @xmath78 with @xmath79 . in this case equations ( [ f ] ) and ( [ g ] ) transform into @xmath80 with the first two terms vanishing if , respectively , @xmath74 and @xmath81 .
the main reason behind this work is that in the expansion ( [ h ] ) the index @xmath82 behaves , from the point of view of the above equation , in a way more symmetric with respect to @xmath3 .
as anticipated it is well known that ( [ d ] ) , ( [ e ] ) and ( [ h ] ) all reduce in the limit @xmath83 to the expansion @xmath84 with @xmath85 identified as the topological amplitude of genus @xmath3 on the local calabi yau that geometrically engineers the gauge theory for @xmath86 . from the left to the right hand side of the arrow it has been performed the holomorphic completion already explained , in this case the dependence by @xmath50 obeys the usual holomorphic anomaly equation ( [ b ] ) .
needless to say on the left the moduli are gauge theory vevs , on the right geometric objects . reversing the derivation that led us to the holomorphic anomaly equation we will start with equation ( [ i ] ) as our `` experimental data '' , assume that it is the direct consequence of some yet unknown worldsheet description , and try to derive it .
it is clear that also @xmath69 and @xmath87 can be associated to proper worldsheet amplitudes .
in fact the authors of @xcite have already advanced a first proposal for @xmath87 as a genus @xmath3 topological amplitude with @xmath82 operator insertions .
given the form of ( [ g ] ) and the constraints from the local special geometry they have identified this operator as the first gravitational descendant of the puncture operator , @xcite .
however the novelty in the choice of the expansion ( [ h ] ) is the peculiar worldsheet interpretation that arises for @xmath88 : as we will see `` @xmath82 '' no longer counting operators but handles .
this new role for the integer @xmath82 is crucial if we want to completely determine the worldsheet definition for the refined amplitudes , thus fixing also the holomorphic ambiguity left from the constrains derived by ( [ i ] ) , by comparison with the exact expressions known from direct computation of @xmath42 .
we will discuss more later on .
let us for the moment discard the minus sign in front of the second term in the right hand side of ( [ i ] ) . in this case
there is an immediate ansatz for @xmath88 that explains the form of holomorphic anomaly equation ( [ i ] ) .
@xmath88 is interpreted as an amplitude of total genus @xmath89 with topological - type measure for the riemann surface moduli space integral , and `` something '' marking as different the handles of `` type @xmath3 '' from the ones of `` type @xmath82 '' ( which is here expressed by the notation @xmath90 ): @xmath91 for the moment we can roughly think to a picture like figure [ figura1 ] , where the riemann surface is divided into two sets of domains , and the sum of all the genera is @xmath89 . in the future
we will refer to this situation as `` domain decomposition '' or `` domain splitting '' .
moreover the amplitude @xmath88 contains all the possible splittings into domains that satisfy certain geometric conditions we will derive later . is interpreted as a riemann surface of genus @xmath89 , with handles of type @xmath82 differentiated by the handles of type @xmath3 in some still undefined way .
pictorially it is shown one of the possible various contributions to @xmath88 for @xmath92 and @xmath93 .
the domains containing the handles of type `` @xmath82 '' are shaded and separated by the domains of type `` @xmath3 '' by border 1-chains.,scaledwidth=80.0% ] the computation of @xmath94 proceeds analogously as in section [ sez1 ] , with the topological charge anticommuting with the enlarged moduli space measure of ( [ i11 ] ) , and thus forcing the nonvanishing contributions to come only from the boundary of the moduli space of @xmath95 . when the degeneration describing these contributions is given by the formation of a node on a non dividing handle belonging to the white `` @xmath3 '' domain we have a contribution , from that domain decomposition , for the first term in equation ( [ i ] ) ; when it belongs to the shaded `` @xmath82 '' domain we have a contribution for the second term ; and finally when the handle is dividing we obtain contributions for the third term .
we want to analyse more in detail the logic that has led us to this first proposal .
first of all the integral measure weighting whatever object is associated to the index @xmath82 still has to contract the beltrami differentials with the supercurrents @xmath8 and @xmath4 instead that , for example , some ghost field .
this is required in order to generate , after the @xmath96 derivative , total derivatives with respect to the riemann surface moduli .
the measure for the @xmath3 handles is further restricted to be exactly the topological string one , as @xmath97 is required to describe the topological amplitude at genus @xmath3 .
however at first sight it would be possible to generalize the `` @xmath82 '' measure as @xmath98 = 0\ ] ] still generating something proportional to a total derivative for riemann surface moduli .
but , if this was the case , after deforming the topological currents around @xmath99 and @xmath100 , one @xmath101 and @xmath102 insertion would remain in the amplitude unpaired to @xmath103 and @xmath104 , and of them we have no trace in the lower @xmath82-order amplitudes inside equation ( [ i ] ) .
the second question is what object should we associate to the integer @xmath82 in such a way to obtain the right hand side of ( [ i ] ) ? for a compact riemann surface
there are only four different possibilities : handles , holes , crosscaps and punctures ( with eventually operators ) .
crosscaps and holes behave similarly and are in any case connected by the mechanism of tadpole cancellation @xcite and @xcite .
as discussed in section [ sez1 ] they degenerate to a single covariant derivative term .
punctures have only the additional term containing the moduli space kahler metric ; a topological amplitude with @xmath82 appropriate operator insertions is eventually compatible with equation ( [ g ] ) , not with ( [ i ] ) .
it remains only to consider handles , which fulfill the requirement due to the same degeneration satisfied by @xmath3 and @xmath82 ( apart the minus sign ) . and
this is the reason for the complex dimension of the moduli space to be @xmath105 .
four are the important points to develop : 1 .
being now on the same footing @xmath3 and @xmath82 it appears that both @xmath106 and @xmath107 inside ( [ h ] ) play the role of coupling constants , neither of them being weighted by the complete euler number but instead sharing it .
2 . we need to explain the minus sign in front of the second term on the right hand side of ( [ i ] ) 3 . some difference between the portion of the riemann surface counted by @xmath3 and
the one counted by @xmath82 should appear .
first of all in order to distinguish among @xmath3 and @xmath82 and among amplitudes with the same @xmath89 value which are not equal in nekrasov computations .
second because , if this was nt the case , the boundary contribution from non dividing handle degeneration to equation ( [ i ] ) would come from all the integers @xmath108 and @xmath109 such that @xmath110 .
4 . we should derive precise geometrical conditions defining the decomposition of the riemann surface into domains of either @xmath3 or @xmath82 phase .
let us proceed in order and discuss first the apparition of a double coupling constant .
this is natural passing to the a - model perspective and using the gopakumar - vafa interpretation of topological strings from type ii a - m theory , @xcite and @xcite .
the coupling constant of the topological partition function is there given by @xmath111 where @xmath57 is the vev of the self - dual part of the graviphoton field strength and @xmath112 the coupling constant of type ii a ( or the radius of m theory ) .
that case is presently generalized for refined amplitudes allowing also to @xmath113 to acquire a vev ( we will discuss this more extensively in section [ sez3 ] ) . and in fact we will see that the self - dual and anti self - dual parts of @xmath114 are respectively @xmath106 and @xmath107 . for the resolution of the second point we consider two different limits : @xmath115 and @xmath116 . in the first one
the expansion ( [ h ] ) reduces to @xmath117 and in the second to @xmath118 . in the first case
@xmath97 is a topological string amplitude of genus @xmath3 and @xmath106 is the coupling constant is made if seen from the gopakumar - vafa point of view ] ; @xmath119 plays the role of the prepotential and it corresponds to the sphere . from the previous discussion @xmath120 should be thought as some genus @xmath82 string amplitude with structure similar to ( [ a ] ) .
we have two problems : the first one is the minus sign appearing in the second term in ( [ i ] ) , when the genus degenerates at the boundary . the second one is the @xmath121 limit where we do not obtain the prepotential , the sphere , but minus it .
we then require that it is not @xmath107 to be the coupling constant but instead @xmath122 .
that is we factorize out an `` @xmath63 '' rewriting the expansion as @xmath123 and considering genus @xmath82 amplitudes the product @xmath124 .
this seems a little pedantic but it solves both problems : first when the new coupling constant goes to zero , @xmath125 , we obtain @xmath119 as it should , and second when we degenerate one non dividing or dividing cycle of `` type @xmath82 '' we obtain @xmath126 @xmath127 but equation ( [ i ] ) is written in terms of @xmath88 alone and this brings our minus sign
. the same conclusions are valid for non zero @xmath3 away from any limit on @xmath128 and @xmath77 , always identifying riemann surfaces with @xmath129 .
@xmath130\ ] ] another consequence of this story is a first @xmath106-@xmath107 or @xmath3-@xmath82 asymmetry in the worldsheet definition of our amplitudes as now they are no longer @xmath88 but instead @xmath131 .
we now move to discuss the third point .
we have seen as the measure for the moduli space integral is fixed to be the one of a purely topological string amplitude of genus @xmath89 .
so for what matters the coupling to 2-d gravity the handles of type @xmath3 are equivalent to the handles of type @xmath82 .
similarly it is not possible to distinguish between them by integrating some operators only on the type @xmath82 domain , as this would not be compatible with equation ( [ i ] ) after degeneration of the index @xmath82 .
the only possibility left is to make use of the 2-d sigma model action ; more specifically we want to glue two different cfts on the riemann surface , one with the usual topological lagrangian @xmath132 integrated over the handles of type @xmath3 , the other one with different lagrangian @xmath133 integrated over the handles of type @xmath82 . explicitly :
@xmath134 with @xmath135 a shortcut notation for all the fields in the path integral and @xmath136 are all the domains containing each one @xmath137 handles of type @xmath82 , @xmath138 . on the boundary between the two cfts
are assumed in the definition of ( [ v1 ] ) appropriate boundary conditions for all the fields . in order to justify , through a process of gauge fixing of the worldsheet metric , the topological measure also for the handles of type @xmath82
, it is required for @xmath139 to not spoil the topological behavior of the theory .
more in detail a topological charge @xmath140 should still be globally conserved on all @xmath141 and the supersymmetry algebra has to hold , especially the anticommutators @xmath142 obviously the explicit realization of @xmath140,@xmath143 and @xmath114 in terms of fields can change passing from a domain to another .
moreover it is necessary that the operators associated to target space antiholomorphic moduli deformations , @xmath144 , remain the same on the whole riemann surface , that is for both cfts : the topological currents are needed for the holomorphic anomaly equation machinery to work and @xmath22 appears on the three point function @xmath67 , which is a common factor in the refined holomorphic anomaly equation for both type @xmath3 and type @xmath82 degenerations .
it is also plausible that operators associated to target space holomorphic moduli deformations @xmath145 should be the same , as the covariant derivatives corresponding to their insertion on the riemann surface are unchanged from ( [ b ] ) to ( [ i ] ) and their form has clear physical justification from the form the operator assumes , see @xcite
. we will soon give additional evidence for this . an important point to note in the definition of our ansatz ( [ v1 ] ) ,
is that there are not moduli associated to the border along which the two cfts are glued on the riemann surface .
this means that the 2-d gravity does not distinguish between the two cfts , the coupling and so the process of gauge fixing being the same .
it still can be that the amplitudes @xmath131 modify their value after deformations of the border keeping unchanged the set of integers @xmath137 . in the following section it will be shown as this clashes with equation ( [ i ] ) ; thus the picture behind our ansatz is the one of a topological interface between two cfts .
we have seen that @xmath131 corresponds to a riemann surface of genus @xmath89 divided into a certain number of domains either of `` type @xmath3 '' or `` type @xmath82 '' defined by the gluing of two different cfts ; one with the usual topological string action , the other a deformation satisfying the above conditions . moreover the moduli space measure is the one that would appear for a topological string action of genus @xmath89 .
we need some rules for the domain decomposition such that the handle degeneration of @xmath131 , at the border of the riemann surface moduli space , is consistent with the results from equation ( [ i ] ) .
decompositions that do not satisfy these rules should vanish .
presently this is just part of the definition of the ansatz ( [ v1 ] ) , derived from consistency with equation ( [ i ] ) . to have
a microscopical derivation would require the knowledge of @xmath139 as we will discuss in the conclusions .
first of all we fix a basis of @xmath89 a - cycles , such that collapsing them corresponds to the shrinking of @xmath89 non dividing handles for the degenerate riemann surface .
there are other non dividing handles that can form a node but for the moment we consider only these .
also we have a set of dividing ( exact ) chains whose collapsing corresponds to a node on a dividing handle .
they are represented in figure ( [ figura1a ] ) .
let us consider first the case of non dividing handle degeneration , that is we concentrate on the first two terms in equation ( [ i ] ) .
we want any domain decomposition to belong univoquely to a single @xmath131 , in such a way that each moduli space boundary term that arises after the formation of a node ( on a non dividing handle ) , will belong either to @xmath146 or @xmath147 and @xmath82 from the ones appearing on the right hand side of equation ( [ i ] ) , for example @xmath148 . ] .
this is achieved by selecting , for a single @xmath89 riemann surface , all the domain decompositions such that the sum of all the handles inside all the type @xmath3 domains is @xmath3 , and the sum of all the handles inside all the type @xmath82 domains is @xmath82 .
said in another way , when we cut along any border the sum of the genera of the disconnected surfaces so obtained should give the total genus of the original surface .
the boundary between any two domains has to be along some connected homologically exact 1-chain .
for example figure [ figura2 ] is not allowed . and @xmath82 associated is not well defined .
the border between any two domains is in this case not connected and they are not allowed in our model , scaledwidth=50.0% ] there is a peculiar case of non dividing handle degeneration that requires additional care , depicted in figure [ figura3a ] .
if we shrink the non dividing handle of a genus one domain sandwiched among two or more domains in the other phase , we end up with a genus zero domain .
we want this to belong to @xmath147 ( or to @xmath146 with inverted colors ) .
this immediately brings us to the requirement depicted in figure [ figura3 ] . in figure [ figura3 ]
we are shifting a border between two domains from an exact connected 1-chain to an homologically equivalent one , such that the region spanned by the shift has total genus zero .
we ask for this shift to not modify the value of the contribution from that domain decomposition .
also the marginal operator insertions are to be unaffected by the change of domain around their location and so transparent with respect to the above border manipulation ) correspond to the same usual marginal operator insertions also on domains of type @xmath82 ] .
then the situation in figure ( [ figura3a ] ) is simply interpreted as the approaching of the two borders for the central annulus domain , until they coincide and the domain disappears .
this is a connected to what we were discussing after equation ( [ v1 ] ) : our ansatz does not have moduli associated to the position and shape of the border between different domains , so from the 2-d gravitational point of view any change as in figure ( [ figura3 ] ) is irrelevant . here
we are asking that also the purely field theory topological amplitude ( that is before the coupling to 2-d gravity ) , does not change under border redefinitions that do not modify the values of the integers @xmath3 and @xmath82 .
a similar issue arises when we consider dividing handle degenerations .
if the dividing handle we are collapsing belongs to a single domain , then the arising of the corresponding contribution from the last terms in equation ( [ i ] ) is immediate .
however imagine to look at the situation in figure [ figura4 ] , where along the dividing handle we are degenerating passes also the border between two domains .
we can have the node of the degenerating handle above the border , on the border or below . then
, unless the node happened to be exactly on the border 1-chain , we will end up having one of the two pieces with a remaining disk in the other phase with one marginal operator on it .
again this is an application of the condition in figure ( [ figura3 ] ) with the shaded domain a disk ; the border shrinks without affecting the value of the amplitude , until it becomes a point and disappears .
this manipulation does not modify the integers @xmath3 and @xmath82 directly applied to the worldsheet ansatz associated to the full partition function .
then this limit can be seen as degenerating all the domains of type @xmath82 until they reach genus zero , that is they become disks or annuli .
the corresponding amplitude is then equivalent to a purely topological one of genus @xmath3 . ] .
the last case is the one of a non dividing handle degeneration besides the ones depicted in figure [ figura1a ] .
this is when we have the shrinking of a cycle like the one in figure ( [ figura4a ] ) .
if along this cycle does not pass any border then the boundary contribution belongs to either @xmath146 or to @xmath147 in the obvious way , depending on the colors .
however if along this handle passes longitudinally the border between two domains , we obtain the result of figure [ figura4a ] .
the domain decomposition on the right hand side of figure [ figura4a ] is obtained by going at the boundary of the moduli space of the riemann surface in the domain decomposition represented by the left hand side .
the latter does not satisfy our conditions .
thus that handle degeneration and the corresponding boundary contribution is vanishing .
this is a first novelty with respect to the usual holomorphic anomaly equation ( [ b ] ) for purely topological amplitudes where every handle degeneration contributed .
obviously which non dividing handles contribute or not depends on the splitting we are considering ; nonetheless it is clear from figure ( [ figura1a ] ) and generalizations that it always exists a basis of @xmath150 a - cycles whose collapsing gives non vanishing result .
this concludes the set of requirements which are part of the definition for the domain decomposition appearing in our worldsheet ansatz ( [ v1 ] ) . they have been obtained asking that every possible handle degeneration will give rise only to a contribution appearing inside the amplitudes on the right hand side of equation ( [ i ] ) .
the next step is to verify the opposite , that is that the most generic object that can appear on the right hand side of equation ( [ i ] ) should come only from degeneration of handles from a domain decomposition satisfying the above geometric conditions . for what matters the last term in equation ( [ i ] ) that is quite evident .
every term of the form @xmath151 is given by two disconnected amplitudes , each one given by the sum of all the allowed domain splitting contributions for integers @xmath152 and @xmath153 , with one marginal operator insertion integrated everywhere .
the domain splitting contribution to the bulk amplitude , whose boundary degeneration is of this kind , is simply given by replacing the two operator insertions by an handle .
by the stated geometric conditions if the two domains on which the operators are located are of the same type , then the handle belongs to the same domain , otherwise somewhere on it passes the border between them . in both cases the bulk amplitude satisfies by construction our conditions .
it is more tricky the situation coming from the first two terms in equation ( [ i ] ) .
consider for example @xmath154 ( the other is treated similarly ) that comes from a degenerating non dividing handle of type @xmath3 .
the location of the two marginal operators is integrated over all the surface so the degenerating handle they replace could have started and ended from anywhere to anywhere , regardless of the current phases ( @xmath3 or @xmath82 ) of the domain on which the marginal operators are .
but in the original bulk amplitude both those locations had to belong to the same @xmath155 domain , otherwise contradicting the geometric conditions derived .
fortunately , when the two marginal operators sits in points belonging to different domains of the riemann surface , it exists an equivalent domain decomposition maintaining the same operator positions , but obtained from the non dividing handle degeneration of an allowed amplitude .
the border manipulation leading to this has been represented in figure [ figura5 ] .
this is a first example of another peculiar fact about our amplitudes .
the same component in the boundary of moduli space can be reached from different types of amplitudes in the bulk . in the present case one of the two bulk amplitudes is not allowed and should consequently be zero .
but the phenomenon is more general and we will see more later . it is possible to reconstruct the two coupling constant expansion in ( [ h ] ) as a weight coming from a sewing construction from different riemann surfaces in different phases , correctly weighted .
this representation makes clear the string role of @xmath156 and @xmath157 as coupling constants for the two different cfts / domains appearing inside our refined topological amplitudes , both arising from the single type iia dilaton and the two different vevs for the self dual and anti self dual part of the graviphoton .
the interpretation is clearly in terms of closed strings either in one phase or the other ( belonging to one type of cft / domain or the other ): passing across the border is represented as the emission of a string in one phase ( an operator insertion for a generic string state in that phase ) times the propagator given by the inverse of the sphere with two string operator insertions , one in one phase and one in the other , times the absorption by the other domain ( an operator insertion for a generic string state in the other phase ) .
the rules are easy and represented in figure [ figura7 ] * for each riemann surface of genus @xmath155 ( resp .
@xmath137 ) in the corresponding phase that is going to be sewed to @xmath158 ( resp .
@xmath159 ) different surfaces in the @xmath82 ( resp .
@xmath3 ) phase we multiply by @xmath160 ( resp .
@xmath161 ) .
* for each sewing between two different surfaces we multiply by the inverse metric @xmath162 ( or the weight of the inverse of a sphere with two operator insertions , one in the @xmath3 and one in the @xmath82 phase ) . *
the sewing procedure should create a composite riemann surface with total genus @xmath89 , @xmath163 and @xmath164 .
it can be verified that proceeding in this way the weight does not change adding disks or intermediate annuli or other genus zero domains , thus automatically satisfying our results . moreover omitting the last rule amounts to reproduce weights that do not fit with the expansion ( [ h ] ) .
last but not least let us look at the enriched moduli space of the new mixed phase riemann surfaces ( it is exactly the moduli space of genus @xmath89 riemann surfaces but subdivided into components belonging to different domain decompositions ) .
we have seen before as different amplitudes in the bulk of the moduli space ( in that case one of them vanishing ) can degenerate to the same boundary component .
the situation is more generic and it involves always non dividing handle degenerations .
the prototype case is @xmath165 ; obviously it can be reached starting from either @xmath166 or @xmath167 .
we have for this reason that the same boundary component is now shared by two bulk portions of the moduli space or , said in another way , a degenerating non dividing handle does not distinguish between its @xmath3 or @xmath82 phase .
what we have seen in general being true for dividing handles , to be able to change their phase without affecting the value of the amplitude ( the irrelevant appearing or disappearing of an annulus domain from the condition in figure ( [ figura3 ] ) ) , is true for non dividing ones only when degenerating . on an infinitely long handle only vacuum states propagates ( as massive states are exponentially suppressed ) .
moreover in topological string theory there is a one to one correspondence between the @xmath140 cohomology class of operators and differential forms of appropriate degree in the de rham or dolbeault cohomology ( in the a or b model respectively )
. then hodge decomposition applies to corresponding states @xmath168 @xmath169 where @xmath170 is annihilated by both @xmath140 and @xmath171 and @xmath172 is generic .
so @xmath168 is in the same cohomology class and thus physically equivalent to the vacuum @xmath170 , and this defines the correspondence between vacua and physical operators for topological strings .
so a long handle can be replaced by a complete set of local topological string operators plus metric , and for conformal invariance to be maintained they also have to be marginal . but a long degenerating handle spanned by usual topological strings , that is in the @xmath3 domain , is equivalent to a one in the @xmath82 domain , as we have seen .
then in both cases it should be replaced by the same set of operator insertions @xmath173 thus representing the vacua also in the @xmath82 domain and so being the operators appearing as substitution for a long handle in the refined holomorphic anomaly equation .
moreover in equation ( [ i ] ) the insertion of these operators is represented by a covariant derivative with respect to the target space moduli , so we conclude that the operators associated to a deformation of the holomorphic target space moduli are the same for the action both on the domain of type @xmath3 and @xmath82 .
this means that these operators are in correspondence with either @xmath174 or @xmath175 on all the riemann surface regardless of the domain to which they belong ( @xmath101 is the calabi yau target space and @xmath176 and @xmath177 refer to the de rham and dolbeault cohomology ) , and the charge @xmath140 maintains the isomorphism with the de rham or dolbeault cohomology ( at least ) when restricted to act on marginal physical operators for the usual topological string theory . before the next section
let me clarify a point which was also the original motivation for the present paper .
we know that in the limit @xmath178 the nekrasov partition function reduces to the topological one , essentially because both of them can be related to the f terms @xmath179 in the four dimensional action .
instead if we consider the limit @xmath180 and we discard for a moment the four dimensional r - symmetry twist which is part of the nekrasov computation ( see expression ( [ c ] ) ) , we can repeat the same reasoning only passing from self - dual to anti self - dual objects , @xmath181 .
but these f terms are now what we would obtain from an antitopological computation , as the bosonized @xmath7 current in the internal part of the anti self - dual graviphoton vertex operator has an opposite sign with respect to the self - dual case , @xcite . and
because this is the piece in the graviton - graviphoton superstring amplitude that resembles the topological twist of the action , we are led to the opposite topological twist , or to an antitopological action .
this has now the holomorphic vevs @xmath40 in the nekrasov result identified as a local expansion around @xmath50 instead that @xmath49 .
so for @xmath131 we would somehow expect a refinement of the topological antitopological fusion construction of @xcite at higher genus @xmath89 , instead of the refinement of our topological genus @xmath89 amplitudes domain , be it in topological or antitopological description . ] . apparently this clashes with the holomorphic anomaly equation ( [ i ] ) as , passing from a topological part to an antitopological one in the same amplitude , the `` anomalous '' modulus changes from @xmath50 to @xmath49 , but our equation derives with respect to the lonely @xmath50 .
we should note however that in our case the meaning of the holomorphic anomaly equation is : we derive with respect to the moduli that correspond to an anholomorphic completion of the nekrasov amplitudes , in order to make them modularly invariant .
these are @xmath50 for amplitudes holomorphic in @xmath49 , but should be @xmath49 themselves for amplitudes considered holomorphic in @xmath50
. the very reason for which antitopological theories are not often considered is that they are completely equivalent to the topological ones , under exchange @xmath182 .
so we can safely identify the nekrasov partition function in the limit @xmath180 ( without the four dimensional r - symmetry twist ) as either the antitopological partition function with @xmath183 or as the usual topological one with @xmath49 replacing @xmath50 . in both cases
the correct answer requires some worldsheet deformation to arise modifying the amplitude , but this is connected to the four dimensional r - symmetry twist we are for the moment omitting .
this section has a double purpose . on one side
we want a little more understanding for the role of @xmath139 inside equation ( [ v1 ] ) . on the other
we should discuss why in the refined case we consider a non vanishing vev for both the self dual and anti self dual part of the graviphoton , as required for our double coupling constant interpretation for @xmath156 and @xmath157 in equation ( [ i1 ] ) .
to this end we relate here the nekrasov partition function to certain terms in the four dimensional effective field theory .
essentially there are two different possibilities which have been explored in the literature .
the first one follows the original idea of nekrasov to generalize the f terms corresponding to purely genus @xmath3 topological amplitudes @xmath184 to @xmath185 inside a not anymore self dual graviphoton background , plus some unknown correction related to a twist in the internal theory , @xcite . the second uses type ii - heterotic duality to study a set of refined f terms from compactified heterotic string theory , @xcite .
these are @xmath186 , where @xmath187 is an arbitrary function of the vector superfield containing the dilaton and @xmath188 is a chiral projector .
it was shown in @xcite that in the large dilaton limit the function @xmath189 satisfies a refinement of the holomorphic anomaly equation which is exactly our ( [ i ] ) without the annoying minus sign on the right side .
this looks very promising but it has two problems : the first one is that its worldsheet interpretation constructed in @xcite and @xcite seems not to give the correct results in reproducing the nekrasov partition function . the second , purely in the spirit of our interpretation , is that it is hard to relate to the philosophy we have followed so far of an @xmath107 - @xmath106 symmetric description plus some deformation .
moreover the dual type ii version is somewhat obscure . a priori it would still be possible to associate the value of @xmath107 to some selected vev given by an appropriate combination of the vevs of the vector fields in the dual type ii description .
however the generalization of the gopakumar - vafa idea really relies on an anti self dual vev for the graviphoton , @xcite , @xcite and @xcite , so from the present point of view we prefer to follow the original nekrasov intuition .
in order to do this we write as an index the nekrasov partition function for rigid @xmath37 supersymmetric gauge theories on a five dimensional space @xmath190 : @xmath191 where @xmath192 is the radius of @xmath193 , @xmath194 are the @xmath43 cartan generators respectively for the @xmath195 acting on @xmath196 and the internal @xmath197 r - symmetry , and the trace is over the four dimensional hilbert space after that the gauge field ( generically @xmath38 ) has acquired diagonal vacuum expectation values @xmath39 . already at first sight we can see that the left - right and @xmath106 - @xmath107 symmetry of ( [ c ] ) is broken by the single operator @xmath198 which in turn is present because of the already mentioned r - symmetry twist performed during the nekrasov computation on the four dimensional @xmath37 gauge theory , @xcite .
the same asymmetry arises in our amplitude interpretation in two fashions : as the @xmath199 appearing in the expansion and as the action deformation @xmath139 that distinguishes between @xmath82 and @xmath3 domains of the full riemann surface .
moreover as in the limit @xmath200 all this disappears leaving us with the correspondence @xmath201 so in the limit @xmath202 we have @xmath203 in this way we isolate both the @xmath199 factorization and the action deformation in the amplitudes @xmath204 from the purely topological string amplitudes @xmath97 , and we identify them as the appearance on the other side of the equation of the operator @xmath198 in the index . in the next section we will give some quantitative measure of this in an easy example .
now we discuss what is the effect of the r - symmetry twist in the four dimensional effective action from the type ii a superstring point of view starting from a review of the effect of the @xmath106 @xmath205 rotation .
if we consider the index in ( [ c ] ) in the limit @xmath200 , we can think the effect of the operator @xmath206 as rotating @xmath205 while moving around the circle , so obtaining a nontrivial fibration of @xmath205 over @xmath207 .
then the logarithm of the nekrasov partition function is equivalent to a vacuum to vacuum amplitude computed on the five dimensional metric called the omega background : @xmath208 with @xmath209 and @xmath210 coordinates along @xmath205 and @xmath207 and the value of the ( field strength of ) @xmath211 to be @xmath212 being @xmath200 , the two form @xmath114 is self - dual . reducing the size of @xmath207 to zero ( @xmath213 )
the index localizes around the fixed points of the twist operators and so , for rotations , around the origin .
so we can limit ourselves to linear approximation in @xmath209 in which we can rewrite the metric ( [ n ] ) as a kaluza - klein ansatz with the kaluza - klein gauge field ( [ o ] ) .
@xmath214 identifying the @xmath207 of the index with the circle of m theory , geometrically engineered to reproduce the four dimensional gauge theory , the nontrivial kaluza - klein background ( [ o1 ] ) is translated in type ii a as a rr two - form self dual field strength background , alternatively referred to as the ( self dual ) graviphoton .
if we assume that the coefficients in ( [ l ] ) can be identified ( modulo field independent factors ) with higher gravitational corrections in the four dimensional effective action , we should look for f terms that are coupled with the rr two - form self - dual background field .
these are the graviphoton f terms @xmath179 already encountered .
actually this reasoning is somehow naive as the identification of ( [ n ] ) with ( [ o1 ] ) is not totally correct even at linear order ( see discussion in @xcite ) .
in fact the exact link with topological string amplitudes comes either from direct computations or from type ii - heterotic duality and the comparison of the contributions of bps multiplets on both sides ( again @xcite and @xcite , @xcite where it is discussed the effect of the additional @xmath215 term to schwinger computations in purely self dual graviphoton background ) .
however in order to have some insight about the general case we will simply follow the linear approximation argument also when @xmath216 . clearly the results will need future confirmation from more solid arguments .
when @xmath128 and @xmath77 are generic the graviphoton ( [ o ] ) is not anymore self dual ; further the index in this case contains also an r - symmetry twist , @xmath198 , that we want to describe in similar way . to obtain an r - symmetry rotation on a four dimensional theory ( here only a phase rotation ) we should rotate the supersymmetry generating parameters .
starting from ten dimensional @xmath37 supersymmetry we can rotate the chiral spinors for the reduced four dimensional @xmath37 by applying an opposite rotation to the covariantly constant spinors of the calabi yau compactification .
this can be achieved by an isometry on the calabi yau ( see for example @xcite ) that we represent infinitesimally as @xmath217 with @xmath218 coordinates on the calabi yau and the index @xmath219 running on all the values @xmath63,@xmath220 .
such isometry exists only on non compact calabi yau .
the @xmath213 limit will localize the index around the fixed points of this isometry ; if we assume these fixed points isolated we can consider the eleven dimensional background reproducing four dimensional r symmetry plus a ( not anymore self dual ) @xmath205 rotation , as the fibration of @xmath205 times the calabi yau over @xmath193 @xmath221 with @xmath222 the internal metric of the calabi yau .
now we perform the linear approximation obtaining again a kaluza - klein background @xmath223 with the contraction for @xmath224 performed by @xmath3 .
so also this time we reach some rr background in type ii a , with part of it in the internal space .
more precisely : consider type ii a geometrically engineered to produce , in the rigid supersymmetry limit , some gauge theory of choice .
if we want connection with the nekrasov partition function with finite generic @xmath225 and @xmath226 , this can be seen as a type ii a computation looking for higher gravitational corrections involving specific background fields . for @xmath200
this is the self dual graviphoton , so leading to f terms of the kind @xmath179 . for generic values of @xmath225 and @xmath226
we are thinking to amplitudes representing f terms containing both self dual and anti self dual graviphoton fields plus a scalar coming from the kaluza - klein four dimensional reduction of the internal rr background one form @xmath227 represented in ( [ z ] ) .
the domains of type @xmath3 are associated to the self dual graviphoton , the domains of type @xmath82 to the rest .
by comparison the role of @xmath139 should induce the appearance of the scalar from the k - k reduction of the internal rr background .
for a compact calabi yau threefold with su(3 ) holonomy @xmath228 , so in this case it is puzzling that , after compactification , a rr one form with internal index does not lead to a four dimensional massless scalar field , but to a massive one .
however in geometric engineering we are dealing with a local non compact internal manifold , local limit of a non well specified calabi yau @xmath229 . as long as one is interested only in the field theory limit , discarding all the gravitational corrections , the exact form of @xmath229 is not important ; here it is .
it would be nice to study in more detail the kaluza klein mechanism on these local spaces as perhaps this can explain the puzzling identification of the correct supergravity term corresponding to the omega background , see for example the discussion in @xcite .
we conclude our discussion with a recipe for a more quantitative analysis of the problem . more specifically we want to produce explicit numbers to associate to our action deformation ; to this end we can look at the gopakumar - vafa formulation of the topological a model , @xcite and @xcite .
the gv formulation expresses the topological partition function in term of integers counting the number of d0-d2 branes bound states wrapped on internal calabi yau 2-cycles , depending on their four dimensional lorentz representations .
d0 branes naturally couple to rr background one - forms and we have seen how the effect of r - symmetry rotation is represented by this kind of field , so gv invariants should be fully sensitive to our deformation ( even if they are performed in the @xmath230 limit so they will not be sensitive to any @xmath50 effect the @xmath139 is going to produce ) .
the above mentioned integers are @xmath231 , corresponding to the number of d0-d2 branes bound states , with the d2 brane wrapping a holomorphic curve in the calabi yau @xmath101 , @xmath232 , and with four dimensional lorentz quantum numbers @xmath233 for @xmath195 .
then a schwinger one loop integral brings the result @xmath234 @xmath235 is the area of the curve @xmath236 , @xmath237 ( @xmath112 is the type ii a coupling constant and @xmath238 is the self dual part of the graviphoton vev ) .
we can rescale on both sides @xmath239 and work with it . finally @xmath240\ ] ] with @xmath241 $ ] the @xmath242 representation of spin @xmath243 and @xmath244 + 2[0])$ ] . in @xcite
it has been proposed that for computing refined topological amplitudes it is enough to pass from the coupling to the purely left spin to both right and left , as @xmath245 with @xmath246\otimes [ j_r]\ ] ] what we are going to do is to define certain numbers @xmath247 , different from the @xmath248 , in term of which we will express our amplitudes stripped by the action deformation , that is reducing to simple @xmath89 topological string amplitudes .
thus a quantitative measure for the effect of @xmath249 will be given by the difference between @xmath248 and @xmath247 at each @xmath250 .
we rewrite the partition function for purely @xmath89 topological amplitudes weighted by @xmath251 as ( we substitute @xmath88 with @xmath252 ) @xmath253 once again this is our ansatz without any deformation distinguishing between @xmath3 and @xmath82 ( apart from the @xmath199 ) .
because of ( [ a1 ] ) we have @xmath254_{coefficient \;of \;\epsilon_{-}^{2g+2n-2}}\ ] ] so we can express the @xmath247 in terms of the @xmath255 and then the @xmath231 and finally the @xmath248 . before giving some numbers in an example for
the simple @xmath256 case , let me point out the situation .
working purely in terms of type ii a - m theory we can compute the refined amplitudes by the knowledge of the numbers @xmath231 .
the only difference between the topological partition function and the refined one is the coupling or not to the right @xmath257 spin , through the appearance of a non vanishing vev for the anti self - dual part of the graviphoton , @xmath107 .
when it is zero we immediately reduce to the usual topological case .
so to speak all the information about the nekrasov partition function is already contained in the gopakumar - vafa invariants @xmath231 .
the choice between computing topological amplitudes or refined ones translates in not coupling or coupling to the right spin . from the present point of view instead we start directly from topological amplitudes ; in this case switching on the vev for the anti self dual graviphoton implies the addition of a new expansion parameter in the refined topological partition function , counting additional handles .
but unlike the type ii a case this is not enough .
we need to consider some additional deformation for the action integrated on the domain of type @xmath82 , as we have seen . as an example we consider the local @xmath256 .
once the @xmath231 are given for some @xmath236 , then it is just an algebraic exercise to construct the @xmath248 and derive the @xmath247 regardless of the original geometric origin .
a basis for @xmath258 is given by @xmath259 and @xmath47 ( base and fiber ) with intersection numbers @xmath260 and @xmath261 . in the following examples we list the values of @xmath248 and @xmath247 for various curves
( computation has been performed also for the curves @xmath262 and @xmath263 but omitted for reasons of space ) . inside the tables @xmath264 span
respectively rows ( the lines ) and columns ( the position on the line ) . [ cols="<,<",options="header " , ] to compare the @xmath247 with the @xmath255 , these matrices can be roughly divided into three parts . the first column , which is equal on both sides .
this is not a surprise as the first column refers to the case @xmath81 where we limit to topological amplitudes both for ( [ r ] ) and for ( [ t]),([u ] ) .
all the columns from a certain number ( depending on the curve ) to infinity and the lower part , from a certain line to infinity , of the remaining columns ( apart from the very first one ) . there
the @xmath255 are zero and this is an issue of the basis ( [ s ] ) . for what matters the @xmath247
they are not zero , but they reduce to very small numbers .
so the action deformation @xmath139 is somehow irrelevant in this zone . the rest , the upper left part of the matrix , apart from the first column . there
the difference between @xmath255 and the @xmath247 is the biggest .
finally note that in general @xmath247 is not an integer .
so there can not be interpretation as counting any object .
in this paper we have constructed a general ansatz for a worldsheet definition of the coefficients of the nekrasov partition function expanded with respect to the parameters @xmath156 and @xmath157 .
this ansatz has the property to reduce to the usual genus @xmath3 topological amplitudes in the limit @xmath265 and to satisfy the refined holomorphic anomaly equation that is known to held from direct computations .
moreover we have given some physical understanding for the interpretation of both @xmath156 and @xmath157 as coupling constant for the refined topological theory .
what is left for transforming this ansatz into an actual uniquely defined worldsheet definition is the direct knowledge of @xmath139 , that is of the cft associated to the type @xmath82 domain .
it is clear that @xmath139 is nothing but the worldsheet counterpart of the holomorphic ambiguity from the integration of the refined holomorphic anomaly equation .
thus its knowledge should be obtained by other means .
indeed a possibility for studying this new conformal theory is to consider amplitudes in a unique type @xmath82 domain , for example @xmath266 .
this from our point of view is nothing but a genus 1 amplitude for the unknown cft with lagrangian @xmath133 for which , in explicit examples , are given exact expressions ( see for example @xcite )
. then the goal would be to derive what cft lagrangian would give rise to such torus amplitude on the local calabi yau considered .
much probably this will need the introduction of new degrees of freedom integrated only on the domains of type @xmath82 and with correct boundary conditions an interesting check for our proposal would be to study the theory around the conifold point in moduli space .
it is known that there the topological string reduces to the @xmath267 string theory at the self dual radius , @xcite , and the refined case generalizes multiplying the radius by the ratio between @xmath225 and @xmath226 .
thus in our case @xmath139 would have to correspond to the operator deforming the compactification radius of the @xmath267 theory , if restricted to the conifold point in moduli space .
a direct relationship between our proposal and the one given by huang and klemm in @xcite is also interesting . roughly speaking it is possible that each dilaton operator of @xcite somehow pops up , in our model , to a genus one domain of type @xmath82 , and higher domains correspond to the case of colliding operators . but being the definitions associated to different objects ( even if linearly related one to the other )
, the dictionary between the two proposal can be more complicated .
finally it is auspicable a direct microscopical derivation of our results .
doing this would probably require the knowledge of some 2-d worldsheet theory reducing in some limit to two different phases described by the two cfts on type @xmath3 and @xmath82 domains , together with the geometrical conditions for the domain decomposition .
to my knowledge something similar has been done only in @xcite .
it is a pleasure to thank giulio bonelli , tohru eguchi , kazuhiro sakai , alessandro tanzini and johannes walcher for various discussions and useful comments , and yu nakayama for email correspondence .
also i would like to thank the cern theory group for hospitality during which i started to tackle the problem , long time ago .
this work was supported by the japanese society for the promotion of science ( jsps ) . | we will show how the refined holomorphic anomaly equation obeyed by the nekrasov partition function at generic @xmath0 values becomes compatible , in a certain two parameters expansion , with the assumption that both parameters are associated to genus counting .
the underlying worldsheet theory will be analysed and constrained in various ways , and we will provide both physical interpretation and some alternative evidence for this model . finally we will use the gopakumar - vafa formulation for the refined topological string in order to give a more quantitative description . #
1#2 # 1 # 2 2ex * double genus expansion for general @xmath1 background * = 3.5ex andrea prudenziati + 3ex _ yukawa institute for theoretical physics , kyoto university _ + _ kyoto 606 - 8502 , japan _ + 1ex [email protected] = 3.5ex |
the properties of @xmath7 transition metal systems have been understood in terms of the mott - hubbard model where charge fluctuations are suppressed owing to a large electronic hubbard gap
. this model forms the basis of our understanding of the electronic properties in high temperature superconductors and also the interactions in multiferroic materials .
@xcite optical techniques have been used to verify these models , but are limited by selection rules requiring @xmath8 ( @xmath9 is the orbital angular momentum ) and have supplied restrictive information about @xmath7 transition materials as @xmath10 transitions can involve a change in the spin or no change in @xmath9 .
@xcite such techniques are also sensitive to a variety of cross sections , including strong multiple phonon scattering , which often mask the subtle @xmath10 transitions .
@xcite here we use neutron scattering to measure the magnetic excitations within the mott - hubbard gap of coo to illustrate that coo is a classical mott - hubbard insulator where the excitations are based upon crystal field excitations within the large mott gap .
coo has a high temperature cubic rock salt structure and below a temperature of 295 k , the magnetic structure becomes antiferromagnetic , accompanied by a structural transition to a predominately tetragonal unit cell with @xmath11 at 10 k. @xcite the electronic structure has both spin and orbital moments making coo unusual when compared with other transition metal materials such as nickelates or cuprates .
@xcite the antiferromagnetic structure is uncertain with both collinear and non - collinear order describing the neutron diffraction data equally well . @xcite inelastic neutron scattering has shown the existence of low lying magnetic branches corresponding to coupled spin and orbital moments .
@xcite however , no complete theory including both spin and orbital moments with nearest and next nearest exchange terms , the magnetic structure , spin - orbit splitting , and crystal field parameters have been developed .
part of the failure to understand the low - energy properties of coo is the ambiguity behind the electronic ground state of the co@xmath12 ion and , in particular , the role of spin - orbit coupling which is difficult to extract and disentangle in the presence of strong exchange terms in the hamiltonian .
original calculations based upon lda+u resulted in energy scales too large to explain the observed spectrum obtained from non - resonant inelastic x - ray scattering .
@xcite further calculations showed that the observed energy scales involved orbital transitions and not the large mott hubbard gap @xmath13 ev .
@xcite however , other studies using arpes suggested the possibility of charge transfer excitations where oxygen states play a key role in the band structure leading to the proposal of an additional zhang - rice " band .
@xcite neutron scattering has a complementary cross section which is sensitive to both orbital and spin transitions with the selection rules determined by matrix elements of the operator @xmath14 .
it is also well suited for low - energy transitions allowing the direct observation of the spin - orbit energy scale .
this paper reports a neutron scattering study of the single - ion properties of coo with the ultimate objective of understanding the relative energy scales of the crystal field imposed by the neighboring oxygen ions and the spin - orbit coupling .
we have used high - energy neutron spectroscopy to probe the @xmath15 1 ev single ion excitations characterizing the crystal field component and energy scale . dilute samples and lower energy neutron scattering
was used to measure the much smaller energy of the spin - orbit coupling term .
the use of dilute samples removes the complex exchange terms which give a strong momentum dependence and a complex energy spectra .
we do not study the exchange terms in this paper which require a detailed measurement of the momentum dependence of the low - energy magnetic fluctuations with neutrons .
the key results of this paper are summarized in the tanabe - sugano diagram illustrated in fig .
[ summary ] .
we will show , using neutron inelastic scattering data , that coo resides in the weak - intermediate crystal field limit and discuss the transitions observed and predicted .
these will be compared with the various theories describing the competing crystal field and electron - electron repulsion energies .
@xmath16 the energy values ( tanabe - sugano diagram ) for co@xmath12 as a function of @xmath17 .
the shaded rectangles represent the excitations measured with neutrons at t=300 k using the maps spectrometer .
the vertical extent is the experimental error .
the states listed on the far left ( @xmath18f , @xmath18p , and @xmath19 g ) are the free ion solutions ( ref . ) . @xmath20 schematic illustration of the ground state for the extreme values of the crystal field splitting @xmath21 ( where @xmath2 is the crystal field splitting and @xmath1 is the intra - orbital exchange defined in ref . )
where @xmath22 ( when @xmath17 is small ) , and @xmath23 , ( when @xmath2 is large).,width=321 ] the paper is divided into 5 sections including this first introductory component . in the second section , the experimental details are outlined including sample preparation , instrument configurations , and also background subtraction and data analysis .
the third section provides an outline of the relevant crystal field theory and also compares the racah parameters derived from tanabe - sugano diagrams to more recently developed lda+u calculations .
the fourth section describes the experimental results aimed at providing a complete experimental understanding of the single - ion properties in coo .
this section discusses the samples including preparation and thermodynamic characterization for the neutron experiments .
a discussion of the methods used to subtract the background are also provided .
the 10 g single crystal of coo was grown using the floating zone technique at oxford university .
the sample was prepared by annealing high purity co@xmath24o@xmath25 ( @xmath2699.99% ) under an atmosphere of high purity argon at a temperature of 1200@xmath27c for 36 hours .
the powder was reground at various intermediate times .
powder diffraction was used to establish the absence of any measurable ( @xmath15 1 - 2 % ) secondary co@xmath24o@xmath25 phase .
the coo powder was then compressed into rods with a hydraulic press and the rods annealed at 1275@xmath27c for 24 hours under an ar atmosphere in a horizontal annealing furnace .
single crystals , 8 mm diameter and 100 mm in length , were then grown in a four - mirror optical floating - zone furnace .
the growth was at a speed of 2 - 4 mm / hr with the feed and seed rods counter rotating at 35 rpm in an ar atmosphere .
initially , a polycrystalline seed rod was used but on subsequent runs was replaced by a single crystal seed from earlier runs . the magnetic susceptibility measured on a 0.28 g sample in an applied magnetic field of 100 g along the [ 100 ] axis .
no measurable co@xmath24o@xmath25 impurity phase is present as evidenced by the lack of any anomaly at low temperatures @xmath15 30 k.,width=283 ] the crystal was cut to a length of 50 mm and used in the experiments described in this paper and for the lower energy studies described earlier .
@xcite the nature of the surface of the coo crystal was examined with optical and scanning electron microscopy and the results showed there was only one growth domain .
x - ray scattering measurements showed that the surface had a mosaic spread of about 0.1@xmath27 and no evidence was found at t = 295 k for multiple magnetic or strain domains .
finally , a small part of the sample was studied with a squid magnetometer and the results for the magnetic susceptibility are shown in fig .
the data agree well with previously published measurements .
@xcite the phase transition at 295 k is the antiferromagnetic transition of coo and the absence of any discontinuities below 20 k shows that there is no significant amount of co@xmath24o@xmath25 ( @xmath28 0.5% ) impurity
. we also used a powdered sample of co@xmath12 and mgo .
the samplewas prepared by mixing the appropriate amounts of co@xmath24o@xmath25 and mgo in a ball mill for 5 hours .
the powder was annealed at 1200@xmath27c in a flowing argon atmosphere for 36 hr with intermediate grinding .
the powder was then examined by x - ray diffraction and no observable ( @xmath15 2% ) secondary phase was present .
the 45.8 g of material was placed in an airtight al can under helium . the pure coo powder used in the experiment
had a mass of 15.7 g , and was also sealed in a similar manner .
the pure mgo powder was annealed at 1250 k for 12 hours and placed in another air - tight al can .
the mgo mass was 32.5 g. these precautions involving high temperature annealing and sealing of the samples were taken because the fine powder was susceptible to water absorption producing spurious high - energy excitations related to hydrogen motion .
the experiments mostly described here were performed using the maps and mari direct geometry ( fixed initial energy ) chopper spectrometers located at the isis spallation neutron source .
the high - energy measurements were made on the maps spectrometer with fixed incident energies of e@xmath29=2 ev and 3.5 ev .
the @xmath30 chopper spun at 100 hz in parallel with the sloppy " fermi chopper which was operated at a frequency of 600 hz .
lower incident energies were used to search for other crystal field excitations and to investigate the spin - orbit splitting in powder samples .
the @xmath30 chopper was spun at the nominal repetition rate of 50 hz for these experiments with incident energies of e@xmath29=750 mev , 250 mev , and 85 mev .
the fermi chopper was spun at 600 hz for all incident energies .
a bottom loading displex cooled the sample to temperatures as low as 15 k. for lower energies , the mari direct geometry spectrometer was utilized to study the powder samples of mgo and mg@xmath4co@xmath5o .
the @xmath30 chopper was operated at 50 hz in parallel with a gd chopper operating at 300 hz with an incident energy of 85 mev .
a thick disk chopper with a frequency of 50 hz reduced the background from high - energy neutrons .
a top loading displex cooled the sample to 5 k. searches for low - energy crystal fields below the gapped fluctuations were performed on powders of coo with e@xmath29=100 and 10 mev with the gd fermi chopper spun at f=350 and 250 hz respectively .
the lowest energy experiments were performed with the iris indirect geometry spectrometer at isis .
the final energy was fixed at e@xmath31=1.84 mev by means of a cooled pg002 analyzer crystals in near backscattering geometry . through the use of three different time windows , the energy transfers of the neutron scattering could be measured up to @xmath15 2 mev .
a top loading displex was used to cool the sample to 11 k. for the high - energy experiments on maps , the single crystal coo sample was aligned such that reflections of the form ( h , h , l ) lay within the horizontal scattering plane .
the magnetic signal was extracted from constant energy cuts as discussed in the next section .
the sample was nominally aligned with the [ 001 ] axis parallel to @xmath32 .
no angular dependence was observed when the sample was rotated as discussed below in the experimental results section and therefore all the angles were averaged to improve statistics .
background originating from the tails of the elastic line and multiple phonon scattering result in a steeply decreasing intensity curve as illustrated through out this text and in particular highlighted in figs .
[ ei_2ev ] @xmath16 and [ ei_3p5ev ] @xmath16 of the experimental section presented below . to remove this background and extract the purely magnetic signal , we relied on the fact that the magnetic contribution to the scattering decreases with wavevector transfer while any background contribution will remain constant or increase with momentum transfer . assuming that the background at large scattering angles is dominated by phonons , the scattering was fitted at each energy transfer to the form ,
@xmath35 where @xmath36 characterizes the strength of the magnetic scattering , @xmath37 is proportional to the full width of the scattering in momentum , @xmath38 is a constant background , and @xmath39 characterizes the phonon contribution from both single and multi - phonon scattering . as seen in fig .
[ ei_250mev ] and the representative scans in figs .
[ ei_2ev ] and [ ei_3p5ev ] , the fit provides a good description of the wave vector dependence of the scattering .
@xmath16 constant energy slice through the paramagnetic scattering of coo at e=45 @xmath0 5 mev .
@xmath20 illustrates a cut along the horizontal detector bank integrating over @xmath0 0.52 @xmath40 in the out of plane detectors .
the magnetic scattering is evidenced by the intensity centered on @xmath41=0 that shows a decrease of intensity with increasing momentum transfer characteristic of magnetic scattering.,width=321 ] the procedure is shown in fig .
[ ei_250mev ] which illustrates a constant energy slice and cut at e=45 mev in the paramagnetic phase of coo projected onto a plane with axes q@xmath42 and q@xmath34 where q@xmath42 is aligned along [ 1 , 1 , 0 ] and q@xmath34 is aligned along [ 1 , @xmath43 , 0 ] .
q@xmath42 is defined to lie in the scattering plane and therefore points in the direction of the high - scattering angle bank on maps which extends out to @xmath15 60@xmath27 .
likewise , q@xmath34 is directed along the vertical direction . the total momentum transfer @xmath41 in terms of q@xmath42 for the in plane detectors ( where the average of q@xmath440 ) is equal to @xmath45 the data shown in figs .
[ ei_2ev ] and [ ei_3p5ev ] , for e@xmath29=2.0 ev and 3.5 ev used an integration in q@xmath34 of @xmath0 0.52 @xmath40 and @xmath0 2.1 @xmath40 , respectively .
the use of this coordinate system is displayed in fig .
[ ei_250mev ] @xmath20 which shows a cut along q@xmath42 integrating over the vertical direction of q@xmath34 by @xmath0 0.52 @xmath40 .
the cut shows a peak centered at q@xmath42=0 with significant width demonstrating it is of magnetic origin .
the solid red curve is a fit to eqn .
[ background ] and provides a good description of the momentum dependence of the data . for the incident energies discussed in figs .
[ ei_2ev ] and [ ei_3p5ev ] , we have performed similar constant energy scans and fits to remove the background .
the background corrected plots in figs .
[ ei_2ev ] and [ ei_3p5ev ] illustrate the integrated magnetic intensity derived from such fits .
in this section of the paper , we outline the relevant aspects of the crystal field theory required to understand the single - ion excitations in coo .
there are two important energy scales - the cubic crystal field and the spin orbit coupling .
the hamiltonian can therefore be written as , @xmath46 coo is cubic at high temperatures and undergoes a small ( @xmath47 ) structural transition to a tetragonal unit cell below 295 k. for the discussion here , which describes data in the cubic phase at t=300 k , we will not consider the effect of the tetragonal distortion on the hamiltonian .
we discuss the crystal field theory of co@xmath12 in coo in terms of the methods outlined in ref . .
there are two possible approaches to the crystal field problem ; the first is to assume that the effect of the crystalline field from the neighboring ions is small compared with the differences between the free ion energy levels ; the second approach is to assume that the crystal field is large compared with the splitting of the atomic levels .
we describe both of these approaches known as the weak and strong crystal field theory respectively .
as found from our experiments , cobalt oxide is an intermediate - weak case for which the crystal field is comparable to the energy differences between the atomic levels , while remaining in a high - spin state .
we first review the free ion solution for an isolated co@xmath12 and write the solution in terms of racah parameters ( @xmath48 ) , slater integrals ( @xmath49 ) , and the hubbard model ( @xmath50 ) .
the cobalt ion , co@xmath12 , has the electronic structure given by @xmath51 .
this structure consists of either filled or empty electronic shells apart from the @xmath7 shell which can accommodate 10 electrons but which is occupied by only 7 .
the energy eigenvalues of the isolated @xmath52 are degenerate in the absence of electron - electron interactions , however are split in the presence of these coulomb interactions .
historically , there have been several ways to parameterize this interaction with the original theory by slater followed by a different parameterization of racah .
more recently , this theory has been cast in terms of the hubbard model .
we outline the three different parameterizations below for the free ion case .
the free ion solution , in the absence of a crystalline electric field that is present in a solid , for @xmath53 can be written in terms of the racah parameters @xmath54 ( ref . ) .
the thee lowest energy states then have the energies , @xmath55 these parameters are a different way of writing the solution that was originally derived in terms of slater integrals .
the racah parameters can be recast in terms of the slater integrals as , @xmath56 however , another definition for the slater terms @xmath57 ( ref . ) is also sometimes used and is related to the slater terms @xmath58 used above by the following relations , @xmath59 the parameters above allow two different means of describing the free ion energy values for @xmath53 co@xmath12 and indeed other transition metal ions .
physically , the terms @xmath38 and @xmath39 represent the effect of the interactions between the electrons in the @xmath7 orbitals . experimentally , it is found that the ratio @xmath60 is close to a constant equal @xmath15 4.6 for the @xmath7 ions and therefore the racah parameter @xmath38 is taken as a means of solely quantifying electron correlations .
based on this observation , it is convenient to introduce a single parameter that captures the effect of electron correlations that characterizes the intra - orbital exchange . in terms of the slater integrals listed above ( which can also be written in terms of racah parameters )
an intra - orbital exchange can be defined which describes the attraction of parallel spins in different @xmath61 orbitals ( ref . ) .
@xmath62 using the above expressions , @xmath1 can be written in terms of either slater ( @xmath49 ) integrals or racah ( @xmath63 ) parameters .
ref . also defines an additional quantity @xmath64 ( not to be confused with the racah parameter @xmath39 ) and it is defined as follows .
@xmath65 this quantity physically represents an angular part of the multiplet splitting . a third quantity known as the hubbard @xmath66
is also defined and can be written in terms of the integral @xmath67 and @xmath1 as outlined in ref . .
this parametrization is particularly useful as the hund s rule ground state energy can be written as @xmath68 for @xmath52 . noting that @xmath69 ( ref
) this reduces to @xmath70 illustrating the fact that a larger @xmath1 favors a hund s rule ground state - a result demonstrated schematically in our tanabe - sugano diagram ( fig . [ summary ] ) .
the above summary demonstrates three ways of presenting the solutions for the free ion levels of @xmath53 .
the first is in terms of racah parameters ( @xmath48 ) , the second in terms of slater terms ( @xmath71 ) , and the third in terms of parameters @xmath72 .
it can be seen that the racah parameter @xmath36 and the hubbard @xmath66 can not be directly probed in our experiment that is only sensitive to transitions between energy levels .
this is obvious from the solutions written in eqn .
[ solutions ] which all depend on @xmath73 . therefore , our neutron scattering experiments are only able to determine racah parameters @xmath63 and also the more recent parameterization values of @xmath74 . in weak crystal fields
the ground state of the electronic wave functions for the 3d electrons is given by hund s rules as for the atomic ground states .
the first rule states that the electrons in any partially occupied shell should be aligned so that the total spin @xmath75 is a maximum , consistent with the pauli principle .
this implies that five of the electrons will have parallel spins with the two additional electrons having oppositely aligned spins giving a total spin of s = @xmath76 as illustrated in the left side of fig .
[ summary ] @xmath20 .
the second of hund s rules states that the total orbital angular momentum , @xmath77 , will be maximized subject to the constraints caused by application of the first of hund s rules .
this implies that the first 5 electrons half - fill the 3d orbitals and that the orbital angular momentum of these electrons is zero .
the remaining two electrons have the maximum total angular momentum so that @xmath78 . in the spectroscopic notation the ground state of the isolated co@xmath12 ion
the ground state is then @xmath79 ( not to be confused with the slater @xmath80 discussed above ) .
the higher energy free ion states are illustrated on the left side of fig .
[ summary ] @xmath16 with the energy scale set by the racah parameters @xmath38 and @xmath39 . taking the @xmath79 ground state for co@xmath12
, we now consider the effect of a crystalline electric field resulting from the symmetry imposed by the lattice of coo . in the absence of mixing between the ground @xmath79 state and the excited levels ( such as the @xmath81 ) ,
the cubic crystal field is given by the following hamiltonian cast in terms of stevens operators ( @xmath82 and @xmath83 ) and numerical coefficients ( @xmath84 ) ( refs . ) , @xmath85 in the high temperature cubic phase of coo , the octahedral cubic symmetry splits the 7 orbital states of the lowest energy hund s rule @xmath79 state into two orbital triplets ( @xmath86 and @xmath87 ) and one singlet ( @xmath6 ) as illustrated in the tanabe - sugano diagram plotted in fig .
1 @xmath16 . for small values of @xmath84 , the energy splitting between the levels
are @xmath88 and @xmath89 . an alternative approach , to that used in the preceding section and fig . 1 @xmath16 , is to assume that the energy splittings caused by the crystal field are much larger than the difference in energy between the free - ion states .
the analysis is similar to that described above , but instead of taking the free ion states as our basis we take the strong crystal field basis where the 3@xmath61 orbital states are split into triplet @xmath90 and doublet @xmath91 states with an energy difference defined by @xmath92 . in an octahedral field ,
the @xmath90 states have the lowest energy ( @xmath93 ) while the @xmath91 state is the highest ( @xmath94 ) .
the ground state electronic configuration can be derived using hund s rules in the context of a large energy gap between the @xmath90 and @xmath91 states .
in such a scenario , it becomes energetically favorable to fill up the @xmath90 states before populating the @xmath91 levels .
this gives rise to a ground state illustrated in the right hand portion of fig .
[ summary ] @xmath20 where only one electron remains unpaired and @xmath95 .
the ground state is a doublet labelled as @xmath96 in fig
. 1 @xmath16 .
the energy scheme as a function of the energy splitting @xmath2 between the @xmath90 and @xmath91 states is also dependent on the internal intra - orbital coulomb interaction between the electrons which can be cast in terms of the quantity @xmath1 outlined above .
detailed calculations have been made by tanabe and sugano ( refs . ) and by liehr ( ref . ) .
the calculations we have done are shown in fig .
[ summary ] @xmath37 ) where we have calculated the co@xmath12 energy levels as a function of crystal field strength ( @xmath2 ) taking the strong crystal field @xmath90 and @xmath91 states as our basis .
the calculations used the electrostatic matrices and character tables supplied in ref . .
the calculations depend on the racah parameters @xmath38 and @xmath39 and we find that values close to those provided by ref . provide a reasonable description of the data .
the energy eigenvalues calculated from a strong crystal field basis link up with the free ion values for small crystal fields as illustrated in fig . [ summary ] @xmath16 .
the calculation displayed in fig .
1 also provides an estimate for the crystal field energy where the ground state switches from the high spin @xmath86 ( predicted based upon a weak crystal field scheme taking the free ion states as the basis functions ) to a low spin @xmath96 state ( expected based upon a strong crystal field basis ) .
the tanabe - sugano diagram is traditionally drawn as a function of @xmath97 , where @xmath38 is the racah parameter .
this is because the ratio @xmath98 is close to a constant of @xmath15 4.6 across the @xmath7 transition metal ions . to make the connection with the hubbard model and recent lda+u models , we have redrawn the tanabe - sugano diagram in terms of the quantity @xmath17 while fixing the quantity @xmath64
this analysis is justified by the fact that we have obtained three different energies allowing us to uniquely derive the three experimental quantities @xmath2 ( the crystal field splitting ) , @xmath1 ( intra - orbital exchange ) , and also @xmath64 ( the angular part ) which can be compared with theory .
we can also work back ( noting the definitions written above for the free ion solutions ) to obtain the racah parameters @xmath38 and @xmath39 which can be compared with works done on dilute and free ion systems and published in refs . .
this comparison is made later in the presentation and parameterization of the experimental results .
a weaker energy scale is given by the spin - orbit coupling , represented by the following term in the hamiltonian , @xmath99 @xmath100 and @xmath101 are the orbital and spin angular momentum respectively .
@xmath3 is the spin - orbit coupling constant . for simplicity , we ignore here extra factors used to account for strong covalency . given that the @xmath86 ground state is separated by @xmath15 1 ev from the first excited state , we consider only the effect of this hamiltonian on the ground orbital triplet within a fixed manifold of constant @xmath100 and @xmath101 .
the effect of the spin orbit term on the 12 fold degenerate ( 4 spin states @xmath102 3 orbital ) @xmath86 ground state of coo . assigning a fictitious orbital angular momentum of @xmath103 results in the crystal field scheme on the right .
the degeneracy of each spin - orbit split level can further be broken by the presence of a molecular field from neighboring ordered moments .
the experiment overcame this complication by studying the spin - orbit transitions in dilute samples of mgo doped with 3% co@xmath12.,width=321 ] for the orbital triplet ground state , it is useful to consider a total angular momentum @xmath104 with an orbital angular moment of @xmath103 and orbital moment projection factor @xmath105 .
the spin - orbit hamiltonian then becomes , @xmath106 with fixed magnitudes @xmath107 and @xmath108 , the energy eigenvalues are then given by @xmath109 .
\label{energies}\end{aligned}\ ] ] for @xmath103 and @xmath110 there are three energy eigenvalues illustrated in fig .
[ spin_orbit_diagram ] .
the ground state corresponds to a doublet with @xmath111 and with excited states of @xmath112 and @xmath113 .
the energy difference between the ground state and the lowest excited state is @xmath114 .
the selection rules for dipolar neutron scattering show that for the allowed transitions , the angular @xmath115 changes by 0 or @xmath0 1 .
this implies that at low temperature when only the lowest energy state is occupied , the spectra will show only transitions within the @xmath116 doublet and from those states to the @xmath117 states at an energy @xmath118 .
the degeneracy of each level displayed in fig .
[ spin_orbit_diagram ] can be split in the presence of a molecular field induced through magnetic order and superexchange .
the field on each co@xmath12 site therefore depends on the magnetic structure and also the values of the nearest neighbor and next nearest neighbor superexchange values . to avoid these extra terms
we have measured a sample of mgo diluted with a small amount of co@xmath12 to isolate and measure the pure spin - orbit term independent of the exchange parameters .
this is described in the following sections .
this section discusses the experimental results and the connection with the crystal field theory outlined above .
as discussed earlier in the paper , the experiments were performed with the isis spallation neutron source that produces a burst of high - energy neutrons from a pulsed proton beam striking a tungsten target . while the high - energy neutrons are moderated , a significant epithermal flux of neutrons above @xmath15 1 ev remains .
@xcite high energy experiments where performed at the maps spectrometer to characterize the crystal field and intra - orbital exchange .
lower energy experiments to determine the spin - orbit coupling were performed on the mari and iris spectrometers .
the electronic ground state of the seven @xmath7 electrons of a co@xmath12 ion situated in coo are largely determined by two energy scales - the intra - orbital exchange and the crystalline electric field .
the exchange can be characterized by a parameter @xmath1 where a large @xmath1 results in a ground state determined by hund s rules .
@xcite the crystalline electric field is characterized by a parameter @xmath2 which has the effect of splitting the degeneracy of the @xmath61 orbitals .
the energy eigenvalues as a function of @xmath17 is illustrated through the tanabe - sugano diagram ( fig .
[ summary ] ) which displays the low - energy eigenvalues as a function of @xmath17 for a co@xmath12 in an octahedral environment .
@xcite the right side of the figure represents the strong crystal field limit where the doubly degenerate @xmath119 levels ( at an energy of @xmath120 ) and triply degenerate @xmath121 ( at an energy of -@xmath122 ) are taken as the basis set .
the critical field where the ground state changes occurs when the crystal field parameter @xmath123 2.5 ( shown by the vertical dashed line in fig . [ summary ] ) . in small crystal fields co@xmath12 adopts a high - spin ( @xmath124 ) configuration as given by hund s rules while for large cubic crystal fields a low - spin configuration ( @xmath125 ) is the ground state if the energy scale of the crystal field splitting dominates over the intra - orbital exchange @xmath1 .
we first consider the high - energy electronic levels and later study the much lower spin - orbit energy scale .
figure [ ei_2ev ] shows data taken using the maps spectrometer on a single crystal of coo with e@xmath29=2.0 ev .
uncorrected data integrating over small angles is plotted in panel @xmath16 and background corrected constant - q scans are shown in panels @xmath126 and @xmath127 at 300 k and 15 k in the disordered cubic and spin - ordered tetragonal phases respectively .
the background and magnetic components of the scattering were extracted through a series of constant energy scans , examples of which are illustrated in @xmath128 .
each constant energy scan was fitted to a gaussian centered at q@xmath42=0 plus a term modelled as @xmath129 to account for multiple scattering resulting from the phonons ( seen in panel @xmath130 .
the momentum dependence allows us to separate magnetic scattering from background inconsistent with any electronic origin .
panels @xmath126 and @xmath127 show magnetic peaks at 0.875 @xmath00.009 ev and 0.917 @xmath00.008 ev at 300 and 15 k respectively . @xmath20 illustrates a plot of the intensity of the 0.88 ev ( t=300 k ) peak as a function of angle between the [ 001 ] axis and the incident beam .
0@xmath27 represents the case with [ 001 ] parallel to @xmath131 and at 90@xmath27 , a [ 110 ] axis is parallel to @xmath32 . no change in the intensity or the energy
is observed .
@xmath16 constant q scans at 15 and 300 k summing over the low - angle detector banks .
@xmath132 illustrate background subtracted scans at 300 and 15 k. @xmath133 plot constant energy scans at 300 k over the energies indicated .
the strong decay with momentum illustrates the magnetic nature of the scattering .
note that the @xmath134 axis is defined in terms of one momentum transfer element @xmath135 as defined above.,width=347 ] figure [ ei_3p5ev ] illustrates a similar series of scans taken with e@xmath29=3.5 ev .
@xmath16 shows the uncorrected data with @xmath20 and @xmath126 plotting background corrected scans in the cubic ( 300 k ) and tetragonal ( 15 k ) phases respectively .
panels @xmath127 and @xmath136 illustrate constant energy scans and show decreasing intensity with increasing momentum transfer .
this proves the electronic origin of these excitations . at 300 k
, peaks occur at [email protected] and [email protected] ev . in the ordered tetragonal phase at 15 k , a shift in the energy spectrum is measured with a broad peak observed at 1.93 @xmath0 0.04 ev ( panel @xmath137 ) .
we shall initially describe the results for the neutron scattering at high temperatures in the cubic phase ( 300 k ) .
three peaks were observed with energies below 3 ev namely ; [email protected] ev , [email protected] ev , and [email protected] ev . a comparison with calculations
is shown in fig .
[ summary ] .
the diagram requires the input of 3 parameters .
two parameters are the racah parameters @xmath63 or hubbard parameters @xmath74 and describe the electronic structure in the absence of a crystal field term .
the third parameter is the crystal field term ( @xmath2 ) and describes the splitting between the @xmath91 and @xmath90 levels ( fig .
[ summary]@xmath20 ) .
the diagram does not depend on the large hubbard @xmath66 and our experiments are not sensitive to this value .
a result of crystal field theoretical and experimental work is that the ratio of @xmath138 and also @xmath139 are roughly a constant for @xmath7 transition metal ions and therefore the diagram is qualitatively fixed through just two parameters - @xmath2 ( crystal field splitting ) and @xmath1 ( intra - orbital exchange ) .
@xcite our experiment is compared with the calculation by varying these two parameters by the shaded areas that show the energies of the excitations while the size of the boxes are representative of the errors in the measured energies and the fitted values of @xmath17 .
the data is inconsistent with the energies on the right hand side of the figure and hence show that the ground state is the high - spin @xmath86 triplet , as expected from the large spontaneous magnetization of the antiferromagnetic state .
the energy scheme is reproduced by the calculation if , @xmath140 and the crystal field splitting is @xmath141 the value of @xmath1 agrees well with the local density approximation ( lda ) calculations that reported 0.92 ev ( ) and the interpolation formula which give a value of @xmath15 1.0 ev ( ref . ) .
the third parameter was also allowed to vary and is in good agreement with theory as @xmath142 comparing well with the calculated values of @xmath15 0.5 .
@xcite again , our experiment is not sensitive to the hubbard @xmath66 estimated at @xmath15 2 ev . the parameters also compare very well with calculations presented in ref . which suggest a value of @xmath1=0.92 ev and also the hartree - fock result - ( @xmath143 ev ) ( where @xmath144 is the atomic number ) and empirical relations ( @xmath145 ev ) stated in ref . .
based on our high energy experiments , we find that coo can best be described in the intermediate to weak crystal field limit in which @xmath86 is the ground state .
coo is therefore far from a high - low spin transition .
however , we note that while the crystal field parameters are in good agreement with theory , the crystal field gap @xmath2 is roughly the same size as @xmath1 . therefore , while coo maybe considered in the weak crystal field limit , the crystal field energy is sizable and comparable to the intra - orbital exchange . using the equations outlined above regarding the relevant crystal field theory ,
we can also write the parameters in terms of racah parameters . we therefore derive the following values for co@xmath12 in coo from the parameterization listed above , @xmath146 this should be compared with @xmath38=0.13 ev and @xmath39=0.64 ev quoted in the summary listed in ref . .
the ratio of @xmath147 is close to the values derived for co@xmath12 in molecules which is typically around 4.8 .
@xcite in the previous section , it was discussed how we calculated the energy eigenvalues in terms of a modified sugano - tanabe diagram shown in fig .
[ summary ] .
our experiment observed energies at [email protected] ev , [email protected] , and [email protected] ev . based on these energies we derived crystal field parameters that are in good agreement with previous experiment and theory .
furthermore , the 0.87 ev and 2.3 ev excitations are in excellent agreement with non - resonant inelastic x - ray experiments .
@xcite we now discuss the neutron intensities and selection rules for the transitions measured above and expected based upon the analysis in fig .
[ summary ] .
transitions based upon matrix elements of the form @xmath148 ( expected in the dipolar approximation ) were calculated so as to determine which transitions have a finite cross section .
note that @xmath14 .
the eigenstates of the initial and final states were determined by relying on the tables provided in refs . .
these give the eigenstates of the levels in question ( figs .
[ summary ] , [ ei_2ev ] , and [ ei_3p5ev ] ) in terms of slater determinants of the basis states @xmath149 .
this corresponds to calculating the matrix elements from both orbital and spin operations of @xmath150 and @xmath151 .
an orbital transition to the excited @xmath87 level is allowed for dipolar transitions and accounts for the [email protected] ev excitation ( shown in fig .
[ ei_2ev ] ) . based on its energy ,
the [email protected] ev excitation originates from either the @xmath96 or @xmath6 level , but both are forbidden for dipolar matrix elements .
however , a transition involving a large quadrupolar matrix element is allowed to the @xmath6 and has been predicted but not resolved with x - ray experiments . @xcite the [email protected] ev peak corresponds to a @xmath86 excitation which has been observed with x - rays previously and also calculated to have a finite quadrupolar matrix element .
while this analysis is for the cubic phase , below 300 k coo is antiferromagnetically ordered and tetragonal and it is expected that the magnetic molecular field will lower the ground state energy .
figures [ ei_2ev ] and [ ei_3p5ev ] show that the two lowest energy excitations increase on cooling by [email protected] ev and [email protected] ev , respectively .
although neither the magnetic molecular field nor the tetragonal distortion are known , the increases in the energy of the excitations are similar to the changes in energy expected from magnetic ordering given the estimated exchange constants .
@xcite having presented the high energy crystal field scheme characterizing the crystal field component of the hamiltonian , we now discuss the second important energy scale referred to above , namely the spin - orbit coupling ; @xmath152 .
as outlined above , coo is cubic in the paramagnetic phase , and the orbital triplet ground state of the co@xmath12 ion can be represented by a fictitious @xmath9=1 with @xmath153 , where @xmath154 . within the manifold of the @xmath86 states ,
the spin - orbit coupling becomes @xmath155 and splits the 12 states into 2 degenerate states with @xmath156 , 4 with @xmath157 and 6 with @xmath158 where @xmath159 .
the energy difference between the ground state with @xmath160 and the excited state with @xmath161 is @xmath162 ( note that @xmath3 is negative ) .
these energy levels can not be directly measured in pure coo because of large magnetic exchange terms which split the degeneracy of the energy levels summarized in fig .
[ spin_orbit_diagram ] therefore greatly complicating the excitation spectrum measured with neutrons .
we have overcome this by studying dilute samples of mg@xmath4co@xmath5o for which most of the co@xmath12 atoms are isolated from another co atom and because mgo remains cubic down to the lowest temperatures measured ( 5 k ) therefore removing the need to include terms due to the crystal field distortion . the inelastic response in mgo and
dilute mg@xmath4co@xmath5o at 5 k. @xmath16 and @xmath20 illustrates the response in pure mgo and dilute mg@xmath4co@xmath5o respectively .
@xmath126 shows a subtraction illustrating the response from dilute co@xmath12 . @xmath127 and @xmath136 plot q and energy averaged cuts demonstrating a peak at 37.1 @xmath0 0.5 mev and that it decays with momentum transfer as expected from the form factor for the co@xmath12 magnetic moment ( solid line ) .
the magnetic peak positions was used to extract the spin - orbit coupling constant for coo .
the use of dilute samples avoids complications due to the exchange terms in the hamiltonian.,width=347 ] fig .
[ spin_orbit ] shows high resolution results for powders of mgo and mg@xmath4co@xmath5o with @xmath16 and @xmath20 plotting energy and momentum slices .
the mgo spectrum is dominated by two phonon bands at @xmath15 40 and 50 mev , however the doped samples also shows additional scattering at small momentum transfers illustrated by the subtraction in fig .
[ spin_orbit ] @xmath126 .
[ spin_orbit ] @xmath127 and @xmath136 show constant wave vector cuts and constant energy scans through the band of excitations with a peak at @[email protected] mev and with a full width of [email protected] mev .
the energy width is considerably broader than the resolution width ( 1.8 mev ) due to the different environments around each co@xmath12 ion .
the momentum dependence ( @xmath136 shows a decrease in the intensity confirming the magnetic origin of this intensity where the solid line is the square of the co@xmath12 form factor . @xcite the energy difference between the @xmath116 and @xmath117 levels , @xmath163 then fixes the spin - orbit coupling as @xmath164= -0.016@xmath0 0.003 ev .
the intensity of the transition from the @xmath165 level to the @xmath166 is forbidden for the neutron scattering cross section , while the transition from @xmath117 to @xmath167 is suppressed at low temperatures .
the theoretical free - ion value for @xmath3 is -0.0234 ev while the experimental value is listed as -0.0221 ev .
@xcite indirect and model dependent attempts at extracting @xmath3 in nominally pure coo have been made and give varied results .
soft x - ray and infrared techniques ( ref . )
give -18.9 mev and a spin - wave analysis based upon the lower magnetic branches gave -12 mev ( ref . ) .
these results need to include the values of the exchange parameters which requires a detailed knowledge of the momentum dependence of the spin excitations and the various exchange constants along with terms accounting for the structural distortion at low temperatures .
all results , including here , show systematically smaller values than the free ion values .
this decrease could be indicative of coupling to the excited @xmath81 state , fig .
[ summary ] or be due to covalency with oxygen ions .
the lack of any new excitations at high - energies , such as charge transfer excitations involving the o@xmath168 ions , is suggestive that the reduced spin - orbit coupling is the result of mixing between the ground state and the excited @xmath81 levels .
the role of charge - transfer excitations and the relevant cross section for neutrons requires further investigation . while the energy scale of the orbital excitations is expected to be large , some neutron scattering results ( ref . ) have suggested the presence of very low - energy peaks of the order @xmath15 1 mev and crystal field excitations with these low energies are inconsistent with the crystal field description provided above along with the comparison with lda+u models .
we have performed a search for these excitations using both direct and indirect geometry neutron spectrometers and discuss the results here . to search for possible low - energy crystal field excitations
, we performed a series of experiments using both powder and single crystals of coo .
figure [ mari_iris ] illustrates the scattering from a powder averaged @xmath169 map of coo measured with the mari spectrometer .
there is a band of excitations from @xmath15 30 - 70 mev , previously investigated in other papers ( refs . ) , and associated with the spin / orbital excitations . a weak phonon band , also previously measured for single crystals ( ref . ) ,
is seen at @xmath15 20 mev . below this band , there is little evidence for magnetic scattering as demonstrated in panels @xmath20 and @xmath126 . while some low - energy peaks ( @xmath15 1 - 2 mev ) were visible , we found these peaks changed with the incident energy and experimental configuration and therefore associate them with double scattering from shielding walls around the sample position .
the peaks are also sharper than both the measured and calculated experimental resolutions in energy .
we therefore have no evidence for the low - energy spin excitations claimed to be seen in ref . below the gapped ( @xmath1525 mev )
one - magnon excitations described previously .
@xcite @xmath16 powder averaged @xmath169 map of the low - energy ( e@xmath28100 mev ) scattering from a powder of coo .
the one - magnon cross section is readily observed between @xmath15 30 mev and 70 mev . @xmath20 and
@xmath126 illustrate the absence of any obvious extra peak at low energies below @xmath15 6 mev.,width=359 ] scans performed on maps using a single crystal sample of coo with @xmath32 aligned along [ 001 ] .
@xmath16 illustrates data taken with e@xmath29=250 mev showing the spin / orbital excitations between @xmath15 20 - 75 mev .
@xmath20 shows a similar scan taken with e@xmath29=750 mev demonstrating no obvious crystal field peaks for energy transfers between 100 and 600 mev.,width=264 ] for completeness , we continued the search for crystal field excitations , above the top of the one - magnon band of @xmath15 70 mev , using the maps chopper spectrometer .
figure [ lowe ] illustrates data taken with a single crystal of coo using the maps spectrometer with @xmath32 aligned along [ 001 ] .
[ lowe ] @xmath16 illustrates data taken integrating over low angles and illustrates the presence of significant magnetic scattering in the range of @xmath15 20 - 75 mev .
we associate this scattering with excitations within the @xmath86 ground state referred to as the one - magnon above and shown in fig .
[ mari_iris ] @xmath16 .
these have been discussed in detail in several other papers but are not discussed in detail here .
@xcite fig .
[ lowe ] @xmath16 does illustrate the absence of any additional peaks between 100 and 200 mev .
a similar scan is illustrated in fig .
[ lowe ] @xmath20 and further indicates a sloping background with no obvious peaks below @xmath15 600 mev .
based upon these two scans performed on maps , we conclude there are no crystal field excitations between 100 and 600 mev . based on the low - energy searches performed on the maps , mari , and iris spectrometers .
we conclude there is no observable magnetic scattering below the gapped one - magnon scattering extending over the energy range from @xmath1520 to 70 mev .
this conclusion agrees with our crystal field analysis presented above and also with corroborating calculations based upon band theory .
the previous section described the single ion excitations and derived the relevant parameters in terms of classic crystal field theory ( racah parameters and tanabe - sugano diagrams ) and also in terms of more recently developed lda+u band theories .
we also used dilute samples to extract the spin - orbit coupling .
the use of dilute samples avoided complications due to spin and orbital exchange which then makes it difficult to understand the neutron cross section and excitation spectrum .
we now summarize these results by comparing them with existing x - ray and optical techniques .
we now compare neutron and inelastic x - ray results .
non - resonant x ray experiments ( ref . ) gave peaks in coo at @xmath15 1.0 and 2.3 ev and similar energies were derived from infrared and soft x - ray experiments ( ref . ) .
the excitations were measured and had significant anisotropy .
they were understood in terms of band structure calculations requiring a large electronic gap @xmath15 3 ev .
@xcite measurements using other x - ray resonant spectroscopy techniques have found further peaks located just below @xmath15 2.0 ev .
we observe three excitations , in the cubic phase , at [email protected] ev , [email protected] , and [email protected] ev and find no angular dependence or dispersion within error .
the 0.870 and 2.3 ev peaks are in agreement with x - rays but the 1.84 ev excitation was not resolved with non - resonant x - rays and is forbidden in the dipolar approximation , however it has been reported in other optical and x - ray studies such as rixs and eels that are sensitive to dipole forbidden cross sections .
@xcite this result highlights the point that the 1.84 @xmath00.03 ev excitation is dipole forbidden and confirms our claim that we are sampling a quadrupolar matrix element in our neutron scattering experiments .
one difference between the neutron results and the inelastic x - ray results is the range of momentum probed . due to kinematic constraints ,
our experiments occur at larger momentum transfers ( approximately a factor 2 ) than those reported using non - resonant inelastic x - rays ( @xmath170 13 @xmath40 here compared with 3.5 @xmath40 with x - rays ) . in this region , as pointed out in ref .
, other terms in the cross section need to be considered beyond the dipolar term and in particular the quadrupolar matrix element .
while there is no allowed excitation in the dipolar approximation , quadrupolar terms do allow a finite intensity for the @xmath6 level .
a similar result is predicted for the [email protected] ev which corresponds to a @xmath86 level .
we note that a neutron study of nio required only the use of dipole terms to understand the spectrum , however , only the lowest energy terms were probed .
@xcite the larger momentum transfers probed with neutrons here consequently imply that our experiments are sensitive to a different lengthscale to that of inelastic x - rays .
this may provide an explanation for the lack of angular dependence in the intensity .
the data can be consistently interpreted in terms of crystal field excitations , in agreement with calculations based upon lda+u as well as early works parameterized in terms of the tanabe - sugano formalism .
the results are orbital excitons within the mott - hubbard gap and do not probe the large u expected in initial x - ray results .
we do not observe evidence of new modes that would be expected for charge transfer excitations ( as observed in mott insulating cuprates @xcite ) or for propagating zhang - rice singlet states .
this suggests that these modes reside at different energies than were probed here or that they are weaker than our sensitivity .
@xcite however , our results show that coo is well described in terms of the class of a large-@xmath66 , hubbard mott insulators with crystal field excitons below the mott hubbard gap .
the observation of dipole forbidden transitions also demonstrates the use of neutrons in measuring such cross sections .
49ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | neutron scattering is used to investigate the single - ion spin and orbital excitations below the mott - hubbard gap in coo .
three excitations are reported at [email protected] ev , [email protected] , and [email protected] ev .
these were parameterized within a weak crystal field scheme with an intra - orbital exchange of @xmath1=1.3 @xmath0 0.2 ev and a crystal field splitting @xmath2=0.94 @xmath0 0.10 ev . a reduced spin - orbit coupling of @xmath3=-0.016@xmath0 0.003 ev is derived from dilute samples of mg@xmath4co@xmath5o , measured to remove complications due to spin exchange and structural distortion parameters which split the cubic phase degeneracy of the orbital excitations complicating the inelastic spectrum .
the 1.84 ev , while reported using resonant x - ray and optical techniques , was absent or weak for non resonant x - ray experiments and overlaps with the expected position of a @xmath6 level .
this transition is absent in the dipolar approximation but expected to have a finite quadrupolar matrix element that can be observed with neutron scattering techniques at larger momentum transfers .
our results agree with a crystal field analysis ( in terms of racah parameters and tanabe - sugano diagrams ) and with previous calculations performed using local - density band theory for mott insulating transition metal oxides .
the results also demonstrate the use of neutron scattering for measuring dipole forbidden transitions in transition metal oxide systems . |
molecular magnets make a wide emerging class of new materials whose properties extend the range of those typically associated with magnets , i.e. they offer low density , transparency , electrical insulation , and a possibility of low - temperature synthesis @xcite .
single - chain magnets ( scm ) belonging to the class of molecular magnets , have been of particular interest since gatteschi et al .
@xcite discovered slow relaxation of magnetization in a chain compound , comprising co(ii ) centers and organic radicals , without any evidence of phase transition to a three - dimensional magnetic ordering .
soon after , clerac et al . discovered similar magnetic properties in an @xmath7 heisenberg ferromagnetic chain comprising mn(iii)-ni(ii)-mn(iii ) trimers with an easy axis parallel to the chain @xcite .
next , other scm systems often based on ferrimagnetic chains containing alternating spins of unequal magnitude were discovered @xcite .
the slow magnetic relaxation arises from large uniaxial magnetic anisotropy , negligible magnetic interactions between the chains and considerable intrachain interactions @xcite .
the last condition is desirable for raising the blocking temperature .
moreover , for some one - dimensional quantum spin systems with competing nearest - neighbor and next - nearest - neighbor interactions @xcite the single - chain magnetic behavior can be affected with frustration @xcite .
the simplest description of ferromagnetic scms is based on the glauber theory if the limit of an ising chain can be explored @xcite .
the ising ferromagnetic chains display slow relaxation of magnetization and the relaxation time @xmath8 depends on the spin - pair correlations .
for these chains the correlation length @xmath9 diverges exponentially at low temperatures as @xmath10 where @xmath11 is the correlation energy and @xmath12 is the boltzmann constant @xcite .
the exponent @xmath11 is proportional to the activation energy which enters the arrhenius law @xcite and can be directly estimated from the experimental static susceptibility of the chain by the relation @xmath13 , plotting ln(@xmath14 ) as a function of 1/@xmath15 .
the correlation energy is also referred to as the energy to create a domain wall .
important fact is that @xmath11 is proportional to the magnetic coupling constant @xcite and similar conclusions remain valid for the antiferromagnetic ising model @xcite but then @xmath16 .
in addition , the correlation energy @xmath11 is also the main part of the additive term determining the relaxation time in the ising - like and anisotropic classical heisenberg chains @xcite . the relations between the correlation length and the zero - field susceptibility as well as between the correlation energy and the magnetic coupling can be used for the classical ferromagnetic heisenberg model in the strong anisotropy limit @xcite . if the hamiltonian is defined by @xmath17 the limit
is reached when the exchange coupling @xmath18 and the single - ion anisotropy @xmath19 satisfy the relation @xmath20 .
then the correlation energy calculated from the product @xmath14 yields the value of the coupling @xmath18 because @xcite @xmath21 for ferromagnetic systems , this scenario is well confirmed @xcite and has been exploited for a number of compounds @xcite , but for antiferromagnetic chains with non - collinear anisotropy axes @xcite the situation is much less clear .
the problem was revealed @xcite for the compound of formula [ mn(tpp)o@xmath22phph]@xmath23h@xmath22o considered a textbook example of scm ( tpp = _ meso_-tetraphenylporphyrin and phph = phenylphosphinate ) and referred to as mn - caf ( canted antiferromagnet ) .
it was possible to rationalise many experimental results performed on mn - caf , assuming the generalised expression for the correlation energy in the form @xmath24 where @xmath25 is the canting angle for two classical spins oriented along the anisotropy axes .
however , the proportionality between the correlation length @xmath9 and the product @xmath14 was ruled out , implying that the correlation energy can not be determined by the magnetic measurement of the static susceptibility . in this paper
we demonstrate that the correlation energy of the quantum model of a canted antiferromagnetic chain can be found from the in - field single - crystal magnetization profile properly chosen and its @xmath25 dependence is given by the classical expression ( [ eq : deltaksi ] ) .
these results suggest that the correlation energy @xmath11 can be estimated from the in - field magnetization measurements .
as soon as the spin value , the geometrical structure and @xmath11 are known , the value @xmath18 of the magnetic coupling follows from eq .
( [ eq : deltaksi ] ) . considering the empirical shape of the magnetization isotherm in the @xmath4 direction for mn - caf
, we directly find the @xmath18 value without any fitting procedure .
the paper is organized as follows : sec .
[ sec:1 ] is devoted to a model describing the anisotropic quantum spin systems and the numerical method for accurate calculations of magnetic properties in a wide temperature range .
[ sec:2 ] is dedicated to presentation of results and their discussion .
[ sec:3 ] concludes our paper , summarizing the main outcomes .
we consider the quantum anisotropic heisenberg model which is needed to get quantitative estimates of the thermodynamic properties @xcite of the chains with non - collinear anisotropy axes @xmath26 where spin @xmath27 and @xmath28 stands for the length of the chain .
the exchange coupling @xmath18 between nearest - neighbor spins is isotropic , uniform and positive .
the tensors representing the single - ion anisotropy @xmath29 and @xmath30-factors are non - diagonal and depend on angles ( @xmath1 , @xmath2 ) or ( @xmath31 , @xmath2 ) for odd and even sites , respectively .
the indices @xmath32 define the global coordination system .
the @xmath4 direction is chosen along the chain axis and plays also the role of the quantization axis . , where the spherical angles ( @xmath1 , @xmath2 ) are defined .
the black arrows represent the canted mn - caf spin arrangement , whereas the gray ones correspond to the collinear antiferomagnetic spin arrangement ( @xmath33 , @xmath34 ) . ]
the explicit form of the tensors in the model ( [ eq:01 ] ) depends on the angles ( @xmath1 , @xmath2 ) defining the anisotropy axes ( fig .
[ fig:00 ] ) and is given in the previous publications @xcite .
we note that for all the pairs of angles ( @xmath1 , @xmath2 ) the canting angle is determined from the relation @xmath35 analysis of our model ( [ eq:01 ] ) is based on the numerical quantum transfer matrix method @xcite , where the partition function of the quantum chain is mapped onto the partition function of the classical @xmath36 system with multi - spin interactions and a finite width @xmath37 @xcite . for different values of @xmath38 , called the trotter number , the classical partition functions form a series of approximants @xmath39 , where the leading errors are of the order of @xmath40 .
the higher the value of @xmath38 , the better the quantum nature of the problem is taken into account . as the hilbert space dimension increases exponentially with increasing trotter number ,
computations are feasible only for relatively small @xmath38 . to overcome this problem
, we employ the density - matrix renormalization group technique ( dmrg ) @xcite for determining an effective hamiltonian representation in the hilbert subspace of a given size @xmath41 which ensures covering the entire experimental temperature range .
the dmrg method is very powerful but it is computationally demanding as far as the resources and programming are concerned .
the transfer matrices are asymmetric and the corresponding density matrices , constructed from the right and left eigenvectors of the largest eigenvalue @xmath42 of a transfer matrix , are non - hermitian .
we were able to overcome these problems by applying complex algebra @xcite , which guaranties the biorthogonality of complex eigenvectors but increases the computational complexity . in some dmrg applications @xcite the hilbert space
is split into subspaces of fixed values of @xmath43 , where @xmath44 denotes a spin variable in a given column .
then the highest eigenvalue @xmath42 can be found in the block with the staggered magnetization @xmath45 .
this simplification follows from some symmetries which may occur in the four - spin dmrg vertices . for the non - collinear model
this is not the case , therefore @xmath42 has to be found in the entire hilbert space . for each temperature
, the free energy of the macroscopic system ( [ eq:01 ] ) per site ( @xmath46 ) is related to the maximum eigenvalue @xmath42 of the transfer matrix @xmath47 @xcite , whereas the magnetization and magnetic susceptibility can be obtained by taking the first and second derivatives of the free energy with respect to external magnetic field .
these quantities are calculated in the subspace @xmath48 and @xmath49 which follows from the hamiltonian symmetry .
the correlation length of the @xmath50 components of spin is given by the following ratio @xcite : @xmath51 where @xmath52 is the tensor product of the @xmath53-component of the spin operator and @xmath54 identity matrices of @xmath55 size .
the @xmath56 vector corresponds to the @xmath42 eigenvalue , whereas the eigenvalue @xmath57 is the highest among all whose eigenvector @xmath58 which satisfy the condition @xmath59 .
the convergence of the approximants with respect to @xmath38 depends on temperature and the size of the optimal basis set @xmath41 .
our results are provided for @xmath60 and the trotter number @xmath38 up to @xmath61 ( the lower temperature the higher @xmath38 is needed ) . according to our estimation
the accuracy of results in the whole parameter space considered is not lower than 0.5% @xcite .
recently several compounds of antiferromagnetic one - dimensional systems containing mn(iii ) ions have been reported @xcite . here
the most interesting example @xcite is mn - caf , where the anisotropy axes are defined by jahn - teller elongation of mn(iii ) octahedra and make the angle @xmath62 deg with the chain direction parallel to the @xmath4 axis ( see fig .
[ fig:00 ] ) .
projections of the anisotropy axes onto the plane perpendicular to the chain , alternate along the @xmath63 crystallographic axis making angles @xmath64 deg .
the single - crystal magnetometry measurements @xcite revealed three patterns of the in - field - magnetization profiles . for the @xmath63 direction
the pattern shows linear field dependence and is referred to as antiferromagnetic ( af ) .
the field dependencies measured along the @xmath65 and @xmath4 are referred to as the weak ferromagnetic ( wf ) and the metamagnetic type ( mm ) , respectively .
taking into account the known values of @xmath1 and @xmath2 angles and performing the monte carlo simulations , the values of exchange coupling , uniaxial anisotropy and @xmath66-factor were obtained ( in short the mn - caf parameters ) for the classical counterpart of the model ( [ eq:01 ] ) , @xmath67 fitting the experimental susceptibility and magnetization curves @xcite .
so , the high value of the ratio @xmath68 corresponding to a strong anisotropy occurred . using the classical mn - caf parameters ( [ eq:00 ] ) and the corresponding angles @xmath69 we have determined the in - field magnetization profiles for three crystallographic directions .
their features agree qualitatively with those published previously in fig . 5 for mn - caf @xcite , including the sequence of patterns af , wf and mm for the @xmath63 , @xmath65 , @xmath4 directions .
the @xmath70-type profile for the field along the @xmath4 axis was calculated earlier within dmrg @xcite and its behavior detected some quantitative deviations with respect to the classical counterpart .
-type patterns occur in the @xmath63 , @xmath65 , @xmath4 directions , respectively .
all the patterns are outlined in the inset in the zero - temperature limit . ] relaxing the angles @xmath1 and @xmath2 , we have noticed that the sequence of the magnetization patterns varies with the easy axes orientation and we have established the domains ( fig . [ fig:03 ] ) in the ( @xmath1,@xmath2 ) parameter space . within a given domain the fixed shape of the low - temperature magnetization isotherms as a function of field for all the crystallographic directions occurs .
the expected zero - temperature patterns are plotted in the inset .
the sectors appearing in the diagram are described in table [ tab:01 ] . in sectors
a , b , c the @xmath70-type patterns are unveiled for the field along the @xmath63 , @xmath65 , @xmath4 axes , respectively .
the sequence of the af and wf profiles accompanying the @xmath70-type shape in a given sector is specified in table [ tab:01 ] . in the b
sectors the af pattern is missing and two wf patterns are present in return .
the crystallographic direction , where the magnetization values are higher in the low fields , has been distinguished by an additional index @xmath63 ( or @xmath65 ) .
.pattern sequences in the diagram sectors .
[ cols="^,^,^,^ " , ] the b sectors in the diagram are separated from the other sectors by the dashed line obtained from the solution of the equation @xmath71 .
this line corresponds to the canted angle between the adjacent anisotropy axes which amounts to @xmath72 .
the solid line in the diagram follows from the symmetry of the hamiltonian and can be determined from the solution of the equation @xmath73 .
the diagram in fig .
[ fig:03 ] provides the criterion whether the model ( [ eq:01 ] ) can describe a particular magnetic system .
it predicts the occurrence of the @xmath70-type profile which is accompanied by a given combination of the @xmath74-type and the @xmath75-type patterns .
we immediately see that the criterion is fulfilled for mn - caf @xcite .
its geometric coordinates ( @xmath1 , @xmath2 ) belong to the sector c and the experimental single - crystal profiles agree with the patterns predicted for this sector .
parallel to the @xmath70 axis ( the mn - caf parameters ) .
( bellow ) our estimations of the @xmath76 vs external magnetic field parallel to the @xmath4 axis plotted by triangles .
the solid line represents the linear regression : @xmath77 .
the magnetization isotherms at @xmath78 and @xmath79 k are given by the dashed and chain lines , respectively . ] as the zero - field susceptibility and the correlation energy are not related for the canted antiferromagnetic chains , we have decided to analyse the field dependence of the correlation energy @xmath76 .
first we checked if the correlation length diverges in the limit of low temperatures . for low temperatures the upper panel of fig .
[ fig:04 ] demonstrates a linear dependence of @xmath80 ( where @xmath81 is determined from eq .
( [ eq:02 ] ) ) on the inverse of temperature , choosing the mn - caf parameters ( [ eq:00 ] ) and the corresponding angles @xmath82 and @xmath83 .
this dependence implies that the low - temperature correlation length @xmath9 exhibits the exponential divergence in the @xmath70-direction @xmath84 but the slope @xmath76 is field dependent .
the field dependence of the correlation energy @xmath76 has been studied numerically and plotted by triangles in the bottom part of fig ( [ fig:04 ] ) .
we have revealed the striking linear behavior which can be very well reproduced by the function @xmath85 where the right hand side is expressed in k and the field is given in t. so , the function @xmath76 vanishes , which implies a finite value of correlation length @xmath9 ( see eq .
( [ eq:03 ] ) ) , for the critical field @xmath86 t , what is close to the value estimated directly from the dmrg correlation length , @xmath87 t. the abrupt change of the correlation length from the infinite to the finite value induced by the field should be accompanied by some changes in the physical properties .
we demonstrate in the bottom panel of fig .
[ fig:04 ] that the @xmath70-type profile displays a rapid jump in the vicinity of @xmath3 t from @xmath88 to a nearly saturation value .
we interpret this behavior as a field - driven transition from the antiferromagnetic to the ferromagnetic order along the @xmath4 direction .
such critical field can be determined from the intersection of the low - temperature magnetization isotherms , from their inflection point or from the maximum of the in - field magnetic susceptibility .
now we analyze the correlation energy @xmath11 for the mn - caf parameters in the absence of a magnetic field . the formula ( [ eq : f ] ) implies @xmath89 k , while the dmrg calculations based on eqs .
( [ eq : ksi ] ) and ( [ eq:02 ] ) yield @xmath90 k. these estimates are in remarkable agreement with the classical prediction @xcite @xmath91 k following from eq .
( [ eq : deltaksi ] ) .
moreover , the coincidence of the correlation energy of the quantum model ( [ eq:01 ] ) and the classical dependence ( [ eq : deltaksi ] ) is confirmed and demonstrated for a number of couplings @xmath18 and the canting angles @xmath25 in fig .
[ fig:05 ] . and @xmath92 @xmath93 plotted by the full and the dashed line , respectively , and the ratios @xmath94 and @xmath95 illustrated by the circles and squares .
the open symbols correspond to @xmath96 k and the full symbols to @xmath97 k. ] in fig .
[ fig:05 ] the upper continuous and lower dashed lines represent the functions 2s@xmath98 and s@xmath98 , respectively .
the open and full circles illustrate the ratios @xmath94 calculated numerically within dmrg for @xmath99=1.26 k and @xmath99=1.36 k , respectively . as to the angles @xmath25 , their values follow from the angles @xmath1 and @xmath2 chosen from different sectors of the diagram in fig .
[ fig:03 ] .
the fields were applied in the proper direction for a sector selected .
we conclude that the agreement between the classical predictions 2s@xmath98 for the ratio @xmath94 and those calculated for our quantum model ( [ eq:01 ] ) and plotted in fig .
[ fig:05 ] is very good .
there are some deviation for small and high values of @xmath25 , but they can be attributed to uncertainties of the extrapolations performed for ln(@xmath9 ) yielding the values @xmath11 .
these uncertainties are of the order of a few percent and are much higher than the accuracy of the dmrg results .
they could be diminished lowering temperature and increasing the cost of our simulations .
concluding , our numerical results in fig .
[ fig:05 ] give a strong evidence that the correlation energy of the canted antiferromagnetic quantum chains can be expressed by the relation ( [ eq : deltaksi ] ) previously found for the classical systems . in fig .
[ fig:05 ] we also plot by the full squares the ratios @xmath95 as a function of @xmath25 , where @xmath100 is the value of the @xmath70-type magnetization profile corresponding to the field @xmath3 .
we emphasise that the ratio is completely determined by the coordinates of the inflection point of the @xmath70-type magnetization profile .
the ratios @xmath95 calculated coincide with the values of the function s@xmath98 drawn by the dashed line and are lower by a factor of 2 than the ratios @xmath94 , implying that @xmath101 we can interpret the result ( [ eq : mh ] ) that the correlation energy is equal to the average zeeman energy per pair of spins calculated for the critical field @xmath3 . the relation ( [ eq : mh ] ) is an important result of our paper because it provides a direct estimate of the correlation energy @xmath11 in terms of the special coordinates of the @xmath70-type magnetization profile .
moreover , the exchange constant @xmath99 can be also obtained from the @xmath70-type single - crystal low - temperature magnetization isotherm .
combining eqs .
( [ eq : deltaksi ] ) and ( [ eq : mh ] ) , we find the formula for the magnetic coupling @xmath102 to validate our conclusions , in fig . [ fig:02b ] the @xmath70-type behavior is drawn for the ratios @xmath103 and @xmath104 in both panels .
for the angles specified in the legends and coming from the sectors a and b , the a and b panels display the profiles found for the magnetic field parallel to the @xmath63 axis and @xmath65 axis , respectively .
the slope of all the @xmath70-type curves in fig .
[ fig:02b ] increases with decreasing temperature and the shape does not change .
the coordinates of the intersections of the corresponding magnetization isotherms in fig .
[ fig:02b ] ( @xmath105 k and @xmath79 k ) inserted into eq .
( [ eq : j ] ) yield the consistent estimates of @xmath99 with the uncertainty equal to @xmath106 k for the input values @xmath96 k and @xmath107 k. when we used lower temperatures ( @xmath108 k and @xmath109 k ) , the uncertainty declined to @xmath110 k for the same @xmath99 , which is the expected trend .
k and the thin lines to @xmath105 k. the solid and the dashed lines belong to the @xmath111 sector ( @xmath112 deg , @xmath113 deg ) and to the @xmath114 sector ( @xmath115 deg , @xmath116 deg ) , respectively . ]
now , taking the values @xmath117 deg , @xmath118 t and @xmath119 ( in @xmath120 ) directly from the experiment for mn - caf compound @xcite , we can evaluate from eqs .
( [ eq : mh ] ) and ( [ eq : j ] ) the value of correlation energy @xmath121 k and the coupling constant @xmath6 k , respectively .
both the values are slightly lower than the @xmath122 k and @xmath123 k reported earlier @xcite .
however , this fact can be understood as the experimental value @xmath123 k was settled on the basis of the classical model and the fitting procedure which usually overestimate the coupling constant @xcite .
if the coupling @xmath18 is overestimated then the relation ( [ eq : deltaksi ] ) may proliferate the inaccuracy and imply somewhat higher @xmath11 .
our investigation has confirmed that the relative positions of the anisotropy axes affect the shape of magnetization curves along different crystallographic directions . the resulting angle - dependent diagram predicts four areas , each characterized by a specific sequence of the in - field magnetization patterns .
if for a compound studied the shapes of magnetization curves in the low - temperature range are consistent with those established in the diagram for a given geometry , then the compound can be analyzed in terms of the quantum heisenberg model with a strong anisotropy
. for the canted antiferromagnet scm the values of correlation energy @xmath76 decrease linearly with the applied field and vanish at @xmath3 which coincides with the intersection point of the @xmath70-type magnetization profiles .
the coordinates @xmath3 and @xmath100 of this special point determine the zero - field correlation energy @xmath11 which in turn reveals the classical dependence on the canting angle @xmath25 and the coupling @xmath18 .
these findings relate the @xmath18 value to the coordinates @xmath3 and @xmath100 and yield a simple recipe to the evaluation of @xmath18 . following these ideas ,
we have estimated for the mn - caf compound the relevant quantities at @xmath121 k and @xmath124 k which agree very well with the values reported earlier .
our calculations provide some insight into the common claim that in the strong anisotropy limit the heisenberg model can be substituted by the ising model .
we confirm that the projections of spins onto the axis displaying the @xmath70-type pattern can be treated as classical ising - type variables .
then the rapid jump in the magnetization profile can be related to formation of an ising domain wall .
anyway , it is striking that the quantity @xmath11 defined in the absence of field can be so easily settled from the experiment performed in the presence of field .
we are grateful to prof .
r. sessoli for many valuable discussions explaining the structure and properties of the compound considered here .
a. b. and g. k. thank to the polish ministry of science and higher education for supporting this work under the iuventus plus program ( ip2010 001170 ) and grant nr 230137 , respectively .
numerical calculations were performed in psnc pozna ( poland ) and wcss wrocaw ( poland , grant 82 ) .
this work was also granted access to the hpc resources within deci by the prace-2ip under grant no ri-283493 .
dante gatteschi , roberta sessoli , and jacques villain , _ molecular nanomagnets _ , oxford university press , 2006 .
a. caneschi , d. gatteschi , n. lalioti , c. sangregorio , r. sessoli , g. venturi , a. vindigni , a. rettori , m. g. pini , m. a. novak , angew .
int . ed . * 40 * , 1760 ( 2001 ) .
r. clerac , h. miyasaka , m. yamashita , c. coulon , j. am . chem .
124 * , 12837 ( 2002 ) . h. miyasaka , t. madanbashi , k. sugimoto , y. nakazawa , w. wernsdorfer , k. sugiura , m. yamashita , c. coulon , r. clerac , chem .
j. * 12 * , 7028 ( 2006 ) . c. coulon , h. miyasaka , and r. clerac , struct
. bond . * 122 * , 163 ( 2006 ) ; h. miyasaka and m. yamashita , dalton trans .
, 399 ( 2007 ) , and references therein .
n. b. ivanov , j. richter , and u. schollwoeck , phys .
b * 58 * , 14456 ( 2007 ) .
zheng , w. xue , w .- x .
zhang , m .-
tong , x .-
chen , f. grandjean , g. j. long , s .- w .
ng , p. panissod , and m. drillon , inorg . chem . * 48 * , 2028 ( 2009 ) .
r. j. glauber , j. math . phys . * 4 * , 294 ( 1963 ) .
h. miyasaka , m. julve , m. yamashita , and r. clerac , inorg . chem . * 48 * , 3420 ( 2009 ) .
m. g. pini and a. rettori , phys .
b * 76 * , 064407 ( 2007 ) .
k. bernot , j. luzon , r. sessoli , a. vindigni , j. thion , s. richeter , d. leclercq , j. larionova , and a. van der lee , j. am .
soc . * 130 * , 1619 ( 2008 ) .
b. barbara , j. phys . * 34 * , 1039 ( 1973 ) . j. m. loveluck , s. w. lovesey , s. j. aubry , phys .
c : solid state phys .
* 8 * , 3841 ( 1975 ) . c. coulon , r. clerac , l. lecren , w. wernsdorfer , and h. miyasaka phys .
b * 69 * , 132408 ( 2004 ) .
a. saitoh , h. miyasaka , m. yamashita , and r. clerac , j. mater . chem . * 17 * , 2002 ( 2007 ) . j. boeckmann , m. wriedt , c.nther , chem .
j. * 18 * , 5284 ( 2012 ) .
y - z . zheng , w. xue , m - l .
tong , x - m .
chen , f. grandjean , g.j .
long , inorg . chem . *
47 * , 4077 ( 2008 ) jung hee yoon , dae won ryu , hyoung chan kim , sung won yoon , byoung jin suh , and chang seop hong , chem .
j. * 15 * , 3661 ( 2009 ) .
a. barasiski , g. kamieniarz , and a. drzewiski , comp .
. comm . *
182 * , 2013 ( 2011 ) .
p. sobczak , a. barasiski , g. kamieniarz , a. drzewiski , phys . rev . * b 84 * , 224431 ( 2011 ) .
g. kamieniarz , r. matysiak , a. c. dauria , f. esposito , and u. esposito , phys . rev . *
b 56 * , 645 ( 1997 ) .
a. caramico dauria , f. esposito , u. esposito , d. gatteschi , g. kamieniarz , and s. wacerz , j. chem . phys . *
109 * , 1613 ( 1998 ) .
g. kamieniarz , m. bieliski , and j .-
renard , phys . rev . *
b 60 * , 14521 ( 1999 ) .
r. matysiak , g. kamieniarz , p. gegenwart , and a. ochiai , phys . rev . *
b 79 * , 224413 ( 2009 ) .
m. suzuki , prog .
. phys . * 56 * , 1454 ( 1976 ) .
t. delica and h. leschke , physica * a 176 * , 736 ( 1990 ) .
x. q. wang and t. xiang , phys
* b 56 * , 5061 ( 1997 ) .
u. schollwoeck , rev . mod
* 77 * , 259 ( 2005 ) ; k. hallberg , adv .
phys . * 55 * , 477 ( 2006 ) ; u. schollwoeck , ann .
( ny ) * 326 * , 96 ( 2011 ) .
a. macioek , a. drzewiski , and r. evans , phys .
b 64 * , 056137 ( 2001 ) .
a. drzewiski , a. macioek , a. barasiski , and s. dietrich , phys . rev . * e 79 * , 041145 ( 2009 ) .
n. shibata , j. phys .
a : math .
gen . * 36 * , r381 ( 2003 ) .
a. barasiski , p. sobczak , a. drzewiski , g. kamieniarz , a. bieko , j. mroziski , and d. gatteschi , polyhedron * 29 * , 1485 ( 2010 ) .
r. j. baxter , _ exactly solved models in statistical mechanics _ , academic press 1982 .
n. hatano and m. suzuki , j. phys .
j. * 62 * , 1346 ( 1993 ) .
g. kamieniarz , p. kozlowski , m. antkowiak , p. sobczak , t. lusarski , d. m. tomecka , a. barasiski , b. brzostowski , a. drzewiski , a. bieko , j. mroziski , acta phys . pol . * a 121 * , 992 ( 2012 ) . | we demonstrate numerically that for the strongly anisotropic homometallic @xmath0 canted single - chain magnet described by the quantum antiferromagnetic heisenberg model the correlation energy and exchange coupling constant can be directly estimated from the in - field - magnetization profile found along the properly selected crystallographic direction . in the parameter space defined by the spherical angles ( @xmath1 , @xmath2 ) determining the axes orientation ,
four regions are identified with different sequences of the characteristic field - dependent magnetization profiles representing the antiferromagnetic , metamagnetic and weak ferromagnetic type behavior .
these sequences provide a criterion for the applicability of the anisotropic quantum heisenberg model to a given experimental system .
our analysis shows that the correlation energy decreases linearly with field and vanishes for a given value @xmath3 which defines a special coordinates in the metamagnetic profile relevant for the zero - field correlation energy and magnetic coupling . for the single - chain magnet formed by the strongly anisotropic manganese(iii ) acetate meso - tetraphenylporphyrin complexes coupled to the phenylphosphinate ligands , the experimental metamagnetic - type magnetization curve in the @xmath4 direction yields an accurate estimate of the values of correlation energy @xmath5 k and exchange coupling @xmath6 k. |
the invariant mass spectrum of dielectrons measured during the 2006 rhic run , after combinatorial background subtraction ( figure [ invmass_fit ] ) reveals peaks corresponding to the , and @xmath0(1s,2s,3s ) quarkonia states .
the continuum spectrum is well described by correlated heavy flavour ( @xmath2 and @xmath3 ) and drell - yan contributions estimated using a pythia simulation @xcite , including the detector acceptance and efficiencies .
the differential cross section shown in figure [ jpsi_psip ] is the first @xmath4-dependent result for an excited charmonium state at rhic .
the to ratio is also shown and compared to results from other experimental facilities @xcite .
one can note the stability of this ratio for a vast collision energy and .
the up - to - date feed - down contributions to from is @xmath5 and from @xmath6 is @xmath7 ( 90% cl ) .
these numbers are in agreement with world average results @xcite .
the new rapidity dependence result ( figure [ jpsi_yield ] ) is in agreement with that already published @xcite .
the uncertainties are dominated by systematics .
both mid and forward rapidity yields are well described by the s - channel cut color singlet model ( csm ) @xcite using two parameters fitted to cdf data at @xmath81.8 tev @xcite over @xmath9 .
phenix started a series of studies of the dependence of the decay angular distribution of the .
the polarization parameter @xmath10 is defined by , where @xmath11 is the angle between the positive lepton and the momentum direction ( helicity frame ) .
initial inclusive ( direct + feed - down ) polarization _ versus _ is shown in figure [ jpsi_pol ] .
although the s - channel cut csm only considers direct ( no feed - down ) @xcite , it is in good agreement with the mid - rapidity results .
however , the @xmath12 result obtained at forward rapidity is @xmath13 away from the model prediction at that rapidity .
upcoming dependent @xmath10 measurements for the forward rapidity will be important to clarify whether the s - channel cut csm describes the data at forward rapidity or not . given the current uncertainties , one can not rule out the possibility of zero polarization , as expected by the color evaporation model ( cem ) @xcite .
yield.,scaledwidth=100.0% ] yield.,title="fig:",scaledwidth=90.0% ] yield.,title="fig:",scaledwidth=90.0% ] we counted 12 unlike - sign and one like - sign dielectron pairs in the @xmath0(1s+2s+3s ) mass range @xmath14~ gev / c^2 $ ] at mid - rapidity .
a preliminary study of possible physical contributions in this region indicates correlated open bottom and drell - yan can contribute up to 15% of the number of net counts and is accounted for in the systematic errors .
the corresponding @xmath0 family cross section in the dielectron channel is this result follows the world trend and estimates based on the cem ( figure [ upsilon_yield ] ) @xcite .
the rapidity dependence of the @xmath0 cross section , adding the preliminary results released at qm06 @xcite , can be used to test production models for bottomonium and estimate the total cross section . given the smaller relativistic contributions , nrqcd calculations for @xmath0 at rhic are supposed to be more reliable than for charmonium .
the new mid - rapidity measurement was also used as a baseline for the first measurement of the @xmath0 nuclear modification factor in au+au collisions @xcite .
the fits performed to the published nuclear modification factors in @xmath1+au collisions ( @xmath15 ) have been revisited taking into account all systematic uncertainties @xcite .
the new fits return a larger uncertainty in the breakup cross sections .
cold nuclear matter `` extrinsic '' calculations , where one treats the production as a 2@xmath162 process ( @xmath17 ) , have shown considerable differences in the breakup cross section compared to those obtained using the standard procedure , using `` intrinsic '' kinematics with 2@xmath161 processes @xcite .
the integrated luminosity accumulated in @xmath1+au collisions during the 2008 run was about thirty times larger than that used in @xcite .
the first nuclear modification factor measurements from the 2008 run were performed using as a reference the yield in the 60 - 80% centrality range , many systematic uncertainties cancel out in a @xmath19 measurements .
figure [ rcp ] shows the rapidity dependence of @xmath19 for three centrality ranges .
the negative rapidity region , which represents large @xmath20 ( in the anti - shadowing region ) of the gluon distribution , shows no nuclear effects within uncertainties , but a considerable suppression is observed at central rapidity and at forward rapidity .
one observes that the strong suppression at forward rapidity and the lack of suppression at backward rapidity can not be described by the existing nuclear modified gluon distributions ( eks98 shown in figure [ rcp ] , eps08 and ndsg ) for any fixed breakup cross section @xcite .
this observation suggests that additional cnm effects besides the standard shadowing and hadronic interactions must be important . ) and comparison with that from eks98 parton modification distributions for different breakup cross sections@xcite.,title="fig:",scaledwidth=40.0% ] ) and comparison with that from eks98 parton modification distributions for different breakup cross sections@xcite.,title="fig:",scaledwidth=40.0% ] | we report new quarkonia measurements necessary to understand production mechanisms and cold nuclear matter effects in the yields observed at rhic energy .
results obtained in collisions collected during the 2006 rhic run include , and @xmath0 differential cross sections as well as polarization .
revisited interpretations of the published nuclear modification factors and statistically improved observations in @xmath1+au collisions taken in the 2008 run are also discussed in the view of the recent understanding of the initial state effects and breakup cross section . |
the tevatron accelerator complex produces @xmath4 collisions with a center of mass energy of @xmath5 , and the cdf and d0 experiments have so far recorded @xmath6 fb@xmath7 of collision data .
just a fraction of this data allowed establishing a successful experimental programme of physics of the top quark and of the electroweak sector . in the following , we summarise the main measurements and report on preliminary results based on the latest data analyses .
top quarks are produced and observed at the tevatron in their pair and singlet modes .
measurements of the cross section allow testing of the production and decay mechanisms of the top and its coupling in the standard model ( sm ) .
the pair production of top and anti - top in @xmath4 collisions has been measured with increasing precision since its discovery in 1995 @xcite at the tevatron .
figure [ fig : d0_top_pair_xsec ] shows the most recent measurements using data collected by the d0 experiment @xcite in various decay channels . as top quarks decay almost exclusively to @xmath8 , the final states of @xmath0 events are labelled as _ dilepton _ , _ lepton plus jets _ and _ all - hadronic _ depending on whether a leptonic decay has occurred in both , one only or none of the two final - state @xmath1 bosons respectively .
all results are in agreement with each other and in agreement with the next - to - leading order ( nlo ) calculation for @xmath0 production in @xmath4 collisions at @xmath9 tev of [email protected] pb @xcite .
preliminary measurements are also shown in figure [ fig : d0_top_pair_xsec ] , performed with 4.35.3 fb@xmath7 of collision data , and yield individually a precision of @xmath11% , similar to the theoretical uncertainty . since a large component ( about 6% ) of the experimental uncertainty on the @xmath0 cross section
is due to the uncertainty on the luminosity determination , higher precision can be obtained by measuring the ratio of @xmath0 to @xmath2-boson cross sections @xmath12 @xcite and using the theoretical @xmath2 boson cross section calculation as input .
with such an approach , the uncertainty on the luminosity largely cancels out and the most accurate preliminary determination of the @xmath0 cross section , combining samples in all decay channels seen with the cdf ii detector @xcite yields : @xmath13 pb ( for a top mass of @xmath14 ) , in excellent agreement with nlo calculations and with an experimental uncertainty of only @xmath15% . the electroweak production of single top quarks has been established at the tevatron in 2009 @xcite .
a preliminary measurement using 4.8 fb@xmath7 of collision data recorded with the d0 detector has now been completed in the @xmath16+jets decay channel , measuring @xmath17 pb . figure [ fig : singletop ] shows the current best measurements of single top cross section ( @xmath18-channel and @xmath19-channel combined ) and the preliminary tevatron combined cross section of @xmath20 pb , in agreement with the nlo theoretical prediction of @xmath21 pb @xcite ( for @xmath22 ) .
this also allows extraction of the cabibbo - kobayashi - maskawa matrix element @xmath23 with a 95% confidence level ( c.l . )
lower limit of @xmath24 .
measurements of single top cross section performed with data collected by the cdf and d0 collaborations . also indicated the preliminary tevatron combination . ]
measurements of single top cross section performed with data collected by the cdf and d0 collaborations . also indicated the preliminary tevatron combination . ]
the mass of the top quark is of interest both because the top is the heaviest known fundamental particle and because a precise measurement helps constraining the mass of the higgs boson .
we present two preliminary measurements using events in the dilepton and lepton plus jets final states reconstructed with the cdf ii detector .
the analyses were based respectively on 4.8 fb@xmath7 and 5.6 fb@xmath7 of collision data and yield experimental top mass values of @xmath25 gev/@xmath26 and @xmath27 gev/@xmath26 .
figure [ fig : topmass ] shows the cdf combination of top mass measurements , @xmath28 gev/@xmath26 , which with an accuracy of @xmath29 is more precise than the previous world average @xcite .
moreover , measurements derived using all the @xmath0 decay channels are consistent with each other , and a preliminary determination of the top and anti - top mass difference using 5.6 fb@xmath7 of collision data recorded with the cdf ii detector gives @xmath30 gev/@xmath26 .
taking into account the preliminary results quoted above , the new tevatron combined experimental top mass determination is @xmath31 gev/@xmath26 @xcite .
projections indicate that the tevatron experiments should reach individually a precision of @xmath100.81.0 gev/@xmath26 with @xmath32 fb@xmath7 of data . the experimental study of physics of top quarks at the tevatron includes the determination of a variety of properties in addition to the production cross section and the top quark mass .
measurement of properties such as the production mechanism , the forward - backward asymmetry and differential cross sections for example are only recently approaching the accuracy needed to test the sm predictions .
we present two preliminary results on the width of the top quark . in the first , using 4.3 fb@xmath7 of tevatron s @xmath4 collision data collected by the cdf ii detector , we exploit the reconstructed top quark mass to set an upper limit on the top quark width ( @xmath33 ) of @xmath34 at 95% c.l .
we also set a central limit of @xmath35 at 68% c.l . in the second result
, we extract the total width of the top quark from the partial decay width @xmath36 and the branching fraction @xmath37 .
@xmath36 is obtained from the measured @xmath19-channel cross section for single top quark production in 2.3 fb@xmath7 of @xmath4 data from the d0 collaboration , and @xmath37 is extracted from a measurement of the ratio @xmath38 in @xmath0 events in 1 fb@xmath7 of integrated luminosity .
assuming @xmath39 , where @xmath40 includes any kinematically accessible quark , the result is : @xmath41 which corresponds to a top quark lifetime of @xmath42 s. a preliminary measurement of the top quark charge , based on 2.7 fb@xmath7 of collision data collected with the cdf ii detector and using soft electron and muon @xmath43-tagging , now excludes a top quark charge of @xmath44 with 95% c.l , extending the significance of a previous result .
measurements of the dynamics of production and decay of the top quark are of importance because they are sensitive to new production and decay mechanisms .
for example , sm top pair production induces a characteristic spin correlation which can be modified by new production mechanisms such as @xmath45 bosons or kaluza - klein gluons @xcite .
we report on a preliminary measurement of @xmath0 helicity fractions and spin correlation in 5.3 fb@xmath7 of reconstructed lepton+jets data with the cdf ii detector . choosing the direction of the top quark momentum as the spin quantization axis ( the _ helicity basis _ )
, we find a spin correlation coefficient @xmath46 , in agreement with the nlo expectation of @xmath47 @xcite .
we also report a preliminary measurement of the forward - backward asymmetry of pair produced top quarks in 5.3 fb@xmath7 of data collected with the cdf ii detector . a small asymmetry of [email protected] in the laboratory rest frame
is expected in quantum chromo dynamics at nlo @xcite and large positive asymmetries with large uncertainties were measured at cdf and d0 @xcite .
we find @xmath48 , in agreement with the sm .
comparison of the @xmath1 boson charge asymmetry measured by the cdf and d0 collaborations as a function of lepton @xmath49 , for lepton s transverse energies @xmath50 .
the data are compared to a simulation based on resbos with cteq6.6 pdfs . ]
comparison of the @xmath1 boson charge asymmetry measured by the cdf and d0 collaborations as a function of lepton @xmath49 , for lepton s transverse energies @xmath50 .
the data are compared to a simulation based on resbos with cteq6.6 pdfs . ]
a precise measurement of the @xmath1 boson mass ( @xmath51 ) provides a measurement of quantum loop corrections , and in the sm these corrections to @xmath51 are contributed predominantly by the top quark and the higgs boson loops .
therefore , @xmath51 in conjunction with the top quark mass constrains the mass @xmath52 of the undiscovered higgs boson .
the most recent combination of @xmath51 measurements at the tevatron is reported in detail elsewhere @xcite , and when combined with measurements at the lep collider yields the value @xmath53 @xcite .
overall , precision electroweak measurements , including the preliminary value presented in this article of @xmath54 , indicate the preferred higgs boson mass of @xmath55 , or @xmath56 at 95% c.l .
in addition to the @xmath1 boson mass , the decay width of the @xmath1 is also of importance towards a stringent test of the sm and can be used to constrain other sm parameters , such as the @xmath57 cabibbo - kobayashi - maskawa matrix element .
a recent combination of the cdf and d0 run i and run ii measurements , which are based on fits to tails of the @xmath1 transverse mass @xcite distributions , yields @xmath58 @xcite , and a preliminary world average including both the tevatron and lep2 gives @xmath59 .
the production of @xmath1 bosons at the tevatron is subject to a well known charge asymmetry , due to the different fraction of the proton momentum carried on average by @xmath60 and @xmath61 quarks : @xmath62 where @xmath63 is the @xmath1 boson rapidity @xcite and @xmath64 is the differential cross section for @xmath65 or @xmath66 boson production .
previous measurements @xcite have been published by the d0 and cdf collaborations , based on candidate @xmath67 events in 0.75 fb@xmath7 and 1 fb@xmath7 of collision data respectively , in which comparisons to monte carlo ( mc ) simulations of @xmath1 production and decay showed generally good agreement .
we present a direct comparison of these analyses with an additional preliminary measurement based on candidate @xmath68 events in 4.9 fb@xmath7 of collision data collected with the d0 detector .
the data are compared to mc event simulation from resbos @xcite with cteq6.6 @xcite parton distribution functions ( pdfs ) , as a function of the lepton pseudo - rapidity @xmath69 , in different ranges of the lepton s transverse energy ( @xmath70 ) . while in the low @xmath70 the data are well reproduced by the mc , for @xmath50 discrepancies
start to emerge and the mc is now seen to overestimate the asymmetry at large @xmath49 ( see figure [ fig : wasymm ] ) with respect to both cdf and d0 experimental data .
the presence of both vector and axial - vector couplings in @xmath71 gives rise to an asymmetry also for @xmath2 bosons in the polar angle of the negatively charged lepton momentum relative to the incoming quark direction in the rest frame of the lepton pair .
we present a preliminary measurement of such forward - backward asymmetry in the kinematic region where the invariant mass ( @xmath72 ) of the lepton pair is between 50 and 600 gev/@xmath26 .
the data sample consists of 4.1 fb@xmath7 of @xmath4 collisions recorded by the cdf ii detector .
figure [ fig : zfb ] shows the data compared with pythia v6.216 @xcite , indicating very good agreement across the entire @xmath72 spectrum .
the forward - backward asymmetry is sensitive to @xmath73 , which is an effective parameter that includes higher order corrections .
the current world average value of @xmath73 is [email protected] @xcite , the d0 collaboration measured @xmath74 from an earlier @xmath2 forward - backward asymmetry measurement @xcite based on 1.1 fb@xmath7 of collision data and we plan to soon report @xmath73 from the cdf preliminary measurement presented here .
invariant mass for the @xmath2 candidates reconstructed in the 4 @xmath75 candidate events . ]
invariant mass for the @xmath2 candidates reconstructed in the 4 @xmath75 candidate events . ]
the measurement of heavy diboson production in @xmath4 collisions is a way of probing the structure of the electroweak sector , as the presence of anomalous triple gauge bosons couplings for instance could indicate new physics at a higher mass scale .
moreover , as these processes are backgrounds to some higgs boson decay signatures , their measurement provides milestones on the way to sensitivity to a higgs boson production measurement .
heavy diboson production has been established at the tevatron in several final states using 13 fb@xmath7 @xcite . as the collision data available increases , the precision of these measurements is being progressively improved .
we report on a preliminary measurement of @xmath77 and @xmath75 production in three- and four - lepton channels , using @xmath78 fb@xmath7 of data collected by the cdf experiment . from 50 @xmath79 candidate events and 4 @xmath80 candidate events ,
we measure @xmath81 and @xmath82 .
the systematic uncertainties are minimised by measuring the experimental ratio to the inclusive @xmath83 cross section and using the theoretical value of @xmath84 production cross section .
figure [ fig : zzmass ] shows the distribution of invariant mass for the @xmath2 candidates reconstructed in the 4 @xmath75 candidate events .
the measurements are in agreement with the lo theoretical prediction of @xmath85 pb and @xmath86 pb , and in agreement with earlier measurements at the tevatron @xcite .
we also report on a preliminary measurement of combined @xmath87 production cross section in the channel with an identified electron or muon , large missing transverse energy and two jets in 4.6 fb@xmath7 of cdf ii data .
the analysis employs a matrix element technique which calculates event probability densities for signal and background hypotheses .
the probabilities are used to form a discriminant variable which is evaluated for signal and background events .
the resulting template distributions are fit to the data using a binned likelihood approach , and yield a cross section @xmath88 .
the measurement increases the precision of previous determinations @xcite and is in agreement with the theory prediction of [email protected] pb @xcite .
we have presented recent preliminary measurements at tevatron of top pair and single production cross section , top quark mass and width , top pair spin correlations and forward - backward asymmetry . in the electroweak sector , we have presented a new tevatron average of the @xmath1 boson width , and preliminary measurements of the @xmath1 and @xmath2 forward - backward asymmetries and @xmath89 diboson production cross sections .
all measurements are based on larger amount of collision data than previously used and are in agreement with the sm .
9 f. abe _ et al . _
( cdf collaboration ) , _ phys .
lett . _ * 74 * 2626 ( 1995 ) .
s. abachi _
( d0 collaboration ) , _ phys .
lett . _ * 72 * 2632 ( 1995 ) .
v.m . abazov _
( d0 collaboration ) , _ nucl .
methods phys .
res . _ a * 565 * 463 ( 2006 ) .
m. cacciari _
_ , _ journal of high energy physics _ * 09 * 127 ( 2008 ) .
n. kidonakis and r. vogt , _ phys .
_ d * 78 * 074005 ( 2008 ) .
u. langenfeld , s. moch and p. uwer , _ phys .
_ d * 80 * 054009 ( 2009 ) .
the cdf ii detector technical design report no .
fermilab - pub-96/390-e ; d. acosta _ et al .
_ , ( cdf collaboration ) _ phys . rev .
_ d * 71 * , 052003 ( 2005 ) .
t. aaltonen _
( cdf collaboration ) , _ phys .
lett . _ * 105 * 012001 ( 2010 ) .
t. aaltonen _
( cdf collaboration ) , _ phys .
_ * 103 * 092009 ( 2009 ) .
( d0 collaboration ) , _ phys .
lett . _ * 103 * 092001 ( 2009 ) .
harris , e. laenen , l. phaf , z. sullivan , and s. weinzier .
_ d * 66 * 054024 ( 2002 ) .
n. kidonakis , _ phys .
_ d * 74 * 114012 ( 2006 ) .
j. campbell , r.k .
ellis , f. tramontano , _ phys .
_ d * 70 * 094012 ( 2004 ) .
the tevatron electroweak working group , fermilab - tm-2427-e , arxiv:0903.2503v1 [ hep - ex ] ( 2009 ) .
the tevatron electroweak working group , fermilab - tm-2466-e , arxiv:1007.3178v1 [ hep - ex ] ( 2009 ) .
m. arai , n. okada , k. smolek , v. simak , _ phys .
_ d * 75 * 095008 ( 2007 ) .
w. bernreuther , a. brandenburg , z.g .
si , p. uwer , cern - ph - th/2004 - 206 , arxiv:0410197 [ hep - ph ] ( 2004 ) .
o. antunano , j.h .
kuhn and g. rodrigo , _ phys .
_ d * 77 * 014003 ( 2008 ) .
t. aaltonen _
et al . _ , ( cdf collaboration ) _ phys .
* 101 * 202001 ( 2008 ) .
v.m . abazov _
et al . _ , ( d0 collaboration ) _ phys . rev
. lett . _ * 100 * 142002 ( 2008 ) .
the tevatron electroweak working group , fermilab - tm-2439-e , arxiv:0908.1374v1 [ hep - ex ] ( 2009 ) .
the lep and tevatron electroweak working groups , cern - ph - ep/2009 - 023 , arxiv:0911.2604v2 [ hep - ex ] ( 2009 ) .
the @xmath1 transverse mass is defined as @xmath90 , where @xmath91 and @xmath92 are the measured transverse momentum of the charged lepton and the transverse momentum of the neutrino respectively .
the tevatron electroweak working group , fermilab - tm-2460-e , arxiv:1003.2826v1 [ hep - ex ] ( 2010 ) .
we assume ( @xmath93 ) coordinate system with the @xmath94 axis in the direction of the proton beam ; @xmath95 and @xmath96 are the azimuthal and polar angle , respectively .
the rapidity ( @xmath63 ) and the pseudorapidity ( @xmath69 ) are defined as @xmath97 and @xmath98 .
v.m . abazov _
( d0 collaboration ) _ phys .
lett . _ * 101 * 211801 ( 2008 ) .
t. aaltonen _
( cdf collaboration ) _ phys .
lett . _ * 102 * 181801 ( 2009 ) . c. balazs and c.p .
yuan , _ phys .
_ d * 56 * , 5558 ( 1997 ) .
p.m. nadolsky _ et al .
d * 78 * 013004 ( 2008 ) .
t. sjostrand _
_ , _ comput .
commun . _ * 135 * , 328 ( 2001 ) .
k. nakamura _
( particle data group ) , _ j. phys .
_ g * 37 * , 075021 ( 2010 ) .
v.m . abazov _
( d0 collaboration ) _ phys .
* 101 * 191801 ( 2008 ) .
( d0 collaboration ) _ phys .
* 94 * 151801 ( 2005 ) ; v.m .
( d0 collaboration ) _ phys .
_ d * 76 * 111104 ( 2007 ) ; v.m .
( d0 collaboration ) _ phys .
* 101 * 171803 ( 2008 ) .
d.e . acosta _
et al _ ( cdf collaboration ) _ phys .
lett . _ * 94 * 211801 ( 2005 ) ; a. abulencia _
( cdf collaboration ) _ phys .
lett . _ * 98 * 161801 ( 2007 ) ; t. aaltonen _
( cdf collaboration ) _ phys .
* 100 * 201801 ( 2008 ) .
t. aaltonen _
( cdf collaboration ) _ phys .
* 98 * 161801 ( 2007 ) .
v.m . abazov _
( d0 collaboration ) arxiv:1006.0761v1 [ hep - ph ] ( 2010 ) .
t. aaltonen _
( cdf collaboration ) _ phys .
_ * 104 * 101801 ( 2010 ) .
v.m . abazov _
( d0 collaboration ) _ phys .
lett . _ * 102 * 161801 ( 2009 ) .
j. m. campbell and r. k. ellis , _ phys .
_ d * 60 * , 113006 ( 1999 ) . | we present recent preliminary measurements at the tevatron of @xmath0 and single top production cross section , top quark mass and width , top pair spin correlations and forward - backward asymmetry . in the electroweak sector , we present the tevatron average of the @xmath1 boson width , and preliminary measurements of the @xmath1 and @xmath2 forward - backward asymmetries and @xmath3 diboson production cross sections .
all measurements are based on larger amount of collision data than previously used and are in agreement with the standard model . |
the study of the five - dimensional einstein - maxwell theory with a chern - simon term plays a important role in elucidating the structure of the string theory since it is the bosonic sector of the minimal supergravity .
all bosonic supersymmetric solutions of minimal supergravity in five dimensions are classified by gauntlett _
et al _ @xcite . many interesting supersymmetric ( bps ) solutions to the five - dimensional einstein - maxwell equations ( with a chern - simon term ) have been found by a lot of authors . following on this classification of the five - dimensional supersymmetric solutions ,
they have been constructed on hyper - khler base spaces .
the first asymptotically flat supersymmetric black hole solution , breckenridge - myers - peet - vafa ( bmpv ) solution , was constructed on the four - dimensional euclid space @xcite . a supersymmetric black hole solution with a compactified extra dimension on the euclidean self - dual taub - nut base space
was constructed by gaiotto _
it was extended to multi - black hole solution with the same asymptotic structure @xcite .
one of the most interesting properties is that the possible spatial topology of the horizon of each black hole is the lens space @xmath1 ( @xmath2:the natural numbers ) in addition to @xmath3 .
similarly , black hole solution on the eguchi - hanson space @xcite was also constructed .
some supersymmetric black ring solutions have been also found , based on the construction of the solutions by gauntlett _ et al _
et al _ found the first supersymmetric black ring solution with asymptotic flatness on the four - dimensional euclidean base space , which is specified by three parameters , mass and two independent angular momentum components @xcite .
gauntlett and gutowski also constructed a multi - black ring solution on the same base space @xcite .
the bps black rings with three arbitrary charges and three dipole charges on the flat space were also constructed in @xcite .
the bps black ring solutions on the taub - nut base space @xcite and the eguchi - hanson space @xcitewere constructed . in recent years
, some non - bps black hole solutions have also been found in addition to supersymmetric black hole solutions .
although no one has found higher - dimensional kerr - newman solutions in einstein - maxwell theory yet , cvetic _
@xcite found a non - extremal , charged and rotating black hole solution with asymptotic flatness in the five - dimensional einstein - maxwell theory with a chern - simon term . in the neutral case ,
the solution reduces to the same angular momenta case of the myers - perry black hole solution @xcite .
exact solutions of non - bps kaluza - klein black hole solutions are found in neutral case @xcite and charged case @xcite .
these solutions have a non - trivial asymptotic structure , i.e. , they asymptotically approach a twisted @xmath0 bundle over the four - dimensional minkowski space - time .
the horizons are deformed due to this non - trivial asymptotic structure and have a shape of a squashed @xmath3 , where @xmath3 is regarded as a twisted bundle over a @xmath4 base space .
the ratio of the radius @xmath4 to that of @xmath0 is always larger than one .
as was proposed by wang , a kind of kaluza - klein black hole solutions can be generated by the from black holes with asymptotic flatness @xcite .
in fact , he regenerated the five - dimensional kaluza - klein black hole solution found by dobiasch and maison @xcite from the five - dimensional myers - perry black hole solution with two equal angular momentum . in the previous work
@xcite , applying the squashing transformation to the cvetic _
et al _ s charged rotating black hole solution @xcite in vanishing cosmological constant case , we obtain the new kaluza - klein black hole solution in the five - dimensional einstein - maxwell theory with a chern - simon term .
this is the generalization of the kaluza - klein black hole solutions in ref .
this solution has four parameters , the mass , the angular momentum in the direction of an extra dimension , the electric charge and the size of the extra dimension .
the solution describes the physical situation such that in general a non - bps black hole is boosted in the direction of the extra dimension . as the interesting feature of the solution , unlike the static solution @xcite
, the horizon admits a prolate shape in addition to a round @xmath3 .
the solution has the limits to the supersymmetric black hole solution and a new extreme non - bps black hole solutions and a new rotating black hole solution with a constant twisted @xmath0 fiber . in this article ,
applying this squashing transformation to kerr - gdel black hole solutions @xcite , we constructed a new type of rotating kaluza - klein black hole solutions to the five - dimensional einstein - maxwell theory with a chern - simon term .
we also investigate the features of the solutions .
though the gdel black hole solutions have closed timelike curves in the region away from the black hole , the new kaluza - klein black hole solutions generated by the squashing transformation have no closed timelike curve everywhere outside the black hole horizons . at the infinity ,
the space - time approaches a twisted @xmath0 bundle over a four - dimensional minkowski space - time .
the solution has four independent parameters , the mass parameter , the size of an extra dimension and two kinds of rotations parameters in the same direction of the extra dimension .
these two independent parameters are associated with the rotations of the black hole and the universe . in the case of the absence of a black hole , the solution describes the gross - perry - sorkin ( gps ) monopole which is boosted in the direction of an extra dimension and has an ergoregion by the effect of the rotation of the universe . the rest of this article is organized as follows . in sec.[sec : solution ] , we present a new kaluza - klein black hole solution in the five - dimensional einstein - maxwell theory with a chern - simon term . in sec.[sec : special ] , we study the special cases of our solution . in sec.[sec : feature ] , we investigate the basic features of the solution . in sec.[sec : summary ] , we summarize the results in this article . in the appendix , we will present more general solution .
first , we present the metric of a new kaluza - klein black hole solution to the five - dimensional einstein - maxwell theory whose action is given by the action : @xmath5 the metric and the gauge potential are given by @xmath6,\end{aligned}\ ] ] and @xmath7 respectively , where the functions in the metric are @xmath8 and the left - invariant @xmath9-forms on @xmath3 are given by @xmath10 the coordinates @xmath11 and @xmath12 run the ranges of @xmath13 , @xmath14 , @xmath15 , @xmath16 , respectively . @xmath17 and @xmath18 are constants .
the space - time has the timelike killing vector fields @xmath19 and two spatial killing vector fields with closed orbits , @xmath20 and @xmath21 . in the limit of @xmath22 ,
i.e. , @xmath23 with the other parameters kept finite , the metric coincides with that of the kerr - gdel black hole solution @xcite with ctcs . in this article
, we assume that the parameters @xmath24 and @xmath18 appearing in the solutions satisfy the following inequalities @xmath25 as will be explained later , these are the necessary and sufficient conditions that there are two horizons and no ctcs outside the horizons .
( [ eq : para1])-([eq : para3 ] ) are conditions for the presence of two horizons , and eq.([eq : para4 ] ) is the condition for the absence of ctcs outside the horizons .
it is noted that in the limit of @xmath23 with the other parameters finite , eq.([eq : para4 ] ) can not be satisfied .
thus , applying the squashing transformation to the kerr - gdel black hole solution , we can obtain such a kaluza - klein black hole solution without ctcs everywhere outside the black hole . here
we derive the parameter region ( [ eq : para1])-([eq : para4 ] ) from the demands for the presence of two horizons and the absence of ctc outside the horizons .
as will be shown later , the killing horizons are located at @xmath26 which are the roots of the equation @xmath27 , i.e. , the quadratic equations with respect to @xmath28 , @xmath29 this equation has two different real roots within the range of @xmath30 if and only if the parameters in the metric obey the inequalities @xmath31>0,\\ & & m^2\left[1 - 4j(a+2jm)\right]^2 - 2a^2>0.\end{aligned}\ ] ] under the condition @xmath32 , these inequalities are equivalent to ( [ eq : para1])-([eq : para3 ] ) .
next , in order to avoid the existence of ctcs outside the horizons , the parameters are chosen so that the two - dimensional @xmath33-part of the metric is positive - definite everywhere outside the outer horizon .
to do so , it is sufficient to require @xmath34 since @xmath35 is automatically assured if it is satisfied
. hence , the following inequality must be satisfied everywhere in the region @xmath36 $ ] .
@xmath37 here , it should be noted that it is sufficient to impose @xmath38 , which gives the equality ( [ eq : para4 ] ) .
consequently , the condition for the absence of ctcs outside the horizons is equal to the demand for the absence of ctcs at the infinity . in order to plot the parameter region ( [ eq : para1])-([eq : para4 ] ) in a two - plane
, we normalize the parameters @xmath39 and @xmath18 as @xmath40 and @xmath41 , respectively and furthermore , we fix the value of @xmath42 .
then , in the cases of @xmath43 , @xmath44 and @xmath45 , the quadratic curve @xmath46 in the condition ( [ eq : para3 ] ) becomes an ellipse , a line and a hyperbola , respectively .
the curve @xmath47 in the condition ( [ eq : para2 ] ) has different shapes in the cases of @xmath48 , @xmath49 and @xmath50 .
hence we consider the cases of ( i ) @xmath51 , ( ii ) @xmath49 ( iii ) @xmath52 ( iv ) @xmath44 and ( v ) @xmath45 .
the shaded regions in fig.[fig : p1]-fig.[fig : p5 ] show the parameter region ( [ eq : para1])-([eq : para4 ] ) for a given @xmath42 in each case of ( i)-(v ) , respectively .
-plane in the case of @xmath53 .
[ fig : p2 ] ] -plane in the case of @xmath53 .
[ fig : p2 ] ] -plane in the case of @xmath54.[fig : p4 ] ] -plane in the case of @xmath54.[fig : p4 ] ] -plane in the case of @xmath55.[fig : p5 ] ] in the coordinate system @xmath56 , the metric diverges at @xmath57 but we see that this is an apparent singularity and corresponds to the spatial infinity . to investigate the asymptotic structure of the solution , we introduce a new radial coordinate defined by @xmath58 where the positive constant @xmath59 is given by @xmath60 moreover , we introduce coordinates @xmath61 so that the metric is in a rest frame at the infinity @xmath62 where two constants @xmath63 and @xmath64 are chosen as @xmath65 then , for @xmath66 , the metric behaves as @xmath67 where the angular coordinate @xmath12 in @xmath68 is replaced by @xmath69 and the size of an extra dimension , @xmath70 , is given by @xmath71 the komar mass and the komar angular momenta at the spatial infinity are given by @xmath72,\\ j_{\phi}&=&0,\end{aligned}\ ] ] respectively .
therefore the space - time has only the angular momentum in the direction of the extra dimension .
in particular , it is noted that @xmath73 in the case of @xmath74 and @xmath75 in the case of @xmath76 .
therefore , two parameters @xmath77 with the opposite signs means the rotations in the same direction , and @xmath77 with the same signs means the rotations in the inverse direction . here
we investigate the near horizon geometry . as is mentioned previously , at the places such that @xmath27 , i.e. , at @xmath78 , the metric seems to be apparently singular but these correspond to killing horizons . to see this , we introduce new coordinates
@xmath79 defined by @xmath80 in the neighborhood of @xmath81 , the metric behaves as @xmath82+{\cal o}((r - r_{\pm})),\end{aligned}\ ] ] where the angular coordinate @xmath12 in @xmath68 is replaced by @xmath83 .
this means that the hypersurfaces @xmath81 are killing horizons since the killing vector field @xmath84 becomes null on @xmath78 and , furthermore , it is hypersurface - orthogonal , i.e. , @xmath85 .
we should also note that in the coordinate system @xmath86 , each component of the metric form is analytic on and outside the black hole horizon .
hence the space - time has no curvature singularity on and outside the black hole horizon .
as will be explained later , the ergosurfaces are located at @xmath87 such that @xmath88 vanishes , where @xmath89 is the cubic equation with respect to @xmath28 given by @xmath90r^4\nonumber\\ & & -[2a^2m+r_\infty^4(1 - 4j^2(2m+r^2_\infty))]^2r^2 + 2mr_\infty^8[1 - 2j(a+2j(2m+r^2_\infty))]^2.\end{aligned}\ ] ]
in this section , we focus on a few simple cases of the solution .
first we consider the simplest case of @xmath91 , where the solution has no event horizon . in terms of the coordinate @xmath92
, the metric can be written as @xmath93 ^ 2\nonumber\\ & & + \left(1+\frac{\rho_0}{\rho}\right)\left[d\rho^2+\rho^2(\sigma_1 ^
2+\sigma_2 ^ 2)\right]+\rho_0 ^ 2\left(1+\frac{\rho_0}{\rho}\right)^{-1}\sigma_3 ^ 2,\label{eq : gps}\end{aligned}\ ] ] where it should be noted that the requirement ( [ eq : para4 ] ) for the absence of ctcs imposes the parameters on the inequality @xmath94 . in the case of @xmath74 , the ( [ eq : gps ] ) exactly coincides with the gps monopole solution in ref.@xcite .
it is noted that the point of @xmath95 is a fixed point of the killing vector field @xmath21 , and corresponds to a kaluza - klein monopole .
this is often called a _
nut_. in the presence of the parameter @xmath96 , @xmath95 is also a fixed point of the killing vector field @xmath21 and the metric is analytic at the point .
hence the presence of the parameter @xmath96 means that although the space - time has no black hole , it is rotating along the direction @xmath21 of the extra dimension . for @xmath97
, the metric behaves as @xmath98\sigma_3 ^ 2,\end{aligned}\ ] ] where it should be noted that ctcs at the infinity in the gdel universe vanish via the squashing transformation . although this _ rotating gps monopole solution _ has no event horizon , we can confirm the presence of an ergoregion .
the @xmath99-component of the metric in the rest frame at the infinity take the form of @xmath100,\end{aligned}\ ] ] where the constant @xmath63 can be written in the form of @xmath101 from eq.([eq : a ] ) . as shown in fig.[fig : gps_ergo ] , in the case of @xmath102 , @xmath103 becomes negative in the region of @xmath104 , where @xmath105
this means that there is an ergoregion , although there is no black hole horizon in the space - time and in the neighborhood of the nut , the ergoregion vanishes . in the case of @xmath102 .
there exists an ergoregion in the region such that @xmath106 .
[ fig : gps_ergo ] ] next , we consider the case of @xmath107 . in this case , the solution is obtained via the squashing transformation for the schwarzschild - gdel black hole solution in ref.@xcite .
the parameter region such that there is a black hole horizon and is no ctc outside the horizon becomes @xmath108 the horizon is located at @xmath109 satisfying @xmath110 the shaded region in fig.[fig : pararegion1 ] shows the region satisfying the inequalities ( [ eq : pp1])-([eq : pp3 ] ) .
it should be noted that @xmath111 and @xmath112 always hold .
hence the ergosurfaces are always located at @xmath87 such that @xmath113 . in particular , when @xmath114 and @xmath115 , remarkably there are two ergoregions @xmath116 and @xmath117 outside the black hole horizon , where @xmath118 and @xmath119 ( @xmath120 ) are the roots of the quadratic equation with respect to @xmath28 , @xmath121 , and @xmath122 are the roots of the cubic equation with respect to @xmath28 , @xmath123 .
the small dark region in fig.[fig : sch_ergo ] denotes a set of the solutions which admits two ergoregions outside the horizon .
the squashed kerr - gdel black hole solution has two independent rotation parameters @xmath124 and @xmath96 , where both @xmath124 and @xmath96 denote the rotation along the killing vector @xmath21 ( or @xmath125 ) associated with a black hole and the ( squashed ) gdel universe , i.e. , the rotating gps monopole as the background , respectively . in this article , we call these parameters _ kerr parameter _ and _ gdel parameter _ , respectively . hence if the black hole is rotating in the inverse direction of the rotation of the ( squashed ) gdel universe , the effect of their rotations can cancel out the komar angular momentum .
in fact , the total angular momentum with respect to the killing vector field @xmath21 , @xmath126 , vanishes if the parameters satisfy @xmath127 as is shown in fig.[fig : angularzero]-[fig : angularzero5 ] , the solid curve denotes a set of the solutions such that the total angular momentum @xmath128 vanishes for various values of @xmath42 . in each case , as is expected , the curve of @xmath129 are located in the regions @xmath130 or @xmath131 . on the other hand ,
the komar angular momentum @xmath132 over the @xmath133 constant surface is obtained as @xmath134.\end{aligned}\ ] ] the dotted curves in fig.[fig : angularzero]-[fig : angularzero5 ] denote the curves of @xmath135 .
the parameter region are decomposed into four regions @xmath136 , @xmath137 , @xmath138 and @xmath139 by two curves @xmath140 and @xmath135 as follows @xmath141 in the solutions within @xmath136 and @xmath142 or @xmath139 and @xmath143 , @xmath144 and @xmath145 have the same sign since the black hole is rotating in the same direction as the squashed gdel universe ( rotating gps monopole ) .
in the solutions within @xmath136 and @xmath143 or @xmath139 and @xmath142 , although two parameters @xmath146 has the same sign , @xmath144 and @xmath145 have the same sign because the effect of the black hole s rotation exceeds that of the gdel rotation .
on the other hand , in the solutions within @xmath137 or @xmath138 , the angular momenta @xmath144 and @xmath145 also have the opposite sign as well as @xmath147 and @xmath148 .
therefore there exist the surface @xmath149 between the horizon @xmath150 and the infinity @xmath151 such that @xmath152 vanishes .
in fact , as is shown in fig.[fig : kg50p ] , the angular momentum vanishes between the horizon and the infinity .
it is expected that the effect of the gdel rotation exceeds that of the black hole s rotation at the large distance . and @xmath135 in the case of @xmath53 .
[ fig : angularzero2 ] ] and @xmath135 in the case of @xmath53 .
[ fig : angularzero2 ] ] and @xmath135 in the case of @xmath54 .
[ fig : angularzero4 ] ] and @xmath135 in the case of @xmath54 .
[ fig : angularzero4 ] ] and @xmath135 in the case of @xmath153[fig : angularzero5 ] ] here we examine the ergoregions of the squashed kerr - gdel black hole solution .
the metric can be rewritten as follows @xmath154 ^ 2\nonumber\\ & & -\frac{\frac{r^2}{4}v(r)}{h(r)+\frac{r^2}{4}}d{\bar t}^2+\frac{k(r)^2}{v(r)}dr^2+\frac{r^2}{4}k(r)(\sigma_1 ^ 2+\sigma_2 ^ 2).\end{aligned}\ ] ] therefore , at the outer horizon @xmath155 and the infinity @xmath57 , the @xmath99-component of the metric takes the non - negative form and the negative - definite form as follows @xmath156 hence , as is mentioned previously , the ergosurfaces are located at @xmath87 such that @xmath113 , i.e. , the cubic equation with respect to @xmath28 , @xmath123 .
like the squashed schwarzschild - gdel black hole solution in the previous section , the space - time admits the presence of two ergoregions outside the black hole horizon .
the spike like dark region @xmath157 in fig.[fig : komar_1 ] correspond to the parameter region such that @xmath158 , @xmath159 , @xmath114 and @xmath160 in the case of @xmath161 . it should be noted that in the case of @xmath162 , such a region does not appear within the parameter region ( [ eq : para1])-([eq : para4 ] ) .
fig.[fig : komar_3 ] shows the close - up of the dark region @xmath157 in fig.[fig : komar_1 ] .
the dotted curve in fig.[fig : komar_3 ] denotes @xmath163 , where the ergoregion near the outer horizon vanishes since in this case , the angular velocity @xmath164 of the black hole horizon vanishes . as is shown in fig.[fig : komar_3 ] , the curve of @xmath163 ( @xmath165 ) decomposes @xmath157 into two regions @xmath166 and @xmath167 . in the region of @xmath166 , the black hole horizon is rotating in the same direction as the gdel universe , while in the region of @xmath167 , they are rotating in opposite directions . here
introduce virtual zamos ( zero angular momentum observer ) located in the ergoregions .
the angular velocity of a zamo , @xmath168 , is given by @xmath169 .
at the horizon , note that @xmath170 .
the dashed curves in fig.[fig : kg50p ] denote the dependence on the radial coordinate @xmath87 of the angular velocity @xmath168 of zamos in the case of @xmath171 in @xmath166 and @xmath172 in @xmath167 , respectively . in the region of @xmath166 , a zamo in the ergoregion @xmath173 ( zamo1 ) and a zamo in
the ergoregion @xmath117 ( zamo2 ) is rotating in the same direction @xmath125 , while in the region of @xmath167 , zamo1 and zamo2 are rotating in the direction of @xmath21 and @xmath125 , respectively . in fig.[fig :
komar_1 ] . ] in fig.[fig : komar_1 ] . ] of the angular velocity of a zamo @xmath174 and the komar angular momentum @xmath175 in the case of @xmath176 and @xmath177 , respectively . ]
applying the squashing transformation to kerr - gdel black hole solutions , we have constructed a new type of rotating kaluza - klein black hole solutions to the five - dimensional einstein - maxwell theory with a chern - simon term .
the features of the solutions have been investigated . though the gdel black hole solutions have closed timelike curves in the region away from the black hole , the new solutions generated via the squashing transformation have no closed timelike curve everywhere outside the black hole horizons .
as the infinity , the space - time approaches a twisted @xmath0 bundle over a four - dimensional minkowski space - time . in the absence of a black hole , i.e. , in the case of @xmath178 , the solution describes the rotating gps ( gross - perry - sorkin ) monopole solution which is boosted in the direction of an extra dimension .
remarkably , in spite of the absence of a black hole , this space - time can have an ergoregion by the effect of the gdel rotation . in the squashed schwarzschild - gdel black hole solution with the parameters @xmath179 and @xmath96 ,
the space - time has one black hole horizon .
it has one ergoregion around the black hole horizon or two disconnected ergoregions , i.e. , an _ inner ergoregion _ located around the horizon and an _
outer ergoregion _ located far away from it . on the other hand , the squashed kerr - gdel black hole solution with the parameters @xmath180 and @xmath96 describes a rotating black hole in the squashed gdel universe ( the rotating gps monopole background ) .
the space - time has two horizons , i.e. , an inner black hole horizon and an outer black hole horizon .
the space - time has two independent rotations along the direction @xmath21 associated with the black hole and the squashed gdel universe . in the case of opposite rotations , the effect of the gdel s rotation and the black hole s rotation can cancel out the angular momentum at the infinity . in such a space - time , the ergoregions have richer and more complex structures . like the squashed schwarzschild - gdel black hole solution ,
the space - time also admits the existence of two ergoregions , an inner ergoregion and an outer ergoregion .
these two ergoregions can rotate in the opposite direction as well as in the same direction .
it should be noted that in addition to the mass and the angular momentum in the extra dimension , this space - time carries a electric charge @xmath181 given by @xmath182 in particular , in the case of the rotating gps monopole solution or the squashed schwarzschild - gdel black hole solution , the electric charge vanishes .
we can generalize our solution to a more general solution with an extra parameter @xmath183 by applying the squashing transformation to a kerr - newman - gdel black hole solution in ref.@xcite .
the solution is given in appendix[app ] .
we leave the analysis of this solution for the future . in the context of the study on dynamical features of bps black holes ,
black hole solutions on the euclid base space @xcite , the taub - nut base space @xcite and on the eguchi - hanson base space @xcite were also constructed in five - dimensional einstein - maxwell theory with a positive cosmological constant .
it is expected that in the vanishing limit of the cosmological constant , the solutions describe the dynamics of bps black holes .
in particular , two - black hole solution on the eguchi - hanson space describes a non - trivial coalescence of black holes . in refs .
@xcite , the authors compared the two - black hole solution on the eguchi - hanson space with the two - black holes solution on the euclid space @xcite , and discussed how the coalescence of five - dimensional black holes depends on the asymptotic structure of space - time .
we also discussed the effect by rotation of black holes @xcite . in the future forthcoming article @xcite
, we will also study the dynamics in the bps case of the squashed kerr - newman - gdel black hole solution by constructing such the solution in the five - dimensional einstein - maxwell theory with a chern - simon term and a positive cosmological constant .
this work is supported by the grant - in - aid for scientific research no.19540305 .
the metric and the gauge potential of the squashed kerr - newman - gdel black hole solution is given by @xmath6,\end{aligned}\ ] ] and @xmath184,\end{aligned}\ ] ] respectively , where the metric functions are @xmath185}{r^2}\nonumber\\ & & + \frac{2(m - q)a^2+q^2(1 - 16ja-8j^2(m+3q))}{r^4},\\ k(r)&=&\frac{v(r_\infty)r_\infty^4}{(r^2-r_\infty^2)^2}.\end{aligned}\ ] ] in the limit of @xmath186 , i.e. , @xmath22 , the solution coincides with the kerr - newman - gdel black hole solution in ref.@xcite . | applying _ squashing transformation _ to kerr - gdel black hole solutions , we present a new type of a rotating kaluza - klein black hole solution to the five - dimensional einstein - maxwell theory with a chern - simon term .
the new solutions generated via the squashing transformation have no closed timelike curve everywhere outside the black hole horizons . at the infinity ,
the metric asymptotically approaches a twisted @xmath0 bundle over a four - dimensional minkowski space - time .
one of the remarkable features is that the solution has two independent rotation parameters along an extra dimension associated with the black hole s rotation and the gdel s rotation .
the space - time also admits the existence of two disconnected ergoregions , an inner ergoregion and an outer ergoregion .
these two ergoregions can rotate in the opposite direction as well as in the same direction . |
water is an anomalous fluid with respect to several thermodynamic properties @xcite . at ordinary pressures
the solid phase ( ice ) is less dense than the corresponding liquid , the liquid phase has a temperature of maximum density , while both isothermal compressibility and isobaric heat capacity display a minimum as a function of temperature . moreover , the heat capacity is unusually large .
there is general agreement , among physicists , that an explanation of such anomalous properties is to be found in the peculiar features of hydrogen bonds , and the ability of water molecules to form such kind of bonds @xcite .
it is also widely believed that the same physics should be responsible of the unusual properties of water as a solvent for apolar compounds @xcite , that is of the hydrophobic effect , of high importance in biophysics @xcite .
nevertheless , a comprehensive theory which explains all of these phenomena has not been developed yet .
a lot of work has been done in `` realistic '' simulations @xcite , based on different interaction potentials , but they generally require a large computational effort , and it is not always easy to understand which detail of the model is important to determine certain properties . on the contrary , simplified models generally need easier numerical calculations and allow quite easily to trace connections between microscopic interactions and macroscopic properties @xcite . a simplified mechanism which has been proposed to describe the relevant physics of hydrogen bonding is the following one ( see for instance refs . ) .
hydrogen bond formation requires that the two involved molecules are in certain relative orientations and stay ( on average ) at a distance which is larger than the optimal distance for van der waals interaction . in other words
there exists a competition between van der waals interaction ( allowing _ higher density _ and _ higher orientational entropy _ , but resulting in a _ weaker bonding _ ) and hydrogen bonding ( requiring _ lower density _ and _ lower orientational entropy _ , but resulting in a _ stronger bonding _ ) .
this simple mechanism has been implemented in different models , both on- @xcite and off - lattice @xcite , in 3 @xcite as well as 2 dimensions @xcite .
one of them is the 2-dimensional mercedes benz model , originally proposed by ben - naim @xcite , in which model molecules possess three bonding arms arranged as in the mercedes benz logo . in recent papers by dill and coworkers @xcite ,
a similar ( off - lattice ) model has been simulated at constant pressure by a monte carlo method , allowing to describe in a qualitatively correct way several anomalous properties of liquid water and also of hydrophobic solvation .
nevertheless , in view of investigations on the behavior of water in contact with other chemical species , as it happens for instance in several biological processes , it would be desirable to obtain an even simpler representation of the physics of hydrogen bonding . in this paper
we investigate a model of the mercedes benz type on the triangular lattice , with a twofold purpose .
as mentioned above , we are first meant to explore the possibility of obtaining a simpler model with the same underlying physical mechanism , and with qualitatively the same macroscopic properties .
moreover , we are interested in extending the model analysis to the global phase diagram and in particular to the supercooled regime , in which water anomalies are thought to find an explanation . such a detailed analysis is just made easier by increased simplicity .
working on a lattice , we have to resort to a trick to describe hydrogen bond weakening , when the two participating molecules are too close to each other .
such a trick is similar to the one proposed by roberts and debenedetti for their 3-dimensional model @xcite .
the energy of any formed bond is increased ( weakened bond ) of some fraction by the presence of a third molecule on a site close to the bond ( i.e. , on the third site of the triangle ) . due to the presence of only three bonding arms
, it is not possible to distinguish between hydrogen bond donors and acceptors , but this seems to be of minor importance to the physics of hydrogen bonding @xcite .
let us notice that the model has the same bonding properties as the early model proposed by bell and lavis @xcite , and the same weakening criterion as the model recently investigated by patrykiejew and coworkers @xcite , but here non - bonding orientations are added .
such a feature is essential to describe directional selectivity of hydrogen bonds .
the paper is organized as follows . in sec .
ii we define the model in detail and analyze its ground state . in sec
iii we introduce the first order approximation in a cluster variational formulation , which we employ for the analysis .
iv describes the results and sec .
v is devoted to some concluding remarks .
the model is defined on a two dimensional triangular lattice .
a lattice site can be empty or occupied by a molecule with three equivalent bonding arms separated by @xmath1 angles .
two nearest - neighbor molecules interact with an attractive energy @xmath2 ( @xmath3 ) representing van der waals forces .
moreover , if two arms are pointing to each other , an orientational term @xmath4 ( @xmath5 ) is added to mimic the formation of a hydrogen ( h ) bond . due to the lattice symmetry , a particle can form three bonds at most and there are only 2 bonding orientations , when the arms are aligned with the lattice , while we assume that @xmath6 non - bonding configurations exist ( @xmath6 is another input parameter of the model ) .
finally , the h bond energy is weakened by a term @xmath7 ( @xmath8 $ ] ) when a third molecule is on a site near a formed bond . in the two dimensional triangular lattice
there are two such weakening sites per bond , so that a fully weakened h bond energy turns out to be @xmath9 .
let us notice that , in the above description , h bonding is a 3-body interaction .
the hamiltonian of the system can be written as a sum over the triangles @xmath10 where @xmath11 is a contribution which will be referred to as triangle hamiltonian , and @xmath12 label site configurations for the 3 vertices @xmath13 , respectively .
possible configurations are empty site ( @xmath14 ) , site with a molecule in one of the 2 bonding orientations ( @xmath15 ) or in one of the @xmath6 non - bonding ones ( @xmath16 ) ( see tab . [ tab - site - conf ] ) .
the triangle hamiltonian reads @xmath17
\nonumber , \label{eq : triham}\end{aligned}\ ] ] where @xmath18 is an occupation variable , defined as @xmath19 for @xmath14 ( empty site ) and @xmath20 otherwise ( occupied site ) , while @xmath21 if the pair configuration @xmath22 forms a h bond , and @xmath23 otherwise .
let us notice that triangle vertices are set on three triangular sublattices , say @xmath24 , and @xmath25 are assumed to denote configurations of sites placed on @xmath24 sublattices respectively .
assuming also that @xmath24 are ordered counterclockwise on up - pointing triangles ( and then clockwise on down - pointing triangles ) , we can define @xmath21 if @xmath26 and @xmath27 and @xmath23 otherwise .
let us notice that both van der waals ( @xmath28 ) and h bond energies ( @xmath29 ) , that are 2-body terms , are split between two triangles , whence the @xmath30 prefactor in eq . . on the contrary the 3-body weakening terms ( @xmath31 )
are associated each one to a given triangle , and the @xmath30 factor is absorbed in the prefactor .
let us denote the triangle configuration probability by @xmath32 , and assume that the probability distribution is equal for every triangle ( no distinction between up- or down - pointing triangles ) .
taking into account that there are 2 triangles per site , we can write the following expression for the internal energy per site of an infinite lattice @xmath33 the multiplicity for the triangle configuration @xmath34 is given by @xmath35 , where @xmath36 for @xmath16 ( non - bonding configuration ) and @xmath37 otherwise ( bonding configuration or vacancy ) . l|cccc config . &
empty & & & @xmath38 & 0 & 1 & 2 & 3 @xmath39 & 1 & 1 & 1 & @xmath6 [ tab - site - conf ] let us now have a look at the ground state properties of the model . in order to do so ,
let us investigate the zero temperature grand - canonical free energy @xmath40 ( @xmath41 being the chemical potential and @xmath42 the density , i.e. , the average site occupation probability ) , which can be formally written in the same way as the internal energy @xmath43 of eq .
, by replacing the triangle hamiltonian @xmath11 by @xmath44 we find an infinitely dilute `` gas '' phase ( g ) with zero density and zero free energy , and an ordered `` open ice '' phase ( @xmath45 ) with maximum number of h bonds per molecule .
the latter configuration is realized through the formation of an open ( honeycomb ) h bond network with density @xmath46 and free energy @xmath47 another possibility is the `` closed ice '' phase ( @xmath48 ) , in which all interstitial sites are occupied and all hydrogen bonds are fully weakened .
the resulting free energy is @xmath49 let us notice that it is never possible to form 3 bonds in a triangle , which means that we have frustration .
it is easy to show that the g phase is stable ( @xmath50 ) for @xmath51 , where @xmath52 the @xmath45 phase is stable ( @xmath53 and @xmath54 ) for @xmath55 , where @xmath56 and the @xmath48 phase is stable ( @xmath57 and @xmath58 ) for @xmath59 .
the @xmath45 phase has actually a stability region , i.e. , @xmath60 , provided @xmath61 which , in the worst case ( @xmath62 ) , reads @xmath63 .
we shall always work in the latter regime , which is the most significant one to describe real water properties .
it is also possible to show that , at the transition point between the open and closed ice phases ( @xmath64 ) , any configuration built up of a honeycomb h bond network with any number of occupied interstitial sites has the same free energy .
hence we expect that the @xmath65 transition does not exist at finite temperature , and actually we shall observe a unique ice ( i ) phase , in which the interstitial site occupation probability gradually increases upon increasing the chemical potential .
let us finally notice that another possible phase is a homogeneous and isotropic one in which the lattice is fully occupied and molecules can assume only bonding configurations ( @xmath15 ) .
this `` bonded liquid '' phase , whose free energy coincides with that of the @xmath48 phase in eq .
, is observed in the @xmath66 case , studied by patrykiejew and others @xcite . in this scenario ,
non - bonding configurations are absent and the bonded liquid ground state has , for @xmath67 , the same degeneracy as the ising triangular antiferromagnet @xcite .
nevertheless , in this work we shall deal with the case @xmath68 , which is relevant to describe h bond directionality . in this case
the closed ice phase is entropically favored with respect to the bonded liquid phase , which can not appear at finite temperature . in conclusion , because of the introduction of non - bonding configurations , the ground state degeneracy is removed at @xmath69 , where only an infinitely dilute ( gas ) phase and a symmetry - broken ( ice ) phase are present .
such a phase behavior is closer to the one of water than the one obtained for @xmath70 .
we shall carry out the finite temperature analysis of the model mainly by means of a generalized first order approximation on a triangle cluster , which we introduce in the framework of the cluster variation method .
the cluster variation method is an improved mean - field theory based on an approximate expression for the entropy . in kikuchi s original formulation @xcite the entropy
is obtained by an approximate counting of the number of microstates . in a modern formulation
@xcite the approximate entropy can be viewed as a truncation of a cluster cumulant expansion .
the truncation is justified by the expected rapid vanishing of the cumulants upon increasing the cluster size , namely when the cluster size becomes larger than the correlation length of the system ( the method necessarily fails near critical points ) @xcite .
the approximation is completely defined by the maximum clusters left in the truncated expansion , usually denoted as basic clusters .
one obtains a free energy functional in the cluster probability distributions , to be minimized , according to the variational principle of statistical mechanics . for our model
we choose up - pointing triangles as basic clusters ( an analogous treatment works for down - pointing triangles ) .
this approximation , which seems to be good in particular for frustrated models @xcite , is easily shown to be equivalent to a first order approximation on a triangle cluster @xcite .
let us notice that the internal energy is treated exactly , because the range of interactions does not exceed the basic cluster size , unlike the ordinary mean - field approximation .
the grand - canonical free energy per site @xmath71 ( @xmath72 being the entropy per site ) , can be written as a functional in the triangle probability distribution as @xmath73 \nonumber , \end{aligned}\ ] ] where @xmath74 ( temperature is expressed in energy units , whence entropy in natural units ) and @xmath75 is the probability of the @xmath38 configuration for a site on the @xmath76 sublattice ( @xmath77 ) .
the latter can be obtained as a marginal of the triangle configuration probability @xmath32 , namely @xmath78 the above expressions show that the only variational parameter in @xmath79 is the triangle probability distribution , that is the 64 variables @xmath80
. the minimization of @xmath79 with respect to these variables , with the normalization constraint @xmath81 can be performed by the lagrange multiplier method , yielding the equations @xmath82 where @xmath83 , related to the lagrange multiplier , is obtained by imposing the constraint eq .
: @xmath84 eq .
is in a fixed point form , and can be solved numerically by simple iteration ( natural iteration method @xcite ) . in our case the numerical procedure can be proved to lower the free energy at each iteration @xcite , and therefore to converge to local minima .
the solution of eq . gives the equilibrium @xmath80 values , from which one can compute the thermal average of every observable .
inserting these values into eqs . and gives respectively the equilibrium internal energy and free energy .
the latter can be also easily expressed through the normalization constant as @xmath85 whence @xmath83 can be viewed as the approximate ( single site ) grand - canonical partition function .
it is also worth mentioning that eq
. preserves homogeneity ( @xmath86 ; @xmath87 ) , due to the invariance of @xmath88 under cycle permutation of the subscripts ( see eqs . and ) .
let us finally notice that the free energy expression eq . can be also derived by considering the model on a triangular husimi tree ( triangle cactus ) @xcite as a bulk free energy density , that is the free energy contribution far enough from the boundary , where an invariance condition for the configuration probability of the triangles is assumed to hold .
in order to provide a first insight into the model , let us report in fig .
[ phasediag ] the phase diagram in the chemical potential - temperature plane , for @xmath89 , @xmath90 , and @xmath91 .
three phases can be observed : an ice ( i ) phase , with broken symmetry among the three sublattices , a liquid ( l ) phase and a gas ( g ) phase . the latter two phases
preserve the sublattice symmetry but the liquid phase has a higher density .
the ice phase has a lower density than the liquid phase , and its structure reminds that of ground state ice , with interstitial sites occupied by molecules in non - bonding configurations . we can observe a triple point ( trp ) , in which the three phases coexist , and a gas - liquid critical point ( cp ) .
all displayed transition lines are first - order .
the above phase diagram shares several properties with the one of real water .
other crystalline phases , such as a real close - packed ice , can not be reproduced by the model .
let us now investigate the role of model parameters , by analyzing phase diagrams obtained for different values . in fig .
[ manyphase]a , @xmath92 and @xmath93 are left unchanged , while the number of non - bonding configurations @xmath6 is varied within the interval @xmath94 $ ] . upon increasing @xmath6
, the liquid phase turns out to be more stable with respect to the ice phase , and the i - l transition temperature decreases . on the contrary , for lower @xmath6 values ,
the i phase is increasingly stabilized and the i - l transition temperature increases . for @xmath95 the whole l - g coexistence and
also the critical point disappears .
such a behavior can be explained by the fact that the l phase is characterized by a higher number of non - bonding molecules than the i phase , in which bonding molecules tend to form an ordered structure .
therefore high @xmath6 values largely increase the liquid phase entropy . in fig .
[ manyphase]b , @xmath6 and @xmath93 are held fixed and the ratio @xmath92 is varied within the interval @xmath96 $ ] .
let us notice that we have restricted the investigation to cases in which the orientational ( h bond ) interaction is stronger than the non - orientational one , which is the case for real water .
it turns out that the ratio @xmath92 affects the stability of the i phase with respect to both the g and l phases .
in fact higher values of @xmath97 means stronger h bond , which favors the i phase , that is the only extensively h - bonded phase . on the contrary
the l and g phases are dominated by non - oriented interactions with coupling constants @xmath98 , therefore both these two phases are unfavored by high @xmath92 values . even in this case
the l - g coexistence may become metastable .
the ice phase at high pressures has maximum density and number of weakening molecules per h bond . raising @xmath93 ,
the stability of this configuration is lowered with respect to the liquid phase with few h bonds .
this is shown in fig .
[ manyphase]c where @xmath92 and @xmath6 are fixed and the weakening parameter @xmath93 is varied in its interval of definition @xmath99 $ ]
. this trend is reversed for low @xmath6 values ( @xmath66 as well ) , because in the latter case the liquid has the maximum number of fully weakened bonds . in the next part of this work we focus on a particular choice of parameters ( @xmath100 , @xmath101 and @xmath102 ) which , from the above analysis , turn out to correspond to a water - like phase diagram .
[ pt_spin_max ] shows the temperature - pressure phase diagram , and fig .
[ density ] the temperature - density phase diagram .
let us notice that pressure @xmath103 is simply given by @xmath104 ( the volume per site is assumed to be equal to 1 , i.e. , pressure is expressed in energy units ) , due to the fact that the free energy has been defined as a grand - canonical potential .
one of the water anomalies that the present model is able to reproduce is the temperature of maximum density ( tmd ) along isobars for the liquid phase .
joining tmd at different pressures defines the so called tmd locus , which is a negatively sloped line in the @xmath105-@xmath103 phase diagram of real water .
we determine the tmd locus numerically , by adjusting the chemical potential in order to fix the pressure and then imposing that the ( isobaric ) thermal expansion coefficient vanishes .
the limit of stability of the liquid phase ( spinodal ) is the locus in which the metastable liquid ceases to be a minimum of the free energy , and becomes a saddle point .
the stability limit can be obtained by studying the eigenvalues of the hessian matrix of the free energy @xcite @xmath106\biggr\ } \nonumber .\end{aligned}\ ] ] let us notice that , when the liquid phase stability is lost ( some eigenvalue of the above matrix vanishes ) , also the corresponding fixed point of the natural iteration equations becomes unstable . in order to determine the stability limit with respect to the symmetry - broken ice phase , it is sufficient to impose homogeneity during the iterative procedure , which is done by replacing eqs . with @xmath107 this trick
can not be applied when the liquid stability is lost with respect to a homogeneous phase , because the liquid fixed point of equations becomes definitely unstable , due to divergence of the density response functions . in the latter case
the spinodal is determined by solving the eigenvalue problem for the hessian matrix rewritten by forcing the homogeneity condition .
the results are shown in fig .
[ pt_spin_max ] .
the stability limit of the liquid with respect to the gas phase starts from the critical point and reaches a minimum in the negative pressure region .
after this point the line becomes negatively sloped and joins continuously the stability limit with respect to the ordered ice phase .
the tmd locus intersects the limit of stability in its minimum in the @xmath105-@xmath103 plane , according to the predictions of speedy and debenedetti @xcite , based on thermodynamic consistency arguments .
in fact the tmd locus causes the liquid limit of stability line to retrace , giving rise to a tensile strength maximum and to a continuous boundary .
let us recall that , while at the stability limit with respect to the gas phase , the density response functions diverge , this is not the case at the stability limit with respect to the ordered phase .
nevertheless we can observe that the density response functions tend to diverge also upon decreasing temperature , as observed experimentally .
the locus of divergence , terminating at some kind of critical point , can be defined , in the framework of a simplified variational free energy forced to describe a homogeneous system , as an additional stability limit with respect to a low density liquid phase .
such `` phase '' corresponds to a saddle point of the original ( not symmetrized ) free energy , unstable with respect to the solid phase . as the low pressure
solid phase reminds the ground state `` open ice '' structure , which is three - fold degenerate , the triangle probability distribution of the low density liquid phase turns out to be essentially an arithmetic average over the three ice distributions .
the unphysical nature of this solution is also reflected in its negative entropy .
the divergence locus , together with the locus at which the liquid phase entropy vanishes ( kauzmann line ) , are shown for completeness in the inset of fig .
[ pt_spin_max ] . upon increasing temperature
the divergence locus meets the spinodal tangentially and they become the same curve ending in the `` true '' gas - liquid critical point .
let us now investigate the density response functions and the specific heat of the liquid at constant pressure @xmath108 ( pressure is kept fixed by numerically adjusting the chemical potential @xmath41 ) .
it turns out that these functions display an anomalous behavior similar to that of real liquid water .
the first response function we consider is the thermal expansion coefficient @xmath109 , which is proportional to the entropy - specific volume cross - correlation . for a typical fluid @xmath110 is always positive because if in a region of the system the specific volume is a little larger then the average , then the local entropy is also larger , i.e. , the two quantities are positively correlated . on the contrary , for our model @xmath110 ( fig .
[ response]a ) displays an anomalous behavior .
as temperature is lowered @xmath110 vanishes ( at the tmd ) , becomes negative , and finally tends to diverge .
as previously mentioned , divergence can be observed only for pressure values less than some `` critical '' pressure .
anyway , before divergence is actually reached , the liquid loses stability with respect to the ice phase .
the trend of the isothermal compressibility @xmath111 is also anomalous ( fig .
[ response]c ) . for a typical liquid @xmath112 decreases as one lowers temperature , because it is proportional to density fluctuations , which decrease upon decreasing temperature . on the contrary , in fig .
[ response]c we can observe that @xmath112 , once reached a minimum , begins to increase upon decreasing temperature .
such a behavior is observed in real liquid water .
an analogous behavior characterizes the constant pressure specific heat @xmath113 ( fig .
[ response]b ) .
we have studied the model in the first order approximation to obtain easily detailed information about phase diagrams and in particular the metastable region . in order to check this approximation and obtain an estimate of its quantitative accuracy ,
we have also performed some ( grand - canonical ) monte carlo simulations on a @xmath114 triangular lattice with periodic boundary conditions . from the very beginning ,
we have chosen quite a low number of non - bonding configurations for our analysis ( @xmath95 ) , in order to increase the speed of simulation dynamics .
in fact a lower @xmath6 value corresponds to a smaller configuration space .
we report some results in the following . in fig .
[ transition ] we show a first order transition between the gas and the liquid phases along a constant temperature path @xmath115 , quite less than the critical temperature . at the critical point
the correlation length increases and the approximation may give worse predictions . fig . [
transition ] suggests that the first order approximation well localizes the transition and that far enough from the critical point its predictions are nearly quantitative .
of course monte carlo simulations display smooth density variations , due to finite size effects , but the binder cumulant ( inset ) , displaying a minimum , gives evidence of a first order transition .
the reentrance of the liquid stability limit , which is one of the striking features of the ( metastable ) phase diagram of this model , is also confirmed by simulations . performing simulations in the metastable region ,
the spinodal has been determined by an arbitrary criterion for the life time of the metastable phase ( 100 monte carlo steps ) , as it has been done in previous studies @xcite .
such a criterion allows us to find the kinetically controlled limit of supercooling ( homogeneous nucleation locus ) , shown in fig .
[ mcspin ] , along with the corresponding first order approximation result .
both methods show a reentrant spinodal forming a continuous boundary .
the simulations also confirm the distinction between liquid limit of stability with respect to the gas or to the ice phase , as in the first order approximation .
in this paper we have investigated a 2-dimensional lattice model in which model molecules possess three equivalent bonding arms , and bonding energy depends on the presence of neighbor molecules , giving rise to a 3-particle interaction .
the observed behavior is qualitatively similar to that of water , exhibiting the correct anomalies . upon supercooling , @xmath112 and
@xmath116 increase and @xmath110 becomes negative and large in magnitude .
nevertheless , at ordinary pressures ( less than the critical pressure ) the density anomaly ( @xmath117 ) is found in the metastable liquid region .
we have also determined the spinodal limits to the liquid state , and pointed out the relationship between these limits and the tmd locus . the growth in the response functions upon decreasing temperature can be interpreted on the basis of a reentrant spinodal scenario .
the liquid - gas spinodal meets the tmd locus at the reentrance point , as required by thermodynamic consistency .
actually the reentrant spinodal conjecture is one of the possible theoretical explanations of water anomalies , and some experimental results are consistent with this explanation @xcite .
nevertheless , it is important to note that , for the specific case of water , alternative interpretations of the stability problem exist , based on the second critical point conjecture @xcite .
the latter , supported by molecular dynamics simulations @xcite , seems to be more consistent with the existence , in the negative pressure region , of a monotonic liquid - gas spinodal and a reentrant tmd locus . on the contrary
, our model displays a metastable liquid state which is bounded by a spinodal both at positive as well as negative pressures , forming a continuous boundary . the lower temperature part of the boundary is the limit of stability with respect to the ordered ice phase , while the higher temperature part is the limit of stability with respect to the gas phase . while the response functions diverge at the liquid - gas spinodal , at the liquid - solid spinodal they do not , even if they tend to higher values .
anyway , in our framework , it is also possible to investigate the behavior of the unstable liquid ( a saddle point of the variational free energy ) and determine the locus of divergence .
the latter always turns out to lie at a temperature less than the limit of stability , according to experiments @xcite .
it also turns out that the divergence locus terminates at some kind of critical point , meaning that response functions should not show divergent - like behavior for pressure values greater than some critical pressure .
let us notice that a previous lattice model on the 3-dimensional body centered cubic lattice had pointed out a qualitatively similar behavior @xcite .
nevertheless , in such a model , orientational degrees of freedom of water are not treated explicitly and two equivalent sublattices are artificially distinguished by the hamiltonian .
this is necessary to favor an open structured phase .
moreover , the analytical treatment is based on the determination of a temperature dependent 2-particle interaction . on the contrary in our model there
exists an explicit , though simplified , modelling of hydrogen bonding and no temperature dependent interaction is introduced .
the open structured phase is favored in principle by the triangular lattice structure .
we have mentioned in the introduction that the present model is actually an extension over an early model proposed by bell and lavis @xcite ( corresponding to the case in which @xmath70 and @xmath62 ) and over a recent model investigated by patrykiejew and coworkers @xcite ( corresponding to @xmath70 ) .
the former model in the same approximation actually displays , for @xmath118 , a density anomaly ( without singularities ) , but we have verified that the anomaly occurs in a negative entropy region .
the latter model shows an unrealistic phase diagram , in which , for high enough pressure , the liquid phase extends its stability region down to zero temperature . in the present work we have shown that the addition of non - bonding configurations to such a simple class of 2-dimensional lattice models allows us to reproduce a qualitatively correct water - like behavior .
moreover , this result has been obtained in a computationally much simpler way than a conceptually similar model with continuous degrees of freedom , that is the mercedes - benz one .
the latter model is highly appealing , because of its ability to explain most phenomena related to hydrophobicity @xcite . therefore it would be interesting to analyze also the properties of the present model for a solution of an inert ( apolar ) solute , whose peculiar properties are thought to be strictly related to hydrogen bonding .
this goes beyond the scope of the present paper and will be the subject of a forthcoming article . | we investigate a lattice - fluid model defined on a two - dimensional triangular lattice , with the aim of reproducing qualitatively some anomalous properties of water .
model molecules are of the `` mercedes benz '' type , i.e. , they possess a @xmath0 ( equilateral triangle ) symmetry , with three bonding arms . bond formation depends both on orientation and local density .
we work out phase diagrams , response functions , and stability limits for the liquid phase , making use of a generalized first order approximation on a triangle cluster , whose accuracy is verified , in some cases , by monte carlo simulations .
the phase diagram displays one ordered ( solid ) phase which is less dense than the liquid one . at fixed pressure
the liquid phase response functions show the typical anomalous behavior observed in liquid water , while , in the supercooled region , a reentrant spinodal is observed . |
in cataclysmic variables ( cvs ) , a white dwarf accretes from a low mass secondary star through roche - lobe overflow .
angular momentum loss drives the binary system to progressively shorter orbital periods until the donor star becomes degenerate .
further mass exchange then results in an increase of the orbital period of the system , which leads to a predicted minimum period of @xmath1070 minutes through which all cvs should eventually evolve ( e.g. kolb & baraffe 1999 ) . with an orbital period of only 82 minutes ,
wz sge is one of the very few candidates among the hundreds of known cvs that may have already evolved past this minimum period ( patterson 1998 ) .
its quiescent magnitude of @xmath11 and distance of approximately 45 pc ( thorstensen 2001 , astrometric parallax , private communication ) , make it one of the lowest luminosity cvs known . despite the accretion light being faint enough for the white dwarf to dominate in quiescence ,
infrared spectroscopy has so far not shown any signs of the mass donor star ( littlefair et al .
2000 ) . instead of the 3 - 5 magnitude outbursts that longer period dwarf novae tend to undergo every few weeks to months ,
wz sge s outbursts have an amplitude of 7 - 8 magnitudes and recur on a timescale of roughly 33 years .
these facts suggest that wz sge is a highly evolved system , in which gigayears of mass transfer have converted a main sequence secondary star to a degenerate brown dwarf .
however , the absence of any direct signatures of the secondary have prevented a robust determination of the system parameters of this unique system .
on july 23rd 2001 , amateur observers reported a sudden and rapid brightening of wz sge ( ishioka et al .
2001 ) , indicating another outburst had started , around 23 years after the previous outburst in 1978 , and therefore around 10 years earlier than anticipated . here
we present phase resolved spectroscopy of wz sge in the first weeks of its 2001 outburst , which reveals a clear signature of the irradiated mass donor star for the first time .
the phase resolved spectra of wz sge were obtained with the 2.5 m isaac newton telescope ( int ) and the 4.2 m william hershel telescope ( wht ) on the island of la palma .
the intermediate dispersion spectrograph on the int in conjunction with the r1200b grating delivered a wavelength coverage of 3800 - 4950 at 0.48 / pixel using an eev ccd detector . on august the 6th ,
1022 spectra were obtained in total on the int using 13 s exposures between 20:46 and 22:14 ut and 00:14 - 04:36 ut . on the wht , the dual arm isis spectrograph was used , covering 4220 - 4975 on the blue arm at 0.22 / pixel ( eev ccd ) and 6380 - 6775 at 0.4 / pixel on the red arm ( tek ccd ) .
the slit width was adjusted in order to project to 2 - 3 pixels on the ccd .
since no sufficiently bright comparison star was available that could be aligned along the slit , the slit was always kept close to the parallactic angle . on august 13th
, 541 red arm spectra using 10 s exposures and 233 blue arm 15 s exposures were acquired between 20:50 and 23:03 ut .
frames were de - biased using the overscan areas of the ccds . a normalised median of tungsten exposures
was then constructed for each night and used to carry out flat field correction .
finally , the wz sge spectra were optimally extracted ( marsh , 1989 ) .
regular arc lamp exposures allowed us to establish an accurate wavelength scale for each spectrum through interpolation between the two nearest arc spectra .
the individual spectra were normalised to the continuum level using a spline fit to selected continuum regions .
in figure [ average ] we present the average normalised spectrum of wz sge on august 6th and 13th .
both observations occured during the initial slow decline phase of the outburst with approximate v magnitudes of 10.0 on the 6th ( 13 days into the outburst ) and 10.5 on the 13th . for comparison , @xmath12 at outburst maximum on july 24th ( we refer to the aavso and vsnet webpages for extensive visual magnitude estimates throughout the outburst ) .
the spectrum of wz sge in outburst is dominated by complex balmer and helium line profiles . except for the heii / bowen blend emission complex ,
all lines show very deep and phase dependent absorption components on top of double peaked emission . in most cases
the absorption goes well below the continuum level , except in h@xmath13 .
we use the orbital ephemeris of patterson et al .
( 1998 ) , hereafter p98 , to calculate the orbital phases throughout this letter .
the ephemeris is a small improvement to the often - used robinson , nather and patterson ( 1978 ; hereafter rnp ) ephemeris amounting to a difference of @xmath100.001 cycles at the time of the current outburst .
the ephemeris zeropoint is based on the sharp eclipses of the bright spot and does not correspond to the inferior conjunction of the secondary star .
spruit & rutten ( 1998 ) , hereafter sr , for example derive a phase offset of @xmath14 between the rnp ephemeris and the mid - point of the accretion disc eclipse .
we will address this phase offset in section 3.3 .
the time dependent h@xmath13 line emission as observed on august 13th is displayed in figure [ trails ] as a trailed spectrogram .
the double - peaked line profile is highly phase dependent and asymmetric .
apart from the two double peaks reflecting emission from the accretion disc , a narrow emission component moves through the peaks and produces a clear s - wave on the orbital period . on much shorter timescales
, rapid transient features can be seen crossing the line profile , almost always from blue to red .
an example is the narrow emission feature rapidly crossing in a few minutes around phase 0.55 . during quiescence ,
the line profiles of wz sge are dominated by the double - peaked disc emission as well as a strong contribution from the bright spot ( sr , skidmore et al .
however , the s - wave in our outburst data is much narrower and traces a near sinusoidal radial velocity curve .
such narrow , sinusoidal s - waves are commonly observed in accreting binaries and are generally attributed to line emission from an irradiated secondary star .
provided a sufficient amount of ionising radiation is received from the compact object and the inner disc regions , optical line emission from the exposed parts of the secondary star is produced ( e.g. marsh & horne 1990 , harlaftis et al .
1996 , steeghs & casares , 2001 ) .
this interpretation is supported by the phasing of the s - wave , which corresponds closely to the expected phasing of the secondary in wz sge .
in addition , its strength is highly phase dependent , reaching maximum emission around orbital phase 0.5 , as expected from an irradiated roche lobe filling star .
given the complexity of the line profiles , we decided to use doppler tomography ( marsh & horne 1988 ) to isolate and study the nature of the emission line components .
doppler tomography requires the systemic velocity ( @xmath15 ) to be supplied as input to the reconstruction algorithm .
gilliland et al .
( 1986 ) measured @xmath16 km / s using radial velocities derived from the h@xmath13 line wings .
skidmore et al .
( 2000 ) derive a mean systemic velocity of @xmath17 km / s based on radial velocity curves of several emission lines .
both of these determinations rely on the assumption that the emission line velocities reflect the motion of the compact object .
on the other hand , sr used doppler tomography techniques to determine the systemic velocity , by minimising the residuals between observed and predicted data .
they found @xmath18 km / s .
we chose to measure the systemic velocity directly using the detected s - wave from the secondary star , since its radial velocity is displaced by the true systemic velocity of the binary system irrespective of the properties of the accretion flow around the white dwarf . to this end
we calculated a series of preliminary doppler maps from the observed data using a filtered back projection method ( marsh 2001 ) . for each map ,
a different systemic velocity was assumed ranging from 0 to -110 km / s in 10 km / s steps .
the doppler maps then provide the strength of all s - waves on the orbital period in the data with a given amplitude and phase , which is maximised when the correct value for @xmath15 is used .
we measured the strength of the secondary star emission in each doppler image , and found that a well defined maximum was achieved for @xmath19 km / s using the observed h@xmath13 emission .
the same analysis applied to the h@xmath20 emission also reveals a clear secondary star contribution which is maximised for @xmath21 km / s .
we thus use a systemic velocity of @xmath22 km / s throughout this letter , based on the mean of the h@xmath13 and h@xmath20 values .
our systemic velocity , based on the radial velocity curve of the mass donor star , is in close agreement with the values determined from the disc emission lines . the final doppler tomogram illustrating the distribution of h@xmath13 emission on august 13th is displayed in figure [ trails ] .
the tomogram was constructed from a regularised fit to the observed line profiles , using maximum entropy regularisation ( marsh 2001 ) .
the secondary star emission maps to a sharp spot with a fwhm of @xmath10130 km / s compared to a resolution element of 36 km / s .
maximum emission occurs at @xmath23 km / s , @xmath24 km / s as derived from a 2d gaussian fit .
if the data was folded on the correct orbital ephemeris , emission from the mass donor should appear on the positive @xmath25-axis , corresponding to the radial velocity ( @xmath4 ) of the mass donor star .
the emission of the secondary thus allows us to calculate the phase of inferior conjunction relative to the photometric ephemeris of p98 .
if we assume that the center of the h@xmath13 emission corresponds to the center of the mass donor we can derive a phase offset of @xmath26 ( @xmath27 in terms of orbital phase ) and an apparent radial velocity amplitude of @xmath28 km / s .
our value for the phase offset appears slightly larger than that of sr based on the disc eclipse during quiescence , but is still within 2 sigma of their value . for comparison ,
the same analysis applied to the doppler maps of the h@xmath20 and @xmath29 emission leads to identical phase offsets for both lines ( @xmath30 ) and radial velocities of @xmath31 km / s ( h@xmath20 ) and @xmath32 km / s ( h@xmath15 ) respectively .
there is no evidence for any secondary star emission in the hei6678 , heii4686 or bowen blend transitions , indicating that the secondary star is exposed to relatively soft ionising radiation .
the quoted uncertainties on these values does not include the systematic errors that affect both @xmath4 and the phase offset because of the unknown distribution of the line emission across the roche lobe ( c.f .
steeghs & casares , 2001 ) .
if the line emission is biased towards either the left or right hemisphere of the lobe , a corresponding bias to the derived phase offset would be introduced .
given that only the front part of the roche lobe is irradiated , and that no intrinsic line emission from the secondary is observed during quiescence , the apparent radial velocity amplitude of the emission @xmath33 will be smaller than the true radial velocity @xmath4 of the secondary
. the observed line emission must originate somewhere between the l1 point and the terminator that separates the irradiated part of the roche lobe from its unirradiated side .
the conservative assumption that all emission originates at the terminator implies that the smallest possible correction between @xmath33 and @xmath4 is around 3% .
this was derived from roche geometry calculations across the allowed mass ratio range .
thus our detection of h@xmath13 emission at 493 km / s implies @xmath34 km / s . on the other hand
, the observed velocities can not be smaller than the velocity of the l1 point , which leads to an upper limit of @xmath35 km / s .
here we have again allowed for a wide range of mass ratios consistent with @xmath36 .
armed with a good estimate for the radial velocity amplitude of the secondary star , the mass ratio @xmath37 of the binary can be determined if the radial velocity of the primary ( @xmath2 ) is also known .
gilliland et al .
( 1986 ) obtained an estimate of @xmath38 km / s from the radial velocities of both h@xmath13 line wings and peaks .
however , the radial velocity curves show phase offsets with respect to the absolute ephemeris which indicate that these radial velocity curves must be severely distorted .
this is a common situation in cvs , and may not be surprising given the strong bright spot emission that is present in wz sge during quiescence .
mason et al .
( 2000 ) also measured emission line velocities using a wide range of spectral lines in the optical and infrared regime .
they found velocity amplitudes between 46 and 121 km / s and large phase offsets depending on the excitation potentials of the lines .
they concluded that a varying degree of bright spot contamination distorts the radial velocity curves of the emission lines .
sr used a different approach and attempted to find the center of symmetry of the disc emission in the doppler map at a given velocity , ignoring areas that are affected by the bright spot .
they found that at large velocities the center of symmetry seems to converge on @xmath39 km / s , even though the phase offset is still considerable ( @xmath40 ) .
we applied a similar method to the outburst h@xmath13 tomogram , and find a convergence to a center of symmetry at @xmath41 km / s at high velocities ( 1200 - 1500 km / s ) before noise starts to dominate .
the optimal center of symmetry , like in the case of sr , is offset from the expected position of the white dwarf corresponding to a phase shift of @xmath42 .
if we force the center of symmetry to be phased with the white dwarf while minimising the residuals at areas not affected by the bright spot , we find @xmath43 km / s .
although the formal uncertainty of this optimal center of symmetry is only a few km / s , our methods may be affected by systematic errors due to the fact that we are relying on a complicated emission structure to reflect the motion of the white dwarf . as indicated by our doppler images ,
the accretion flow is highly asymmetric , and a significant amount of distortion may be expected .
we therefore consider 37 km / s to be an upper limit to the true radial velocity amplitude of the white dwarf .
the bottom - right panel of figure [ trails ] plots the asymmetric part of the h@xmath13 emission , after the symmetric part with respect to the optimal center of symmetry was subtracted .
significant asymmetries are clearly present , and the resemblance between our outburst map and the quiescent @xmath44 doppler map of sr is both striking and surprising .
it appears a substantial contribution to the line flux originates from the bright spot region .
it has been proposed in the past ( smak 1996 ; hameury , lasota & hure , 1997 ) , that heating of the secondary during outburst may lead to an increase in the mass transfer rate and thereby prolong the outburst duration .
we refrain from speculating about the nature of the disc asymmetries until a more thorough comparison with other outburst tomography throughout the 2001 campaign can be made .
white dwarf mass estimates have led to a wide range of published white dwarf masses in wz sge ranging from 0.3@xmath45 to 1.2@xmath45 .
our lower limit to the radial velocity of the secondary star ( @xmath34 km / s ) leads to a mass function of ; @xmath46 thus the low white dwarf mass values of , for example , smak ( 1993 ) , rnp and cheng et al .
( 1997 ) , are not compatible with our @xmath4 measurements , and a more massive white dwarf is required . with @xmath47
km / s and @xmath48 km / s we have @xmath49 implying @xmath50 since @xmath51 .
thus , a non - degenerate secondary star is formally not yet ruled out .
however , the lack of any contribution of the mass donor to the j and k bands ( littlefair et al .
2000 ) is difficult to reconcile with a late main sequence mass donor around @xmath52 . even for highly evolved main sequence stars ,
the predicted j and k band magnitudes are significantly too bright ( e.g. leggett et al .
if , as expected for a system that has evolved through the period minimum , the secondary star is in fact a degenerate star with @xmath53 , @xmath2 must be less than @xmath54 km / s .
this clearly illustrates the need for an accurate determination of the radial velocity of the primary in wz sge .
given that the accretion flow is clearly asymmetric both during quiescence as well as outburst , this may only be possible through the use of photospheric white dwarf line velocities .
cheng et al .
( 1997 ) did not detect a systematic velocity shift in their hst data of wz sge .
the measured phase offset ( @xmath55 ) between inferior conjunction of the secondary and bright spot eclipse provides another constraint to the allowed mass ratio range .
we calculated gas stream trajectories in order to determine the predicted @xmath55 as a function of mass ratio and disc radius .
allowing for the uncertainty in disc radius measurements , we can then rule out mass ratios smaller than @xmath56 since the predicted phase offset would be larger than 0.049 compared to our value of @xmath57 , and @xmath58 as derived by sr .
we have detected balmer emission originating from the irradiated mass donor in the cv wz sge during the second and third weeks of its 2001 outburst .
this is the first time a direct detection of the low mass secondary in wz sge has been made .
the doppler maps of wz sge on august 13th are markedly different from those in the first few days of the outburst .
orbit resolved spectroscopy on july 28th , only 5 days into the outburst revealed an accretion disc dominated by two spiral arms ( steeghs et al .
2001 ) , and no sign of any secondary star emission in either h@xmath20 , hei or heii ( h@xmath13 was not observed ) . by august 13th , not only is the secondary star present in emission , the accretion flow also has made a major transition .
the disc emission is dominated by a strong extended bright spot , very similar to its quiescent structure even though the system is still 5 magnitudes brighter than its quiescent level .
the implications of this in terms of varying mass transfer and bright spot contribution throughout the 2001 outburst will be pursued in a future paper .
a reliable determination of the component masses in wz sge awaits an accurate determination of the radial velocity of the white dwarf .
accounting for the possible systematic errors affecting both @xmath2 and @xmath4 measurements , we conservatively derive @xmath59 km / s , @xmath60 km / s and @xmath61 . in terms of component masses , this corresponds to @xmath62 , while @xmath63 .
ds is supported by a pparc fellowship .
the william herschel and isaac newton telescopes are operated on the island of la palma by the isaac newton group in the spanish observatorio del roque de los muchachos of the instituto de astrofisica de canarias .
many thanks to the isaac newton group staff for obtaining part of the data through the ing service programme .
we would also like to thank the many collegues and amateur observers who are contributing to the wz sge campaign during its 2001 outburst .
cheng , f.h . ,
sion , e.m . ,
szkody , p. , huang , m. , 1997 , apjl , 484 , 149 gilliland , r.l .
, kemper , e. , suntzeff , n. , 1986 , apj , 301 , 252 hameury , j.m . , lasota , j.p . ,
hure , j.m . , 1997 , mnras , 287 , 937 harlaftis , e.t . , horne k. , filippenko , a.v . , 1996 , pasp , 108 , 762 ishioka , r. , uemura , m. , matsumoto , k.t . , yamaoka , h. , 2001 , iauc 7669 kolb , u. , baraffe , i. , 1999 , mnras 309 , 1034 leggett , s. k. , allard , f. , geballe , t. r. , hauschildt , p. h. , schweitzer , a. , 2001 , apj , 548 , 908l littlefair , s.p . ,
dhillon , v.s . ,
howell , s.b . ,
ciardi , d.r . , 2000 ,
mnras , 313 , 117 marsh , t.r . , 1989 , pasp , 101 , 1032 marsh , t.r . , 2001 , in : astrotomography , lecture notes in physics series volume 573 , springer verlag marsh , t. , horne k. , 1988 , mnras , 235 , 269 marsh , t. , horne k. , 1990 , apj , 349 , 593 mason , e. , skidmore , w. , howell , s.b . ,
ciardi , d.r . ,
littlefair , s. , dhillon , v.s . , 2000 ,
mnras , 318 , 440 patterson , j. , 1998 , pasp 110 , 1132 patterson , j. , richman , h. , kemp , j , mukai , k. , 1998 , pasp , 110 , 403 robinson , e.l .
, nather r.e . ,
patterson , j. , 1978 , apj , 219 , 168 skidmore , w. , mason , e. , howell , s.b . , ciardi , d.r .
, littlefair , s. , dhillon , v.s .
, 2000 , mnras , 318 , 429 smak , j. , 1993 , act.astr . , 43 , 101 smak , j. , 1996 , in : evans , a . , wood.j , eds , iau coll . 185 , cataclysmic variables and related objects , kluwer spruit , h. , rutten , r.g.m . , 1998 , mnras , 299 , 768 steeghs , d. , casares , j. , 2001 , apj , in press [ astro - ph/0107343 ] steeghs , d. , marsh , t. , kuulkers , e. , skidmore , w. , 2001 , iauc 7675 | we present the first detection of the mass donor star in the cataclysmic variable wz sge .
phase resolved spectroscopy reveals narrow balmer emission components from the irradiated secondary star during the 2001 outburst .
its radial velocity curve indicates a systemic velocity of @xmath0 km / s and an apparent velocity amplitude of @xmath1 km / s .
doppler tomography reveals a highly asymmetric accretion disc including a significant bright spot contribution 20 days into the outburst .
we estimate the primary radial velocity @xmath2 using a center of symmetry method and find @xmath3 km / s .
accounting for the likely systematic errors affecting both @xmath2 and @xmath4 measurements , we conservatively derive @xmath5 km / s and @xmath6 km / s .
this implies a massive white dwarf with @xmath7 . a non - degenerate mass donor , implying wz sge has not yet evolved through its minimum orbital period ,
is not ruled out by our observations .
this would require an improved estimate of @xmath2 .
together with the measured phase offset between bright spot eclipse and inferior conjuction of the secondary star , we can bracket the allowed mass ratio ( @xmath8 ) to lie between 0.040 and 0.073 .
this provides a firm upper limit to the mass of the secondary of @xmath9 . |
the formation of massive stars , due to their large chemical and energetic feedback , is an important issue not just for the field of star formation but also for many other areas of astronomy from the interstellar medium to galaxy formation .
the difficulties in forming massive stars are two - fold .
one , how does that much mass get accumulated by one star , and two , can the accumulation of this matter overcome the large radiation pressure from the star .
the formation of low - mass stars appears trivial in comparison as the stellar masses are not vastly different from estimates of the jeans mass in molecular clouds and as the stellar luminosities are not sufficiently large to affect the accretion process .
massive stars , in contrast , have masses up to 100 times the estimated jeans mass in molecular clouds .
this implies that the majority of their mass is probably due to subsequent accumulation rather than the initial collapse ( see a. maeder , this volume ) .
of potentially more of an impediment is the radiation pressure from stars more massive then @xmath2 which can act on the infalling dust grains to reverse the infall and halt the accretion process ( yorke & krugel 1977 ; yorke 1993 ; beech & mitalas 1994 ) .
there are possible solutions to this impediment involving changing the properties of the dust grains ( wolfire & cassinelli 1987 ) or perhaps a non - isotropic radiation flux such that matter can more easily accrete onto the star s equator from an accretion disc ( c.f .
h. yorke , this volume ) . in this paper ,
i consideran alternative model whereby the accumulation of mass comes from stellar collisions .
the accumulation of such optically - thick , dust - free entities as stars elliminates any problem due to radiation pressure . in constructing a theory for massive star formation
, we should make use of all the available observational clues .
one such clue is that massive stars are not generally found in isolation but appear to form in the central regions of rich young stellar clusters ( zinnecker , mccaughrean & wilking 1993 ; hillenbrand 1997 ; clarke , bonnell & hillenbrand 2000 ; mermilliod 2001 ) .
even runaway ob stars are probably most easily explained as having been ejected from such a stellar cluster ( clarke & pringle 1992 ) .
testi ( 1997 ) have argued that there is a strong correlation between the mass of the most massive star in a cluster and the cluster s mean stellar density .
although this data is still compatible with random pairing from an imf ( bonnell & clarke 1999 ) , such a correlation would imply a potential causal relationship between cluster density and the mass of the most massive star .
in considering the clustered environment for massive star formation , there are a number of observations that need to be considered .
in addition to the large number of accompanying low - mass stars , the relatively large stellar densities and that the mass spectrum is similar to a field star imf , one feature that needs to be explained is the observed mass segregation ( hillenbrand 1997 ; carpenter 1997 ; hillenbrand & hartmann 1998 ) . for example , the orion nebula cluster ( onc ) shows a segregation of massive stars towards the centre of the cluster .
this appears to be the case not only for the most massive o stars but also for the intermediate - mass stars present in the cluster .
lastly , but perhaps most importantly , the fact that massive stars form in stellar clusters implies that we need to consider a dynamical formation scenario .
the dynamical time of the cluster as a whole is comparable to the formation time of the individual stars .
that implies that the stars move around in the cluster while they are forming .
we have to thus abandon a static formation scenario where the environment can be ignored .
although this causes significant complications , it can also provide simple physical processes that combine to determine the resultant stellar properties .
the most straight forward solution for massive star formation would be that a massive clump of @xmath3 collapses to form the star .
this scenario is complicated by the dense cluster environment .
as the massive stars are located in the cluster core , the pre - stellar clump would have to fit in between the neighbouring stars , which , in the core of the onc are separated by @xmath4 au . in order for the clump
to be gravitationally bound , it must have a jeans length smaller than this separation : r_j = ( ) ^1/2 ( ) ^-1/2 5000 a.u .
, where @xmath5 is the gas density and @xmath6 the temperature , @xmath7 is the gas constant , @xmath8 is the gravitational constant , and @xmath9 is the mean molecular weight .
this implies that for typical temperatures in molecular clouds , the gas density must be very high .
this also makes sense as one would expect the highest gas densities in the deepest part of the potential well ( zinnecker 1993 ) .
the mass of the resultant star can then be estimated from the jeans mass , the minimum mass to be gravitationally bound : [ jeans_mass ] m_j = ( ) ^3/2 ( ) ^-1/2 . in the core of the onc ,
the jeans mass is then of order @xmath10 , much smaller then the most massive star of @xmath11 , or the mean stellar mass of @xmath12 ( zinnecker 1993 ; bonnell , bate & zinnecker 1998 ) .
something else is needed in order to explain the observed masses .
furthermore , from the above equations it is obvious that the fragment mass should be smallest in the centre of clusters as the gas density is likely to be largest there .
the initial masses from fragmentation in a stellar cluster should therefore result in an inverse mass segregation , with small stellar masses in the core ( bonnell , bate & zinnecker 1998 ) .
investigations of the dynamics of young stellar clusters have shown that dynamical mass segregation , the sinking of massive stars due to two - body relaxation , can not explain the location of the massive stars ( bonnell & davies 1998 ) .
clusters such as the onc are too young for the massive stars that form the trapezium to have sunk from a significant distance from the cluster core .
we are thus left with a formation process that occurs in situ but that does not reflect the initial fragment or jeans mass .
if stellar collisions are responsible for the formation of massive stars ( eg .
bonnell , bate & zinnecker 1998 ) , then the timescale for multiple collisions has to be less than the observed age of the massive stars .
although it is difficult to ascertain the exact ages of massive stars , we know that in orion , the massive stars are younger than a few million years and potentially as young as a few @xmath13 years .
the collisional timescale @xmath14 is related to the stellar properties as ( binney & tremaine 1987 ) @xmath15 where @xmath16 is the density of stars in the cluster , @xmath17 is the velocity dispersion , and @xmath18 are the mass and radius of the star undergoing the collision .
figure [ fig1 ] plots the collisional timescale as a function of stellar density for collisions involving a 10 star assuming a 10 collisional radius .
the velocity dispersion is assumed to be 2 km / s .
we can see from figure [ fig1 ] that for densities typical of young stellar clusters ( @xmath19 stars pc@xmath20 ) , the collisional timescale is @xmath21 years . in order to get a collisional timescale of order @xmath22 years , a stellar density of @xmath23 stars pc@xmath20 is required .
young stellar clusters are relatively dense agglomerations of stars .
the orion nebula cluster ( onc ) , for example , has a mean stellar density of @xmath24 stars pc@xmath20 ( hillenbrand 1997 ) .
the core of the onc is considerably more dense with @xmath25 stars pc@xmath20 ( mccaughrean & stauffer 1994 ) . from figure 1
we see that the corresponding collisional timescale is @xmath26 years .
clearly collisions can not play an important role in the present conditions of the onc .
if stellar mergers are an important process in massive star formation , then the required stellar density of @xmath27 stars pc@xmath20 implies that the onc would have had to undergo a high density phase .
the main question is then what process could have lead to such a short - lived but high - density phase . in the following sections we consider how accretion in stellar clusters may lead to such a scenario .
observations of young stellar clusters reveal that they generally contain significant amounts of mass in the form of gas .
this gas can amount up to 90 % of the total cluster mass ( lada 1991 ) .
accretion of this gas can thus be an important factor in determining the masses of individual stars as well as the properties of the stellar cluster .
numerical simulations of accretion in clusters have found that accretion naturally results in a mass segregated cluster ( bonnell 1997 , 2001a ) .
this occurs as the cluster potential funnels the gas down to the centre of the cluster where it is accreted by whatever stars happen to be there .
the non - uniform accretion rates then result in the most massive stars being located in the centre of the cluster .
in general , the resulting cluster demonstrates a degree of mass segregation comparable to that found in young stellar clusters .
another finding of these numerical simulations is that there are two different physical regimes for the accretion , depending on whether the gas or the stars dominate the gravitational potential ( bonnell 2001a ) .
when the gas dominates the potential , the motions of the stars and the gas are determined by the changing gravitational potential as the gas collapses . in this regime ,
the stellar and gas motions are similar and the accretion rate is determined by each star s tidal radius relative to the cluster potential .
material that moves inside the tidal radius gets accreted by the star .
this assumes that the gas is initially unsupported , as is expected if it is able to fragment to form the large number of stars in the cluster .
the second regime occurs when the stars dominate the gravitational potential . in this case , the stars virialise and have motions uncorrelated to those of the gas .
the gas velocity relative to a star is therefore large such that it is the determining factor as to whether the gas is bound to the star .
the accretion rate in this case is the common bondi - hoyle accretion which depends on the local gas velocity and the star s mass , but not on the external potential .
using these two different regimes for the gas accretion , bonnell ( 2001b ) showed how accretion in stellar clusters can result in a double power - law imf where the low - mass stars have a shallow @xmath28 ( where salpeter is @xmath29 ) power - law due to the tidal lobe accretion while the high - mass stars have a steeper power - law ( @xmath30 ) due to the subsequent bondi - hoyle accretion .
the high - mass stars accrete the majority of their mass in the stellar dominated regime as they form in the core of the cluster where the stars ( due to their high accretion rates ) first come to dominate the potential .
the simulations in the previous section also found that accreting clusters tend to shrink as mass is added onto the stars ( bonnell 2001a ) . in order to understand why this happens let us consider the energy of the stars in the cluster potential ( c.f .
bonnell , bate zinnecker 1998 ) : @xmath31 where @xmath32 is the total momentum of the stars , @xmath33 is the mass in stars , @xmath34 is the total mass in both stars and gas while @xmath35 is the size of the system . in what follows
we assume that the accreting gas has no net momentum and that therefore the stellar momenta is conserved during accretion .
this assumption is justified as in general the accreting gas will be spherically infalling and , once the stars are virialised , there is no correlation between gas motions and stellar motions .
as mass is added while momentum is conserved , the kinetic energy of the stars decreases at the same time as the stars become more bound to the potential .
the stars are then no longer in virial equilibrium with the potential and thus shrink until they revirialise at a smaller radius .
the new radius reflects the new , lower energy as : @xmath36 as long as the accretion timescale is longer than the dynamical timescale , then the cluster can accrete and revirialise many times and thus shrink substantially . if , on the other hand , the accretion timescale is shorter than the dynamical timescale , then the cluster can only shrink by a factor of 2 after the accretion has finished . under the condition of slow accretion
, the rate of cluster shrinkage can be estimated by considering that accretion occurs at a constant radius and then shrinks to revirialise , reflecting the new energy .
thus @xmath37 solving the above equation yields ( bonnell 1998 ) @xmath38 where the value of @xmath39 depends on whether the total mass is conserved ( @xmath40 or whether it is proportional to the mass in stars ( @xmath41 ) .
the latter possibility reflects the case when accretion occurs onto the core of the cluster and gas continues to infall from outside the core .
thus , the increase in the stellar density can be very large , given by @xmath42 such that for @xmath41 and an increase in the stellar mass by a factor of three results in an increase in the density as @xmath43 .
we therefore take as our model the core of a cluster similar to the onc .
this core , containing @xmath44 initially low - mass stars within @xmath45 pc .
the stars in the core accrete until they have reached intermediate masses ( @xmath46 ) by which time they have attained a sufficiently high density that stellar collisions and mergers occur to form the massive stars .
figure [ crittime ] shows the evolution of such a system as the stars accrete at a constant accretion rate of @xmath47 year@xmath48(from bonnell 1998 ) .
the half - mass radius of the system is plotted as a function of time as are the collisional timescale and the accretion timescale .
once the cluster core has shrunk sufficiently such that the collisional timescale is shorter than the accretion timescale , stellar mergers dominate the mass buildup process .
two collisional timescales are plotted reflecting collisional radii of @xmath49 and @xmath50 au .
the transition from accretion to mergers occurs after a few crossing times , corresponding roughly to @xmath51 years .
the actual cross sections for collisions depends on the dynamics and the presence of circumstellar material .
one of the crucial factors is the formation of binaries as they then have a much larger cross section for further interactions .
there are two ways in which binary systems can form .
firstly , they can form through tidal capture encounters .
this is most important for near - equal mass encounters .
most encounters involving the more massive stars in the core are with lower - mass stars
. these encounters do not generally result in tidal captures due to the disparate interior densities of the stars . instead
, the lower - mass star can be tidally disrupted to form a circumstellar disc .
this results in the second binary formation mechanism as a subsequent encounter that passes through the disc gets captured due to the binding energy of the disc ( davies , in preparation ) .
such binary systems will either merge on subsequent encounters or exchange in a higher - mass star to form a near equal - mass binary , that may itself merge in further encounters . in either case , a high frequency of near - equal mass binary systems should result from a system where stellar collisions are occurring .
this agrees with the finding that many o stars are members of short - period binary systems ( mason 1996 ) .
this is also the case in o - stars found in the centre of open clusters where they all appear to be in close binary systems with stars of comparable masses ( mermilliod 2001 ) .
in order to study the dynamics and evolution of an accreting cluster we have performed a numerical simulation of gas accretion in a cluster containing 1000 stars ( bonnell & bate , in preparation ) .
this simulation , performed on the ukaff supercomputer with @xmath52 sph particles , follows the gas and stellar dynamics as the stars accrete .
the primary aims of this simulation were to investigate whether an actual stellar and gas dynamical system would react according to our simple prescription and to see if the stellar interactions would destroy the cluster core before the necessary densities for collisions occurs .
the maximum density in the cluster , as well as the density at the half - mass radius are plotted as a function of time in figure [ clusdenevol ] .
we see that the density increases dramatically after about one free - fall time . the mean density of the cluster increases by a factor of @xmath53 while the peak density increases by over @xmath54 up to @xmath0 .
this large increase in stellar density implies a large decrease in the timescale for collisions and mergers .
if we take the cluster s initial conditions as being those present in the onc , then we have an initial mean density of @xmath55 stars pc@xmath20 and a peak density of @xmath56 stars pc@xmath20 ( mccaughrean & stauffer 1994 ) . in the ukaff simulation the peak density increases by a factor of @xmath57 times the initial mean density .
this then corresponds to stellar densities of @xmath58 stars pc@xmath20 .
this stellar density is sufficient to produce significant numbers of stellar collisions in the cluster core .
accretion results in a centrally condensed cluster with a @xmath59 stellar density profile .
figure [ stellrad ] plots the distribution of stellar densities calculated as the volume necessary to contain 10 stars .
the maximum density is considerably higher than the mean density of @xmath60 , but is contained in a core which occupies only the innermost @xmath61 of the half - mass radius . in the ukaff simulation , collisions
are assumed to occur when two stars pass within 1 au of each other .
this value is much larger than the stellar radii but comparable to the binary separations considered above . in any case , as shown in bonnell ( 1998 ) , the actual collisional radius affects only the delay until the cluster has shrunk sufficiently for collisions . using the above collisional radius ,
15 collisions occur during the ukaff simulation .
about half of these involve the more massive stars including a near equal - mass collision which results in the most massive star having a mass of @xmath44 times the initial stellar masses .
we thus can conclude that an accreting cluster can contract far enough such that significant stellar collisions occur which result in the formation of massive stars .
the evolution of the mean and maximum masses during the simulation are plotted in figure [ maxmassevol ] .
we see that the maximum mass increases dramatically through both accretion ( smooth increase ) and collisions ( jumps ) .
as noted above , the maximum mass reaches a value @xmath44 times the initial stellar mass . at the same time the mean stellar mass only increases by @xmath62% during the evolution .
many of the mergers involving the most massive star involve lower - mass impactors as is expected .
two mergers near the end of the simulation involve near - equal mass components and thus significantly contribute to the mass .
the mean stellar mass in the cluster core is also plotted in figure [ maxmassevol ] .
this mean mass in the core increases dramatically but is only half due to the most massive star .
thus the cluster has been mass segregated due to the accretion and collisions .
it is also worth noting that the cluster stars tend to be in small groups , and along filaments ( see figure [ gasstreams ] below ) .
these configurations increase collisional rates and thus the ability for collisions to form massive stars . furthermore , if the absence of collisions , the maximum stellar densities would have been considerably higher .
massive stars are generally associated with energetic outflows .
these outflows are often seen as evidence for a circumstellar disc which would most likely be destroyed in regions sufficiently dense to allow for stellar mergers .
the implication of a stellar disc for these outflows is not at all certain as the outflows are generally poorly collimated and thus of a fundamentally different type than the well - collimated outflows from low - mass stars .
an alternative to disc collimation is that there is present some structure in the accretion flow or general circumstellar environment which tends to collimate ( poorly ) any outflow from the massive stars .
figure[gasstreams ] shows the configuration of the stars and the gas towards the end of the ukaff run .
the gas ( and stars ) are far from uniformly distributed . instead , the gas displays filamentary shapes pointing towards the cluster centre . the stars are generally situated along the filaments and tend to be in small groups . both the gas and the stars move along these filaments which increases the rate of collisions . the filamentary structures present in the accretion flows in the cluster centre are due to gravitational instabilities .
the gas initially contains some 1000 jeans mass , in agreement with its previous fragmentation to form the 1000 stars .
in such a system , gravity tends to increase any over densities and to reduce the dimensions of the structure . in this way
, the over densities can grow and form into filamentary structures .
such structures are a possible source for the collimation of the massive stars outflows .
further numerical simulations will have to be performed in order to test this speculation .
it should be noted , however , that any collimation that does result from this structure is likely to be fairly open and as such this mechanism is not likely to result in the formation of jet - like outflows .
any such highly collimated outflows around young o stars would thus be difficult to explain with this stellar merger mechanism .
there are many implications of this collisional formation mechanism for massive stars which need to be considered .
the most significant of these is the effect of the energy involved in the collisions .
as two @xmath63 stars collide , the energy involved is of order @xmath64 ergs .
this much energy will puff up the stars which will contract on their kelvin - helmhotz timescale .
the increased size of the star will make it more susceptible to further collisions .
if we assume that the energy is radiated over some @xmath54 years , then the energy flux will be @xmath65 ergs s@xmath48 .
indeed , a fraction of the kinetic energy may be released on much shorter timescales and result in an x - ray flare .
such an event should be observable ( see chapters by h. zinnecker and j. bally ) .
it is alkso probable that the energy released in this way would contribute to the re - expansion of the cluster core .
the formation of massive stars is a complicated process in which the environment is likely to play a crucial role .
massive stars form in the cores of stellar clusters where the expected fragment mass is low .
accretion in clusters forms higher - mass stars from low mass stars and naturally results in a mass segregated cluster as the higher accretion rates in the core result in more massive stars there .
this process also results in a two - slope imf similar to that observed in stellar clusters and in the field . the accretion onto the stars in the cluster forces it to shrink and revirialise at a smaller radius .
if the accretion is slower or comparable to the local dynamical time ( say of the cluster core ) , then accretion can force the core to continue shrinking until it has reached sufficient densities where stellar collisions occur .
these collisions can then result in the formation of massive stars .
a recent simulation using the ukaff supercomputer has shown that an accreting cluster can indeed attain the required stellar densities and that mergers can play an important role in the formation of massive stars .
the accretion flow is susceptible to the formation of filamentary structure which may play a crucial role in collimating any outflow from massive stars .
it would be unlikely to be able to collimate a jet - like outflow .
the results presented here are partially based on numerical simulations performed with the united kigndom s astrophysical fluid facility ( ukaff ) supercomputer . | in this review , i present the case for how massive stars may form through stellar collisions .
this mechanism requires very high stellar densities , up to 4 orders of magnitude higher than are observed in the cores of dense young clusters . in this model ,
the required stellar densities arise due to gas accretion onto stars in the cluster core , including the precursers of the massive stars .
this forces the core to contract until the stellar densities are sufficiently high for collisions to occur .
gas accretion is also likely to play a major role in determining the distribution of stellar masses in the cluster as well as the observed mass segregation .
one of the main advantages of this mechanism is that it explicitly relies on the cluster environment in order to produce the massive stars .
it is thus in a position to explain the relation between clustered and massive star formation which is not an obvious outcome of an isolated accretion mechanism .
a recent numerical simulation supports this model as the cluster core increases its density by @xmath0 during gas accretion .
approximately 15 stellar collisions occur ( with @xmath1 au ) in the cluster core , making a significant contribution to the mass of the most massive star . |
fractional laplacian operators are usually considered as the prototype of the non - local operators @xcite and have recently garnered much attention in many applications in mathematics and physics .
related problems usually lie at the interface of probability , stochastic processes , partial differential equations and spectral theory such as @xcite . for some applications ,
we refer the reader to graphene models @xcite , obstacle problems @xcite , non - local minimal surfaces @xcite . in this article
, we study eigenvalues of the fractional laplacian operator @xmath6 defined by @xmath7 where @xmath8 is a bounded connected domain with smooth boundary in @xmath9 , for either @xmath2 @xmath10 or @xmath11 and @xmath12 since @xmath8 is bounded , the spectrum of the fractional laplacian is discrete and eigenvalues @xmath13 ( including multiplicities ) can be sorted in an increasing order .
unlike laplacian , fractional laplacian is a non - local operator such that for suitable test functions , including all functions in @xmath14 , it is defined as @xmath15 where @xmath16 is a well - known positive normalizing constant . from a probabilistic point of view , @xmath6 can be considered as the infinitesimal generator of the semigroup of the symmetric @xmath17stable process , denoted by @xmath18 , with the characteristic function @xmath19 where @xmath20 is called the transition density of the stable process ( or the heat kernel of the fractional laplacian ) . while explicit formulae for the transition density of symmetric @xmath21-stable processes are only available for the cauchy process ( @xmath22 ) and the brownian motion ( @xmath23 ) , these processes share many of the basic properties of the brownian motion .
another process of importance is the holtsmark distribution ( @xmath24 ) which is used to model gravitational fields of stars ( see e.g. , @xcite ) .
stable processes do not have continuous paths which is related to non - locality of the fractional laplacian operator @xcite .
when fractional laplacians involved , some of the known methods fail because of the fractional power and non - locality of such operators .
this drawback can be evaded by using the fourier transform definition .
recall that the fourier transform and its inverse are defined as @xmath25(\mu ) = \hat{u}(\mu)=c_d\int_{{\mathbb r}^d}e^{-i\mu\cdot z}\,u(z)\,dz , \quad{\mathcal f}^{-1}[u](z)=c_d\int_{{\mathbb r}^d}e^{i\mu\cdot z}\,u(\mu)\,d\mu,\ ] ] where @xmath26 is the normalizing constant .
interestingly , the fractional laplacian operator on @xmath1 can be defined as a pseudo - differential operator as @xmath27\right],\qquad 0<\alpha \le2 , \quad u\in h_0^{\alpha/2}(\omega).\ ] ] here , @xmath28 denotes the sobolev space of order @xmath29 .
when @xmath30 , one can look at proposition 3.3 .
@xcite for the proof of the equivalence between the definitions in and .
there is an extensive literature devoted to the inequalities involving the eigenvalues of the dirichlet laplacian operator , which can be regarded as the fractional laplacian when @xmath23 .
one may consult the articles @xcite and references therein for a through literature review .
it is worth pausing here for a moment to consider the dirichlet laplacian results relevant to our main result .
the first such result is the li - yau inequality that provides a lower bound for the sums of eigenvalues , sharp in the sense of weyl asymptotics @xcite .
the authors generalized this result in @xcite by obtaining the following li - yau type inequality for the eigenvalues of the diriclet fractional laplacian operator : @xmath31 where @xmath32 represents the volume of @xmath8 and @xmath33 denotes the gamma function @xmath34 for @xmath35 one may also look at @xcite for a li - yau type inequality involving the eigenvalues of the ( massless ) klein gordon square root operators @xmath36 , ( i.e. , the case @xmath22 ) . to look at this inequality from a different perspective , one may take the legendre transform of the following result by laptev @xcite and obtain : @xmath37 when we set @xmath23 in , we recover an earlier result by berezin @xcite , which supplies the li - yau inequality after an application of the legendre transform .
thus , in what follows , we call as the berezin - li - yau inequality . in @xcite , melas proved the following improvement to the berezin - li - yau inequality ( @xmath23 ) : @xmath38 where @xmath39 , the moment of inertia , is defined by @xmath40 by a translation of the origin and a rotation of axes if necessary , in the sequel , we assume that the origin is the center of mass of @xmath8 and that @xmath41 melas type bounds and their many variants and extensions have recently attracted a lot of attention , see for instance @xcite . in particular , in @xcite , the authors obtained a refinement of , stating that @xmath42 remark that @xmath23 in recovers melas bound in @xcite for the dirichlet laplacian eigenvalues .
unfortunately , it is not easy to take legendre transform of to find a similar improved berezin type bound in the case of the fractional laplacian . for @xmath43 , authors also proved in @xcite that @xmath44 the main purpose of this article is to obtain analogous but sharper bounds with dirichlet laplacian replaced by dirichlet fractional laplacian .
the first step , inspired by results in @xcite , is establishing the following result : [ impbly ] for @xmath45 , and either @xmath46 and @xmath47 or @xmath48 and @xmath49 the eigenvalues @xmath13 of the fractional laplacian operator defined on @xmath50 satisfy @xmath51 note that the constants in the leading terms on the right side of , which is a fractional version of weyl s law , are optimal due to a classical result of blumenthal and getoor @xcite .
while the constant in the second term in is still sub - optimal , the estimate in is substantially stronger than previous known results in @xcite .
in addition , we recover the lower bounds in the case of the dirichlet laplacian @xcite when we set @xmath23 in .
in @xcite , chen and song obtained that @xmath52 for each @xmath53 and any constant @xmath54 $ ] .
thus , proposition [ impbly ] along with an application of leads to our principal result : [ impblyt ] for @xmath45 , @xmath55 and either @xmath46 and @xmath47 or @xmath48 and @xmath49 the eigenvalues @xmath13 of the fractional laplacian operator defined on @xmath50 satisfy @xmath56 on a side note , the proof of proposition [ impbly ] consists of a very delicate application of steffensen s type inequalities , which is mainly about the comparison of the integrals on the subsets of interval @xmath57 .
the article is structured as follows : in section [ sec : bly2 ] , we revisit the relevant facts about the eigenvalues and eigenfunctions of the fractional laplacian operator . after providing the proof of an auxilliary lemma that plays a crucial role in proving proposition [ impbly ] , we present the proof of our main results in section [ sec : impbly ] . finally , we end section [ sec : impbly ] with a remark which extends the main result even further for some elliptic operators studied in @xcite .
please see remark [ rmkgen ] for details .
in this section , we give an overview of the definitions and tools that are essential to establish the estimates in . even though , these were previously studied in @xcite , we include them so that the article is self - contained . by using plancherel s theorem , one can show that the set of fourier transforms @xmath58 of @xmath59 forms an orthonormal set in @xmath60 since the set of eigenfunctions @xmath59 is an orthonormal set in @xmath61 to ease the notation , we set @xmath62 because the support of @xmath63 is @xmath64 the integral is taken over @xmath8 instead of @xmath65 .
interchanging the sum and integral and using @xmath66 we derive @xmath67 observe that @xmath68\rangle = \sum_{j=1}^k \int_{{\mathbb r}^d}|\mu|^{\alpha } \,|\hat{\phi}_j(\mu)|^2\,d\mu \\ & = \int_{\mathbb{r}^d } |\mu|^{\alpha } \ , \phi_k(\mu)\ , d\mu .
\end{split}\ ] ] application of bessel s inequality gives an upper bound for @xmath69 : @xmath70 next , we find an estimate for @xmath71 .
note that @xmath72 in view of h " older s inequality and utilizing and , we arrive at the following uniform bound : @xmath73 let @xmath74 designate the ball of radius @xmath75 centered at @xmath76 in @xmath65 and @xmath77 denotes the volume of @xmath78 dimensional unit ball @xmath79 in @xmath65 given by @xmath80 now assume that @xmath75 is such that @xmath81 .
that is , @xmath82 is the symmetric rearrangement of @xmath83 note that @xmath84 roughly giving @xmath85 let @xmath86 denote the decreasing radial rearrangement of @xmath87 there exists a real valued absolutely continuous function @xmath88 $ ] such that @xmath89 by using plya - szeg inequality , one can show that @xmath90 for more details , see for example @xcite .
before diving into the proof of the main results , we present the following surprising sharper inequality which will be the main ingredient in the proof of the sharper lower bound in .
our method of proof has been previously explored in several articles , for instance @xcite , with crucial differences .
* case 1 : * assume that @xmath99 and @xmath100 .
observe that @xmath101 can be written as @xmath102 where @xmath103 differentiating we get @xmath104 @xmath105 note that @xmath106 where @xmath107 when @xmath108 we have @xmath109 , implying that @xmath110 is convex on @xmath111 .
thus , @xmath112 since @xmath113 and @xmath114 .
that is , @xmath115 on @xmath111 .
in particular , @xmath116 on the other hand , when @xmath117 we have @xmath118 yielding that @xmath119 is decreasing on @xmath120 $ ] .
this implies that @xmath121 meaning that @xmath110 is decreasing on @xmath120.$ ] this leads to @xmath122 for @xmath123.$ ] hence , @xmath115 for @xmath124 .
therefore , we deduce that @xmath125 for @xmath126 * case 2 : * now , assume that @xmath47 and @xmath3 . then @xmath101 becomes @xmath127 as before , we can write @xmath102 where @xmath128 note that if @xmath129 then @xmath130 differenting again , we obtain @xmath131 notice that for @xmath3 , @xmath132 meaning that @xmath110 is convex . since @xmath113 and @xmath114 , @xmath133 implies that @xmath115 and , therefore , @xmath134 assume that - hold and @xmath138 consider the decreasing , absolutely continuous function @xmath139 defined by .
we know that @xmath140 for @xmath141 where @xmath142 is given by . since @xmath143
due to let us first define @xmath144 note that @xmath145 is positive , @xmath146 and @xmath147 . to simplify the notation , we also set @xmath148 for @xmath149 .
hence , @xmath150 for @xmath141 and @xmath151 now , set @xmath152 using we get @xmath153 moreover , since the map @xmath154 is radial and increasing , by , we obtain that @xmath155 substitution of into yields @xmath156 observe that application of fubini s theorem together with @xmath157 leads to @xmath158 which together with @xmath159 and @xmath160 respectively yields @xmath161 notice that @xmath162}(a)\big)\ge 0 , \qquad a\in[0,\infty ) .
\label{ineq:00}\ ] ] integrating from @xmath163 to @xmath164 gives @xmath165 where @xmath166 is defined by @xmath167 since @xmath168 is continuous and non - decreasing and @xmath169 as @xmath170 the intermediate value theorem provides us with the existence of @xmath171 such that @xmath172 which , by , concludes that @xmath173 now consider the polynomial @xmath174 where @xmath175 are chosen so that @xmath176 and @xmath177 and @xmath178 remains negative on @xmath179 and positive on @xmath180.$ ] it is immediate to observe that @xmath181}(a)-\theta_k(a)\right)\le 0\quad\mbox{on}\quad [ 0,\infty ) .
\label{ineq:01}\ ] ] integration of on @xmath57 leads to @xmath182 simplifying to @xmath183 using , we infer that @xmath184 observe that @xmath185 notice that gives the key inequality in the proof of this lemma . indeed , integrating in @xmath186 from @xmath187 to @xmath188 we obtain @xmath189 note that @xcite @xmath190 @xmath191 and so , we have @xmath192 for any @xmath193 and @xmath194 since @xmath195 due to , setting @xmath196 and using and , we deduce that yields to @xmath197 equations in combined with and turn into @xmath198 to finish the proof of proposition [ impbly ] we shall minimize the right side of over @xmath199 to prove , we show that the function , which we call @xmath200 , on the right - hand side of with @xmath201 decreases on @xmath202.$ ] by we know that @xmath203 first , define @xmath204 where @xmath205 and @xmath206 thus , it is enough to show that @xmath207 and @xmath208 defined by and are also decreasing on the interval @xmath202 $ ] . differentiating @xmath207 and @xmath209
we observe that @xmath210 and @xmath211 are decreasing on the interval @xmath212 and @xmath213 respectively , where @xmath214 hence , we particularly obtain that @xmath215 is decreasing on @xmath202 $ ] when we have @xmath216 for any @xmath217 since @xmath218 is increasing for @xmath219 we obtain that @xmath220 in view of , @xmath221 and definition of @xmath77 given in we observe that @xmath222^{\frac{2}{d } } \,|\omega|^{\frac{d+1}{d } } k^{\frac{1}{d}}}{(2\pi)^{d+1 } } \right)^{\frac{d}{d+1}}\ge \omega_d,\ ] ] as @xmath45 and @xmath223 similarly , we obtain that @xmath224^{\frac{2}{d}}\,|\omega|^{\frac{d+1}{d } } \,k^{\frac{1}{d}}}{(2\pi)^{d+1}\ , } \right)^{\frac{d}{d+1}}\ge \omega_d,\ ] ] as @xmath45 and @xmath225 in conclusion , we obtain that @xmath226 as @xmath215 is decreasing on @xmath202 $ ] .
substitution of @xmath227 given in together with @xmath228 turns into .
[ rmkgen ] in view of the recent work @xcite , it is worth noting that one can easily extend this for elliptic operators @xmath229 defined by a kernel @xmath230 @xmath231 where @xmath230 satisfies @xmath232 @xmath16 is the normalizing constant in the fractional laplacian definition and @xmath233 to this end , let us consider the eigenvalue problem defined by @xmath234 it is shown in @xcite that the spectrum of @xmath229 is also discrete and the eigenvalues @xmath235 ( including multiplicities ) can be sorted in an increasing order .
also , the set of fourier transforms @xmath58 of @xmath59 forms an orthonormal set in @xmath60 since the set of eigenfunctions @xmath59 is an orthonormal set in @xmath61 note that we use the same notation for eigenvalues and eigenfunctions to illuminate the striking similarities though they might be different for each @xmath236 defining @xmath69 as in , we obtain immediately .
however , needs to be re - written as the following inequality @xmath237 where we used the fact ( proposition 3.3 . in @xcite ) that @xmath238 having in hand , changes as follows : @xmath239 and proceeding exactly as before using in place of , we immediately arrive at the following remarkable estimate : m. ashbaugh , r. benguria and r. laugesen , _
inequalities for the first eigenvalues of the clamped plate and buckling problems _ , general inequalities 7 , international series of numerical mathematics , vol .
123 , c. bandle , w. n. everitt , l. losonczi , and w. walter , eds .
, ( birkhuser verlag , basel , 1997 ) 95110 .
r. el hajj and f. mhats , _ analysis of models for quantum transport of electrons in graphene layers _ , math .
models methods appl .
* 24 * ( 2014 ) 2287 .
e. m. harrell ii and s. yildirim yolcu , _ eigenvalue inequalities for klein - gordon operators _ , j. funct .
analysis , * 256*(12 ) ( 2009 ) 3977 - 3995 .
a. laptev and t. weidl , _ recent results on lieb - thirring inequalities _ , journes quations aux drives partielles ( 2000 ) 114 .
p. li and s .- t .
yau , _ on the schrdinger equation and the eigenvalue problem _ , comm .
, * 88 * ( 1983 ) 309318 .
s. yildirim yolcu and t. yolcu , _ sharper estimates on the eigenvalues of dirichlet fractional laplacian _ , discrete and continuous dynamical systems - series a , * 35*(5 ) ( 2015 ) 22092225
. t. yolcu , _ refined bounds for the eigenvalues of the klein - gordon operator _ , proceedings of the american mathematical society , * 141*(12 ) ( 2013 ) 43054315 . | the purpose of this article is to establish new lower bounds for the sums of powers of eigenvalues of the dirichlet fractional laplacian operator @xmath0 restricted to a bounded domain @xmath1 with @xmath2 @xmath3 and @xmath4 @xmath5 .
our main result yields a sharper lower bound , in the sense of weyl asymptotics , for the berezin - li - yau type inequality improving the previous result in @xcite .
furthermore , we give a result improving the bounds for analogous elliptic operators in @xcite . |
one of the most intriguing predictions of einstein s general relativity is the existence of black holes which are formed from gravitational collapse in the last stage of heavy stars life or in high - density regions of the density perturbations in the early universe @xcite .
the gravitational collapse of an incoherent spherical dust cloud is described by the einstein equation @xmath1 with the energy density @xmath2 and the time - like velocity vector @xmath3 . in a reference frame comoving with the collapsing matter
, it can be proved that the metric of spherically symmetric space - time reads @xcite @xmath4 where @xmath5 is the comoving radial coordinate and @xmath6 the proper time of freely falling shells . as well known , the solution of this metric is the lemaitre - tolman - bondi ( ltb ) solution @xcite , which has been extensively studied not only in spherical collapse , but also in cosmology as well @xcite@xcite .
it is well known that the end state of gravitational collapse is a singularity of space - time curvature with infinite density of matter@xcite@xcite .
but , it is not known clearly whether such a singularity will be naked or covered by a event horizon . at the singularity , the laws of science and the ability to predict the future would break down .
it led roger penrose to propose the ( weak and strong ) cosmic censorship hypothesis ( cch ) @xcite@xcite , which to date remain unproven .
the weak cch asserts that there can be no singularity visible from future null infinity and light from singularity is completely blocked by the event horizon .
it protects external observers from the consequences of the breakdown of predictability that occurs at the singularity , but dose nothing at all to observers who are cofalling with the collapsing massive body , while the strong cch prohibits singularity s visibility by any observer . despite almost 40 years of effort we still do nt have a general proof of cch @xcite . on the contrary
, it was shown that the seminal work of oppenheimer and snyder @xcite is not a typical model and the central singularities formed in generic collapse are naked and observable @xcite@xcite . up to now ,
many quantum theories to unify all fundamental interactions including gravity have been proposed , the most promising candidate among them is superstring / m - theory . in string theory ,
extra dimensions were promoted from an interesting curiosity to a theoretical necessity since superstring theory requires a ten - dimensional space - time to be consistent from the quantum point of view @xcite@xcite .
it caused a renewed interest about the general relativity in more than 4 dimensions .
several ltb - like solutions to einstein equations in higher dimensions have been obtained in recent years @xcite@xcite . on the other hand , in recent years
another renewed interest has grown in higher order gravity , which involves higher derivative curvature terms . among the higher curvature gravities ,
the most extensively studied theory is so - called lovelock gravity @xcite , which naturally emerged when we want to generalize einstein s theory in higher dimension by keeping all characteristics of usual general relativity excepting the linear dependence of riemann tensor .
in addition , lovelock terms naturally occur in the effective low - energy action of superstring theory @xcite@xcite .
hence , the lovelock gravity provides a promising framework to study curvature corrections to the einstein - hilbert action .
since we are interested in what influence to the theory of gravity will be caused by lovelock terms , for simplicity , we only consider pure lovelock gravity and neglect the other terms in the low - energy expansion string theory .
the lagrangian of lovelock theory is the sum of dimensionally extended euler densities @xcite @xmath7 where arbitrary constants @xmath8 are the lovelock coefficients , and @xmath9 is the euler density of a @xmath10-dimensional manifold @xmath11 here the generalized delta function @xmath12 is totally antisymmetric in both sets of indices and @xmath13 is the riemann tensor . though the lagrangian of lovelock gravity consists of some higher derivative curvature terms , its field equations of motion contain the most symmetric conserved tensor with no more than two derivative of the metric .
the @xmath14 is assumed to be identity and @xmath15 the cosmological constant .
@xmath16 gives the usual einstein - hilbert action term .
@xmath17 is called gauss - bonnet term , it is a correction term of the action @xcite .
so far , the exact static and spherically symmetric black hole solutions in third order lovelock gravity were first found in @xcite , and the thermodynamics have been investigated in @xcite@xcite@xcite .
it should be interesting to discuss the gravitational collapse in higher dimensional lovelock gravity .
the natural questions would be , for instance , how does the lovelock terms affect final fate of collapse ? what horizon structure will be formed ?
whether solutions leads to naked singularities ?
whether cch can be hold ? recently , maeda considered the spherically symmetric gravitational collapse of a inhomogeneous dust with the @xmath18-dimensional action including the gauss- bonnet term .
he discussed its effects on the final fate of gravitational collapse without finding the explicit form of the solution @xcite .
then , jhingan and ghosh considered the 5d action with the gauss - bonnet terms for gravity and give a exact model of the gravitational collapse of a inhomogeneous dust @xcite@xcite .
it s interesting to investigate the effect of higher order terms of lovelock gravity in gravitational collapse .
we explore the gravitational collapse in comoving coordinates , and seek ltb - like solutions in the third order lovelock gravity .
the dimensions of space - time concerned is @xmath19 because the third order term yields nontrivial effect in dimensions greater than or equal to 7 @xcite@xcite . using our solutions
, we discuss the formation of singularities , and analyze the nature of them . in particular , we consider that whether such singularities would be hidden or be visible to outside observers . since the cch would be violate by naked gravitational strong singularities , we investigate the gravitational strength of naked singularities , and discuss that whether the naked singularities in third order theory is a serious threat to cch .
this paper is organized as follows . in sec [ 22s ] , for the @xmath19-dimensional space - time , we give the field equations in third order lovelock gravity without a cosmological constant , and derive the ltb - like solutions . in sec [ 33s ] ,
we investigate the final fate of the spherically symmetric gravitational collapse of a dust cloud .
the subject of sec [ 44s ] is to analyze horizons ( both apparent horizons and event horizons ) and trapped surfaces , and explore that whether the singularity formed by gravitational collapse is hidden or visible to outside observers .
the strength of the singularity is demonstrated in sec [ 55s ] .
sec [ 66s ] is devoted to conclusions and discussions . throughout this paper
we use units such that @xmath20 .
the action of third order lovelock gravity in the presence of matter field can be written as @xmath21 where @xmath22 and @xmath23 are coupling constants of the second order ( gauss - bonnet ) and the third order terms , respectively . in the low - energy limit of the heterotic string theory ,
@xmath24 is regard as the inverse string tension and positive define @xcite .
hence we restrict ourselves to the case @xmath25 in this paper . for future simplicity , we take coefficients @xmath26 and @xmath27 . the gauss - bonnet term @xmath17 is @xmath28 and the third order terms of lovelock lagrangian is of the form @xmath29 varying the action eq .
( [ eq:3a ] ) , we obtain the equation of gravitation field @xmath30 where @xmath31 the solution we search is collapse of a spherically symmetric dust in @xmath19 space - time .
following ltb solutions , we assume that the system consists of a freely falling perfect fluid , it requires that the mean free path between collisions is small compared with the scale of lengths used by observer .
the energy - momentum tensor for the perfect fluid in comoving coordinates is @xmath32 where @xmath33 is the velocity vector field .
we assume that the energy density @xmath2 on the initial surface is smooth , that is , it is can be extended to a @xmath34 function on the entire real line .
the metric in comoving coordinates is written in the form @xmath35 where @xmath5 is the comoving radial cooedinate , and @xmath6 is the proper time of freely falling shells . plugging the metric eq .
( [ eq:6a ] ) into eq .
( [ eq:4a ] ) , we have @xmath36= 0 , \label{eq:7a}\end{aligned}\ ] ] where an over - dot and prime denote the partial derivative with respect to @xmath6 and @xmath5 , respectively . based on eq .
( [ eq:7a ] ) , we arrive at two families of solutions which satisfy @xmath37 and @xmath38 respectively .
the second equation involves the lovelock coupling constants , and leads to a trivial solution if @xmath39 .
we will neglect it since we want to explore the lovelock corrections to einstein gravity .
the solution of eq.([eq:8aa ] ) reads @xmath40 with an arbitrary function @xmath41 .
as it is possible to make an arbitrary re - labeling of spherical dust shell by @xmath42 , we fix the labeling by requiring that , on the hypersurface @xmath43 , @xmath5 coincides with the area radius @xmath44 apparently , every @xmath45 and @xmath46 slice of the space - time is a sphere of radius @xmath47 . the radius @xmath47 can be given an absolute significance by following interpretation . considering two particles @xmath48 and @xmath49 distribution along the radial direction in comoving coordinates
, the space distance between them at the coordinate time @xmath6 can be obtained on the hypersurface @xmath45 as @xmath50 here @xmath51 is the induced metric on the hypersurface , @xmath52 and we assume @xmath53 . in last step
the fact @xmath54 at a constant @xmath6 is used .
thus , the line element of actual space distance along radial direction is @xmath55 .
the signature of @xmath56 corresponds to three types of solutions , namely hyperbolic , parabolic and elliptic , respectively .
@xmath57 is the marginally bound case in which the metric takes the form of minkowski metric on the hypersurface @xmath43 . in comoving coordinates ,
the equations of momentum conservation @xmath58 are automatically satisfied , and the @xmath6-component reads @xmath59 which gives the solution @xmath60 thus , the mass function is defined as @xmath61 it is positive and increases with increasing @xmath5 for the nonnegativity of energy density @xmath62 . based on eq .
( [ eq:8a ] ) and @xmath52 , the @xmath63 component of equations is given by @xmath64 ' = - \varepsilon ( t , r).\label{eq:15a}\end{aligned}\ ] ] substituting eq .
( [ eq:13a ] ) and eq .
( [ eq:14a ] ) into eq .
( [ eq:15a ] ) , we get the real solution that @xmath65,\label{eq:001c}\end{aligned}\ ] ] where @xmath66 with @xmath67 .
this equation governs the time evolution of @xmath47 in d - dimensional third lovelock gravity . for gravitational collapse , the solution of @xmath68 takes @xmath69}.\label{eq:02c}\end{aligned}\ ] ] it is straightforward to check that other field equations are automatically satisfied when eq .
( [ eq:001c ] ) is satisfied . for arbitrary initial data of energy density @xmath70 , the eq .
( [ eq:02c ] ) completely specify the dynamical evolution of collapsing dust shells . in the general relativistic limit @xmath71 with marginally bound case , eq .
( [ eq:02c ] ) can be integrated to yield @xmath72 where @xmath73 is an function of integration , and can be formulated as @xmath74 consequently , the function @xmath75 is @xmath76^{\frac{2}{d-1}}}.\label{eq:20a}\end{aligned}\ ] ] it is similar to the form of @xmath75 in ltb and ltb - like - solutions @xcite@xcite .
we can consider that our ltb - like solution is attached at the boundary of the dust cloud , which is represented by a finite constant comoving radius @xmath77 , to the outside vacuum region .
the outside vacuum region is represented by the solution whose metric is @xmath78 where @xmath79 takes form of @xmath80,\label{eq:03c}\end{aligned}\ ] ] where @xmath81 with @xmath82 . if we define that @xmath83 we can prove that the ltb - like solution attached at a finite constant comoving radius @xmath84 , where we represent this hypersurface as @xmath85 , to the outside vacuum solution smoothly .
_ proof : _ as seen from inside of @xmath85 , the metric on @xmath85 is obtained by @xmath86 as seen from outside of @xmath85 , it is @xmath87 it can be checked that the induced metric is the same on both sides of the hypersurface @xmath85 with the definition eq .
( [ eq:04c ] ) .
it implies @xmath88 we can define a function @xmath89 by @xmath90 as seen from inside , the nonzero components of the extrinsic curvature @xmath91 of @xmath85 are calculated as @xmath92 and @xmath93 . as seen from outside of @xmath85
, we find @xmath94 and @xmath95 . with the definition eq .
( [ eq:04c ] ) , @xmath91 is the same on both sides of @xmath85 .
it implies @xmath96 finally , we combine eq .
( [ eq:05c ] ) and eq .
( [ eq:06c ] ) giving the equation of motion for the hypersurface @xmath85 as @xmath97 it takes the same form as eq .
( [ eq:01c ] ) with the definition eq .
( [ eq:04c ] ) .
thus , the ltb - like solution attached at a finite constant comoving radius @xmath84 to the outside vacuum solution smoothly .
in this section , we consider the final fate of the gravitational collapse in @xmath19 space - time .
as pointed earlier , in general relativity , a collapse leads to a singularity , and the conjecture that such a singularity must be covered by an event horizon is the weak cch .
there are two kinds of singularities : shell crossing singularity and shell focusing singularity which is defined by @xmath99 and @xmath100 , respectively .
the characteristic of a singularity in space - time manifold is the divergence of the riemann tensor and the energy density @xcite . in our case , the kretschmann invariant scalar @xmath101 for the metric eq .
( [ eq:6a ] ) is @xmath102 it can be certified that the kretschmann scalar is finite on the initial data surface .
according to eq .
( [ eq:14a ] ) , the energy density of fluid dust sphere is @xmath103 clearly the kretschmann scalar and the energy density diverge when @xmath99 and @xmath100 .
hence , we have both shell crossing and shell focusing singularities . shell crossing singularities can be naked , but they are inessential .
although the mass density and curvature invariants blow up there , the metric is in fact continuous .
this is seen from that the riemann curvature tensor and the energy momentum tensor are well defined , if we replace the coordinates @xmath104 by @xmath105 . on the other hand ,
shell focusing singularities are considered to be the genuine singularities in space - time manifold @xcite@xcite .
henceforth , we only concern the shell focusing singularity here . in third order lovelock gravity ,
equations and solutions of gravitational collapse are quite different from counterparts in general relativity .
hence it is necessary to investigate whether the evolution of collapsing dust cloud leads to the formation of the shell crossing singularity .
we assume each shell satisfies @xmath106 initially .
if @xmath107 is satisfied for some @xmath5 after the initial moment , this shell ceases to collapse and then bounces ( @xmath108 ) , and the shell focusing singularity would not be formed .
indeed , it can be proven that in the case @xmath109 , @xmath110 at the initial moment , the shell focusing singularity will be inevitably formed .
_ proof : _ from eq .
( [ eq:15a ] ) , we find that @xmath111 where @xmath112 .
clearly , @xmath113 if @xmath109 and @xmath114 .
here we have used the condition @xmath25 which is mentioned in the last section .
if @xmath115 is non - negative at the initial surface , it will be always non - negative as increasing @xmath116 .
hence , @xmath117 increase as decreasing @xmath47 therefore the collapsing of the shell @xmath118 from the finite initial data @xmath119 is accelerated .
thus , @xmath47 will vanish in finite proper time for the comoving observer , the shell focusing singularity is inevitably formed .
@xmath98 physically ,
the condition @xmath120 is satisfied for @xmath121 . for @xmath122 , one should take in mind eq.([eq:001c ] ) with a real solution , which gives a lower limit of @xmath123 . within this region ,
the numerical figure show the condition valid , but it is hard to analytically prove it .
moreover , one can show the condition @xmath124 valid forever if it is satisfied at the beginning .
thus , the final fate of a freely falling fluid sphere , which have initial density @xmath125 and zero pressure , is a state of infinite energy density and curvature . in the rest of this paper we will only consider the case that the shell focusing singularity could be formed , and assume @xmath109 . as demonstrate above
, our solution can be attached to the outside vacuum solution at the boundary of the dust cloud .
the outgoing property of the central singularity depends on the dominant term of @xmath126 in vacuum solution for @xmath127 @xcite .
if the metric function @xmath128 , then the tortoise coordinate defined by @xmath129 has finite negative value , the singularity is space - like . if @xmath130 , the singularity is time - like . if @xmath131 , the singularity is null since @xmath132 . from eq .
( [ eq:03c ] ) , we find that the metric function behaves around the central singularity as @xmath133 where @xmath134 is the boundary of the collapsing dust cloud . since we assume @xmath135 , if @xmath136 , we have @xmath137 hence the singularity is space - like . if @xmath138 and @xmath139 , the singularity is also space - like . if @xmath138 and @xmath140 , it is null . if @xmath138 and @xmath141 , it is time - like .
this conclusion is shown in the penrose diagram fig.1 .
the event horizon of such vacuum solution would be discussed in the next section .
if the energy density is independent on @xmath5 , namely the energy density and the space are homogeneous , the time of the formation of the singularity @xmath142 is a constant , as in the general relativity case .
a homogeneous space that is isotropic about some point is maximally symmetric and the curvature in such a space is a constant . in homogeneous case , the metric eq .
( [ eq:6a ] ) takes the form as the robertson - walker metric @xcite@xcite @xmath143 where @xmath144 is a constant . in order to satisfy eq
( [ eq:10a ] ) , the radial coordinate @xmath5 can be normalized so that @xmath145 .
when the shell hits the shell focusing singularity , the time is completely determined by @xmath146 . considering metric eq .
( [ eq:25a ] ) and @xmath147 , we get @xmath148 } , \label{eq:01c}\end{aligned}\ ] ] where @xmath149 with @xmath150 .
this equation does not contain the variable @xmath5 , that is , @xmath151 is a function of @xmath6 independent on @xmath5 .
it implies that every shell will collapse into the shell focusing singularity at the same time . obviously ,
this conclusion is hold in general relativity @xcite , gauss - bonnet gravity @xcite and third order lovelock case . from the geometric perspective
, it can be proven that in the homogeneous case the shell crossing singularity is ingoing null when @xmath138 and is ingoing space - like when @xmath136 , as shown in fig.1 .
_ proof : _ considering eq .
( [ eq:01c ] ) we have @xmath152 ( { { \dot{a}^2 } - k})a^{d-7}.\label{eq:28a}\end{aligned}\ ] ] since the energy density @xmath153 have a finite non - negative value , we find that the factor @xmath48 behaves as @xmath154 near the singularity @xmath155 , where @xmath156 is a coefficient .
we take the line element of the frw solutions with the factor @xmath157 to the conformally flat form as @xmath158 where @xmath159 the range of @xmath160 for @xmath161 is @xmath162 when @xmath138 and is @xmath163 for @xmath136 , where @xmath164 is a constant .
thus , the shell focusing singularity is ingoing null for @xmath138 , while it is ingoing space - like for @xmath136 . @xmath98
an important construction in general relativity is that of the trapped surface , which is indispensable in proving null - geodesics incompleteness in context of gravitational collapse @xcite . in general relativity ,
the trapped surface is defined as a @xmath165 closed space - like two - surface that two families of null geodesics orthogonal to this surface are converging .
physically , it captures the notion of trapping by implying that if two - surface @xmath166 ( @xmath167 ) is a trapped surface , then its entire future development lies behind a horizon . the apparent horizon is the outermost
marginally trapped surface for the outgoing null geodesics @xcite , therefore the trapped surface could not be formed during collapse without the occurrence of the apparent horizon .
the main advantage of working with the apparent horizon is that it is local in time and can be located at a given space - like hypersurface .
instead , the event horizon coincide in case of static or stationary space - time , it is non - local
. moreover , in the vacuum region the apparent horizon coincides with the event horizon of the vacuum solution @xcite .
that is , without the presence of apparent horizons there is no event horizon .
demanding the presence of the trapped surface in our spherically symmetric case implies @xmath168 and the condition for the existence of the apparent horizon with outward normals null is @xmath169 using eq .
( [ eq:8a ] ) , the apparent horizon condition becomes @xmath170 combing eq .
( [ eq:15a ] ) and eq .
( [ eq:08c ] ) , we obtain @xmath171 where @xmath172 clearly , coupling constants @xmath173 and @xmath174 produces a change in the location of apparent horizons . such a change could have a signification effect in the dynamical evolution of these horizons .
it has been shown that in the 5d gauss - bonnet gravity case , positive @xmath24 forbids apparent horizon from reaching the coordinate center thereby making the singularity massive and eternally visible @xcite , which is forbidden in the corresponding general relativistic scenario and @xmath175 gauss - bonnet gravity @xcite@xcite@xcite . in our case ,
positive @xmath173 leads to noncentral naked singularities when @xmath138 . from eq .
( [ eq:32a ] ) , we find the condition that the apparent horizon is earlier than singular shell ( @xmath176 ) for positive @xmath173 in 7d case is @xmath177 .
oppositely , eq . ( [ eq:32a ] ) could not be satisfied before the formation of the singularity in the 7d space - time in the case @xmath178 . furthermore ,
the apparent horizon touches the singularity when @xmath179 .
thus , there is no shell can reach the apparent horizon if @xmath180 with @xmath181 the boundary of the dust cloud .
it implies if the mass function @xmath182 takes sufficiently small value , the shell focusing singularity is eternally visible from infinity during collapse , and leave open even the weak form of the cch .
the condition of the formation of the eternally visible shell focusing singularity in 7d case is completely determined by the initial data of the energy density . with the help of eq .
( [ eq:14a ] ) , we can find the initial data of the energy density that condition eq .
( [ eq:33a ] ) requires . for homogeneous case ,
such initial data satisfies @xmath183 we can consider a more realistic model that @xmath184 $ ] , which is a density profile where energy density decreases as an observer move away from the center , as is expected inside a star .
this form of initial data leads to @xmath185 on the other hand , it is clear to see that in the @xmath0 space - time , apparent horizons lies earlier than the singularity unless @xmath186 .
thus , the solution forbids the formation of the naked singularity except the singular point at the coordinate center , as same as in the general relativity and the gauss - bonnet gravity . as pointed by christodoulou
, the center singularity can be naked for suitable initial data of @xmath187 in the inhomogeneous case @xcite .
the physical picture captured here is that the coordinate center hits the singularity so early that other shells have not reached apparent horizons at this moment , hence the light from the singular coordinate center can escape to the infinity . in gauss - bonnet gravity
, the formation of the eternally visible shell focusing singularity is forbidden in @xmath175 space - time , and is permitted when @xmath188 and @xmath189 in the 5d case . comparing results in second and third order lovelock gravity
, one can conjectures : for @xmath190 order lovelock gravity , such a naked singularity does not exist in @xmath191 space - time , and would be formed when @xmath192 in @xmath193 case , where @xmath8 is the coefficient of the highest order lovelock term and is positive defined .
now we consider the event horizon of the vacuum solution which describes the final fate of the collapse .
based on eq .
( [ eq:03c ] ) , we find the metric function @xmath126 satisfies @xmath194 .\nonumber\end{aligned}\ ] ] the event horizon occurs when @xmath195 , thus we obtain @xmath196 this equation leads to that @xmath197 comparing this equation to eq .
( [ eq:32a ] ) , it is obvious to see the condition for the existence of the event horizon is the same to the condition for the presence of the apparent horizon , which is : it exist in the @xmath0 case , and in the 7d case with @xmath139 .
if @xmath140 in 7d space - time , the event horizon touches the singularity at the center of the vacuum solution coordinate . this result
is also shown in fig.1 .
thus , the @xmath138 and @xmath141 case indicates the existence of naked singularities which are eternally visible from infinity , and violates even the weak form of cch .
such naked singularities are forbidden in standard general relativity , and is allowed in collapse in 5d second order lovelock gravity with positive @xmath24 @xcite .
such conclusion coincides with the result which is obtained by investigating the apparent horizon and the outgoing property of the singularity .
in the case of formation of a naked singularity in gravitational collapse , one of the most significant aspects is the strength of such a singularity in terms of the behavior of the gravitational tidal forces in its vicinity @xcite .
the importance of the singularity strength lies in the fact that even if a naked singularity occurs , if it is gravitationally weak in some suitable sense , it may not have any physical implications and it may perhaps be removable by extending the space - time through the same @xcite . the gravitational strength of a space - time singularity is characterized in terms of the behavior of the linearly independent jacobi fields along the time - like or null geodesics which terminate at the singularity . in particular , a causal geodesic @xmath198 , incomplete at the affine parameter value @xmath199 , is said to terminate in a strong curvature singularity at @xmath199 , if the volume three - form @xmath200 defined as a two - form in the case of a null geodesic ) vanishes in the limit as @xmath201 for all linearly independent vorticity free jacobi fields @xmath202 along @xmath203 .
a sufficient condition for a strong curvature singularity in 4d space - time is that in the limit of approach to the singularity , we must have along at least one causal geodesic @xmath198 @xcite@xcite , @xmath204 where @xmath205 is the tangent vector to the geodesic . essentially , the idea captured here is that in the limit of approach to such a singularity , the physical objects get crushed to a zero size , and so the idea of extension of space - time through it would not make sense , characterizing this to be a genuine space - time singularity .
this condition has been applied to higher dimensional case by ghosh , beesham and jhingan @xcite@xcite . here
we follow them to assume that such condition can be applied to our case .
we consider radial time - like causal geodesics @xmath206 with the marginally bound case @xmath207 , here the affine parameter @xmath208 is the proper time along particle trajectories .
according to such definition , @xmath209 satisfies @xmath210 , that is , @xmath211 using the geodesic equation , we have @xmath212 substituting eq .
( [ eq:38a ] ) into eq .
( [ eq:39a ] ) , we obtain that radial time - like geodesics must satisfy @xmath213 the simplest solution is the worldline of a freely falling particle , which is @xmath214 . in terms of proper time we can describe it as @xmath215 eq .
( [ eq:33a ] ) shows that the naked shell focusing singularity in 7d space - time occurs when the region of the collapsing dust is sufficiently small .
moreover , as mentioned earlier , in @xmath0 space - time only the central singularity can be naked .
hence , our purpose is to discuss the strength of the central singularity .
we consider the expansion of @xmath216 near @xmath186 @xmath217 it specify that @xmath68 behaves as @xmath218 ( homogeneous case ) near the coordinate center , where @xmath219 is a function of @xmath6 and vanishes at @xmath220 .
thus , we get @xmath221 clearly , @xmath222 thus we have @xmath223 from eq .
( [ eq:001c ] ) , we obtain that @xmath224 where @xmath225 is defined as @xmath226 therefore has finite value . substituting eq .
( [ eq:45a ] ) into eq .
( [ eq:46a ] ) we have @xmath227 the strong curvature condition eq .
( [ eq:37a ] ) is not satisfied on the singularity near @xmath186 in the 7d space - time , while it is satisfied in the @xmath0 space - time .
hence , the central shell focusing singularity is gravitational strong in the @xmath0 space - time and may be gravitational weak in the 7d space - time since we chosen a special solution of eq .
( [ eq:41a ] ) .
the difference between two cases is that the singularity can be eternally visible in the 7d case while it is forbidden in the @xmath0 case .
that is , the singularity is gravitational strong when the solution represents the black hole formation for arbitrary initial data .
if a naked singularity is gravitational weak , it may not have any significant physical consequences so may not be a serious threat to the cch .
thus , the naked singularity in the @xmath0 case violates the cch seriously , as in the general relativity @xcite@xcite , while the naked singularity may not be regarded as an essential counter example to the cch in the 7d case .
nevertheless , such a central singularity in the 7d case is gravitational strong in the general relativistic and second order lovelock gravity . in the general relativistic limit @xmath228 , eq .
( [ eq:37a ] ) takes the form @xmath229,\nonumber\end{aligned}\ ] ] with @xmath230 thus , @xmath231 the strong curvature condition is satisfied in @xmath232 space - time . in second order lovelock gravity limit @xmath233
, we find that @xmath234 the strong curvature condition is satisfied in @xmath235 space - time .
these results indicate that the lovelock interaction weaken the strength of the singularity .
moreover , a generic expression for @xmath190 order lovelock gravity could be @xmath236 it shows that the strong curvature condition is satisfied in @xmath237 space - time .
in this paper , we have investigated the gravitational collapse of a spherically symmetric dust cloud in @xmath19 space - time in third order lovelock gravity without a cosmological constant . from field equations given by the third order lovelock action , we discussed the solutions of three families @xmath238 , @xmath239 and
@xmath240 . we discussed the final fate of the collapse , and gave a condition for the formation of the shell focusing singularity .
we also analysed the global structure of the space - time and gave the penrose diagram for the homogeneous case .
it turns out that , the contribution of high order curvature corrections has a profound influence on the nature of the singularity , and the whole physical picture of the gravitational collapse changes drastically .
high order lovelock terms alters the course of collapse and the time of formation of singularities , modifies apparent horizon formation and the location of apparent horizons , and changes the strength of singularities .
the most attractive consequence is that an massive naked shell focusing singularity is inevitably formed in 7d space - time , which is quite different from that in the general relativity and in the gauss - bonnet gravity .
however , as we shown , the strength of the naked singularity in 7d case is weaker than that in the general relativistic limit , therefore this may not be a serious threat to the cch . on the other hand , unlike the 7d case , there is a serious threat to the cch caused by the naked , gravitational strong singularity in the @xmath0 inhomogeneous collapse , thus the cch which is violated in general relativity could not be protected in lovelock gravity .
when revising our paper in lines with reviewer s suggestions , we have received a paper by ohashi , shiromizu and jhingan@xcite discussing the gravitational collapse in the lovelock theory with arbitrary order which partially covers the result of present paper .
in fact , they did not investigate the ingoing and outgoing properties of the singularity , the penrose diagram , and whether the condition of strong singularity is satisfied .
one may study such properties in a general lovelock gravity similarly .
j. r. oppenheimer and h. snyder , phys .
rev . * 56 * , 455 ( 1939 ) .
g. lemaitre , gen .
* 29 * , 641 ( 1997 ) [ annales soc .
brux . ser .
phys . a * 53 * , 51 ( 1933 ) ] ; r. c. tolman , proc . nat
. acad .
sci . * 20 * , 169 ( 1934 ) ; h. bondi and t. gold , mon . not .
soc . * 108 * , 252 ( 1948 ) .
r. maartens , n. p. humphreys , d. r. matravers and w. r. .
stoeger , class .
* 13 * , 253 ( 1996 ) [ erratum - ibid .
* 13 * , 1689 ( 1996 ) ] [ arxiv : gr - qc/9511045 ] .
d. garfinkle , class .
* 23 * , 4811 ( 2006 ) [ arxiv : gr - qc/0605088 ] .
s. hawking and d. sciama , comments astrophys .
space phys .
* 1 * , 1 ( 1969 ) . s. w. hawking and g. f. r. ellis , the large scale structure of space - time , _ cambridge university press , cambridge , 1973 _ r.penrose , riv .
nuovo cimento 1 , 252 ( 1969 ) p. s. joshi , pramana * 55 * , 529 ( 2000 ) [ arxiv : gr - qc/0006101 ] .
d. christodoulou , commun .
phys . * 93 * , 171 ( 1984 ) .
b. c. nolan and f. c. mena , class . quant .
* 19 * , 2587 ( 2002 ) [ arxiv : gr - qc/0203078 ] .
p. horava and e. witten , nucl . phys .
b * 475 * , 94 ( 1996 ) [ arxiv : hep - th/9603142 ] ; a. lukas , b. a. ovrut and d. waldram , phys . rev .
d * 60 * , 086001 ( 1999 ) [ arxiv : hep - th/9806022 ] .
l. randall and r. sundrum , phys .
lett . *
83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ] ; l. randall and r. sundrum , phys .
* 83 * , 4690 ( 1999 ) [ arxiv : hep - th/9906064 ] .
a. banerjee , a. sil and s. chatterjee , astrophys .
j. * 422 * , 681 ( 1994 ) ; a. sil and s. chatterjee , gen .
* 26 * , 999 ( 1994 ) .
s. g. ghosh and a. beesham , phys . rev .
d * 64 * , 124005 ( 2001 ) [ arxiv : gr - qc/0108011 ] ; s. g. ghosh , d. w. deshkar and n. n. saste , int .
j. mod .
d * 16 * , 53 ( 2007 ) .
d. lovelock , j. math .
* 12 * , 498 ( 1971 ) .
d. lust and s. theisen , lect .
notes phys . *
346 * , 1 ( 1989 ) ; s. o. alekseev and m. v. pomazanov , grav .
cosmol . * 3 * , 191 ( 1997 ) ; k. i. maeda , n. ohta and y. sasagawa , phys . rev .
d * 80 * , 104032 ( 2009 ) [ arxiv:0908.4151 [ hep - th ] ] .
e. elizalde , s. jhingan , s. nojiri , s. d. odintsov , m. sami and i. thongkool , eur .
j. c * 53 * , 447 ( 2008 ) [ arxiv:0705.1211 [ hep - th ] ] .
m. c. bento and o. bertolami , phys .
b * 368 * , 198 ( 1996 ) [ arxiv : gr - qc/9503057 ] .
r. c. myers and j. z. simon , gen .
* 21 * , 761 ( 1989 ) .
x. h. ge , s. j. sin , s. f. wu and g. h. yang , phys .
d * 80 * , 104019 ( 2009 ) [ arxiv:0905.2675 [ hep - th ] ] .
m. h. dehghani and r. pourhasan , phys .
d * 79 * , 064015 ( 2009 ) [ arxiv:0903.4260 [ gr - qc ] ] .
m. h. dehghani and m. shamirzaie , phys .
d * 72 * , 124015 ( 2005 ) [ arxiv : hep - th/0506227 ] .
h. maeda , phys .
d * 73 * , 104004 ( 2006 ) [ arxiv : gr - qc/0602109 ] .
s. jhingan and s. g. ghosh , phys .
d * 81 * , 024010 ( 2010 ) ; s. g. ghosh and s. jhingan , phys .
024017 ( 2010 ) [ arxiv:1004.0071 [ gr - qc ] ] . m. h. dehghani and n. bostani , phys .
d * 75 * , 084013 ( 2007 ) [ arxiv : hep - th/0612103 ] . m. h. dehghani and m. shamirzaie , phys .
d * 72 * , 124015 ( 2005 ) [ arxiv : hep - th/0506227 ] .
d. g. boulware and s. deser , phys .
* 55 * , 2656 ( 1985 ) .
s. s. deshingkar , i. h. dwivedi and p. s. joshi , phys .
d * 59 * , 044018 ( 1999 ) [ arxiv : gr - qc/9805055 ] .
h.p.robertson , ap.j.82,284(1935 ) .
a.g.walker , proc.lond.math.soc(2),42,90(1936 ) .
s. a. hayward , phys .
d * 49 * , 6467 ( 1994 ) .
f. i. cooperstock , s. jhingan , p. s. joshi and t. p. singh , class .
* 14 * , 2195 ( 1997 ) [ arxiv : gr - qc/9609051 ] .
t. torii and h. maeda , phys .
d * 71 * , 124002 ( 2005 ) [ arxiv : hep - th/0504127 ] .
f. j. tipler , phys .
lett . a * 64 * ( 1977 ) 8 .
c. j. s. clarke , the analysis of space - time singularities , _ cambridge , uk : univ .
( 1993 ) 175 p. ( cambridge lecture notes in physics , 1 ) _ c. j. s. clarke and a. krolak , j. geom .
phys . 2 , 127 , ( 1985 ) .
s. g. ghosh and a. beesham , class.quant.grav .
17 , 4959 , ( 2000 ) .
seiju ohashi , tetsuya shiromizu and sanjay jhingan , [ arxiv : gr - qc/1103.3826 ] . | we investigate the spherically symmetric gravitational collapse of an incoherent dust cloud by considering a ltb type space - time in third order lovelock gravity without cosmological constant , and give three families of ltb - like solutions which separately corresponding to hyperbolic , parabolic and elliptic .
notice that the contribution of high order curvature corrections have a profound influence on the nature of the singularity , and the global structure of space - time changes drastically from the analogous general relativistic case .
interestingly , the presence of high order lovelock terms leads to the formation of massive , naked and time - like singularities in the 7d space - time , which is disallowed in general relativity .
moveover , we point out that the naked singularities in the 7d case may be gravitational weak therefore may not be a serious threat to the cosmic censorship hypothesis , while the naked singularities in the @xmath0 inhomogeneous collapse violate the cosmic censorship hypothesis seriously . |
among the many problem areas in mechanics the study of instabilities and irreversible processes seem particularly interesting .
engineering mechanics exists as a discipline because failures of structures cost so many lives .
the analysis of local lyapunov instability gives a means for localizing and predicting catastrophic failures so that there is a decidedly practical engineering aspect of this fascinating scientific research area .
the goal we pursue here is to develop models that help us to understand .
it is profoundly interesting that the small scale microscopic models of material behavior ( ordinary newtonian mechanics ) are time - reversible while the macroscopic models of the same thing ( finite - element and finite - difference fluid mechanics and solid mechanics ) are irreversible .
the tools from nonlinear dynamics and chaos are useful in analyzing these two kinds of description . here
we describe the basic building blocks for particle simulation and point out the ways that these time - reversible simulations already lead to time - irreversible behavior .
most of the examples treated here are also described in our three books on computational statistical mechanics , smooth - particle applied mechanics , and time reversibility , computer simulation , algorithms , and chaos@xcite .
no special tricks are necessary to get started with particle - based simulations .
microscopic mechanics can provide us with accurate particle trajectories @xmath0 .
all we need to do is to integrate newton s ordinary differential equations of motion , @xmath1 here @xmath2 is the potential energy , a function of the coordinates @xmath3 .
alternatively , we can obtain an equivalent coordinate - momentum description @xmath4 by integrating hamilton s first - order ordinary differential equations : @xmath5 both these approaches are time - reversible .
that is , a movie of the motion , played backwards , satisfies exactly the _ same _ equations ( with the values of @xmath6 and @xmath7 in reversed order and with the sign of @xmath7 changed also ) .
a movie is an excellent analog of numerical simulation .
both the simulation and the movie are sets of discrete records of coordinates at discrete values of the time , separated by the `` timestep '' @xmath8 .
in addition to the basic algorithm keep in mind that three crucial questions remain to be answered : [ 1 ] what are the initial conditions , [ 2 ] what are the boundary conditions , and ( most important of all ) [ 3 ] what is the problem to be solved ?
macroscopic continuum mechanics is based on the three conservation laws for mass , momentum , and energy : @xmath9 here @xmath10 is density , @xmath11 is velocity , @xmath12 is energy per unit mass , @xmath13 is the pressure tensor ( force per unit area , necessarily a symmetric second - rank tensor ) , and @xmath14 is the heat flux vector ( energy flow per unit area ) .
all these variables are continuous functions of space and time .
both @xmath13 and @xmath14 , as well as @xmath15 , @xmath16 , and @xmath17 , are defined in the _ comoving _ frame , a coordinate frame moving with the local velocity @xmath18 .
finite - difference approximations to the gradients on the righthand sides of the three continuum equations , evaluated at a discrete set of spatial mesh points , reduce the partial differential equations to ordinary ones , which can then be solved with runge - kutta integration .
again , the hard part of the problem is the same : what to do and how to implement the initial and boundary conditions .
different materials can be described by different types of constitutive relations ( elastic , plastic , viscous , ... ) giving @xmath13 and @xmath14 in terms of the basic @xmath19 set , together with their time derivatives and spatial gradients .
time - reversed movies of solved macroscopic problems look `` funny '' and make no sense .
this is because the underlying phenomenological constitutive relations are typically irreversible .
the simplest most familiar irreversible examples are newtonian viscosity and fourier heat conduction : @xmath20i - \eta [ \ \nabla
u + \nabla u^t \ ] \ ; \ q = -\kappa \nabla t \ .\ ] ] in the symmetrized velocity gradient @xmath21 is the transpose of @xmath22 .
@xmath23 is the unit tensor . a boxed conducting fluid , with that fluid initially in motion , ( a rayleigh - bnard flow , for instance , but with the box suddenly insulated and with the accelerating gravitational field suddenly switched off ) described with a shear viscosity @xmath24 and a heat conductivity @xmath25 eventually comes to an isothermal state of rest . evidently the reversed movie of this decay makes no sense and would correspond to an illegal `` something from nothing '' contradicting the second law of thermodynamics . of course
, with the right initial conditions and the right boundary conditions , one can indeed observe tornadoes ! for an @xmath26 ( two - dimensional ) system , with kinematic viscosity and thermal diffusivity of order @xmath27 the initial gradients decay exponentially , @xmath28 .
the reversed movie , with its exponential growth , @xmath29 , is simply wrong . in applications of mechanics to simulation we strongly recommend the use of the fourth - order runge - kutta integrator because it is easy to use and to modify for the treatment of open systems interacting with their environments .
if the focus is on the time - reversibility of conservative newtonian systems it is useful to consider a very simple , yet rigorously time - reversible , integrator discovered by levesque and verlet@xcite . in order
better to understand the coexistence of the reversible microscopic and irreversible macroscopic views we adopt levesque and verlet s `` bit - reversible '' algorithm .
this approach generates a numerical trajectory in an _ integer _ coordinate space , by rounding off the acceleration terms : @xmath30_{\rm integer } \
\ } \ .\ ] ] the subscripts indicate the time , in units of the ( integer ) timestep @xmath8 . the initial conditions to start this algorithm are the coordinates at two successive times .
a simple illustration of the algorithm follows a harmonic oscillator trajectory , using unit mass , force constant , and timestep @xmath8 : @xmath31 the solution of repeating coordinates @xmath32 , is typical , and illustrates the fact that no matter what the initial conditions , the solution is both periodic ( for chaotic problems , the length of the period is of order the square root of the number of states ) and reversible .
the algorithm is a faithful analog of classical deterministic time - reversible mechanics .
if momenta are desired they too can be approximated accurately from the coordinate values : @xmath33\frac{(q_+ - q_-)}{2\delta t } - \left[\frac{1}{3}\right]\frac{(q_{++ } - q_{--})}{4\delta t } \ .\ ] ] figure 1 compares the energy calculated with these momenta with calculations based on the beeman and velocity - verlet algorithms .
our formulation ( small dots in the figure ) is clearly an improvement .
it is evident that a promising research area lies in the development of higher - order bit - reversible algorithms combining coordinates , velocities , and accelerations from more than three successive times .
is there any chance of detecting _ irreversibility _ with such a time - reversible algorithm ?
oddly enough , there is !
it is based on the analysis of lyapunov instability , looking in the _ neighborhood _ of the trajectory , not just at the trajectory itself .
such a nonlocal analysis necessarily depends upon the imagination and the information contributed by an observer of the motion .
let us take an example , a _ maximally irreversible _
situation described by time - reversible , even bit - reversible , newtonian mechanics .
consider the pair of shockwaves launched by the collision of two mirror - image fluid samples .
see figure 2 for four snapshots of such a problem .
_ initially _ the velocities are @xmath34 .
_ eventually _ the periodic @xmath26 system shows no more systematic motion
the initial kinetic energy has been completely converted to internal energy ( heat ) : @xmath35 just as before , any portion of the developing trajectory can be reversed precisely and exactly despite the lyapunov instability ( exponential growth of perturbations ) of the dynamics . there is a vast literature@xcite on the quantification of lyapunov instability , the exponentially sensitive deformation of comoving hypervolumes in @xmath6 space , @xmath7 space , or @xmath36 phase space . for @xmath37 particles in two dimensions
the @xmath38-dimensional phase - space motion defines @xmath38 local lyapunov exponents .
the fact that these `` local '' exponents depend upon the chosen coordinate system can be viewed as a disadvantage or as an opportunity .
again , there are many promising research problems suggested by this observation .
optimizing the analysis is certainly a useful and stimulating activity . the largest lyapunov exponent
we will call it @xmath39 associated with the motion can be found by following two nearby trajectories in time . the primary or `` reference '' trajectory
can be generated with bit - reversible dynamics , so that it is possible to extend it as far as desired into the future or the past@xcite .
the dynamics of a nearby `` satellite '' trajectory is restricted by constraining the satellite trajectory to stay within a fixed distance of the reference trajectory .
the satellite dynamics can readily be generated with runge - kutta integration , rescaling the separation between the two trajectories at the end of each time step .
the local lyapunov exponent is : @xmath40 \ln \left[\frac{(\delta r)_{\rm before}}{(\delta r)_{\rm after}}\right ] \ ; \ \delta
r \equiv |r_{\rm satellite } - r_{\rm reference}| \ .\ ] ] although the motion equations are perfectly reversible for both the reference and the satellite , the reversed satellite trajectory turns out to be totally unlike the forward one , _ if _ the system is a nonequilibrium system . both the local lyapunov exponent associated with the instability and the identities of those particles making above - average contributions to the offset vector separating the trajectories , @xmath41 , are qualitatively different . in a typical shockwave simulation of the type shown in figure 2
the number of these more - heavily - weighted particles is about twice as great in the reversed motion as in the forward one .
this is an extremely interesting result .
no doubt it suggests various `` arrows of time '' which can be constructed based on the structure of nearby trajectories ( which react to the past , not the future ) .
a study of irreversible flows from this standpoint should shed light on the reversibility paradox for simple newtonian and hamiltonian systems .
in order to control nonequilibrium states , in particular nonequilibrium _ steady _ states , it is necessary to do work and/or to exchange heat , with the system of interest .
the dynamics becomes a little more complicated due to these interactions , but the interpretation compensates by becoming simpler .
here we take up the description of `` open systems '' .
`` open '' systems have mechanical and/or thermal connections to their environment , opening up the possibility of simulating processes including thermodynamic work and the flow of heat .
analysis of these systems requires microscopic analogs for all the continuum variables .
density , velocity , and energy are the simplest of these .
we adopt the smooth - particle averaging method discovered by lucy and monaghan in 1977@xcite : @xmath42 the particle energies @xmath43 are necessarily defined in the comoving frame , which moves with velocity @xmath18 .
a useful weight function , with a range @xmath44 which can be optimized , is lucy s , here normalized for two space dimensions : @xmath45 this weight - function approach guarantees the continuity of the first and second spatial derivatives of the field variables and lends itself to optimization studies .
in addition to the basic mass , momentum , and energy , several other variables need to be considered in order to compare microscopic and macroscopic simulations . unlike their continuum cousins ,
the pressure tensor and the heat flux vector from molecular dynamics are respectively even and odd functions of the time : @xmath46 + \sum_i(ep / m)_i \ .\ ] ] here @xmath47 and @xmath48 is the force on particle @xmath49 due to its interaction with particle @xmath50 . the individual particle energies @xmath43 include half of each particle s pair interactions with its neighbors .
temperature needs a definition too .
the usual equilibrium definition , based on entropy , is useless away from equilibrium where entropy has no consistent definition@xcite .
_ at _ equilibrium _ temperature _ can be defined in many ways , all based on gibbs statistical mechanics or maxwell and boltzmann s kinetic theory .
the even moments of the velocity distribution are examples .
in addition to these there are also configurational definitions .
the simplest `` configurational temperature '' is based on an identity from landau and lifshitz text@xcite : @xmath51 this definition follows from an integration by parts in gibbs canonical ensemble .
if the differentiation indicated by @xmath52 is carried out in momentum space the landau - lifshitz formula gives the usual kinetic - theory definition of temperature , @xmath53 . if instead the gradient is carried out in coordinate space the `` configurational temperature '' depends on the first and second derivatives of the potential function governing the motion : @xmath54 one - body or many - body configurational temperatures , either scalar or tensor , can be defined in this way . but
an evaluation of them for the shockwave problem reveals divergences .
typically particle values of @xmath55 frequently alternate between positive or negative values , so that the corresponding configurational temperatures frequently diverge !
configurational temperature also has unphysical undesirable contributions arising from rotation whenever coriolis or centrifugal forces are significant .
the simplest definition for temperature is the kinetic second - moment one .
it is based on a mechanical model of a working ideal - gas thermometer . in that instance
a relatively heavy mass-@xmath56 `` system atom '' interacts with a collection of light - weight mass-@xmath57 `` ideal - gas thermometer '' particles characterized by an unchanging equilibrium maxwell - boltzmann distribution with temperature @xmath58 .
kinetic theory shows that the averaged effect of such collisions causes the system - atom velocity to decay while its mean - squared velocity approaches the equilibrium value for the temperature @xmath58 : @xmath59/\tau \ ; \ [ \ { \rm for } \
m<<m \ ] \ .\ ] ] accordingly , we adopt the kinetic definition of temperature in what follows . with temperature
defined we can proceed to devise `` thermostats '' able to control it .
the first of the deterministic mechanical thermostats was woodcock s isokinetic thermostat@xcite , implemented by rescaling the velocities at the end of each timestep .
much later it was discovered@xcite that a continuous time - reversible version of this thermostat could be implemented with a _ time - reversible _ friction coefficient @xmath60 : @xmath61 this `` isokinetic '' thermostat can be applied to one or many particles and to one or many space directions .
an illustrative application is the `` galton board''@xcite , in which a single particle is accelerated through a lattice of scatterers but constrained to move at constant speed .
overall , the potential energy drops . because the mean value of the friction coefficient is necessarily positive , the phase - space probability density collapses onto a multifractal strange attractor , quantifying the rarity of nonequilibrium phase - space states .
this approach to temperature control is often termed the `` gaussian '' thermostat because gauss principle ( of least constraint ) gives this thermostat when applied to the problem of constraining the kinetic energy@xcite .
reference 15 is a detailed discussion of the model ( summarized in references 1 and 2 ) .
this work clearly shows the fractal nature of the phase space ( with vanishing phase volume ) that results when the dynamics is thermostated .
fancier thermostats , based on statistical mechanics , can be found . in 1984
shuichi nos discovered a precursor of the best of them , a thermostat@xcite with a more elaborate basis in lagrangian and hamiltonian mechanics , but somewhat disfigured by a novel `` time - scaling variable '' @xmath62 .
his thermostat imposed gibbs canonical phase space distribution at equilibrium rather than the less - usual isokinetic one .
a simplification of his equation of motion , without the useless time - scaling , likewise contained a friction coefficient , which itself obeyed an evolution equation depending upon _ past _ values of the kinetic energy : @xmath63/\tau ^2 \ [ \ { \rm nos\acute e - hoover } \ ] \ .\ ] ] the relaxation time @xmath64 is a free parameter determining the time required for the thermostat forces @xmath65 to bring the kinetic temperature @xmath66 to the desired thermostat temperature @xmath67 . just as in the isokinetic case the nonequilibrium averaged friction coefficient for this `` nos - hoover '' mechanics is positive , leading once again to multifractal phase space distributions away from equilibrium .
there is an extensive somewhat mathematical literature having to do with picking the `` right '' relaxation time or the `` right '' thermostat . the gaussian and nos - hoover thermostats are particularly useful for controlling nonequilibrium problems , such as shear flows and heat flows , and the rayleigh - bnard problem combining them .
the definition of temperature depends upon the definition of the local velocity @xmath18 .
a straightforward definition of velocity , which nicely satisfies the continuity equation exactly@xcite , can be based on smooth - particle weighting functions : @xmath68 the gauss , nos , and nos - hoover thermostats can _ all _ be related to hamiltonian mechanics .
dettmann , together with morriss , carried out much of this work , with later contributions by bond , laird , and leimkuhler , and then by campisi , hnggi , talkner , and zhan@xcite .
all of them helped to clarify the connections of time - reversible thermostats with standard hamiltonian mechanics .
this work leads to the conclusion that _ many _ different thermostats can be used at equilibrium but that some of them fail in nonequilibrium situations , even in situations close to equilibrium .
just as in real life the failures , rather than the successes , are the more newsworthy subjects .
let us turn to some examples .
a very stimulating `` log - thermostat '' has just been described by campisi , hnggi , talkner , and zhan@xcite .
they pointed out that the microcanonical ( constant energy ) ensemble distribution for a logarithmic potential generates ( at least formally ) the maxwell - boltzmann velocity distribution : @xmath70 \propto \exp [ \ ( h_0/kt ) - ( p^2/2mkt ) \ ] \\ ] ] because the dynamics of this thermostat is unstable , there being nothing to keep @xmath6 away from the origin , in applications they recommend using @xmath71 , where @xmath72 is sufficiently small .
our effort to use this thermostat for a nonequilibrium heat flow problem failed . connecting a cold and
a hot log - thermostat to opposite ends of a two - particle @xmath69 chain gave different temperatures at the two ends , but _ no heat flux at all_. the problem is that the hamiltonian log - thermostat is unable to replicate the phase - space contraction associated with dissipative systems .
there are some other examples of such failures .
leete and hoover s hamiltonian@xcite , @xmath73 keeps the kinetic energy , @xmath74 constant , equal to @xmath75 .
configurational _ temperature can alternatively be kept constant using a special hamiltonian . in both these cases
a cold and a hot thermostated region , in contact with newtonian regions , gives _ no heat flux at all _ despite huge temperature differences .
the lesson is that hamiltonian mechanics is not able to describe dissipation properly .
a `` good '' thermostat should , for instance , be able to provide good solutions of the rayleigh - bnard problem , heat transfer through a compressible fluid in a gravitational field .
it should also be useful in treating small - scale `` toy problems '' .
the simplest thermostat test problems are [ 1 ] a harmonic oscillator@xcite with a coordinate - dependent temperature : @xmath76 and [ 2 ] the flow of heat through a @xmath69 chain of particles : @xmath77 + \sum_{i < j } \phi_{ij } \ , \ ] ] where @xmath78 is a nearest - neighbor hookes - law potential and where the first few and last few particles in the chain are thermostated with a gaussian or nos - hoover or another thermostat .
this model is a specially good one to study because it is known to satisfy fourier s law , even in one dimension .
a comparison of seven different thermostat methods showed that the @xmath69 problem is well - posed and relatively easy to solve@xcite .
some hamiltonian - based thermostats are ineffective for nonequilibrium problems@xcite and it is useful to understand why . at equilibrium a given temperature and volume imply corresponding values of the kinetic and potential energies .
this is also true for particular states away from equilibrium , even where there is no longer a unique equation - of - state relation .
using a hamiltonian thermostat away from equilibrium one can independently specify the kinetic energy _ and
_ the potential energy or the temperature _ and _ the energy .
this additional freedom contradicts the notion of thermodynamic state and can lead to very strange results@xcite . constraining the configurational temperature or using a version of hamiltonian mechanics to constrain the kinetic energy discovered by hoover and leete provide temperature profiles that make no sense .
the log - thermostat is another demonstration that hamiltonian mechanics is a poor choice for thermostats .
this paradoxical situation is the symptom of two incompatible requirements on the dynamics : [ 1 ] liouville s theorem requires that the phase - space motion be incompressible ; [ 2 ] heat flow consistent with the second law of thermodynamics requires that the phase volume decrease to zero .
several of the thermostats have no problem with generating heat flows and solve the problem of decreasing phase volume by generating strange attractors in the phase space .
let us consider what might appear to be the simplest of these problems , the nos - hoover oscillator@xcite with a temperature gradient@xcite : @xmath79 for @xmath80 the motion is a one - dimensional limit cycle with @xmath81 positive .
the mean value of the friction coefficient @xmath60 in the range @xmath82 increases from about 0.15 to 1.35 .
figure 3 shows the gradual expansion of the hysteretic limit cycle as the maximum temperature gradient is increased from 0.44 to 1.00 .
figures 4 and 5 show a bit of the complexity associated with smaller values of the temperature gradient .
using an initial momentum of unity gives more regular attractors , of the type shown to the left . on the other hand ,
much higher initial momenta give chaotic distributions like those shown to the right .
this complexity is no doubt related to that seen without any temperature gradient at all@xcite . in that latter case
the phase - space distribution is divided into an _
infinite _ number of coexisting distributions , whose union is gibbs canonical distribution , @xmath83/2 } \ .\ ] ] the figures show projections of a strange attractor that forms with @xmath84 .
the lyapunov spectrum in this case is nearly symmetric , so that it is difficult to compute an accurate information dimension of the attractor .
fortunately , the complex dynamics of the thermostated oscillator can be greatly simplified by adding another control variable , a friction coefficient controlling the fourth velocity moment@xcite : @xmath85 at equilibrium the extra control variable allows the oscillator to sample the complete canonical distribution .
this works at nonequilibrium too .
figure 6 compares the distributions of the two friction coefficients @xmath86 for @xmath87 equal to 0.5 and 1.0 .
even in the latter case the chaos induced by the two coefficients is enough to prevent collapse of the dynamics onto a limit cycle .
although counterintuitive , it appears to be true that a four - dimensional attractor is actually much _ simpler _ than its three - dimensional counterpart .
larger systems fit the pattern to which the small systems hint .
the phase - space distribution shrinks to a strange attractor . in a system with several thermostated degrees of freedom
liouville s theorem gives the details of the shrinkage@xcite : @xmath88 \ .\ ] ] here @xmath89 is the external entropy production , the heat extracted from the controlled system by the thermostats , divided by the thermostat temperature .
@xmath90 is a small comoving phase volume .
@xmath90 has three possible evolutions : it can expand ; it can shrink ; or it can remain the same . the last possibility is the equilibrium one , with no net heat transfer to the outside world .
the first possibility ( expansion ) is ruled out for steady states , as a continually expanding phase volume implies catastrophic instability .
only the possibility of continual shrinkage , dissipation , is left .
the accessible phase - space states for a nonequilibrium steady state continually decrease in number as the volume shrinks ( exponentially fast ) toward zero .
the deterministic time - reversible thermostats make possible a simple geometric interpretation of the second law of thermodynamics .
nonequilibrium steady states necessarily collapse to a zero - volume strange attractor . thus nonequilibrium states are vanishingly rare .
any attempt to reverse the ( time - reversible ) dynamics would lead to divergence , with a positive lyapunov sum , and a violation of the second law .
what happens in fact is that , when reversed , the dynamics soon breaks its time symmetry and seeks out again the attractor .
time - reversible thermostats have deepened our understanding of the second law@xcite .
the paradoxical reversibility properties of newtonian and hamiltonian mechanics can be modeled with bit - reversible algorithms .
such algorithms do nt exist in cases where the phase volume changes , where the mechanics is thermostated . in the latter case lyapunov instability
seeks out the unstable strange attractor , more stable still than is its repeller twin , leading to a simple geometric understanding of the second law of thermodynamics for open systems .
the symmetry breaking revealed by strong shockwaves suggests that a deepened understanding of isolated systems can come from study of the local lyapunov spectrum . both of these problem areas , nonequilibrium
conservative systems and nonequilibrium open systems , suggest many interesting research opportunities for the future .
this work represents a continuing effort which has been stimulated by many colleagues .
most recently , we thank francesco ginelli and massimo cencini for inviting our contribution to a special issue of the journal of physics a ( mathematical and theoretical ) , `` lyapunov analysis from dynamical systems theory to applications '' .
anton krivtsov and vitaly kuzkin encouraged our participation in this meeting .
we are also specially grateful to michele campisi and his colleagues and coworkers for many stimulating emails which helped us to identify some of the many interesting problem areas presented here .
g. hoover , c. g. hoover , h. a. posch , and j. a. codelli , `` the second law of thermodynamics and multifractal distribution functions : bin counting , pair correlations , and the kaplan - yorke conjecture '' , communications in nonlinear science and numerical simulation * 12 * , 214 - 231 ( 2007 ) . | molecular dynamics and statistical mechanics make possible a particle - based understanding of thermodynamics and hydrodynamics , including the fascinating loschmidt contradiction between time - reversible atomistic mechanics and the time - irreversible thermodynamic dissipation incorporated into macroscopic fluid and solid mechanics . |
possessing two higgs doublets is one of the characteristics of the minimal supersymmetric standard model ( mssm ) in order to give masses to the up - quark sector and the down - quark sector separately [ 1 ] .
this property enables in principle cp violation to occur in the higgs sector of the mssm through the mixing between the cp even and the cp odd states [ 2 ] .
the mechanism of cp violation in the mssm has been studied by many authors [ 3 - 12 ] .
it is found that at the tree level , neither explicit nor spontaneous cp violation is possible in the higgs sector of the mssm , because any complex phases in it can always be eliminated by rotating the higgs fields . even at the one - loop level ,
spontaneous cp violation is disfavored because it requires a very light neutral higgs boson , which has already been ruled out by experiments [ 3,4 ] . on the other hand ,
explicit cp violation is viable in the mssm at the one - loop level by virtue of the radiative corrections due to the loops of relevant particles , such as quarks and squarks [ 5 - 12 ] .
the radiative corrections by these particles yield the mixing between the cp even and the cp odd neutral higgs bosons .
thus , it is possible to achieve the explicit cp violation in the radiatively corrected higgs sector of the mssm . however , a drawback of the mssm is that it contains a higgs mixing term @xmath1 , where @xmath2 is a parameter having mass dimension , and @xmath3 ( @xmath4 ) are higgs doublets .
it is known that this @xmath2 parameter causes a hierarchy problem with respect to the symmetry breaking scale [ 13 ] . a natural framework to solve
this problem is to generate the @xmath2 parameter via the vacuum expectation value ( vev ) of a higgs singlet [ 14 ] .
a possibility for the framework is to extend simply the higgs sector of the mssm by introducing a higgs singlet , like the next - to - minimal supersymmetric standard model , where the @xmath2 parameter is practically replaced by a dimensionless parameter times the vev of the higgs singlet [ 14 ] . a more plausible possibility to solve the @xmath2 problem is to extend not only the higgs sector but also the gauge symmetry of the mssm by introducing a nonanomalous @xmath0 symmetry broken at the tev scale [ 15 ] . extending the gauge symmetry to include an additional @xmath0 symmetry is widely considered in various theoretical models , such as string models or gut models .
the mssm with an extra @xmath0 symmetry can actually forbid the @xmath1 term but allow @xmath5 , by assigning appropriate @xmath0 charges , where @xmath6 is the higgs singlet and @xmath7 is a dimensionless parameter .
after @xmath6 develops a vev of the electroweak scale , this @xmath7 term would effectively generate the @xmath2 term at the electroweak scale [ 16 ] . in the literature
, a number of studies have been performed on the mssm with an extra @xmath0 symmetry [ 17 - 20 ] .
several authors have calculated the higgs boson masses from the tree - level higgs potential of this model [ 17 ] as well as from the radiatively corrected higgs potential [ 18,19 ] .
it is found that , at the tree level , the lightest scalar higgs boson in this model can be heavier than @xmath8 boson .
that is , the upper bound on the mass of the lightest scalar higgs boson in this model is larger than the one in the mssm without an extra @xmath0 symmetry .
moreover , it is also found that this tree - level mass of the lightest scalar higgs boson in this model can be significantly affected by radiative corrections .
recently , this model has been further investigated within the context of the supersymmetric cp problem [ 20 ] , as it can have explicit cp violating phases in the higgs sector . at the tree level , like the mssm
, this model has no cp mixing between the cp even and the cp odd states in the higgs sector because the single cp phase arising from the tree - level higgs potential can always be eliminated by rotating the higgs fields . in ref . [ 20 ] , the one - loop contribution from the top and stop quark loops are considered in the explicit cp violation scenario .
it is observed that the lightest neutral higgs boson remains essentially cp even for a wide parameter space , with a fixed value of @xmath9 , where @xmath10 is the ratio of the vevs of two higgs doublets .
meanwhile , large cp mixtures for the other heavier higgs bosons may be realized as the size of the effective @xmath2 terms becomes large . in this paper , motivated by the results of ref .
[ 20 ] , we would like to study this model in some detail in the explicit cp violation scenario .
the radiative cp mixing may be generated by means of the complex phases coming from the one - loop effective potential due to top quarks , bottom quarks , the exotic quarks , and their superparticles , where the exotic quarks and squarks with electric charges @xmath11 are introduced into this model in order to cancel gauge anomaly [ 21 ] . in particular , we are interested in the contributions of exotic quark sector in this model . at the one - loop level ,
the exotic quarks and squarks are found to play some recognizable role in the scalar - pseudoscalar mixings .
we find that the relevant cp phase from the exotic quark sector may change the size of the cp mixing between the heaviest scalar and the pseudoscalar higgs bosons by up to 20 % .
also we find that this model yields a negative or extremely small axion mass , as well as other unacceptable predictions , at the one - loop level for full parameter space in the spontaneous cp violation scenario .
thus , we note that the present model is impossible to realize spontaneous cp violation at the one - loop level .
gut , such as the string - inspired @xmath12 model , might be a natural motivation to enlarge the gauge symmetry in order to accommodate an extra @xmath0 symmetry into the mssm .
the @xmath12 gauge group may be decomposed into @xmath13 where @xmath14 is further broken down to the standard model ( sm ) gauge group , @xmath15 . at the electroweak scale
, the desired extra @xmath0 symmetry may be given as an orthogonal linear combination of @xmath16 and @xmath17 as @xmath18 where @xmath19 , @xmath20 , and @xmath21 are the @xmath0 , @xmath16 , and @xmath17 charges , respectively .
sometimes , for four different values of the angle @xmath22 , the four particular combinations of the @xmath16 and @xmath17 are called as follows : the @xmath23-model for @xmath24 , the @xmath25-model for @xmath26 , the @xmath27-model for @xmath28 , and the @xmath29-model for @xmath30 .
the gauge group we consider is thus @xmath31 .
the higgs sector of the mssm with an extra @xmath0 symmetry consists of two higgs doublets @xmath32 and @xmath33 , and a neutral higgs singlet @xmath6 .
this model possesses an extra pair of @xmath34 singlet quarks , @xmath35 and @xmath36 , with electric charges @xmath37 and @xmath38 , respectively , introduced by the requirement of the gauge anomaly cancellation . for the fermion matter fields in this model ,
we take into account only the third generation of quarks , besides the exotic quarks .
then , the superpotential for the model we consider may be expressed as @xmath39 where @xmath40 is the left - handed quark doublet , @xmath41 and @xmath42 are the charge conjugate of the right - handed top quark and bottom quark , respectively , with @xmath43 and @xmath44 being their respective yukawa coupling coefficients , and @xmath45 .
all the coupling coefficients are dimensionless .
this superpotential is symmetric under @xmath46 , and the @xmath0 symmetry is broken down by the vev of @xmath6 .
the higgs potential of this model at the tree level may expressed as a sum of the @xmath47-term , the @xmath48-term , and the soft breaking term , that is , @xmath49 where @xmath50 \ , \cr v_d & = & { g_2 ^ 2 \over 8 } ( h_1^{\dagger } \vec\sigma h_1 + h_2^{\dagger } \vec\sigma h_2)^2 + { g_1 ^ 2 \over 8 } ( |h_1|^2 - |h_2|^2)^2 \cr & & \mbox{}+ { g^{'2}_1 \over 2 } ( { \tilde q}_1 |h_1|^2 + { \tilde q}_2 |h_2|^2 + { \tilde q}_3 |s|^2)^2 \ ,
\cr v_{\rm s } & = & m_1 ^ 2 |h_1|^2 + m_2 ^ 2 |h_2|^2 + m_3 ^ 2 |s|^2 - [ \lambda a_{\lambda } ( h_1 h_2 ) s + { \rm h.c . } ] \ , \end{aligned}\ ] ] where @xmath51 ( @xmath52 = 1,2,3 ) are soft masses , @xmath53 , @xmath54 , and @xmath55 are respectively the coupling coefficients of @xmath34 , @xmath56 , and @xmath0 ; @xmath57 are the pauli matrices ; @xmath58 , @xmath59 , and @xmath60 are respectively the @xmath0 charges of @xmath61 , @xmath62 , and @xmath6 , satisfying the condition of @xmath63 to obey the gauge invariance of the superpotential under @xmath0 . at the electroweak scale , the three neutral higgs fields develop vevs as @xmath64 , @xmath65 , and @xmath66 , where @xmath67 is the relative phase between @xmath68 and @xmath6 . in general , @xmath69 in the soft breaking term may be complex , like @xmath70 . however , it is always possible to make the phase @xmath71 and the phase @xmath67 cancel out each other by redefining the higgs singlet @xmath6 , and thus the tree - level higgs potential can be made free of any complex phase .
therefore , the cp symmetry is not violated in this model at the tree level .
now let us consider the one - loop radiative corrections to the tree - level higgs potential . in supersymmetric models ,
the incomplete cancellation between ordinary particles and their superpartners yield the one - loop corrections to the tree - level higgs boson masses .
generally , the most dominant part of the one - loop corrections to the tree - level higgs potential come primarily from the top and stop quark loops . by considering the top quark sector only
, it has been observed that explicit cp violation in this model is viable through the radiatively corrected higgs potential [ 20 ] . for large @xmath72
, the contribution of the bottom and sbottom quark loops can also be significantly large .
we would like to consider in this paper the one - loop contributions from both the top and bottom quark sector , as well as from the exotic quark sector . to start with
, we calculate the tree - level masses of the relevant particles in order to evaluate the radiative corrections to the neutral higgs sector .
the fermion matter fields obtain their masses after the electroweak symmetry breaking as @xmath73 for top quark , @xmath74 for bottom quark , and @xmath75 for the exotic quark .
the tree - level masses of their superpartners are given as @xmath76 where @xmath77 , @xmath78 .
further , @xmath79 , @xmath80 , and @xmath81 are the soft susy breaking masses respectively for stop quarks , sbottom quarks , and the exotic squarks , and likewise @xmath82 , @xmath83 , and @xmath84 are respectively their trilinear soft susy breaking parameters of the mass dimension .
one can notice that there are three phases @xmath85 , @xmath86 , and @xmath87 in the above expressions .
these complex phases are determined by the generally complex @xmath88 ( @xmath89 ) and the phase of @xmath90 , the vev of the higgs singlet .
we note that the @xmath48-terms do not contribute to the squark masses .
the full higgs potential at the one - loop level may be written as @xmath91 where @xmath92 is the radiative corrections due to the relevant particles and their superpartners . according to the effective potential method [ 22 ] , @xmath92 is given as @xmath93 \ , \ ] ] where @xmath94 is the renormalization scale in the modified minimal subtraction scheme , and the subscript @xmath95 stands for various participating particles : @xmath96 , @xmath97 , @xmath98 , @xmath99 , @xmath100 , @xmath101 , @xmath102 , @xmath103 and @xmath104 .
the degrees of freedom for these particles including the sign convention are @xmath105 and @xmath106 ( @xmath4 ) , since in the above formula enter fermions with a negative sign while bosons with a positive sign .
if cp violation takes place in the higgs sector , the neutral higgs bosons would not have definite cp parity , hence mixings among them . in case of explicit cp violation ,
the non - trivial tadpole minimum condition with respect to the pseudoscalar component of the higgs field is given as @xmath107 where the dimensionless function @xmath108 arising from radiative corrections is defined as @xmath109 + 1 \ , \ ] ] and the phase @xmath110 is given by @xmath111 , with @xmath71 being the phase of @xmath69 in @xmath112 and @xmath67 being the phase of @xmath90 .
note that @xmath110 is generally nonzero at the one - loop level whereas at the tree level it can be made zero by rotating the higgs singlet . in the above tadpole minimum condition
, the first term comes from the tree - level higgs potential , and the remaining three terms come respectively from the top squark , bottom squark , and exotic squark contributions . at the tree level ,
the higgs sector in the mssm with an extra @xmath0 has ten real degrees of freedom , which are decomposed by two neutral goldstone bosons , a pair of charged goldstone bosons , four neutral higgs bosons and a pair of charged higgs bosons . after the electroweak symmetry breaking
, the two neutral goldstone bosons and a pair of charged goldstone bosons will eventually be absorbed into the longitudinal component of @xmath8 , @xmath113 , and @xmath114 gauge bosons .
since the cp symmetry is conserved in the higgs sector at the tree level , the four neutral higgs bosons can be divided into three scalar higgs bosons and one pseudoscalar higgs boson , according to the cp parity .
the squared mass matrix @xmath115 of the four neutral higgs bosons is given as a symmetric @xmath116 matrix that is obtained by the second derivatives of the higgs potential with respect to the four higgs fields , in the ( @xmath117 ) basis .
at the tree level , @xmath115 is given by @xmath118 as @xmath119 where the tree - level squared mass of the pseudoscalar higgs boson is given as @xmath120 with @xmath121 standing for the splitting between an extra @xmath0 symmetry breaking scale and the electroweak scale .
note that @xmath122 is divided into two blocks of the upper - left @xmath123 submatrix and the single element @xmath124 , corresponding to the scalar part and the pseudoscalar part .
there is no scalar - psuedoscalar mixing in @xmath122 , since @xmath125 ( @xmath126 ) are zero , and hence no cp violation .
explicit calculations yield @xmath127 at the one - loop level , the squared mass matrix may be decomposed as @xmath128 where @xmath129 has exactly the same appearance as @xmath122 but @xmath130 in @xmath122 should be replaced by @xmath131 in @xmath129 , where @xmath131 is given as @xmath132 \ , \end{aligned}\ ] ] one may notice that in the expression for @xmath131 , the first term comes from the tree - level higgs potential and the remanning three terms comes from the contributions of the squarks .
therefore , @xmath129 contains not only the tree - level results but also the contributions from radiative corrections .
now , let us calculate @xmath133 , which is obtained from @xmath92 .
we may conveniently decompose @xmath133 as @xmath134 where symmetric matrices @xmath135 , @xmath136 , and @xmath137 are obtained respectively from the top quark sector , the bottom quark sector , and the exotic quark sector contributions to @xmath92 , after imposing the tadpole minimum condition .
somewhat lengthy calculations yield the matrix elements of @xmath135 as @xmath138 where @xmath139 and the dimensionless function @xmath140 is defined as @xmath141 likewise , the matrix elements of @xmath136 as @xmath142 where @xmath143 and the matrix elements of @xmath137 as @xmath144 where @xmath145 notice that the matrix elements @xmath146 , @xmath147 , and @xmath148 ( @xmath126 ) are proportional to @xmath149 , @xmath150 , and @xmath151 , respectively .
if the three cp phases are not all zero , these matrix elements would not be all zero , and therefore the matrix elements @xmath152 ( @xmath126 ) would not be zero , which are responsible for the scalar - psuedoscalar mixing and hence cp violation .
in short , these three cp phases in the radiative corrections generate the cp mixing between the scalar and pseudoscalar components , implying that the four neutral higgs bosons are not states of definite cp parity .
the four physical neutral higgs bosons at the one - loop level are defined as the eigenstates of the squared mass matrix @xmath115 , with their squared masses as the eigenvalues of @xmath115 .
let us denote the physical four neutral higgs bosons as @xmath153 ( @xmath52= 1 to 4 ) and their squared masses as @xmath154 ( @xmath52 = 1 to 4 ) .
we assume that @xmath155 for @xmath156 . among them , the upper bound on @xmath157 , the squared mass of the lightest neutral higgs boson , may be obtained by noticing that the smallest eigenvalue of a positive symmetric matrix can not exceed the smaller eigenvalue of its upper left @xmath158 submatrix .
thus , the upper bound on @xmath157 at the one - loop level is given as @xmath159 where the first three terms come from the tree - level higgs potential while the other terms come from the one - loop corrections due to top quark , bottom quark , the exotic quark and their superparticles .
for numerical analysis , we assume that the extra @xmath0 emerges from the @xmath12 group , as described in the previous section . to be concrete , we take the @xmath160-model [ 19 ] , where @xmath0 is a mixture of @xmath16 and @xmath17 with the mixing angle @xmath30 . at the electroweak scale ,
the analysis of renormalization group equation leads to @xmath161 for @xmath0 gauge coupling constant [ 19 ] .
then , the @xmath0 charges of @xmath61 , @xmath62 , and @xmath6 are given respectively as @xmath162 , @xmath163 , and @xmath164 .
we set quark masses as @xmath165 = 175 gev , @xmath166 = 4 gev , and @xmath167 = 700 gev .
the renormalization scale is set as @xmath94 = 700 gev in the one - loop effective potential .
we assume that the lighter squarks are larger than top quark mass . for the remaining parameters
, we set the ranges for their variations as follows : @xmath168 , @xmath169 , @xmath170 , @xmath171 , @xmath172 , and @xmath173 .
note that one need not vary @xmath110 since it appears always together with @xmath174 in the definition of @xmath175 .
also note that the upper bound on @xmath7 is chosen by considering that the the maximum value of @xmath7 is about 0.83 in the analysis of the renormalization group equation . a general comment on @xmath90
may be that a large @xmath90 leads to a heavy @xmath113 in the present model .
moreover , the large @xmath90 can satisfy the experimental constraint that the mixing angle between @xmath8 and @xmath113 should be smaller than 2 - 3 @xmath176 .
we calculate the upper bound on @xmath177 by varying all the relevant parameters within their allowed ranges for a given @xmath178 .
the results are shown in fig .
1 , where the upper bound on @xmath177 is plotted as a function of @xmath72 , the solid curve is obtained by considering only the tree - level higgs potential , whereas the dashed curve is obtained by considering the full one - loop higgs potential with explicit cp violation .
it is observed in the mssm that the lightest neutral higgs boson mass increases steadily as @xmath72 increases . in the present model , where the mssm is incorporated with an extra @xmath0 ,
the lightest neutral higgs boson mass does not increase in accordance with the increase of @xmath72 , because of both the higgs singlet and @xmath179 contributions due to an extra @xmath0 .
one can see that in fig .
1 the lightest neutral higgs boson mass becomes maximum at @xmath180 for both curves . as we have remarked in the previous section , the scalar - pseudoscalar mixing , which is responsible for cp violation in this model ,
is triggered by three cp phases , @xmath85 , @xmath86 , and @xmath87 .
the size of the cp mixing between the scalar and pseudoscalar higgs bosons could be regarded as maximal if @xmath181 = 1 , while there would be no cp mixing if @xmath181 = 0 . in order to see a more clear picture about the amount of the cp mixing among the scalar and pseudoscalar higgs bosons , we introduce ( @xmath52 = 1,2,3 ) @xmath182 which depend implicitly on other parameters as well .
these @xmath183 measure by what amount @xmath87 contributes to the size of cp mixing . in order to express the relative contributions by @xmath87
, we further introduce @xmath184 by evaluating @xmath185 , we may study the role of the exotic quark sector in the cp violation in the present model . in fig .
2 , we plot @xmath186 as a function of @xmath87 , where we fix @xmath187 gev and @xmath188 , and we set for simplicity @xmath189 . the four curves in fig .
2 correspond to four different sets of @xmath189 and @xmath72 , namely , the solid curve is obtained for @xmath190 and @xmath191 , the dashed curve for @xmath190 and @xmath192 , the dotted curve for @xmath193 and @xmath191 , and the dash - dotted curve for @xmath193 and @xmath192 .
it is trivial to notice that @xmath194 at @xmath195 , @xmath196 , or @xmath197 . for other values of @xmath87 , we find that @xmath186 may change up to as much as about @xmath198 for some parameter values .
this implies that the effect of @xmath87 upon @xmath186 , or @xmath199 , is very significant . in other words ,
the matrix element @xmath200 depends critically on the complex phase of the exotic quark sector .
we study the behaviors of @xmath201 and @xmath202 for @xmath203 , fixing the values of other parameters as in fig .
2 . unlike @xmath186
, they do not fluctuate widely against the variation of @xmath87 .
thus , it can be said that the complex phase of the exotic quark sector has a strong influence on the magnitude of the mixing between @xmath204 and @xmath205 , whereas it has relatively weak influence on the magnitudes of the mixing between @xmath206 and @xmath205 or between @xmath207 and @xmath205 .
we calculate the masses of the neutral higgs bosons for the same parameter setting as in fig . 2 : @xmath187 gev , @xmath188 , together with four different sets of @xmath189 and @xmath72 , for @xmath203 .
we obtain approximately 126 , 1000 , 1002 , and 1043 gev , respectively , for @xmath177 , @xmath208 , @xmath209 , and @xmath210 .
we find that these values are almost unchanged for @xmath203 , and nearly the same within a few gev for the four different sets of @xmath189 and @xmath72 .
the mass of the extra gauge boson @xmath113 is also estimated .
we find that @xmath211 is stable at approximately 1039 gev for the parameter values in fig .
the mixing angle between @xmath8 and @xmath113 , denoted as @xmath212 , depends on both @xmath72 and @xmath90 .
we find that @xmath213 for @xmath214 , and and slightly increases for larger values of @xmath178 as @xmath215 for @xmath216 .
we have studied the dependence of the mixing element between @xmath204 and @xmath205 on the complex phase of the exotic quark sector .
meanwhile , one can evaluate effectively the absolute size of the cp violation by calculating the dimensionless parameter @xmath217 where @xmath218 are the elements of the orthogonal matrix which diagonalizes the squared mass matrix for the neutral higgs bosons .
the range of @xmath219 goes from 0 to 1 since the elements of the orthogonal matrix satisfy the orthogonality condition of @xmath220 .
if @xmath221 , there would be no cp violation , whereas the cp symmetry would be maximally violated if @xmath222 . the maximal cp violation that leads to @xmath219 = 1 takes place when @xmath223 . in order to figure out the dependence of @xmath219 on @xmath87 , we introduce @xmath224 we calculated @xmath225 as a function of @xmath87 , for the same parameter setting as fig .
2 . if @xmath225 remains at 1 , it would imply that @xmath225 does not depend on @xmath87 .
on the other hand , @xmath87 would contribute more significantly to cp violation if @xmath225 moves farther away from 1 .
the result is shown in fig .
the four curves correspond to the four sets of parameter values .
3 shows that the fluctuation of @xmath225 is larger for @xmath192 than for @xmath191 .
this implies that @xmath87 play an important role in @xmath225 for large @xmath72 .
one can see that @xmath225 fluctuates by up to 20 % for nonzero @xmath87 .
since @xmath226 is an absolute measure for the cp violation in our model , it is reasonable to expect that the complex phase of the exotic quark sector may change up to 20 % of the cp violation .
in this section , let us examine briefly whether it is possible to realize the spontaneous cp violation in the present model . at initial stage
, the lagrangian density of the model is assumed to be invariant with respect to cp property .
after the electroweak symmetry breaking , the cp symmetry may spontaneously be broken by complex phases in the vevs of the neutral higgs fields . in the mssm with an extra @xmath0
, one can conjecture from the georgi - pais theorem that the spontaneous cp violation leads to a very light pseudoscalar higgs boson , namely , the axion .
it has been addressed that the spontaneous cp violation through radiative corrections can be realized when at the tree level there exist massless higgs bosons other than goldstone boson [ 23 ] . in the present model , there is no pseudo - glodstone boson , which is referred to as axion , and thus the spontaneous cp violation can not occur at the one - loop level . in order to examine the possibility of spontaneous cp violation scenario at the one - loop level
, we can quantitatively analyze the axion mass by using the formulae which are derived for explicit cp violation scenario .
in spontaneous cp violation scenario , the one - loop effective potential of the present model may have only one physical cp phase ( @xmath67 ) arising from the vev of the higgs singlet by considering the radiative corrections due to top quark , bottom quark , the exotic quark , and their superpartners . at the one - loop level , we obtain the minimum condition for the vacuum with respect to @xmath67 as @xmath227 for @xmath228 .
we obtain the axion mass at the one - loop level as @xmath229 the massless function @xmath140 in the above expression is found to have negative or very small values for the entire parameter space we consider , in the case of explicit cp violation scenario . employing this result from the explicit cp violation scenario ,
the squared mass of axion is almost always negative for the entire parameter space , which is unacceptable .
therefore , the spontaneous cp violation is not viable for the mssm with an extra @xmath0 .
furthermore , the lightest neutral higgs boson mass also faces difficulty in the spontaneous cp violation scenario . calculate the lightest neutral higgs boson mass from the @xmath116 squared mass matrix for neutral higgs bosons , we find that the squared mass of the lightest higgs boson has negative values for the above parameter space .
in fact , the minimum condition for the vacuum with respect to @xmath67 has a very strong constraint , and thus the vacuum becomes unstable for the full parameter space we consider , in the spontaneous cp violation scenario .
we consider cp violation in the mssm with an extra @xmath0 symmetry , which is assumed to be originated from a string - inspired @xmath12 model .
the charge of the extra @xmath0 is defined as @xmath230 with @xmath231 .
this model possesses one higgs singlet and exotic quarks and squarks , in addition to the particle content of the mssm . in this model ,
the cp violation is impossible to occur either explicitly or spontaneously in its the tree - level higgs sector .
the mixing between the scalar and pseudoscalar higgs bosons , which induce the cp violation , may occur via the radiative corrections at the one - loop level .
for the radiative corrections , we consider the loop contributions due to top quark , bottom quark , the exotic quark and their superpartners .
the radiative corrections contain in general three complex phases originated from the top squark sector , the bottom squark sector , and the exotic squark sector .
we pay attention to the complex phase @xmath87 that comes from the exotic squark sector .
we calculate the lightest neutral higgs boson mass . in explicit cp violation scenario
, we find that the mass of the lightest neutral higgs boson is smaller than about 135 gev and 165 gev , respectively , at the tree level and at the one - loop level , for the whole parameter space that we considered . for four different sets of parameters , we calculate the magnitude of the cp mixing among the scalar and pseudoscalar higgs bosons for @xmath203 .
the matrix element , @xmath200 , representing the cp mixing between the heaviest scalar and the pseudoscalar higgs bosons , is found to depend heavily on the complex phase from the exotic squark sector , @xmath87 . in particular , the magnitude of the cp mixing between the heaviest scalar and the pseudoscalar higgs bosons is found to change as much as about 20 % for @xmath232 when @xmath193 and @xmath192 , whereas the cp mixing among the other scalar higgs bosons and the pseudoscalar higgs boson is found to be relatively stable against the variation of @xmath87 . also , we have investigated the possibility of spontaneous cp violation at the one - loop level in the neutral higgs sector , where the tree - level higgs potential possesses cp symmetry .
we argue that it is impossible to achieve spontaneous cp violation in the radiatively corrected higgs potential of this model .
0.3 in 0.2 in this research is supported by kosef through chep .
the authors would like to acknowledge the support from kisti ( korea institute of science and technology information ) under `` the strategic supercomputing support program '' with dr .
kihyeon cho as the technical supporter .
the use of the computing system of the supercomputing center is also greatly appreciated .
nilles , phys .
rep . * 110 * , 1 ( 1984 ) ; j.f .
gunion , h.e .
haber , g.l .
kane , and s. dawson , _ the higgs hunters guide _ ( addison - wesley redwood city , ca , 1990 ) .
s. weinberg , phys .
37 * , 657 ( 1976 ) .
a. pomarol , phys .
b * 287 * , 331 ( 1992 ) ; n. maekawa , phys .
b * 282 * , 387 ( 1992 ) ; n. haba , phys .
b * 398 * , 305 ( 1997 ) .
o. lebedev , eur . phys .
j. c * 4 * , 363 ( 1998 ) .
a. pilaftsis , phys .
b * 435 * , 88 ( 1998 ) ; a. pilaftsis , phys .
d * 58 * , 096010 ( 1998 ) .
demir , phys .
d * 60 * , 055006 ( 1999 ) ; s.y .
choi , m. drees , and j.s .
lee , phys .
b 481 , 57 ( 2000 ) . g.l . kane and l.t .
wang , phys . lett .
b * 488 * , 383 ( 2000 ) .
s. heinemeyer , eur .
. j. c * 22 * , 521 ( 2001 ) .
t. ibrahim and p. nath , phys .
d * 63 * , 035009 ( 2001 ) ; t. ibrahim and p. nath , phys .
d * 66 * , 015005 ( 2002 ) ; s.w .
ham , s.k .
yoo , c.m .
kim , and d. son , phys .
d * 68 * , 055003 ( 2002 ) .
lee , a. pilaftsis , m. carena , s.y .
choi , m. drees , j.r .
ellis , and , c.e.m .
wagner , comput .
* 156 * , 283 ( 2004 ) .
ellis , j.s .
lee , and a. pilaftsis , phys .
d * 72 * , 095006 ( 2005 ) .
a. pilaftsis and c.e.m .
wagner , nucl . phys . *
b553 * , 3 ( 1999 ) ; m. carena , j. ellis , a. pilaftsis , and c.e.m .
wagner , nucl .
b586 * , 92 ( 2000 ) ; s.w .
ham , s.k .
yoo , and h.k .
lee , j. phys .
g * 27 * , 1 ( 2001 ) ; m. carena , j.r .
ellis , a. pilaftsis , and c.e.m .
wagner , phys . lett .
b * 495 * , 155 ( 2000 ) .
kim and h.p .
nilles , phys .
b * 138 * , 150 ( 1984 ) .
j. ellis , j.f .
gunion , h.e .
haber , l. roszkowski , and f. zwirner , phys .
d * 39 * , 844 , ( 1989 ) .
j.l . hewett and t.g .
rizzo , phys .
rep . * 183 * , 193 ( 1989 ) ; a. leike , phys .
rep . * 317 * , 143 ( 1999 ) .
m. cvetic and p. langacker , phys .
d * 54 * , 3570 ( 1996 ) .
m. cvetic , d.a .
demir , j.r .
espinosa , l. everett , and p. langacker , phys .
d * 54 * , 3570 ( 1996 ) .
d.a . demir and n.k .
pak , phys .
d * 57 * , 6609 ( 1998 ) ; y. daikoku and d. suematsu , phys .
d * 62 * , 095006 ( 1998 ) ; h. amini , new j. phys .
* 5 * , 49 ( 2003 ) .
s.f . king , s. moretti , and r. nevzorov , phys .
d * 73 * , 035009 ( 2006 ) ; s.f . king , s. moretti , and r. nevzorov , phys .
b * 634 * , 278 ( 2006 ) .
d.a . demir and l.l .
everett , phys .
d * 69 * , 015008 ( 2004 ) .
j. erler , nucl .
* b586 * , 73 ( 2000 ) . s. coleman and e. weinberg , phys .
d * 7 * , 1888 ( 1973 ) .
h. georgi and a. pais , phys .
d * 10 * , 1246 ( 1974 ) . 0.3 in fig . 1
: the upper bound on the lightest neutral higgs boson mass as a function of @xmath72 , where other relevant parameters are allowed to vary within their respective ranges : @xmath168 , @xmath169 , @xmath170 , @xmath171 , @xmath172 , and @xmath173 .
the quark masses are set as @xmath165 = 175 gev , @xmath166 = 4 gev , @xmath167 = 700 gev , and the renormalization scale is set as @xmath94 = 700 gev .
the solid and dashed curves are respectively obtained from the tree - level and the one - loop level higgs potential with explicit cp violation . 0.3 in fig .
plot of @xmath186 as a function of @xmath87 , for four different sets of @xmath189 and @xmath72 : @xmath190 and @xmath191 ( solid curve ) , @xmath190 and @xmath192 ( dashed curve ) , @xmath193 and @xmath191 ( dotted curve ) , and @xmath193 and @xmath192 ( dash - dotted curve ) .
the remaining parameters are set as @xmath187 gev and @xmath188 . , where other relevant parameters are allowed to vary within their respective ranges : @xmath168 , @xmath169 , @xmath170 , @xmath171 , @xmath172 , and @xmath173 .
the quark masses are set as @xmath165 = 175 gev , @xmath166 = 4 gev , @xmath167 = 700 gev , and the renormalization scale is set as @xmath94 = 700 gev .
the solid and dashed curves are respectively obtained from the tree - level and the one - loop level higgs potential with explicit cp violation . ] as a function of @xmath87 , for four different sets of @xmath189 and @xmath72 : @xmath190 and @xmath191 ( solid curve ) , @xmath190 and @xmath192 ( dashed curve ) , @xmath193 and @xmath191 ( dotted curve ) , and @xmath193 and @xmath192 ( dash - dotted curve ) .
the remaining parameters are set as @xmath187 gev and @xmath188 . ] | we study that a minimal supersymmetric standard model with an extra @xmath0 gauge symmetry may accommodate the explicit cp violation at the one - loop level through radiative corrections .
this model is cp conserving at the tree level and can not realize the spontaneous cp violation for a wide parameter space at the one - loop level . in explicit cp violation scenario , we calculate the higgs boson masses and the magnitude of the scalar - pseudoscalar mixings in this model at the one - loop level by taking into account the contributions of top quarks , bottom quarks , exotic quarks , and their superpartners .
in particular , we investigate how the exotic quarks and squarks would affect the scalar - pseudoscalar mixings .
it is observed that the size of the mixing between the heaviest scalar and pseudoscalar higgs bosons is changed up to 20 % by a complex phase originated from the exotic quark sector of this model . |
arp & madore ( 1987 , hereafter am87 ) undertook a fairly comprehensive survey of the southern sky in search of peculiar and interacting galaxies visible on the iiia - j photographic survey plates as part of the uk schmidt sky survey .
they inspected some 100,000 galaxies and culled out about 7,000 objects that were noteworthy enough to be cataloged , and in a number of cases , individually illustrated . based on
this statistically robust sample am87 went beyond simple cataloging and added a broad list of descriptors covering about two dozen properties .
one of these descriptor / classifications was category 6 : ring galaxy .
as already noted by am87 `` ring galaxy '' , in the isolated context of this particular catalog , was a purely morphologically defined descriptor ; as a consequence this descriptor encompasses several physically distinct types of galaxies .
first , there are relatively normal galaxies with `` outer rings '' . in the rc3 system of de vaulcouleurs et al .
these are typed as * rs * galaxies , and called `` outer ringed '' galaxies in the compilation by buta ( 1995 ) .
outer rings are not uncommon and are generally well understood ( see buta for an overview ) .
second , there are `` polar ring galaxies '' ; systems thought to be advanced mergers with the disrupted satellite now orbiting the parent galaxy over the pole and forming the ring - like structure .
these are exceedingly rare objects ( see whitmore et al .
1990 for a recent compilation and references therein . ) finally , there are the collisional systems ( called `` ring '' in rc3 ) ; and they are the focus of this paper . in an earlier paper ( which was based on an incomplete , early - release version of the arp - madore catalogue )
few , madore & arp ( 1986 ) ( hereafter fma ) pointed out the two dominant types of rings in the catalog and separated them into two distinct classes : * o - types * for the outer ( `` resonant '' ) rings , and * p - types * for the classical ( `` collisional '' ) rings .
the main conclusion of that paper was that the * p - type * ( ring ) galaxies were consistent with the head - on collision scenario , being championed in that same time period by theys & spiegel ( 1976 ) and independently by lynds & toomre ( 1976 ) .
the full success of that theory has broader implications which are explored in madore et al .
all three of the current authors independently inspected and reclassified each of the 500 + category 6 ( ring galaxies ) in am87 . only * p - type
* galaxies ( on the fma classification system ) were retained for this study .
that is , any object with a crisp , relatively high - surface - brightness ring , with the nucleus often asymmetrically placed ( or in some cases absent ) , was given the p classification and retained .
objects with more diffuse or feathery outer rings ( of generally somewhat lower surface brightness ) , with fairly large nuclear regions or bulges symmetrically placed with respect to the ring , were deemed to be the * o types*. these objects were returned to the general pool . out of 552 class 6 ( ring ) objects in am87 we culled out 104 * p - type * ring galaxies ; and for only a handful of objects was there any disagreement among the authors as to their classification .
erring on the conservative / inclusive side , all of the contentious objects were retained in this study .
table 1 lists all 104 pure * p - type * ( ring ) galaxies found in am87 .
we note here that pure ring galaxies constitute less than 2% of the am87 peculiar galaxy population , and as such they are found in the general population of galaxies only 0.01% of the time ( as normalized by the entire population of @xmath0100,000 galaxies inspected by am87 in the course of constructing the catalog as a whole . )
ring galaxies are a rare type of collision with short - lived evidence .
it is of course interesting to contemplate what ring galaxies look like after the collisionally - induced burst of star formation fades .
we leave that as an exercise for the future . rather than simply illustrating a few of the most spectacular examples of ring galaxies , as was done in am87 , we have chosen to re - catalog and present the entire sample .
however , before doing that we decided to additionally survey the literature for previously cataloged ring galaxies and add them to the _ atlas _ for completeness sake . the am87 sample is given in table 1 , while the ring galaxies from the literature are collected in table 2 .
clearly the northern and equatorial sky is incompletely sampled .
we estimate that about 200 more ring galaxies await discovery once those regions are inspected to the same apparent - size and surface - brightness limit as am87 .
table 1 and 2 share the same basic formating .
column 1 contains the name of the host ring galaxy complex ; the second column contains the name or names of the ring nucleus followed by as many companions as are suspected of being colliders .
the name of the ring nucleus is composed of the name of the host object followed by rn ( e.g. , am 0147 - 350:rn ) ; by way of contrast , companion / colliders are identified by the host object name followed by c1 , c2 , etc .
( e.g. , am 0147 - 350:c1 , am 0147 - 350:c2 , am 0147 - 350:c3 ) .
columns 4 and 5 respectively contain ra(2000 ) and dec(2000 ) positions , each measured to a precision of 5 arcsec . in column 5
the first entry ( d = diameter ) , corresponds to the diameter of the ring , measured in arcsec , while the following entries in that same column but now for the companions are marked by s , and give the separation of the companion from the nucleus of the ring ( again , measured in arcsec ) .
finally , column 6 gives the heliocentric velocity for the object / component when available from the literature as captured by ned in its december 2008 release .
all of the objects in this new listing consist of at least two objects in the immediate field of view that can reasonably be identified with the collider and the ( collidee ) ring .
we mark those objects in each of the illustrations and for uniformity sake we have measured the positions of all objects given in tables 1 and 2 using the display capabilities of aladin interacting with the ned image archive .
the presence of nearby companions was by no means a foregone conclusion given that the rings were chosen purely on the basis of their morphology , not on the basis of their nearby or distant environment .
we consider it highly likely that these optical companions are the colliders , plausibly responsible for the `` gravitational splash '' and the ensuing ring - like , star - formation response predicted / post - dicted by the simulations . indeed , of the 64 companions with published redshifts ( found from a search of the january 2008 release of the ned database ) 60 have redshifts that place them at the same distance as the ring itself .
in all four of the other instances where one of the apparent companions proved to be a background object there was always an additional candidate collider in the same field of view .
although common radial velocities are not conclusive proof that any given object is the collider it does clearly demonstrate that the vast majority of those objects are not chance line - of - sight background ( or foreground ) objects .
111 apparent companions still need redshifts ; 43 of the rings themselves are without published velocities .
we are working to ameliorate that situation . while all of our ring galaxies have one or more apparent companions , as a class ring galaxies tend not to be found in high - density ( cluster ) environments .
indeed , only 7 of the 132 rings cataloged here have more than three companions in their immediate vicinity .
they are am 0401 - 641 , am 0417 - 391 , am 0455 - 465 , am 1003 - 215 , am 1251 - 283 , am 2100 - 725 and am 2200 - 715 .
83 of the rings have only one nearby optical companion , 29 have two companions , and only 13 have three .
it seems reasonable to suppose that in dense systems the more frequent , but off - center encounters with other group members will disrupt and/or destroy any ring that might form from a direct hit more quickly than in the quiescent field .
perhaps the increased velocity dispersion of all the members associated with the cluster environment also plays a role .
and it should not be forgotten that very rich clusters are _ currently _ depleted of disk galaxies in any case : with no disks to be hit , no rings can be formed . however , at high redshift the situation is expected to be quite different . a search for rings in various `` deep fields '' ( udf , goods and gems ) has been made ( elmegreen & elmegreen 2006 ) . according to those authors and judging from the published images of their pure `` rings '' ( their figure 1 ) these are almost exclusively * o - type * ( outer - ring ) galaxies with well - centered bulges and no obvious companions .
some collisional objects may be found in their sample of fifteen `` partial rings '' ; however , the area shown around each partial ring is too small to judge whether these objects have interacting companions or not .
figure 1 shows the wide range of absolute ( k - band ) magnitudes that the ring galaxies in this catalog have .
spanning more than a factor of 100 in luminosity galaxies with ring - like morphology can be as faint as m@xmath1 = -19 mag and as bright as m@xmath1 = -25 mag ; typical ring galaxies have m@xmath1 = -23 mag which is somewhat fainter than the knee in the general - field luminosity function . the colliders are typically half a magnitude fainter than the central ( target ) galaxy ( figure 2 ) , but their dispersion is large ( @xmath20.75 mag ) .
the current diameters of the rings cataloged here span a wide range being broadly peaked at 30 - 40 kpc , but extending up to values as large as 70 kpc ( figure 3 ) .
none of these plots argue for separate populations of ring galaxies based solely on these physical properies .
some ring galaxies ( e.g. , am 0058 - 220 , am 1953 - 260 , arp 147 & vii zw 466 ) do not have any obvious component that can unambiguously be identified as a nucleus .
the morphology of such objects has been aptly likened to `` smoke rings . ''
the apparent lack of a nucleus may be because the target galaxy simply never had a particularly noteworthy nucleus to begin with ( i.e. , even before the collision ) or it may be that the timing of the interaction , combined with our particular viewing angle places the nucleus in the ring ( see ngc 4774 , arp 146 ) and/or even confused with the intruder . in other instances the nucleus may have been disrupted by the collision .
detailed modeling would illuminate these possibilities .
we draw attention to a sub - class of objects in this compilation of ring galaxies : double rings . am 2136 - 492 is noteworthy .
the cartwheel ( am 0035 - 335 ) is another example ( with its celebrated compact nuclear ring ) ; while am 0339 - 625 , am 1323 - 222 and am 1354 - 250 may be other less obvious examples of this class .
figure 6 shows a deeply stretched image of am 0339 - 625 illustrating its fragmented , outer ring - like feature .
although multiple rings structures seem to be a common feature in the numerical simulations ( e.g. antunes & wallin 2007 ) they appears not to be not so common in nature .
a simple calculation concerning the apparent frequency of ring galaxies leads to a rather surprising / interesting result .
the relative frequency of the collisonal ring phenomenon is at least 1 in 1000 , given that am87 inspected approximately 100,000 galaxies and out of those cataloged only 100 rings .
( and this is a lower limit given that edge - on rings are probably under - represented in the catalog . ) if the ring structure can unambiguously be detected as such for one dynamical timescale ( @xmath3 yrs , say ) and if the collision rate is constant over cosmic time , then about one in every ten galaxies in this same volume is expected to have been perturbed in this same way over a hubble time .
given the restricted parameter space available for a head - on collision to occur this strongly suggests that , if the orbits of these types of companions are isotropic ( and only a small fraction of them produce rings ) then the majority of galaxies in the local volume have presumably been hit ( with a different impact parameter ) by a companion comparable in mass to that of a typical ring - galaxy collider . in figure 4
we show the histogram of relative velocities from those ring galaxies with redshift measurements for both the ring itself and at least one of its putative colliders .
the median velocity difference for this sample is rather small ( @xmath0100 km / s ) . excluding obvious background objects with @xmath4v @xmath5 1,000 km / s , the largest two remaining velocity differences are 293 and 419 km / s .
we conclude that the majority of these systems are bound and that if simulations are a guide then there will be a rapid merging of these pairs in the course of the next passage .
computer simulations of galaxy interactions first showed that ring galaxies could convinciungly be created by the head - on collision of a satellite galaxy with a primary disk .
observations of prototypical ring galaxies , such as the cartwheel ( fosbury & hawarden 1977 ) and the lindsay - shapley ring , each confirmed the `` prediction '' in as much as colliders were found in evidence , and within the range of distances , velocities and position angles expected for head - on collisional events .
the present study follows the early work of few , madore & arp ( 1986 ) and extends this confirmation of theory from a handful of galaxies to a relative complete sample of over 100 suspected and previously cataloged ring galaxies .
however , there is another prediction that is fundamentally at odds with what is observed for this sample of ring galaxies .
that prediction comes from @xmath6-cdm cosmology in which many more satellite galaxies are expected to be orbiting central galaxies than are in fact observed .
making these satellites non - luminous ( be it by supernova events , by interactions , by initial conditions , or by fiat ) does not prevent them from interacting gravitationally , and as such they too should be responsible for creating rings through collisions in direct proportion to their numbers with respect to their luminous colliding counterparts .
the astonishing fact is that _ there is not a single compelling example of a ring galaxy without a plausible optical collider .
_ there is no isolated cartwheel in the sky ; there are no smoke rings without the smoking gun . and there should be many .
we state this simple fact and leave a more detailed discussion of this point and its implications for @xmath6-cdm cosmology to the companion paper by madore et al .
( 2009 ) .
we wish to acknowledge and thank the creators of saoimage ds9 and aladin both of which were used extensively in the preparation of this _
atlas_. and , of course , this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration . finally , we thank ian steer for bringing the ring galaxy ic 1908 to our attention .
am 0034 - 351&am 0034 - 351:rn & 00:37:09.3 & -34:56:15 & d = 31 & 8,891 + & am 0034 - 351:c1 & 00:37:04.7 & -34:56:24 & s = 58 & . .
+ & & & & & + am 0035 - 335&am 0035 - 335:rn & 00:37:41.1 & -33:42:59 & d = 58 & 9,050 + & am 0035 - 335:c1 & 00:37:44.7 & -33:42:20 & s = 61 & 9,104 + & am 0035 - 335:c2 & 00:37:43.1 & -33:42:08 & s = 60 & 8,639 + & & & & & + am 0053 - 353&am 0053 - 353:rn & 00:56:19.3 & -35:14:53 & d = 47 & 14,494 + & am 0053 - 353:c1 & 00:56:15.7 & -35:15:15 & s = 43 & 14,420 + & & & & & + am 0054 - 634&am 0054 - 634:rn & 00:56:51.1 & -63:28:51 & d = 27 & 11,557 + & am 0054 - 634:c1 & 00:56:45.5 & -63:29:16 & s = 20 & 11,526 + & am 0054 - 634:c2 & 00:56:48.8 & -63:28:38 & s = 45 & . . . + & & & & & + am 0058 - 220&am 0058 - 220:rn & 01:00:46.0 & -21:43:30 & d = 17 & . . .
+ & am 0058 - 220:c1 & 01:00:47.2 & -21:43:26 & s = 17 & 12,524 + & & & & + am 0058 - 311&am 0058 - 311:rn & 01:00:44.5 & -33:55:49 & d = 34 & 23,650 + & am 0058 - 311:c1 & 01:00:48.1 & -33:55:49 & s = 46 & 23,684 + & & & & & + am 0058 - 665&am 0058 - 665:rn & 01:00:15.1 & -66:37:07 & d = 36 & 21,055 + & am 0058 - 665:c1 & 01:00:32.3 & -66:37:49 & s = 112 & . . .
+ & am 0058 - 665:c2 & 01:00:29.4 & -66:36:11 & s = 103 & . . . + &
& & & & + am 0102 - 610&am 0102 - 610:rn & 01:04:56.3 & -60:48:58 & d = 22 & . .
+ & am 0102 - 610:c1 & 01:04:53.8 & -60:49:17 & s = 26 & . . .
+ & & & & & + am 0115 - 463&am 0115 - 463:rn & 01:18:09.9 & -46:14:32 & d = 33 & 17,688 + & am 0115 - 463:c1 & 01:18:09.7 & -46:15:07 & s = 34 & . .
+ & & & & & + am 0126 - 525&am 0126 - 525:rn & 01:28:24.6 & -52:38:08 & d = 29 & 3,700 + & am 0126 - 525:c1 & 01:28:25.2 & -52:38:11 & s = 06 & 3,450 + & & & & & + am 0147 - 350&am 0147 - 350:rn & 01:49:43.5 & -34:49:53 & d = 52 & 8,244 + & am 0147 - 350:c1 & 01:49:47.2 & -34:48:57 & s = 73 & . . .
+ & am 0147 - 350:c2 & 01:49:41.4 & -34:49:25 & s = 39 & . . .
+ & am 0147 - 350:c3 & 01:49:39.2 & -34:50:08 & s = 55 & bckgrnd + & & & & & + am 0154 - 353&am 0154 - 353:rn & 01:57:08.8 & -35:18:47 & d = 23 & . . .
+ & am 0154 - 353:c1 & 01:57:08.9 & -35:18:51 & s = 06 & . . .
+ am 0155 - 815&am 0155 - 815:rn & 01:54:41.4 & -81:37:53 & d = 38 & . .
+ & am 0155 - 815:c1 & 01:53:00.8 & -81:37:57 & s = 43 & . .
+ & am 0155 - 815:c2 & 01:52:39.7 & -81:37:50 & s = 05 & . . .
+ & & & & & + am 0157 - 311&am 0157 - 311:rn & 01:59:39.5 & -31:01:11 & d = 29 & 17,742 + & am 0157 - 311:c1 & 01:59:37.8 & -31:01:22 & s = 24 & . .
+ & & & & & + am 0200 - 581&am 0200 - 581:rn & 02:01:50.8 & -57:56:52 & d = 27 & . .
+ & am 0200 - 581:c1 & 02:01:50.7 & -57:57:01 & s = 10 & . . .
+ & am 0200 - 581:c2 & 02:01:51.4 & -57:56:56 & s = 06 & . . .
+ & & & & & + am 0240 - 242&am 0240 - 242:rn & 02:42:32.0 & -24:13:52 & d = 26 & 25,310 + & am 0240 - 242:c1 & 02:42:32.0 & -24:13:57 & s = 05 & . .
+ & am 0240 - 242:c2 & 02:42:33.4 & -24:13:05 & s = 46 & . . .
+ & & & & & + am 0305 - 824&am 0305 - 824:rn & 03:01:44.3 & -82:35:30 & d = 34 & . .
+ & am 0305 - 824:c1 & 03:02:01.8 & -82:35:26 & s = 34 & . . .
+ & & & & & + am 0318 - 610&am 0318 - 610:rn & 03:19:37.5 & -60:57:01 & d = 17 & . .
+ & am 0318 - 610:c1 & 03:19:37.9 & -60:57:10 & s = 09 & . . .
+ am 0322 - 374&am 0322 - 374:rn & 03:23:54.6 & -37:30:36 & d = 43 & 1,405 + & am 0322 - 374:c1 & 03:23:53.7 & -37:30:47 & s = 15 & 1,397 + & & & & & + am 0339 - 625&am 0339 - 625:rn & 03:40:37.8 & -62:46:49 & d = 30/62 & 14,860 + & am 0339 - 625:c1 & 03:40:39.6 & -62:46:36 & s = 18 & . .
+ & & & & & + am 0401 - 641&am 0401 - 641:rn & 04:01:36.9 & -64:04:03 & d = 40 & . .
+ & am 0401 - 641:c1 & 04:01:43.4 & -64:05:29 & s = 16 & . . .
+ & am 0401 - 641:c2 & 04:01:54.9 & -64:06:10 & s = 85 & 12,457 + & am 0401 - 641:c3 & 04:01:37.2 & -64:04:38 & s = 78 & . . .
+ & am 0401 - 641:c4 & 04:01:46.7 & -64:05:54 & s = 28 & . .
+ & & & & & + am 0403 - 555&am 0403 - 555:rn & 04:04:16.4 & -55:46:23 & d = 40 & . .
+ & am 0403 - 555:c1 & 04:04:17.0 & -55:46:06 & s = 18 & 17,825 + & am 0403 - 555:c2 & 04:04:12.1 & -55:45:20 & s = 80 & . . .
+ & & & & & + am 0404 - 271&am 0404 - 271:rn & 04:06:28.8 & -27:09:21 & d = 25 & 26,443 + & am 0404 - 271:c1 & 04:06:29.4 & -27:09:24 & s = 08 & . . .
+ & & & & & + & & & & & + am 0413 - 590&am 0413 - 590:rn & 04:14:25.9 & -58:57:45 & d = 43 & 17,143 + & am 0413 - 590:c1 & 04:14:27.0 & -58:57:40 & s = 09 & . .
+ & & & & & + am 0415 - 475&am 0415 - 475:rn & 04:16:29.9 & -47:50:54 & d gt 84 & 10,079 + & am 0415 - 475:c1 & 04:16:18.6 & -47:49:12 & s = 152 & . .
+ & am 0415 - 475:c2 & 04:16:31.6 & -47:50:43 & s = 23 & . . .
+ & & & & & + am 0417 - 391&am 0417 - 391:rn & 04:19:39.6 & -39:10:31 & d = 31 & 15,255 + & am 0417 - 391:c1 & 04:19:38.9 & -39:10:26 & s = 09 & . . .
+ & am 0417 - 391:c2 & 04:19:39.9 & -39:10:23 & s = 10 & 15,255 + & am 0417 - 391:c3 & 04:19:40.5 & -39:08:40 & s = 114 & 14,820 + & am 0417 - 391:c4 & 04:19:40.4 & -39:11:39 & s = 70 & . . .
+ & am 0417 - 391:c5 & 04:19:51.3 & -39:10:05 & s = 140 & 15,185 + & & & & & + am
0425 - 421&am 0425 - 421:rn & 04:26:36.6 & -42:05:38 & d = 55 & 4,491 + & am 0425 - 421:c1 & 04:26:44.6 & -42:05:42 & s = 91 & . .
. + & & & & & + am 0437 - 314&am 0437 - 314:rn & 04:39:40.0 & -31:42:39 & d = 28 & . .
+ & am 0437 - 314:c1 & 04:39:40.3 & -31:42:38 & s = 05 & . . .
+ & & & & & + am 0437 - 644&am 0437 - 644:rn & 04:37:54.9 & -64:38:59 & d = 14 & . .
+ & am 0437 - 644:c1 & 04:37:56.9 & -64:39:06 & s = 33 & . . .
+ & & & & & + am 0438 - 503&am 0438 - 503:rn & 04:39:20.5 & -50:31:50 & d = 38 & . .
+ & am 0438 - 503:c1 & 04:39:22.1 & -50:31:45 & s = 15 & . . .
+ & & & & & + am 0438 - 661&am 0438 - 661:rn & 04:38:37.1 & -66:13:03 & d = 34 & 14,650 + & am 0438 - 661:c1 & 04:38:37.9 & -66:13:49 & d = 12 & . .
+ & & & & & + am 0455 - 465&am 0455 - 465:rn & 04:57:01.7 & -46:45:00 & d = 15 & . .
+ & am 0455 - 465:c1 & 04:57:04.1 & -46:44:54 & s = 26 & . . .
+ & am 0455 - 465:c2 & 04:57:06.5 & -46:45:03 & s = 49 & . . .
+ & am 0455 - 465:c3 & 04:57:00.1 & -46:46:00 & s = 60 & . . .
+ & am 0455 - 465:c4 & 04:56:55.8 & -46:44:10 & s = 83 & . . .
+ & am 0455 - 465:c5 & 04:57:06.5 & -46:46:08 & s = 80 & . . .
+ & & & & & + am 0507 - 512&am 0507 - 512:rn & 05:08:28.0 & -51:21:29 & d = 34 & . .
+ & am 0507 - 512:c1 & 05:08:27.6 & -51:20:38 & s = 43 & . . .
+ & am 0507 - 512:c2 & 05:08:27.3 & -51:21:22 & s = 11 & . . .
+ & am 0507 - 512:c3 & 05:08:29.5 & -51:21:42 & s = 20 & . . .
+ am 0520 - 390&am 0520 - 390:rn & 05:22:42.5 & -39:03:48 & d = 43 & 15,114 + & am 0520 - 390:c1 & 05:22:43.4 & -39:03:25 & s = 26 & . .
+ & am 0520 - 390:c2 & 05:22:48.3 & -39:03:38 & s = 68 & . . .
+ & am 0520 - 390:c3 & 05:22:44.3 & -39:03:41 & s = 23 & . . .
+ & & & & & + am 0521 - 434&am 0521 - 434:rn & 05:22:30.5 & -43:46:17 & d = 30 & 24,291 + & am 0521 - 434:c1 & 05:22:30.4 & -43:45:51 & s = 26 & . .
+ & & & & & + am 0545 - 355&am 0545 - 355:rn & 05:47:22.0 & -35:49:29 & d = 42 & 13,885 + & am 0545 - 355:c1 & 05:47:19.4 & -35:49:45 & s = 36 & . . .
+ & & & & & + am 0545 - 434&am 0545 - 434:rn & 05:47:00.0 & -43:44:56 & d = 38 & . .
+ & am 0545 - 434:c1 & 05:47:01.2 & -43:44:33 & s = 26 & . .
+ & & & & & + am 0642 - 645&am 0642 - 645:rn & 06:42:34.2 & -64:58:28 & d = 20 & . .
+ & am 0642 - 645:c1 & 06:42:32.7 & -64:59:24 & s = 05 & . . .
+ & am 0642 - 645:c2 & 06:42:33.4 & -64:59:28 & s = 05 & . . .
+ & & & & & + & & & & & + & & & & & + am 0642 - 801&am 0642 - 801:rn & 06:38:38.4 & -80:14:50 & d = 67 & 4,777 + & am 0642 - 801:c1 & 06:38:36.6 & -80:14:49 & s = 08 & . . .
+ & am 0642 - 801:c2 & 06:38:27.9 & -80:14:59 & s = 29 & . . .
+ & & & & & + am 0643 - 462&am 0643 - 462:rn & 06:45:03.3 & -46:26:56 & d = 60 & 12,056 + & am 0643 - 462:c1 & 06:45:00.2 & -46:26:43 & s = 36 & . . .
+ & & & & & + am 0644 - 741&am 0644 - 741:rn & 06:43:05.6 & -74:14:12 & d = 97 & 6,505 + & am 0644 - 741:c1 & 06:43:06.1 & -74:12:55 & s = 78 & 7,000 + & am 0644 - 741:c2 & 06:43:25.5 & -74:15:27 & s = 110 & 6,750 + & am 0644 - 741:c3 & 06:44:17.5 & -74:16:37 & s = 326 & 6,430 + & & & & & + am 0755 - 785&am 0755 - 785:rn & 07:53:31.5 & -79:06:28 & d = 27 & . . .
+ & am 0755 - 785:c1 & 07:53:35.6 & -79:06:28 & s = 13 & . . .
+ & & & & & + am 0814 - 760&am 0814 - 760:rn & 08:13:31.6 & -76:17:29 & d = 28 & . .
+ & am 0814 - 760:c1 & 08:13:30.9 & -76:17:41 & s = 13 & . . . + & & & & & + & & & & & + & & & & & + am 1003 - 215&am 1003 - 215:rn & 10:05:43.2 & -22:05:43 & d = 28 & 26,329 + & am 1003 - 215:c1 & 10:05:43.3 & -22:05:53 & s = 12 & . . . + &
am 1003 - 215:c2 & 10:05:43.4 & -22:05:03 & s = 41 & . .
+ & am 1003 - 215:c3 & 10:05:44.1 & -22:06:19 & s = 38 & . . .
+ & am 1003 - 215:c4 & 10:05:46.2 & -22:05:52 & s = 43 & . . .
+ & & & & & + am 1006 - 380&am 1006 - 380:rn & 10:09:05.2 & -38:24:35 & d = 80 & 4,943 + & am 1006 - 380:c1 & 10:09:08.0 & -38:23:48 & s = 58 & 4,481 + & & & & & + am 1025 - 370&am 1025 - 370:rn & 10:27:13.8 & -37:25:17 & d = 27 & . .
+ & am 1025 - 370:c1 & 10:27:13.8 & -37:25:33 & s = 16 & . . .
+ & & & & & + am 1133 - 245&am 1133 - 245:rn & 11:35:30.8 & -25:08:44 & d = 67 & 11,704 + & am 1133 - 245:c1 & 11:35:17.6 & -25:08:45 & s = 180 & 11,682 + & am 1133 - 245:c2 & 11:35:34.5 & -25:05:41 & s = 189 & 11,601 + & & & & & + am 1135 - 284&am 1135 - 284:rn & 11:37:42.2 & -29:05:01 & d = 24 & 5,558 + & am 1135 - 284:c1 & 11:37:36.2 & -29:04:11 & s = 97 & . .
+ & & & & & + & & & & & + am 1152 - 421&am 1152 - 241:rn & 11:55:21.7 & -42:29:19 & d = 46 & 4,617 + & am 1152 - 241:c1 & 11:55:21.3 & -42:29:17 & s = 07 & . .
+ & & & & + am 1159 - 530&am 1159 - 530:rn & 12:01:35.9 & -53:21:28 & d = 55 & . .
+ & am 1159 - 530:c1 & 12:01:32.3 & -53:20:47 & s = 52 & . . .
+ & & & & + am 1249 - 462&am 1249 - 462:rn & 12:52:38.7 & -46:38:59 & d = 50 & . .
+ & am 1249 - 462:c1 & 12:52:34.9 & -46:40:04 & s = 78 & . . .
+ & & & & & + am 1251 - 283&am 1251 - 283:rn & 12:54:42.4 & -28:51:57 & d = 29 & 17,003 + & am 1251 - 283:c1 & 12:54:43.1 & -28:52:17 & s = 24 & 16,242 + & am 1251 - 283:c2 & 12:54:43.1 & -28:50:17 & s = 100 & 16,359 + & am 1251 - 283:c3 & 12:54:36.5 & -28:51:28 & s = 82 & bckgrnd + & am 1251 - 283:c4 & 12:54:40.9 & -28:53:03 & s = 69 & 16,576 + & am 1251 - 283:c5 & 12:54:49.7 & -28:52:37 & s = 105 & foregrnd + & & & & & + am 1300 - 412&am 1300 - 412:rn & 13:03:00.3 & -41:42:16 & d = 18 & 3,422 + & am 1300 - 412:c1 & 13:03:00.4 & -41:42:20 & s = 05 & . .
+ & & & & & + & & & & & + am 1308 - 253&am 1308 - 253:rn & 13:11:08.7 & -25:54:00 & d = 44 & 13,556 + & am 1308 - 253:c1 & 13:11:08.5 & -25:54:00 & s = 05 & . .
+ & am 1308 - 253:c2 & 13:11:07.4 & -25:54:28 & s = 85 & . . .
+ & & & & & + am 1323 - 222&am 1323 - 222:rn & 13:26:20.3 & -22:37:54 & d = 88 & 4,627 + & am 1323 - 222:c1 & 13:26:29.4 & -22:33:49 & s = 272 & bckgrnd + & & & & & + am 1325 - 251&am 1325 - 251:rn & 13:27:58.1 & -25:31:49 & d = 18 & 10,936 + & am 1325 - 251:c1 & 13:27:57.7 & -25:31:44 & s = 07 & . .
+ & am 1325 - 251:c2 & 13:27:57.1 & -25:30:50 & s = 52 & . . .
+ & & & & & + am 1354 - 250&am 1354 - 250:rn & 13:57:13.7 & -25:14:44 & d = 53 & 6,150 + & am 1354 - 250:c1 & 13:57:17.8 & -25:13:30 & s = 97 & 6,175 + & am 1354 - 250:c2 & 13:57:24.6 & -25:13:59 & s = 157 & . . .
+ & & & & & + am 1358 - 221&am 1358 - 221:rn & 14:01:08.0 & -22:33:35 & d = 53 & 10,940 + & am 1358 - 221:c1 & 14:01:13.2 & -22:32:45 & s = 87 & . . .
+ & am 1358 - 221:c2 & 14:01:12.8 & -22:31:52 & s = 121 & . . .
+ & am 1358 - 221:c3 & 14:01:25.6 & -22:34:20 & s = 247 & 10,786 + & & & & & + am 1413 - 243&am 1413 - 243:rn & 14:16:09.3 & -24:50:30 & d = 35 & 13,740 + & am 1413 - 243:c1 & 14:16:04.9 & -24:49:21 & s = 94 & . .
+ & am 1413 - 243:c2 & 14:16:09.3 & -24:51:09 & s = 37 & . . .
+ & am 1413 - 243:c3 & 14:16:13.3 & -24:50:47 & s = 56 & . . .
+ & & & & + am 1425 - 234&am 1425 - 234:rn & 14:55:09.8 & -23:58:09 & d = 61 & . .
+ & am 1425 - 234:c1 & 14:55:09.4 & -23:58:16 & s = 09 & . . .
+ & am 1425 - 234:c2 & 14:55:08.0 & -23:56:24 & s = 112 & . . .
+ & am 1425 - 234:c3 & 14:55:00.6 & -23:55:44 & s = 195 & . . .
+ & & & & & + am 1434 - 783&am 1434 - 783:rn & 14:40:27.5 & -78:48:35 & d = 40 & 4,625 + & am 1434 - 783:c1 & 14:40:32.2 & -78:48:28 & s = 40 & . .
+ & & & & & + am 1452 - 234&am 1452 - 234:rn & 14:55:09.8 & -23:58:09 & d = 61 & 12,304 + & am 1452 - 234:c1 & 14:55:09.4 & -23:58:16 & s = 09 & . .
+ & am 1452 - 234:c2 & 14:55:08.0 & -23:56:24 & s = 112 & . . .
+ & am 1452 - 234:c3 & 14:55:00.6 & -23:55:44 & s = 195 & . . .
+ & & & & & + am 1514 - 362&am 1514 - 362:rn & 15:17:48.2 & -36:34:56 & d = 24 & . .
+ & am 1514 - 362:c1 & 15:17:47.8 & -36:35:00 & s = 06 & 7,108 + am 1627 - 824&am 1627 - 824:rn & 16:35:56.7 & -82:50:17 & d = 32 & . .
+ & am 1627 - 824:c1 & 16:36:01.0 & -82:50:12 & s = 11 & . . .
+ & am 1627 - 824:c2 & 16:35:17.2 & -82:51:17 & s = 94 & . .
+ & & & & & + am 1724 - 622&am 1724 - 622:rn & 17:29:09.6 & -62:26:45 & d = 125 & 4,641 + & am 1724 - 622:c1 & 17:29:25.3 & -62:28:50 & s = 167 & 4,800 + & & & & & + am 1827 - 625&am 1827 - 625:rn & 18:32:02.7 & -62:55:44 & d = 33 & 4,086 + & am 1827 - 625:c1 & 18:32:02.6 & -62:55:47 & s = 04 & . .
. + & & & & & + am 1854 - 490&am 1854 - 490:rn & 18:58:06.8 & -49:00:41 & d = 14 & 4,250 + & am 1854 - 490:c1 & 18:58:05.5 & -49:00:52 & s = 18 & . .
+ & & & & & + am 1947 - 445&am 1947 - 445:rn & 19:51:19.8 & -44:52:40 & d = 62 & 5,805 + & am 1947 - 445:c1 & 19:51:26.3 & -44:52:36 & s = 141 & 5,672 + & & & & & + am 1953 - 260&am 1953 - 260:rn & 19:56:28.4 & -25:54:56 & d = 40 & . . .
+ & am 1953 - 260:c1 & 19:56:29.0 & -25:55:16 & s = 24 & 14,802 + & & & & & + & & & & & + am 1957 - 394&am 1957 - 394:rn & 20:00:36.0 & -39:39:02 & d = 51 & . .
+ & am 1957 - 394:c1 & 20:00:35.4 & -39:39:02 & s = 05 & . . .
+ & & & & & + am 2012 - 282&am 2012 - 282:rn & 20:15:13.9 & -28:18:25 & d = 25 & 7,051 + & am 2012 - 282:c1 & 20:15:13.6 & -28:18:24 & s = 06 & . .
+ & am 2012 - 282:c2 & 20:15:14.2 & -28:18:19 & s = 08 & . . .
+ & & & & & + am 2021 - 724&am 2021 - 724:rn & 20:26:29.2 & -72:36:46 & d = 28 & . .
+ & am 2021 - 724:c1 & 20:26:33.7 & -72:36:08 & s = 42 & 22,706 + & & & & & + am 2024 - 544&am 2024 - 544:rn & 20:27:53.9 & -54:38:00 & d = 61 & 8,100 + & am 2024 - 544:c1 & 20:27:57.0 & -54:37:51 & s = 28 & . . .
+ & am 2024 - 544:c2 & 20:27:48.5 & -54:38:28 & s = 55 & . . .
+ & & & & & + am 2026 - 424&am 2026 - 424:rn & 20:29:32.2 & -42:30:24 & d = 61 & . .
+ & am 2026 - 424:c1 & 20:29:33.5 & -42:30:24 & s = 17 & 15,486 + & & & & & + am 2033 - 260&am 2033 - 260:rn & 20:36:20.5 & -25:56:52 & d = 23 & 12,500 + & am 2033 - 260:c1 & 20:36:23.9 & -25:57:29 & s = 60 & 12,257 + & am 2033 - 260:c2 & 20:36:20.7 & -25:56:41 & s = 10 & . . . + am 2034 - 483&am 2034 - 483:rn & 20:38:08.7 & -48:23:16 & d = 22 & . .
+ & am 2034 - 483:c1 & 20:38:06.2 & -48:23:16 & s = 26 & . . .
+ & am 2034 - 483:c2 & 20:38:02.8 & -48:22:55 & s = 63 & . . .
+ & & & & & + am 2044 - 691&am 2044 - 691:rn & 20:48:56.8 & -69:05:32 & d = 42 & 11,413 + & am 2044 - 691:c1 & 20:48:56.6 & -69:05:29 & s = 03 & . .
+ & & & & & + am 2056 - 392&am 2056 - 392:rn & 21:00:07.3 & -39:17:57 & d = 48 & . .
+ & am 2056 - 392:c1 & 21:00:15.8 & -39:20:05 & s = 150 & . . .
+ & & & & & + am 2100 - 725&am 2056 - 392:rn & 21:05:38.2 & -72:47:09 & d = 27 & . .
+ & am 2056 - 392:c1 & 21:05:54.5 & -72:47:20 & s = 72 & 20,984 + & am 2056 - 392:c2 & 21:05:52.9 & -72:47:22 & s = 65 & . . .
+ & am 2056 - 392:c3 & 21:05:49.5 & -72:46:21 & s = 69 & . . .
+ & am 2056 - 392:c4 & 21:05:58.8 & -72:47:59 & s = 103 & . . .
+ & am 2056 - 392:c5 & 21:05:16.2 & -72:47:58 & s = 109 & . .
+ & am 2056 - 392:c6 & 21:05:11.5 & -72:47:04 & s = 119 & . . .
+ & am 2056 - 392:c7 & 21:05:48.2 & -72:47:29 & s = 48 & . . . + & & & & & + & & & & & + am 2107 - 474&am 2107 - 472:rn & 21:10:31.0 & -47:30:32 & d = 31 & 5,060 + & am 2107 - 472:c1 & 21:10:31.4 & -47:30:36 & s = 05 & . . .
+ & & & & & + am 2128 - 302&am 2128 - 302:rn & 21:31:08.2 & -30:16:27 & d = 84 & 10,037 + & am 2128 - 302:c1 & 21:31:01.5 & -30:19:40 & s = 208 & 9,199 + & & & & & + am 2131 - 254&am 2131 - 254:rn & 21:34:29.2 & -25:28:38 & d = 33 & 16,218 + & am 2131 - 254:c1 & 21:34:31.1 & -25:28:46 & s = 29 & . . .
+ & & & & & + am 2131 - 495&am 2131 - 495:rn & 21:34:37.6 & -49:42:46 & d = 27 & 21,645 + & am 2131 - 495:c1 & 21:34:37.2 & -49:42:32 & s = 15 & . .
+ & & & & & + am 2132 - 535&am 2132 - 535:rn & 21:36:11.1 & -53:41:26 & d = 32 & . .
+ & am 2132 - 535:c1 & 21:36:13.0 & -53:42:54 & s = 91 & . . .
+ & am 2132 - 535:c2 & 21:36:27.1 & -53:43:09 & s = 176 & . . .
+ & am 2132 - 535:c3 & 21:36:26.7 & -53:43:31 & s = 188 & . . .
+ & & & & & + am 2134 - 471&am 2134 - 471:rn & 21:37:28.0 & -47:02:09 & d = 52 & 9,254 + & am 2134 - 471:c1 & 21:37:31.1 & -47:00:38 & s = 98 & . . .
+ & am 2134 - 471:c2 & 21:37:35.9 & -47:02:40 & s = 90 & . . .
+ am 2136 - 492&am 2136 - 492:rn & 21:39:25.6 & -49:09:36 & d = 40/72 & 15,990 + & am 2136 - 492:c1 & 21:39:31.6 & -49:08:11 & s = 104 & . .
+ & & & & & + am 2141 - 515&am 2141 - 515:rn & 21:45:15.1 & -51:41:54 & d = 33 & . .
+ & am 2141 - 515:c1 & 21:45:14.9 & -51:41:49 & s = 05 & . . .
+ & & & & & + am 2145 - 543&am 2145 - 543:rn & 21:49:08.5 & -54:18:14 & d = 22 & 19,544 + & am 2145 - 543:c1 & 21:49:08.2 & -54:17:53 & s = 22 & . .
+ & & & & & + am 2152 - 592&am 2152 - 592:rn & 21:56:19.4 & -59:07:15 & d = 15 & 22,351 + & am 2152 - 592:c1 & 21:56:17.5 & -59:07:17 & s = 15 & . . .
+ & am 2152 - 592:c2 & 21:56:20.8 & -59:06:51 & s = 27 & . . .
+ & & & & & + am 2155 - 263&am 2155 - 263:rn & 21:58:38.4 & -26:22:41 & d = 27 & 21,255 + & am 2155 - 263:c1 & 21:58:37.2 & -26:22:25 & s = 23 & . . .
+ & & & & & + & & & & & + & & & & & + & & & & & + & & & & & + am 2200 - 715&am 2200 - 715:rn & 22:04:57.5 & -71:42:14 & d = 32 & . . .
+ & am 2200 - 715:c1 & 22:04:22.4 & -71:43:20 & s = 177 & 9,350 + & am 2200 - 715:c2 & 22:04:12.7 & -71:35:08 & s = 478 & . .
+ & am 2200 - 715:c3 & 22:04:47.8 & -71:35:59 & s = 379 & . .
+ & am 2200 - 715:c4 & 22:05:51.5 & -71:39:51 & s = 288 & . .
+ & am 2200 - 715:c5 & 22:06:23.9 & -71:43:53 & s = 418 & . .
+ & am 2200 - 715:c6 & 22:05:39.0 & -71:46:45 & s = 330 & . .
+ & & & & & + am 2201 - 230&am 2201 - 230:rn & 22:04:19.3 & -22:47:33 & d = 24 & 21,732 + & am 2201 - 230:c1 & 22:04:19.5 & -22:47:48 & s = 17 & 21,675 + & & & & & + am 2220 - 493&am 2220 - 493:rn & 22:23:13.4 & -49:17:46 & d = 32 & . . .
+ & am 2220 - 493:c1 & 22:23:12.7 & -49:17:16 & s = 31 & 17,947 + & & & & & + am 2230 - 481&am 2230 - 481:rn & 22:33:47.0 & -48:01:30 & d = 52 & 10,476 + & am 2230 - 481:c1 & 22:33:43.3 & -48:01:31 & s = 37 & 10,361 + & & & & & + am 2238 - 541&am 2238 - 541:rn & 22:41:55.1 & -53:58:43 & d = 32 & . . .
+ & am 2238 - 541:c1 & 22:41:57.5 & -53:59:45 & s = 67 & . . .
+ & & & & & + am 2240 - 304&am 2240 - 304:rn & 22:43:36.9 & -30:28:50 & d = 22 & 17,449 + & am 2240 - 304:c1 & 22:43:36.3 & -30:28:55 & s = 08 & . .
+ & & & & & + am 2302 - 322&am 2302 - 322:rn & 23:05:39.9 & -32:09:07 & d = 25 & 18,137 + & am 2302 - 322:c1 & 23:05:41.9 & -32:08:11 & s = 66 & 18,077 + & am 2302 - 322:c2 & 23:05:43.9 & -32:09:46 & s = 59 & bckgrnd + & & & & & + am 2308 - 324&am 2308 - 324:rn & 23:11:17.3 & -32:27:07 & d = 33 & 11,327 + & am 2308 - 324:c1 & 23:11:16.9 & -32:27:21 & s = 15 & . . .
+ & & & & & + am 2313 - 261&am 2313 - 261:rn & 23:15:46.5 & -25:54:20 & d = 20 & 6,472 + & am 2313 - 261:c1 & 23:15:46.5 & -25:54:28 & s = 09 & . . .
+ & & & & & + am 2317 - 672&am 2317 - 672:rn & 23:20:09.7 & -67:08:58 & d = 32 & . .
+ & am 2317 - 672:c1 & 23:20:11.2 & -67:09:00 & s = 10 & . . .
+ & am 2317 - 672:c2 & 23:20:11.6 & -67:08:28 & s = 33 & . . .
+ & & & & & + am 2322 - 671&am 2322 - 671:rn & 23:25:42.8 & -67:03:29 & d = 19 & . .
+ & am 2322 - 671:c1 & 23:25:45.3 & -67:03:33 & s = 15 & . . .
+ & & & & & + am 2323 - 512&am 2323 - 512:rn & 23:26:31.8 & -51:08:07 & d = 21 & . .
+ & am 2323 - 512:c1 & 23:26:31.2 & -51:07:45 & s = 23 & . . .
+ & & & & & + am 2338 - 312&am 2338 - 312:rn & 23:41:26.0 & -31:03:22 & d = 54 & 17,538 + & am 2338 - 312:c1 & 23:41:19.2 & -31:03:01 & s = 91 & 17,268 + & & & & & + am 2343 - 703&am 2343 - 703:rn & 23:46:39.3 & -70:22:51 & d = 44 & . . .
+ & am 2343 - 703:c1 & 23:46:38.5 & -70:22:25 & s = 27 & . . .
+ & & & & & + am 2353 - 291&am 2353 - 291:rn & 23:56:23.8 & -29:01:25 & d = 36 & 8,931 + & am 2353 - 291:c1 & 23:56:25.1 & -29:01:24 & s = 17 & 8,953 + arp 146&arp 146:rn & 00:06:44.3 & -06:38:08 & d = 19 & 22,616 + & arp 146:c1 & 00:06:44.7 & -06:38:13 & d = 09 & . . .
+ & & & & & + arp 318&arp 318:rn & 02:09:24.4 & -10:08:10 & d = 40 & 4,073 + & arp 318:c1 & 02:09:20.7 & -10:08:00 & s = 55 & 3,864 + & arp 318:c2 & 02:09:38.3 & -10:08:48 & s = 208 & 3,851 + & arp 318:c3 & 02:09:42.7 & -10:11:02 & s = 319 & 3,934 + & & & & & + arp 010&arp 010:rn & 02:18:26.2 & + 05:39:14 & d = 44 & 9,108 + & arp 010:c1 & 02:18:28.6 & + 05:40:08 & s = 65 & . .
+ & & & & & + arp 273&arp 273:rn & 02:21:28.6 & + 39:22:32 & d = 100 & 7,563 + & arp 273:c1 & 02:21:32.7 & + 39:21:24 & s = 82 & 7,335 + & & & & & + arp 145&arp 145:rn & 02:23:07.7 & + 41:21:11 & d = 57 & 5,425 + & arp 145:c1 & 02:23:11.1 & + 41:22:04 & s = 36 & 5,852 + & & & & & + ngc 0985&ngc 0985:rn & 02:34:37.6 & -08:47:15 & d = 36 & 13,168 + & ngc 0985:c1 & 02:34:38.6 & -08:48:03 & s = 47 & . . .
+ & & & & & + arp 118&arp 118:rn & 02:55:12.0 & -00:11:03 & d = 30 & 8,648 + & arp 118:c1 & 02:55:09.5 & -00:10:40 & s = 45 & 8,459 + & arp 118:c2 & 02:55:06.5 & -00:09:46 & s = 114 & 8,417 + & & & & & + arp 147&arp 147:rn & 03:11:18.6 & + 01:18:51 & d = 17 & 9,656 + & arp 147:c1 & 03:11:19.4 & + 01:18:48 & s = 14 & 9,415 + & & & & & + ic 1908&ic 1908:rn & 03:15:05.4 & -54:49:09 & d = 30 & 8,234 + & ic 1908:c1 & 03:15:05.7 & -54:49:20 & s = 13 & . .
+ & & & & & + eso 200-ig 009&eso 200-ig 009:rn & 03:21:40.5 & -51:39:31 & d = 29 & 17,328 + & eso 200-ig 009:c1 & 03:21:41.0 & -51:39:49 & s = 21 & 20,980 + & eso 200-ig 009:c2 & 03:21:43.6 & -51:39:42 & s = 32 & . .
+ & & & & & + arp 219&arp 219:rn & 03:39:53.2 & -02:06:47 & d = 46 & 10,488 + & arp 219:c1 & 03:39:54.8 & -02:07:25 & s = 44 & . .
+ & & & & & + arp 141&arp 141:rn & 07:14:20.0 & + 73:28:25 & d = 75 & 2,728 + & arp 141:c1 & 07:14:20.2 & + 73:28:51 & s = 26 & 2,735 + & & & & & + arp 143&arp 143:rn & 07:46:55.0 & + 39:00:56 & d = 87 & 4,257 + & arp 143:c1 & 07:46:52.9 & + 39:01:55 & d = 65 & 4,048 + & arp 143:c2 & 07:46:55.5 & + 39:00:27 & d = 30 & 4,002 + & & & & & + ngc 2793&ngc 2793:rn & 09:16:47.1 & + 34:25:56 & d = 46 & . . .
+ & ngc 2793:c1 & 09:16:47.2 & + 34:25:48 & s = 11 & 1,687 + & ngc 2793:c2 & 09:16:40.8 & + 34:26:52 & s = 97 & bckgrnd + & ngc 2793:c3 & 09:16:46.6 & + 34:26:14 & s = 18 & 1,667 + & & & & & + arp 142&arp 142:rn & 09:37:44.1 & + 02:45:39 & d = 53 & 7,225 + & arp 142:c1 & 09:37:45.0 & + 02:44:51 & s = 52 & 6,806 + & arp 142:c2 & 09:37:41.2 & + 02:46:45 & s = 81 & . .
+ & & & & & + ic 0614&ic 0614:rn & 10:26:51.8 & -03:27:53 & d = 34 & 10,258 + & ic 0614:c1 & 10:26:51.9 & -03:27:39 & s = 16 & . .
+ & ic 0614:c2 & 10:26:49.0 & -03:29:19 & s = 97 & 18,097 + & & & & & + arp 107&arp 107:rn & 10:52:14.8 & + 30:03:28 & d = 55 & 10,372 + & arp 107:c1 & 10:52:18.4 & + 30:04:20 & d = 70 & 10,665 + & & & & & + arp 148&arp 148:rn & 11:03:52.4 & + 40:50:54 & d = 18 & 10,551 + & arp 148:c1 & 11:03:53.8 & + 40:51:00 & s = 17 & . .
+ & & & & & + arp 335&arp 335:rn & 11:04:23.5 & + 04:49:44 & d = 130 & 7,738 + & arp 335:c1 & 11:04:23.5 & + 04:49:29 & s = 15 & 7,763 + & arp 335:c2 & 11:04:16.5 & + 04:52:10 & s = 180 & 7,647 + & & & & & + vii zw 466&vii zw 466:rn & 12:32:04.9 & + 66:24:13 & d = 21 & 14,490 + & vii zw 466:c1 & 12:32:10.6 & + 66:24:19 & s = 34 & . .
+ & vii zw 466:c2 & 12:32:13.1 & + 66:23:59 & s = 50 & 14,100 + & vii zw 466:c3 & 12:32:11.6 & + 66:23:21 & s = 65 & 14,360 + & & & & & + ngc 4774&ngc 4774:rn & 12:53:06.4 & + 36:49:10 & d = 23 & 8,373 + & ngc 4774:c1 & 12:53:05.8 & + 36:49:34 & s = 25 & 8,373 + & & & & & + vv 256&vv 256:rn & 14:00:54.4 & + 40:59:19 & d = 64 & 3,738 + & vv 256:c1 & 14:00:56.2 & + 41:00:22 & s = 68 & 3,834 + & & & & & + kig 0686&kig 0686:rn & 15:32:56.9 & + 46:27:07 & d = 11 & . .
+ & kig 0686:c1 & 15:32:57.5 & + 46:27:14 & s = 09 & 656 + & & & & & + kig 0700&kig 0700:rn & 15:45:14.8 & + 00:46:24 & d = 19 & . .
+ & kig 0700:c1 & 15:45:14.3 & + 00:46:21 & s = 08 & 3,839 + & & & & & + arp 125&arp 125:rn & 16:38:13.9 & + 41:56:19 & d = 45 & 8,572 + & arp 125:c1 & 16:38:13.6 & + 41:55:51 & s = 31 & 8,404 + & & & & & + arp 150&arp 150:rn & 23:19:30.0 & + 09:30:30 & d = 34 & 11,879 + & arp 150:c1 & 23:19:31.0 & + 09:30:11 & s = 26 & 11,562 + & arp 150:c2 & 23:19:33.7 & + 09:29:42 & s = 75 & 11,637 + & arp 150:c3 & 23:19:27.7 & + 09:29:40 & s = 59 & 12,350 + & & & & & + arp 284&arp 284:rn & 23:36:14.2 & + 02:09:17 & d = 44 & 2,798 + & arp 284:c1 & 23:36:21.8 & + 02:09:25 & s = 111 & 2,771 + | we present a catalog and imaging atlas of classical ( collisional ) ring galaxies distilled from the arp - madore _ atlas of southern peculiar galaxies and associations _ and supplemented with other known ring galaxies from the published literature .
the catalog lists the original host object , compiles available redshifts and presents newly determined positions for the central ( target ) galaxy and its nearest companion(s ) .
127 collisional ring systems are illustrated and their components identified .
all of the rings have plausible colliders identified ; many are radial - velocity confirmed companions . finally , we make note of the existence of a rare sub - class of ring galaxies exemplified by am 2136 - 492 , double / concentric ring galaxies . these objects are predicted by numerical simulations , but they appear to be quite rare and/or short - lived in nature . |
it has been suggested that observations of black holes may be affected by general relativity ( gr ) and the frequency dependent properties of plasma @xcite and that these effects may be relevant for black holes in x - ray binary systems @xcite .
lensing effects modified by plasma surrounding galaxy scale lenses were also investigated by @xcite , requiring observations at mhz frequencies . in this work
we will study ray
tracing under the influence of both gravitational lensing and refractive plasma effects following @xcite , and examine frequency
dependent orbits for light rays .
deflection by both gravity and plasma was discussed by @xcite for the case of radio waves passing by the sun , using the product of the gravitational and plasma refractive indices to derive a weak field approximation .
@xcite measured the deflection of the radio loud quasar 3c273 as it was occulted by the sun and obtained an estimate of the electron density in the solar corona . in this work ,
we make use of the geometric optics approach developed by @xcite , which has application to a diverse range of phenomena @xcite .
fully covariant radiative transfer through a refractive medium was treated in depth by @xcite , and @xcite discuss dispersion relations applicable to ray propagation in relativistic plasmas near black holes and rapidly rotating neutron stars ( nss ) .
however , even the most basic dispersion effects and simple distributions of plasma near relativistic objects provide novel examples that differ from the vacuum behaviour of rays in the schwarzschild space time in interesting ways .
we review the theory and background of relativistic light bending including plasma effects in section [ sec : theory ] . in section [ sec : orbits ] , we study the trajectories of light rays through the plasma and consider waves near the plasma frequency with density varying as an inverse power law with index @xmath2 . when @xmath3 , the plasma frequency is constant and light rays behave like massive particles with mass given by the plasma frequency @xcite . when @xmath4 , we find that the plasma term acts like an additive contribution to the angular momentum .
the @xmath5 case has similarities with the behaviour of massless scalar fields on the schwarzschild space time .
section [ circularorbits ] includes a study of circular orbits for a general power - law index .
we apply the analytical results to the pulse profiles of a compact object sheathed in plasma in section [ sec : pulse ] , and review relevant properties of compact objects in section [ discussion ] .
in this work we use natural units , with @xmath6 , unless otherwise stated .
as we proceed through the derivations of the main equations , we keep the @xmath7 factors , but the numerical calculations are performed with @xmath8 .
we follow the method and notation of @xcite , and review the derivation of @xcite to find the generalized trajectories of photons in the presence of an isotropic , inhomogeneous plasma distribution , though more general possibilities exist @xcite .
let us assume spherical symmetry and adopt the line element @xmath9 with @xmath10 and @xmath11 for the schwarzschild metric .
we write the coordinate indices as @xmath12 , with the index @xmath13 to sum over the spatial coordinates and @xmath14 to run over all four . without loss of generality
we consider orbits in the equatorial plane and set @xmath15 . now , suppose a distribution of plasma surrounds the mass @xmath16 , which is located at the origin .
the plasma has an index of refraction @xmath17 where the photon frequency is @xmath18 , with effective energy @xmath19 with respect to the velocity of the plasma medium @xmath20 .
the linear momentum of the photon is @xmath21 , where @xmath22 are the elements of the wave four - vector .
the medium equation @xcite gives the relation between the components of the photon four - momentum and the refractive index of the medium , @xmath23 in the vacuum case @xmath24 and the standard condition for null geodesics @xmath25 is recovered .
we describe photon paths through the plasma in the schwarzschild space
time by the hamiltonian in the geometric optics limit . using the variational principle and the medium equation we have @xmath26=0 .
\label{generalhamiltonian}\ ] ] following @xcite , we assume the plasma is a static , inhomogeneous medium such that @xmath27 , @xmath28 .
this static condition simplifies the effective photon energy @xmath29 . with this condition
there is no time or @xmath30 dependence of the hamiltonian , so @xmath31 and @xmath32 are constant , and we restrict @xmath33 so that the orbit remains in the equatorial plane by symmetry .
the time - like component @xmath31 is related to the vacuum photon energy @xmath34 and for a static medium this gives @xmath35 with the effective redshift relationship @xmath36 we have @xmath37 the radial dependence of the photon frequency is due to the gravitational redshift supplied by the space
time metric , and we denote the photon frequency far from the origin @xmath38 as @xmath39 . for the sake of notational simplicity
we will suppress the radial dependence when referring to the frequency variables . in order to make analytical progress at this point
we will introduce a specific form for the plasma frequency .
let us assume that the refractive index is of the general form @xmath40 where @xmath41 . with equations ( [ nfreq ] ) and ( [ p^t ] ) ,
the hamiltonian in equation ( [ generalhamiltonian ] ) simplifies such that @xmath42 the paths of light rays are then described in terms of the affine parameter @xmath43 along the trajectory by @xmath44 and @xmath45 these equations give the positions and momenta of photons through the space
time under the influence of gravity and the optical medium .
the @xmath46 and @xmath30 components from equation ( [ vel ] ) can be combined to eliminate the affine parameter @xmath43 , resulting in @xmath47 the denominator can be simplified by using equation ( [ nullhamiltonian ] ) to give @xmath48 in terms of the constants @xmath31 and @xmath32 , such that @xmath49 here we take the positive solution corresponding to the case when both the @xmath30 and @xmath46 coordinates increase , and the negative solution with decreasing radial coordinate . to proceed ,
we write the constant @xmath32 in terms of the impact parameter @xmath50 , such that @xmath51 then we have @xmath52^{1/2}}. \label{dphidr}\ ] ] as expected , for @xmath53 , equation ( [ dphidr ] ) reduces to the vacuum case .
this equation gives the paths of photons in the equatorial plane in the general case of strong deflection which is appropriate for orbits near the surface of the star .
the refractive index of the plasma in this expression introduces frequency dependence to null geodesics @xcite . with the substitution @xmath54 , outgoing rays have @xmath55^{1/2 } } , \label{dphidu}\ ] ] where the potential dependence of @xmath56 on @xmath57 is explicitly stated .
now we consider an observer a great distance away at @xmath58 , such that the fiducial direction from the centre of the star to the observer corresponds to the impact parameter @xmath59 .
since we are interested in describing rays which leave the surface of the star ( @xmath60 ) and reach the observer ( @xmath61 ) , we define @xmath62^{1/2 } } , \label{delta}\ ] ] with the additional simplification @xmath63 , and taking the sign into account through the integration limits .
this expression gives us the maximum viewing angle on the stellar surface that can be seen by the observer .
in flat space @xmath64 always , which means exactly half the surface of the star is visible .
this is not the case in gr since gravitational lensing acts to redirect light rays such that a distant observer sees a larger apparent area and more of the stellar surface than would be possible in flat space .
in fact , in extreme cases the observer can see portions of the surface multiple times ( for @xmath65 ) .
it is also useful to describe the impact parameters of the rays which connect the surface of the star with the observer .
let us define @xmath66 as the angle between the radial and angular components of the photon momentum .
at the surface of the star @xmath67 , this is the angle with respect to the surface normal that the ray is emitted @xmath68 using equation ( [ pr ] ) leads to @xmath69 where the maximum impact parameter @xmath70 occurs for @xmath71 .
this formula reduces to the standard expression for the emission angle in vacuum when @xmath24 @xcite .
the presence of the refractive index within the impact parameter expression predicts that the portion of the stellar surface visible to a distant observer shrinks as @xmath72 .
specific numerical examples require a choice of @xmath73 .
the simplest option is to pick @xmath74 where @xmath75 is the plasma frequency , @xmath76 and @xmath77 the electron charge and mass respectively , and @xmath78 the plasma number density .
together with equations ( [ nfreq ] ) and ( [ generalhamiltonian ] ) , equation ( [ plasmafreqdef ] ) gives the dispersion relation for rays in a cold , non - magnetized plasma .
we will make use of this form in the following sections , but we stress that our results to this point are general for @xmath79 .
propagation through this medium requires that the photon has @xmath80 everywhere . in the case
when @xmath81 or the plasma density vanishes @xmath82 , we recover the vacuum case @xmath24 .
the condition for rays to propagate from the surface to infinity is @xmath83 consider the situation illustrated in fig .
[ fig : raytrace ] . in this figure
we show the behaviour of light rays with a range of asymptotic frequencies @xmath39 .
we use the surface value @xmath84 , with equations ( [ omegax ] ) and the propagation condition ( [ propcond ] ) to label the frequency ratio .
tracing rays from the stellar surface numerically requires a specific plasma density distribution , so we choose a physically motivated example based on the goldreich
julian ( gj ) density @xcite .
normal pulsar magnetospheres contain plasma with densities that are sustained by gj currents , transferring charge carriers from the surface of the ns to the magnetosphere .
these currents therefore provide a lower limit for the density of charge carriers in a standard magnetosphere . using the plasma frequency ( equation [ plasmafreqdef ] ) ,
let us consider a radial power - law density @xmath85 where @xmath1 , such that @xmath86 since the gj density depends on the strength of the polar magnetic field we use @xmath5 giving a @xmath87 form .
let us consider a highly relativistic star with @xmath88 , and assume a plasma frequency @xmath89 with @xmath90 as an arbitrary constant .
we plot the rays with positive impact parameter ranging from @xmath91 to @xmath70 as thick dotted lines .
all frequencies satisfy the condition for propagation given in equation ( [ propcond ] ) .
classical turning points exist for the ray paths where @xmath92 .
for all rays with impact parameter @xmath93 these turning points are within the spherical surface of the star and have no effect on the ray paths . for rays with the maximum impact parameter
the turning point is exactly on the surface of the star , indicating a ray that just grazes the stellar surface .
the rays which have larger impact parameters than @xmath70 do not connect the stellar surface to a distant observer .
in this section , we are interested in the gravitational lensing effect on rays passing by the star .
these rays approach from infinity and have impact parameters @xmath94 .
the turning points for these trajectories , @xmath95 , are found when the denominator of equation ( [ delta ] ) vanishes : @xmath96 using a power - law density distribution gives rise to a polynomial with order that depends on the power - law exponent @xmath97 the solutions of this equation signify the boundary between regions in which @xmath98 is real and complex valued .
a trajectory that reaches a turning point has vanishing radial momentum and subsequently reverses direction .
when the effect of refraction is significant , even light rays near the line of sight with small impact parameters can be strongly affected by the interaction with the medium and turned away . in this case
the plasma behaves like a scattering medium , and it is only far from the origin that the vacuum - type lensing behaviour is recovered .
we define the energy and effective potential in analogy with the vacuum case by associating the @xmath32 with the angular momentum of a photon , and the time component @xmath31 with the vacuum energy of a photon .
we call the radial rate of change with respect to the curve parameter @xmath99 , and equations ( [ vel ] ) and ( [ pr ] ) can be used to write @xmath100 the effective potential of the schwarzschild space
time with a plasma distribution is @xmath101 where @xmath102 , with @xmath103 and @xmath104 the vacuum and plasma contributions , respectively . in the absence of plasma @xmath105 ( @xmath106 )
the potential reduces to the vacuum case for null geodesics . for the @xmath3 case
@xmath89 is constant , and photons behave like massive particles with effective mass equal to the plasma frequency , a scenario that was explored by @xcite , @xcite and @xcite .
these photon paths are identical to the well - studied time - like paths in the vacuum schwarzschild case .
however , it is far more interesting to explore the behaviour for other , less trivial plasma density distributions . the @xmath107 case is plotted in fig .
[ raytraceh1 ] for the trajectories of @xmath108 rays past the mass @xmath109 with radius @xmath110 , and shows the corresponding effective potential for two of these rays .
we plot the square of the photon effective energy @xmath111 , the effective potential @xmath112 , and the contributions from the vacuum and plasma parts @xmath103 and @xmath104 .
the effect of the plasma dominates over gravity for rays near the plasma frequency , and this scattering behaviour is more pronounced at lower @xmath2 . for the @xmath5 case , shown in fig .
[ raytrace ] , the plasma density drops off quickly enough that we recover behaviour similar to lensing in vacuum for rays further from the compact object even at low ratios of photon to plasma frequency . in the figures we arbitrarily set @xmath90 for these numerical examples .
the @xmath4 case has the plasma term act like an additional contribution to the angular momentum . to make this connection explicit ,
consider the turning points from equation ( [ turningpts ] ) with the angular momentum @xmath113 : @xmath114 the solutions are known for this specific form of third order polynomial @xmath115\end{aligned}\ ] ] with @xmath116 , analogous to the vacuum case with @xmath117 .
circular orbits with radii @xmath118 occur at extrema of the potential , where the first derivative vanishes . for general @xmath2
this is @xmath119 .
\label{vdiff}\ ] ] stability is determined from the sign of the second derivative @xmath120.\ ] ] we summarize the behaviour of the circular paths in table [ table1 ] for integer @xmath121 .
it is also easy to find the corresponding angular momentum necessary for circular motion in general .
setting equation ( [ vdiff ] ) equal to zero and solving for the angular momentum gives @xmath122.\ ] ] this is the angular momentum that is required to keep a photon in a circular path around a compact object surrounded by a plasma distribution with a power - law density .
the radii of the circular orbits is listed in table [ table1 ] as @xmath118 , along with the corresponding density index @xmath2 .
only the @xmath3 case admits both solutions , which are physical and become stable when they coincide at @xmath123 . for the other cases ,
only one of the roots is positive .
all of the solutions have @xmath124 as @xmath125 .
introducing terms into the effective potential through refractive plasma brings new behaviour to the trajectories of light rays , some of which correspond analogously to known cases .
for example , the @xmath5 case is related to the effective potential found in the klein
gordon equation that describes the evolution of scalar field perturbations on the schwarzschild space
time @xcite . to proceed , we assume the energy density of the scalar field to be small and its influence on the space time geometry negligible . due to spherical symmetry , the radial and angular parts of the klein gordon equation can be separated @xcite , and the field written in terms of vector and spherical harmonics . this radial part is the regge
wheeler equation @xcite for massless , electromagnetic ( @xmath126 ) and gravitational ( @xmath127 ) perturbations on the schwarzschild background @xmath128 where radial coordinates with an asterisk denote the use of a ` tortoise coordinate ' , @xmath129 the potential governing the behaviour of the perturbation is @xmath130,\ ] ] with the orbital quantum number @xmath131 and spin given by @xmath132 , along with the condition @xmath133 @xcite .
writing @xmath134 and @xmath135 , the solution for circular orbits is then @xmath136 now consider the case of a plasma surrounding a compact object with a density profile that drops off as @xmath87 , such that @xmath5 in equation ( [ effpot ] ) .
the resulting effective potential is a direct analogy to the potential that arises in the regge - wheeler equation .
the solutions for circular orbits @xmath137 and @xmath118 for @xmath5 in table [ table1 ] can be transformed into one another given the substitutions @xmath138 and @xmath139 .
the analogy works directly for @xmath140 , and reproduces the vacuum case for @xmath126 ( @xmath106 ) as expected for photons . for @xmath127 we must allow @xmath141 which corresponds to a refractive index that has @xmath142 everywhere , but is an unphysical density distribution for plasma .
however , with this caveat we are able to reproduce the potential and circular orbits of massless scalar fields of arbitrary spin .
while we work in the geometric optics limit , our trajectories point along the propagation directions of the corresponding solutions to the regge - wheeler equation .
thus , the interaction of passing electromagnetic radiation with a gravitational field and a plasma that has a density distribution that drops off as @xmath87 is in some sense analogous to the more general interaction of a massless boson and scalar field perturbation on the schwarzschild background .
as a physically motivated scenario , let us consider an ns embedded within a plasma with bright polar emission regions near the surface that produce observable broad - band pulsations .
this toy model provides an example of the significance refraction can have on the self - lensing of a compact object .
the basic model and viewing geometry used here were thoroughly discussed by @xcite , and the interested reader should consult that work for full details .
let us define the observer s coordinate system with respect to a plane perpendicular to the line of sight ( the ` detector ' ) which collects the rays from a distant object .
the detector surface records the flux received from the object ( the ` image ' ) over the angles @xmath143 and @xmath144 such that an element of solid angle in the observer coordinate system is given by @xmath145 under the appropriate small angle assumption for @xmath143 .
this angular patch is related to the impact parameter in a simple way by @xmath146 which gives @xmath147 where @xmath50 is given by equation ( [ b ] ) and varies from @xmath91 to @xmath70 . since the star is spherically symmetric @xmath148 .
thus we can imagine tracing each patch of image back to some location on the surface of the star given by the angles @xmath98 ( equation [ delta ] ) and @xmath30 . in this sense @xmath98
translates the two dimensional solid angle element of the observer to a corresponding surface area element on the star . to generate pulse profiles
we must find the flux from a bright emission region at the pole of the star and describe the projection of this feature on the observer s sky .
the line of sight passes through the origin of the detector plane and the centre of the star , taken to be the origin of a spherical coordinate system fixed relative to the surface of the star ( the ` object ' coordinates ) .
let a circular emission region ( a ` hotspot ' ) be centred on the polar cap with an opening angle @xmath149 , and the centre of the polar cap be denoted by a unit vector @xmath150 .
let us also define a vector along the line of sight , @xmath151 , such that the angle between @xmath151 and the cap centre @xmath150 is called @xmath152
. in general @xmath152 is a time
dependent quantity since the star rotates . to express this as a function of time
, we define the phase @xmath153 where @xmath154 with @xmath155 the rotation period and @xmath156 an arbitrary phase constant . then the angle between the rotation axis @xmath157 and line of sight is @xmath158 and the angle between the rotation axis and the centre of the hotspot is @xmath159 . the orientation of the cap
can then be expressed with respect to these angles as @xmath160 \label{theta0time}\ ] ] where the time dependence of @xmath152 has been explicitly stated . in the object
coordinates , an arbitrary point on the cap boundary is given by ( @xmath149 , @xmath161 ) , and the corresponding point on the observer s detector is ( @xmath162 , @xmath163 ) . in general the two coordinate systems
are rotated about the @xmath151-@xmath150 plane .
using the transformation between the cartesian components of the two systems , we arrive at an expression for @xmath164 in terms of the observer coordinates , @xmath165 for a given @xmath162 .
thus we are able to determine the boundary of the hotspot in a slice along @xmath162 in the observer coordinates for a given orientation of the star ( i.e. , a given @xmath152 ) and a given cap size @xmath149 .
the hotspot extends over the range @xmath166 to @xmath163 , and we denote this range as @xmath167 , simplifying the observer s solid angle element . determining the observable flux from a hotspot
therefore only requires a one dimensional integral over the impact parameter @xmath50 . in general the definition of @xmath163
is more complicated than described above .
in fact , @xmath163 is not defined for all arbitrary orientations of the cap since equation ( [ phib ] ) requires @xmath168 and @xmath169 for example .
however , the conditions necessary to continuously extend @xmath163 are given in a variety of sources including @xcite and @xcite who include a thorough discussion for cases when multiple imaging is important . to compare our results with the vacuum case
we introduce one final substitution . in equation ( [ delta ] ) we set @xmath63 , which brings us to @xmath170 ; \theta_0 , \theta_{\text{c } } ) x \text{d}x.\ ] ] to find the flux of a surface bright spot we must use the specific intensity . @xcite
have shown that the generalization of the invariant relationship between specific intensity and frequency along a ray in a refractive medium is @xmath171 where the labels obs and em denote the observed and emitted quantities , respectively . taking the observation point to be infinitely far from the star , @xmath172 , @xmath173 , and assuming the emission point on the surface @xmath174 and @xmath175 . by equating the observed and emitted intensity through this relationship
we obtain @xmath176 for simplicity we take @xmath177 where @xmath178 is a constant and @xmath179 a function of emission angle to accommodate anisotropic emission effects @xmath180 other choices for @xmath181 are also possible , such as approximating the specific intensity via the brightness temperature @xcite . the emission angle at the surface is a function of @xmath182 since @xmath183 where @xmath184 .
the flux at a given frequency @xmath18 is @xmath185 ) h(\theta[x ] ; \theta_0 , \theta_{\text{c } } ) x \text{d}x.\end{aligned}\ ] ] the result looks similar to that obtained by @xcite ; however , the key differences come from @xmath98 and the range of @xmath182 , since it depends on the maximum impact parameter which is now dependent on frequency .
the presence of an antipodal hotspot simply requires a second component with @xmath186 . to generate light curves
we recall the time dependency of @xmath152 and plot @xmath187 for a variety of frequency ratios @xmath84 and mass - radius relationships @xmath188 .
the ratios we used start with a case that demonstrates multiple imaging of the stellar surface .
we decrease this ratio in each subsequent plot to show how the pulse profiles change for larger ratios of radius to mass .
we consider a constant size polar cap fixed with aperture @xmath189 in an orthogonal configuration with @xmath190 .
we plot the @xmath5 case in fig .
[ pulsesingle ] , and an analogous example with two antipodal hotspots in fig .
[ pulsedouble ] .
we emphasize the change in morphology of the pulse profiles as a function of frequency by assuming isotropic emission , setting @xmath191 .
the calculations are done in terms of the ratio @xmath84 , so the specific value of @xmath192 will affect only the frequency at which plasma effects become significant , and will leave the morphology of the calculated pulse profiles unchanged . .
the pulses are plotted for an orthogonal configuration with @xmath193 , and frequency ratios @xmath194 ( dashed line ) , @xmath195 ( dashed - dotted ) , @xmath196 ( dotted ) .
the aperture angle is @xmath189 .
the vacuum case is plotted as a thin , solid line .
the upper left panel has @xmath197 , the upper right @xmath198 , lower left @xmath199 and lower right @xmath200 .
the pulse profiles narrow as @xmath201 approaches @xmath202 . ] .
the pulse profiles for two antipodal hotspots are plotted .
parameters are identical to those used in fig .
[ pulsesingle ] . ]
we vary the angles @xmath203 and @xmath204 in fig .
[ pulseclasses ] to demonstrate the frequency dependent pulse morphologies for a variety of viewing angles and system geometries . with the convenient classification scheme of @xcite
we show separate examples of each pulse class for antipodal configurations . in this scheme ,
class i has a primary spot which is always visible , and an antipodal spot which is never seen . in class
ii profiles , the primary spot is always seen and the antipodal spot is only visible for part of the time .
class iii profiles are generated when the primary and antipodal spots are both visible for fractions of the time and class iv has both spots seen at any time , producing a relatively constant profile .
we plot the profiles for the @xmath5 case which shows how the profiles evolve as a function of @xmath18 . .
all other parameters are as in fig .
[ pulsesingle ] .
we plot plasma frequency ratios @xmath194 ( dashed line ) , @xmath195 ( dashed - dotted ) , @xmath196 ( dotted ) , which is similar to the vacuum profile for all classes .
the upper left panel has @xmath205 , @xmath206 , the upper right panel @xmath207 , @xmath208 , the lower left panel has @xmath209 , @xmath210 and the lower
right panel has @xmath205 , @xmath210 . except in class
i , the profiles completely vanish nearest the plasma frequency . ]
another significant effect for pulse profiles is the time delay , which is also affected by the medium .
consider the arrival times of photons leaving the stellar surface and travelling to a distant observer @xmath211 the time delay @xmath212 for a given impact parameter is found by comparing to the fiducial value @xmath59 . for photons emitted from the surface at @xmath213 , we have @xmath214 dr\ ] ] which reduces to the vacuum case when @xmath24 .
the diverse population of nss is defined by a wide range of magnetic fields , varying from the rapidly spinning millisecond pulsars ( msps ) with fields from @xmath215 to @xmath216 g @xcite to the slowly rotating magnetars which have super critical qed scale fields between @xmath217 and @xmath218 g @xcite .
generally , nss possess complex magnetospheres containing plasma that can have a significant effect on pulse profiles @xcite , though the exact mechanism responsible for radio emission in the magnetosphere is not well understood @xcite .
the gj number density provides a lower limit for the charge density in the neighbourhood around an ns , and is given by @xmath219 in units of particles per @xmath220 , with @xmath155 the rotation period in seconds and @xmath221 ( g ) the field component along the rotation axis in the approximation above @xcite . since @xmath222 is a lower limit we use the proportionality @xmath223 to discuss the magnetospheric density @xmath224 @xcite .
this expression gives a physical motivation for the @xmath5 plasma density index discussed throughout this work . the ray tracing discussed in section [ sec : theory ] can be used for rays passing through a pulsar wind , introducing relativistic and plasma effects into models of binary eclipses for a double ns or ns black hole binary that is nearly edge on with respect to the observer . in the case of an aligned rotator
we expect a wind density proportional to @xmath225 , such that @xmath107 @xcite , shown in fig .
[ raytraceh1 ] .
it is plausible that both lensing and plasma may affect the eclipse mechanism for as - yet undiscovered compact binary systems with an edge on orientation @xcite .
normal radio pulsars show a rich variety of frequency dependent behaviours @xcite . a proportionality constant @xmath226 is needed to describe the dense environment surrounding these objects @xcite .
generally the simplification of cold plasma is unrealistic since matter within the light cylinder is strongly magnetized in corotation with the ns and highly relativistic .
particles moving along the field lines have a lorentz factor @xmath227 , giving an effective plasma frequency @xmath228 where the right - hand side is summed and averaged over the positron and electron distribution functions @xcite .
the relativistic case gives a much lower cutoff frequency , well below @xmath229 mhz @xcite .
thus , despite the low frequencies involved the interplay between gr and plasma effects in pulsars may potentially be within observable ranges @xcite .
it is also significant to note the existence of a subset of radio pulsars with pulse widths that broaden with increasing frequency @xcite , also seen in figs .
[ pulsesingle ] and [ pulsedouble ] for the toy light curves generated here . due to their fast rotation rate , the radio emission region
is thought to be closest to the ns surface in msps @xcite .
if a low - frequency component of the radio emission from msps does originate sufficiently near the surface ( ie , @xmath230 ; * ? ? ?
* ; * ? ? ?
* ) , it may be plausibly affected by both relativistic and plasma effects . the significance of rotation for a quickly spinning msp , especially due to induced oblateness of the surface , becomes increasingly important for the pulse profiles and time delays , making the schwarzschild case a first approximation @xcite . at the opposite extreme from msps ,
anomalous x - ray pulsars ( axps ) and soft gamma repeaters ( sgrs ) appear to be nss with magnetic fields above the qed critical field @xmath231
g @xcite , most commonly described by the magnetar model @xcite .
the lack of persistent radio emission above @xmath232 mhz @xcite and the transient , highly - variable radio emission observed from a subset of these objects @xcite suggests that the density of plasma in the magnetospheres of magnetars is lower than in normal radio pulsars @xcite .
while gravitational effects are expected to dominate photon trajectories outside of magnetars @xcite , it is generally believed that plasma effects should also become important deep within magnetar atmospheres @xcite
. a particularly interesting alternative to the magnetar model describes axp and sgr behaviour in terms of a quark star @xcite .
the environments around these objects are expected to be complex , and some models explain magnetar - like behaviour through accretion from fall - back material on to the surface of the object @xcite .
interest in the quark nova scenario has grown considerably due in part to the _ nustar _ observations of the cassiopeia a supernova remnant @xcite which detected a discrepancy in the distribution of titanium-44 and iron-56 abundances that can be explained by the quark nova scenario @xcite .
considerations of more realistic dispersion relations coupled with gr may be relevant in the dynamic environment around an energetic , highly compact quark star .
ray tracing through the schwarzschild space
time including plasma demonstrates light ray paths where refraction can dominate over gravity , scattering incoming rays . figs . [ raytraceh1 ] and [ raytrace ]
show the effects for density power - law index @xmath107 and @xmath5 , respectively .
we have explored the orbits discussed in section [ sec : orbits ] with both the integral formalism ( equation [ delta ] ) and by solving the system of differential equations given by equations [ vel ] and [ accel ] directly using a fourth order runge kutta method .
the integral approach is valuable for analytical results like the pulse profiles given in section [ sec : pulse ] , while the rk4 scheme is suited to handling general cases when spherical symmetry is not assumed ( for example , with the kerr metric ) , and will be useful for more complicated plasma distributions . in section [ circularorbits ] , we found the radii of circular orbits for arbitrary @xmath2 and derived the turning points for trajectories that do not reach an observer at infinity .
depending on the details of the density distribution the plasma frequency can act as an effective photon mass ( @xmath3 ; * ? ? ?
* ) , an additional contribution to the angular momentum ( @xmath4 ) , or reproduce the potential of a massless scalar field perturbation on the schwarzschild background with a judicious choice of scaling constant ( @xmath5 ) .
the interaction between a refractive medium and gravitational lensing has been considered at a variety of scales in the literature @xcite .
the effect of plasma on the formation of gravitational lensed images of background sources will be difficult to detect due to the small angular size of the images .
observing frequency
dependent magnification from transient microlensing events requires multi - waveband monitoring @xcite . in this work we investigated another scenario in which dispersion and general relativistic effects may overlap : the pulse profiles generated by a compact object embedded in a continuous and smoothly varying refractive medium .
we made use of the schwarzschild metric to study the frequency dependent effects , but a similar analysis can be done using the reissner nordstrom and kerr @xcite solutions .
both charge and rotation influence the space time geometry and therefore would produce unique lensing effects @xcite .
this work shows that inclusion of even the simplest plasma effects generates achromatic behaviour in pulse profiles and novel trajectories for light rays near massive compact objects .
i acknowledge and thank samar safi - harb for numerous contributions to this work including discussion , proofreading and support through the natural sciences and engineering research council of canada ( nserc ) canada research chairs program .
i would also like to acknowledge the contributions of t. a. osborn , charlene pawluck and brian corbett for proofreading and helpful suggestions for improving the text .
further thanks to andrew senchuk for many stimulating conversations during the development of this work .
finally , i would like to thank the anonymous referee whose constructive feedback improved the clarity and flow of the text .
alpar m. a. , cheng a. f. , ruderman m. a. , shaham j. , 1982 , nature , 300 , 728 arons j. , barnard j. j. , 1986 , apj , 302 , 120 backer d. c. , kulkarni s. r. , heiles c. , davis m. m. , goss w. m. 1982 , nature , 300 , 615 barnard j. j. , arons j. , 1986 , apj , 1 , 302 beloborodov a. m. , 2002 , apj , 566 , l85 bhattacharya d. , van den heuvel e. p. j. , 1991 , phys . rep . , 203 , 1 bisnovatyi - kogan g. s. , tsupko o. yu . , 2009 , gravit .
cosmol . , 15 , 1 .
20 bisnovatyi - kogan g. s. , tsupko o. yu . , 2010 , mnras , 404 , 1790 bick j. , hadrava p. , 1975
, a&a , 44 , 389 broderick a. , blandford r. , 2003 , mnras , 342 , 1280 cadeau c. , morsink s.m .
, leahy d. , campbell s. s. , 2007 , apj , 654 , 458 camilo f. , ransom s. m. , halpern j. p. , reynolds j. , 2007 , apj , 666 , l93 chakraborty s. , sen a. k. , 2014 , preprint ( arxiv:1406.1295 ) chandrasekhar s. , detweiler s.l . , 1975 , proc .
r. soc . a , 344 , 441 chen j. l. , wang , h. g. , 2014 , apjs , 215 , 11 cheng a. f. , ruderman m. a. , 1979 , apj , 229 , 348 contopoulos i. , kazanas d. , fendt c. , 1999 , apj , 511 , 351 dressel j. , rajeev s. g. , howell j. c. , jordan a. n. , 2009 , phys . rev .
a , 79 , 013834 duncan r. c. , thompson c. , 1992 , apj , 392 , l9 dbrowski m. p. , osarczuk j. , 1995 , ap&ss , 229 , 139 er x. , mao s. , 2014 , mnras , 437 , 2180 gedalin m. , melrose d. b. , 2001 , phys .
e , 64 , 027401 gedalin m. , melrose d. b. , gruman e. , 1998 , phys .
e , 57 , 3399 gil j. a. , kijak j. , 1993 , a&a , 273 , 2 goldreich p. , julian w. h. , 1969 , apj , 157 , 869 grefenstette b. w. et al .
, 2014 , nature , 506 , 339 hankins t. h. , rankin j. m. , eilek j. a. , 2009 , astro2010 : the astronomy and astrophysics decadal survey , science white papers , no . 112 , national academy of sciences , washington , dc . harding a. k. , contopoulos i. , kazanas d. , 2013 , apj , 525 , l125 heyl j. s. , shaviv n. j. , 2002 , phys .
d , 66 , 023002 istomin y. , 2001 , ap&ss , 278 , 77 istomin y. n. , sobyanin d. n. , 2007 , astron .
lett . , 33 , 660 keane e. f. et al . , 2014 , to be published in advancing astrophysics with the square kilometre array , pos(aaska14)040 kijak j. & gil j. a. , 1998 , mnras , 299 , 3 kim j. y. , 2012 , j. cosmol .
10 , 56 kirk j. g. , lyubarsky y. , petri j. , 2009 , in becker w. , ed , astrophysics and space science library , vol . 357 , neutron stars and pulsars .
springer , berlin , p. 697
kulsrud r. , loeb a. , 1992 , phys .
d , 45 , 2 lai d. , ho w. c. g. , 2003 , apj , 588 , 962 lazarus p. , kaspi v. m. , champion d. j. , hessels j. w. t. , dib r. , 2012 , apj , 744 , 97 leahy d. , ouyed r. , 2009 , adv .
, 306821 linet b. , 2009 , preprint ( arxiv:0907.2181 ) lorimer d. r. , kramer m. , 2005 , the handbook of pulsar astronomy .
cambridge university press , new york lun a. w .- c .
, fackerell e. d. , 1974 , lett .
nuovo cimento , 9 , 15 malov i. f. , 2014 , astron
58 , 139 malov o. i. , malofeev v. m. , 2010 , astron . rep .
, 54 , 210 melrose d. b. , 2000 , in kramer m. , wex n. , wielebinski n. eds , asp conf .
202 , pulsar astronomy - 2000 and beyond .
, san francisco , p. 721
melrose d. b. , gedalin m. e. , 1999 , apj , 521 , 351 mereghetti s. , 2013 , braz .
j. phys . , 43 , 356 morozova v. s. , ahmedov b. j. , tursunov a. a. , 2013 , ap&ss , 346 , 513 muhleman d. o. , johnston i. d. , 1966 , phys .
17 , 8 , 455 muhleman d. o. , ekers r.d .
, fomalont e. b. , 1977 , phys .
24 , 1377 natario j. , schiappa r. , 2004 , adv .
phys . , 8 , 1001 noonan t. w. , 1982 , apj , 262 , 344 ouyed r. , elgaroy o. , dahle h. , keranen p. , 2004
, a&a , 420 , 20125 ouyed r. , leahy d. , koning n. , 2015 , res .
astrophys . , 15 , 483 ouyed r. , leahy d. , niebergal b. , 2007 , a&a , 473 , 357 paczynski b. , 1992 , acta astron . , 42 , 145 pechenick k. r. , ftaclas c. , cohen j. m. , 1983 , apj , 274 , 846 perlick v. , 2000 , ray optics , fermat s principle and applications to general relativity .
springer - verlag , heidelberg .
popov m. v. et al . , 2006 ,
, 50 , 562 regge t. , wheeler , j. a. , 1957 , phys .
d. , 108 , 4 shao l. et al .
, 2014 , preprint ( arxiv:1501.00058 ) skakala j. , visser m. , 2010 , j. high energy phys . , 1008 , 61 stappers b. w. et al . , 2011 , a&a , 530 , a80 sturrock p. a. , 1971 , apj , 164 , 529 synge j. l. , 1960 , relativity : the general theory .
north - holland , amsterdam tong h. , xu r. x. , song l. m. , qiao g. j. , 2013 , apj , 768 , 144 tsupko o. y. , bisnovatyi - kogan g. s. , 2013 , phys . rev .
d , 87 , 124009 usov v. v. , 1992 , nature , 357 , 472 wang p. f. , wang c. , han j. l. , 2014 , mnras , 441 , 1943 zerelli f. j. , 1970 , phys . | the interaction between refraction from a distribution of inhomogeneous plasma and gravitational lensing introduces novel effects to the paths of light rays passing by a massive object .
the plasma contributes additional terms to the equations of motion , and the resulting ray trajectories are frequency dependent .
lensing phenomena and circular orbits are investigated for plasma density distributions @xmath0 with @xmath1 in the schwarzschild space time . for rays passing by the mass near the plasma frequency refractive effects can dominate , effectively turning the gravitational lens into a mirror .
we obtain the turning points , circular orbit radii , and angular momentum for general @xmath2 .
previous results have shown that light rays behave like massive particles with an effective mass given by the plasma frequency for a constant density @xmath3 .
we study the behaviour for general @xmath2 and show that when @xmath4 the plasma term acts like an additional contribution to the angular momentum of the passing ray . when @xmath5 the potential and radii of circular orbits are analogous to those found in studies of massless scalar fields on the schwarzschild background .
as a physically motivated example we study the pulse profiles of a compact object with antipodal hotspots sheathed in a dense plasma , which shows dramatic frequency
dependent shifts from the behaviour in vacuum . finally , we consider the potential observability and applications of such frequency dependent plasma effects in general relativity for several types of neutron star .
[ firstpage ] gravitation
plasmas stars : neutron pulsars : general |
one of the most interesting new developments in hadron physics is the application of the ads / cft correspondence @xcite to nonperturbative qcd problems @xcite .
it is well know that ads / cft gives an important insight into the viscosity and other global properties of the hadronic system formed in heavy ion collisions @xcite .
the essential ansatz for the application of ads / cft to hadron physics is the indication that the qcd coupling @xmath1 becomes large .
therefore conformal symmetry can be applied to , for example solutions of the qcd dyson schwinger equations and phenomenological studies of qcd couplings based on physical observables such as @xmath2 decay and the bjorken sum rule show that the qcd @xmath3 function vanishes and @xmath1 become constant at small virtuality ; i.e. , effective charges develop an infrared fixed point . fully exploiting the gauge / gravity correspondence to produce a model for real strong interaction physics- a method called `` holographic qcd '' or `` ads / qcd''-may
be attempted either through a top - dawn approach starting with a particular string theory and choosing a background that naturally produces qcd - like properties , or a bottom - up approach starting with real qcd properties and using them to obtain constraints on viable dual gravity theories .
the first attempts have been made for constructing phenomenological holographic models of qcd @xcite .
surprisingly simple models consisting of gauge theory in an anti - de - sitter space interval have turned out to provide a remarkably good description of the meson sector of qcd .
therefore , it will be interesting that the calculation of the higher twist effects within holographic qcd in proton - proton collisions in the running coupling approach .
the large - order behavior of a perturbative expansion in gauge theories is inevitably dominated by the factorial growth of renormalon diagrams @xcite . in the case of quantum chromodynamics ( qcd ) , the coefficients of perturbative expansions in the qcd coupling @xmath4 can increase dramatically even at low orders .
this fact , together with the apparent freedom in the choice of renormalization scheme and renormalization scales , limits the predictive power of perturbative calculations , even in applications involving large momentum transfer , where @xmath4 is effectively small . in this work
we apply the running coupling approach @xcite in order to compute the effects of the infrared renormalons on the pion production in proton - proton collisions within holographic qcd .
this approach was also employed previously @xcite to calculate the inclusive meson production in proton - proton and photon - photon collisions . for the calculations of the higher - twist cross sections on the dependence of wave functions of pion
, we used the holographic qcd prediction @xmath5 @xcite and pion asymptotic wave functions @xmath6
@xcite from the perturbative qcd evolution .
theoretically obtained predicted results within holographic qcd were compared with results of the perturbative qcd which obtained by the running coupling and frozen coupling constants approaches . the frozen coupling constant approach in refs .
@xcite was used for calculation of integrals , such as @xmath7 it should be noted that , in pqcd calculations , the argument of the qcd coupling constant ( or the renormalization and factorization scale ) @xmath8 should be taken equal to the square of the momentum transfer of a hard gluon in a corresponding feynman diagram . but definition of @xmath9 suffers from infrared singularities .
therefore , in the soft regions as @xmath10 , and @xmath11 , integrals ( 1.1 ) diverge and for their calculation some regularization methods are needed for @xmath12 in these regions .
investigation of the infrared renormalon effects in various inclusive and exclusive processes is one of the most important and interesting problem in the perturbative qcd .
it is known that infrared renormalons are responsible for factorial growth of coefficients in perturbative series for the physical quantities .
but , these divergent series can be resummed by means of the borel transformation @xcite and the principal value prescription @xcite .
studies of higher - twist and renormalon effects also opened new prospects for evaluation of power suppressed corrections to processes characteristics .
we organize the paper as the follows . in section [ ht
] , we provide some formulas for the calculation of the contributions of the higher twist and leading twist diagrams . in section [ ir ] ,
we present the formulas and analysis of the higher - twist effects on the dependence of the pion wave function by the running coupling constant approach , and in section [ results ] , the numerical results for the cross section and discuss the dependence of the cross section on the pion wave functions are presented . finally , some concluding remark are stated in section [ conc ] .
the higher - twist feynman diagrams , which describe the subprocess @xmath13 for the pion production in the proton - proton collision are shown in fig.1 .
the amplitude for this subprocess can be found by means of the brodsky - lepage formula @xcite @xmath14 in eq.(2.1 ) , @xmath15 is the sum of the graphs contributing to the hard - scattering part of the subprocess . the mandelstam invariant variables for subprocesses @xmath16 are defined as @xmath17 the pion wave functions predicted by ads / qcd @xcite and the pqcd evolution @xcite has the form : @xmath18 @xmath19 where @xmath20 is the pion decay constant .
the cross section for the higher - twist subprocess @xmath21 is given by the expression @xmath22
^ 2}{{\hat s}^3}\left[\frac{1}{{\hat u}^2}+\frac{1}{{\hat t}^2}\right]\ ] ] where @xmath23+e_2\hat u\int_{0}^{1}dx\left[\frac{\alpha_{s}(q_2 ^ 2)\phi_{\pi}(x , q_2 ^ 2)}{1-x}\right].\ ] ] in the eq.(2.5 ) @xmath24and @xmath25 represent the momentum squared carried by the hard gluon in fig.1 , @xmath26 is the charge of @xmath27 and @xmath28 .
the higher - twist contribution to the large-@xmath29 pion production cross section in the process @xmath30 is @xcite @xmath31 we denote the higher - twist cross section obtained using the frozen coupling constant approach by @xmath32 . regarding the higher - twist corrections to the pion production cross section , a comparison of our results with leading - twist contributions is crucial .
we take two leading - twist subprocesses for the pion production:(1 ) quark - antiquark annihilation @xmath33 , in which the @xmath34 and ( 2 ) quark - gluon fusion , @xmath35 , with subsequent fragmentation of the final quark into a meson , @xmath36 @xcite .
the main problem in our investigation is the calculation of integral in ( 2.5 ) by the running coupling constant approach within holographic qcd and also discussion of the problem of normalization of the higher twist process cross section in the context of the same approach .
therefore , it is worth noting that , the renormalization scale ( argument of @xmath37 ) according to fig.1 should be chosen equal to @xmath38 , @xmath39 .
the integral in eq.(2.5 ) in the framework of the running coupling approach takes the form @xmath40 the @xmath41 has the infrared singularity at @xmath42 , for @xmath43 or @xmath44 , for @xmath45 and so the integral @xmath46 diverges . for the regularization of the integral , we express the running coupling at scaling variable @xmath41 with the aid of the renormalization group equation in terms of the fixed one @xmath12 .
the solution of renormalization group equation for the running coupling @xmath47 has the form @xcite @xmath48^{-1}.\ ] ] then , for @xmath49 , we get @xmath50 where @xmath51 . having inserted eq.(3.3 ) into eq.(2.5 ) we obtain @xmath52 @xmath53 where @xmath54 and @xmath55 .
although the integral ( 3.4 ) is still divergent , it is recast into a suitable form for calculation . making the change of variable as @xmath56
, we obtain @xmath57 in order to calculate ( 3.5 ) we will apply the integral representation of @xmath58 @xcite .
@xmath59 gives @xmath60 in the case @xmath61 for the @xmath62 it is written as @xmath63 and for @xmath64 wave function @xmath65 + \frac{4\sqrt{3}\pi f_{\pi}e_{2}\hat u}{\beta_{0 } } \int_{0}^{\infty}du e^{-t_{2}u } \left[\frac{1}{1-u}-\frac{1}{2-u}\right].\ ] ] where @xmath66 is beta function .
the structure of the infrared renormalon poles in eq.(3.8 ) and eq.(3.9 ) strongly depend on the wave functions of the pion . to remove them from eq.(3.8 ) and eq.(3.9 )
we adopt the principal value prescription .
we denote the higher - twist cross section obtained using the running coupling constant approach by @xmath67 .
in this section , we discuss the higher - twist contributions calculated in the context of the running and frozen coupling constant approaches on the dependence of the chosen pion wave functions in the process @xmath68 . in numerical calculations for the quark distribution function inside the proton , the mstw distribution function @xcite , and the gluon and quark fragmentation @xcite functions into a pion have been used .
the results of our numerical calculations are displayed in figs.2 - 14 .
firstly , it is very interesting comparing the higher - twist cross sections obtained within holographic qcd with the ones obtained within perturbative qcd . in fig.2 and fig.3
we show the dependence of higher - twist cross sections @xmath69 , @xmath70 calculated in the context of the frozen and running coupling constant approaches as a function of the pion transverse momentum @xmath29 for different pion wave functions at @xmath71 .
it is seen from fig.2 and fig.3 that the higher - twist cross section is monotonically decreasing with an increase in the transverse momentum of the pion . in fig.4-fig.7
, we show the dependence of the ratios @xmath72/@xmath73 , @xmath67/@xmath69 , @xmath69/@xmath74 and @xmath70/@xmath74 as a function of the pion transverse momentum @xmath29 for @xmath75 , @xmath76 and @xmath77 pion wave functions . here
@xmath78 is the leading - twist cross section , respectively .
as shown in fig.4 , in the region @xmath79 resummed higher - twist cross section for @xmath75 is suppress by about half orders of magnitude relative to the resummed higher - twist cross section for @xmath76 also higher - twist cross section for @xmath76 is suppress by about half orders of magnitude relative to the higher - twist cross section for @xmath77 . in fig.5 and fig.6 , we show the dependence of the ratios @xmath67/@xmath69 , and @xmath69/@xmath74 as a function of the meson transverse momentum @xmath29 for the @xmath75 , @xmath76 and @xmath77 pion wave functions .
it is observed from fig.5 that , the ratios for all wave functions decrase with an increase in the transverse momentum of pion , but in the region @xmath80 resummed higher - twist cross section for @xmath75 is suppress by about 2 - 3 and 2 - 4 orders of magnitude relative to the higher - twist cross sections calculated in the context of the frozen coupling method for @xmath76 and @xmath77 , respectively . also , resummed higher - twist cross section for @xmath76 is suppress by about 1 - 3 orders of magnitude relative to the higher - twist cross sections calculated in the context of the frozen coupling method for @xmath77 .
but , as shown in fig.6 , the ratios decrease with increasing in the @xmath29 transverse momentum of the pion and has a minimum approximately at the point @xmath81 , then the ratios increase with increasing in the @xmath29 transverse momentum of the pion . in fig.7
is similar to fig.6 with an exception that there in a minimum approximately at the point @xmath82 . in fig.8 - fig.10
, we have depicted higher - twist cross sections @xmath83 , and ratio @xmath84 , as a function of the rapidity @xmath85 of the pion at @xmath86 and @xmath87 .
figures show that higher - twist cross section and ratios have a different distinctive .
the resummed higher - twist cross section for @xmath75 has a maximum approximately at the point @xmath88 .
however in this point higher - twist cross section for @xmath77 has a minimum . as shown in fig.9 in the region ( @xmath89 ) the ratios
@xmath90 and @xmath91 increase with an increase of the @xmath85 rapidity of the pion and has a maximum approximately at the point @xmath92 , but ratio @xmath93 has a minimum in this point . as is seen from fig.10 ratios @xmath94 , for all wave functions increase with an increase of the @xmath85 rapidity of the pion and has a maximum approximately at the point @xmath88 . in the region
@xmath95 resummed higher - twist cross section is suppress by about half order of magnitude relative to the higher - twist cross section calculated in the framework of the frozen coupling approach .
besides that , the ratio decreases with an increase in the @xmath85 rapidity of the pion .
as is seen from fig.10 , the ratio @xmath96 depends on the choice of the pion wave function .
analysis of our calculations shows that @xmath69 , @xmath70 higher - twist cross sections and ratio sensitive to pion wave functions predicted holographic and perturbative qcd .
we have also carried out comparative calculations in the center - of - mass energy @xmath97 and obtained results are displayed in fig.11-fig.14 .
analysis of our calculations at the center - of - mass energies @xmath86 and @xmath98 , show that with increasing in the beam energy contributions of higher - twist effects to the cross section decrease by about one - two order . as is seen from fig.5 , fig.7 , fig.10 and fig.13 that infrared renormalon effects enhance the perturbative predictions for the pion production cross section in the proton - proton collisions about 2 - 3 order .
this feature of infrared renormalons may help the explain theoretical interpretations with future experimental data for the pion production cross section in the proton - proton collisions
. in our calculations , the higher - twist cross section of the process the dependence of the transverse momentum of pion appears in the range of @xmath99 .
therefore , higher - twist cross section obtained in our work should be observable at rhic .
proton - proton collisions are known are known to be the most elementary interactions and form the very basis of our knowledge about the nature of high energy collisions in general .
physicists , by and large , hold the view quite firmly that the perturbative quantum - chromodynamics provides a general framework for the studies on high energy particle - particle collisions .
obviously , the unprecedented high energies attained at large hadron collider offer new window and opportunities to test the proposed qcd dynamics with its pros and cons .
however , we should remember that lhc opens a new kinematical regime ah high energy , where several questions related to the description of the high - energy regime of the qcd .
consequently , studies of proton - proton interactions at the rhic and lhc could provide valuable information on the qcd dynamics at high energies . in this work the single meson inclusive production via higher twist mechanism within holographic qcd
are calculated . for calculation of the cross section ,
the running coupling constant approach is applied and infrared renormalon poles in the cross section expression are revealed .
infrared renormalon induced divergences is regularized by means of the principal value prescripton and the resummed expression ( the borel sum ) for the higher - twist cross section is found .
it is observed that , the resummed higher - twist cross section differs considerably from that found using the frozen coupling approximation , in some region .
the following results can be concluded from the experiments ; the higher - twist contributions to single meson production cross section in the proton - proton collisions have important phenomenological consequences , the higher - twist pion production cross section in the proton - proton collisions depends on the form of the pion model wave functions and may be used for their study .
also the contributions of renormalons effects within holograpich qcd in this process are essential and may help to analyse experimental results .
further investigations are needed in order to clarify the role of higher - twist effects in qcd .
especially , the forthcoming rhic and lhc measurements will provide further tests of the dynamics of large-@xmath0 hadron production beyond the leading twist .
one of author a.i.ahmadov is grateful to all members of the department of physics of karadeniz technical university for appreciates hospitality extended to him in trabzon . financial support by tubitak under grant number 2221(turkey ) is also gratefully acknowledged .
99 j. m. maldacena , adv .
phys . 2 , 231 ( 1998 ) ; int .
38 , 1113 ( 1999 ) ; s. s. gubser , i. r. klebanov and a. m. polyakov , phys .
b 428 , 105 ( 1998 ) ; e. witten , adv .
phys . 2 , 253 ( 1998 ) .
j. polchinski and m. j. strassler , phys .
88 , 031601 ( 2002 ) .
r. a. janik and r. peschanski , nucl .
b 565 , 193 ( 2000 ) .
j. erlich , e. katz , d. t. son and m. a. stephanov , phys .
95 , 261602 ( 2005 ) [ arxiv : hep - ph/0501128 ] .
a. karch , e. katz , d. t. son and m. a. stephanov , phys . rev .
d 74 , 015005 ( 2006 ) .
s. j. brodsky and g. f. de teramond , phys.lett * b582 * , 211 ( 2004 ) . g. f. de teramond and s. j. brodsky , phys .
94 , 201601 ( 2005 ) .
g. f. de teramond and s. j. brodsky , phys .
96 , 201601 ( 2006 ) .
p. kovtun , d. t. son and a. o. starinets , phys . rev .
94 , 111601 ( 2005 ) .
hooft , in the why s of subnuclear physics , erice , 1977 , edited by a. zichichi ( plenum , new york , 1979 ) , p.94 a. h. mueller , nucl . phys . *
b250 * , 327 ( 1985 ) ; phys . lett . *
b308 * , 355 ( 1993 ) .
v. i. zakharov , nucl .
phys . * b385*,452 ( 1992 ) .
m. beneke , phys . rep .
* 317 * , 1 ( 1999 ) .
agaev , phys . lett .
* b360 * , 117 ( 1995 ) ; * b369 * , 379(e ) ( 1996 ) .
a. i. ahmadov , coskun aydin , sh . m. nagiyev , yilmaz a. hakan , and e. a. dadashov , phys . rev . *
d80 * , 016003 ( 2009 ) .
a. i. ahmadov , coskun aydin , e. a. dadashov and sh .
m. nagiyev , phys . rev . *
d81 * , 054016 ( 2010 ) .
a. i. ahmadov , r. m. burjaliyev , int . j. mod
e20 * , 1243 ( 2011 ) .
a. i. ahmadov , sh .
m. nagiyev , and e. a. dadashov , int . j. mod
e21 * , 1250014 ( 2012 ) . s. j. brodsky and g. f. de teramond , phys
. rev . * d77 * , 056007 ( 2008 ) .
s. j. brodsky , arxiv : hep - ph/0707.2643 .
a. vega , i. schmidt , t. branz , t.gutsche,v .
lyubovitskij , phys . rev . *
d80 * , 055014 ( 2009 ) .
g. p. lepage and s. j. brodsky , phys.lett . * b87 * , 359 ( 1979 ) .
a. i. ahmadov , i. boztosun , r. kh .
muradov , a. soylu , and e. a. dadashov , int . j. mod.phys .
* e15 * , 1209 ( 2006 ) .
a. i. ahmadov , i. boztosun , a. soylu , and e. a. dadashov , int . j. mod.phys .
* e17 * , 1041 ( 2008 ) .
j. a. bagger and j. f. gunion , phys
* d25 * , 2287 ( 1982 ) .
v. n. baier and a. grozin , phys . lett .
* b96 * , 181 ( 1980 ) ; s. gupta , phys . rev . *
d24 * , 1169 ( 1981 ) .
h. contopanagos and g. sterman , nucl . phys . * b419 * , 77 ( 1994 ) . g. l. lepage and s. j. brodsky , phys . rev . *
d22 * , 2157 ( 1980 ) .
j. f. owens , rev .
* 59 * , 465 ( 1987 ) .
w. greiner , s. schramm and e. stein , quantum chromodynamics , 2nd edn.(berlin , springer , 2002 ) , pp.551 . j. zinn - justin , phys .
* 70 * , 109 ( 1981 ) .
a. erdelyi , higher transcendental functions ( mcgrow - hill book company , new york , 1953 ) , vol.2 .
a. d. martin , w. j. stirling , r. s. thorne , and g. watt , hep - ph/0901.0002 .
s. albino , b. a. kniehl , g. kramer , nucl.phys . *
b725 * , 181 ( 2005 ) . | we calculate the contribution of the higher twist feynman diagrams to the large-@xmath0 inclusive pion production cross section in proton - proton collisions in the case of the running and frozen coupling approaches within holographic qcd . the structure of infrared renormalon singularities of the higher twist subprocess cross section and it s resummed expression are found .
we compare the resummed higher twist cross sections with the ones obtained in the framework of the frozen coupling approximation and leading twist cross section .
we discuss the phenomenological consequences of possible higher - twist contributions to the pion production in proton - proton collisions within holographic qcd . |
dutch physicist hendrick brugt gerhard casimir published his famous paper @xcite in 1948 , where he found a simple yet profound explanation for the retarded van der waals interaction between polarizable molecules .
he showed that there exists a net force between grounded infinite metallic plates , placed a few micrometers apart , in a vacuum without any external electromagnetic field .
the casimir effect is essentially a direct consequence of the change in the vacuum energy of a quantum field when some boundary condition or background field is imposed . since the casimir s work
, many papers have been written on the casimir energy or the resulting force due to the presence of boundary conditions .
they have calculated this effect for parallel plates @xcite , cylinders @xcite , and spheres @xcite ( first considered by casimir @xcite , deraad and milton @xcite , and boyer @xcite , respectively ) , and many other geometries @xcite . in addition , many authors have considered this effect for various fields , different boundary conditions , dimensions and topologies @xcite .
there are various methods to calculate the casimir energy , e.g. the green function formalism @xcite , the heat - kernel method @xcite and the multiple - scattering expansions @xcite .
also , as mentioned before , the presence of the nontrivial background fields such as solitons affect the zero - point energy .
for example , to compute the quantum corrections to the mass of the soliton , the change in the quantum fluctuations of the vacuum energy caused by the presence of the soliton , should be taken into account . in 1974 ,
dashen _ et al _ @xcite computed the one - loop correction to the mass of the kink of the @xmath5 theory for the first time .
after this work , several authors have used various methods to compute the corrections to the soliton mass , such as the scattering phase shift method , zeta function analytic continuation technique , dimensional regularization technique , etc .
@xcite . moreover , the casimir effect appears in supersymmetric models . to investigate the validity of the bps saturation by supersymmetric solitons
, one should compute the change in the fluctuations of the vacuum energy to obtain the quantum corrections to the soliton mass ( see for example @xcite ) .
one of the first experimental attempts to verify the existence of the casimir effect was conducted by marcus sparnaay for two parallel metallic plates in 1958 @xcite .
the results was not in contradiction with the casimir theory , but had a very poor accuracy . in 1997 , using a plate and a spherical lens covered by a metallic sheet , steve k. lamoreaux measured the casimir energy with high accuracy @xcite .
it was the first `` successful '' experiment to verify the casimir effect .
since then , many different experiments have been performed to measure the casimir effects for various geometries , such as two parallel plates and a sphere in front of a plane , with ever increasing accuracy ( see for example @xcite ) . in this paper
we consider a pseudoscalar field as a prescribed , spatially varying background field which is coupled chirally to a fermi field .
the background field in our model has two adjustable parameters , one related to its scale of variation and the second to its asymptotic value at spatial infinity .
by varying the latter from zero to @xmath6 , we can build a field configuration with topological charge @xmath7 .
as it is well - known the nomenclature `` soliton '' refers to a static field configuration which is a solution to the classical field equations with finite energy and localized energy density .
topological solitons by definition carry nonzero topological charge . in a field theory
in which this background field interacts with a fermion , such as the one considered in this paper , the exact shape of the soliton depends on the back - reaction of the fermi field , which in turn depends on the particular fermionic state to which the soliton is coupled .
therefore , the soliton can in principle acquire infinitely many different shapes
. however , since the back - reaction is usually small , all these different shapes are close to each other .
however , we have obtained the exact ( to within numerical approximation ) non - perturbative solutions for the coupled system treating both fields as dynamical , verifying the perturbative result that the back - reaction is small @xcite .
the problem of obtaining the casimir energy for any of these particular field configurations is also not exactly solvable .
therefore , in this paper we choose a simple field configuration which has proper topological charge and renders the problem exactly solvable .
although the field configuration that we choose is not a solution to the field equations , it possesses all other properties of a proper soliton .
we shall henceforth refer to this configuration as a soliton for brevity of notation .
we expect the exact casimir energy for this topologically non - trivial configuration to illuminate many of the qualitative features of the problem with an exact soliton . since this model is exactly solvable and we have the complete spectrum of the fermi field @xcite , we are able to calculate the vacuum expectation value of the hamiltonian directly and exactly for all values of the aforementioned parameters .
we obtain the casimir energy for this system by subtracting the zero - point energy in the presence and absence of the background field . in section
ii we introduce the model and present a derivation for the expression of the casimir energy . in section
iii we obtain an exact expression for the casimir energy and its density for our model .
then , we thoroughly explore the properties and behavior of the casimir energy as a function of the parameters of our model .
first , we depict the casimir energy as a function of the scale of variation of the soliton with winding number one . in general
there is a sharp maximum whenever the fermi field has a zero - mode bound state .
we obtain the expected results that the casimir energy tends to zero in the limits of extremely wide and sharp solitons .
second , we depict the casimir energy as a function of @xmath2 . in general @xmath8
is an increasing function of @xmath2 with very mild maxima in the form of cusps whenever there is an energy level crossing zero .
then , we exhibit the energy densities of the bound and continuum states and their sum for states with @xmath4 and @xmath3 , separately .
these two energy densities are exact mirror images of each other , and from this we compute the casimir energy density . in section
iv we summarize and discuss the results .
consider the following lagrangian describing a fermion interacting non - linearly with a pseudoscalar field @xmath9 , in ( @xmath10 ) dimensions , @xmath11 in order to have an exactly solvable model , gousheh and mobilia @xcite chose @xmath9 to be piecewise linear as follows @xmath12 where @xmath13 is the slope in the middle region .
one could interpolate between the extreme adiabatic and non - adiabatic regimes by changing @xmath1 from zero to infinity , respectively . when @xmath14 , the soliton has winding number @xmath7 .
figure ( [ fig.1 ] ) shows the form of @xmath9 . as mentioned earlier this simple topological configuration is used as an approximation to the kink of the @xmath15 theory .
although our simple static configuration is not a solution to the euler - lagrange equation , its total energy can be easily obtained from the corresponding hamiltonian as follows @xmath16dx=\int_{-\infty}^{\infty } \left[\frac{1}{2}(\partial_x \phi ) ^2 + \frac{\lambda}{4}\left(\phi^2-\frac{m'^2}{\lambda}\right)^2\right]dx=\frac{23}{15}\mu\theta_0,\ ] ] where @xmath17 , the value of @xmath18 when @xmath19 , and @xmath20 , the slope at @xmath21 . the form of the background pseudoscalar field @xmath9.,width=238 ] the energies of the bound states as a function of @xmath2 for @xmath22 obtained from @xmath23 derived in @xcite .
solid ( dashed ) lines show the bound states with positive ( negative ) parity .
, width=306 ] the procedure is to find the spectrum of the fermi field in the presence of this background , for arbitrary values of @xmath1 and @xmath2 .
then , for fixed @xmath1 , one could observe the changes in the spectrum of the fermion as a soliton with the topological charge @xmath7 is built by changing @xmath2 from zero to @xmath6 .
this method was first introduced by mackenzie and wilczek @xcite and is based upon the adiabatic method first introduced by goldstone and wilczek @xcite . since this model is exactly solvable , the complete spectrum of the fermi field , including the eigenfunctions and their corresponding eigenenergies , has been obtained exactly for arbitrary values of @xmath1 and @xmath2 , and are given in @xcite . here
we do not present the solutions in detail and the interested reader is referred to @xcite .
if we denote the fermi field in ( @xmath10 ) dimensions by @xmath24 , then we can define @xmath25 and solve the equations of motion in terms of @xmath26 .
since the hamiltonian commutes with the parity operator , the eigenfunctions of the hamiltonian can be chosen to be eigenfunctions of the parity operator as well . in the first representation ( @xmath27 , @xmath28 , @xmath29 ) , the parity operation is given by @xmath30 . in the new representation i.e. eq.([e3 ] ) ( @xmath31 , @xmath32 , @xmath33 ) , it becomes @xmath34 .
figure ( [ fig.2 ] ) shows the energies of the bound states as a function of @xmath2 for @xmath35 as an example .
as shown in the figure , the bound states have alternating parity signs , starting with the positive sign .
this graph will be important for the upcoming discussions . in order to obtain the casimir energy
, we should subtract the zero point energy in the absence from the presence of the background field .
therefore , the vacuum energies for both cases should be calculated .
mackenzie and wilczek @xcite assumed that the solutions in the presence of the soliton are complete .
later on gousheh and mobilia proved this assumption .
therefore , the fermi field operator can be expanded in terms of either the modes in the absence or in the presence of the background field , as follows @xmath36\nonumber\\ & = & \int_{0}^{+\infty}\frac{dp}{2\pi}\sum\limits_{j=\pm}\left[a_p^j \mu_p^j(x , t)+c_p^{j\dag } \nu_p^j(x , t)\right ] + \sum\limits_{i}\left[e_i \chi_{1b_i}(x , t)+f_i^\dag\chi_{2b_i}(x , t ) \right]\end{aligned}\ ] ] where @xmath37 ( @xmath38 ) are the positive ( negative ) energy continuum eigenstates in the absence of the background field . also , @xmath39 ( @xmath40 ) are the positive ( negative ) energy continuum eigenstates and @xmath41 ( @xmath42 ) are the positive ( negative ) energy bound states in the presence of the soliton .
here , @xmath43 refers to the possibility of multiple bound states and the superscript @xmath44 refers to the parity of the continuum states .
mackenzie and wilczek chose the coefficients of both positive and negative energy bound states to be annihilation operators .
however , we choose annihilation ( creation ) operator for the positive ( negative ) energy bound states , as we do for the continuum states .
we shall see one of the advantages of this choice when we compute the casimir energy .
in general the hamiltonian can be expressed as follows @xmath45.\ ] ] substituting the expression for the field operator @xmath46 in eq.([e6 ] ) in the presence of the background field into eq.([e9 ] ) , we obtain @xmath47+\sum\limits_{i}\left[e_i^\dag \chi_{1b_i}^\dag(x , t)+f_i\chi_{2b_i}^\dag(x , t ) \right]\bigg\}\nonumber\\ & \times&\bigg\{\int_{0}^{+\infty}\frac{dq}{2\pi}\sum\limits_{n=\pm}\left[\sqrt{q^2+m^2}a_q^n \mu_q^n(x , t)-\sqrt{q^2+m^2}c_q^{n\dag } \nu_q^n(x , t)\right]\nonumber\\ & + & \sum\limits_{l } \left[e^{l+}_{\text{bound}}e_l \chi_{1b_l}(x , t)+e^{l-}_{\text{bound}}f_l^\dag\chi_{2b_l}(x , t ) \right ] \bigg\},\end{aligned}\ ] ] where , the @xmath48 superscripts on @xmath49 denote the sign of the bound state . upon calculating the expectation value of the hamiltonian in the vacuum state ,
all the terms vanish except the terms with an annihilation operator on the left side and the corresponding creation operator on the right side .
there are only two terms in this form which can be written in the following form by the use of the anticommutation relations @xmath50 therefore , we obtain the following expression for the vacuum expectation value of the hamiltonian @xmath51 where @xmath52 is the vacuum state in the presence of the background field . from this expression
we conclude that the entire contribution to the vacuum energy for fermionic systems originates from the negative energy eigenstates .
note that our separation of the bound states into the positive and negative energy ones has the advantage of the automatic inclusion of the negative energy bound states in the expression for the casimir energy . to obtain the casimir energy ,
one should subtract the zero point energy in the absence from the presence of the background field .
thus , we obtain @xmath53 note that our problem possesses charge conjugation symmetry and therefore our expression for the casimir energy given in eq.([e12 ] ) is equivalent to the conventional one , where one would sum over all modes symmetrically with a factor of @xmath54 , while preserving the sign .
by substituting the expressions for the eigenstates into eq.([e12 ] ) , the casimir energy can be calculated .
our prescription for subtracting two quadratically divergent integrals appearing in this equation , is to subtract the integrands with corresponding values of @xmath55 and @xmath56 , and then performing the integral .
the change in the energy densities of the dirac sea ( @xmath57 ) and sky ( @xmath58 ) ( i.e. the difference between the energy densities in the presence and absence of the background field ) are @xmath59 where @xmath60 , @xmath61 ( @xmath48 are for the dirac sky and sea , respectively ) , @xmath62 for interacting dirac sky ( sea ) and @xmath63 for free dirac sky ( sea ) .
the bound state energy densities are @xmath64 where the @xmath48 refers to the sign of the bound state energies , and @xmath65 for the positive ( negative ) bound states . substituting the explicit expressions for the states given in @xcite into eqs.([r4 ] ) and ( [ r5 ] )
, we obtain the energy densities of the two continua and the bound states , shown bellow . for @xmath66 :
@xmath67\left(|k_1|^2-p^2\right)\nonumber\\ & \times&\cosh\left[\text{im}\left(\zeta\right)l\right ] + \sin\left[2p ( |x|-l)\right]2\,p\ , \text{im}(k_1)\sinh\left[\text{im}\left(\zeta\right)l\right]\big ] \nonumber\\ & + & \left(|n_{c+}|^2-|n_{c-}|^2\right)\big[\cos\left[2p ( |x|-l)\right][e \left(|k_1|^2+p^2\right)+2\,p^2\,\text{re}(k_1)]\cos\left[\text{re}\left(\zeta\right)l\right]\nonumber\\ & -&\sin\left[2p ( |x|-l)\right]p\ , [ |k_1|^2+p^2 + 2e\text{re}\left(k_1\right)]\sin\left[\text{re}\left(\zeta\right)l\right]\big]\bigg\},\\\label{r62 } \varepsilon_{\text{bound}\pm}(x)&= & 2n^2e_{\text{bound}}^{\pm } \text{e}^{-2\lambda\left(|x|-l\right)},\end{aligned}\ ] ] and for @xmath68 : @xmath69\nonumber\\ & \times \cosh\left[\text{im}\left(\zeta\right)x\right ] + 2m\big[e+\text{re}(k_1)\big]\cos\left[\text{re}\left(\zeta\right)x\right ] \left(|n_{c+}|^2-|n_{c-}|^2\right)\bigg\},\\\label{r72 } \varepsilon_{\text{bound}\pm}(x)=&\frac { 4n^2e_{\text{bound}}^{\pm}}{\left[\text{re}\left(\zeta\right)\right]^2 }
\bigg\{\left(\mu e_{\text{bound}}^{\pm}-2\lambda^2\right)\cos\left[\text{re}\left(\zeta\right)(x - l)\right]\nonumber\\ & -\text{re}\left(\zeta\right)\lambda\sin\left[\text{re}\left(\zeta\right)(x - l)\right]+\frac{1}{2}\mu^2+\mu e_{\text{bound}}^{\pm}\bigg\},\end{aligned}\ ] ] where @xmath70 , @xmath71 ( where @xmath72 depending on the case ) , @xmath73 and the normalization constants are @xmath74 ^ 2}\bigg[\frac{\mu e_{\text{bound}}^{\pm } -2 \lambda^2}{\text{re}\left(\zeta \right)}\sin\left[2
l \text{re}\left(\zeta \right)\right]-\lambda\cos\left[2 l \text{re}\left(\zeta \right)\right]+\lambda \nonumber\\&+
l\mu \left(\mu + 2 e_{\text{bound}}^{\pm}\right)\bigg]\bigg\}^{-1/2},\nonumber\\n_{c\pm}&\equiv \bigg\{\cosh\left[\text{im}\left(\zeta\right)l\right]\left[\left(\frac{|k_1|^2}{p^2}+1\right)2 e^2 + 4 e \text{re}\left(k_1\right)\right]\mp\cos \left[\text{re}\left(\zeta\right)l\right]2 m e\nonumber\\&\times\left(\frac{|k_1|^2}{p^2}-1\right)\bigg\}^{-1/2}.\end{aligned}\ ] ] all of the spatial integrals can be performed analytically .
however , the remaining integrals over @xmath55 in eqs.([r6],[r7 ] ) can not be done analytically and we integrate them numerically . figure ( [ fig.3 ] ) shows the casimir energy as a function of @xmath1 at @xmath75 , that is a soliton of winding number one . the casimir energy as a function of @xmath1 , the scale of variation of the soliton , at @xmath75 .
the points displayed are the results of the numerical integration and the solid line represents the best fit to the points . in the illustrated box we focus on small values of @xmath1 to show the details of the maximum .
this maximum is precisely a cusp . ]
this graph has a sharp maximum at the critical value @xmath76 , where a bound state crosses @xmath77 precisely at @xmath75 .
this crossing is similar to what is depicted in fig.([fig.2 ] ) for @xmath22 where one bound state crosses @xmath77 at @xmath78 .
the appearance of the maximum at the critical point implies that constructing a soliton with this slope needs more energy than the ones with other slopes .
according to jackiw and rebbi s profound work @xcite , when the system has particle conjugation symmetry and there is a zero - energy fermionic mode , the soliton carries fermion number @xmath79 . in our problem which lacks this symmetry , our analysis shows that this point is energetically unfavorable .
however , this property might be shared by problems possessing this symmetry .
one can easily obtain the behavior of the casimir energy at @xmath75 numerically for large values of @xmath1 and the result is `` @xmath80 '' , where @xmath81 .
therefore , we can conclude that to within our numerical accuracy , it tends to zero as @xmath82 .
this is the expected result , as we shall describe shortly .
gousheh and mobilia obtained the difference between the total number of levels in the dirac sea in the presence and absence of the soliton , and in the limit @xmath82 their result is as follows @xmath83}.\ ] ] using this relation , the casimir energy for the case @xmath84 would be @xmath85 } + \sum\limits_{i}\left(e_{\text{bound}}^{i-}\right).\end{aligned}\ ] ] the last term appears automatically due to our choice of the expansion of the fermi field in the presence of the background field ( we choose the annihilation ( creation ) operator for positive ( negative ) bound states ) .
when @xmath82 one bound state just reaches the dirac sea at @xmath86 , i.e. it becomes a threshold bound state .
therefore , like all threshold bound states in one spatial dimension , it should be counted as one half bound state . since its energy is @xmath87 , the last term in eq.([r3 ] ) is @xmath88 .
therefore , the casimir energy is zero in this limit at @xmath86 , since the integrand term vanishes identically there .
for small values of @xmath1 the graph of the casimir energy can be approximated by `` @xmath89 '' .
this shows that the casimir energy tends to zero , to within our numerical precision as @xmath90 .
therefore , as the slope of the soliton decreases to zero , the vacuum energy approaches that of the trivial vacuum , in spite of the residual non - trivial boundary conditions .
these two limits occur at any @xmath14 i.e. when we have proper solitons with winding number @xmath7 .
however , for arbitrary values of @xmath2 , although the result @xmath91 as @xmath90 still holds , for @xmath92 the casimir energy is in general non - zero and has the following form @xmath93\right)\cos \theta_0 \right ] + m
\cos \theta_0 \theta\left(-\cos \theta_0\right ) \theta \left(\theta_0-\left[\frac{\theta_0}{\pi}\right]\pi\right ) , \end{split}\ ] ] where @xmath94 denotes the usual step function defined by , @xmath95 and @xmath96 $ ] extracts the integer part of @xmath97 . in general the jumps in the energy of continuum and bound states at the points where threshold bound states appear cancel out so that the casimir energy is continuous . in fig.([fig.4 ] ) we show the casimir energy as a function of @xmath2 for @xmath22 .
note that the casimir energy is , on the average , an increasing function of @xmath2 for fixed @xmath1 . comparing fig.([fig.4 ] ) with fig.([fig.2 ] ) , we can conclude that there are local maxima in the form of mild cusps in the graph of the casimir energy when bound states cross @xmath77 , as we vary @xmath2 . in eq.([e12 ] ) the integrand of the spatial integral represents the casimir energy density . in what follows
we shall compute and display the changes for both the positive and negative parts of the spectrum for comparison .
in general the change in the energy density naturally divides into two pieces .
first is the contribution coming from the continuum states ( @xmath98 ) and second the contribution coming from the bound states ( @xmath99 ) . the casimir energy as a function of @xmath2 for @xmath22 .
the mild cusps in the graph are actually mild maxima and occur when bound states cross @xmath77 .
this also occurs in fig.([fig.3 ] ) , except that the maximum there is much more pronounced.,width=306 ] the left graph in fig.([fig.5 ] ) shows the energy densities of negative and positive bound and continuum states for the case @xmath22 at @xmath75 , given in eqs.([r6],[r62],[r7],[r72 ] ) .
note that an overall positive energy density in the sea and overall negative energy density in the sky indicate that both of the continua contain spectral deficiency as compared to the free case .
the sum of these deficiencies is precisely canceled by the sum of bound state densities , at each point of space @xcite .
the sum of @xmath100 and @xmath101 ( solid and dashed lines in the left graph , respectively ) is the total distortion of the states with @xmath4 , and is depicted in the right graph in fig.([fig.5 ] ) by the solid line .
also , in this graph we show the sum of @xmath102 and @xmath103 ( dotted and dotdashed lines in the left graph , respectively ) , which is the total distortion of the states with @xmath3 , by a dashed line .
this figure shows that these two total densities are exactly the mirror images of each other .
it is important to note that in problems which possess charge conjugation symmetry , such as ours , we can express the casimir energy by the following equations . @xmath104 the energy densities as a function of @xmath105 for @xmath22 at @xmath75 .
the left graph : the solid ( dotted ) line shows the energy density of the negative ( positive ) continuum states and dashed ( dotdashed ) line shows the energy density of the negative ( positive ) bound state .
the right graph : the solid line shows the sum of the negative bound and continuum energy states and the dashed line shows the sum of the positive bound and continuum energy states .
the casimir energy density can be represented by the solid line . ]
now we can consider the effect of the casimir energy on the total energy of the system with a real fermion present at the ground state level .
for such a system the total energy is the sum of the soliton mass and the energy of the single fermion , if the casimir energy is not taken into account .
since the ground state of the system is distorted by the presence of the soliton , this portion should be added .
note that for those values of @xmath1 for which the energy of the fermion is negative , we should not add this energy , since this portion has already been taken into account by the casimir energy .
figure ( [ fig.7 ] ) shows the total energy ( solid line ) along with the casimir energy ( dashed line ) , the soliton mass ( dotdashed line ) and the energy of the fermion in the ground state ( dotted line ) . in this figure we just depict the binding energy as long as it is positive .
the main portion in the total energy originates from the soliton mass .
both the fermion and casimir energies are multiplied by @xmath106 to make them more visible on the graph . the total energy ( the sum of the energy of a real fermion , the soliton mass and the casimir energy ) as a function of @xmath1 .
the solid line shows the total energy , dashed line the casimir energy , dotdashed line the energy of the soliton ( @xmath107 given in eq.([e1r ] ) ) , and the dotted line the energy of a real fermion . the fermion energy and casimir energy
are multiplied by @xmath106 to make them more visible . ]
in this paper we have investigated a system consisting of a pseudoscalar field interacting non - linearly with a fermi field .
the pseudoscalar field in this model is static and prescribed , and by varying its parameters , that is the slope @xmath1 and its asymptotic values @xmath0 , we can form different configurations including the ones with a proper topological charge . using the exact fermionic spectrum for this model
, we have calculated the exact value of the casimir energy by subtracting the vacuum energy in the presence and absence of the soliton .
we have displayed the casimir energy as a function of @xmath2 and @xmath1 and discovered that this energy has a sharp maximum at the points where the fermi field has zero - mode bound states which would correspond to a fractional fermion number for the soliton if particle conjugation symmetry was not broken .
therefore , our analysis shows that the configurations which have zero fermionic modes are less likely to occur .
as expected , the casimir energy tends to zero as @xmath82 when @xmath2 is an integer multiple of @xmath108 and as @xmath90 for all values of @xmath2 .
furthermore , we have shown the casimir energy as a function of @xmath2 at @xmath22 .
this graph has two mild cusps at the points where a bound state crosses @xmath77 , and this is the generic behavior of the casimir energy for any values of @xmath1 .
then , we have exhibited the contributions of the bound and continuum states to the total energy densities for the states with @xmath4 and @xmath3 , separately and found out that they are the mirror images of each other at each point of space .
finally , we have added the casimir energy to the total energy of a system consisting of a soliton and a real fermion in the ground state .
we have concluded that although the contribution of the casimir energy is small compared to the soliton mass , it is not negligible in comparison with the fermion energy .
we would like to thank the research office of the shahid beheshti university for financial support . 9 h.b.g .
casimir , proc .
* 51 * , 793 ( 1948 ) ; h.b.g .
casimir and d. polder , phys . rev . *
73 * , 360 ( 1948 ) .
bezerra , g. bimonte , g.l .
klimchitskaya , v.m .
mostepanenko and c. romero , eur .
j. c * 52 * , 701 ( 2007 ) ; r. moazzemi , m. namdar and s.s .
gousheh , jhep * 09 * , 029 ( 2007 ) ; r. moazzemi and s.s .
gousheh , phys .
b * 658 * , 255 ( 2008 ) ; s.s .
gousheh , r. moazzemi and m.a .
valuyan , phys .
b * 681 * , 477 ( 2009 ) ; h. cheng , phys .
d * 82 * , 045005 ( 2010 ) .
f.d . mazzitelli , m.j .
sanchez , n.n .
scoccola and j. von stecher , phys .
a * 67 * , 013807 ( 2003 ) ; d.a.r .
dalvit , f.c .
lombardo , f.d .
mazzitelli and r. onofrio , europhys . lett . *
67 * , 517 ( 2004 ) ; p.a.m .
neto , j. opt .
b : quantum semiclass . opt .
* 7 * , s86 ( 2005 ) ; d.a.r .
dalvit , f.c .
lombardo , f.d .
mazzitelli and r. onofrio , phys .
a * 74 * , 020101 ( 2006 ) ; e.k .
abalo , k.a .
milton and l. kaplan , phys .
d * 82 * , 125007 ( 2010 ) .
r. balian and b. duplantier , ann .
* 112 * , 165 ( 1978 ) ; k.a .
milton , l.l .
deraad and j. schwinger , ann .
* 115 * , 388 ( 1978 ) ; c.m .
bender and k.a .
milton , phys .
d * 50 * , 6547 ( 1994 ) ; m. bordag , e. elizalde , k. kirsten and s. leseduarte , phys .
d * 56 * , 4896 ( 1997 ) .
casimir , proc .
* 51 * , 793 ( 1948 ) .
deraad and k.a .
milton , ann .
* 136 * , 229 ( 1981 ) .
boyer , phys . rev . * 174 * , 1764 ( 1968 ) .
w. lukosz , physica * 56 * , 109 ( 1971 ) ; j.r .
ruggiero , a. villani and a.h .
zimerman , j. phys .
* 13 * , 761 ( 1980 ) ; s. hacyan , r. jauregui and c. villarreal , phys .
a * 47 * , 4204 ( 1993 ) ; g.j .
maclay , phys .
a * 61 * , 052110 ( 2000 ) ; x. li and x. zhai , j. phys . a : math
. gen . * 34 * , 11053 ( 2001 ) ; h. cheng , j. phys . a : math
35 * , 2205 ( 2002 ) ; m.a .
valuyan , r. moazzemi and s.s .
gousheh , j. phys .
b : at . mol .
. phys . * 41 * , 145502 ( 9pp ) ( 2008 ) ; a. seyedzahedi , r. saghian and s.s .
gousheh , phys . rev .
a * 82 * , 032517 ( 2010 ) .
f. caruso , n.p .
neto , b.f .
svaiter and n.f .
svaiter , phys .
d * 43 * , 1300 ( 1991 ) ; x.z .
cheng , j.m .
li and x.h .
zhai , phys .
d * 56 * , 2155 ( 1997 ) ; r.m .
cavalcanti , phys .
d * 69 * , 065015 ( 2004 ) ; m.p .
hertzberg , r.l .
jaffe , m. kardar and a. scardicchio , phys .
lett . * 95 * , 250402 ( 2005 ) ; h. alnes , f. ravndal , i.k .
wehus and k. olaussen , phys .
d * 74 * , 105017 ( 2006 ) ; a. edery and i. macdonald , jhep * 09 * , 005 ( 2007 ) .
milton , l.l .
deraad and j. schwinger , ann .
( n.y . ) * 115 * , 388 ( 1978 ) .
t.p . branson and p.b .
gilkey , commun .
partial differential eqs . *
15 * , 245 ( 1990 ) ; m. bordag and k. kiresten , int . j. mod
a * 17 * , 813 ( 2002 ) .
r. balian and b. duplantier , ann .
( n.y . ) * 112 * , 165 ( 1978 ) .
dashen , b. hasslacher and a. neveu , phys .
d * 10 * , 4130 ( 1974 ) .
gervais and a. neveu(eds . ) , phys .
* 23 * , 237 ( 1976 ) ; h.j .
vega , nucl .
* b115 * , 411 ( 1976 ) ; j. verwaest , nucl . phys . * b123 * , 100 ( 1977 ) ; l.d . faddeev and v.e .
korepin , phys . rept . * 42 * , 1 ( 1978 ) ; r. rajaraman , _ solitons and instantons : an introduction to solitons and instantons in quantum field theory _
( north - holland , amsterdam , 1982 ) ; m. bordag , a.s . goldhaber , p. van nieuwenhuizen and d. vassilevich , phys . rev .
d * 66 * , 125014 ( 2002 ) ; a. rebhan , p. van nieuwenhuizen and r. wimmer , new j. phys . * 4 * , 31 ( 2002 ) ; g. mussardo , v. riva and g. sotkov , nucl .
b699 * , 545 ( 2004 ) .
j.a . bagger and s.g .
naculich , phys .
lett . * 67 * , 2252 ( 1991 ) ; phys .
d * 45 * , 1395 ( 1992 ) ; s.g .
naculich , phys .
d * 46 * , 5487 ( 1992 ) ; e. farhi , n. graham , r.l .
jaffe and h. wiegel , nucl .
* b585 * , 443 ( 2000 ) ; e. farhi , n. graham , r.l . jaffe and h. wiegel , phys .
b * 475 * , 335 ( 2000 ) .
a. rebhan and p. van nieuwenhuizen , nucl
b508 * , 449 ( 1997 ) ; a.s .
goldhaber , a. litvintsev and p. van nieuwenhuizen , phys .
d * 67 * , 105021 ( 2003 ) ; a. rebhan , p. van nieuwenhuizen and r. wimmer , nucl .
* b679 * , 382 ( 2004 ) .
sparnaay , physica * 24 * , 751 ( 1958 ) .
lamoreaux , phys .
* 78 * , 5 ( 1997 ) ; s.k .
lamoreaux , phys .
lett . * 81 * , 5475(e ) ( 1998 ) .
u. mohideen and a. roy , phys .
* 81 * , 4549 ( 1998 ) ; a. roy and u. mohideen , phys .
lett . * 82 * , 4380 ( 1999 ) ; g. bimonte , d. born , e. calloni , g. esposito , u. huebner , e. ilichev , l. rosa , f.tafuri and r. vaglio , j. phys . a : math
. theor . * 41 * , 164023 ( 8pp ) ( 2008 ) . l. shahkarami and s.s .
gousheh , jhep * 06 * , 116 ( 2011 ) .
s.s . gousheh and r.l .
mobilia , nucl .
b428 * , 189 ( 1994 ) .
r. mackenzie and f. wilczek , phys .
d * 30 * , 2194 ( 1984 ) .
j. goldstone and f. wilczek , phys .
lett . * 47 * , 986 ( 1981 ) .
r. jackiw and c. rebbi , phys .
d * 13 * , 3398 ( 1976 ) . | in this paper we compute the casimir energy for a coupled fermion - pseudoscalar field system . in the model considered in this paper
the pseudoscalar field is _ static _ and _ prescribed _ with two adjustable parameters .
these parameters determine the values of the field at infinity ( @xmath0 ) and its scale of variation ( @xmath1 ) .
one can build up a field configuration with arbitrary topological charge by changing @xmath2 , and interpolate between the extreme adiabatic and non - adiabatic regimes by changing @xmath1 .
this system is exactly solvable and therefore we compute the casimir energy exactly and unambiguously by using an energy density subtraction scheme .
we show that in general the casimir energy goes to zero in the extreme adiabatic limit , and in the extreme non - adiabatic limit when the asymptotic values of the pseudoscalar field properly correspond to a configuration with an arbitrary topological charge .
moreover , in general the casimir energy is always positive and on the average an increasing function of @xmath2 and always has local maxima when there is a zero mode , showing that these configurations are energetically unfavorable . we also compute and display the energy densities associated with the spectral deficiencies in both of the continua , and those of the bound states .
we show that the energy densities associated with the distortion of the spectrum of the states with @xmath3 and @xmath4 are mirror images of each other .
we also compute and display the casimir energy density .
finally we compute the energy of a system consisting of a soliton and a valance electron and show that the casimir energy of the system is comparable with the binding energy . |
to explain our intended approach to integrating high and low level planning , we introduce the high level motion plan random _ trajectory _ variable @xmath5 that is governed by the distribution @xmath6 and conditioned on symbolic data @xmath7 .
we treat the high level plan as a random variable because of the following : the high level planner must be able to accommodate local disturbances returned by the low level motion planner . in turn
, high level motion plans must be able to adjust to online goal changes ; these high level changes must then trickle down to low level behavior .
conceptually then , the high level plan and the low level plan are _ coupled _ variables ; if either is restricted to a single hypothesis ( as is typical in conventional approaches to hierarchical planning ) , then the high and low level plans are unable to influence each other .
similarly , we represent the low level motion plan with a random trajectory variable @xmath8 that is governed by the joint distribution @xmath9 over the platform and environmental agents @xmath10 , where @xmath11 $ ] is platform state data ( such as localization information ) and agent state data respectively ( as in @xcite ) . because the high and low level plans are dependent , we seek models for @xmath12 , the joint distribution over the high level plan , low level plan , and environmental agents . in taking a probabilistic approach , our first challenge lies in modeling the distribution in a faithful yet tractable way . in figure
[ fig : single - level - hierarchy ] we present the graphical model of our approach , inspired by the approach taken in @xcite ; similarly , undirected graphical models of this sort ( markov random fields ) have enjoyed great success in the image and natural processing literature . according to the graphical model in figure [ fig : single - level - hierarchy ] , the distribution factors as @xmath13 where @xmath14 encodes the `` agreeability '' of the low level plan and the high level plan and @xmath15 encodes the most likely path through the dynamic , responsive environment ( e.g. , a crowd of humans ) . as was argued in @xcite , our model immediately suggests a natural way to perform navigation : at time @xmath16 , find the _ maximum a - posteriori _
( map ) assignment for the joint distribution @xmath17 , and then take @xmath18 as the next action in the path . since @xmath19 captures the most probable joint value of the high level and low level plan , @xmath18 captures the next actuator command that is most in agreement with the high level plan and most likely to succeed in navigating through the dynamic , responsive environment . in the following case studies , we present scenarios that move from the most limited form of high level guidance ( designating a goal in a map ) to a scenario where a human operator interacts arbitrarily ( sets as many waypoints as desired / grabs and releases the joystick whenever he pleases ) .
* the environment is static with a single known obstacle .
thus , we have @xmath20 ( i.e. , there is no crowd ) . *
the operator designates the goal @xmath21 in the predefined map @xmath22 .
* we can thus utilize a standard global planner ; call it a*. we let @xmath23 .
thus , @xmath24 . *
our local planner is trivial , since there are no local disturbances .
* we compute `` actuator inputs '' by finding @xmath25 where @xmath26 is the kinematic model of the robot , and @xmath27 is the interaction function between the robot and the global plan . the operator - robot interaction function
@xmath27 could take a number of forms ; we choose @xmath28[{{\boldsymbol \eta}}_0-{{\mathbf{f}}^r}]^\top)$ ] .
thus , @xmath29 converges to @xmath30 in the situation of figure [ fig : gotog ] . recovering the ros navigation stack with this approach is trivial : at each time step @xmath16 , sample local paths @xmath31 , and weight each sample according to @xmath32the first factor encodes global compatibility , while the second factor encodes kinematic feasibility .
choose the sample @xmath33 with the highest weight as the inputs to the actuators .
the probabilistic formulation allows us to approach the dwa ros navigation stack in a more general manner : in the ros navigation stack , sampling from @xmath26 and then weighting amounts to straightforward importance sampling .
however , the distribution @xmath34 can be approximately inferred using a host of methods : markov chain monte carlo , laplace approximation , hybrid monte carlo , etc any approximate inference technique is at our disposal .
in contrast , the ros navigation stack ( http://wiki.ros.org/base_local_planner ) , does not pose the high level to low level path planning problem as a probability distribution , so it is not immediately clear how to employ approximate inference techniques to find more accurate solutions in a more efficient manner . [ [ single - global - operator - instruction - multiple - static - and - dynamic - obstacles ] ] single global operator instruction , multiple static and dynamic obstacles ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in figure [ fig : gotog_multiple ] we introduce the notion of multiple global plans , each of which have nontrivial value .
in particular , global plans @xmath35 have values in the static map of @xmath36 .
the global plan distribution thus takes the form @xmath37 in figure [ fig : gotog_multiple_dynamic ] , we introduce a local crowd disturbance in the bottom right of the map .
we assume that the crowd enters into the robot s field of view near the center corridor ; thus , the robot has to make a planning decision according to @xmath38 when the crowd is not in the robot s field of view , @xmath20 , and the low level planner stays close to the optimal global plan @xmath30 .
however , when @xmath39 , it is no longer obvious which global plan to follow . with our probabilistic approach , which global path to follow
is determined by balancing the capabilities of the low level planner in the crowd ( effectively , how much probability is in @xmath15 near the global plan @xmath40 ) against how much more efficient @xmath30 is than @xmath41 ( or , how @xmath42 compares to @xmath43 ) .
heuristically , one can think of the distribution @xmath44 as having three modes , each ( roughly ) centered around the global plans . the relative probability mass in each mode ends up determining the map value of the distribution .
thus , the global plan s fitness represented by @xmath45is balanced against the challenge of the local situation , which is represented by @xmath15 . now
, suppose that the operator has provided a global goal @xmath21 ( and thus high level plans are generated ) , but intervenes via a joystick at random times according to @xmath46 and @xmath47 , as in figure [ fig : intervene ] ( the difference between this scenario and the scenario in figure [ fig : gotog_multiple_dynamic ] is the presence of the joystick data ) .
now , not only do we have to balance global considerations ( the weights of the global plans ) against local disturbances , but also the online desires of the operator .
in particular , the robot will move through the environment in the same manner as in figure [ fig : gotog_multiple_dynamic ] , until the operator intervenes with the joystick at @xmath46 . at this point ,
the global plan distribution will become @xmath48 , and thus influence local decision making by `` pulling '' @xmath15 towards the more peaked regions of @xmath48we are able to simultaneously represent high level operator desires with online refinements .
our full joint distribution now becomes @xmath49 .
importantly , in the absence of a global goal @xmath21 , the formulation reduces to @xmath50 .
this is the case of fully assistive shared control , where the absence of a global map or corrupted localization data renders the global goal @xmath21 meaningless . in this case
, the global plan @xmath5 is revealed incrementally via local user input data @xmath46 .
this capability becomes important when , for instance , the robot enters a crowd , and standard localization techniques start to fail at this point the robot must `` share awareness '' with the operator by inferring global destinations from local operator input data .
while the success of previous experiments and the utility of the markov random field factorization lend credence to our model above , we point out that results from formal methods ( and thus provably correct constructions ) can guide how we model our joint distribution ( courtesy of ufuk topcu ) . to see how
, we refer to figure [ fig : pgm - to - hier ] , where we have plotted the state of the art formal methods decision stack next to its corresponding graphical model decomposition .
note that the results from formal methods suggests that a `` tactical variable '' , which we call @xmath51 , is used to mediate information between the high level @xmath5 and the low level @xmath8 ( we assume that some form of tactical data , @xmath52 informs the governing distribution @xmath53 ) .
this graphical model in turn represents the factorization one of our ongoing research objectives is to fully understand how formal methods can guide our probabilistic decompositions while the probabilistic approach is well suited to capture dependencies between variables and flexible enough to capture the vagaries of human behavior ( or online manipulation of robot objectives ) , balancing tractability and fidelity in the factorization of the joint distribution is more of an anecdotal art than a science .
results from formal analysis , however , can provide guidance on our decomposition and potentially insight into the form of our `` cooperation functions '' @xmath55 and @xmath56 that link the mission , tactical , and trajectory levels .
furthermore , it is not immediately clear how to relate data coming in at various levels to planning level variables ( e.g. , high level symbolic data @xmath57 clearly relates to @xmath5 ; however , the introduction of other planning levels necessitates understanding of how lower level data such as joystick commands in the form of @xmath58measures lower level planning variables ) .
the hierarchical decision stack illustrated in figure [ fig : pgm - to - hier ] was tied to a specific application , and so is not in general the correct hierarchical decomposition .
however , the approach of finding provably correct hierarchical decompositions for arbitrary scenarios , and then reading off the corresponding graphical model ( and thus probabilistic decomposition ) is fully general ; we depict this approach in figure [ fig : arbitrary - stacks ] . in combination with human - learning and symbolic planner approaches ( which guide how we model and adapt distributions , such as @xmath59 and @xmath60 , at specific levels of the planning stack )
, our approach has the potential to be both flexible enough to accommodate a wide variety of online manipulation of global robot objectives while maintaining the rigor of formal analysis . | we present a possible method for integrating high level and low level planning . to do so , we introduce the global plan random _ trajectory _ @xmath0
\to \mathbb r^2 $ ] , measured by goals @xmath1 and governed by the distribution @xmath2 .
this distribution is combined with the low level robot - crowd planner @xmath3 ( from @xcite ) in the distribution @xmath4 .
we explore this _ integrated planning _
formulation in three case studies , and in the process find that this formulation 1 ) generalizes the ros navigation stack in a practically useful way 2 ) arbitrates between high and low level decision making in a statistically sound manner when unanticipated local disturbances arise and 3 ) enables the integration of an onboard operator providing real time input at either the global ( e.g. , waypoint designation ) or local ( e.g. , joystick ) level .
importantly , the integrated planning formulation @xmath4 highlights failure modes of the ros navigation stack ( and thus for standard hierarchical planning architectures ) ; these failure modes are resolved by using @xmath4 .
finally , we conclude with a discussion of how results from formal methods can guide our factorization of @xmath4 . |
nonsingular spacetimes and those containing black holes are qualitatively quite different . nevertheless , it is possible to find sequences of spacetimes that , while remaining nonsingular , come arbitrarily close to having horizons @xcite . in a previous paper
@xcite we studied a class of such solutions that are associated with self - gravitating monopoles in a spontaneously broken yang - mills theory .
the emphasis there was on the detailed behavior of the fields as one approaches the critical solution in which a horizon first appeared . in this paper
, we concentrate instead on the geometrical aspects of the spacetimes associated with these objects near criticality , and on using these to gain insights into the properties of true black holes . as in ref .
@xcite , we restrict ourselves to spherically symmetric spacetimes and write the metric in the form @xmath0 a horizon corresponds to a zero of @xmath1 ; the horizon is extremal if @xmath2 also vanishes .
we work in the context of an su(2 ) gauge theory with gauge coupling @xmath3 and a triplet higgs field @xmath4 whose vacuum expectation value @xmath5 breaks the symmetry down to u(1 ) . in flat spacetime
this theory possesses a finite energy monopole solution with magnetic charge @xmath6 and mass @xmath7 .
it has a core region , of radius @xmath8 , with nontrivial higgs and massive vector boson fields . beyond this core
is a coulomb region in which all massive fields approach their vacuum values exponentially rapidly , leaving only the coulomb magnetic field .
the effects of adding gravitational interactions depend on the value of @xmath5 . for @xmath5
much less than the planck mass @xmath9 , one finds that @xmath1 is equal to unity at the origin , decreases to a minimum at a radius of order @xmath10 , and then increases again with @xmath11 . as @xmath5
is increased , this minimum becomes deeper , until an extremal horizon develops at a critical value @xmath12 of the order @xmath9 . as we describe in more detail in sec .
ii , two distinct types of critical behavior are possible , depending on the ratio of the higgs and gauge boson masses . for lower values of this ratio
, one finds `` coulomb - type '' critical solutions , in which the horizon occurs in the coulomb region of the monopole at @xmath13 . outside the horizon
, the metric is that of an extremal reissner - nordstrom black hole , with @xmath14 while the massive fields take on their vacuum values . as the higgs self - coupling increases
, there is a transition to `` core - type '' critical solutions that have a horizon inside the monopole core and nontrivial matter fields ( or hair ) outside the horizon . in both types of critical monopole solutions
the fields remain nonsingular at @xmath15 .
however , it is also possible to have solutions with singularities at @xmath15 that can be viewed as self - gravitating monopoles with schwarzschild black holes at their center .
as long as the mass of the central black hole is not too great , the variation of these solutions with @xmath5 is quite similar to that of the nonsingular monopoles , and one finds the same two types of critical behavior @xcite . after reviewing these solutions ,
we discuss in sec .
iii how near - critical monopoles might appear to an outside observer .
one would expect the measurements made by such an observer to vary continuously with the parameters of the monopole and to show no discontinuity at the critical solution .
the external observer could probe the monopole with either particles or waves . in the case of the particle , we find that the time needed for the particle to emerge from the interior ( as measured by a static external observer using `` schwarzschild time '' ) diverges as the critical solution is approached . when a wave is sent in , there is a reflected wave due to the gravitational field just outside the horizon and a transmitted wave that passes through the interior and then emerges with a time delay . as before , the time delay diverges as @xmath16 , while the reflected wave becomes indistinguishable from that due to a black hole .
using either type of probe , an observer whose lifetime is finite can not distinguish between a true black hole and a nonsingular , subcritical solution that is sufficiently close to being critical .
we discuss the implications for our understanding of black hole entropy .
we also find that the near - critical coulomb - type solutions display what horowitz and ross @xcite have termed naked - black - hole behavior , even though there is no black hole at all .
this is characterized by the fact that a freely - falling observer passing through the minimum of @xmath1 ( we shall refer to the location of this minimum as the quasi - horizon ) feels a tidal force that diverges as the critical solution is approached . for core - type solutions , on the other hand ,
no such behavior is observed . in sec .
iv , we consider the effect of having additional matter fall into a near - critical solution , addressing in particular the question of whether this process could produce an extremal black hole .
extremal black holes are especially interesting from the standpoint of black hole thermodynamics because they have vanishing hawking temperature .
the analogies between black hole dynamics and thermodynamics thus suggest that they should be rather difficult , if not impossible , to create .
indeed , one of the formulations @xcite of the third law of black hole dynamics asserts the impossibility ( under certain technical assumptions ) of making a nonextremal black hole extremal .
one could also envision producing an extremal black hole starting from a nonsingular spacetime .
boulware @xcite showed that this can be done by the collapse of a charged shell of matter .
however , this mechanism relies critically on the shell being infinitely thin ; shells of finite thickness and density do not collapse to an extremal configuration .
it is easy to understand the difficulty of making an extremal black hole if one recalls that the extremal reissner - nordstrom black hole is characterized by having a mass and a charge that ( in appropriately rescaled planck units ) are equal .
forming such an object by the collapse of a shell with equal charge and mass densities involves a delicate balance between electromagnetic and gravitational forces .
one could instead try to achieve extremality by adding matter to a pre - existing nonextremal reissner - nordstrom black hole ( i.e. one with greater mass than charge ) .
however , because the added matter would have to have more charge than mass , the coulomb repulsion between the black hole and the infalling matter would tend to overcome their gravitational attraction .
the situation is rather different in our case , because the nonsingular monopole solutions are _ overcharged _ ; i.e. , their long range fields are those of a reissner - nordstrom solution with greater charge than mass . allowing uncharged matter to fall into these objects increases their mass and should bring them closer to criticality . if the amount of matter entering is just sufficient to create a zero of @xmath1 , one would expect an extremal solution to result .
we will present numerical arguments that support this expectation .
finally , we make some concluding remarks in sec . v.
as in ref .
@xcite , we consider an su(2 ) gauge theory that is spontaneously broken to u(1 ) by the vacuum expectation value @xmath5 of a triplet higgs field @xmath4 .
this theory has magnetic monopole solutions that can be described by the spherically symmetric ansatz @xmath17 finiteness of the energy requires that @xmath18 and @xmath19 . if the solutions are also required to be nonsingular at @xmath15 , then @xmath20 and @xmath21 . in a spacetime with a metric of the form of eq .
( [ metric ] ) , the static field equations for these matter fields can be derived from a @xmath22-dimensional action of the following form @xcite @xmath23 \label{matteraction}\ ] ] where @xmath24 , involves the fields but not their derivatives , while @xmath25 the euler - lagrange equations for the matter fields that follow from eq .
( [ matteraction ] ) must be supplemented by the gravitational field equations . for static , spherically symmetric field configurations these reduce to @xmath26 = - 4\pi g \left({k\over a } + u \right)\ ] ] and @xmath27 here
carets indicate orthonormal components .
equation ( [ abeqn ] ) can be immediately integrated to obtain @xmath28\ . \label{bsolution}\ ] ] using this to eliminate @xmath29 , one is left with two second order and one first order equation for the functions @xmath30 , @xmath31 , and @xmath32 .
these equations must be solved numerically . up to a rescaling of distances ,
the solutions of these equations depend only on the two dimensionless parameters @xmath33 and @xmath34 . for small values of @xmath35 ( roughly @xmath36 for a quartic higgs ) @xcite one finds coulomb - type solutions in which the minimum of @xmath1 is located outside the monopole core
. this minimum decreases . for a detailed description ,
see @xcite . ] with increasing @xmath37 , until the critical solution is reached at @xmath38 . in the critical solution , the matter fields @xmath30 and @xmath31 reach their asymptotic values @xmath39 and @xmath40 at the horizon and are then constant for all @xmath41 ; because both fields fall as fractional powers of @xmath42 , the derivatives @xmath43 and @xmath44 both diverge as @xmath45 approaches @xmath46 from below .
( this nonanalytic behavior is possible because an extremal horizon is a singular point of the matter field equations . )
the metric of the critical solution is identical to the extremal reissner - nordstrom metric outside the horizon , but differs from it for @xmath47 .
the metric function @xmath1 varies relatively smoothly , falling monotonically from unity at the origin to a zero at the horizon .
just inside the horizon @xmath48 , with @xmath49 being larger than for the corresponding reissner - nordstrom solution .
the behavior of @xmath29 contrasts sharply with this .
equation ( [ bsolution ] ) shows that the product @xmath50 ( which is identically equal to unity in both the schwarzschild and reissner - nordstrom solutions ) is given by an integral of the functional @xmath51 .
the singularities in the derivatives of @xmath30 and @xmath31 at the horizon are strong enough to cause this integral to diverge , so that the ratio @xmath52 is infinite in the critical limit .
if we adopt the conventional normalization @xmath53 , then @xmath29 vanishes identically inside the horizon .
if we instead set @xmath54 , then @xmath29 is finite and varying inside the horizon and infinite for @xmath55 ; depending on the value of @xmath35 , the minimum of @xmath29 may be at @xmath15 or at some finite radius , but in neither case does @xmath29 have a zero . for near - critical solutions where the minimum value @xmath56 is small but nonzero , we find that the ratio @xmath57 varies as @xmath58 , where @xmath59 ranges from about 0.7 to unity .
a rather different type of critical solution is found for larger @xmath35 .
for these core - type solutions the horizon occurs at a radius @xmath60 , with the values @xmath61 and @xmath62 of the matter fields at this point being different than their asymptotic values .
although the solutions are still nonanalytic at the horizon , this nonanalyticity occurs only in subdominant terms .
thus , @xmath1 again vanishes as @xmath63 as one approaches the horizon , but the coefficient is now the same inside and outside the horizon . ,
subcritical monopoles exhibit both core - type and coulomb - type quasi - horizons .
however , as one approaches criticality for a given value of @xmath35 , only one quasi - horizon actually becomes a horizon .
the other quasi - horizon , though interesting , is essentially irrelevant for our purposes . ]
the radial derivatives of both matter fields are finite at the horizon , so @xmath64 remains finite and there is no sharp change in @xmath50 at the horizon . because @xmath50 remains finite and nonzero
, @xmath29 has a zero at the horizon that coincides with the zero of @xmath1 .
these solutions can be generalized to include a black hole in the center of the monopole . instead of requiring that the fields be nonsingular at @xmath15 , one instead
requires that there be a zero of @xmath1 at a nonzero radius @xmath65 . at this zero , the equations for the matter fields become constraint equations relating the fields and their first derivative ; solving these constraints yields enough boundary conditions to determine a solution .
if @xmath65 is not too large , the effect of increasing @xmath37 is similar to what it is in the absence of a central black hole .
, see the discussion in @xcite .
] there is an outer minimum of @xmath1 that moves downward , finally reaching zero and becoming an extremal horizon at some critical value @xmath66 . for small values of @xmath35
the solutions are coulomb - type , while for large @xmath35 one finds core - type critical solutions . rather than increasing @xmath37 with @xmath65 fixed
, one can instead increase @xmath65 with @xmath37 held fixed ; this is much more analogous to the process of actually dropping matter into a near - critical solution that we will consider in sec .
iv . for initial values of @xmath37
that are sufficiently close to @xmath67 , this gives a family of solutions with a critical limit . in figs .
[ fig : low ] and [ fig : big ] we illustrate this with coulomb - type solutions with @xmath68 and core - type solution with @xmath69 .
for any @xmath70 , the self - gravitating monopole solution is a nonsingular spacetime with a penrose diagram of the same form as that of minkowski spacetime ( fig .
[ fig : penrose]a ) .
the critical solution , on the other hand , can be extended beyond the original coordinate patch to yield a spacetime with the penrose diagram shown in fig .
[ fig : penrose]b .
this diagram is quite similar to that of an extremal reissner - nordstrom black hole , but differs from it by not having a singularity at @xmath15 .
the difference between the two diagrams is striking and seems to indicate a discontinuity at @xmath71 , in contradiction with the usual expectation that physical quantities should vary continuously with the parameters of a theory .
however , this discontinuity can been seen as an artifact of the conformal transformation that produces the penrose diagram from an infinite spacetime .
this can be illustrated by considering the points a and b that lie on the curves of constant @xmath45 that are shown in figs .
[ fig : penrose]a and [ fig : penrose]b .
these have been chosen so that it is possible for an object to start at a , move in to @xmath15 at c , and then travel out again to b. the total proper time along this worldline ( or the total affine parameter , if the worldline is lightlike ) is finite .
this should be compared with the proper time along the worldline of constant @xmath45 .
this is finite for the subcritical case , whereas in the critical case the proper times along the segments ad and db are both infinite , corresponding to the fact that an observer in the exterior region containing a can not detect objects behind the horizon pd . in order to obtain consistency with our physical expectations of continuity
, we should require that the proper time along the world of constant @xmath45 should diverge as @xmath72 .
more generally , the time required for an external observer to receive information from a probe of the interior regions solution should diverge in the critical limit . to see how this works out
, we consider probing the interior region ( i.e. , the region @xmath73 , where @xmath74 is the quasi - horizon ) of a near - critical solution by sending in either a particle or a wave . in both cases , we assume that the probe interacts only gravitationally , and has no direct interaction with the monopole fields .
to begin , imagine releasing a massive particle from an initial radius @xmath75 that is large enough that we may approximate @xmath76 .
the rotational and time - translation symmetries of the metric allow us to take the motion to lie in the @xmath77 plane and guarantee the conservation of the angular momentum per unit mass @xmath78 and the energy per unit mass @xmath79 these , together with eq .
( [ metric ] ) , imply that @xmath80^{1/2 } \label{drdt}\ ] ] if @xmath81 , the particle falls radially in , pass through the origin , and emerge on the other side of the monopole .
if instead @xmath82 , the particle turns around after reaching a minimum radius @xmath83 and return to @xmath84 with its trajectory advanced by an angle @xmath85^{-1/2}\ .\ ] ] we are interested in the proper time measured by an observer who remains at @xmath86 . assuming that the observer is not moving at relativistic speeds , this is approximately the same as the schwarzschild coordinate time @xmath87 , which over the course of the entire trajectory increases by an amount @xmath88^{-1/2}\ . \label{dt}\ ] ] there are two potential sources of divergences in this integral as @xmath89 and the critical solution is approached . in both types of critical solutions
there is a contribution from @xmath90 associated with the growth of @xmath91 . in the coulomb - type solutions
the near - vanishing of @xmath92 [ see eq .
( [ bsolution ] ) ] gives a second contribution from the entire region @xmath93 .
let us examine these in more detail . for core - type solutions , in the region @xmath90 we can write @xmath94^{-1}\ ] ] with @xmath95 of order unity , while @xmath92 is roughly constant and independent of @xmath96 . because @xmath97 is small ,
the @xmath98-dependent term in @xmath99 can be neglected for any @xmath98 such that the particle could have reached @xmath74 . by a similar argument , we see that any particle that reaches @xmath74 goes through the peak of @xmath32 before turning around . hence , @xmath100 where the ellipses represent subdominant terms . for coulomb solutions the dominant effect is due to the fact that @xmath101 is almost vanishing throughout the interior region .
because our numerical solutions show that @xmath59 ranges between 0.7 and unity , the divergence due to this effect is greater than that from the region near @xmath74 . furthermore , the near - vanishing of @xmath29 in the interior implies that any particle that enters the interior almost reaches the origin , so that @xmath102 .
thus , @xmath103 where @xmath104 is of order unity . rather than sending in a particle , one can also probe the near - black hole by sending in a wave packet . as an example , let us consider a free massive scalar field @xmath4 , whose field equation in curved spacetime takes the form @xmath105 + m^2 \phi\ . \label{scalar}\ ] ] this can be put into a more tractable form by writing @xmath106 and defining a new coordinate @xmath107 satisfying @xmath108 equation ( [ scalar ] ) then takes the form of a one - dimensional wave equation @xmath109 \psi\ ] ] with a scattering potential @xmath110 + { j(j+1 ) b\over r^2}\ .\ ] ] when a wave packet incident from large @xmath45 reaches the region near the quasi - horizon , a portion is reflected by the scattering potential , while a portion is transmitted and emerge with some time delay .
if a near - critical solution is to appear essentially indistinguishable from a black hole to an outside observer , two conditions must hold .
first , the reflection coefficient as a function of wave number must approach that of the black hole as @xmath111 .
second , the time delay in the emergence of the transmitted wave should diverge in the critical limit . to see how the first of these conditions comes about ,
let us use eq .
( [ abeqn ] ) to rewrite the scattering potential as @xmath112\ .\ ] ] for both core- and coulomb - type solutions the quantity @xmath113 remains finite in the critical limit , while the second term in the brackets is zero at the quasi - horizon . since @xmath114 , it is clear that @xmath115 vanishes at least as fast as @xmath96 as the critical limit is approached .
hence , in the limit the scattering potential splits into two parts , one inside and one outside the horizon .
the outer potential is equal either to that of an extremal reissner - nordstrom black hole ( in the coulomb case ) or that of a black hole with hair ( in the core case ) . because the variation of the outer potential with @xmath96 is smooth in both cases , our conditions on the reflection coefficients are satisfied if we can ignore reflection from the inner part of the potential .
this can be understood by noting that the natural distance variable in which to discuss the motion of the waves is @xmath116 . by integrating eq .
( [ newcoord ] ) inward from some reference point @xmath75 , we obtain @xmath117 the behavior of this integral as the critical limit is approached is very similar to that of the integral in the expression for @xmath99 , eq . ( [ dt ] ) . for either type of solution
, the region near the quasi - horizon gives a contribution that diverges at least as fast as @xmath118 .
there is a corresponding growth in both the effective distance from the inner portion of the potential to any external point and in the time delay of the corresponding reflected wave .
as @xmath96 is increased , an external observer at fixed @xmath45 first sees the reflections from the inner and outer parts of the potential split into two distinct reflected waves , and then finds that the time delay of the second reflected wave ( from the inner potential ) grows without bound . the portion of the wave that is transmitted through the region near the quasi - horizon either continues through the origin and then outward or reflects off a central centripetal barrier , according to whether or not @xmath98 vanishes . in either case ,
the time delay accumulated by this wave before it returns to the quasi - horizon grows in essentially the same manner as the travel time for a massive particle traversing the same path : as @xmath118 for a core - solution and as @xmath58 for a coulomb - type solution .
thus , regardless of the type of probe used , an external observer at fixed @xmath119 must wait for at least a time @xmath120 before the probe emerges from the region inside the quasi - horizon . to leading order , this time delay is independent of the energy or angular momentum of the probe , and is instead determined solely by the spacetime geometry .
hence , to an observer with a finite lifetime @xmath121 , the interior region of any near - critical configuration with @xmath122 is inaccessible .
is less than @xmath123 , the the boundary of the inaccessible region depends only very weakly on @xmath121 and @xmath119 , and is essentially indistinguishable from the quasi - horizon . ]
he would most naturally describe any larger system containing this configuration by a density matrix @xmath124 obtained by tracing over the degrees of freedom inside the quasi - horizon .
from this density matrix one can derive an entropy @xmath125 that can be associated with the interior of the quasi - black hole .
one could , of course , proceed in this manner to define an entropy for any arbitrary region in space , just as one can choose to make the information in any subsystem inaccessible by putting the subsystem behind a locked door .
the crucial difference here is that the inaccessibility is due to the intrinsic properties of the spacetime , and that the boundary of the inaccessible region is defined by the system itself rather than by some arbitrary external choice .
it is thus reasonable to define @xmath126 as _ the _ entropy of the quasi - black hole .
a precise calculation of this entropy is clearly infeasible . among other problems
, such a calculation would require a correct implementation of an ultraviolet cutoff , which presumably would require a detailed understanding of how to perform the calculation in the context of a consistent theory of quantum gravity .
as an initial effort , one can take the ultraviolet cutoff as the planck mass @xmath127 and ask for an order of magnitude calculation .
such a calculation was done by srednicki @xcite , who showed that tracing over the degrees of freedom of a scalar field inside a region of flat spacetime with surface area @xmath32 led to an entropy @xmath128 where @xmath129 is the ultraviolet cutoff and @xmath130 is a numerical constant .
furthermore , although the precise calculations depend on the details of the theory , srednicki gave general arguments suggesting that such an entropy should always be proportional to the surface area .
this leads us to expect that @xmath131 .
this result is , of course , consistent with the possibility that in the critical limit @xmath126 goes precisely to the standard black hole result @xmath132 .
however , in contrast with the usual black hole case , our spacetime configurations are topologically trivial .
the `` interior '' region enclosed by the quasi - horizon is nonsingular and static .
furthermore , this region can be unambiguously defined , so that it is conceptually clear what is meant by tracing over the interior degrees of freedom , even though it may not yet be possible to implement this calculation in complete detail .
we find it quite striking that by this approach one can arrive so nearly at the standard entropy result . in our discussion above of the trajectory of a particle probe
, we focussed on the coordinate time that elapses over the course of the particle s passage through the monopole .
however , it also of interest to consider the elapsed proper time , which can be found by integrating @xmath133 [ see eq .
( [ drdt ] ) ] . for core - type solutions
this gives a finite nonzero result with no unusual behavior as the critical limit is approached .
the situation with coulomb - type solutions is , on the other hand , quite striking .
the sharp decrease in @xmath50 at the quasi - horizon leads to a corresponding decrease in @xmath133 , so that the proper time elapsed while the probe is within the quasi - horizon is also has consequences for the shape of the trajectory through the interior . by combining eqs . (
[ j ] ) and ( [ drdt ] ) , one finds that in the critical limit all probes follow a straight line passing through the origin , regardless of the incident angle with which they hit the horizon . ]
@xmath134 in the critical limit @xmath50 vanishes identically for @xmath73 , and @xmath135 .
this vanishing of @xmath136 is related to another interesting property of these solutions .
it is well known that the riemann tensor is nonsingular at a black hole horizon .
it therefore does not seem surprising that in the most familiar black hole solutions , the schwarzschild and reissner - nordstrom , a particle suffers no unusual effects as it crosses the horizon .
however , horowitz and ross @xcite showed that this is not always the case .
because of the acceleration of a particle as it approaches the horizon , the components of the riemann tensor in a coordinate frame that is freely falling with the particle can be quite different from the components measured in a static frame . with a metric of the form of eq .
( [ metric ] ) , the components @xmath137 ( where @xmath49 denotes a transverse spatial direction and @xmath138 the time in the boosted frame ) are given by @xmath139\ , \label{nakedbh}\ ] ] where @xmath140 is the energy per unit mass of the infalling particle .
the fact that this curvature component is never large ( with @xmath140 of order unity ) for the schwarzschild and reissner - nordstrom black holes is a consequence of the fact that @xmath50 is constant in both cases .
this is not true in general .
horowitz and ross found several examples of dilaton black holes for which @xmath137 , and thus the tidal forces felt by an infalling particle , could be made arbitrarily large near the horizon by taking the solution to be sufficiently close to extremality .
they introduced the term naked black hole to indicate the fact that this ( almost ) singular behavior occurs outside the horizon .
subsequently @xcite , they showed that in these examples @xmath137 was inversely proportional to the square of the proper time remaining before the particle reached the singularity at @xmath15 . applying their results to our solutions
, we find that near - critical coulomb - type solutions display naked black hole behavior , even though they are not black holes at all .
this can be seen by noting that eq .
( [ abeqn ] ) implies that @xmath141 inside the quasi - horizon @xmath142 , while the radial derivatives of @xmath30 and @xmath31 , and hence @xmath64 , are of order unity . inserting this result into eq .
( [ nakedbh ] ) , gives @xmath143 note that this is proportional to @xmath144 , giving a relationship between tidal forces and proper time reminiscent of the examples described in ref .
to gain an insight into the third law of thermodynamics , as noted in sec .
i , it is of interest to determine whether systems with either initially nonsingular spacetimes or initially non - extremal black holes can evolve into systems with an extremal horizon .
it appears that under reasonable conditions of finiteness and causality this can not be done by adding charge to an undercharged object @xcite .
the discussion in the previous two sections , however , suggests an alternative . recall that our subcritical monopoles are overcharged ; i.e. , they have a charge larger than their mass . in the normal reissner - nordstrom case , such a system would exhibit a naked singularity .
however , in the monopole the coulomb core is screened by the massive particles in such a way that no gravitational singularity exists .
this suggests a scenario by which an extremal black hole is dynamically formed from a monopole by dropping in uncharged matter .
let us add to our theory a chargeless matter field that is coupled to the monopole fields only through gravity .
we then allow a spherical shell ( of small , but finite thickness ) of this matter fall into the monopole .
if the mass of the shell is sufficiently small , we do not expect a horizon to be formed . on the other hand ,
if the shell contains enough matter , the system should collapse to form a black hole .
it seems plausible that threshold case between these two regimes should produce an extremal horizon . in the case of a coulomb - type solution
, one might run into difficulties because of the naked - black - hole behavior it exhibits .
this concern , however , should not be an issue for the case of a core - type solution . here
the fields of the critical solution are much better behaved ; what nonanalyticities exist at the horizon are mild , and become increasingly so as one increases @xmath35 . moreover
, nothing unusual happens to the near - critical core - solutions as one parametrically approaches criticality .
adding a small schwarzschild black hole at the center of the self - gravitating monopole should not significantly change the scenario .
the infall of a spherical shell of appropriate mass should still turn the quasi - horizon of a near - critical solution into an extremal horizon .
the black hole would essentially play a spectator role , as its presence is largely irrelevant to the dynamics of the system .
we have carried out numerical simulations to test these ideas .
we begin with a core - type monopole solution that , for numerical convenience , has a small schwarzschild black hole with horizon radius @xmath145 at its center .
the parameters are chosen to be such that the solution is near criticality , so that only a small amount of additional matter is needed for the quasi - horizon at @xmath146 to collapse to a true horizon .
we add to the theory a massive scalar field @xmath147 that is coupled only to gravity .
we then send a spherically symmetric gaussian pulse of @xmath148 field in toward the monopole and watch the system evolve . to simplify the computation , we freeze the matter field variables @xmath30 and @xmath31 at their initial values ; because the fields for near - critical core - type monopoles are not very sensitive to the metric ( cf .
fig [ fig : big ] ) , this approximation should cause little error .
when the pulse amplitude is larger than some threshold value , the pulse falls into the monopole until the metric function @xmath1 develops a simple zero near the location of the quasi - horizon of the initial monopole configuration .
this newly formed horizon is non - extremal .
more interesting is the situation where the pulse amplitude is at , or just below , this threshold value .
figure [ fig : t1 ] shows a sequence of snapshots illustrating this scenario at four different points during the pulse s motion inward . as @xmath149 , the horizon of the internal black hole ; this corresponds to using a time variable appropriate to an observer in the interior of the monopole .
if we had used the more conventional normalization with @xmath50 equal to unity at spatial infinity , the time coordinate would be that appropriate to an external observer and would grow rapidly as the pulse approached the quasi - horizon . ]
( the distortion of the pulse is due to the backreaction of the monopole metric . )
[ fig : t1 ] [ fig : t2 ] [ fig : t3 ] as required by birkhoff s theorem , @xmath1 is is undisturbed ahead of the pulse , but undergoes a shift as the pulse passes .
thus , in the region ahead of the pulse @xmath1 is the same as it was in the original configuration , while behind the pulse it has the form corresponding to a static monopole with a central black hole whose horizon radius exceeds @xmath145 by an amount determined by the energy - momentum of the infalling pulse .
these two are joined by a kink at the pulse position . as the pulse passes through the quasi - horizon
, @xmath1 reaches its minimum .
similarly , the plot of @xmath92 appears as nearly a step function centered at the pulse position , with the inner and outer regions corresponding to the initial and final configurations .
the jump in @xmath92 at the pulse position varies with time .
it reaches its maximum when the pulse is passing through the quasi - horizon , and then decreases .
( the resulting variation in the value of @xmath92 at large @xmath45 may appear to violate causality , but is actually a just a consequence of the gauge choice implicit in our choice of coordinates . ) as the pulse continues past the quasi - horizon into the monopole core , the metric function @xmath1 remains static outside the pulse .
however , @xmath150 decreases from its maximum value as the pulse continues inward .
eventually , the pulse bounces off the central black hole ( with a small amount of its energy being absorbed ) and the process reverses itself .
the pulse passes by the quasi - horizon and retreats to infinity .
as it does so , one sees the metric variables restore themselves to almost the original values they had before the insertion of the pulse . as the initial pulse amplitude
is increased towards its threshold value , the minimum value achieved by @xmath1 approaches zero , while the maximum value of @xmath92 appears to grow without bound .
the time @xmath151 at which the pulse passes through the quasi - horizon does not vary appreciably . at spatial infinity ; @xmath151 would then diverge as the threshold amplitude was approached . ] a numerical simulation will not , of course , be able to produce a precisely extremal horizon . in our simulations , we have been able to adjust the pulse amplitude to make the minimum @xmath1 be as small as @xmath152 . in analyzing the behavior of the solutions
as the pulse amplitude is varied , we see no indication of any singularity as the threshold is approached .
we therefore expect that a pulse precisely at threshold would produce a nonsingular extremal horizon . in this case of critical collapse , the subsequent evolution of the system would be quite similar to the subcritical case , with the pulse continuing inward , bouncing off the central black hole , and then retreating outward .
however , in this case the spacetime into which it moves is causally distinct from the one where the pulse had originated ; i.e. , it is a new sector of the penrose diagram . as a final comment , in all this analysis , dropping in pressureless dust should give analogous results .
one can invoke birkhoff s law so that the metric behind the ( radially thick ) dust shell must be represented by a static metric .
but one should expect the same sorts of naked singularity behavior since this results from the interaction of metric variables .
in this paper we have used near - critical self - gravitating monopoles as tools for studying the transition from a nonsingular spacetime to one with a horizon . by analyzing the properties of trajectories that pass through the quasi - horizon and
then emerge again , we have seen that the observations made by an external observer vary continuously and show no evidence of discontinuity when the critical limit is reached .
this analysis also shows how the many causally distinct regions of the extremal black hole spacetime naturally emerge from the simple penrose diagram of the subcritical monopole .
a somewhat unexpected result from this analysis is that for the coulomb - type solutions the proper time required to traverse the interior region vanishes in the critical limit .
this is closely associated with the fact that near - critical coulomb solutions display naked black hole behavior ; these are the first examples of configurations without horizons that do so .
however , the absence of this behavior in core - type solutions shows that this is not a universal property of near - critical solutions .
our analysis also sheds light on some aspects of black holes themselves .
we have seen that the region bounded by the quasi - horizon becomes effectively inaccessible to outside observers when the solution is sufficiently close to criticality .
the interior degrees of freedom thus become unmeasurable .
tracing over them then leads a naturally defined entropy that can be attributed to this configuration .
an order of magnitude estimate of this entropy agrees with the hawking - bekenstein formula for the entropy of a black hole ; it seems plausible that in the critical limit the agreement would be precise .
these nonsingular near - critical solutions thus provide a concrete and unambiguous framework for implementing the old idea that black hole entropy might be understood in terms of the degrees of freedom hidden behind the horizon .
finally , we have argued that an extremal black hole can be produced by allowing additional matter to fall into a near - critical monopole .
we have illustrated this by numerical simulations .
starting with an initially nonsingular monopole , this leads to a zero - temperature black hole where there had previously been no horizon at all .
alternatively , one can start with a small black hole at the center of the monopole . in this latter case
, a configuration with a nonzero hawking temperature evolves into one with @xmath153 .
the existence of these possibilities gives additional clues for , and constraints on , a more precise formulation of the third law of black hole thermodynamics . | although nonsingular spacetimes and those containing black holes are qualitatively quite different , there are continuous families of configurations that connect the two . in this paper
we use self - gravitating monopole solutions as tools for investigating the transition between these two types of spacetimes .
we show how causally distinct regions emerge as the black hole limit is achieved , even though the measurements made by an external observer vary continuously .
we find that near - critical solutions have a naturally defined entropy , despite the absence of a true horizon , and that this has a clear connection with the hawking - bekenstein entropy .
we find that certain classes of near - critical solutions display naked black hole behavior , although they are not truly black holes at all . finally , we present a numerical simulation illustrating how an incident pulse of matter can induce the dynamical collapse of a monopole into an extremal black hole .
we discuss the implications of this process for the third law of black hole thermodynamics .
epsf
@=11 versim#1#2.7@ @=12 # 1#2#3#4phys . rev .
d * # 1 * , # 2 ( 19#3#4 ) # 1#2#3#4phys .
rev .
lett . *
# 1 * , # 2 ( 19#3#4 ) # 1#2#3#4phys . rev . * # 1 * , # 2 ( 19#3#4 ) # 1#2#3#4nucl .
phys . *
b#1 * , # 2 ( 19#3#4 ) # 1#2#3#4phys .
lett . * # 1b * , # 2 ( 19#3#4 ) + + + + + + plus 2pt minus 2pt |
networked storage systems have gained prominence in recent years . these include various genres , including decentralized peer - to - peer storage systems , as well as dedicated infrastructure based data - centers and storage area networks .
because of storage node failures , or user attrition in a peer - to - peer system , redundancy is essential in networked storage systems .
this redundancy can be achieved using either replication , or ( erasure ) coding techniques , or a mix of the two .
erasure codes require an object to be split into @xmath1 parts , and mapped into @xmath3 encoded fragments , such that any @xmath1 encoded fragments are adequate to reconstruct the original object .
such coding techniques play a prominent role in providing storage efficient redundancy , and are particularly effective for storing large data objects and for archival and data back - up applications ( for example , cleversafe @xcite , wuala @xcite ) .
redundancy is lost over time because of various reasons such as node failures or attrition , and mechanisms to maintain redundancy are essential .
it was observed in @xcite that while erasure codes are efficient in terms of storage overhead , maintenance of lost redundancy entail relatively huge overheads .
a naive approach to replace a single missing fragment will require that @xmath1 encoded fragments are first fetched in order to create the original object , from which the missing fragment is recreated and replenished .
this essentially means , for every lost fragment , @xmath1-fold more network traffic is incurred when applying such a naive strategy .
several engineering solutions can partly mitigate the high maintenance overheads .
one approach is to use a ` hybrid ' strategy , where a full replica of the object is additionally maintained @xcite .
this ensures that the amount of network traffic equals the amount of lost data .
a spate of recent works @xcite argue that the hybrid strategy adds storage inefficiency and system complexity .
another possibility is to apply lazy maintenance @xcite , whereby maintenance is delayed in order to amortize the maintenance of several missing fragments .
lazy strategies additionally avoid maintenance due to temporary failures .
procrastinating repairs however may lead to a situation where the system becomes vulnerable , and thus may require a much larger amount of redundancy to start with .
furthermore , the maintenance operations may lead to spikes in network resource usage @xcite .
it is worth highlighting at this juncture that erasure codes had originally been designed in order to make communication robust , such that loss of some packets over a communication channel may be tolerated .
network storage has thus benefitted from the research done in coding over communication channels by using erasure codes as black boxes that provide efficient distribution and reconstruction of the stored objects .
networked storage however involves different challenges but also opportunities not addressed by classical erasure codes .
recently , there has thus been a renewed interest @xcite in designing codes that are optimized to deal with the vagaries of networked storage , particularly focusing on the maintenance issue . in a volatile network where nodes may fail , or come online and go offline frequently
, new nodes must be provided with fragments of the stored data to compensate for the departure of nodes from the system , and replenish the level of redundancy ( in order to tolerate further faults in future ) . in this paper
, we propose a new family of codes called _ self - repairing codes _
( src ) , which are tailored to fit well typical networked storage environments . in @xcite ,
dimakis et al .
propose regenerating codes ( rgc ) by exposing the need of being able to reconstruct an erased encoded block from a smaller amount of data than would be needed to first reconstruct the whole object .
they however do not address the problem of building new codes that would solve the issue , but instead use classical erasure codes as a black box over a network which implements random linear network coding and propose leveraging the properties of network coding to improve the maintenance of the stored data .
network information flow based analysis shows the possibility to replace missing fragment using network traffic equalling the volume of lost data .
unfortunately , it is possible to achieve this optimal limit only by communicating with all the @xmath2 remaining blocks .
consequently , to the best of our knowledge , regenerating codes literature generally does not discuss how it compares with engineering solutions like lazy repair , which amortizes the repair cost by initiating repairs only when several fragments are lost .
furthermore , for rgcs to work , even sub - optimally , it is essential to communicate with at least @xmath1 other nodes to reconstruct any missing fragment .
thus , while the volume of data - transfer for maintenance is lowered , rgcs are expected to have higher protocol overheads , implementation and computational complexity .
for instance , it is noted in @xcite that a randomized linear coding based realization of rgcs takes an order of magnitude more computation time than standard erasure codes for both encoding and decoding .
the work of @xcite improves on the original rgc papers in that instead of arguing the existence of regenerating codes via deterministic network coding algorithms , they provide explicit network code constructions . in @xcite ,
the authors make the simple observation that encoding two bits into three by xoring the two information bits has the property that any two encoded bits can be used to recover the third one .
they then propose an iterative construction where , starting from small erasure codes , a bigger code , called hierarchical code ( hc ) , is built by xoring subblocks made by erasure codes or combinations of them .
thus a subset of encoded blocks is typically enough to regenerate a missing one .
however , the size of this subset can vary , from the minimal to the maximal number of encoded subblocks , determined by not only the number of lost blocks , but also the specific lost blocks .
so given some lost encoded blocks , this strategy may need an arbitrary number of other encoded blocks to repair . while motivated by the same problem as rgcs and hcs , that of efficient maintenance of lost redundancy in coding based distributed storage systems , the approach of self - repairing codes ( src ) tries to do so at a somewhat different point of the design space .
we try to minimize the number of nodes necessary to reduce the reconstruction of a missing block , which automatically translates into lower bandwidth consumption , but also lower computational complexity of maintenance , as well as the possibility for faster and parallel replenishment of lost redundancy .
we define the _ concept of self - repairing codes _ as @xmath0 codes designed to suit networked storage systems , that encode @xmath1 fragments of an object into @xmath3 encoded fragments to be stored at @xmath3 nodes , with the properties that : + ( a ) _ encoded fragments can be repaired directly from other subsets of encoded fragments without having to reconstruct first the original data_. + more precisely , based on the analogy with the error correction capability of erasure codes , which is of any @xmath4 losses independently of which losses , + ( b ) _ a fragment can be repaired from a fixed number of encoded fragments , the number depending only on how many encoded blocks are missing and independent of which specific blocks are missing .
_ to do so , srcs naturally require more redundancy than erasure codes .
we will see more precisely later on that there is a tradeoff between the repair ability and this extra redundancy .
consequently , srcs can recreate the whole object with @xmath1 fragments , though unlike for erasure codes , these are not arbitrary @xmath1 fragments , though many such @xmath1 combinations can be found ( see section [ sec : static ] for more details ) .
note that even for traditional erasure codes , the property ( a ) may coincidentally be satisfied , but in absence of a systematic mechanism this serendipity can not be leveraged . in that respect , hcs @xcite
may be viewed as a way to do so , and are thus the closest example of construction we have found in the literature , though they do not give any guarantee on the number of blocks needed to repair given the number of losses , i.e. , property ( b ) is not satisfied , and has no deterministic guarantee for achieving property ( a ) either .
we may say that in spirit , src is closest to hierarchical codes - at a very high level , src design features mitigate the drawbacks of hcs . in this work , we make the following _ contributions _ : + ( i ) we propose a new family of codes , self - repairing codes ( src ) , designed specifically as an alternative to erasure codes ( ec ) for providing redundancy in networked storage systems , which allow repair of individual encoded blocks using only few other encoded blocks .
like ecs , srcs also allow recovery of the whole object using @xmath1 encoded fragments , but unlike in ecs , these are not any arbitrary @xmath1 fragments
. however , numerous specific suitable combinations exist .
+ ( ii ) we provide a deterministic code construction called _ homomorphic self - repairing code _ ( hsrc ) , showcasing that src codes can indeed be realized .
+ ( iii ) hsrc self - repair operations are computationally efficient .
it is done by xoring encoded blocks , each of them containing information about all fragments of the object , though the encoding itself is done through polynomial evaluation , not by xoring .
+ ( iv ) we show that for equivalent static resilience , marginally more storage is needed than traditional erasure codes to achieve self - repairing property .
+ ( v ) the need of few blocks to reconstruct a lost block naturally translates to low overall bandwidth consumption for repair operations .
srcs allow for both eager as well as lazy repair strategies for equivalent overall bandwidth consumption for a wide range of practical system parameter choices .
they also outperform lazy repair with the use of traditional erasure codes for many practical parameter choices .
+ ( vi ) we show that by allowing parallel and independent repair of different encoded blocks , srcs facilitate fast replenishment of lost redundancy , allowing a much quicker system recovery from a vulnerable state than is possible with traditional codes .
since this work aims at designing specifically tailored codes for networked storage systems , we first briefly recall the mechanisms behind erasure codes design . in what follows , we denote by @xmath5 the finite field with @xmath6 elements , and by @xmath7 the finite field without the zero element . if @xmath8 , an element @xmath9 can be represented by an @xmath10-dimensional vector @xmath11 where @xmath12 , @xmath13 , coming from fixing a basis , namely @xmath14 where @xmath15 forms a @xmath16-basis of @xmath5 , and @xmath17 is a root of an irreducible monic polynomial of degree @xmath10 over @xmath16 .
the finite field @xmath16 is nothing else than the two bits 0 and 1 , with addition and multiplication modulo 2 .
a linear @xmath18 erasure code over a @xmath6-ary alphabet is formally a linear map @xmath19 which maps a @xmath1-dimensional vector @xmath20 to an @xmath3-dimensional vector @xmath21 .
the set @xmath22 of codewords @xmath21 , @xmath23 , forms the code ( or codebook ) .
the third parameter @xmath24 refers to the minimum distance of the code : @xmath25 where the hamming distance @xmath26 counts the number of positions at which the coefficients of @xmath27 and @xmath28 differ .
the minimum distance describes how many erasures can be tolerated , which is known to be at most @xmath4 , achieved by maximum distance separable ( mds ) codes .
mds codes thus allow to recover any codeword out of @xmath1 coefficients .
let @xmath29 be an object of size @xmath30 bits , that is @xmath31 , and let @xmath1 be a positive integer such that @xmath1 divides @xmath30 .
we can write @xmath32 which requires the use of a @xmath0 code over @xmath33 , that maps @xmath29 to an @xmath34-dimensional binary vector @xmath27 , or equivalently , an @xmath3-dimensional vector @xmath35 since the work of reed and solomon @xcite , it is known that linear coding can be done via polynomial evaluation . in short , take an object @xmath36 of size @xmath30 , with each @xmath37 in @xmath33 , and create the polynomial @xmath38.\ ] ] now evaluate @xmath39 in @xmath3 elements @xmath40 , to get the codeword @xmath41 [ ex : exrs23 ] suppose the object @xmath42 has 4 bits , and we want to make @xmath43 fragments : @xmath44 , @xmath45 .
we use a @xmath46 reed - solomon code over @xmath47 , to store the file in 3 nodes . recall that @xmath48 where @xmath49 .
thus we can alternatively represent each fragment as : @xmath50 , @xmath51 .
the encoding is done by first mapping the two fragments into a polynomial @xmath52 $ ] : @xmath53 and then evaluating @xmath39 into the three non - zero elements of @xmath47 , to get a codeword of length 3 : @xmath54 where @xmath55 , @xmath56 , @xmath57 , so that each node gets two bits to store : @xmath58 at node 1 , @xmath59 at node 2 , @xmath60 at node 3 .
encoding linearly data as explained in section [ sec : lincod ] can be done with arbitrary polynomials .
we now first describe a particular class of polynomials that will play a key role in the construction of homomorphic codes , a class of self - repairing codes presented in subsection [ subsec : src ] .
since we work over finite fields that contains @xmath16 , recall that all operations are done in characteristic 2 , that is , modulo 2 .
let @xmath61 , for some @xmath62 .
then we have that @xmath63 and consequently @xmath64 recall the definition of a linearized polynomial .
a _ linearized polynomial _ @xmath39 over @xmath5 , @xmath8 , has the form @xmath65 we now define a _ weakly linearized polynomial _ as a _ weakly linearized polynomial _ @xmath39 over @xmath5 , @xmath8 , has the form @xmath66 we will see below why we chose this name .
we use the notation @xmath1 since later on it will indeed correspond to the number of data symbols that can be encoded with the proposed scheme .
we start with a useful property of such polynomials .
[ lem : sr ] let @xmath61 and let @xmath39 be a weakly linearized polynomial given by @xmath67 .
we have @xmath68 note that if we evaluate @xmath39 in an element @xmath69 , we get , using ( [ eq : frob ] ) , that @xmath70 we can strengthen the above lemma by considering instead a polynomial @xmath39 over @xmath5 , @xmath8 , of the form : @xmath71 where @xmath72 , @xmath73 ( @xmath74 makes @xmath39 a linearized polynomial ) .
we now get : let @xmath61 and let @xmath39 be the polynomial given by @xmath75 , @xmath72 , @xmath76 .
we have @xmath77 if we evaluate @xmath39 in @xmath78 , we get @xmath79 again by ( [ eq : frob ] ) , and @xmath80 using the property that @xmath81 for @xmath82 .
we now mimic the way encoding works for reed - solomon codes ( see subsection [ subsec : encodpoly ] ) for weakly linearized polynomials . note that neither the encoding nor the decoding process described below are actual efficient algorithms .
implementations of these processes is a separate issue to be dealt with . 1 .
take an object @xmath29 of length @xmath30 , with @xmath1 a positive integer that divides @xmath30 .
decompose @xmath29 into @xmath1 fragments of length @xmath83 : @xmath84 2 .
take a linearized polynomial with coefficients in @xmath33 @xmath85 and encode the @xmath1 fragments as coefficients , namely take @xmath86 , @xmath87 .
3 . evaluate @xmath39 in @xmath3 non - zero values @xmath88 of @xmath33 to get a @xmath3-dimensional codeword @xmath89 and each @xmath90 is given to node @xmath91 for storage .
in particular , we need @xmath92 1 . given @xmath1 linearly independent fragments , the node that wants to reconstruct the file computes @xmath93 linear combinations of the @xmath1 fragments , which gives @xmath93 points in which @xmath94 is evaluated .
lagrange interpolation guarantees that it is enough to have @xmath95 points ( which we have since @xmath96 for @xmath97 ) to reconstruct uniquely the polynomial @xmath94 and thus the data file .
this requires @xmath98 a codeword constructed with the above procedure is of the form @xmath99 , where each coefficient is in @xmath33 and @xmath100 .
we will denote by @xmath101 the maximum value that @xmath3 can take , namely @xmath102 .
we know that @xmath33 contains a @xmath16-basis @xmath103 with @xmath83 linearly independent elements .
if @xmath104 , the @xmath105 , @xmath106 , can be expressed as @xmath16-linear combinations of the basis elements , and we have from lemma [ lem : sr ] that @xmath107 in words , that means that an encoded fragment can be obtained as a linear combination of other encoded fragments . in terms of computational complexity , this further implies that the cost of a block reconstruction is that of some xors ( one in the most favorable case , when two terms are enough to reconstruct a block , up to @xmath108 in the worst case ) . on the other hand ,
if @xmath88 are contained in @xmath109 , then the code has no self - repairing property . for any choice of a positive integer @xmath1 that divides @xmath30
, we work in the finite field @xmath33 . to do explicit computations in this finite field , it is convenient to use the generator of the multiplicative group @xmath110 , that we will denote by @xmath17 .
a generator has the property that @xmath111 , and there is no smaller positive power of @xmath17 for which this is true .
[ ex : complete]take a data file @xmath112 of @xmath113 bits , and choose @xmath114 fragments .
we have that @xmath115 , which satisfies ( [ eq : boundk ] ) , that is @xmath116 .
the file @xmath29 is cut into 3 fragments @xmath117 , @xmath118 , @xmath119 .
let @xmath17 be a generator of the multiplicative group of @xmath120 , such that @xmath121 .
the polynomial used for the encoding is @xmath122 the @xmath3-dimensional codeword is obtained by evaluating @xmath39 in @xmath3 elements of @xmath123 , @xmath124 by ( [ eq : boundn ] ) . for @xmath125 , if we evaluate @xmath39 in @xmath126 , @xmath127 , then the 4 encoded fragments @xmath128 are linearly independent and there is no self - repair possible .
now for @xmath129 , and say , @xmath130 , we get : @xmath131 note that @xmath132 suppose node 5 which stores @xmath133 goes offline .
a new comer can get @xmath133 by asking for @xmath134 and @xmath135 , since @xmath136 table [ tab : enumerate ] shows other examples of missing fragments and which pairs can reconstruct them , depending on if 1 , 2 , or 3 fragments are missing at the same time .
.ways of reconstructing missing fragment(s ) in example [ ex : complete ] [ cols="^,^ " , ] a potential schedule to download the available blocks at different nodes to recreate the missing fragments is as follows : in first time slot , @xmath137 , @xmath138 , @xmath139 , nothing , @xmath140 , @xmath141 and @xmath142 are downloaded separately by seven nodes trying to recreate each of @xmath143 respectively . in second time
slot @xmath139 , @xmath142 , @xmath141 , @xmath138 , @xmath137 , @xmath140 and @xmath144 are downloaded .
note that , besides @xmath145 , all the other missing blocks can now already be recreated . in third time slot ,
@xmath139 can be downloaded to recreate it .
thus , in this example , six out of the seven missing blocks could be recreated within the time taken to download two fragments , while the last block could be recreated in the next time round , subject to the constraints that any node could download or upload only one block in unit time .
even if a full copy of the object ( hybrid strategy @xcite ) were to be maintained in the system , with which to replenish the seven missing blocks , it would have taken seven time units . while , if no full copy was maintained , using traditional erasure codes would have taken at least nine time units .
this example demonstrates that src allows for fast reconstruction of missing blocks .
orchestration of such distributed reconstruction to fully utilize this potential in itself poses interesting algorithmic and systems research challenges which we intend to pursue as part of future work .
we propose a new family of codes , called self - repairing codes , which are designed by taking into account specifically the characteristics of distributed networked storage systems .
self - repairing codes achieve excellent properties in terms of maintenance of lost redundancy in the storage system , most importantly : ( i ) low - bandwidth consumption for repairs ( with flexible / somewhat independent choice of whether an eager or lazy repair strategy is employed ) , ( ii ) parallel and independent ( thus very fast ) replenishment of lost redundancy .
when compared to erasure codes , the self - repairing property is achieved by marginally compromising on static resilience for same storage overhead , or conversely , utilizing marginally more storage space to achieve equivalent static resilience .
this paper provides the theoretical foundations for srcs , and shows its potential benefits for distributed storage .
there are several algorithmic and systems research challenges in harnessing srcs in distributed storage systems , e.g. , design of efficient decoding algorithms , or placement of encoded fragments to leverage on network topology to carry out parallel repairs , which are part of our ongoing and future work .
99 r. bhagwan , k. tati , y. cheng , s. savage , g. voelker , `` total recall : system support for automated availability management '' , _ networked systems design and implementation ( nsdi ) _ , 2004 .
a. g. dimakis , p. brighten godfrey , m. j. wainwright , k. ramchandran , `` the benefits of network coding for peer - to - peer storage systems '' , _
workshop on network coding , theory , and applications ( netcod ) _ , 2007 .
a. g. dimakis , p. brighten godfrey , y. wu , m. o. wainwright , k. ramchandran , `` network coding for distributed storage systems '' , available online at
_ http://arxiv.org / abs/0803.0632_. a. datta , k. aberer , `` internet - scale storage systems under churn a study of the steady - state using markov models '' , _ peer - to - peer computing ( p2p ) _ , 2006 .
a. duminuco , e. biersack , `` hierarchical codes : how to make erasure codes attractive for peer - to - peer storage systems '' , _ peer - to - peer computing ( p2p ) _ , 2008 .
a. duminuco , e.w .
biersack , `` a practical study of regenerating codes for peer - to - peer backup systems '' , _ intl .
conference on distributed computing systems ( icdcs ) _ , 2009 .
x. liu , a. datta , `` redundancy maintenance and garbage collection strategies in peer - to - peer storage systems '' , _ intl .
symposium on stabilization , safety , and security of distributed systems ( sss ) _ 2009 .
d. grolimund , `` wuala - a distributed file system '' , google tech talk _
k. v. rashmi , n. b. shah , p. v. kumar and k. ramchandran , `` explicit construction of optimal exact regenerating codes for distributed storage '' , _ allerton conf . on control , computing and comm .
i. s. reed and g. solomon , `` polynomial codes over certain finite fields '' , _ journal of the society for industrial and appl . mathematics _ , no 2 , vol . 8 , siam , 1960 .
r. rodrigues and b. liskov , `` high availability in dhts : erasure coding vs. replication '' , _ workshop on peer - to - peer systems ( iptps ) _ 2005 . | erasure codes provide a storage efficient alternative to replication based redundancy in ( networked ) storage systems .
they however entail high communication overhead for maintenance , when some of the encoded fragments are lost and need to be replenished .
such overheads arise from the fundamental need to recreate ( or keep separately ) first a copy of the whole object before any individual encoded fragment can be generated and replenished .
there has been recently intense interest to explore alternatives , most prominent ones being regenerating codes ( rgc ) and hierarchical codes ( hc ) .
we propose as an alternative a new family of codes to improve the maintenance process , which we call _ self - repairing codes _ ( src ) , with the following salient features : ( a ) encoded fragments can be repaired directly from other subsets of encoded fragments without having to reconstruct first the original data , ensuring that ( b ) a fragment is repaired from a fixed number of encoded fragments , the number depending only on how many encoded blocks are missing and independent of which specific blocks are missing .
these properties allow for not only low communication overhead to recreate a missing fragment , but also independent reconstruction of different missing fragments in parallel , possibly in different parts of the network .
the fundamental difference between srcs and hcs is that different encoded fragments in hcs do not have symmetric roles ( equal importance ) .
consequently the number of fragments required to replenish a specific fragment in hcs depends on which specific fragments are missing , and not solely on how many .
likewise , object reconstruction may need different number of fragments depending on which fragments are missing .
rgcs apply network coding over @xmath0 erasure codes , and provide network information flow based limits on the minimal maintenance overheads .
rgcs need to communicate with at least @xmath1 other nodes to recreate any fragment , and the minimal overhead is achieved if only one fragment is missing , and information is downloaded from all the other @xmath2 nodes .
we analyze the _ static resilience _ of srcs with respect to traditional erasure codes , and observe that srcs incur marginally larger storage overhead in order to achieve the aforementioned properties .
the salient src properties naturally translate to _ low communication overheads _ for reconstruction of lost fragments , and allow reconstruction with lower latency by facilitating _
repairs in parallel_. these desirable properties make self - repairing codes a good and practical candidate for networked distributed storage systems .
* keywords : * coding , networked storage , self - repair |
topological defects occur in many topics of field theory , cosmology and condensed matter physics .
there is a particular interest to study models with additional nonabelian internal degrees of freedom localized on solitons ( domain walls , strings , monopoles etc . ) , see @xcite .
the next possible step is to consider theories containing solitons on solitons ( as in @xcite ) . in this paper
we study q - lumps on domain wall ( see similar construction in @xcite , full numerical solution in @xcite and generalization to higher dimension in @xcite ) .
q - lumps are q - ball type solutions of nonlinear o(3 ) @xmath0-model with an additional potential term of some special form .
they were discovered by r. leese @xcite and have nontrivial topological charge in addition to the nonzero conserved noether charges .
properties of such configurations differ significantly from those of pure @xmath0-model solution even when coupling constant is small , e.g. in case of small perturbation to the action , that may be important from physical point of view as every physical system is unlikely to contain no perturbation at all .
the main result of r. leese was that he first managed to obtain explicit solutions to o(3 ) @xmath0-model with a particular potential term and to investigate stability and scattering effects of them , later the mechanism of building of such configurations was generalized to arbitrary @xmath0-models by e. abraham @xcite .
one special feature of q - lumps is that the model admits stationary manysoliton solutions that can be interpret as non - interacting particle ( relevant to initially motionless configuration ) . in this paper
we investigate whether and what interaction appears if one adds spin - orbit interaction term in the bulk ( as in @xcite,@xcite,@xcite ) .
the main body of the paper looks at the interaction of two q - lumps on a domain wall .
2 presents the model we deal with and introduces all the objects ( the wall , q - lumps , spin - orbit interaction term ) in details .
3 investigates interaction effects perturbatively and through adiabatic approximation , finally , effects found are outlined in sect .
the theory considered includes scalar fields @xmath1 and @xmath2 in 3 + 1dimensional spacetime and admits domain wall with additional nonabelian internal degrees of freedom localized on it .
lagrangian of the theory is @xmath3 to build a domain wall let us look for solutions with separated variables @xmath4 , @xmath5 , where @xmath6 , and fields @xmath7 correspond to o(3 ) @xmath0-model .
thus the lagrangian takes form @xmath8 where @xmath9 , differentiation over @xmath10 is denoted by prime and @xmath11 motion equations derived from action ( [ larg0 ] ) are : @xmath12 the model with lagrangian ( [ lagr ] ) in three - dimensional spacetime was studied in detail by robert leese in paper @xcite : he managed to find explicit form of solitonic solutions having both topological and noether charges related respectively to homotopic group @xmath13 and to internal rotation with constant velocity around @xmath14 ( one can notice that potential term breaks o(3 ) symmetry of the model to o(2 ) ) .
these solitons were named `` q - lumps '' ( by analogy with s. coleman s `` q - balls '' @xcite ) .
they are not forbidden by derrick s theorem ( e.g. @xcite ) as they are not static and are stabilized by charge .
let s write them and their properties explicitly . for convenience
we introduce complex scalar fields @xmath15 , @xmath16 instead of @xmath7 through stereographic projection : @xmath17 , then @xmath18 topological and noether charges and total energy of configuration look like @xmath19d^2x=\frac{i}{2 \pi}\int \frac{\partial_x u^*\partial_y u-\partial_x u\partial_y u^*}{(1+u u^*)^2 } d^2x , \\ & q = \int \frac{1}{2}(s_2 \partial_t s_1 - s_1 \partial_t s_2 ) d^2x = i \int \frac{u^ * \partial_t u - u\partial_t u^*}{(1+u u^*)^2 } d^2x,\\ & \mathcal{e}_{lump}=\int \frac{\partial_t u \partial_t u^*+ \partial_i u \partial_i u^*+\alpha^2 u u^ * } { ( 1+u u^*)^2 } d^2x\end{aligned}\ ] ] and are linked by bogomolny bound @xmath20 thus configurations saturating bogomolny bound have minimal energy among solutions with given charges and are classically stable because of charge conservation
. conditions of saturations are : @xmath21 and that immediately gives explicit form of solutions : @xmath22 here function @xmath23 must be ( anti-)rational for energy to be finite . in that case degree of the function
gives topological charge @xmath24 of configuration , whereas noether charge @xmath25 takes finite value when @xmath26 .
for instance , ansatz @xmath27 corresponds to radially symmetric soliton of topological charge @xmath28 , while @xmath29 includes all solutions with top .
configurations having several distant poles could be interpret as manyparticle solutions and one can study scattering processes on moduli space in the limit of low energies , e.g. papers @xcite , @xcite considering scattering of monopoles and vortices , and in particular @xcite and @xcite about solitons of o(3 ) @xmath0-model .
term @xmath30 allows to construct static domain wall of fields @xmath31 and @xmath32 : firstly one can see that in case of @xmath33 motion equation ( [ phieq ] ) for field @xmath31 has ordinary kink solution @xmath34 , but such a configuration is unstable ( @xcite ) .
the stable one has nonzero expectation value @xmath35 inside domain wall ( i.e. around @xmath36 ) .
profiles of functions @xmath31 , @xmath32 shown in fig
. [ pic1 ] , [ pic2 ] were derived numerically in the same paper @xcite for some choice of parameters @xmath37 of lagrangian @xmath38 . ] ] fields @xmath7 like field @xmath32 become also localized on the wall around plane @xmath36 and in this way describe effectively two - dimensional ( to be more precise , 2 + 1dimensional ) theory on domain wall .
now we add spin - orbit interaction in the bulk ( for more detail about its origin see , for example , @xcite ) .
it breaks lorentz - invariance of lagrangian , leads to entanglement between fields @xmath39 and coordinates and is effectively described by term @xmath40 or , in case of separated variables , @xmath41 , @xmath42 where @xmath43 to obtain effective action for theory on the wall in presence of this term one should integrate ( [ larg0 ] ) + ( [ so ] ) over @xmath10 . then in 2 + 1 dimensions we get @xmath44 where constant values @xmath45 are introduced and in case of symmetric wall ( function @xmath32 ) @xmath46 after division by a constant @xmath47 and up to a constant term @xmath48 thereby , of the three terms in ( [ so ] ) the first one leads to the correction of @xmath49 @xmath50 the second one vanishes due to the symmetry and only the third one can significantly influence the system .
to find interaction effects let s consider two - particle solution of non - perturbated theory ( without spin - orbit interaction ) .
two - particle configurations correspond to solutions of topological charge 2 : @xmath51 here complex parameters @xmath52 and @xmath53 give size of particles @xmath54 and distance between them @xmath55 . here
we look for correction to ansatz ( [ nonper ] ) by perturbation theory assuming coupling @xmath56 to be small .
to simplify calculations let s assume parameters @xmath57 , this corresponds to poles lying on axis @xmath58 .
equation for the first correction looks like @xmath59 where @xmath60 in case of distant particles ( @xmath61 ) one can expand the equation in powers of @xmath62 , that in the leading order leads to @xmath63 and gives asymptotes of the first correction at the spatial infinity , i.e. far from lumps : @xmath64 it is necessary to notice some peculiarities of the correction found : firstly , it breaks analyticity of the solution ; secondly , the phase rotates in the opposite direction ; and finally one can see linear growth with time , that means size increase .
now we say parameters ( moduli ) @xmath65 , @xmath66 of ansatz ( [ nonper ] ) slowly depend on time .
then effective action for such a four - dimensional dynamical system after integrating over @xmath67 , @xmath58 takes the following form ( here and below we denote @xmath68 just as @xmath49 ; moreover in the very ansatz we include rotation with constant velocity @xmath68 into the definition of @xmath69 ) : @xmath70 where @xmath71 and functions @xmath72 , @xmath73 are complete elliptical integrals of the first and of the second kind respectively .
this way we obtain explicit form of motion equations : @xmath74 @xmath75 where @xmath76 equations ( [ gequation])([pequation ] ) can be solved numerically .
then @xmath77 give initial configuration , @xmath78 , @xmath79 a small perturbation .
in our case @xmath80 , @xmath81 .
equations were solved using runge - kutta 4th order method for following parameter values : couplings @xmath82 , @xmath83 , lumps size @xmath84 , distance between lumps @xmath85 .
integrals @xmath86-@xmath87 were also calculated numerically .
several effects of spin - orbit interaction on two - particle solution were discovered . 1 .
size increase : this effect is in good correspondence to the growth found with perturbation theory .
+ 2 . attraction : distance between particles declines , that was not seen from correction ( [ correction ] ) . the fact , however , is due to expression ( [ correction ] ) gives only asymptotes at spatial infinity , which remain the same .
+ 3 . dependence on phases : figures [ size ] , [ distance ] show numerical solutions for two choice of parameter @xmath88 and @xmath89 that determine initial positions of q - lumps .
the first one corresponds to particles lying on the real axis , the second one on the imaginary one .
while system behaviour remains qualitatively the same , one can notice that velocities of processes depend significantly on the parameter .
thus in contrast to functions ( [ f])-([h ] ) not depending on phases @xmath90 , @xmath91 and allowing to consider just functions @xmath92 , @xmath93 in absence of spin - orbit interaction ( in case of the same initial @xmath94 , @xmath95 ) , addition of @xmath96 results in appearance of such a dependence . in fig . [ profilepicture ] one can see how potential @xmath97 depends on phases : indeed for every fixed @xmath90 , @xmath91 profile @xmath97 qualitatively remains the same and results in growth and attraction of particles .
in this paper we considered o(3 ) @xmath0-model with an additional potential term on domain wall in 3 + 1dimensional spacetime admiting solitonic solutions ( q - lumps ) .
effects caused by spin - orbit interaction in the bulk on two - particle solutions were studied .
asymptotic behaviour at spatial infinity of first corrections to two - particle ansatz was found perturbatively ; evolution of initially motionless configuration was numerically studied through adiabatic approximation .
we discovered that in presence of spin - orbit interaction q - lumps begin to interact : they grow and attract ; also entanglement between internal degrees of freedom and coordinates leads to dependence on phases of configuration .
one important feature of q - lumps in ( 2 + 1 ) discovered by leese was that the additional term , which significantly change behaviour of the system , can be treated as just external perturbation .
the interaction effects outlined in sect .
4 suggest that another perturbation to the action in media ( in bulk ) can `` break '' some properties of leese s non - interacting lumps living on the ( 2 + 1 ) dimensional wall , and all this happens even when taking small both couplings .
the author is grateful to a. gorsky for suggesting the problem and for numerous inspirational discussions and also would like to thank k. pavlenko for valuable comments . | the nonlinear o(3 ) @xmath0model in ( 2 + 1 ) dimensions with an additional potential term admits solutions called q - lumps , having both topological and noether charges .
we consider in 3 + 1dimensional spacetime the theory with q - lumps on a domain wall in presence of spin - orbit interaction in bulk and find interaction effects for two - particle solution through perturbation theory and adiabatic approximation .
q - lumps , o(3 ) @xmath0model , domain wall , spin - orbit interaction |
the rest - frame optical / uv emission of radio quiet qsos is dominated by strong broad emission lines and by extremely blue continuum emission ( the big blue bump , which peaks in the far - uv , eg . malkan & sargent 1982 ) .
the emission of flat radio spectrum radio - loud quasars is often quite different .
it is now well established that many have much redder colours than traditional optically selected qsos ( eg .
rieke , lebofsky & wisniewski 1982 , webster et al .
a few also have extremely low equivalent width emission lines , and are known as bl lac objects . what is responsible for these differences ? in this paper , we address this question , using the spectra of a representative sub - set of flat radio spectrum sources from the parkes half - jansky flat - spectrum survey ( phfs ) .
this survey consists of 323 sources with flux densities at 2.7 ghz of greater than 0.5 jy , and radio spectral indices @xmath7 ( @xmath8 ) with @xmath9 as measured between 2.7 and 5.0 ghz ( drinkwater et al . 1997 ) .
the phfs quasars lie at redshift @xmath10 .
francis , whiting and webster ( 2000 , hereafter fww ) obtained quasi - simultaneous optical and near - ir photometry for a random subset of 157 phfs sources .
they showed that the spectral energy distributions of red quasars are diverse .
they divided the sources , on the basis of their colours , into three classes . 1 . dusty quasars .
about 10% of the sample fall into this category .
they have the strongly curved spectral energy distributions expected from dust obscuration , and show reddened line ratios .
their spectra are discussed by fww .
2 . galaxies .
another 10% of the sources are spatially resolved , and have spectra typical of elliptical galaxies .
these sources are discussed by masci , webster and francis ( 1997 ) .
3 . pseudo power - law sources . about 80% of the sources
have approximately power - law spectral energy distributions , extending from the rest - frame uv to the near ir .
these power - law continua vary in slope between @xmath11 and @xmath12 .
these are the sources we discuss in this paper .
what are these pseudo power - law sources , and what is responsible for their diversity of continuum slopes ?
three models have been proposed .
firstly , red pseudo - power - law sources might have intrinsically weaker big blue bumps ( eg .
mcdowell et al .
secondly , all quasars might have similar big blue bumps , but in the redder quasars this emission might be partially absorbed by dust ( eg .
webster et al . 1995 ) .
the dust properties would have to be unusual not to induce curvature in the spectra .
thirdly , all quasars might have similar big blue bumps , but in the redder quasars , this big blue bump emission is swamped by a red synchrotron emission component from the radio jet ( eg .
serjeant & rawlings 1997 ) . in this paper , we test these models against the _ spectra _ of the phfs pseudo power - law sources . in section [ sample ]
we discuss our spectra of the phfs quasars , and show that they represent a reasonably representative sub - sample . in section [ analysis ] we describe in detail our measurements of the spectra , and in section [ results ] we show the results .
the predictions of the three models for the emission - line properties are discussed in section [ models ] .
we show , in section [ discussion ] , that the spectra are mostly consistent with the synchrotron model , and derive constraints on this model .
finally , conclusions are drawn in section [ conclusions ] .
we obtained spectra for 77 of the phfs sources with the anglo - australian telescope ( aat ) and the siding spring 2.3 m telescope ( drinkwater et al .
these were combined with 25 unpublished spectra drawn from the aat archives , and 86 spectra drawn from the compilation of wilkes et al .
we thus have spectra of 187 of the 323 sources in the phfs . in fig [ completeness ]
, we show the number of phfs sources for which we have spectra as a function of @xmath13 magnitude .
note that the fraction with spectra is reasonably uniform .
the quality and wavelength coverage of the spectra is very diverse .
the signal to noise ratio of most quasars was too poor to allow measurements of line widths or profiles .
as most are not spectrophotometric , we did not attempt to extract continuum parameters or line ratios from the spectra . the quality and diversity of the spectra
can best be appreciated by looking at the sample spectra in drinkwater et al .
( 1997 ) .
spatially extended sources were excluded from this paper .
images from the uk schmidt and palomar sky survey images ( iiia - j ) were automatically classified , using the cosmos and apm scanners .
the automated classification was checked against our own deeper imaging .
we classified 18 of the phfs sources with spectra as spatially extended : they are discussed by masci et al .
( 1997 ) .
the remaining point - like sources were analysed on the basis of their optical / near - ir colours .
the quasi - simultaneous broad - band magnitudes of fww were used .
this excluded half of the quasars with spectra , as no photometry was available for them .
sources with @xmath14 were excluded from the analysis .
these sources have the convex continuum spectral energy distributions expected from dust absorption , and their spectra were discussed in fww . in fig [ completeness_bk ]
we show the percentage of these remaining pseudo power - law sources for which we have spectra , as a function of their @xmath15 colour .
note once again that the subsample with spectra is reasonably representative .
we defined two subsamples of these pseudo - power - law sources .
they were defined on the basis of their rest - frame colours .
we measured these colours from our broad - band photometry , and not from the ( non - spectrophotometric ) spectra .
the flux densities at 340 nm and 750 nm were calculated by extrapolation between the photometric data points from fww . * the blue sample .
22 sources with spectra .
flux density ( @xmath16 ) at rest - frame 340 nm which exceeds that at 750 nm by more than 0.45 dex ( a factor of 2.8 ) * the red sample .
20 sources with spectra
. flux density ratio of less than 0.45 dex between these wavelengths .
this dividing line corresponds to a power - law of index @xmath17 between these two wavelengths , and was chosen to bisect the sample . as a control sample , the large bright qso sample ( lbqs )
was used ( morris et al . 1991 and
refs . therein ) .
this sample of 1053 optically selected qsos is well matched in both redshift and optical luminosity to the phfs .
broad absorption line qsos were excluded from the comparison .
error arrays were not available for most of the sample spectra .
we therefore estimated them as follows . for each pixel
, we fit a straight line to the fluxes in the 15 pixels centred on it .
the sum of the squares of the residuals from this fit was divided by 13 to estimate the error variance in the pixel .
we divided by 13 rather than 15 to allow for the two degrees of freedom in the straight - line fit .
we iteratively fit a 4th order chebyshev polynomial to the continuum of each quasar .
wavelength regions with strong emission - lines were excluded from the fitting .
the fits were iterated three times , with regions more than @xmath18 above or @xmath19 below each fit excluded from the next .
the results were checked by eye : these fits were subjectively judged to be good for 95% of the quasars .
the remaining 5% were fit interactively .
emission - line equivalent widths were obtained by dividing the flux above these continuum fits by the continuum level , integrated over the wavelength regions in table [ linelimits ] .
only the strongest emission - lines were measured , as the signal - to - noise ratio was too poor to allow measurement of the weaker lines .
ly@xmath7 was not measured due to the difficulty in defining the continuum under it , and the small number of red quasars at high enough redshift .
the [ o iii ] equivalent widths are for the stronger component ( the 500.7 nm line ) only .
all the continuum fits and line equivalent widths were checked by eye , and about 10% of the automated measurements were replaced with interactive ones .
lcc line & lower rest wavelength limit & upper rest wavelength limit + + c iv & 149 nm & 159 nm + c iii ] & 183 nm & 196 nm + mg ii & 268 nm & 290 nm + h@xmath1 & 476 nm & 493 nm + @xmath20o iii@xmath21 $ ] & 498 nm & 505 nm + + composite quasar spectra were obtained as follows .
each individual spectrum was divided through by the continuum fit and shifted to its rest - frame , using interactively defined redshifts .
the spectra were then rebinned and co - added .
all spectra were equally weighted in the co - add , regardless of their signal - to - noise ratio .
this degrades the signal - to - noise ratio of the final composite spectrum , but ensures that it is not biassed towards the properties of the quasars with better quality spectra .
the fluxes at very blue and very red wavelengths were given a low weighting , to reflect their generally poor signal - to - noise ratios .
only wavelength regions to which at least five individual quasars contributed were included in the composite spectra .
three sources have no redshift in drinkwater et al .
( pks 0048@xmath22097 , pks 0829@xmath23046 and pks1156@xmath22094 ) .
falomo ( 1991 ) measured a redshift of 0.18 for pks 0829@xmath23046 , which we adopt .
both the other sources have red colours and would lie in the red sample regardless of their redshift .
we estimated upper limits on the emission - line equivalent widths of pks 0048@xmath22097 and pks 1156@xmath22094 as follows .
the biggest wavelength gap between strong emission lines is between mg ii and h@xmath1 .
the ratio of their wavelengths is 1.74 .
thus any wavelength range with an end wavelength more than 1.74 times its starting wavelength must contain at least one strong line . in both spectra , the least noisy wavelength range of this width was taken .
the @xmath24 limit for a line of width @xmath25 was computed as a function of wavelength throughout this region , and the highest value taken as our limit .
we place a @xmath24 upper limit of 0.17 nm for pks 0048@xmath22097 .
the spectrum of pks 1156@xmath22094 is poor , and we can only place a @xmath24 upper limit of 6.6 nm on the equivalent width of any emission line .
continuum luminosities were calculated at particular rest - frame wavelengths , by linear extrapolation between adjacent redshifted photometric bands from fww .
we assume that @xmath26 , @xmath27 and @xmath28 throughout .
radio luminosities were calculated in the same way , using radio flux densities at 1600 , 2468 , 4800 and 8640 mhz obtained by webster et al .
( in preparation ) .
emission - line luminosities were calculated by scaling the spectra to match the broad - band photometry of fww extrapolated to the line wavelength .
the composite spectra of blue and red phfs quasars ( excluding the two without measured redshifts ) , and of the optically selected lbqs qsos , are shown in figs [ bluecomp ] and [ redcomp ] .
a number of differences can be seen .
the red phfs composite has much weaker mg ii , balmer and [ o iii ] lines than the blue phfs composite .
the c iii ] and c iv lines , however , are more similar in strength .
the lbqs composite more closely resembles the blue phfs composite , but its ly@xmath7 , c iv , h@xmath1 and [ o iii ] lines are weaker .
the reality of these differences will be addressed in the next section .
the equivalent width measurements can be found in a machine readable table , included in the electronic edition of this paper .
equivalent widths are plotted against continuum slope in fig [ ewbk ] .
histograms of equivalent widths for the blue and red sub - samples , and for the lbqs , are shown in figs [ hist1 ] , [ mghist ] and [ hist2 ] .
the two phfs quasars without measured redshifts ( both in the red sample ) are not included .
are the differences between the red and blue sub - samples real ?
the correlation between the equivalent width of h@xmath1 and the continuum slope is significant at the 99.2% level , using spearman s rank correlation test .
a kolmogorov - smirnov test confirms that the h@xmath1 equivalent width distributions of the two sub - samples are different at the 97% confidence level .
the c iii ] equivalent width distributions are marginally inconsistent ( at the 95% level ) .
there are no other differences significant at the 95% confidence level .
the two quasars without measured redshifts were excluded from this analysis .
if we assume that they are of low enough redshift that h@xmath1 would have been seen , they increase the significance of the difference .
no correlation was found between line flux ratios and colour .
the correlation predicted by a dust model is , however , weak enough to be consistent with the observations .
we have wide wavelength coverage , spectrophotometric spectra for three of the reddest pseudo power - law sources , however , from which we can measure balmer decrements . in all three ,
the ratio of h@xmath7 to h@xmath1 is around 3.5 , consistent with no dust reddening .
the continuum slope of the phfs samples correlates with the optical continuum luminosity ( fig [ slope_lum ] ) .
spearman s rank correlation test shows that this correlation is significant with 99.999% confidence .
no significant correlation is seen between continuum slope and _ radio _ luminosity . as the phfs is a flux limited sample , redshift and luminosity
are strongly correlated .
is this really a correlation against luminosity , or could it be a correlation against redshift ? to test this , the sample was subdivided into redshift bins .
the median luminosity of the sources in each bin was calculated , and subtracted from all the sources in that bin .
the resultant differential luminosities still correlated with colour , with 95% confidence . when the process was repeated with luminosity bins , no significant residual correlation was seen .
we therefore conclude that the physical correlation is between colour and luminosity .
a stronger correlation is seen between the _ emission - line _ luminosity and the continuum slope ( fig [ slope_linelum ] ) .
the correlation with c iii ] luminosity is only significant at the 97% level , but the other correlations shown are significant with 99.99% confidence or better .
how should the strength of the big blue bump affect the emission - line properties ?
the big blue bump is normally ascribed to thermal emission from an accretion disk .
the disk could be different in red and blue quasars .
alternatively , the rest - frame uv / optical emission may be entirely synchrotron .
red and blue quasars could have different electron energy spectra in the jet .
either way , the emission - lines of red and blue quasars should be quite different , as the big blue bump emission dominates their photoionisation .
we modelled the effect of varying the big blue bump size on the emission line , using the cloudy photoionisation code ( ferland 1996 ) .
the broad emission - line region was modelled as a single plane - parallel slab of uniform hydrogen number density @xmath29 , and solar metallicity .
the flux incident on the front of the slab at 500 nm wavelength was fixed at @xmath30 ( low ionisation model ) , @xmath31 ( medium ionisation model ) and @xmath32 ( high ionisation model ) .
all these parameters are typical of published broad - line region models .
cloudy s standard agn photoionising continuum was used .
this is a power - law in the uv and optical , with exponential upper- and lower cut - offs , together with a second power - law in the x - rays : @xmath33 where @xmath34k , @xmath35 ryd and @xmath36 .
the constant @xmath37 is set such that the slope of a power - law extrapolated between 2 kev and 250 nm rest - frame , @xmath38 , is @xmath39 .
thus the hard x - ray fluxes of red and blue quasars are same , as suggested by the rosat observations of siebert et al .
( 1998 ) . all that we varied was the strength of the big blue bump , parameterised by @xmath40 .
the equivalent widths of the strongest emission lines were calculated as a function of @xmath40 , for each of the three ionisation models .
results are shown in fig [ bbb ] .
a few clear patterns emerge .
h@xmath1 is weaker in weak bump quasars .
this is expected , as it is predominantly photoionised by flux just shortward of the lyman limit , which is dominated by the big blue bump . c iii ] and c iv can be either weaker or stronger in weak bump quasars , depending on the ionisation state of the broad - line region .
this too is expected , as the strength of the big blue bump affects the ionisation parameter , and hence the population of the different ionisation states of carbon .
mg ii becomes much stronger in weak bump quasars .
mg ii is a relatively low ionisation line , and is predominantly excited by x - rays penetrating deep into the emission - line clouds .
thus an ionising continuum weak in uv photons but relatively strong in harder x - rays is ideal for mg ii formation
. a more realistic model would allow for the stratified nature of the emission - line regions .
it is also unlikely that the physical properties of the emission line region are uncorrelated with the colour of the continuum emission .
normal dust models predict a strongly curved continuum in red quasars , and hence are not consistent with the photometry of the pseudo power - law sources ( fww ) .
if the dust were deficient in small grains , however , the uv absorption would be reduced , and the predicted continuum emission could be brought into agreement with that observed .
dust should affect the emission - line ratios , but should not affect the emission - line equivalent widths , as it should absorb line and continuum radiation equally .
the exception to this would be where the dust is either patchy on scales comparable to the broad - line region , or where it is mixed in with the broad - line region .
unified models suggest that we are observing flat radio spectrum quasars from close to the axis of their jets .
any emission from these jets would thus be relativistically boosted .
if these jets sometimes emitted a synchrotron component with very red colours in the optical , it could produce the red colours .
whiting , webster & francis ( 2001 , hereafter wwf ) showed that a model of the continuum emission with two components ( a big blue bump and a red synchrotron component ) can reproduce the colours of the pseudo power - law sources as measured by fww , including the small ( but significant ) deviations from a pure power - law .
the model also correctly predicts the polarisation properties of the small fraction of phfs quasars for which this data is available .
if the synchrotron model is correct , there should be an anti - correlation between the equivalent widths of the emission lines and the redness of the continuum emission .
this is because the emission - line flux of a quasar should be proportional to the strength of the big blue bump that photoionises it , and should not be affected by the strength of any red synchrotron emission component .
the red component will have too little flux in the uv to ionise emission lines , and it is probably beamed , and hence will only illuminate a small part of the emission - line region . thus if the observed red continuum emission greatly exceeds the big blue bump , it will also greatly exceed the fluxes of the emission lines , which will hence have very low equivalent widths .
the effect should be most marked for longer wavelength lines , as the red synchrotron component will be relatively stronger at longer wavelengths .
line ratios should not correlate with colour .
at short wavelengths , there are no significant differences between the emission - line equivalent widths of the red and blue sub - samples .
h@xmath1 , however , shows a very significant difference , in the sense that most red sources have very weak lines .
the data from [ o iii ] are also consistent with this ( table [ medians ] ) .
lccc line & blue phfs quasars & red phfs quasars & ratio blue / red + + c iv & 3.3 nm & 2.7 nm & 1.2 + c iii ] & 2.2 nm & 1.1 nm & 2.0 + mg ii & 2.6 nm & 2.3 nm & 1.2 + h@xmath1 & 7.9 nm & 0.9 nm & 9.1 + @xmath20o iii@xmath21 $ ] & 4.4 nm & 0.4 nm & 12.6 + + these results are consistent with the red synchrotron component model , if this red component is about ten times stronger than the blue component at @xmath41 nm wavelength , but no longer swamps the blue component at 300 nm .
this implies a spectrum at least as red as a power - law of the form @xmath42 , which is consistent with the theoretical modelling of wwf .
these results are not consistent with the dust model , which predicts no correlation between equivalent width and colour . neither are they consistent with the weak blue bump model , which predicts that mg ii should be stronger in the red quasars , and that h@xmath1 should only be weaker by a factor of @xmath43 , not @xmath3 .
let us therefore assume that the weakness of the lines of the red sub - sample is due primarily to dilution by a red synchrotron component .
we detect the h@xmath1 line at @xmath24 confidence or better in all but one of the red quasars with decent data .
the equivalent widths span roughly two orders of magnitude at rest frame 500 nm , with no obvious bimodality . in at most 10% of the red quasars can the red continuum component exceed the blue component by a factor of more than 100 , at rest - frame 500 nm .
this too is consistent with the modelling of wwf .
if the synchrotron model is correct , all red quasars should have low equivalent width emission - lines , and all quasars with very low equivalent widths should be red .
the latter is consistent with our data , but as fig [ ewbk ] shows , there are a small number of extremely red quasars with high equivalent width mg ii lines .
the spectra of the two most extreme red , strong - lined quasars are shown in fig [ strongred ] . indeed , these red quasars have the highest equivalent width mg ii emission - lines in the sample .
these sources clearly do not fit into the simple synchrotron model .
dust could redden the continuum without changing the equivalent widths .
the balmer decrements of these sources are , however , @xmath44 , which is typical of unreddened agn .
a weak big blue bump model fits the red colours and high mg ii equivalent widths of these sources well .
we therefore hypothesise that these quasars have intrinsically weak big blue bumps .
note , however , that while the equivalent widths of these sources are anomalous , their line luminosities fit on the correlation with continuum slope ( fig [ slope_linelum ] ) .
the synchrotron model appears to best fit most of the spectra .
what then determines the relative strength of the synchrotron and big blue bump components , and hence the quasar colours ?
the continuum slope measured between 340 and 750 nm should be a good measure of the relative strength of the synchrotron and big blue bump components .
emission - line luminosities should correlate strongly with the luminosity of the big blue bump which photoionises them .
thus the correlations shown in both figs [ slope_lum ] and [ slope_linelum ] suggest that the ratio of synchrotron emission to big blue bump flux is relatively high in quasars with low luminosity big blue bumps .
this implies that in quasars with low luminosity big blue bumps , the fraction of the accretion energy going into the jet is relatively high , or that the jet emission extends to shorter wavelengths in these sources .
the @xmath45 scatter in this correlation is 0.4 dex , which is better than the scatter in the baldwin effect .
this correlation may thus be useful for cosmological distance scale work .
bl lac objects are normally defined as having no observed emission lines with rest - frame equivalent widths greater than 0.5 nm ( eg .
stickel et al .
1991 , rector et al .
2000 ) . in our model
, this implies that the red synchrotron emission component exceeds the blue one at the line wavelength by a factor of @xmath3 .
we do not see any bimodality in the distribution of equivalent factors , though our sub - sample is small .
if confirmed , this would imply that radio selected bl lac objects are simply the tail of the normal population of flat radio spectrum quasars , and not a different class with intrinsically weaker emission lines .
our analysis suggests that the synchrotron emission component is very red , and hence that it dilutes long wavelength lines far more effectively than short wavelength lines .
this implies that bl lac samples are biassed against finding high redshift objects
. how significant could this bias be ? consider a model in which the red continuum component has a spectrum of the form @xmath2 , and the logarithm of the ratio of red to blue components at the wavelength of h@xmath1 is uniformly distributed between 0 and 2 , as suggested by our observations .
the ratio of red to blue components at the wavelength of mg ii will be 3.4 times smaller than at h@xmath1 , while at c iii ] it will be 7.8 times smaller .
it is assumed that the ratio of red to blue components does not correlate with redshift or luminosity .
now consider , for example , a sample of bl lac objects chosen on the basis of spectra covering 400 600 nm .
at redshift zero , the strongest line in this wavelength range will be h@xmath1 , which in blue phfs quasars has a rest - frame equivalent width of @xmath46 nm .
for this line to have an observed equivalent width of less than 0.5 nm , we require that the red component at the wavelength of h@xmath1 exceed the blue component by a factor of ten . according to our model , 50% of all red quasars meet this criterion and would hence be classed as bl lac objects . at redshift one ,
however , mg ii will be the strongest line within this observed wavelength region . in blue phfs quasars
, mg ii has a median rest - frame equivalent width of 3 nm . for the line to have a rest - frame equivalent width of less than 0.5 nm , we require that the red continuum component exceed the blue one by a factor of 6 at the wavelength of mg ii .
this implies that the red component must exceed the blue by a factor of @xmath47 at the wavelength of h@xmath1 . according to our model
, only 35% of red quasars will meet this criterion . repeating this calculation at redshift two ( where c iv is the strongest line ) ,
only 6% of red phfs quasars would be classed as bl lac objects .
thus any sample using spectra at a uniform observed - frame wavelength will miss @xmath48% of the bl lacs at redshift one , and @xmath49% at redshift two .
our small sub - sample size makes these estimates highly uncertain .
nonetheless , this is probably sufficient to cancel out the negative evolution claimed by some authors ( eg . rector et al .
2000 ) , but not to match the enormous positive evolution seen in other classes of quasar .
it has long been known that the optical spectra of radio - loud and radio - quiet quasars are remarkably similar , despite the enormous difference in their radio luminosities . a number of small differences in the emission - line properties have been claimed ( eg .
boroson & green 1992 , francis , hooper & impey 1993 , corbin & francis 1994 , corbin 1997 ) , but only on the basis of small , poorly matched and/or incomplete samples .
these studies found that c iv , c iii ] and [ o iii ] had higher equivalent widths in flat radio spectrum quasars than in radio - quiet quasars .
the optical luminosity and redshift distribution of the lbqs are quite similar to those of the phfs quasars , so any differences between the emission - lines of the two samples must be physical . if we compare the spectra of the lbqs qsos with the blue phfs sub - sample , we find that the equivalent width distributions of c iv , h@xmath1 and [ o iii ] are significantly different , with 95% confidence .
c iii ] is different with 94.8% confidence . in all cases ,
the equivalent widths of the blue phfs quasars are greater than those of the lbqs qsos .
the red phfs quasars have significantly weaker mg ii and h@xmath1 equivalent widths than the lbqs qsos , presumably due to dilution by the red synchrotron component .
we conclude that the spectra of the red phfs quasars are significantly different from those of the blue phfs quasars .
small number statistics make our conclusions tentative , though they are formally significant .
the probable differences are consistent with a model in which the red colours are due to the addition of a featureless red synchrotron continuum component to an otherwise normal blue quasar spectrum , as proposed by wwf .
the red component must have a spectrum at least as red as a power - law of the form @xmath2 .
this red component contributes no more than half the continuum flux at rest - frame 300 nm , but at rest - frame 500 nm it contributes about 90% of the continuum flux of the red quasars .
the physics of such a component is discussed by wwf .
the relative strengths of the blue and red components span two orders of magnitude at rest - frame 500 nm .
the blue component is relatively weaker in low optical luminosity sources .
if this model is correct , then existing bl lac surveys are biassed against high redshift objects , though this bias is insufficient to bring the evolution of bl lacs into consistency with the evolution of other quasars .
a few phfs quasars , however , have both very red colours and very high equivalent width emission - lines .
these quasars do not easily fit the synchrotron model , but their properties can be fit by an intrinsically weak big blue bump model .
we confirm that the emission - lines of optically selected qsos have significantly weaker equivalent widths than those of radio - loud , flat radio spectrum quasars with similar optical luminosities .
digital copies of the composite spectra and of the measured equivalent widths are available from the on - line copy of this paper .
we wish to thank belinda wilkes for making her spectra available to us in digital format , and chriss wallwork for doing a literature search for new redshifts .
boroson , t. & green , r. 1992 , apj , 338 , 630 | we measure the spectral properties of a representative sub - sample of 187 quasars , drawn from the parkes half - jansky , flat - radio - spectrum sample ( phfs ) .
quasars with a wide range of rest - frame optical / uv continuum slopes are included in the analysis : their colours range from @xmath0 .
we present composite spectra of red and blue subsamples of the phfs quasars , and tabulate their emission - line properties . the median h@xmath1 and [ o iii ] emission - line equivalent widths of the red quasar sub - sample are a factor of ten weaker than those of the blue quasar sub - sample .
no significant differences are seen between the equivalent width distributions of the c iv , c iii ] and mg ii lines .
both the colours and the emission - line equivalent widths of the red quasars can be explained by the addition of a featureless red synchrotron continuum component to an otherwise normal blue quasar spectrum .
the red synchrotron component must have a spectrum at least as red as a power - law of the form @xmath2 .
the relative strengths of the blue and red components span two orders of magnitude at rest - frame 500 nm .
the blue component is weaker relative to the red component in low optical luminosity sources .
this suggests that the fraction of accretion energy going into optical emission from the jet is greater in low luminosity quasars .
this correlation between colour and luminosity may be of use in cosmological distance scale work .
this synchrotron model does not , however , fit @xmath3% of the quasars , which have both red colours and high equivalent width emission - lines .
we hypothesise that these red , strong - lined quasars have intrinsically weak big blue bumps . there is no discontinuity in spectral properties between the bl lac objects in our sample and the other quasars .
bl lac objects appear to be the red , low equivalent width tail of continuous distribution .
the synchrotron emission component only dominates the spectrum at longer wavelengths , so existing bl lac surveys will be biassed against high redshift objects .
this will affect measurements of bl lac evolution .
the blue phfs quasars have significantly higher equivalent width c iv , h@xmath1 and [ o iii ] emission than a matched sample of optically selected qsos .
= 2em psfig = 17.5 cm = 24.6 cm = -2.5 cm = -1.0 cm = -1.0 cm @xmath4 research school of astronomy and astrophysics , australian national university , canberra act 0200 + pfrancis,[email protected] + @xmath5 joint appointment with the department of physics , faculty of science , australian national university + @xmath6 school of physics , university of melbourne , victoria 3010 + mwhiting , m.drinkwater , [email protected] + [ 3 mm ] * keywords : * quasars : general quasars : emission lines bl lacertae objects : general |
the atomic scale structure , e.g. the pair distribution function or , equivalently , the density - density equal time correlation function , of both equilibrium and nonequilibrium fluids is directly accessible to experiment by means of light scattering and has been used to study the behavior of complex systems such as sheared colloidal suspensions@xcite .
however , the subject of nonequilibrium fluid structure remains obscure , particularly when compared to the structure of equilibrium fluids which is one of the most advanced areas of equilibrium statistical mechanics@xcite .
while kinetic theory can provide some asymptotic results , see e.g. @xcite,@xcite,@xcite , models of the small - scale structure remain phenomenological .
this is not surprising : in equilibrium , one starts with the exact n - body distribution and the problem is to integrate out uninteresting degrees of freedom so as to get the pair distribution function ( pdf ) .
away from equilibrium , even in steady states , the n - body distribution function is not known and the only starting point is either a complex dynamical equation , the liouville equation , or equivalently a coupled set of equations , the bbgky hierarchy relating the n - body to the ( n+1)-body distribution functions .
another method used to derive equilibrium integral equations is based on the fact that from the bbgky hierarchy it is easy , essentially by integrating over the velocities , to derive the ybg hierarchy which relates the pair distribution function to the triplet distribution function , the triplet to the quartic distribution function , etc . some sort of closure hypothesis , like the neglect of three body correlations , then yields a closed equation for the pdf , see e.g.@xcite,@xcite . however , away from equilibrium , the distribution of velocities is not known and this procedure can not be followed in detail .
one approach to nonequilibrium fluid structure is that of hess and coworkers@xcite,@xcite .
the object there was to describe sheared colloidal suspensions as well as computer simulations of sheared simple fluids .
the theory , based on a kind of relaxation approximation , is phenomenological and will be considered in more detail below .
fluctuating hydrodynamics offers another possibility@xcite,@xcite,@xcite but is restricted to length - scales much larger than atomic length scales .
another approach , developed about a decade later by eu et al@xcite,@xcite , was to try to generalize the equilibrium integral equations , like the kirkwood and the percus - yevick equations , to nonequilibrium fluids . for the reasons just mentioned , the only way to carry out this program was to start with an ansatz for the n - body distribution which allows the development to parallel that of the equilibrium equations . whether or not this is a good approach is difficult to assess since the approximations made are uncontrolled , but it has been previously shown@xcite , and will be argued in more detail below , that velocity correlations are the crucial determinant of atomic - scale deviations from equilibrium structure and , it would follow , that approaches which do consider them may be missing an essential contribution to the structure . indeed , the role played by velocity correlations in determining one aspect of the structure of sheared simple fluids , namely the value and angular dependence of the pair distribution function at contact has already been demonstrated@xcite and the purpose of this paper is to show that the knowledge of such velocity correlations can be used to extend equilibrium structural models into the nonequilibrium domain . in order to avoid a number of complexities associated with the spatial anisotropy of sheared fluids , the system used here to illustrate
this model is the dissipative hard sphere model used in the study of granular fluids .
this fundamentally nonequilibrium system presents a unique opportunity for the application and development of nonequilibrium statistical mechanics in a system which is both of practical relevance and yet sufficiently simple to be amenable to theoretical analysis .
the basis for such analysis has recently been formulated by brey , dufty and santos@xcite who have constructed a liouville description of the dynamics of the dissipative hard - sphere model . from this ,
boltzmann- and enskog - level kinetic equations for the one - body distribution function ( the dense - fluid generalization of the boltzmann equation ) follow and their properties have been studied theoretically , in particular , the chapman - enskog solution to it has been developed , giving explicit expressions for the pressure and transport properties of the system@xcite , and it is possible to solve it numerically@xcite . despite the simplicity of the dynamics , simulations have shown that this system exhibits a rich phenomenology which is only partially understood at present ( see , e.g. , @xcite ) . the starting point for describing the phenomenology
is the so - called homogeneous cooling state ( hcs ) : since the collisions dissipate energy , the analog of the equilibrium state is one which is spatially homogeneous but in which the temperature decreases algebraically with time .
however , this state is relatively unstable as it is subject to a number of forms of spontaneous symmetry breaking .
the simplest example of this is the so - called clustering instability in which sufficiently large systems , subject to no external forces , exhibit a spatial clustering .
this has been shown to be due to a small wave - vector hydrodynamic instability@xcite .
small systems for which the maximum wavevector ( as determined by the size of the simulation cell ) is too large , do not exhibit the transition .
other phenomena include the shearing state , in which the system remains structurally homogeneous but long - ranged momentum correlations develop , and inelastic collapse in which a few particles collide with one - another many times until all relative momentum is dissipated and the particles come to rest in contact with one another .
one motivation for trying to understand the structure of the hcs is that it is needed as input to various kinetic theory calculations which could be useful in understanding some of these features .
the remainder of this paper is organized as follows . in section
ii , the relevant elements of kinetic theory are reviewed and it is shown that the pdf at contact can in general be explicitly calculated with the same level of approximation as is used to get the enskog equation .
it is noted that both the pdf at contact , as well as the pressure , can be calculated in the hcs with no further approximations even though the exact solution for the one - body distribution is not known . in section iii , the dynamics of the dissipative hard sphere model is reformulated , through a change of variables , so that the hcs is mapped onto a steady state . although mathematically , equivalent , the steady state formulation is convenient since it avoids numerical problems associated with the rapid decay of kinetic energy in the hcs and also allows statistical properties , like the pressure , to be computed by replacing ensemble averages by time averages ( the assumption of ergodicity ) .
section iv consists of a presentation of a model for the structure of the hcs and section v presents a comparison of this model to the results of simulations .
the paper concludes with a discussion of the results and their bearing on the questions raised above .
the model granular fluid consists of a collection of @xmath0 classical hard - spheres of diameter @xmath1 having positions @xmath2 and momenta @xmath3 which experience inelastic binary collisions . between collisions ,
the atoms freely stream so the state evolves according to : @xmath4 where we allow for the possibility of an external force acting on the particles .
henceforth , we will consider only identical atoms and will use units in which the mass is equal to one .
when two particles , say @xmath5 and @xmath6 , collide , their total momentum is unaffected but their relative momentum is altered according to @xmath7 where , in general , relative quantities are denoted as @xmath8 , unit vectors as @xmath9 and where this equation serves to define the momentum transfer operator @xmath10 . the parameter @xmath11 is called the coefficient of restitution and takes on values between one ( elastic hard spheres ) and zero ( completely dissipative inelastic collisions ) . under this dynamics , the one - body distribution function , @xmath12 and the two - body distribution function @xmath13 are related by the first equation of the bbgky hierarchy@xcite @xmath14 \theta \left ( -\widehat{{\bf q}}_{12}\cdot { \bf p}% _ { 12}\right ) f_{2}(x_{1},x_{2},t ) \label{2-body}\end{aligned}\ ] ] where the notation indicates that integration of @xmath15 is over the unit sphere , @xmath16 for @xmath17 and is otherwise zero and where @xmath18 is the inverse of @xmath10 and may be easily seen to be @xmath19 finally , the two - body distribution must satisfy the identity , a kind of boundary condition@xcite , @xmath20 the origin of which - basically , the conservation of probability during a collision - is discussed in the appendix . here , i only note that this identity is independent of , and provides information additional to , the first bbgky hierarchy equation given above .
in fact , as discussed in the appendix , it can actually be derived from the second equation of the bbgky hierarchy . using an abbreviated notation , @xmath21 @xmath22
, it is easily seen that the full distribution at contact can be written as @xmath23 the combination @xmath24 appearing here , as well as in the collisional term in eq.([2-body ] ) refers to the probability for two atoms just prior to a collision and we refer to it as the pre - collisional part of the distribution .
equation ( [ bc ] ) thus expresses the two - body distribution at contact solely in terms of the pre - collisional distribution . as discussed in ref .
@xcite , the assumption of molecular chaos used to obtain the boltzmann equation and its generalization to dense hard - sphere fluids , the enskog equation , is that this pre - collisional distribution can be approximated by neglecting velocity correlations between the particles so that one writes @xmath25 where @xmath26 is the _ pre - collisional _ pair distribution function ( pdf ) at time @xmath27 which is normally taken to be the local - equilibrium form@xcite,@xcite . using this in eq .
( [ 2-body ] ) , the generalized enskog equation immediately results @xmath28 \theta \left ( -\widehat{{\bf q}}_{12}\cdot { \bf p}_{12}\right ) g_{0}({\bf q}% _ { 1},{\bf q}_{2})f_{1}(x_{1},t)f_{1}(x_{2},t ) \label{enskog}\end{aligned}\ ] ] while from the boundary condition , we get @xmath29 thus expressing the distribution at contact as the sum of an uncorrelated term , the first on the right , and a term expressing the corrections due to momentum correlations . in equilibrium , the second term vanishes . naturally , the momenta of the atoms are correlated _
after _ a collision and , in fact , ( [ final_bc ] ) allows us to determine these correlations in a manner consistent with the degree of approximation of the generalized enskog equation .
although such an evaluation might be used in a number of different ways , i will here focus one particular application of it to the problem of understanding the structure of the non - equilibrium state . specifically , integrating over momenta gives @xmath30 where the nonequilibrium density is @xmath31 .
( [ bc2 ] ) thus gives us an approximate evaluation of the contribution of momentum correlations to the structure of the fluid as expressed through the pdf .
this relation was first used in ref .
@xcite to characterize velocity correlations in a sheared fluid and has recently been used in a study of velocity correlations in granular fluids near equilibrium@xcite . the macroscopic balance equations for the density , momentum field @xmath32 , and energy density @xmath33 follow from ( [ 2-body ] ) by multiplying by @xmath34 , @xmath35 , and @xmath36 respectively and integrating over @xmath35 to get @xcite @xmath37 and explicit forms of the pressure tensor @xmath38 and heat flux vector @xmath39 are given in the literature@xcite .
the source term in the temperature equation is due to the cooling caused by the inelastic collisions .
it is easy to see that a spatially homogeneous solution to these equations is possible with @xmath40 where @xmath41 is a constant . because there is no potential energy in hard - sphere systems , the time and energy scales are set by the temperature and so ,
on dimensional grounds , it is clear that @xmath42 is independent of temperature : the temperature is therefore given by @xmath43 ^{-2}$ ] thus explicitly demonstrating the cooling .
a linear stability analysis of the equations ( [ balance ] ) expanded about the homogeneous cooling state shows that the state is unstable against small wavevector fluctuations@xcite .
the pdf at contact for hcs can also be evaluated without explicit knowledge of the one body distribution ( see the appendix ) and is @xmath44 where @xmath45 is defined by an analogous expression to this with @xmath46 replaced by @xmath26 . finally .
since it will be of use below , we quote the expression for the collisional contribution to the pressure which is @xmath47 where the last line follows from the spatial isotropy and time invariance of the hcs and also does not require explicit knowledge of the one - body distribution .
this is useful since an exact solution to the enskog equation for the one - body distribution for hcs is not known .
since the time scale is determined by the temperature , we can simplify the description of this state by making a change of variables . specifically , define a scaled time
coordinate via @xmath48 where @xmath49 and @xmath50 are arbitrary constants , and the corresponding equations of motion are @xmath51 where @xmath52 is the momentum ( velocity ) in the new coordinates . to carry through the statistical description
, we must define a new set of probability densities .
if we denote the set of positions and velocities of atoms @xmath53 in the original system by @xmath54 and of those in the scaled system by @xmath55 then the m - body distribution will transform according to @xmath56 where @xmath57 is the jacobian of the transformation for a d - dimensional system .
we then find , e.g. , that eq.([2-body ] ) becomes @xmath58 \theta \left ( -\widehat{{\bf q}}_{12}\cdot { \bf c}_{12}\right ) \widetilde{f}_{2}(x_{1}^{\prime } x_{2}^{\prime } , t)\end{aligned}\ ] ] so that the only change is that a new term appears in the streaming operator .
this term has the same form as the artificial thermostats used in the study of shear flow@xcite : however , here it arises solely from a change of variables .
the balance equations become @xmath59 where @xmath60 , @xmath61 , etc . now
, the temperature of the homogeneous state is given by @xmath62 so that any initial temperature will equilibrate to a final stable value of @xmath63 .
thus , the original homogeneous cooling state is mapped by this change of variables onto a superficially steady state .
it is worth noting that if one were to return to the original form of the dynamics and were to periodically rescale the momenta of the atoms so as to restore the temperature to its original value , then it is easy to see that as the time interval between rescalings goes to zero , the resulting equations of motion can be written in the form [ scaled_eom ] with the constant replaced by a complicated function of the momenta .
periodic rescaling is commonly used in simulations of sheared fluids and has recently been employed in the simulation of hcs@xcite,@xcite .
this mapping also makes the hydrodynamic instability apparent .
if we expand ( [ balance2 ] ) about the steady state and retain terms to second order in the gradients ( i.e. , the navier - stokes equations ) and to first order in the densities and transform to fourier space ( with wavevector @xmath64 ) , it is easy to see that the vorticity , @xmath65 , satisfies @xmath66 where @xmath67 is the shear viscosity evaluated at density @xmath41 and temperature @xmath68 .
it is obviously unstable for sufficiently small wavevectors and for sufficiently large systems , we therefore expect a spontaneous shear to develop ( note that @xmath69 so that the arbitrary constant , @xmath49 , plays no role in the stability criterion ) .
it should also be noted that there is a closely related instability in the total ( @xmath70 ) velocity in the mapped system which , in the linear stability analysis , obeys @xmath71 and so is clearly unstable for all system sizes . since we are at liberty to choose the initial conditions to be those for which @xmath72 , we will find that this is represents a minor problem in the simulations and is of no theoretical significance .
in this section , we will assume spatial homogeneity so that the pdf depends only on the scalar separation between atoms .
the simplest realistic model for the structure of the equilibrium hard sphere fluid is the percus - yevick approximation(see e.g. , refs.@xcite,@xcite ) .
this consists of the ornstein - zernike equation @xmath73 where @xmath74 and @xmath75 is the direct correlation function , together with the boundary conditions @xmath76 where the first condition is exact while the second defines the approximation .
comparison with computer simulation shows that the py approximation for the pair distribution function is quite accurate for separations greater that about two hard - sphere diameters but less accurate near contact .
the description of the small - separation structure can be significantly improved by considering the yukawa closure for the orstein - zernicke equation which replaces the boundary condition on the direct correlation function by @xmath77 and choosing the constants @xmath78 and @xmath79 to reproduce known properties : for example , taking @xmath80 and fitting the carnahan - starling equation of state as calculated by both the pressure equation and the compressibility equation .
the original mean spherical approximation , for arbitrary pair potentials @xmath81 , consists of requiring that @xmath82 where the effective hard sphere diameter is fit according to some criterion .
the py approximation is then seen to be the msa for hard spheres .
equation ( [ yc_bc ] ) may therefore be viewed as either the msa for a potential which is the sum of yukawas or as a general expansion with coefficients to be fitted in which case it is termed the generalized msa or gmsa and can be viewed as being systematic since any function could be fitted as a sum of yukawas .
note that eq.([oz ] ) defines the direct correlation function and that the first of the boundary conditions in eq.([py_bc ] ) is an exact requirement .
the only way in which this model uses the assumption that the system being modeled is in equilibrium is through the arguments that lead to the conclusion that the direct correlation function is short - ranged and hence the justification for the boundary conditions in eq.([py_bc ] ) and eq.([yc_bc ] ) .
this connection to the equilibrium state is made even weaker in a reformulation of the model due to yuste and santos@xcite , @xcite,@xcite .
they begin by noting that the laplace transform of the quantity @xmath83 , in the py approximation , is naturally written as @xmath84 with @xmath85 they go on to point out that given the second equality of eq.([gt ] ) and making a pad approximation for the function @xmath86 one can deduce the correct order of the numerator and denominator of @xmath86 as well as the py expressions for the coefficients based solely on the asymptotic properties of the pdf .
specifically , they note that a)@xmath87 at contact is given by @xmath88 @xmath89 thus fixing the relative number of terms in the numerator and denominator ; b ) @xmath90 implies that @xmath91 and c ) the fact that the static structure factor , given by @xmath92 , is finite at @xmath93 implies , given the previous condition , that for small @xmath27 , @xmath94 .
( the last condition is equivalent to assuming that long - range correlations do not exist . )
the minimal approximant satisfying these conditions is that given in eq.([ft ] ) with the py value for the coefficients .
they also note that the extension of the pad approximant to include one more term in both the numerator and denominator is exactly equivalent to the one - yukawa closure while the further extension of the approximation corresponds to a closure consisting of a sum of yukawas .
this method is also shown to give the py solution for sticky hard spheres as well as the exact structure for both ordinary and sticky hard spheres in one - dimension .
a straight forward extension of these ideas has also been used to model the square - well fluid .
thus , in this formulation , the py form of @xmath95 is taken as an ansatz characteristic of hard - core systems and the function @xmath86 modeled as a pad approximant subject to whatever knowledge exists about the structure . with this justification in mind
, we consider the application of this approach to the nonequilibrium hcs . in equilibrium ,
the next inclusion of an additional term in the numerator and denominator of @xmath96 introduces two new parameters which are used to fit a known equation of state ( normally the carnahan - starling equation of state ) through both the pressure equation and the compressibility equation .
the pressure equation for hard spheres in three dimensions reads@xcite @xmath97 and so allows to calculate the pdf at contact , recall @xmath98 , from the equation of state . in a nonequilibrium state
, the collisional boundary condition can be used , together with assumption of molecular chaos , to give the same information .
the equilibrium compressibility equation is @xmath99 and for this , there is no obvious substitute for the nonequilibrium state . with nothing to use in its place , i will continue to apply this even in the nonequilibrium state , calculating the pressure from eq.([pressure ] ) , which might be viewed as a local - equilibrium approximation .
in fact , the resulting model is relatively insensitive to the value used for the pressure since this only fixes the area of the structure function whereas the results are quite sensitive to the value of the pdf at contact . using @xmath100
as is normally done in enskog theory , the model is given by @xmath101 with @xmath102 and @xmath103 with the carnahan - starling expression@xcite @xmath104 together with eqs.([pressure ] ) and ( [ chi ] ) completes the model .
this then reduces to the gmsa in equilibrium and can be seen as its natural generalization to a spatially isotropic nonequilibrium state .
to determine the structure of the model granular system , i have performed molecular dynamics simulations of a small systems of 108 and 500 particles in 3 dimensions governed by the steady - state dynamics described by eq.([scaled_eom ] ) and subject to periodic boundary conditions .
the density was taken to be @xmath105 : high enough that finite density effects are important but low enough that the enskog approximation is expected to be valid . the choice to simulate the steady - state dynamics , rather than to simulate the
real dynamics of the cooling system , was made on the basis that the systems cool very rapidly so that the time scales involved in the simulations become very large , the velocities and energies very small and numerical inaccuracies due to round - off error are a significant problem .
this could be dealt with by periodically rescaling the velocities ( i.e. , redefining the time unit as in @xcite,@xcite ) but it is far more elegant and efficient to directly simulate the steady - state dynamics .
furthermore , the changes needed to implement this starting with a code for simulating equilibrium hard spheres are minimal .
one point that does require attention is the instability with respect to the total momentum . even if initial conditions are chosen so that the total momentum is zero at the start of the simulation , round - off errors lead to the spontaneous appearance of a non - zero total momentum which then quickly goes due to the instability .
this effect is , however , benign and is easily suppressed by calculating the total momentum during each propagation step and subtracting @xmath106 of its value from the momentum of each particle .
the starting point for the simulations was an equilibrium configuration of velocities and positions .
the value of the thermostat constant , @xmath49 , was set arbitrarily .
for each value of @xmath11 , the simulations were equilibrated for a total of @xmath107 collisions and then statistics gathered over a second series of @xmath107 collisions . to obtain steady - state averages of one - body properties ,
quantities were time - averaged over periods of @xmath108 collisions throughout the simulations ; these samples were then treated as statistically independent estimates and their average and standard error computed@xcite .
the errors in all quantities reported below are found to be small , less than 1% .
the determination of collisional effects , such as in the pressure and the pdf at contact , is somewhat different . for any collisional quantity of the form
@xmath109 the ergodic assumption gives @xmath110 where the first two lines integrate the total time dependence of the function @xmath111 .
the third line follows from a change of variable in the delta - function @xmath112(@xmath113 ) is the relative momentum of the colliding pair immediately before(after ) the collision and @xmath114 is the time at which the pair ( i , j ) collide ( which could be imaginary or outside the range of integration indicating in either case that they do not collide ) .
the last expression obviously reduces to a sum over collisions : @xmath115 where @xmath116 represents the colliding pair and which is the form used to evaluate the collisional part of the pdf at contact , @xmath117 . using @xmath118 ,
this becomes @xmath119 while for the pressure one finds @xmath120 and in both cases , the step function indicates that the expression is evaluated with the pre - collisional momenta .
finally , we present below determinations of the pdf for finite separations
. these are determined in the obvious way by looping over all pairs of atoms and creating a histogram of the separations .
the bin size used was 0.025 ( hard sphere diameters ) and these were compiled every @xmath108 collisions and the results averaged to obtain the final histogram .
figure [ fig1 ] shows the pdf at contact as determined from the simulations and from the collisional boundary condition . for @xmath121
, the agreement is seen to be good but it becomes worse values below this : furthermore , there appears to be a strong number dependence to the results with the larger system diverging more rapidly from the prediction . in both cases , simulations are only possible for @xmath11 above some threshold : below this , the simulation code fails due to the time between collisions becoming smaller and smaller until the machine precision is reached .
a detailed analysis of the sequence of collisions shows that this is due to a small number of particles with virtually no momentum relative to one another colliding over and over again - in other words , this is the phenomena of elastic collapse described by mcnamara and young@xcite .
the threshold for the collapse is in the range @xmath122 for the 500 particle system and @xmath123 for the 108 particle system . in both cases ,
prior to the collapse , the value of the pdf at contact is several times the highest values shown in figure [ fig1 ] .
similar behavior is seen in the pair distribution function . for reference ,
[ fig2 ] shows the pdf at equilibrium as determined from the 500 atom simulations and from the model : the agreement is seen to be excellent as is also the case for the data coming from the 108 atom system .
figures [ fig3 ] and [ fig4 ] show the nonequilibrium part of the pdf ( i.e. , @xmath124 ) for the 108 atom system as determined by simulation and by from the nonequilibrium gmsa for @xmath125 and @xmath126 respectively and in both cases , the extended msa is seen to give a good quantitative description of all features of the nonequilibrium structure .
( in fact , since the md results are , by their nature , binned , the theoretical curve is obtained by integrating the model pdf over bins of the same size and position as used in the simulations . )
figures [ fig5 ] and [ fig6 ] show that , not surprisingly , the agreement is less good for the 500 atom system , particularly at the smaller value of @xmath11 .
the disparity between the results for the two systems is much greater than one finds in equilibrium and suggests that a qualitative difference between them .
one obvious possible source of such a difference is the hydrodynamic instability discussed above .
using the values for the transport coefficients given by @xcite , one finds the critical size curve shown in fig .
[ fig7 ] which indicates that the 108 atom system is always stable but that the 500 atom system becomes unstable around @xmath127 however , knowledge of the transport coefficients at small @xmath11 is only approximate so these numbers may only be indicative of the position of the instability . nevertheless , the importance of the instability in the larger system is easily confirmed and fig .
[ fig8 ] , showing the velocities in one direction versus the positions along another as taken from a snapshot of the 500 atom system with @xmath128 , shows a spontaneously formed shearing profile . in a larger system
, this would manifest itself in the form of vortices .
further evidence of the instability can be found in the kinetic contributions to the pressure tensor where , beginning at @xmath125 in the 500 atom system , an oscillation develops whereby a large fraction , on the order of @xmath129 , of the kinetic energy is concentrated in first one component of the pressure tensor and then another indicating that macroscopic flows are forming .
the obvious interpretation is that around @xmath130 the shear mode is soft or unstable .
there is no evidence of such an unstable mode for any value of at @xmath11 in the 108 atom system .
this picture is thus in qualitative agreement with the predictions based on the calculations described above and has recently been observed in other studies@xcite .
we can suppress the instability in a crude way by periodically adjusting the velocities of the atoms . in these constrained simulations we interpose
correction whereby after every 100 collisions , we calculate the amplitude of the longest wavelength fourier modes of the systems ( i.e. , @xmath131 for @xmath132 , etc . ) and we then subtract the mode from each atom s velocity ( @xmath133 ) .
this is a crude procedure in that the amplitudes of the modes are only approximately set to zero and it also has the effect of removing kinetic energy from the system ( which is , however , masked by the input of kinetic energy coming from the equations of motion ) .
a more elegant procedure could be devised based on standard nonequilibrium molecular dynamics techniques such as gauss principle of least constraint@xcite , but as the present purpose is only to control the unstable mode , the crude method was deemed sufficient . as shown in fig .
[ fig9 ] , the result is to give better agreement in the measured value of the pdf at contact between the two systems while having relatively little effect in the smaller system except at the highest values of alpha .
figure [ fig10 ] shows that the pdf as determined from the constrained simulation of the larger system is in considerably better agreement with the model .
the main purpose of this paper has been to show that the gmsa can be extended to nonequilibrium systems by replacing the equilibrium input required by the gmsa with accessible nonequilibrium information coming from the collisional boundary condition and , incidently , to elucidate the atomic - scale structure of the hcs of granular fluids . in order to compare the predicted values of the pdf at contact and the pressure with the results of simulations ,
the equations of motion of the dissipative hard sphere system were mapped onto those that describe a steady state thus allowing us to use standard methods of steady state simulation such as the replacement of ensemble averages by time averages .
the comparisons with molecular dynamics simulations also show that the pair distribution function at contact can be used as a signal of the onset of elastic collapse - its value steadily diverges from the predicted value as the elastic collapse threshold is approached and its value in the simulations that feature the collapse is very large .
the significant differences observed between the 108 and 500 atom systems were seen to be largely due to the soft hydrodynamic modes present in the larger system . nevertheless , even when these are accounted for , there remains a significant deviation of the simulation results from the various predictions of the enskog theory .
it is tempting to conclude that this is due to a poor estimation of the quantity @xmath45 , for which we use the equilibrium value , but examination of the pressure shows that the answer can not be this simple : the pdf at contact would require a _ larger _
value for @xmath45 that increases with @xmath11 in order to be in agreement with the simulations whereas fig .
[ fig11 ] shows that pressure would require a _
smaller _ value that decreased with @xmath11 ( in both the unconstrained and constrained simulations ) .
we thus conclude that the deviations are due to the enskog approximation itself and could probably be described via mode coupling .
nevertheless , that @xmath45 should depend on @xmath11 is intuitively clear : atoms moving slowly away from a collision will be more likely to be knocked , by a third atom , into a second collision with one another leading to such a dependence . as mentioned in the introduction , one phenomenological approach to the description of nonequilibrium structure
is that of hess and coworkers@xcite,@xcite . in its simplest form
, this reduces to a relaxation model for the nonequilibrium contribution to the pdf : @xmath134 where @xmath135 is a relaxation time and @xmath136 is taken to be the equilibrium pdf .
it is clear that for a homogeneous steady state with no flow this gives the trivial result that @xmath137 .
however , this model is intended only as a simplification of a more complex model given by @xmath138 where @xmath139 is a kind of diffusion constant .
again , for a homogeneous steady state with no flow , only the last term survives so that this reduces to @xmath140 which has the solution @xmath141 where we have used the boundary condition @xmath142 .
although this has an interesting structure , the function @xmath143 is positive - definite so that for hcs the difference @xmath144 is also which precludes a description of the oscillatory nature of the differences found in the simulation .
another approach is the theory of eu et al@xcite,@xcite , also mentioned in the introduction .
this theory is based on an ansatz for the n - body nonequilibrium distribution and takes the form of an integral equation : @xmath145 \end{aligned}\ ] ] where @xmath146 y({\bf q}_{1},{\bf q}_{2};t ) \nonumber \\
f_{ne}\left ( { \bf q}_{1},{\bf q}_{3};t\right ) & = & \exp \left [ -v_{ne}\left ( { \bf q}_{1},{\bf q}_{2};t\right ) \right ] -1\end{aligned}\ ] ] and where the nonequilibrium potential @xmath147 is a sum of the equilibrium potential and terms related to the moments of the velocity .
this formulation has the desirable property that , if linearized in the density by replacing @xmath148 , it reduces to the py approximation in equilibrium .
it is difficult , however , to see how to incorporate the exact requirement of the information coming from the collisional boundary condition and it is therefore an open question whether this can give a model comparable to the nonequilibrium gmsa described above . indeed , in the derivation of this model , the authors explicitly neglect velocity correlations of the kind used here to control the model of the structure .
nevertheless , the solution of this model to allow for such a comparison would be of some interest .
it is also appropriate to comment on the applicability of the enskog - level description of the system .
the density chosen for the simulations is one at which the enskog description of elastic hard spheres is very good with most quantities being accurately predicted to within a few percent .
furthermore , neither the calculation of the pressure nor the pdf at contact requires explicit knowledge of the one - body distribution at contact which is good , because no exact solution to the enskog equation for hcs exists .
none the less , the results described above show that there are systematic deviations from the enskog predictions at all values of the coefficient of restitution .
evidence has also been given that these are at least partly due to the presence of soft modes within the system .
the conclusion is therefore mixed : while the enskog description seems to be qualitatively accurate , all but the very smallest systems will contain soft or unstable modes that result in significant deviations from it .
this , more than the separation of time scales discussed in ref.@xcite would seem to be the greatest obstacle to using a boltzmann / enskog description of the one body function or of using a hydrodynamic description of the macroscopic state . indeed , the enskog and hydrodynamic descriptions are successful in predicting fairly well the location of the hydrodynamic instability in the vorticity
. a better test of these issues would be to study a related system , such as a sheared granular fluid , which may be more stable .
the nonequilibrium gmsa described here works surprisingly well .
encouraging results have also been found when this model was applied to simple sheared fluids@xcite and a systematic study of this system is in progress .
it is also of interest to try to improve on the use of the compressibility equation in the nonequilibrium model : one substitute would be information coming from kinetic theory such as the asymptotic behavior of the pdf for which various approaches are possible .
bruce j. ackerson , jos van der werff and c. g. de kruif , phys .
a * 17 * , 4819 ( 1988 ) .
hansen and i.r .
mcdonald , _ theory of simple liquids _
( academic press , san diego , ca . ,
1986 ) .
t. kirkpatrick , e.d.g .
cohen and j. r. dorfman , phys .
lett . * 42 * , 862 ( 1980 ) .
j. f. lutsko , j. w. dufty and s. p. das , phys .
a * 39 * , 1311 ( 1989 ) .
m. lee , ph.d .
thesis , univ . of fla .
, 1996 .
reichl , _ a modern course in statistical mechanics _
( university of texas press , austin , texas , 1980 ) .
j.c . rainwater and s. hess ,
physica * 118a * , 371 ( 1983 ) . s. hess , journal de physique * 46 * , c3 ( 1985 ) .
d. ronis , phys .
* 52 * , 473 ( 1984 ) .
lutsko and j.w .
dufty , phys .
a * 32 * , 3040 ( 1985 ) .
t. p. c. van noije , m. h. ernst , r. brito , phys .
e * 57 * , r4891 ( 1998 ) . h.h .
gan and b.c .
eu , phys .
a * 45 * , 3670 ( 1992 ) .
gan and b.c .
eu , phys .
a * 46 * , 6344 ( 1992 ) .
lutsko , submitted to phys .
lett . .
lutsko , phys .
lett . * 77 * , 2225 ( 1996 ) .
brey , j.w .
dufty and a. santos , j. stat .
* 87 * , 1051 ( 1997 ) .
v. garzo and j. dufty , phys .
e * 59 * , 5895 ( 1999 ) .
montanero and a. santos , phys .
e * 54 * , 438 ( 1996 ) .
s. mcnamara and w. young , phys . rev .
e * 53 * , 5089 ( 1996 ) .
i. goldhirsh and g. zanetti , phys .
* 70 * , 1619 ( 1993 ) .
h. beijeren and m.h .
ernst , j. stat . phys . * 21 * , 125 ( 1979 ) .
r. evans , adv .
* 28 * , 143 ( 1979 ) .
r. soto and m. mareschal , submitted to phys .
r. soto , m. mareschal and m. mansour , to appear in phys .
j. hoye and l. blum , j. stat . phys .
* 16 * , 399 ( 1977 ) .
yuste and a. santos , phys .
a * 43 * , 5418 ( 1991 ) .
yuste and a. santos , j. stat .
72 * , 703 ( 1993 ) .
yuste and a. santos , j. chem .
* 101 * , 2355 ( 1994 ) .
j. erpenbeck , physica a * 118 * , 144 ( 1983 ) . s. mcnamara and w.r .
young , phys .
e * 50 * , r28 ( 1994 ) .
denis evans , william g. hoover , bruce h. failor , bill moran and anthony j. c. ladd , phys .
a * 28 * , 1016 ( 1983 ) .
l . tan and i. goldhirsch , phys .
* 81 * , 3022 ( 1998 ) .
it is instructive to derive the collisional boundary condition from the bbgky hierarchy , the first equation of which appears above as eq.([2-body ] ) .
the second bbgky equation is @xmath149 f^{(2)}(x_{1},x_{2};t ) \\ & = & \int dx_{3}\;\left [ \overline{t}_{-}(x_{1},x_{3})+\overline{t}% _ { -}(x_{2},x_{3})\right ] f^{(3)}(x_{1},x_{2},x_{3};t)\end{aligned}\ ] ] where the collision operator is @xmath150 \\ & = & \int d{\bf \sigma } \;\delta \left ( \sigma -\sigma _ { 0}\right ) \theta \left ( { \bf p}_{12}\cdot \widehat{{\bf \sigma } } \right ) \;\left ( { \bf p}% _ { 12}\cdot \widehat{{\bf \sigma } } \right ) \left [ \delta \left ( { \bf q}_{12}-% { \bf \sigma } \right ) \alpha ^{-2}\widehat{b}_{12}^{-1}-\delta \left ( { \bf q}% _ { 12}+{\bf \sigma } \right ) \right ] \\ & = & \delta \left ( q_{12}-\sigma _ { 0}\right ) \left ( { \bf p}_{12}\cdot { \bf q}% _ { 12}\right ) \left [ \theta \left ( { \bf p}_{12}\cdot { \bf q}_{12}\right ) \;\alpha ^{-2}\widehat{b}_{12}^{-1}+\theta \left ( -{\bf p}_{12}\cdot { \bf q}% _ { 12}\right ) \;\right ] \\ & = & \delta \left ( q_{12}-\sigma _ { 0}\right ) \left ( { \bf p}_{12}\cdot { \bf q}% _ { 12}\right ) \left [ \alpha ^{-2}\widehat{b}_{12}^{-1}+1\;\right ] \theta \left ( -{\bf p}_{12}\cdot { \bf q}_{12}\right ) \end{aligned}\ ] ] we now observe that no matter what the state , the atoms can not overlap so we must be able to write the distribution as @xmath151 so that @xmath152 \theta \left ( q_{12}-\sigma _ { 0}\right ) \widetilde{f}^{(2)}(x_{1},x_{2};t ) \\ & = & { \bf p}_{12}\cdot { \bf q}_{12}\delta \left ( q_{12}-\sigma _ { 0}\right ) \widetilde{f}^{(2)}(x_{1},x_{2};t)+\theta \left ( q_{12}-\sigma _ { 0}\right ) % \left [ { \bf p}_{1}\cdot \frac{\partial } { \partial { \bf q}_{1}}+{\bf p}% _ { 2}\cdot \frac{\partial } { \partial { \bf q}_{2}}\right ] \widetilde{f}% ^{(2)}(x_{1},x_{2};t)\end{aligned}\ ] ] so that there are two sources of singularities in the second bbgky equation : one coming from the collision operator and one from the streaming operator .
if we integrate @xmath153 over a vanishingly small interval centered at @xmath154 , only the singular terms will contribute so that we conclude they must cancel independently of the regular terms .
this gives @xmath155 \theta \left ( -{\bf p}_{12}\cdot { \bf q}_{12}\right ) \widetilde{f}^{(2)}(x_{1},x_{2};t)\ ] ] or @xmath156 which is the desired result relating the post - collisional distribution , on the left , to the pre - collisional distribution on the right .
it is thus apparent that this identity carries part of the information of the second bbgky hierarchy .
it is clear that the probability to find two atoms moving towards each other with a given relative momentum , the left hand side , is the same as that to find two atoms in contact moving away from one another with the corresponding post - collision momentum : @xmath157 which simply states that the probability to find two atoms in contact moving towards each other with a given relative momentum , the left hand side , is the same as that to find two atoms in contact moving away from one another with the corresponding pre - collision momentum ( the right hand side ) .
using @xmath158 this can be written as @xmath159 .
the factors of @xmath11 in the jacobian arise because of the change from pre- to post - collisional momenta ( @xmath160 gives one factor of @xmath11 ) and because the relation between positions @xmath161 after the collision and positions before the collision also involves the momenta giving a second factor of @xmath11 . from the identity and the assumption of molecular chaos
, we have @xmath162 so that if averages over @xmath163 are denoted by @xmath164 , then for any two - body function @xmath165 one finds @xmath166 and , in particular , if @xmath167 then @xmath168 as reported in the text . | it is shown that the equilibrium generalized mean spherical model of fluid structure may be extended to nonequilibrium states with equation of state information used in equilibrium replaced by an exact condition on the two - body distribution function .
the model is applied to the homogeneous cooling state of granular fluids and upon comparison to molecular dynamics simulations is found to provide an accurate picture of the pair distribution function . |
the compactness properties of the gauss equation on a surface and the liouville equation in @xmath21 are important tools in prescribing the gaussian curvature on a surface , a very famous problem .
building on a previous work of the authors with t. rivire @xcite , we study the compactness properties of the fractional liouville equation @xmath22 in @xmath23 under fairly weak and natural geometric assumptions .
let us first recall some classical result . if @xmath24 is a smooth , closed riemann surface with gauss curvature @xmath25 , for any @xmath26 the conformal metric @xmath27 has gaussian curvature @xmath28 determined by the gauss equation : @xmath29 where @xmath30 is the laplace - beltrami operator on @xmath24 , ( see e.g. @xcite for more details ) . in particular
when @xmath31 or @xmath32 the equation ( [ ld2 ] ) reads respectively @xmath33 and @xmath34 both equations ( [ ld2r ] ) and ( [ ld2s ] ) have been largely studied in the literature . for what concerns e.g.
the compactness properties of , in a seminal work h. brzis and f. merle @xcite showed among other things the following blow - up behavior : [ thbm ] given an open subset of @xmath35 of @xmath21 , assume that @xmath36 is a sequence of weak solutions to ( [ ld2r ] ) with @xmath37 and such that for some @xmath38 @xmath39 then up to subsequences either 1 .
@xmath8 is bounded in @xmath40 , or 2 .
there is a finite ( possibly empty ) set @xmath41 ( the blow - up set ) such that @xmath42 locally uniformly in @xmath43 , and @xmath44 here it is important to notice the requirement that @xmath45 .
particularly interesting is the case @xmath46 , in which case theorem [ thbm ] tells us that the amount of concentration of curvature @xmath47 at each blow - up point is at least @xmath48 , which is _ half _ of the total curvature of @xmath49 . on the other hand , as shown by y - y .
li and i. shafrir @xcite , if one assumes that @xmath50 in @xmath51 , then a stronger and subtle quantization result holds , namely @xmath47 is an integer multiple of @xmath52 .
this result was then extended to higher even dimension @xmath53 in the context of @xmath54-curvature and gjms - operators by several authors @xcite , always under the strong assumption that the curvatures are continuous and converge in @xmath55 ( sometimes even in @xmath56 ) , but , at least in @xcite giving up the requirement that the curvatures are non - negative .
the main ingredient here is that for uniformly continuous curvatures , in case of blow - up one can dilate the metrics and reduce to the case of constant curvature , proving that the curvature at a blow - up point is necessarily positive ( by results of @xcite ) . finally positive and constant
curvature leads to spherical metrics , thanks to various classification results ( e.g. @xcite ) , and this is in turn responsible for the constant @xmath52 ( or higher - dimensional analogs ) in the quantization results , i.e. at each blow - up point the total curvature of a sphere ( of the appropriate dimension ) concentrates .
it is natural to ask what happens if we remove _ both _ the positivity and uniform continuity assumptions on the curvature , only relying on an arbitrary @xmath4 bound .
we will address this question in dimension @xmath0 , where the analog of is @xmath57 whose geometric interpretation will play a crucial role in having a precise understanding of the blow - up behaviour .
the definitions of @xmath58 and @xmath59 will be given in the appendix .
[ trm2 ] let @xmath60 be a sequence of solutions to @xmath61 and assume that for some @xmath62 and for every @xmath7 it holds @xmath63 up to a subsequence assume that @xmath64 in @xmath65 , @xmath66 as radon measures .
then there exists a finite ( possibly empty ) set @xmath67 such that , up to extracting a further subsequence , one of the following alternatives holds : 1 .
@xmath10 in @xmath11 for @xmath68 , where @xmath69 ( compare to fig . [ fcase1 ] ) .
2 . @xmath70 locally uniformly in @xmath14 and @xmath71 for some @xmath16 , @xmath72 ( compare to fig .
[ fpinch ] ) .
let us compare the above theorem with the result of brzis and merle .
the cost to pay for allowing @xmath3 to change sign is that even in case 1 , in which @xmath8 has a non - trivial weak - limit , there can be blow - up , and in this case a _ half - quantization _ appears : the constant @xmath73 in is half of the total - curvature of @xmath74 . in case 2 , instead we are able to recover the analogue of case 2 in the brzis - merle theorem . on the other hand ,
the proof is now much more involved , as near a blow - up point regions of negative and positive curvatures can ( and in general do ) accumulate , and one needs a way to take into account the various cancelations .
a direct blow - up approach does not seem to work because there can be infinitely many scales at which non - trivial contributions of curvature appear .
in general it would be easy to prove that @xmath75 for some @xmath16 , but removing the absolute values we need to prove that there is `` more '' positive than negative curvature concentrating at each blow - up point .
this is turn will be reduced to a theorem of differential topology about the degree of closed curves in the plane , inspired by a classical work of s. j. blank @xcite , and to the blow - up analysis provided in @xcite , which will allow us to choose a suitable blow - up scale and estimate the curvature left in the other scales .
things simplify and the above theorem can be sharpened if we assume that @xmath45 , hence falling back into a statement of brzis - merle type . with @xmath76 , in the interpretation given by theorem [ geomequiv ] .
@xmath77 is the inverse of the stereographic projection . from the function
@xmath78 blowing up at @xmath79 and @xmath80 ( and possibly at infinity , in the sense that some curvature vanishes at infinity ) on construct @xmath81 blowing up at @xmath82 , @xmath83 and possibly @xmath84 , but converging to an immersion @xmath85 away from @xmath86.,width=604 ] [ trm3 ] let @xmath87 and @xmath20 be as in theorem [ trm2 ] and additionally assume that @xmath45 .
then up to a subsequence in case 1 of theorem [ trm2 ] we have @xmath88 and in case 2 we have @xmath89 for @xmath72 .
the proofs of theorems [ trm2 ] and [ trm3 ] are strongly based on the following geometric interpretation of equation .
[ geomequiv ] let @xmath90 with @xmath91 satisfy @xmath92 for some function @xmath93 .
then there exists @xmath94 with @xmath95 for every @xmath68 such that @xmath96 is a holomorphic immersion of @xmath97 into @xmath98 , @xmath99 is @xmath100 , where @xmath101 is the stereographic projection given by @xmath102 .
theorem [ geomequiv ] in turn is an extension of the following result proven in @xcite : [ geomequiv0 ] a function @xmath103 with @xmath104 satisfies @xmath105 for some function @xmath106 , if and only if there exists a closed curve @xmath107 , with @xmath108 , a holomorphic immersion @xmath109 a diffeomorphism @xmath110 , such that @xmath111 for all @xmath112 @xmath113 is @xmath114 .
given by theorem [ geomequiv ] is in general singular at @xmath84 .
because of this the curve @xmath115 can have rotation index greater than @xmath0 ( it is @xmath18 in the above example ) .
the image of the curve @xmath116 should facilitate the intuition of the geometry of @xmath96 near @xmath84.,width=529 ] indeed it was shown in @xcite that if @xmath90 , @xmath91 and @xmath78 satisfies , then @xmath117 satisfies @xmath118 then theorem [ geomequiv ] can be seen as an extension of theorem [ geomequiv0 ] to the case in which @xmath119 .
another ingredient in the proof of theorem [ trm2 ] comes from differential geometry and roughly speaking says the if a closed positively oriented curve @xmath120 of class @xmath56 except at finitely many points can be extended to a function @xmath121 which is a @xmath56-immersion except at finitely many boundary points , then the rotation index of @xmath122 is at least @xmath0 ( see definition [ defrot ] and theorem [ trmblank ] ) .
this is obvious if @xmath123 is an immersion everywhere ( no corners on the boundary ) , and in fact the rotation index of @xmath124 is @xmath0 in this case , but in the general case the rotation index can be arbitrarily high and the proof that it must be strictly positive rests on ideas introduced by blank to study which regular closed curves can be extended to an immersion of the disk into the plane . mixing this result ( rotation index at least @xmath0 ) with theorem [ geomequiv ] ( via proposition [ proppull2 ] and corollary [ corangle ] ) one obtains at once : [ thmpi ] let @xmath90 with @xmath91 satisfy for some function @xmath93 and let @xmath125 then @xmath126 . if @xmath127 then @xmath128 . when @xmath127 , the proof of theorem [ thmpi ] is fairly easy , since
@xmath78 can be written as @xmath129 for a constant @xmath130 ( compare to proposition 5.5 in @xcite ) , and for @xmath127 one easily obtain an asymptotic behavior of the form @xmath131 for some contant @xmath130 and this contradicts the assumption @xmath132 if @xmath133 .
more delicate is the case when @xmath28 is allowed to change sign , because of possible cancelations . in any case ,
more than result itself , we would like to emphasize that it has a geometric interpretation in terms of the rotation index of the curve @xmath115 given by theorem [ geomequiv ] ( see fig .
[ findex2 ] ) . whether an analogue interpretation can be used for the liouville equation in dimension @xmath18 or higher
is unknown .
another consequence of theorem [ geomequiv ] is a new and geometric proof , not relying on a pohozaev - type identity , nor on the moving plane technique ( moving point in this case ) , of the classification of the solutions to the non - local equation @xmath134 under the integrability condition @xmath135 [ trm1 ] every function @xmath136 solving - is of the form @xmath137 for some @xmath138 and @xmath139 .
the proof will be given at the end of section [ sec:2 ] .
similar higher - dimensional , also in the fractional case have appeared in @xcite .
[ [ acknowledgements ] ] acknowledgements + + + + + + + + + + + + + + + + we would like to thank carlos tomei and tristan rivire for referring us to the works of blank @xcite and ponaru @xcite , and to the monograph @xcite .
it is easy to verify that for @xmath140 we have @xmath141 where @xmath142 is the hilbert transform on @xmath74 defined by @xmath143 the following lemma is well - known .
[ lemmahilbert ] the hilbert transform @xmath144 is bounded from @xmath145 into itself , for @xmath146 , and it is of weak type @xmath147 . a function @xmath148 with @xmath149
can be extended to a holomorphic function in @xmath150 if and only if @xmath151 for some @xmath152 .
consider the conformal map @xmath153 given by @xmath154 we will use on the domain @xmath150 the coordinate @xmath155 and on the target @xmath21 the coordinates @xmath156 or @xmath157 .
notice that @xmath158 hence @xmath159 is the stereographic projection .
its inverse is @xmath160 if we write @xmath161 we get the following useful relation @xmath162 which follows easily from @xmath163 .
the following proposition is proven in ( * ? ? ?
[ proppull2 ] given @xmath164 set @xmath165 as @xmath166 or equivalently and using @xmath167 then @xmath168 if and only if @xmath169 , and @xmath170 if and
only if @xmath171 . in this case @xmath78 solves if and only if @xmath165 solves @xmath172 * proof of theorem [ geomequiv ] * given @xmath173 as in the statement of the theorem , set @xmath169 as in , so that by proposition [ proppull2 ] we have @xmath174 with @xmath175 by lemma [ lemmafund4 ] the function @xmath176 satisfies @xmath177 hence by regularity theory we have @xmath178 , and @xmath179 for every @xmath68 . then @xmath180 and by lemma [ lemmahilbert ] one can define the holomorphic extension @xmath181 and @xmath182 , and @xmath183 where @xmath184 is the oriented segment joining @xmath185 to @xmath186 , or any other piecewise @xmath56 path in @xmath187 joining @xmath185 to @xmath186 . that @xmath96 can be continuously extended to the point @xmath84 depends on the following facts .
we have @xmath188 since @xmath189 .
we also have @xmath190 since @xmath191 is obtained by convolving @xmath192 with the poisson kernel @xmath193 , see .
in particular @xmath194 similarly , using that @xmath195 we obtain @xmath196 so that @xmath197 using these facts one can easily prove that for any sequence @xmath198 , the sequence @xmath199 is a cauchy sequence , since one can join @xmath200 to @xmath201 by paths @xmath202 made of arcs as the one described above ( pieces of radii and arcs of circles ) , in such a way that @xmath203 and one can extend @xmath96 continuously to @xmath84 .
for instance , consider the sequence @xmath204 . for
any @xmath205 one case choose the path @xmath202 made by joining the two arcs @xmath206 and the segment @xmath207 where @xmath208 slowly enough to have @xmath209 ( one can not simply join @xmath200 to @xmath201 with the segment @xmath210 because at this stage we can not use for @xmath211 . ) we now want to prove that @xmath114 is the curvature @xmath212 of the curve @xmath213 .
we compute for @xmath214 @xmath215,\\ \end{split}\ ] ] and , using a well - known formula for the curvature of a curve in the plane , @xmath216 where we used that the vectors @xmath217 and @xmath218 are orthogonal , and .
finally , follows at once from and : @xmath219 for @xmath220 .
@xmath221 * proof of theorem [ trm1 ] * given @xmath78 solving according to proposition [ proppull2 ] the function @xmath165 defined by satisfies @xmath222 with @xmath223 by theorem [ geomequiv ] the function @xmath165 determines a holomorphic immersion @xmath224 with the property that @xmath225 is a curve of curvature @xmath0 which extends continuously to a closed curve @xmath115 . then up to translations @xmath115 is a parametrization of the unit circle , possibly with degree @xmath226 different from @xmath0 . but
with lemma [ lemmabla ] , together with the fact that @xmath96 is holomorphic and @xmath227 never vanishes in @xmath150 , we immediately get that @xmath228 , @xmath229 and @xmath96 is a mbius diffeomorphism of @xmath230 . from that
, the explicit form of @xmath165 and @xmath78 can be computed as in the proof of theorem 1.8 in @xcite .
in order to prove theorems [ trm2 ] and [ trm3 ] we pull back the liouville equation from @xmath23 into @xmath74 using theorem [ geomequiv ] , and then prove the following theorem about a singular liouville equation on @xmath74 , whose statement can be easily pulled back onto @xmath23 .
we remark that in dimension @xmath18 singular liouville equations arise in the study of chern - simons vortices , see e.g. @xcite .
[ mainth ] let @xmath231 be a sequence with @xmath232 satisfying @xmath233 where @xmath234 is a bounded sequence and @xmath235 satisfies @xmath236 assume w.l.o.g . that @xmath237 and @xmath238 in @xmath239
. then , up to a subsequence , @xmath240 as radon measures on @xmath74 , and @xmath241 in @xmath242 where @xmath243 is a ( possibly empty ) subset of @xmath244 . setting @xmath245 @xmath246 one of the following alternatives holds : 1 .
@xmath247 and @xmath248 , @xmath249 with @xmath250 , @xmath251 and @xmath252 2 . @xmath253 as @xmath254 , @xmath255 , @xmath256 where @xmath257 solves @xmath258 moreover @xmath259 for @xmath72
. if @xmath45 for every @xmath7 , then in case 1 we have @xmath88 and in case @xmath18 we have @xmath89 for @xmath72 . more precisely , if @xmath45 is a neighbourhood of @xmath260 for some @xmath17 and for every @xmath7 , then we are in case @xmath18 and @xmath89 . the terms @xmath261 and @xmath262 multiplying @xmath263 in and have different origins : @xmath262 is the limit of the angles @xmath264 , while @xmath261 is the amount of curvature concentrating at @xmath84 .
both coefficients can be @xmath185 of course .
the following theorem is a preliminary step towards the proof of theorem [ mainth ] .
[ convergencephi ] let @xmath265 be a sequence as in theorem [ mainth ] , and let @xmath266 be holomorphic immersions with @xmath267 and @xmath268 , as given by theorem [ geomequiv ] ) .
then , up to extracting a subsequence , the following set is finite @xmath269 and for functions @xmath270 and @xmath271 we have for @xmath272 @xmath273 and @xmath274 weakly in @xmath275 , strongly in @xmath276 , and in @xmath277 .
moreover , one of the following alternatives holds : 1 .
the sequence @xmath278 is bounded , @xmath279 , and @xmath85 is a holomorphic immersion of @xmath280 .
moreover @xmath281 for some @xmath282 , where @xmath283 .
2 . @xmath284 locally uniformly in @xmath285 as @xmath286 and @xmath287 for some constant @xmath288
. moreover @xmath289 for some @xmath290 .
the proof of theorem [ convergencephi ] is essentially identical to the proof of theorem 3.5 is @xcite , with ( * ? ? ?
* lemmata 3.3 , 3.4 ) replaced by lemmata [ elemma ] and [ l2infty ] below , which take into account the presence of the term @xmath291 in the equation .
the convergence of @xmath292 uniformly in @xmath230 is not proven in @xcite , but follows easily from the convergence in @xmath293 and lemma [ l2infty ] below .
indeed shows that the oscillation of @xmath294 in a neighborhood of the blow - up points must be uniformly bounded , so that one can apply the theorem of ascoli - arzel to the sequence @xmath295 .
the proof of and is standard , but we sketch it . by the bounds @xmath296 we have ( up to a subsequence ) @xmath297 as radon measures , where @xmath298 is supported on @xmath299 and is therefore a linear combination of dirac s deltas , supported on @xmath300 .
[ elemma ] let @xmath169 be a solution of @xmath301 with @xmath302 , @xmath303 , @xmath304 .
assume that for some arc @xmath305 @xmath306 for some @xmath307 .
then for every arc @xmath308 with @xmath309 @xmath310 set @xmath311 and define @xmath312 where @xmath313 is as in lemma [ lemmafund4 ] . from and it follows that @xmath314 choose now an arc @xmath315 with @xmath316 and @xmath317 . with we easily bound @xmath318 it follows from and theorem [ mt4 ] that @xmath319 for some @xmath320 , and consequently also @xmath321 .
then for @xmath322 we have @xmath323 where in @xmath324 we use that @xmath325 for @xmath326 and in @xmath327 we use that @xmath328 finally notice that for @xmath329 the term @xmath330 in is bounded depending only on @xmath331 .
[ l2infty ] let @xmath332 satisfy @xmath333 for some @xmath334 , and let @xmath335 be the harmonic extension of @xmath165 to @xmath150
. then @xmath336 and for any ball @xmath337
@xmath338 set @xmath339 , so that @xmath340 we can then write @xmath341 where @xmath342 and @xmath343 are the harmonic extentions to @xmath230 of @xmath344 and @xmath345 respectively . then using from the appendix we write @xmath346 where @xmath313 is the green function associated to the neumann problem .
it is know that @xmath347 ( see e.g. @xcite ) .
therefore @xmath348 and since we also have @xmath349 ( [ stimalo1 ] ) follows . the proof of follows from oneil s inequality @xcite @xmath350 for any @xmath351 .
let @xmath352 be a diffeomorphism fixing @xmath84 such that @xmath353 is in normal parametrization , i.e. @xmath354 is constant for every @xmath7 . since the lengths @xmath355 of the curves @xmath356 satisfy @xmath357 up to adding ( upper and lower bounded ) constants to @xmath358
we can assume that @xmath359 and @xmath360 where @xmath361 this will not change the statement of theorems [ mainth ] and [ convergencephi ] .
also define an angle function @xmath362 for @xmath356 namely a continuous function @xmath363 such that @xmath364 by the boundedness of the curvatures the following limits exist : @xmath365 up to a translation we can assume that @xmath366 , so that by arzel - ascoli s theorem a subsequence converges to a @xmath367-curve with a possible angle at @xmath84 . also for such curve
an angle function @xmath368 can be defined as above . up to
adding multiples of @xmath48 , or requiring that @xmath369 for every @xmath7 , we also have @xmath370 following @xcite we introduce the following distance function @xmath371 .
@xmath372,\bar d^2 ) , ~\delta(0)=\sigma_k(p ) , ~\delta(1)=\sigma_k(q)\bigg\},\end{aligned}\ ] ] the infimum in ( [ dk ] ) is attained by a path @xmath116 such that @xmath373 so that @xmath374 in particular , when bounding @xmath375 from above , it is sufficient to bound the _ length _ of curves @xmath376 . in the sequel we will often construct such curves by patching together different curves and arcs , and
it is more convenient to estimate the length than the corresponding @xmath377-norms .
[ pinched ] a point @xmath378 is called _ pinched point _ for the sequence @xmath379 if there exists @xmath380 , @xmath381 such that @xmath382 .
we call @xmath383 conjugate " or `` pinched '' to @xmath384 .
we denote by @xmath385 the sets of the pinched points of @xmath386 the definition of pinched point is independent of @xmath387 and @xmath388 , as can be easily verified . moreover being pinched is an equivalence relation .
contrary to the case studied in @xcite ( @xmath389 for every @xmath7 ) in which every pinched point has a unique conjugate ( ( * ? ? ?
* lemma 3.12 ) ) , in our case a pinched point can have several conjugates ( see e.g. fig .
[ fpinch ] ) .
[ pinchdist ] let @xmath390 pinched to @xmath391
. then the arc @xmath392 joining @xmath393 and @xmath394 has length @xmath395 .
consider the curve @xmath396 .
then @xmath356 has bounded curvature and length independent of @xmath7 , equal to the length of the arc joining @xmath393 to @xmath394 ( because we assumed that @xmath359 , so that @xmath397 ) . moreover , since @xmath393 and @xmath394 are pinched , @xmath398 we can now easily see that the length of @xmath356 is at least @xmath399 by considering the direction of the tangent vector @xmath400 . as @xmath254 ,
the arc covered by @xmath401 will have length at least @xmath402 ( compare for instance the proofs of lemmata 3.10 and 3.11 in @xcite ) , which completes our proof .
contrary to the case @xmath389 studied in @xcite , two pinched points can be arbitrarily close to @xmath84 ( see figure [ fpinchclose ] ) .
.,width=529 ] the set @xmath385 of pinched points is closed in @xmath403 .
consider a sequence of points @xmath404 and for each of them let @xmath405 be a conjugate .
up to a subsequence we can assume that @xmath406 and by lemma [ pinchdist ] we have @xmath407 .
then since @xmath408 it is easy , with a diagonal argument , to construct a sequence of paths @xmath409 joining to @xmath410 to @xmath411 made of paths @xmath412 joining @xmath413 to @xmath414 and small arcs joining @xmath410 to @xmath413 and @xmath411 to @xmath414 , hence proving that @xmath415 , so that @xmath416 . [ lemmapinch ] assume that we are in case @xmath18 or theorem [ convergencephi ] . then there are points @xmath417 such that @xmath418 ( with the convention @xmath419 ) . the points @xmath420 are pinched to each other and distinct , with the possible exception of @xmath421 and @xmath422 which might coincide , in which case @xmath423 ( compare to fig .
[ fpinch ] ) .
if this is the case and @xmath424 , in general it is not true that @xmath84 is a pinched point .
finally @xmath425 in this formula , if @xmath426 , we replace it by @xmath427 and if @xmath428 we replace it by @xmath429 , where the definitions of @xmath430 are given by .
we have @xmath431 in @xmath432 , which implies ( considering that @xmath433 is arc - length parametrized ) that @xmath434 hence @xmath435 for points @xmath436 . by and
@xmath437 one has @xmath438 since @xmath439 , and @xmath440 , we have that @xmath441 for @xmath442 , except possibly the case @xmath443 , @xmath444 . to prove that the @xmath420 are pinched , fix @xmath445 with @xmath446 ( unless @xmath424 and @xmath447 , in which case we can not prove that @xmath421 and @xmath448 are necessarily pinched ) , choose points @xmath449 , @xmath450 .
let @xmath116 a path joining @xmath451 to @xmath452 ( for instance a straigth line ) .
since @xmath453 , we have @xmath454 and similarly for @xmath455 . joining these two arcs with @xmath116 with obtain a path @xmath409 with @xmath456 , @xmath457 and length going to zero , so that @xmath458 and @xmath420 is pinched to @xmath459 . for the last statement
set @xmath460 for @xmath461 sufficiently small , so that @xmath462 and @xmath463 .
we have @xmath464 and @xmath465 .
then , recalling , we compute @xmath466 where in the last identity we used .
we now introduce the following equivalence relation on the set @xmath467 , which can geometrically understood by looking at figure [ fpinchequiv ] .
[ equiv ] given @xmath468 we say that @xmath469 if there exists a sequence of paths @xmath470\to \bar d^2 $ ] with @xmath471 such that @xmath472}{p\in{\mathcal}p}}d_k(\delta_k(t),\sigma_k(p)),\ ] ] where @xmath473 is the distance defined as @xmath474,\bar d^2 ) , ~\delta(0)=z , ~\delta(1)=w\bigg\}. \end{split}\ ] ] , but @xmath475 because any path @xmath476 joining @xmath477 and @xmath478 passes close to the pinching at @xmath479 and @xmath480 . here
@xmath481 has @xmath482 connected components and @xmath483 equivalence classes.,width=415 ] [ eqcl ] let @xmath484 , @xmath485 and @xmath486 be respectively the equivalence class and the connected component containing @xmath487 then @xmath488
. see proposition 3.15 in @xcite . with @xmath489 . in the limit
@xmath490 converges to an immersion @xmath491 of @xmath492 into the region bounded by @xmath493 . here
the equivalence class @xmath494 has the @xmath483 connected components @xmath495 , while @xmath496 are the other equivalence classes of @xmath497 . in this example
the pinched set is finite , but in general it can be countable , or even have positive measure ( see figure [ fpinchequiv ] ) .
this figure also exemplifies lemma [ lemmanest ] : the connected component @xmath498 of the equivalence class @xmath499 is contained in the arc joining @xmath500 and @xmath500 .
therefore the whole equivalence class @xmath499 is contained in the same arc between @xmath500 and @xmath501.,width=604 ] [ eqclbis ] let @xmath502 be an equivalence class in @xmath503 . then there exists a sequence @xmath504 of mbius transformations such that @xmath505 in @xmath506 ) , @xmath507 , and letting @xmath508 be such that @xmath509 , one has @xmath510 in @xmath511 , @xmath512 and @xmath513 .
in fact @xmath514=(\tilde \phi_\infty)_*[s^1\setminus \tilde b]$ ] .
in particular each equivalence class @xmath494 is a finite union of open arcs @xmath515 ( ordered in the counter - clockwise direction ) , with @xmath516 pinched to @xmath517 and @xmath518 pinched to @xmath519 .
see figure [ fmobius ] .
this follows with the same proof of proposition 3.16 in @xcite .
[ lemmanest ] the equivalence classes of @xmath497 are nested , in the sense that if an @xmath498 is an arc of the equivalence class @xmath499 included in an arc @xmath520 joining two arcs @xmath500 and @xmath501 of the equivalence class @xmath494 , then @xmath521 , see e.g. figure [ fmobius ] .
this follows easily from the definition of equivalence classes , since assuming that @xmath499 contains an arc @xmath522 , any path @xmath409 connecting @xmath523 to @xmath524 would cross any path @xmath525 connecting @xmath526 to @xmath527 . by assumptions one can choose such paths so that is satisfied . mixing these paths , one can easily find new paths joining @xmath523 to @xmath526 and still satisfying , hence proving that @xmath528 .
[ lemmaleqpi ] in case 1 of theorem [ convergencephi ] we have @xmath529 for @xmath72 .
assume that we are in case 1 of theorem [ convergencephi ] .
then up to a subsequence @xmath530 in @xmath531 and @xmath532 for some @xmath533 , @xmath534 . choose a blow - up point @xmath535 .
we want to prove that @xmath529 .
can be extended to an immersion @xmath536 only singular at @xmath79 , one has @xmath537,width=566 ] let @xmath538 , where @xmath539 is fixed but so small that @xmath540 .
let now @xmath541 be a given regular @xmath542-jordan curve with @xmath543 , positively oriented .
assume also that @xmath544 for @xmath545 ( in other words @xmath116 extends the arc @xmath546 to a regular jordan curve ) .
consider now the closed curves @xmath547 .
since @xmath116 can be extended to an immersion @xmath548 of @xmath230 into @xmath230 , then @xmath549 can be extended to an immersion @xmath550 of @xmath230 into @xmath98 , with rotation index @xmath551 , see definition [ defrot ] .
in particular , since @xmath549 does not have corners , it follows easily that @xmath552 where @xmath553 denotes the curvature of @xmath549 . on the other hand , our definitions and imply that @xmath554 since away from @xmath79 the convergence of @xmath549 to @xmath555 is regular enough , we obtain @xmath556 i.e. @xmath557 on the other hand , since @xmath558 can be extended to an immersion @xmath559 with @xmath560 in @xmath561 , by theorem [ trmblank ] we infer that @xmath537 . looking at the definition of @xmath562 ( see definition [ defrot ] with @xmath424 , @xmath563 ) , we obtain that for a function @xmath564 such that @xmath565 ( here @xmath566 is the derivative with respect to the arc - length on @xmath74 , taken with the usual counter - clockwise orientation ) , then @xmath567 on the other hand @xmath568 where @xmath569 $ ] is the exterior angle formed between @xmath570 and @xmath571 .
then we conclude that @xmath572 [ lemmagepi ] assume that we are in case 2 of proposition [ convergencephi ] .
then @xmath16 for @xmath573 .
moreover if for some @xmath72 and some @xmath574 we have @xmath575 then @xmath89 .
according to lemma [ lemmapinch ] we have to prove that @xmath576 let us first assume that for some @xmath461 we have @xmath577 ( that this is not always the case can be seen e.g. in figure [ fpinchcomp ] , where two pinched points @xmath448 and @xmath578 belong to arcs of pinched points of positive measure . ) since the equivalence classes in @xmath579 are nested by lemma [ lemmanest ] , we have that @xmath580 for an equivalence class @xmath494 of points in @xmath497 ( compare to figure [ fpinchcomp ] , without the arcs @xmath581 and @xmath582 , but it could also be that @xmath494 covers the whole @xmath583 ) .
let @xmath584 and @xmath585 be as in proposition [ eqclbis ] , corresponding to this equivalence class @xmath494 . then @xmath586 in @xmath587 , and @xmath588 where @xmath589 and @xmath590 .
up to relabelling we can assume that @xmath591 is the connected component component of @xmath494 starting at @xmath592 and @xmath593 is the one ending at @xmath594 ( see figure [ fmobius ] ) .
then , by lemma [ lemmapinch ] applied to @xmath595 , we have ( by the definition of @xmath368 ) @xmath596 which is to say that @xmath595 is the amount of curvature , including the singularity @xmath597 at @xmath84 , accumulated along the curve from @xmath598 to @xmath599 , which is @xmath48 minus the curvature on accumulated along the curve from @xmath599 to @xmath598 . according to lemma [ lemmaleqpi ] , we have @xmath600 , hence @xmath601 can also contain arcs .
its complement contains the equivalent class @xmath602 nested between @xmath603 and @xmath604.,width=604 ] this completes the proof , under the assumption that @xmath605 and @xmath606 are not pinched for @xmath607 small . if they are pinched for _ every _
@xmath608 for a small @xmath461 , one has @xmath609 for @xmath608 for a maximal @xmath461 ( see e.g. figure [ fpinchcomp ] ) . in this case
@xmath610 and these contributions to the value of @xmath611 can be ignored .
then we can reduce to the case in which a right neighbourhood of @xmath592 and a left neighbourhood of @xmath594 contain a sequence of connected components of @xmath497 .
since the equivalence classes of @xmath497 are nested , we can find arcs @xmath612 arbitrarily close to @xmath592 and @xmath613 arbitrarily close to @xmath594 , say @xmath614 , @xmath615 , both belonging to an equivalence class @xmath616 . applying the previous reasoning ( proposition [ eqclbis ] in particular )
to @xmath494 one finds again @xmath617 , but because of the curvature lost in the arcs joining @xmath592 to @xmath612 and @xmath594 to @xmath613 , which is at most @xmath618 , one has @xmath619 and since @xmath574 is arbitrarily small we conclude again that @xmath16 . finally , assume .
then , up to possibly taking @xmath574 smaller , so that @xmath620 we have @xmath621 so that the curves @xmath622 have the following properties : 1 .
they have non - negative curvatures uniformly bounded by @xmath623 and lengths uniformly bounded by @xmath624 ; 2 .
they are almost closed , in the sense that , writing @xmath625 , we have @xmath626 .
we now claim that these properties imply that for constant @xmath627 depending on @xmath624 and @xmath623 one has @xmath628 with @xmath629 as @xmath630 , which implies @xmath631 in order to prove one can argue by contradiction as follows .
assume that for a sequence of @xmath367-curves @xmath632\to \c$ ] , which we may assume arc - length parametrized with @xmath633 and non - negative curvatures @xmath634 uniformly bounded by @xmath635 , and with @xmath636 we have @xmath637 up to translation we can assume that @xmath638 and by arzel - ascoli s theorem we have convergence of a subsequence to a @xmath367-curve @xmath639\to \c$ ] with curvature @xmath640 and @xmath641 which is impossible ( see e.g. lemma 3.10 in @xcite ) .
* proof of theorem [ mainth ] ( completed ) .
* we start by applying theorem [ convergencephi ] .
if case 1 occurs , then by lemma [ lemmaleqpi ] we have @xmath529 for every @xmath573 .
we claim that in fact @xmath19 for @xmath573 .
to do so , we compose @xmath387 with mbius diffeomorphisms @xmath642 with @xmath643 sufficiently slow , so that for @xmath644 , @xmath645 , @xmath646 we have @xmath647 now @xmath490 is as in case 2 of theorem [ convergencephi ] because @xmath292 with @xmath85 an immersion ( away from @xmath648 ) , while @xmath584 concentrates .
it follows from lemma [ lemmagepi ] that @xmath649 , hence @xmath650 moreover this can not happen if @xmath45 , since in this case lemma [ lemmagepi ] would imply @xmath651 . in particular ,
if @xmath45 for every @xmath7 and case 1 of theorem [ convergencephi ] occurs , then @xmath88 , i.e. there is no blow - up .
if case @xmath18 occurs , we apply directly lemma [ lemmagepi ] to obtain @xmath16 for @xmath72 , and in fact @xmath89 if @xmath45 in a neighbourhood of @xmath260 . the convergence of @xmath652 and the formula for @xmath653 follow exactly as in the proof of theorem [ convergencephi ] of @xcite .
a curve @xmath654 is called piecewise of class @xmath56 if @xmath655 and there exist @xmath656 [ defrot ] given a closed piecewise @xmath56-curve @xmath657 arc - length parametrized , with finitely many self - intersections , with @xmath658 ordered in the counter - clockwise direction , ( and w.l.o.g .
we can consider @xmath659 ) define @xmath660 as follows : 1 .
setting @xmath661 , we have @xmath662 and @xmath663 according to the quantities @xmath664 are also well defined .
2 . for @xmath72
set @xmath665 $ ] ( @xmath666 ) ; this condition uniquely determines @xmath667 ( called exterior angle ) and @xmath668 given @xmath660 , unless @xmath669 , and if @xmath669 we prescribe @xmath670 if the curve turns left at @xmath671 and @xmath672 if the curve turns right ( see figure [ findex2 ] ) . turning left or right
is well defined since @xmath122 has finitely many self - intersections , so it is injective in a sufficiently small neighbourhood of @xmath673 .
then we define the rotation index of @xmath122 as @xmath674 , case @xmath424 .
the curve has rotation index @xmath18.,width=529 ] a similar definition applies if @xmath122 is regular ( @xmath675 in @xmath676 ) but not arc - length parametrized , by replacing @xmath677 with @xmath678 , or if we replace @xmath74 by the image @xmath679 of a piecewise @xmath56 jordan curve @xmath680 , so that the rotation index of @xmath681 is well defined as @xmath682 .
the following theorem generalizes a result of s. j. blank @xcite . in the result of blank an immersion @xmath683
is considered and properties of the curve @xmath684 are studied , such as algebraic conditions on its self - intersections . if we allow that immersion @xmath96 to be singular at finitely many points on the boundary , things become more comples . for instance
it is not true anymore that @xmath685 , but we can at least prove that @xmath686 , which is the statement of theorem [ trmblank ] below .
this will use many ideas of the original work of blank , which we will recall during the proof .
[ trmblank ] let @xmath687 satisfy the following : 1 .
@xmath688 for a finite set of points @xmath689 .
@xmath690 everywhere in @xmath691 .
3 . @xmath692 is piecewise @xmath56 and regular .
then the rotation index of @xmath122 is positive , i.e. @xmath693 . moreover any word of blank associated to @xmath122 can be completely contracted ( see definitions in the proof ) .
up to slightly deforming @xmath96 we can assume that @xmath122 has finitely many self - intersections , each of them is transversal and away from the points @xmath694 , @xmath72 .
this of course does not change @xmath695 .
it is now possibly to locally deform @xmath96 near each point @xmath451 in a way that @xmath696 , i.e. it is possibly to flatten the angles between @xmath697 and @xmath698 , for instance by composing with suitable functions @xmath699 with @xmath700 away from @xmath451 , in a way that the rotation number @xmath695 remains unchanged , therefore we can work with this stronger assumptions .
following the work of blank @xcite ( see also @xcite - ( * ? ? ?
* section 4 ) ) , to every bounded connected component of @xmath701 we associate a letter ( @xmath702 ) , and construct a segment ( @xmath703 ) from the interior of the component and extending until it reaches the unbounded component of @xmath701 .
we can do that in such a way that the segment crosses @xmath704 in generic points ( not at self - intersection points ) and transversally . on each segment
we move from the extreme in the interior of the corresponding component towards the other extreme and to each crossing with @xmath704 we associate a natural number ( starting from @xmath185 for first crossing , then @xmath0 , etc . ) and an orientation ( @xmath705 if @xmath122 crosses the segment from the right , @xmath706 otherwise ) . for instance
@xmath707 determines finitely many points on @xmath704 named @xmath708 ( see fig .
[ fblank ] , where @xmath707 determines @xmath709 and @xmath710 ) .
since @xmath711 we can also assume that each segment stays away from @xmath712 for a small @xmath574 and @xmath72 , so that @xmath96 is a local @xmath56-diffeomorphism near the segments .
we construct now the `` word of blank '' associated to @xmath122 and this ( non - unique ) choice of segments by choosing a starting point for @xmath122 , following the curve and collecting all the letters corresponding to the crossing of the segments with @xmath704 .
for instance a word of blank could be @xmath713 ( as in fig .
[ fblank ] ) . as in the work of blank ,
we now contract such word as follows . whenever a letter with minus sign appears ( for instance @xmath709 ) in the above example , if we can find the same letter with plus sign so that between them ( seeing the word in a periodic sense , i.e. up to a cyclic permutation : @xmath714 ) there is no other letter with minus sign , we cancel all such letters : for instance @xmath715 can be contracted to @xmath716 by removing @xmath710 , @xmath709 and the letters in between . a word is completely contracted when no minus sign is left . in general a word of blank associated to a curve can not be completely contracted ( to the curve @xmath717 one can associate the words @xmath718 or @xmath719 ( see fig . [ fblank ] ) , neither of whom can be contracted ) , but using that @xmath720 for an immersion @xmath96 as in the statement of the theorem we will contract the word of blank in a way that each contraction corresponds to a slice of the domain @xmath230 . indeed consider @xmath721 for any segment @xmath707 as above . because away from @xmath722 the map @xmath96 is a locally invertible @xmath56-map
, we see that @xmath723 is a union of @xmath56 curves @xmath724 with extremes , not intersecting , and with an orientation induced by the orientation of the segments @xmath725 .
because of the orientation , all curves end on @xmath74 on points which are preimages of letters with positive sign , some curves start in the interior of @xmath230 , an those who start on @xmath74 , start from preimages of points with negative sign .
this implies that if negative signs exist ( so the word of blank is not completely contracted ) , there is at least a curve , say @xmath726 starting and ending on @xmath74 splitting @xmath230 into two compact components @xmath727 and @xmath728 ( intersecting on @xmath726 ) , one of which , say @xmath727 does not contain preimages of points with negative sign ( the proof of this fact is easy ) , and this leads to the first contraction of the word of blank , by removing all letters in @xmath729 .
call @xmath730 and consider @xmath731 . a word of blank associated to it , is easily seen to be given by the letters of the previous word of blank with pre - image in @xmath732 , which by assumption have all positive sign . by deforming the domain @xmath733 into a disk @xmath230 ( flattening the two angles at which @xmath726 intersects @xmath74 )
we do not change the rotation index of @xmath734 , and by lemma [ lemmapositive ] below , we have that @xmath735 .
we can now repeat the same procedure on @xmath728 if at least a letter with preimage in @xmath736 has negative sign . in
finitely many step we end up with a decomposition @xmath737 so that the @xmath738 and we can also order them so that @xmath739 for some letter @xmath740 . a brief look at the definition shows that @xmath741 and similarly when gluing several components , so that finally @xmath742 this completes the proof . and @xmath743 . the first one contracts , so that the first curve can be extended to an immersion of the disk , the second one does not and the curve can not be extended to an immersion of the disk.,width=529 ] .,width=529 ] [ corangle ] let @xmath96 and @xmath122 be as in theorem [ trmblank ] , with @xmath424 and let @xmath744 be as in definition [ defrot ] .
then @xmath745 in particular if @xmath746 , then @xmath747 as a consequence , given @xmath169 solution of @xmath748 with @xmath749 , @xmath189 , @xmath750 and @xmath751 , one always has @xmath752 .
indeed by theorem [ trmblank ] we have @xmath753 with @xmath754.$ ] the last statement follows by first applying theorem [ geomequiv ] to construct @xmath96 and @xmath122 , to which we can then apply the first part of the corollary . into @xmath483 simple oriented curves.,width=529 ] [ lemmapositive ] if a closed piecewise @xmath56-curve @xmath755 admits a word of blank without negative signs , then @xmath693 .
this ( probably well - known ) fact is based on what is called seifert decomposition ( fig .
[ fseifert ] ) . up to a small perturbation of @xmath122
we can assume that it has finitely many self - intersections , avoiding corners , and transversal .
decompose @xmath704 into finitely many piecewise @xmath56 jordan curves @xmath756 , @xmath757 ( with no self - intersections but new angles at the self - intersection points of @xmath704 , which can be smoothed out , although this is not necessary in our setting ) endowed with the same orientation given by @xmath122 . each @xmath758 is positively oriented , otherwise the segment leaving the component bounded by @xmath759 would give rise to a negative letter in the word of blank .
then @xmath760 for every @xmath17 and one easily sees that @xmath761 we shall collect here some useful results and definitions about half - laplacians and holomorphic maps .
given @xmath762 we define its fourier coefficients as @xmath763 if @xmath78 is smooth we can define@xmath764 for @xmath762 we can define @xmath765 as distribution as @xmath766 if @xmath762 , set @xmath767 ( in polar coordinates ) as @xmath768 this is the poisson integral of @xmath78 and it is harmonic .
then @xmath769 for every @xmath770 , where @xmath771 is the harmonic extension of @xmath772 in @xmath773 for @xmath779 ( the schwarz space of rapidly decaying functions ) we set @xmath780 one can prove that it holds @xmath781}\frac{u(x)-u(y)}{(x - y)^2}dy,\ ] ] from which it follows that @xmath782 then one can set @xmath783 and for every @xmath136 one defines the tempered distribution @xmath784 as @xmath785 an alternative definition of @xmath59 can be given via the poisson integral . for @xmath136 define the poisson integral @xmath786 which is harmonic in @xmath787 and whose trace on @xmath788 is @xmath78 .
then we have @xmath789 where the identity is pointwise if @xmath78 is regular enough ( for instance @xmath790 ) , and has to be read in the sense of distributions in general , with @xmath791 [ lapeq ] if @xmath792 for some interval @xmath793 and some @xmath794 , then the tempered distribution @xmath795 defined in coincides on the interval @xmath796 with the functions given by and . for general @xmath90 the definitions and are equivalent , where the right - hand side of is defined by .
[ lemmafund4 ] the function @xmath797 belongs to @xmath798 , can be decomposed as @xmath799\sim s^1 , \quad \text{with } h\in c^0(s^1),\ ] ] and satisfies @xmath800 and for every function @xmath762 with @xmath801 one has @xmath802 s. j. blank , _ extending immersions and regular homotopies in codimension one _ , phd thesis , brandeis university , 1967 .
h. brezis , f. merle , _ uniform estimates and blow - up behaviour for solutions of @xmath813 in two dimensions _ , comm .
partial differential equations * 16 * ( 1991 ) , 1223 - 1253 .
r. burckel , _ an introduction to classical complex analysis , vol.1 _ , pure and applied mathematics , vol .
82 , academic press , new york , 1979 .
a. chang , non - linear elliptic equations in conformal geometry , zurich lecture notes in advanced mathematics , ems ( 2004 ) .
a. chang , _ conformal invariants and partial differential equations , _ bull .
( n.s . ) * 42 * , 2005 , 365 - 393 .
s - y . a. chang , p. yang , _ on uniqueness of solutions of @xmath226-th order differential equations in conformal geometry _
lett . * 4 * ( 1997 ) , 91 - 102 .
w. chen , c. li , _ classification of solutions of some nonlinear elliptic equations _ , duke math . j.
* 63 * ( 3 ) ( 1991 ) , 615 - 622 .
f. da lio , t. rivire , _ sub - criticality of non - local schrdinger systems with antisymmetric potentials and applications to @xmath814-harmonic maps , _ advances in mathematics * 227 * , ( 2011 ) , 1300 - 1348 .
f. da lio , l. martinazzi , t. rivire , _ blow - up analysis of a nonlocal liouville - type equations _ , analysis and pde * 8 * , no . 7 ( 2015 ) , 1757 - 1805
. o. druet , f. robert , _ bubbling phenomena for fourth - order four - dimensional pdes with exponential growth _ ,
soc * 3 * ( 2006 ) , 897 - 908 .
a. hyder , _ structure of conformal metrics on @xmath815 with constant q - curvature _ , ` arxiv:1504.07095 ` ( 2015 ) .
t. jin , a. maalaoui , l. martinazzi , j. xiong , _ existence and asymptotics for solutions of a non - local q - curvature equation in dimension three _ , calc .
* 52 * ( 2015 ) , 469 - 488 . c. kenig , _ harmonic analysis techniques for second order elliptic boundary value problems , _ cbms , 83 , 1994
. y. li , i. shafrir , _ blow - up analysis for solutions of @xmath816 in dimension @xmath18 _ , indiana univ .
* 43 * ( 1994 ) , 1255 - 1270 .
lin , _ a classification of solutions of a conformally invariant fourth order equation in @xmath817 _ , comment .
* 73 * ( 1998 ) , no . 2 , 206 - 231 .
a. malchiodi _ compactness of solutions to some geometric fourth - order equations _
, j. reine angew .
* 594 * ( 2006 ) , 137 - 174 .
i. p. malta , n. c. saldanha , c. tomei , _ geometria e anlise numrica de funes do plano no plano _
, 19o colquio brasileiro de matematica , impa 1993 . l. martinazzi , _ classification of solutions to the higher order liouville s equation on @xmath818 _ , math . z.
* 263 * ( 2009 ) , 307329 .
l. martinazzi , _ conformal metrics on @xmath818 with constant @xmath54-curvature _ , rend .
lincei . mat .
* 19 * ( 2008 ) , 279 - 292 .
l. martinazzi , _ concentration - compactness phenomena in higher order liouville s equation _ , j. funct .
256 * ( 2009 ) , 3743 - 3771 . l. martinazzi , _
quantization for the prescribed q - curvature equation on open domains _ , commun .
contemp . math .
* 13 * ( 2011 ) , 533 - 551 . c. b. ndiaye
_ constant @xmath54-curvature metrics in arbitrary dimension _ , j. func .
analysis * 251 * no.1 ( 2007 ) , 1 - 58 .
r. oneil , _ convolution operators and @xmath819 spaces _ , duke math .
j. 30 ( 1963 ) , 129 - 142 .
v. ponaru , _ extension des immersions en codimension @xmath0 _ , sminaire bourbaki 1967/1968 , expos 342 .
f. robert , _ quantization effects for a fourth order equation of exponential growth in dimension four _ , proc .
edinburgh sec .
a * 137 * ( 2007 ) , 531 - 553 .
g. tarantello , , progress in nonlinear differential equations and their applications , 72 .
birkhuser , boston 2008 .
x. xu , _ uniqueness and non - existence theorems for conformally invariant equations _ , j. funct .
222 ( 2005 ) , no . 1 , 1 - 28 | in this paper we perform a blow - up and quantization analysis of the fractional liouville equation in dimension @xmath0 . more precisely , given a sequence @xmath1 of solutions to @xmath2 with @xmath3 bounded in @xmath4 and @xmath5 bounded in @xmath6 uniformly with respect to @xmath7 , we show that up to extracting a subsequence @xmath8 can blow - up at ( at most ) finitely many points @xmath9 and either ( i ) @xmath10 in @xmath11 and @xmath12 , or ( ii ) @xmath13 uniformly locally in @xmath14 and @xmath15 with @xmath16 for every @xmath17 .
this result , resting on the geometric interpretation and analysis of provided in a recent collaboration of the authors with t. rivire and on a classical work of blank about immersions of the disk into the plane , is a fractional counterpart of the celebrated works of brzis - merle and li - shafrir on the @xmath18-dimensional liouville equation , but providing sharp quantization estimates ( @xmath19 and @xmath16 ) which are not known in dimension @xmath18 under the weak assumption that @xmath20 be bounded in @xmath4 and is allowed to change sign . |
the minimal supersymmetric standard model ( mssm ) @xcite is affected by the so - called ` @xmath0-problem ' .
the mssm superpotential can be written as @xcite @xmath1 ( hereafter , hatted variables describe superfields while un - hatted ones stand for the corresponding scalar superfield components ) .
the last term in the above equation contains a dimensionful parameter , @xmath0 . upon electro - weak symmetry breaking ( ewsb ) ,
it provides a contribution to the masses of both higgs bosons and higgsino fermions .
furthermore , the associated soft supersymmetry ( susy ) breaking term @xmath2 mixes the two higgs doublets .
now , the presence of @xmath0 in the superpotential before ewsb indicates that its natural value would be either 0 or the planck mass @xmath3 . on the one hand , @xmath4 would mean no mixing is actually generated between higgs doublets at any scale and the minimum of the higgs potential occurs for @xmath5 , so that one would have in turn massless down - type fermions and leptons after su(2 ) symmetry breaking . on the other hand , @xmath6 would reintroduce a ` fine - tuning problem ' in the mssm since the higgs scalars would acquire a huge contribution @xmath7 to their squared masses ( thus spoiling the effects of susy , which effectively removes otherwise quadratically divergent contributions to the higgs mass from sm particles ) .
therefore , the values of this ( theoretically arbitrary ) parameter @xmath0 are phenomenologically constrained to be close to @xmath8 or @xmath9 @xcite .
the most elegant solution to the @xmath0-problem is to introduce a new singlet higgs field @xmath10 and replace @xmath2 by an interaction @xmath11 .
when the extra scalar field @xmath10 acquires a vacuum expectation value ( vev ) , an effective @xmath0 term , naturally of the ew scale , is generated automatically .
this idea has been implemented in the next - to - minimal supersymmetric standard model ( nmssm ) @xcite , described by the superpotential @xmath12 where @xmath13 is an extra higgs iso - singlet superfield , @xmath14 and @xmath15 are dimensionless couplings and the last ( @xmath16 invariant ) term is required to explicitly break the dangerous peccei - quinn ( pq ) u(1 ) symmetry @xcite group , so that the @xmath16 symmetry is embedded in the local gauge symmetry @xcite . ] .
( see ref .
@xcite for nmssm higgs sector phenomenology with an exact or slightly broken pq symmetry . )
furthermore , a @xmath17 symmetry can be imposed to avoid the so - called ` domain - walls problem ' of the nmssm @xcite , through harmless tadpoles breaking the global @xmath16 symmetry @xcite .
( alternative formulations known as the minimal non - minimal supersymmetric standard model ( mnssm ) and new minimally - extended supersymmetric standard model ( nmssm ) vetoing the presence of the @xmath18 term and allowing instead suitable linear tadpole terms through the enforcement of discrete @xmath19-symmetries also exist @xcite . )
another positive feature of all these non - minimal susy models is that they predict the existence of a ( quasi-)stable singlet - type neutralino ( the singlino ) that could be responsible for the dark matter ( dm ) of the universe @xcite .
finally , notice that , in these extended susy models , the singlet superfield @xmath13 has no sm gauge group charge ( so that mssm gauge coupling unification is preserved ) and that one can comfortably explain the baryon asymmetry of the universe by means of a strong first order ew phase transition @xcite ( unlike the mssm , which would require a light top squark and higgs boson barely compatible with current experimental bounds @xcite ) . clearly , in eq .
( [ wnmssm ] ) , upon ewsb a vev will be generated for the real scalar component of @xmath20 ( the singlet higgs field ) , @xmath21 , alongside those of the two doublets @xmath22 and @xmath23 ( related by the parameter @xmath24 ) . in the absence of fine - tuning , one should expect these three vevs to be of the order of @xmath25 or @xmath9 , so that now one has an ` effective @xmath0-parameter ' @xmath26 of the required size , thus effectively solving the @xmath0-problem .
in the nmssm , the soft susy - breaking higgs sector is described by the lagrangian contribution @xmath27 where @xmath28 and @xmath29 are dimensionful parameters of order @xmath25 . as a result of the introduction of an extra complex singlet scalar field , which only couples to the two mssm - type higgs doublets ,
the higgs sector of the nmssm comprises of a total of seven mass eigenstates : a charged pair @xmath30 , three cp - even higgses @xmath31 ( @xmath32 ) and two cp - odd higgses @xmath33 ( @xmath34 ) .
consequently , higgs phenomenology in the nmssm may be plausibly different from that of the mssm and extremely rich of new signals . for a start
, the mass expressions for the cp - even higgses in the nmssm can be translated into a modified upper bound on the lightest higgs mass , @xmath35 , as @xcite @xmath36 as the higher order corrections are similar to those in the mssm , the upper bound on the lightest higgs boson is different in the nmssm , reaching 135140 gev , for maximal stop mixing and @xmath37 @xcite ( a configuration indeed excluded in the mssm by lep data ) .
more in general , the ` little fine tuning problem ' , resulting in lep failing to detect a light cp - even higgs boson , predicted over most of the mssm parameter space , is much attenuated in the nmssm , because the mixing among more numerous cp - even or cp - odd higgs fields enables light mass states being produced at lep yet they can remain undetected because of their reduced couplings to @xmath38 bosons @xcite .
as for future machines , chiefly the cern large hadron collider ( lhc ) , quite some work has been dedicated to probing the nmssm higgs sector over recent years .
primarily , there have been attempts to extend the so - called ` no - lose theorem ' of the mssm stating that at least one mssm higgs boson should be observed via the usual sm - like production and decay channels at the lhc throughout the entire mssm parameter space @xcite to the case of the nmssm @xcite . from this perspective , it was realised that at least one nmssm higgs boson should remain observable at the lhc over the nmssm parameter space that does not allow any higgs - to - higgs decay .
however , when the only light non - singlet ( and , therefore , potentially visible ) cp - even higgs boson , @xmath39 or @xmath40 , decays mainly to two very light cp - odd higgs bosons , @xmath41 , one may not have a higgs signal of statistical significance at the lhc @xcite .
in fact , further violations to the theorem may well occur if one enables higgs - to - susy decays ( e.g. , into neutralino pairs , yielding invisible higgs signals ) . while the jury is still out on whether a ` no - lose theorem ' can be proved for the nmssm , we are here concerned with an orthogonal approach .
we asked ourselves if a , so to say , ` more - to - gain theorem ' can be formulated in the nmssm .
that is , whether there exist regions of the nmssm parameter space where more higgs states of the nmssm are visible at the lhc than those available within the mssm . in our attempt to overview all such possibilities ,
we start by considering here the case of the di - photon decay channel of a neutral higgs boson .
this mode can be successfully probed in the mssm , but limitedly to the case of one higgs boson only , which is cp - even and rather light
. we will show that in the nmssm one can instead potentially have up to three di - photon signals of higgs bosons , involving not only cp - even but also cp - odd states , the latter with masses up to 600 gev or so .
in fact , even when only one di - photon signal can be extracted in the nmssm , this may well be other than the @xmath39 state .
when only the latter is visible , finally , it can happen that its mass is larger than the maximum value achievable within the mssm . in all such cases then , the existence of a non - minimal susy higgs sector would be manifest .
the choice of parameter space largely depends on the version of nmssm under consideration , or more specifically , on the implication of unification of parameters at some very high scale .
this , in turn , leans on the technique adopted for the breaking of susy , since it still remains undetected in nature . the advantage of assuming unification of masses and couplings in a susy model is twofold .
firstly , it enormously reduces the labour from a phenomenological perspective by minimising the number of parameters required to extrapolate physical information from the model .
secondly , and theoretically more crucially , it caters to the fundamental objective of constructing a grand unification theory ( gut ) , which was , to a great extent , responsible for devising susy in the first place .
a particular case of a low energy nmssm with gravity mediated susy breaking was studied in @xcite . as usual , in such a model , susy is assumed to be broken in some hidden sector and then mediated to the physical sector through gravitational interactions .
such a model implies unification of the couplings and soft masses at the gut scale , which are then run down , using renormalisation group equations , to the weak scale or to some other scale at which the theory is being tested .
this results in only a handful of parameters to deal with . for a more general study of the nmssm higgs sector with a wider range of parameters
, we used here the publicly available fortran code nmhdecay ( version 1.1 ) developed in ref .
this program computes the masses , couplings and decay widths of all the higgs bosons of the nmssm in terms of its parameters at the ew scale . for our purpose , instead of postulating unification , and without taking into account the susy breaking mechanism , we fixed the soft susy breaking terms to a very high value , so that they have little or no contribution to the outputs of the parameter scans .
consequently , we are left with six free parameters .
our parameter space includes the yukawa couplings @xmath14 and @xmath15 , the soft trilinear terms @xmath28 and @xmath29 , plus tan@xmath42 and @xmath43 .
the computation of the spectrum includes leading two loop terms , ew corrections and propagator corrections .
the decay widths , however , do not include three body decays .
the nmhdecay program also takes into account theoretical as well as experimental constraints from negative higgs searches at lep , along with the unconventional channels relevant for the nmssm .
we have used the nmhdecay code to scan over the nmssm parameter space defined through the aforementioned six parameters taken in the following intervals : @xmath14 : 0.0001 0.75 , @xmath15 : @xmath440.65 + 0.65 , @xmath45 : 1.6 54 , + @xmath0 , @xmath46 , @xmath47 : @xmath441000 + 1000 gev .
+ remaining soft terms which are fixed in the scan include : + @xmath48 tev , + @xmath49 tev , + @xmath50 tev , + @xmath51 tev .
+ in line with the assumptions made in @xcite , the allowed decay modes for neutral nmssm higgs bosons are to signify any of the neutral cp - even(odd ) higgs bosons of the nmssm . ] : @xmath52 ( notice that for the pseudoscalar higgses , the decay into vector boson pairs is not allowed due to cp - conservation . ) here , ` higgses ' refers to any final state involving all possible combination of two higgs bosons ( neutral and/or charged ) or of one higgs boson and a gauge vector .
the aforementioned range of parameters is borrowed from ref .
@xcite , where the authors singled out the most difficult scenarios for nmssm higgs discovery at the lhc ( using the mentioned code ) . in this paper
they also concluded , as a follow up of the nmssm ` no - lose theorem ' , that the region where no higgs is observable , due to its decays into the unconventional channel @xmath41 , comprises of only about 1% of the entire parameter space of the nmssm . in short ,
our idea is to explore the same range of parameter space and look for the ` best case scenarios ' where the discovery of one or more higgses at the lhc is possible above and beyond what is predicted in the mssm .
we have performed our scan over several millions of randomly selected points in the specified parameter space .
the output , as stated earlier , contains masses , branching ratios ( brs ) and couplings of the nmssm higgses , for all the points which are not forbidden by the various constraints .
the points which violate the constraints are eliminated .
the surviving data points are then used to determine the cross - sections for nmssm higgs hadro - production by using an adapted version of the codes described in @xcite . as the susy mass scales have been arbitrarily set well above the ew one ( see above ) ,
the nmssm higgs production modes exploitable in simulations at the lhc are those involving couplings to heavy ordinary matter only , i.e. , ( hereafter , @xmath53 and @xmath54 ) for neutral higgs production ( where the last two channels are only allowed for cp - even higgs production ) : @xmath55 @xmath56 @xmath57 @xmath58 ( these are the so - called ` direct ' higgs production modes . ) here , ` higgs ' refers to any possible neutral higgs boson .
production and decay rates for nmssm neutral higgs bosons have then been multiplied together to yield inclusive event rates , assuming a lhc luminosity of 100 fb@xmath59 throughout .
as an initial step towards the analysis of the data , we computed the total cross - section times br into @xmath60 pairs against each of the six parameters of the nmssm , for each higgs boson .
we have assumed all production modes described in the above section and started by computing total ( i.e. , fully inclusive ) rates .
we are here focusing on the @xmath60 decay mode since it is the most promising channel for the discovery of a ( neutral ) higgs boson at the lhc in the moderate higgs mass range ( say , below 130 gev ) .
in fact , since the tail of the @xmath60 background falls rapidly with increasing invariant mass of the di - photon pair , signal peaks for heavier higgses could also be visible in addition to ( or instead of ) the lightest one , although the cross - section for these processes is relatively very low @xcite . as the starting point of our numerical study ,
based on the atlas analysis of ref .
@xcite , we argue that cross - section times br rates of 10 fb or so are potentially interesting from a phenomenological point of view , in the sense that they may yield visible signal events , the more so the heavier the decaying higgs state ( also because the photon detection efficiency grows with the higgs mass @xcite ) .
( see ref .
@xcite for a preliminary account in this respect . )
[ tab:1 ] recaps the potential observability of one or more nmssm higgs states in the di - photon mode at the lhc , under the above assumptions .
it is obvious from the table that one light cp - even higgs should be observable almost throughout the nmssm parameter space ( in line with the findings of ref .
@xcite ) . however , there is also a fair number of points where two higgses may be visible simultaneously ( @xmath39 and @xmath40 or more rarely
@xmath39 and @xmath61 ) , while production and decay of the three lightest higgses ( @xmath39 , @xmath40 and @xmath61 ) at the same time , although possible , occurs for only a very small number of points in the parameter space .
furthermore , the percentage of points for which only the second lightest higgs state is visible is also non - negligible .
these last two conditions are clearly specific to the nmssm , as they are never realised in the mssm . furthermore ,
while the lone detection of the lightest cp - even nmssm higgs boson may mimic a similar signal from the corresponding state in the mssm , the reconstructed mass may well be beyond the upper mass limit in the mssm , this possibility also pointing towards the evidence of a nmssm higgs sector .
finally , none of the two heaviest nmssm neutral higgs states ( @xmath62 and @xmath63 ) will be visible in the di - photon channel at the lhc ( given their large masses ) .
we have then plotted the cross - sections times br for the three potentially observable higgses against the various nmssm parameters .
these plots , shown in figs . 16 ,
reveal that the distribution of possibly visible points ( i.e. , of those yielding cross - section time br rates in excess of 10 fb ) is quite homogeneous over the nmssm parameter space and not located in some specific parameter areas ( i.e. , in a sense , not ` fine - tuned ' ) .
the distribution of the same points as a function of the corresponding higgs masses can be found in fig .
7 . of particular relevance is the distribution of points in which only the nmssm @xmath39 state is visible , when its mass is beyond the upper mass limit for the corresponding cp - even mssm higgs state , which is shown in fig . 8 from nmhdecay version 1.1 , of @xmath64 gev , based on the leading two - loop approximations described in @xcite , is a few gev lower than the value declared in sect . 1 .
besides , for consistency , we use the value of 120 gev ( obtained at the same level of accuracy ) as upper mass limit on the lightest cp - even higgs boson of the mssm .
( notice that a slightly modified @xmath65 value is obtained for the nmssm from nmhdecay version 2.1 @xcite , because of the improved mass approximations with respect to the earlier version of the program adopted here . )
eventually , when the lhc is on line , the exercise that we are proposing can be performed with the then state - of - the - art calculations . ] .
this plot reveals that about 93% of the nmssm @xmath39 masses visible alone are expected to be within 23 gev beyond the mssm bound , hence the two models would be indistinguishable .
nonetheless , there is a fraction of a percent of such points with @xmath35 values even beyond 125 gev or so ( the higher the mass the smaller the density , though ) , which should indeed allow one to distinguish between the two models .
moreover , by studying the cross - section times br of the higgses when two of them are observable against their respective mass differences , figs .
911 , one sees that the former are larger than the typical mass resolution in the di - photon channel , so that the two decaying objects should indeed appear in the data as separate resonances .
( we have also verified , though not shown here , that their decay widths are small compared to the detector resolution in @xmath66 . )
next , we have proceeded to a dedicated parton level analysis of signal and background processes , the latter involving both tree - level @xmath67 and one - loop @xmath68 contributions .
we have adopted standard cuts on the two photons @xcite : @xmath69 gev and @xmath70 2.4 on transverse energy and pseudorapidity , respectively . as illustrative examples of a possible nmssm higgs phenomenology appearing at the lhc in the di - photon channel
, we have picked up the following three configurations : 1 .
@xmath71 0.6554 , @xmath72 0.2672 , @xmath73 gev , @xmath74 2.68 , @xmath75 gev , @xmath76 gev ; 2 .
@xmath71 0.6445 , @xmath72 0.2714 , @xmath77 gev , @xmath74 2.62 , @xmath78 gev , @xmath79 gev ; 3 .
@xmath71 0.4865 , @xmath72 0.3516 , @xmath80 gev , @xmath74 2.35 , @xmath81 gev , @xmath82 gev .
the first is representative of the case in which only the nmssm @xmath39 boson is visible , but with mass larger than the mssm upper limit on the corresponding higgs state .
the second and third refer instead to the case when also the @xmath40 or @xmath61 state are visible , respectively .
the final results are found in fig .
the corresponding mass resonances are clearly visible above the continuum di - photon background and discoverable beyond the @xmath83 level .
indeed , similar situations can be found for each of the combinations listed in tab .
[ tab:1 ] and most of these correspond to phenomenological scenarios which are distinctive of the nmssm and that can not be reproduced in the mssm .
in summary , we have shown that there exists the possibility of establishing a ` more - to - gain theorem ' within the nmssm , as compared to what is expected in the mssm , in terms of novel higgs signals appearing in the di - photon discovery channel which can be ascribed to the former but not to the latter .
we have showed this to be the case for a few selected nmssm parameter points , by performing a proper signal - to - background analysis at the partonic level . however , a similar numerical study can easily be extended to encompass sizable regions of the nmssm parameter space .
while the bulk of the latter is in a configuration degenerate with the mssm case ( as far as di - photon signals at the lhc are concerned ) , non - negligible areas exist where further phenomenological studies have the potential to unveil a non - minimal nature of the underlying susy model . to this end ,
nmssm benchmark scenarios , amenable to experimental investigation in the context of full monte carlo ( mc ) analyses , based on event generation and detector simulation , are currently being devised @xcite . also , un upgraded version of herwig @xcite , suitable for mc event generation in the context of non - minimal susy models , is currently being prepared @xcite .
we acknowledge useful email exchanges with cyril hugonie . sm would like to thank the british council under the ` alliance :
franco british partnership programme 2004 ( project number : pn 04.051 ) ' for travel support in connection with this research .
he is also grateful to fawzi boudjema and genevieve belanger for their kind hospitality at lapp - th . for reviews
see , e.g. : h.e . haber and g.l .
kane , phys .
* 117 * ( 1985 ) 75 ; p. fayet and s. ferrara , phys . rep .
* 32 * ( 1977 ) 249 ; h.p .
nilles , phys .
* 110 * ( 1984 ) 1 ; r. barbieri , riv .
* 11 * ( 1988 ) 1 .
p. fayet , nucl .
b * 90 * ( 1975 ) 104 ; phys .
b * 64 * ( 1976 ) 159 ; phys . lett .
b * 69 * ( 1977 ) 489 and phys . lett .
b * 84 * ( 1979 ) 416 ; h.p .
nilles , m. srednicki and d. wyler , phys .
b * 120 * ( 1983 ) 346 ; j.m .
frere , d.r .
jones and s. raby , nucl .
b * 222 * ( 1983 ) 11 ; j.p .
derendinger and c.a .
savoy , nucl .
b * 237 * ( 1984 ) 307 ; k. inoue , a. komatsu and s. takeshita , prog .
phys * 68 * ( 1982 ) 927 ; ( e ) _ ibid . _
* 70 * ( 1983 ) 330 ; m. dine , w. fischler and m. srednicki , phys .
b * 104 * ( 1981 ) 199 ; a.i .
veselov , m.i .
vysotsky and k.a .
ter - martirosian , sov .
jetp * 63 * ( 1986 ) 489 [ zh .
* 90 * ( 1986 ) 838 ] ; b.r .
greene and p.j .
miron , phys .
b * 168 * ( 1986 ) 26 ; j.r .
ellis , j.f .
gunion , h.e .
haber , l. roszkowski and f. zwirner , phys .
d * 39 * ( 1989 ) 844 ; m. drees , int .
a * 4 * ( 1989 ) 3635 ; u. ellwanger , m. rausch de traubenberg and c.a .
savoy , phys .
b * 315 * ( 1993 ) 331 ; p.n .
pandita , z. phys .
c * 59 * ( 1993 ) 575 ; s.f .
king and p.l .
white , phys .
d * 52 * ( 1995 ) 4183 .
m. cvetic , d.a .
demir and l. everett , phys .
d * 56 * ( 1997 ) 2861 ; d.a .
demir and l. everett , phys .
d * 69 * ( 2004 ) 015008 ; t. han , p. langacker and b. mcelrath , phys .
d * 70 * ( 2004 ) 115006 ; s.f .
king , s. moretti and r. nevzorov , phys .
b * 634 * ( 2006 ) 278 , phys .
d * 73 * ( 2006 ) 035009 and hep - ph/0601269 ( and references therein ) .
nilles , m. srednicki , and d. wyler , phys . lett .
b * 124 * ( 1983 ) 337 ; a.b .
lahanas , phys .
b * 124 * ( 1983 ) 341 ; u. ellwanger , phys . lett .
b * 133 * ( 1983 ) 187 ; j. bagger and e. poppitz , phys .
* 71 * ( 1993 ) 2380 ; j. bagger , e. poppitz and l. randall , nucl .
b * 426 * ( 1994 ) 3 ; v. jain , phys .
b * 351 * ( 1995 ) 481 ; s.a .
abel , nucl .
b * 480 * ( 1996 ) 55 ; c.f .
kolda , s. pokorski and n. polonsky , phys . rev
* 80 * ( 1998 ) 5263 ; s.a .
abel , s. sarkar and p.l .
white , nucl .
b * 454 * ( 1995 ) 663 . c. panagiotakopoulos and k. tamvakis , phys .
b * 469 * ( 1999 ) 145 ; c. panagiotakopoulos and a. pilaftsis , phys .
d * 63 * ( 2001 ) 055003 ; a. dedes , c. hugonie , s. moretti and k. tamvakis , phys .
d * 63 * ( 2001 ) 055009 a. menon , d.e .
morrissey and c.e.m .
wagner , phys .
d * 70 * ( 2004 ) 035005 ; v. barger , p. langacker and h .- s .
lee , phys .
b * 630 * ( 2005 ) 85 ; j.f .
gunion , d. hooper and b. mcelrath , hep / ph0509024 ; g. belanger , f. boudjema , c. hugonie , a. pukhov and a. semenov , jcap * 0509 * ( 2005 ) 001 .
m. bastero - gil , c. hugonie , s.f .
king , d.p .
roy and s. vempati , phys .
b * 489 * ( 2000 ) 359 ; s.w .
ham , s.k .
oh , e.j .
yoo , c.m .
kim and d. son , phys .
d * 70 * ( 2004 ) 075001 ; k. funakubo , s. tao and f. toyoda , prog .
* 114 * ( 2005 ) 369 .
.[tab:1 ] higgs events potentially visible at the lhc through the @xmath60 decay mode ( i.e. , those yielding cross - section time br rates of order 10 fb or upwards ) .
percentage refers to the portion of nmssm parameter space involved for each discovery scenario . [
cols="^,<,<,^",options="header " , ] and @xmath85 mentioned in the text , for 100 fb@xmath59 of luminosity , in the case of the background ( solid ) and the sum of signal and background ( dashed ) , for the example points 1.3 . described in the text ( from top to bottom , in correspondence).,title="fig : " ] + and @xmath85 mentioned in the text , for 100 fb@xmath59 of luminosity , in the case of the background ( solid ) and the sum of signal and background ( dashed ) , for the example points 1.3 . described in the text ( from top to bottom , in correspondence).,title="fig : " ] + and @xmath85 mentioned in the text , for 100 fb@xmath59 of luminosity ,
in the case of the background ( solid ) and the sum of signal and background ( dashed ) , for the example points 1.3 . described in the text ( from top to bottom , in correspondence).,title="fig : " ] + | the nmssm contains a higgs singlet in addition to the two higgs doublets typical of the mssm , thus resulting in a total of seven physical higgs mass states .
therefore , the phenomenology of the nmssm higgs sector can vary considerably from that of the mssm and there are good prospects of finding in regions of the nmssm parameter space higgs signals that can not be reproduced in the mssm .
we examined here the two - photon decay mode of a higgs boson and found that up to three neutral higgs states , heavy and/or light , could be simultaneously observable at the lhc , a possibility precluded to the mssm .
there are also some possibilities that only the lightest nmssm higgs boson be detectable via this mode , with a mass beyond the upper limit of the corresponding mssm state , thus also allowing to distinguish between the two scenarios .
however , in most of the nmssm parameter space the configurations of the non - minimal model are not very different from those arising in the minimal case .
preprint shep-06 - 01 + * di - photon higgs signals at the lhc in the + next - to - minimal supersymmetric standard model * + + |
the top quark certainly is one of the most interesting of the presently known elementary particles , partly because it is the one we know least of , but also because its large mass implies that we can use it for a variety of unprecedented qcd studies .
a future 500 @xmath0 collider will provide a clean laboratory where we can investigate the top quark s properties to a high precision by performing a detailed analysis of its production close to threshold .
it is thus not surprising to find a long series of papers concerned with this subject , i.e. @xcite@xcite , leading to the following conclusions : 1 .
the excitation curve @xmath1 , i.e. the total cross - section for @xmath2-production , allows for a precise measurement of @xmath3 and the top quark mass .
this is in analogy to the charmonium and bottomonium systems with the difference that perturbative qcd is far more reliable in the @xmath2 system because of the considerably higher scale and , as will be explained below , the large top quark width .
2 . the momentum distribution @xmath4 and the forward - backward asymmetry will provide a second , independent determination of @xmath5 and also measure the top quark width @xmath6 and consequently give us a handle on @xmath7 or physics beyond the standard model .
the size of the top quark polarization , which is predicted to amount to @xmath8 even for unpolarized beams , is governed by the top quark s couplings to the @xmath9-boson and the photon .
hence its measurement will give us information about these presently unknown parameters .
we will also gain some information about the qcd potential , which determines the dynamics of the nonrelativistic `` toponium '' system .
when compared with charmonium and bottomonium there seems to be a big problem connected to the top quark , caused by its enormous width of about 1.5 : the heavy quarks will decay before resonances can build up and thus we wo nt see narrow peaks in @xmath1 . furthermore , the width can not be neglected in calculations .
but as fadin and khoze pointed out several years ago @xcite , this obstacle in fact turns out to be an advantage , because it implies that top decays before hadronization effects can spoil the purely perturbative behaviour . the width acts as an infrared cutoff and protects us from the poorly known physics at large distances .
in addition , the study of momentum distributions and top polarization is only made possible by the dominance of the single quark decay . @xmath10 and @xmath11 for example mainly decay via @xmath12 annihilation and thus all information about momentum distributions is lost . in the following section
i will briefly describe the calculation of the cross - section for @xmath2 production including terms of the order @xmath13 .
these results are however inconsistent without the inclusion of @xmath14 corrections , which i will present in section 3 for the differential cross - section and in section 4 for the top quark polarization .
i will not dwell too much on the technical parts of the calculation .
the reader who is interested in more details is advised to read @xcite .
close to threshold it is convenient to employ a nonrelativistic approximation and use the corresponding variables : the top quark three - momentum in the laboratory frame @xmath15 , with @xmath16 and @xmath17 , and the energy measured from nominal threshold @xmath18 .
we will also need the angle @xmath19 between the top and electron beam directions and the three unit vectors @xmath20 , @xmath21 and @xmath22 which are defined in fig . [ fig1 ]
. in the kinematic regime @xmath23 one usually encounters the problem that coulomb gluon exchange is not suppressed because @xmath24 is not a small parameter and thus all terms of the form @xmath25 have to be summed up .
this is usually solved by using the schrdinger equation to calculate the wave - functions for the bound states and expressing the observables in terms of the latter . in the @xmath2 system
there is however the additional complication already mentioned : the large width which can not be neglected . as a consequence
there will be no isolated resonances in the cross - section and thus the usage of wave - functions is very inconvenient , because lots of them would have to be summed , including interferences .
the idea of fadin and khoze @xcite to circumvent this was to use the green function instead , which by construction is a sum over all wave - functions : @xmath26 a method to see this function appear in a diagrammatical approach in momentum space is as follows : one can express the fact that all gluon exchanges have to be summed ( in ladder approximation ) in the form of self - consistency equations for the vertex functions @xmath27 , with @xmath28 or @xmath29 in the case of interest for @xmath0 annihilation , @xmath30 which are also shown in fig .
[ lseq ] . in the nonrelativistic approximation , neglecting terms of the order @xmath31 , they can essentially be reduced to the integral equations @xmath32 the functions @xmath33 in turn are tightly connected to the green function @xmath34 which solves the lippmann - schwinger equation @xmath35 { \cal g } ( \bfp,\bfx , e ) = e^{i \bfp \cdot \bfx } + \int { d^3 k \over ( 2 \pi)^3 } v(\bfp - \bfk ) { \cal g } ( \bfk,\bfx , e ) : \ ] ] expanding @xmath34 in a taylor series around @xmath36 we find @xmath37 with @xmath38 and @xmath39 @xmath40 and @xmath41 thus are the @xmath42- and @xmath43-wave green functions , respectively .
they form the nontrivial dynamical ingredients for the production cross - section which is otherwise completely determined by electroweak coupling constants . for a realistic potential they have to be determined numerically , and it turns out convenient to perform this calculation directly in momentum space using a method introduced in @xcite .
we will employ the following conventions for the fermion couplings @xmath44 and the useful abbreviations @xmath45 @xmath46 denotes the longitudinal positron / electron polarization and the variable @xmath47 is defined as @xmath48 with these parameters the differential cross - section for the production of a @xmath2 pair with a top quark of three - momentum @xmath15 and spin projection @xmath49 on direction * n * can be written in the form @xmath50
@xmath51 denotes the ( momentum dependent ) forward - backward asymmetry and * p * can be interpreted as top quark polarization vector .
these quantities are given by @xmath52 with @xmath53 , @xmath54 @xmath55 and @xmath56 , @xmath57 .
they have first been derived for polarized beams in @xcite , apart from p@xmath58 which has been supplemented in @xcite .
the forward - backward asymmetry has been calculated first , for unpolarized beams , in @xcite .
the coefficient functions @xmath59 are displayed in figure [ cx ] . it is not difficult to see that for small @xmath13 the top quark spin is mainly aligned with the electron beam direction , apart from corrections of order @xmath13 that depend on the production angle @xmath19 .
however , the size of these @xmath60 corrections is not really given by the expressions above , although they are correct in a formal sense .
close to threshold @xmath61 and consequently @xmath14 terms might be as large , or put another way , they should be included in the @xmath60 analysis .
some of these qcd corrections are not too difficult to implement because they are universal .
the @xmath62 and @xmath63 vertex corrections caused by the exchange of transverse gluons for example can be incorporated to the required order by simply multiplying with the following factors @xcite : @xmath64 resulting in a corresponding modification in the definition of @xmath65 .
corrections to the individual top decays can be accounted for by using the 1-loop result for @xmath6 @xcite .
there is however a type of corrections which is more difficult because it connects the different stages of the production and subsequent decay process .
i will call them `` rescattering corrections '' , the corresponding diagrams are depicted in figure [ fig : resc ] ( in principle there is a third diagram with gluon exchange between the two @xmath66-quarks , which can however be neglected to the required order @xcite ) . [ cols="^,^ " , ] in addition i should mention that @xmath67 , @xmath68 and @xmath69 could be interpreted as additional @xmath42- , @xmath43- and @xmath70-wave contributions , a viewpoint motiviated by the way they are defined and the way they enter the expressions .
figures [ plong ] and [ pperp ] show the predictions for the subleading term of the longitudinal polarization and for the transverse part , respectively .
the solid and dash - dotted lines correspond to the full result for unpolarized / fully polarized beams , the dashed and dotted lines to the @xmath60 term only .
it is quite evident that rescattering has a strong influence on the result , which is due to the fact that the @xmath71-terms are suppressed by ratios of electroweak couplings appearing in the coefficients @xmath72 and @xmath73 .
@xmath74 and @xmath75 are of comparable size as they should be because @xmath76 .
in this paper i have discussed recent results on observables in top quark pair production near threshold at @xmath0 colliders .
i have sketched the calculational procedure used to derive the predictions for the differential cross - section and top quark polarization for arbitrary beam polarization which are correct up to terms of the order @xmath77 because they also include the rescattering corrections .
the method of moments proved to be a very useful tool , which could of course also be used for other production mechanisms or far above threshold .
it turned out that rescattering corrections are not generically small and are indeed required for a consistent prediction of the components of the top quark polarization vector lying in the production plane .
they are of minor importance for the forward - backward asymmetry but might nevertheless be visible in precision studies , especially if polarized beams are used .
an important result was that there are no corrections to the normal polarization .
this component is of special interest because it is the place where non - standard cp violating interactions would manifest themselves . in connection with the leading term of p@xmath78 , which can be measured by averaging over the top direction , i.e. via @xmath79
, it provides an opportunity to determine the top quark couplings @xmath80 .
i would like to express my gratitude to the organizers of the xxxvi cracow school of theoretical physics , especially m. jeabek , for their invitation and for providing such a nice atmosphere .
i would also like to thank j.h .
khn for carefully reading the manuscript and giving useful advice . # 1#2#3_nucl .
_ * b # 1 * ( # 2 ) # 3 # 1#2#3_phys .
* b # 1 * ( # 2 ) # 3 # 1#2#3_phys . rev .
_ * d # 1 * ( # 2 ) # 3 # 1#2#3_phys .
rev . _ * # 1 * ( # 2 ) # 3 # 1#2#3_phys .
reports _ * c#1 * ( # 2 ) # 3 # 1#2#3_phys
. reports _ * # 1 * ( # 2 ) # 3 # 1#2#3_phys .
lett . _ * # 1 * ( # 2 ) # 3 # 1#2#3_sov .
j. nucl .
phys . _ * # 1 * ( # 2 ) # 3 # 1#2#3_yad .
fiz . _ * # 1 * ( # 2 ) # 3 # 1#2#3_jetp lett .
_ * # 1 * ( # 2 ) # 3 # 1#2#3_z .
_ * c # 1 * ( # 2 ) # 3 000 v.s .
fadin and v.a .
khoze , ; .
m.j . strassler and m.e .
peskin , y. sumino , k. fujii , k. hagiwara , h. murayama and c .- k .
m. jeabek , j.h .
khn and t. teubner , .
y. sumino , phd thesis , university of tokyo 1993 ( unpublished ) ; + h. murayama and y. sumino , .
m. jeabek and t. teubner , .
k. fujii , t. matsui and y. sumino , .
r. harlander , m. jeabek , j.h .
khn and t. teubner , .
r. harlander , m. jeabek , j.h .
khn and m. peter , karlsruhe preprint ttp95 - 48 , hep - ph/9604328 ; to be published in _ z. phys . _ * c*. r. barbieri , r. gatto , r. kgerler and z. kunszt , _ phys
. lett . _ * 57b * ( 1975 ) 455 , + j.h .
khn and p. zerwas , .
m. jeabek and j.h .
k. melnikov and o. yakovlev , . | top quark pair production close to threshold at a next linear collider is discussed , with special focus on the influence of rescattering corrections .
numerical results are shown for the differential cross - section , i.e. the momentum distribution and the forward backward asymmetry , and for moments of the angular distribution of leptons arising from the semileptonic top quark decay .
* ttp 96 - 36 * + * august 1996 * + * hep - ph/9608436 * |
heavy quarkonium is a multiscale system covering all regimes of quantum chromodynamics ( qcd ) , which make it an ideal place to study strong interactions @xcite . despite the success of qcd in the high energy region , due to asymptotic freedom
, the nonperturbative effect dominates at low energies and brings problems to perturbative calculation .
one tool to study this non - perturbative effect is lattice qcd .
however , due to its huge calculation work , it is still unable to calculate all the physical quantities with the current computation power .
another important approach is to develop various phenomenological models . among these phenomenological models ,
the quark model is a prominent one . under the quark model framework ,
various types of interactions have been suggested by various groups , and they have achieved many impressive successes ( see e.g. refs .
. however , these potential models can not be the whole story .
one important missing ingredient is the mechanism to generate quark - antiquark pairs which enlarge the fock space of the initial state , i.e. , the initial state contains multiquark components .
these multiquark components will change the hamiltonian of the potential model , causing mass shift and mixing between states with the same quantum numbers or directly contributing to an open channel strong decay if the initial state is above the corresponding threshold .
these consequences can be summarized as unquenched effects or coupled - channel effects . through various approaches , such as the @xmath7 model @xcite , flux - tube breaking model @xcite , microscopic decay models @xcite , @xmath8 matrix analysis @xcite , coupled - channel effects
are extensively studied in many literatures ( e.g. , refs .
@xcite ) despite these pioneering works , we find that at least two factors have the potential to jeopardize the calculations of the coupled - channel effects .
one is the widely used simple harmonic oscillator ( sho ) wave function which approximates the realistic wave function in the wave function overlap integration .
as already pointed out in our previous work @xcite , the sho approximation is not good , especially for states near thresholds , and it is essential to treat the wave functions precisely .
another is the assumption that the contribution of the excited meson loops is negligible .
take the bottomonium , which is the system we study in this paper , as an example .
two reasons may explain why this approximation is widely used .
one is that the threshold corresponding to the excited @xmath0 mesons is higher than the ground state @xmath9 threshold ; thus the coupled - channel effects is expected to be small .
another is that the calculations of the excited @xmath0 mesons contributions are more complicated , an effective method to do the calculation is still not widely known . as explained in@xcite , the gaussian expansion method plus the techniques of transformation between the cartesian and spherical basis
can , in principle , do the sophisticated calculations .
another observation is that the first excited @xmath2 wave @xmath0 meson is only around @xmath10mev heavier than the ground state @xmath0 meson .
this mass difference is less than one tenth of the @xmath0 mesons mass .
given the fact that the quantum numbers of @xmath11 are different from the ground state @xmath0 mesons , and the coupled - channel effects do rely on the quantum numbers , the suppression purely originates from the larger mass may be not as large as what has been taken for granted . the sum of all the intermediate meson loops is more than just a calculation challenge , it may also lead to profound physics . in the light sector , a series of works by geiger _
et al . _
@xcite show us that even though different intermediate meson loops contribute to the breaking of okubo - zweig - iizuka ( ozi ) rule , under some simplifications ( such as neglecting the mass difference in the denominator ) , they contribute destructively , leaving us a perfect ozi rule .
if one only sums over some of the meson loops , one may leave with a wrong conclusion .
one should not confuse their calculation with what we are going to do in this work .
they studied the flavor - changing process , such as @xmath12 .
however , we mainly focus on the mass shift and their cancellation does not happen to our case . for states below the threshold , the intermediate loops always contribute a negative mass shift . even though for the above threshold case , where the mass shift may add destructively , the mass difference does matter in the real calculation . in ref .
@xcite , under the flux - tube breaking model , geiger and isgur also showed that if one adds up all the intermediate states , the mass shift caused by coupled - channel loops does not converge when two assumptions are adopted ; i ) the meson wave function is the sho wave function , and ii ) the string length between the generated quark pairs is zero . in the @xmath7 model , this conclusion needs to be checked or recalculated , because the sho approximation is far from true , which was already stressed before and the vertex of the quark generation is different from the flux - tube breaking model .
there are also some studies of the excited meson loops for heavy quarkonium . in the charmonium sector , a lattice calculation in ref .
@xcite shows that @xmath13 and @xmath14 has small but non - negligible components of @xmath15 . in bottomonium sector , loops involving a @xmath11 meson are proven to be critical to explain the large breaking of the heavy quark spin symmetry of @xmath16 @xcite .
however , these studies focus on some specific states , and systematic studies of the excited meson loops is still missing .
to summarize , it remains to be answered or clarified whether the ground state approximation is good or not , and the general properties of the excited meson loops still need to be systematically exploited . in this work
, we try to answer these questions under the framework of the @xmath7 model .
this paper is organized as follows . in sec .
[ modelsec ] , we briefly describe the ingredients of the coupled - channel effects and the calculation techniques , including the @xmath7 model , the cornell potential model , the gaussian expansion method , and the transformations between the spherical and cartesian basis .
[ resultsec ] is devoted to the results of the higher excited @xmath0 mesons up to @xmath4 , and we explain the necessity of the form factor and the limitation of the @xmath7 model .
finally , we give a short summary in sec .
[ summary ] .
the calculation methods and tools are described in details in ref .
@xcite ; thus , we only sketch the key steps in this section
. model for the bottomonium .
@xmath17 and @xmath18 , respectively , denote the initial and final states with the same @xmath19 and @xmath9 stands for all possible @xmath0 meson pairs , including the excited @xmath0 mesons . ,
title="fig:",scaledwidth=60.0% ] + the coupled - channel effects in the @xmath7 model @xcite can be described by fig .
[ ccdiagram ] . in this model ,
the generated quark pairs have a vacuum quantum number @xmath20 . in the notation of @xmath21
, it is @xmath7 that explains the model s name .
the hamiltonian to generate the quark - antiquark pairs is expressed as h_i=2 m_q d^3x |_q _ q , where @xmath22 is the produced quark mass and@xmath23 is the dimensionless coupling constant . since the probability to generate heavier quarks
is suppressed , we use the effective strength @xmath24 in the following calculation , where @xmath25 is the constituent quark mass of the up ( or down ) quark and @xmath26 is the strange quark mass . as sketched in fig .
[ ccdiagram ] , the experimentally observed state should be a mixture of a pure quarkonium state @xmath27 and a @xmath0 meson continuum state @xmath28 . the experimentally observed state @xmath29 should be expressed as = c_0 + _ bc d^3p c_bc(p ) , which is the eigenstate of the full hamiltonian @xmath30 defined in eq .
[ fullh ] .
@xmath31 and @xmath32 stand for the normalization constants of the bare state and @xmath0 meson continuum , respectively . in order to work out @xmath33 , as a first step
, one has to solve the wave function @xmath34 for the heavy quarkonium . in this work ,
it is obtained by solving the schrdinger equation h_0=(2m_b++v(r)+v_s(r))=m_0 where @xmath35 and @xmath36 represent the mass of @xmath37 quark and the bare mass of the bottomonium , respectively . in the above equation ,
@xmath38 is the well - known cornell potential @xcite v(r)=- + r+c , [ potential ] where @xmath39 and @xmath40 stand for the strength of color coulomb potential , the strength of linear confinement and mass renormalization , respectively .
@xmath41 stands for the spin dependent interactions which restores the hyperfine or fine structures of the bottomonium , v_s(r)=(- ) + ( r ) _
b_|b + ( + ) , [ finestructure ] where @xmath42 denotes relative orbital angular momentum , @xmath43 is the total spin of the @xmath37 quark pairs and @xmath35 is the @xmath37 quark mass .
@xmath44 is the smeared delta function and is written as @xmath45 @xcite .
we treat the spin dependent term as a perturbation , and the spatial wave functions are obtained by solving schrdinger equation numerically using numerov s method @xcite .
combine the cornell potential and the dynamics of quark pair generation , we get the full hamiltonian , h = h_0+e_bc+h_i,[fullh ] where @xmath46 the @xmath47 term in eq .
[ fullh ] is mainly responsible for the mass shift .
as the name tells us , it is naturally defined as m m - m_0 [ inteqn ] and it can be obtained by solving the integral equation m=_bcd^3p .
[ mshift ] note that the @xmath48 term is added to handle the situation when @xmath49 .
in this case , @xmath50 will pick up an imaginary part ( m)=_bc p_b h_i ^2,[decay ] which is equal to one half of the the decay width . @xmath51 and @xmath52 denote the momentum and energy of @xmath0 meson , respectively .
the wave function overlap integration lies in the term h_i = _ d^3k _ 0(+_b ) _ b^*(+x _ b)_c^*(+x _ b ) @xmath35 and @xmath22 denote the @xmath37 quark and the light quark mass , respectively .
once @xmath53 is solved , the coefficient of different components can be worked out either . for states below the threshold ,
the probability of the @xmath54 component is expressed as p_b|b:=c_0 ^2=(1+_bc ls_0^dp ) ^-1 , [ 1prob ] where @xmath55 is represented as ^ls^2=d_b h_i ^2 .
the main calculation work lies in the eq .
[ overlap ] .
in order to evaluate it precisely , we use the gaussian expansion method ( gem)@xcite and the transformation between the spherical and cartesian basis to make the gem automatic applicable to the excited @xmath0 meson .
in this work , we do not intend to reproduce the spectrum and decay widths of the bottomonium family , but to study the general contributions of the excited b meson to coupled - channel effects .
there are two simplifications in our calculation .
firstly , since the constituent quark mass of @xmath56 quarks are set to be the same to get the wave functions , we further make the approximation @xmath57 . notwithstanding
this simplification , there are still 42 channels to be calculated , i.e. , @xmath58 and @xmath59 . in principle , one can still treat their mass precisely , however , based on our experiences , this simplification is quite precise , and it saves a lot of the calculation work .
second simplification comes from @xmath11 meson multiplets . from the perspective of the quark model
, there are four @xmath60 wave @xmath0 mesons . nevertheless , the predicted @xmath61 and one of the @xmath62 are still not experimentally observed .
as far as the coupled - channel is concerned , the missing of these two states is not a real problem . because the mass of @xmath37 quark is very large , it is safe to assume that the heavy quark spin symmetry does not break . under this limit ,
it is reasonable to set the mass of @xmath63 and missing partner of @xmath64 to be the same as @xmath64 .
there is a mixing between @xmath65 and @xmath66 to form the experimentally observed @xmath64@xcite .
however , since we sum up all the possible combinations of @xmath0 mesons , mixing between the two @xmath67 mesons can also be neglected .
as far the mass of the intermediate states are well measured , we use the value from the particle data group @xcite .
the coupled - channels are much more involved than the case , which only includes the ground state @xmath0 mesons . for illustration purposes
, we sum up the contributions of the states in the same multiplet and @xmath68 flavors and their charge conjugate partners , classifying the channels into three groups ( see fig [ shiftcompare ] ) , e.g. , @xmath69 stands for ground state @xmath70 , and @xmath71 mesons .
also , in the rest of this paper , we define @xmath72 to represent the sum of the mass shift due to all the possible combinations of @xmath17 and @xmath73 wave @xmath0 mesons .
the parameters are given in table [ paratab ] , which are the same as our previous work @xcite .
.the parameters used in our calculation . [ cols="^,^,^,^,^,^,^,^,^,^ " , ] the mass shift @xmath50 of both @xmath69 and @xmath3 mesons are depicted in fig .
[ shiftcompare ] . in the rest of this section
, we are going to discuss some interesting structures of the plots .
when the states are far below the threshold , the mass shift is always negative [ as indicated in eq .
[ mshift ] ] , and closer to the threshold means an increase of @xmath74 .
when the mass goes higher , the magnitude of @xmath50 is no longer monotonically increasing but oscillating , which reflects the node structure of the wave functions .
this phenomenon happens to both cases whether @xmath11 mesons contributions are considered or not .
a specific example would help to clarify this point .
for @xmath16 , which we treat it as @xmath5 , @xmath75 is generally the largest but there is a big dip around @xmath76 gev with parameters in tab .
[ paratab ] ; however , in ref.@xcite s parameters , this channel s contribution always grows . in these cases
, it is difficult to make a solid conclusion about the spectrum behavior , i.e. , the spectrum is sensitive to the potential parameters .
this conclusion also agrees with ref .
@xcite .
the most notable impact of @xmath11 family is the unexpected large contributions to the mass shift , and generally , @xmath75 is the largest .
even for @xmath77 , which is far below the threshold , @xmath78 is 14 times larger than @xmath79 .
compared with the case that only considers the ground states @xmath0 mesons contributions , threshold effect is more clearly reflected when the excited @xmath0 mesons is included , because open channels due to different multiplets are well separated .
the first open bottom threshold is @xmath80gev .
when the mass approaches to this value , the denominator @xmath81 approaches to 0 in eq .
[ mshift ] ; thus , there is an enhancement of @xmath50 from the @xmath82 channel . from the @xmath83 case in fig .
[ shiftcompare ] , one can clearly see the sharp increase of @xmath79 between @xmath84 and @xmath85gev , while the slop of the other channels does not change much because their threshold is 450mev larger .
the threshold effect is broken in some degree by the nodes of the wave functions .
the peaks and valleys of the wave function are more likely to cancel with each other for higher excited states , leaving a relatively small slope of @xmath50 .
one can compare the result of @xmath86 with @xmath83 verify this conclusion .
of different coupled - channels for @xmath87 . the results with parameters in tab .
[ paratab ] and the results recalculated with parameters of ref .
@xcite are represented by thick and thin curves , respectively .
@xmath88 , and @xmath89 are represented by black dot - dashed , red dashed and blue solid curves , respectively .
, title="fig : " ] + of different coupled - channels for some selected representatives of the @xmath90 families . the results with parameters in tab .
[ paratab ] , and results recalculated with parameters of ref .
@xcite are represented by thick and thin curves , respectively .
@xmath88 , and @xmath89 are represented by black dot - dashed , red dashed , and blue solid curves , respectively . here
we omit the results of @xmath91 , and @xmath92 when @xmath93 , because these results are quite similar to those of @xmath77 or @xmath94 .
, title="fig:",scaledwidth=90.0% ] + we need to stress that @xmath95 is not the only family who couples strongly with @xmath96 loop . as shown in fig [ shiftcompare2 ] , all the families of @xmath90 share these general properties .
an direct consequence of the large @xmath75 is that the parameters only considering ground state @xmath0 mesons to reproduce the experimental data are somewhat incomplete or even misleading .
this conclusion is independent of the parameters to some degree , since the results are based on two different sets of parameters ( refs .
@xcite and @xcite ) , both give large @xmath75 .
it seems feasible that to fit the spectrum , one may tune the parameters , such as the mass renormalization term @xmath40 in eq .
[ potential ] , to absorb this unexpected large mass shift , or take a further step , weaken the confine potential to reflect the fact that closer to the threshold generally means more suppression of the mass .
this could be a possible way out , if someone only focuses on the spectrum .
however , as pointed out in our previous work @xcite , the mass shift only reveals one aspect of the coupled - channel effects .
to be specific , the big contribution of @xmath3 not only brings us the large mass shift @xmath50 , but also the large fraction of meson pairs ( or equivalently , small @xmath97 ) .
this renormalization of the wave function can not be accommodated in the framework of the potential model .
the large contribution of @xmath11 mesons reminds us that it is essential to do the _ ab initio _ calculations of coupled - channel effects .
however , before really carrying out these calculations , one may naturally ask this question , how large are the contributions of the states beyond @xmath11 or how to evaluate all the intermediate meson loops ? " it is very difficult to offer a complete answer to this question .
nevertheless , it turns out that we can still estimate the contributions up to @xmath4 .
furthermore , by analyzing the mass shift behavior carefully , we can draw some model independent conclusions . first of all , we are trying to precisely evaluate the loops .
of course , since the @xmath4 mass and its wave functions are not known , the calculation has to be model dependent , which is inevitable .
so far , the @xmath4 meson has not been experimentally well determined and its mass is model dependent . for our purposes , it is sufficient to take an average of several theoretical estimations in refs .
@xcite , and for the consistency treatment of the wave functions , the @xmath4 wave functions are still deduced from our parameters .
we define the ratio r:= [ rdefine ] to tell how large are the contributions of @xmath4 .
as will be explained below , in the calculation when @xmath4 are involved , we adopt the gaussian form factor which reveals the size of the generated quark pairs ( see , e.g. , ref .
this form factor modifies @xmath98 to be h_i = _
d^3k e^-2 r^2 k^2/3 _ 0(+_b ) _ b^*(+x _ b)_c^*(+x _ b ) || y_1^m(_,_),[overlapff ] in this work , we fit @xmath99 to be @xmath100 fm , which minimizes @xmath101 of @xmath77 .
this value slightly larger than the value @xmath102 fm used in ref .
@xcite . from fig .
[ 2sratiocompare ] , one can clearly observe that the form factor is crucial to suppress the @xmath4 s contributions . without it ,
@xmath4 can contribute an additional @xmath103 of the @xmath50 .
this deviation is not negligible if one wants to make a precise fit of the spectrum .
however , with a form factor @xmath104 fm , the @xmath50 of @xmath4 is suppressed to less than @xmath105 .
the previous question is partly answered , the form factor is indispensable if one needs to suppress the contributions of @xmath4 . however , with the increase of @xmath106 in @xmath87 family , the form factor works less and less efficiently to suppress @xmath101 .
[ defined in eq .
[ rdefine ] ] of the @xmath107 family .
results corresponding to @xmath108 ( no form factors ) , @xmath102fm ( taken from @xcite ) , and @xmath104fm ( our best fit ) are denoted by black dot - dashed , blue dashed , and red solid curves , respectively .
, title="fig : " ] + another effect of the form factor is that it adds more peaks to the oscillation of @xmath50 .
this result has a pictorial explanation .
since the form factor effectively serves as a cutoff , this term will generally suppress every channel s contribution .
the key point is that different channels are suppressed by a different magnitude . with the increase of the nodes of the wave functions of bottomonia and @xmath0 mesons , at some specific energy point , @xmath4
s relative contributions may increase . of @xmath5 and @xmath6 with different form factors .
solid curves represent the denominator of eq .
[ rdefine ] , i.e. , @xmath109 , and dashed curves represent the numerator , i.e. @xmath110 . @xmath50 corresponding to @xmath108 , @xmath102fm , and @xmath104fm are denoted by black , blue , and red curves , respectively .
the value of the @xmath111 can not compare within different form factors because the @xmath7 coupling constants are not fit to reproduce experimental data .
, title="fig : " ] + clearly the sharp peaks of @xmath5 and @xmath6 deserves a special analysis .
both of them locate around @xmath112gev .
given the fact that the wave functions of @xmath5 and @xmath6 are quite different , yet they both have the same sharp peaks at the same energy points , it turns out to be more than just a coincidence .
@xmath101 is convenient to estimate @xmath4 s relative contributions ; however , it can not tell whether the peak is due to the suppression of the denominator or the increase of the numerator .
the complete information is encoded in @xmath50 itself . in fig .
[ 5s6sshift ] , we show @xmath50s for @xmath5 and @xmath6 with different form factors . from fig .
[ 5s6sshift ] , one can see that the enhancement of @xmath101 around @xmath112 gev is in fact the bigger suppression of @xmath113 , [ or more precisely , the suppression of @xmath75 ] .
we have scanned @xmath99 in quite a wide range @xmath115 fm , which all show that the form factor does not shift the positions of peaks or valleys of @xmath50 . around this region
, the form factor fails to suppress the @xmath4 s contribution is somewhat a coincidence , because this suppression does not happen to @xmath5 and @xmath6 with ref.@xcite s parameters ; their @xmath75 always goes up . even in the low energy range where all the coupled - channels contribute a negative @xmath50 ,
no matter whether the form factor is considered or not , the @xmath4 still contribute around @xmath116 of the total mass shift for @xmath5 and @xmath6 .
in other words , the form factor fails to suppress @xmath4 s contributions .
this strongly indicates that only adding the gaussian form factor is not adequate to result in a complete calculation of the coupled - channel effects under the @xmath7 framework .
one may argue that the coupled channel effect can be absorbed into the smeared potential which returns wave functions broader than those of the cornell model .
these broad wave functions indeed somewhat suppress the higher excited states contributions .
however , based on our experience , changing the size of the wave functions is also not able to solve the convergence issue .
we conclude this section with some comments on the @xmath7 model . in the classical @xmath7 model , the suppression due to the natural cutoff from wave functions and the increase of the denominator of eq .
[ mshift ] are too weak .
even with the remedy of the gaussian form factor , if one does not choose the cutoff @xmath99 carefully , one will enhance instead of suppressing the @xmath4 contribution .
compared with other models , such as the microscopic decay model or the flux - tube breaking model , which have rich microscopic details of the decay vertices , the @xmath7 model replaces these fine structures as an overall coupling constant @xmath23
. this approximation may be not appropriate ( see e.g. ref .
@xcite ) . in our case
, we show that this approximation leads to the bad convergence of the sum of the excited meson loops for highly excited bottomonia .
the rich structure of the quark pair generation vertices may help to solve the convergence issue .
for example , the hamiltonian of the flux - tube breaking model suppresses the generation of the farther quarks , one will expect that it is harder to generate excited intermediate states . + as illustrated in fig .
[ ccofftb ] , contributions from the regions that are farther from the flux - tube line can be dropped , as a result , the excited mesons contribution is naturally suppressed . without this dynamical suppression , it is more difficult to suppress @xmath4 s contributions even with the modified version of the @xmath7 model , where the form factor is added .
in this paper , under the @xmath7 model framework , we explicitly calculated the excited @xmath0 mesons contributions to the coupled - channel effects for the bottomonium .
we reveal the fact that compared to the ground state @xmath0 mesons , contributions from @xmath3 mesons are generally the largest . up to this partial wave ,
it is necessary to do the _ ab initio _ calculations of the coupled - channel effects .
when we push the calculation beyond @xmath3 , we find some fundamental difficulties of the @xmath7 model .
even with the carefully chosen form factor , it still can not efficiently suppress intermediate state contributions of higher partial waves . since we do not fit our parameters with experimental data , and we have exploited several different sets of parameters , we have enough reasons to believe that the difficulties are independent of the wave functions or the potential models .
we suggest that an efficient suppression mechanism such as a dynamically suppression is needed to evaluate the coupled - channel effects .
how to effectively sum up all the intermediate loops of coupled - channels still remains to be an open issue .
yu lu is grateful to professor feng - kun guo for various discussions and suggestions and to mr .
bao - dong sun and professor yu - bing dong for lending their cluster computing resources .
m. naeem anwar acknowledges several discussions with professor cristoph hanhart during the hadron physics summer school ( hpss 2016 ) at forschungszentrum jlich .
this work is supported by the national natural science foundation of china under grants no .
11261130311 ( crc110 by dfg and nsfc ) and no .
m. naeem anwar is supported by cas - twas president s fellowship for international ph.d .
students .
n. brambilla _ et al .
_ , `` heavy quarkonium : progress , puzzles , and opportunities , '' http://dx.doi.org/10.1140/epjc/s10052-010-1534-9[_eur .
phys . j. _
* c71 * ( 2011 ) 1534 ] , http://arxiv.org/abs/1010.5827[arxiv:1010.5827 [ hep - ph ] ] .
a. martin , `` a simultaneous fit of @xmath54,@xmath117 , @xmath118 , ( bcs pairs ) and @xmath119 spectra , '' http://dx.doi.org/10.1016/0370-2693(81)90617-1[_phys .
lett . _ * b100 * ( 1981 ) 511 ] .
r. a. bertlmann and a. martin , `` inequalities on heavy quark - antiquark systems , '' http://dx.doi.org/10.1016/0550-3213(80)90279-5[_nucl .
phys . _ * b168 * ( 1980 ) 111136 ] .
e. eichten , k. gottfried , t. kinoshita , k. d. lane , and t .- m .
yan , `` charmonium : the model , '' http://dx.doi.org/10.1103/physrevd.17.3090 , 10.1103/physrevd.21.313[_phys .
rev . _ * d17 * ( 1978 ) 3090 ] .
[ erratum : phys .
rev.d21,313(1980 ) ] . w. buchmuller and s. h. h. tye , `` quarkonia and quantum chromodynamics , '' http://dx.doi.org/10.1103/physrevd.24.132[_phys .
rev . _ * d24 * ( 1981 ) 132 ] .
s. godfrey and n. isgur , `` mesons in a relativized quark model with chromodynamics , '' http://dx.doi.org/10.1103/physrevd.32.189[_phys
* d32 * ( 1985 ) 189231 ] . l. micu , `` decay rates of meson resonances in a quark model , '' http://dx.doi.org/10.1016/0550-3213(69)90039-x[_nucl .
phys . _ * b10 * ( 1969 ) 521526 ] .
a. le yaouanc , l. oliver , o. pene , and j. c. raynal , `` naive quark pair creation model of strong interaction vertices , '' http://dx.doi.org/10.1103/physrevd.8.2223[_phys .
_ * d8 * ( 1973 ) 22232234 ] .
a. le yaouanc , l. oliver , o. pene , and j. c. raynal , `` naive quark pair creation model and baryon decays , '' http://dx.doi.org/10.1103/physrevd.9.1415[_phys .
* d9 * ( 1974 ) 14151419 ] .
h. g. blundell , _ meson properties in the quark model : a look at some outstanding problems_. phd thesis , carleton u. , 1996 .
http://arxiv.org/abs/hep-ph/9608473[arxiv:hep-ph/9608473 [ hep - ph ] ] .
t. barnes and e. s. swanson , `` hadron loops : general theorems and application to charmonium , '' http://dx.doi.org/10.1103/physrevc.77.055206 [ _ phys .
rev . _ * c77 * ( 2008 ) 055206 ] , http://arxiv.org/abs/0711.2080[arxiv:0711.2080 [ hep - ph ] ] .
j. carlson , j. b. kogut , and v. r. pandharipande , `` a quark model for baryons based on quantum chromodynamics , '' http://dx.doi.org/10.1103/physrevd.27.233[_phys .
* d27 * ( 1983 ) 233 ] . j. carlson , j. b. kogut , and v. r. pandharipande , `` hadron spectroscopy in a flux tube quark model , '' http://dx.doi.org/10.1103/physrevd.28.2807[_phys .
_ * d28 * ( 1983 ) 2807 ] .
r. kokoski and n. isgur , `` meson decays by flux - tube breaking , '' http://dx.doi.org/10.1103/physrevd.35.907[_phys .
d _ * 35 * no . 3 , ( feb . ,
1987 ) 907933 ] .
http://link.aps.org/doi/10.1103/physrevd.35.907 .
p. geiger and n. isgur , `` the quenched approximation in the quark model , '' http://dx.doi.org/10.1103/physrevd.41.1595[_phys .
_ * d41 * ( 1990 ) 1595 ] . f. e. close and p. r. page , `` the production and decay of hybrid mesons by flux tube breaking , '' http://dx.doi.org/10.1016/0550-3213(95)00085-7[_nucl .
_ * b443 * ( 1995 ) 233254 ] , http://arxiv.org/abs/hep-ph/9411301[arxiv:hep-ph/9411301 [ hep - ph ] ] .
e. eichten , k. gottfried , t. kinoshita , k. d. lane , and t .- m .
yan , `` charmonium : comparison with experiment , '' http://dx.doi.org/10.1103/physrevd.21.203[_phys . rev . _ * d21 * ( 1980 ) 203 ] .
e. s. ackleh , t. barnes , and e. s. swanson , `` on the mechanism of open flavor strong decays , '' http://dx.doi.org/10.1103/physrevd.54.6811 [ _ phys .
rev . _ * d54 * ( 1996 ) 68116829 ] , http://arxiv.org/abs/hep-ph/9604355[arxiv:hep-ph/9604355 [ hep - ph ] ] .
t. barnes , s. godfrey , and e. s. swanson , `` higher charmonia , '' http://dx.doi.org/10.1103/physrevd.72.054026[_phys .
_ * d72 * ( 2005 ) 054026 ] , http://arxiv.org/abs/hep-ph/0505002[arxiv:hep-ph/0505002 [ hep - ph ] ] . v. baru , c. hanhart , yu . s. kalashnikova , a. e. kudryavtsev , and a. v. nefediev , `` interplay of quark and meson degrees of freedom in a near - threshold resonance , '' http://dx.doi.org/10.1140/epja/i2010-10929-7[_eur . phys . j. _ * a44 * ( 2010 ) 93103 ] , http://arxiv.org/abs/1001.0369[arxiv:1001.0369 [ hep - ph ] ] .
c. hanhart , yu .
s. kalashnikova , and a. v. nefediev , `` interplay of quark and meson degrees of freedom in a near - threshold resonance : multi - channel case , '' http://dx.doi.org/10.1140/epja/i2011-11101-9[_eur .
phys . j. _ * a47 * ( 2011 ) 101110 ] , http://arxiv.org/abs/1106.1185[arxiv:1106.1185 [ hep - ph ] ] . f. k. guo , c. hanhart , yu .
s. kalashnikova , p. matuschek , r. v. mizuk , a. v. nefediev , q. wang , and j. l. wynen , `` interplay of quark and meson degrees of freedom in near - threshold states : a practical parametrization for line shapes , '' http://dx.doi.org/10.1103/physrevd.93.074031 [ _ phys .
d93 * no . 7 , ( 2016 ) 074031 ] , http://arxiv.org/abs/1602.00940[arxiv:1602.00940 [ hep - ph ] ] . i. k. hammer , c. hanhart , and a. v. nefediev , `` remarks on meson loop effects on quark models , '' http://dx.doi.org/10.1140/epja/i2016-16330-8 [ _ eur .
* a52 * no .
11 , ( 2016 ) 330 ] , http://arxiv.org/abs/1607.06971[arxiv:1607.06971 [ hep - ph ] ] .
k. heikkila , s. ono , and n. a. tornqvist , `` heavy @xmath117 and @xmath54 quarkonium states and unitarity effects , '' http://dx.doi.org/10.1103/physrevd.29.110[_phys .
_ * d29 * ( 1984 ) 110 ] .
[ erratum : phys .
rev.d29,2136(1984 ) ] .
s. ono and n. a. tornqvist , `` continuum mixing and coupled channel effects in @xmath120 and @xmath121 quarkonium , '' http://dx.doi.org/10.1007/bf01558041[_z .
phys . _ * c23 * ( 1984 ) 59 ] . n. a. tornqvist and p. zenczykowski , `` ground state baryon mass splittings from unitarity , '' http://dx.doi.org/10.1103/physrevd.29.2139[_phys
. rev . _ * d29 * ( 1984 ) 2139 ] . s. ono , a. i. sanda , and n. a. tornqvist , `` @xmath0 meson production between the @xmath107 ( 4s ) and @xmath107 ( 6s ) and the possibility of detecting @xmath122 mixing , '' http://dx.doi.org/10.1103/physrevd.34.186[_phys . rev . _
* d34 * ( 1986 ) 186 ] .
s. ono , a. i. sanda , n. a. tornqvist , and j. lee - franzini , `` where are the @xmath122 mixing effects observable in the @xmath107 region ?
, '' http://dx.doi.org/10.1103/physrevlett.55.2938[_phys .
* 55 * ( 1985 ) 2938 ]
. i. v. danilkin and yu .
a. simonov , `` channel coupling in heavy quarkonia : energy levels , mixing , widths and new states , '' http://dx.doi.org/10.1103/physrevd.81.074027[_phys .
rev . _ * d81 * ( 2010 ) 074027 ] , http://arxiv.org/abs/0907.1088[arxiv:0907.1088 [ hep - ph ] ] .
y. lu , m. n. anwar , and b .- s .
zou , `` coupled - channel effects for the bottomonium with realistic wave functions , '' http://dx.doi.org/10.1103/physrevd.94.034021[_phys
_ * d94 * no . 3 , ( 2016 ) 034021 ] , http://arxiv.org/abs/1606.06927[arxiv:1606.06927 [ hep - ph ] ] .
p. geiger and n. isgur , `` how the okubo - zweig - iizuka rule evades large loop corrections , '' http://dx.doi.org/10.1103/physrevlett.67.1066[_phys .
* 67 * ( 1991 ) 10661069 ] .
p. geiger and n. isgur , `` reconciling the ozi rule with strong pair creation , '' http://dx.doi.org/10.1103/physrevd.44.799[_phys .
rev . _ * d44 * ( 1991 ) 799808 ] .
p. geiger and n. isgur , `` when can hadronic loops scuttle the ozi rule ?
, '' http://dx.doi.org/10.1103/physrevd.47.5050[_phys .
_ * d47 * ( 1993 ) 50505059 ] .
g. s. bali , s. collins , and c. ehmann , `` charmonium spectroscopy and mixing with light quark and open charm states from @xmath123=2 lattice qcd , '' http://dx.doi.org/10.1103/physrevd.84.094506[_phys .
* d84 * ( 2011 ) 094506 ] , http://arxiv.org/abs/1110.2381[arxiv:1110.2381 [ hep - lat ] ] .
guo , u .- g .
meiner , and c .- p .
shen , `` enhanced breaking of heavy quark spin symmetry , '' http://dx.doi.org/10.1016/j.physletb.2014.09.043[_phys .
_ * b738 * ( 2014 ) 172177 ] , http://arxiv.org/abs/1406.6543[arxiv:1406.6543 [ hep - ph ] ] .
li , c. meng , and k .-
chao , `` coupled - channel and screening effects in charmonium spectrum , '' http://dx.doi.org/10.1103/physrevd.80.014012[_phys . rev . _
* d80 * ( 2009 ) 014012 ] , http://arxiv.org/abs/0904.4068[arxiv:0904.4068 [ hep - ph ] ] .
b. numerov , `` note on the numerical integration of @xmath124 , '' http://dx.doi.org/10.1002/asna.19272301903[_astron .
* 230 * ( 1927 ) 359 ] .
e. hiyama , y. kino , and m. kamimura , `` gaussian expansion method for few - body systems , '' http://dx.doi.org/10.1016/s0146-6410(03)90015-9[_prog
phys . _ * 51 * ( 2003 ) 223307 ] .
collaboration , c. patrignani _ et al .
_ , `` review of particle physics , '' http://dx.doi.org/10.1088/1674-1137/40/10/100001[_chin .
* c40 * no . 10 , ( 2016 ) 100001 ] .
liu and g .- j .
ding , `` bottomonium spectrum with coupled - channel effects , '' http://dx.doi.org/10.1140/epjc/s10052-012-1981-6[_eur .
phys . j. _ * c72 * ( 2012 ) 1981 ] , http://arxiv.org/abs/1105.0855[arxiv:1105.0855 [ hep - ph ] ] . n. a. tornqvist , `` the @xmath107 ( 5s ) mass and @xmath125 , @xmath126 , @xmath127 as sensitive tests of the unitarized quark model , '' http://dx.doi.org/10.1103/physrevlett.53.878[_phys . rev . lett . _
* 53 * ( 1984 ) 878 ] .
j. zeng , j. w. van orden , and w. roberts , `` heavy mesons in a relativistic model , '' http://dx.doi.org/10.1103/physrevd.52.5229[_phys .
rev . _ * d52 * ( 1995 ) 52295241 ] , http://arxiv.org/abs/hep-ph/9412269[arxiv:hep-ph/9412269 [ hep - ph ] ] .
m. di pierro and e. eichten , `` excited heavy - light systems and hadronic transitions , '' http://dx.doi.org/10.1103/physrevd.64.114004 [ _ phys .
rev . _ * d64 * ( 2001 ) 114004 ] , http://arxiv.org/abs/hep-ph/0104208[arxiv:hep-ph/0104208 [ hep - ph ] ] .
d. ebert , r. n. faustov , and v. o. galkin , `` heavy - light meson spectroscopy and regge trajectories in the relativistic quark model , '' http://dx.doi.org/10.1140/epjc/s10052-010-1233-6[_eur .
phys . j. _
* c66 * ( 2010 ) 197206 ] , http://arxiv.org/abs/0910.5612[arxiv:0910.5612 [ hep - ph ] ] . t. a. lahde , c. j. nyfalt , and d. o. riska , `` spectra and m1 decay widths of heavy light mesons , '' http://dx.doi.org/10.1016/s0375-9474(00)00154-8[_nucl . phys . _ * a674 * ( 2000 ) 141167 ] , http://arxiv.org/abs/hep-ph/9908485[arxiv:hep-ph/9908485 [ hep - ph ] ] .
pan , y .- y .
wang , e. wang , and d .-
li , `` excited bottom and bottom - strange mesons in the quark model , '' http://dx.doi.org/10.1103/physrevd.94.074012[_phys .
rev . _ * d94 * no . 7 , ( 2016 ) 074012 ] , http://arxiv.org/abs/1607.02812[arxiv:1607.02812 [ hep - ph ] ] . b. silvestre - brac and c. gignoux , `` unitary effects in spin orbit splitting of p wave baryons , '' http://dx.doi.org/10.1103/physrevd.43.3699[_phys .
_ * d43 * ( 1991 ) 36993708 ] .
w. roberts and b. silvestre - brac , `` meson decays in a quark model , '' http://dx.doi.org/10.1103/physrevd.57.1694[_phys .
* d57 * ( 1998 ) 16941702 ] , http://arxiv.org/abs/hep-ph/9708235[arxiv:hep-ph/9708235 [ hep - ph ] ] . j. ferretti and e. santopinto , `` higher mass bottomonia , '' http://dx.doi.org/10.1103/physrevd.90.094022[_phys .
* d90 * no . 9 , ( 2014 ) 094022 ] , http://arxiv.org/abs/1306.2874[arxiv:1306.2874 [ hep - ph ] ] .
a. m. badalian , v. d. orlovsky , and yu . a. simonov , `` microscopic study of the string breaking in qcd , '' http://dx.doi.org/10.1134/s1063778813080048[_phys . atom .
* 76 * ( 2013 ) 955964 ] , http://arxiv.org/abs/1210.4674[arxiv:1210.4674 [ hep - ph ] ] . | we study the excited @xmath0 mesons contributions to the coupled - channel effects under the framework of the @xmath1 model for the bottomonium .
contrary to what has been widely accepted , the contributions of @xmath2 wave @xmath0 mesons are generally the largest , and to some extent , this result is independent of the potential parameters .
we also push the calculation beyond @xmath3 and carefully analyze the contributions of @xmath4 .
a form factor is a key ingredient to suppress the contributions of @xmath4 for low lying bottomonia .
however , this suppression mechanism is not efficient for highly excited bottomonia , such as @xmath5 and @xmath6 .
we give explanations why this difficulty happens to the @xmath1 model and suggest analyzing the flux - tube breaking model for the full calculation of coupled - channel effects . |
it is generally believed that molecular clouds are turbulent ( see the reviews of @xcite ) .
observationally ( e.g. * ? ? ?
* ; * ? ? ?
* ) this turbulence appears to be highly supersonic with rms mach numbers of anything up to 20 or more .
it is also thought that the alfvnic mach number is not significantly less than one , and may be much greater ( e.g. * ? ? ?
the amplitude of this turbulence makes it likely to be an important ingredient in both the dynamics of molecular clouds and the process of star formation @xcite and hence developing an understanding of this phenomenon is of considerable interest .
many authors have addressed the issue of mhd turbulence in the context of molecular clouds using both the ideal mhd approximation @xcite and , more recently , various flavours of non - ideal mhd @xcite .
three dimensional mhd turbulence involves the transfer of energy from an energy injection scale to ever smaller scales until the dissipation length scale of the system is reached .
given that in molecular clouds the lengthscales at which multifluid effects become important are much larger than the viscous lengthscale , it is inevitable that these effects will have an impact on the energy cascade .
when ambipolar diffusion is included in mhd turbulence simulations it has been found that it causes greater temporal variability in the turbulence statistics @xcite .
although clearly of lesser significance @xcite , the hall effect is capable of inducing topological changes in the magnetic field which are quite distinct to any influence caused by ambipolar diffusion . in particular , it introduces a handedness into the flow which is of particular interest when considering the conversion of kinetic energy into magnetic energy .
researchers working on reconnection and the solar wind have studied the hall effect in the context of turbulence and found that , although the overall decay rate appears not to be affected , the usual coincidence of the magnetic and velocity fields seen in ideal mhd does not occur at small scales @xcite .
* hereafter paper i ) performed simulations of decaying , non - ideal mhd , molecular cloud turbulence incorporating parallel resistivity , the hall effect and ambipolar diffusion .
they found that the hall effect has surprisingly little impact on the behaviour of the turbulence : it does not affect the energy decay at all and has very limited impact on the power spectra of any of the dynamical variables , with the exception of the magnetic field at very short length - scales .
* hereafter paper ii ) extended the results of paper i to properly multifluid mhd turbulent decay .
a system of three fluids , 1 neutral and two charged species , was simulated and , intriguingly , it appears that for the chosen system a full multifluid simulation is unnecessary and simple non - ideal mhd with temporally and spatially constant resistivities can give quite reliable results . in this paper
we use the hydra code @xcite to investigate _ driven _ , isothermal , multifluid mhd turbulence in a system consisting of 4 fluids ( 1 neutral and 3 charged ) .
the aim of this work is to determine the influence of multifluid effects on the behaviour of driven turbulence in molecular clouds .
an added benefit of our properly multifluid approach is that we can self - consistently deduce the behaviour of the charged species and , in principle , make links with observations such as those of @xcite .
the structure of this paper is as follows : in section [ sec : model ] we outline the equations and numerical method ; in section [ sec : initial - conditions ] we specify our parameters , initial conditions and how we drive the turbulence in our simulations ; section [ sec : resolution - study ] contains the results of a resolution study demonstrating that the results we present are reasonably well converged ; section [ sec : results ] contains the results and analysis of our simulations and finally we draw conclusions from our work in section [ sec : conclusions ] .
we note here that we omit a discussion of the differences in line widths between the charged and the neutral species .
the results associated with these differences require detailed discussion in their own right and are the subject of a forthcoming paper .
molecular clouds are , to a great extent , weakly ionised systems .
this allows us to make several simplifying assumptions , as follows : * the bulk flow velocity _ is _ the neutral velocity * the majority of collisions experienced by each charged species occur with neutrals * the charged species inertia is unimportant * the charged species pressure gradient is unimportant making these assumptions allows us to derive a relationship between the electric field and the current density .
this generalised ohm s law ( e.g. * ? ? ?
* ; * ? ? ?
* ) allows us to avoid having to calculate the electric field explicitly from the charge distribution .
once we have our generalised ohm s law , the system of equations for our weakly ionised flow is as follows ( e.g. * ? ? ?
* ) @xmath0 where @xmath1 denotes the species ( with @xmath2 denoting the neutral species ) , @xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath7 , @xmath8 and @xmath9 are the mass density of species @xmath1 , the velocity of species @xmath1 , the isothermal sound speed , the magnetic field , the identity matrix , the current density and the electric field in the neutral rest frame , respectively . the parameters @xmath10 and @xmath11 are the charge - to - mass ratios and the collision coefficient with the neutrals for species @xmath1 .
these equations describe conservation of mass for all species , the neutral momentum equation , the induction equation , force balance for the charged species , the absence of magnetic monopoles , faraday s law and charge neutrality . in the work presented here we set @xmath12 .
the electric field in the frame of the fluid , @xmath9 , is calculated from the generalised ohm s law for weakly ionised fluids ( e.g. @xcite ) and is given by @xmath13 where @xmath14 using the definitions @xmath15 , @xmath16 , @xmath17 , where @xmath18 , @xmath19 , @xmath20 .
@xmath21 , @xmath22 and @xmath23 are the parallel , hall and ambipolar resistivities respectively and are given by @xmath24 with the conductivities given by @xmath25 where @xmath26 is the hall parameter for species @xmath1 and is given by @xmath27 equations ( [ mass ] ) ( [ charge_neutrality ] ) are a nonlinear system of pdes and , in general , must be tackled using simulations . with modern techniques
the system is relatively straightforward , with the exception of dealing with the right - hand side of equation ( [ b_eqn ] ) .
the method used by hydra to integrate these equations has been elucidated previously in the literature ( e.g. * ? ? ?
* ; * ? ? ?
* ) so we only give a brief recapitulation of this here and refer the reader to these papers for a more in - depth discussion .
the system is solved using an operator - split approach with three operators : * the equations governing the neutrals are solved using a standard second order , shock capturing godunov scheme .
the induction equation is also solved as part of this operator , but without incorporation of the non - ideal effects .
the divergence of the magnetic field is controlled using the method of @xcite . *
the magnetic diffusion terms are evaluated and applied using the super - time - stepping method ( for the ambipolar and parallel resistivity terms ) and the hall diffusion scheme ( for the hall term ) . *
the velocities of the charged species are determined from equation ( [ charged_mom ] ) and the distribution of their densities are updated using a second order upwind scheme .
the whole system of operators is stepped forward in time using strang operator splitting in order to ensure second order accuracy in time . since molecular cloud material has very low kinematic viscosity ( e.g. * ? ? ?
* ) , flows , these systems have a high prandtl number and thus we can consider them to be effectively inviscid . as with all higher - order schemes , our simulations incorporate artificial viscosity which smears discontinuities over around 3 5 grid zones .
this artificial viscosity length scale is much less than that over which the multifluid effects act ( see , for example , figure [ fig : v - spec - res ] ) and thus our multifluid effects dominate at all scales above around 3 5 grid zones . formally , the prandtl number of the highest resolution simulations in this work is around 15 .
the method is explicit and scales extremely well on massively parallel systems , displaying strong scaling from 8192 up to 294,912 cores with 73% efficiency on the jugene bluegene / p system in jlich , germany .
we run our turbulence simulation in a cube of side 0.2pc with periodic boundary conditions on all faces .
the density is initially uniform and has the value of @xmath28@xmath29 , assuming that the average neutral particle has a mass of 2.33@xmath30 ( where @xmath30 is the mass of the proton ) .
the magnetic field is also initially uniform of strength 20@xmath31 g respectively .
the ( isothermal ) sound - speed is chosen to be @xmath32cms@xmath33 or , equivalently , the temperature is 10k .
the four fluids in the system are initially at rest . these parameters were chosen to be representative of molecular clouds on scales of 0.2pc and are based on the data in table 1 of @xcite . the grid used is uniform and made up of @xmath34 grid points with @xmath35 .
we emphasise here that this work is directed at understanding the nature of multifluid mhd turbulence .
a physical system with the properties just described will have a mass just below the jeans mass and so may well be subject to dynamically important self - gravity .
however , it seems reasonable that the optimal way of approaching understanding the behaviour of molecular clouds is to first attempt to understand the turbulence , and then to introduce self - gravity .
introducing all the physical effects at once runs the risk of producing results which , while possibly valid , are very difficult to interpret .
the properties of the charged species are based on the data in @xcite .
we make the approximation that the dust density is 1% that of the neutrals .
the metal ions are assumed to be singly ionised and to have an average mass of 24@xmath30 and a mass density of @xmath36g@xmath29 . the electron density is then chosen to ensure local charge neutrality .
this set - up results in an ionisation fraction of approximately @xmath37 and the system is therefore clearly weakly ionised . in order to drive turbulence
we add velocity increments to the neutral velocity at each time - step .
the incremental velocity field , @xmath38 , is defined in a manner similar to @xcite .
each component of @xmath38 is generated from a set of waves with wave numbers , @xmath39 , @xmath40 .
the amplitudes of these waves are drawn from a gaussian random distribution with mean 1.0 and deviation 0.33 while the phases of the waves are drawn from a uniform distribution between 0 and @xmath41 .
the waves are chosen such that the forcing is solenoidal .
it is known that the precise nature of the forcing will influence the turbulence statistics ( e.g. * ? ? ?
in this paper we focus on gaining some insight into the influences of multifluid effects on turbulence and leave a study of multifluid mhd turbulence under differing driving to a later date . finally , to ensure a constant energy injection rate and zero net momentum injection , a quadratic equation must be solved to normalise @xmath38 . the resulting normalised incremental velocity field
is then added on to the neutral velocity field .
the rate of injection of energy is the same for all simulations in this work and is given by @xmath42 where @xmath43 is the initial ( uniform ) density , @xmath44 is the length of side of the computational domain and @xmath5 is the ( isothermal ) sound speed . in our simulations
this energy injection rate yields steady state turbulence with an rms mach number of around 4.5 and an alfvnic mach number of between 3 and 3.5 .
table [ table : nomenclature ] contains the definition of the nomenclature of the various simulations used in this work .
.definition of the nomenclature used to refer to the various simulations presented in this work . [
cols="<,^,^",options="header " , ]
we have presented the results of 4-fluid , mhd turbulence in molecular clouds .
the parameters chosen were matched to the observations of molecular clouds reported in @xcite .
we performed a resolution study to ensure that the results presented , in terms of both the power spectra and the rms mach number , were not strongly influenced by resolution effects .
we found that the rms mass - weighted mach number was not significantly decreased with the inclusion of multifluid effects , in spite of multifluid effects providing an effective pathway for removal of magnetic energy .
the overall energy dissipation , then , appears not to be significantly affected by multifluid effects .
this phenomenon is a result of the fact that the majority of the energy in a driven , turbulent system with the properties of a molecular cloud resides in kinetic energy and so an enhanced removal of magnetic energy does not significantly effect the global energetics of the system .
we found that both the velocity power spectra of the neutrals and the magnetic field power spectra were strongly influenced by multifluid effects at all length - scales up to the driving scale ( 0.05pc ) . in multifluid mhd
, the velocity power spectrum of the neutrals is found to have considerably less structure at all length - scales than those of the charged species .
the electron and ion fluids behave almost identically to each other , while the dust displays some slight differences with the other charged species .
this latter point arises largely from the requirement of local charge neutrality and the fact that the electrons and ions are the dominant charge carriers in the system .
we also find , at very short length - scales , some signs of the hall effect in the power spectra of the charged species .
thus we find , in common with the results presented in papers i and ii that multifluid effects are important up to quite large length - scales .
this result appears to be in conflict with the observations of @xcite and we deal with this apparent contradiction in a forthcoming paper
. the differences in the power spectra of the bulk densities in the multifluid and ideal mhd systems are not as dramatic as those in the other variables , although there is somewhat less power in the distribution of the neutral density in the multifluid simulation than the bulk density in the ideal mhd one .
interestingly , there is _ less _ power in the spectrum of the charged species density than the neutral species in the multifluid system .
this is the opposite of what is found for the velocity power spectra . for all species , and for all cases ,
the density pdfs are found to be very close to log - normal .
the neutral species in the multifluid simulation has a slightly less platykurtic ( flattened ) distribution than that in the ideal mhd case .
the charged densities in the multifluid simulation have , in turn , less platykurtic distributions than the neutrals in that simulation with the dust actually having a leptokurtic distribution .
each of these results is explained in terms of the ambipolar diffusion experienced by the multifluid system , and the fact that the charged species are more directly influenced by the magnetic field than the neutrals .
finally , we find that multifluid effects do not significantly influence the centroid velocity increment structure functions .
the computations for this work were carried out as part of the deisa - deci 6 project `` d - sturb '' .
we also acknowledge the results presented in this paper were partly achieved using the prace research infrastructure resource jugene based in germany at the jlich supercomputing centre . | it is believed that turbulence may have a significant impact on star formation and the dynamics and evolution of the molecular clouds in which this occurs .
it is also known that non - ideal magnetohydrodynamic effects influence the nature of this turbulence .
we present the results of a numerical study of 4-fluid mhd turbulence in which the dynamics of electrons , ions , charged dust grains and neutrals and their interactions are followed .
the parameters describing the fluid being simulated are based directly on observations of molecular clouds .
we find that the velocity and magnetic field power spectra are strongly influenced by multifluid effects on length - scales at least as large as 0.05pc .
the pdfs of the various species in the system are all found to be close to log - normal , with charged species having a slightly less platykurtic ( flattened ) distribution than the neutrals .
we find that the introduction of multifluid effects does not significantly alter the structure functions of the centroid velocity increment .
[ firstpage ] ism : kinematics and dynamics
ism : magnetic fields magnetohydrodynamics ( mhd ) methods : numerical turbulence |
[ s : uno ] in quantum mechanics , quantum channels describe evolution of an open system @xmath1 interacting with external degrees of freedom . in the schdinger picture
these transformations are described by completely positive trace preserving ( cpt ) linear maps @xmath2 acting on the set @xmath3 of the density matrices @xmath4 of the system .
it is a well known ( see e.g. @xcite , @xcite ) that @xmath2 can be described by a unitary coupling between the system @xmath1 in input state @xmath4 with an external ancillary system @xmath5 ( describing the _ environment _ ) prepared in some fixed _ pure _ state .
this follows from stinespring dilation @xcite of the map which is unique up to a partial isometry .
more generally , one can describe @xmath2 as a coupling with environment prepared in some _ mixed _ state @xmath6 , i.e. @xmath7 \ ; , \label{hgcuno}\end{aligned}\ ] ] where @xmath8 $ ] is the partial trace over the environment @xmath5 , @xmath9 is a unitary operator in the composite hilbert space @xmath10 .
we call eq .
( [ hgcuno ] ) a `` physical representation '' of @xmath2 to distinguish it from the stinespring dilation , and to stress its connection with the physical picture of the noisy evolution represented by @xmath2 .
any stinespring dilation gives rise to a physical representation .
moreover from any physical representation ( [ hgcuno ] ) one can construct a stinespring dilation by purifying @xmath6 with an external ancillary system @xmath11 , and by replacing @xmath9 with the unitary coupling @xmath12 .
equation ( [ hgcuno ] ) motivates the following @xcite for any physical representation _ ( [ hgcuno ] ) _ of the quantum channel @xmath2 we define its _ weakly complementary _ as
the map @xmath13 which takes the input state @xmath4 into the state of the environment @xmath5 after the interaction with @xmath1 , i.e. @xmath14 \;. \label{duedue}\end{aligned}\ ] ] the transformation ( [ duedue ] ) is cpt , and it describes a quantum channel connecting systems @xmath1 and @xmath5 .
it is a generalization of the _ complementary ( conjugate ) channel _ @xmath15 defined in ref .
in particular , if eq .
( [ hgcuno ] ) arises from a stinespring dilation ( i.e. if @xmath6 of eq .
( [ duedue ] ) is pure ) the map @xmath16 coincides with @xmath15 .
hence the latter is a particular instance of a weakly complementary channel of @xmath2 . on the other hand , by using the above purification procedure , we can always represent a weakly complementary map as a composition @xmath17 where @xmath18 is the partial trace over the purifying system ( here _ `` @xmath19 '' _ denotes the composition of channels ) .
as we will see , the properties of weakly complementary and complementary maps in general differ .
let @xmath20 be a pair of mutually weakly - complementary channels such that @xmath21 for some channel @xmath22 and all density matrix @xmath23 .
then @xmath2 is called _ weakly - degradable _ while @xmath16 _ anti - degradable _ ( cf .
@xcite ) . similarly if @xmath24 for some channel @xmath25 and all density matrix @xmath26 , then @xmath2 is _ anti - degradable _ while @xmath16 is _ weakly - degradable_. in ref .
@xcite the channel @xmath2 is called _ degradable _ if in eq .
( [ deg ] ) we replace @xmath16 with a complementary map @xmath15 of @xmath2 .
clearly any degradable channel @xcite is weakly degradable but the opposite is not necessarily true .
notice , however , that due to eq .
( [ comp ] ) , in the definition of anti - degradable channel we can always replace weakly complementary with complementary ( for this reason there is no point in introducing the notion of weakly anti - degradable channel ) .
this allows us to verify that if @xmath2 is anti - degradable ( [ antideg ] ) then its complementary channel @xmath15 is degradable @xcite and vice - versa .
it is also worth pointing out that channels which are unitarily equivalent to a channel @xmath2 which is weakly degradable ( anti - degradable ) are also weakly degradable ( anti - degradable ) .
finally an important property of anti - degradable channels is the fact that their quantum capacity @xcite is null .
as discussed in @xcite this is a consequence of the no - cloning theorem @xcite ( more precisely , of the impossibility of cloning with arbitrary high fidelity @xcite ) .
it is useful also to reformulate our definitions in the heisenberg picture . here
the states of the system are kept fixed and the transformation induced on the system by the channel is described by means of a linear map @xmath27 acting on the algebra @xmath28 of all bounded operators of @xmath1 so that @xmath29 = { \mbox{tr}}_a [ \rho_a \ ; \phi_h ( \theta_a ) ] \label{identity}\;,\end{aligned}\ ] ] for all @xmath30 and for all @xmath31 . from this it follows that the heisenberg picture counterpart of the physical representation ( [ hgcuno ] ) is given by the unital channel @xmath32\;. \label{physicalheis}\end{aligned}\ ] ] similarly , from ( [ duedue ] ) it follows that in the heisenberg picture the weakly complementary of the channel is described by the completely positive unital map @xmath33 \label{wconjheis } \;,\end{aligned}\ ] ] which takes bounded operators in @xmath34 into bounded operators in @xmath35 .
within this framework the weak - degradability property ( [ deg ] ) of the channel @xmath27 requires the existence of a channel @xmath36 taking bounded operators of @xmath34 into bounded operators of @xmath35 , such that @xmath37 for all @xmath38 .
similarly we say that a quantum channel @xmath27 is anti - degradable , if there exists a channel @xmath39 from @xmath40 to @xmath41 , such that @xmath42 for all @xmath31 .
gaussian channels arise from linear dynamics of open bosonic system interacting with gaussian environment via quadratic hamiltonians . loosely speaking
, they can be characterized as cpt maps that transform gaussian states into gaussian states @xcite . here
we focus on one - mode bosonic gaussian channels which act on the density matrices of single bosonic mode @xmath1 .
a classification of such maps obtained recently in the paper @xcite allows us to simplify the analysis of the weak - degradability property . in the following we start by reviewing the result of ref .
@xcite , clarifying the connection with the analysis of ref .
@xcite ( cf . also ref .
then we pass to the weak - degradability analysis of these channels , showing that with some important exception , they are either weakly degradable or anti - degradable . consider a single bosonic mode characterized by canonical observables @xmath43 obeying the canonical commutation relation @xmath44=i$ ] .
a consistent description of the system can be given in terms of the unitary weyl operators @xmath45 $ ] , with @xmath46 being a column vector of @xmath47 .
in this framework the canonical commutation relation is written as @xmath48\ ; v_a(z+z^{\prime } ) \ ; , \label{weyl}\ ] ] where @xmath49 is the symplectic form @xmath50 with @xmath51 being the second pauli matrix .
moreover the density operators @xmath4 of the system can be expressed in terms of an integral over @xmath52 of the @xmath53 s , i.e. @xmath54 with @xmath55\ ; , \label{characte}\end{aligned}\ ] ] being the characteristic function of @xmath4 ) holds also for all trace class operators of @xmath1 @xcite . ] .
consequently a complete description of a quantum channel on @xmath1 is obtained by specifying its action on the operators @xmath53 , or , equivalently , by specifying how to construct the characteristic function @xmath56 of the evolved states . in the case of gaussian channels
@xmath2 this is done by assigning a mapping of the weyl operators @xmath57 \label{linbos}\;,\ ] ] in the heisenberg picture , or the transformation of the characteristic functions @xmath58 , \label{linbos1 } \end{aligned}\ ] ] in the schrdinger picture . here
@xmath59 is a vector , while @xmath60 and @xmath61 are real matrices ( the latter being symmetric and positive ) .
equation ( [ linbos1 ] ) guarantees that any input gaussian characteristic function will remain gaussian under the action of the map .
a useful property of gaussian channels is the fact that the composition of two of them ( say @xmath62 and @xmath63 ) is still a gaussian channel .
indeed one can easily verify that the composite map @xmath64 is of the form ( [ linbos1 ] ) with @xmath59 , @xmath60 and @xmath61 given by @xmath65 here @xmath66 , @xmath67 , and @xmath68 belongs to @xmath69 while @xmath70 , @xmath71 , and @xmath72 belongs to @xmath63 .
not all possible choices of @xmath60 , @xmath61 correspond to transformations @xmath2 which are completely positive .
a necessary and sufficient condition for this last property ( adapted to the case of one mode ) is provided by the nonnegative definiteness of the following @xmath73 hermitian matrix @xcite @xmath74 this matrix reduces to @xmath75 -1 ) \ ; \sigma_2 $ ] and its nonnegative definiteness to the inequality @xmath76 \geqslant \left(\frac{{\mbox{det}}[k]-1}{2}\right)^2 \;. \label{inequality}\end{aligned}\ ] ] within the limit imposed by eq .
( [ inequality ] ) we can use eq .
( [ linbos1 ] ) to describe the whole set of the one - mode gaussian channels .
an important subset of one - mode gaussian channels is given by the maps @xmath2 which possess a physical representation ( [ hgcuno ] ) with @xmath6 being a gaussian state of a _ single _ external bosonic mode @xmath5 and with @xmath9 being a canonical transformation of @xmath77 , @xmath78 , @xmath79 and @xmath80 ( the latter being the canonical observables of the mode @xmath5 ) .
in particular let @xmath6 be a thermal state of average photon number @xmath81 , i.e. @xmath82 = \exp [ - ( n+1/2 ) be such that @xmath83 with @xmath84 being a @xmath85 symplectic matrix of block form @xmath86 this yields the following evolution for the characteristic function @xmath87 , @xmath88 = { \mbox{tr}}_a [ \rho_a \;\phi_h(v_a(z ) ) ] \nonumber \\ & = & { \mbox{tr}}_{ab } \left [ u_{ab}^\dag \ ; ( v_a(z ) \otimes { \leavevmode\hbox{\small1\kern-3.8pt\normalsize1 } } ) u_{ab } \ ; ( \rho_a \otimes \rho_b ) \right ] \nonumber \\ & = & { \mbox{tr}}_{ab } \left[\big ( v_a(m_{11 } \cdot z ) \otimes v_b(m_{12}\cdot z ) \big ) \ ; ( \rho_a \otimes \rho_b ) \right ] \nonumber \\ & = & \phi(\rho_a ; m_{11 } \cdot z ) \ ; \exp [ - ( n+1/2 ) | m_{12}\cdot z|^2 /2]\ ; , \label{output}\end{aligned}\ ] ] which is of the form ( [ linbos1 ] ) by choosing @xmath89 , @xmath90 and @xmath91 .
it is worth stressing that in the case of eq .
( [ output ] ) the inequality ( [ inequality ] ) is guaranteed by the symplectic nature of the matrix @xmath84 , i.e. by the fact that eq .
( [ couplingn ] ) preserves the commutation relations among the canonical operators .
indeed we have @xmath76 & = & ( n+1/2)^2 \ ; { \mbox{det}}[m_{12}]^2 = ( n+1/2)^2 \ ; ( { \mbox{det}}[m_{11}]-1)^2\nonumber \\ & \geqslant & ( { \mbox{det}}[k]-1)^2/4 \label{ineq1}\;,\end{aligned}\ ] ] where in the second identity the condition ( [ simpcond ] ) was used . as we shall see , with certain important exception one - mode gaussian channels ( [ linbos ] ) are unitarily equivalent to transformations which admit physical representation with @xmath6 and @xmath9 as in eqs .
( [ sigmab ] ) and ( [ couplingn ] ) .
following ref .
@xcite any gaussian channel ( [ linbos1 ] ) can be transformed ( through unitarily equivalence ) into a simple canonical form .
namely , given a channel @xmath2 characterized by the vector @xmath59 and the matrices @xmath60 , @xmath61 of eq .
( [ linbos1 ] ) , one can find unitary operators @xmath92 and @xmath93 such that the channel defined by the mapping @xmath94 is of the form ( [ linbos1 ] ) with @xmath89 and with @xmath60 , @xmath61 replaced , respectively , by the matrices @xmath95 , @xmath96 of table [ t : table ] , i.e. @xmath97\;. \label{linbos2 } \end{aligned}\ ] ] an important consequence of eq .
( [ linbos2 ] ) is that to analyze the weak - degradability properties of a one - mode gaussian channel it is sufficient to focus on the canonical map @xmath98 which is unitarily equivalent to it ( see remark at the end of sec .
[ s : uno ] ) . here
we will not enter into the details of the derivation of eqs .
( [ equivalent ] ) and ( [ linbos2 ] ) , see ref .
@xcite .
cl|c|cc channel @xmath2 & & class & canonical form @xmath98 & + @xmath99 $ ] & & & @xmath95 & @xmath96 + @xmath100 & @xmath101 = 0 $ ] & @xmath102 & @xmath100 & @xmath103 + @xmath100 & @xmath101 = 1 $ ] & @xmath104 & @xmath105 & @xmath103 + @xmath106&@xmath107 = 1 $ ] & @xmath108 & @xmath109 & @xmath110 + @xmath106 & @xmath107 \neq 1 $ ] & @xmath0 & @xmath109 & @xmath111 + @xmath112 & & @xmath11 & @xmath113 & @xmath114 + @xmath115 & & @xmath116 & @xmath117 & @xmath118 the dependence on the matrix @xmath95 of @xmath98 upon the parameters of @xmath2 can be summarized as follows , @xmath119 } \ ; { \leavevmode\hbox{\small1\kern-3.8pt\normalsize1 } } & & \mbox{det}[k]\geqslant 0 \\ \sqrt{|\mbox{det}[k]| } \ ; \sigma_3 & & \mbox{det}[k]<0 \end{array } \right . & & \mbox{rank}[k ] \neq 1 \\ \\ ( { \leavevmode\hbox{\small1\kern-3.8pt\normalsize1}}+ \sigma_3)/2 & & \mbox{rank}[k ] = 1 \ ; , \end{array } \right.\label{kappacan}\end{aligned}\ ] ] with @xmath120 being the third pauli matrix .
analogously for @xmath96 we have @xmath121 } \ ; { \leavevmode\hbox{\small1\kern-3.8pt\normalsize1 } } & & \mbox{rank}[\alpha ] \neq 1 \\ \\ \ ; n_c \ ; ( { \leavevmode\hbox{\small1\kern-3.8pt\normalsize1}}- \sigma_3)/2 & & \mbox{rank}[\alpha ] = 1 \;. \end{array } \right . \label{alphacan}\end{aligned}\ ] ] the quantity @xmath122 is a free parameter which can set to any positive value upon properly calibrating the unitaries @xmath123 and @xmath93 of eq .
( [ equivalent ] ) .
following ref .
@xcite we will assume @xmath124 .
notice also that from eq .
( [ inequality ] ) , @xmath107=1 $ ] is only possible for @xmath125=1 $ ] .
equations ( [ kappacan ] ) and ( [ alphacan ] ) show that only the determinant and the rank of @xmath60 and @xmath61 are relevant for defining @xmath95 and @xmath96 .
indeed one can verify that @xmath95 and @xmath96 maintain the same determinant and rank of the original matrices @xmath60 and @xmath61 , respectively .
this is a consequence of the fact the @xmath2 and @xmath98 are connected through a symplectic transformation for which @xmath126 $ ] , @xmath127 $ ] , @xmath128 $ ] , and @xmath107 $ ] are invariant quantities . [ in particular @xmath129 $ ] is directly related with the invariant quantity @xmath130 analyzed in ref .
@xcite . ]
the six inequivalent canonical forms of table [ t : table ] follow by parametrizing the value of @xmath131}$ ] to account for the constraints imposed by the inequality ( [ inequality ] ) .
it should be noticed that to determine which class a certain channel belongs to , it is only necessary to know if @xmath125 $ ] is null , equal to @xmath106 , negative or positive ( @xmath132 ) . if @xmath125=0 $ ] the class is determined by the rank of the matrix . if @xmath125=1 $ ] the class is determined by the rank of @xmath61 ( see fig .
[ f : scheme ] ) . within the various classes ,
the specific expression of the canonical form depends then upon the effective values of @xmath125 $ ] and @xmath133 $ ] .
we observe also that the class @xmath102 can be obtained as a limiting case ( for @xmath134 ) of the maps of class @xmath11 or @xmath116 .
analogously the class @xmath0 can be obtained as a limiting case of the maps of class @xmath11 .
indeed consider the channel with @xmath135 and @xmath136 with @xmath137 , with @xmath138 and @xmath139 positive ( @xmath140 ) .
for @xmath139 sufficiently close to @xmath106 , @xmath141 is positive and the maps belongs to the class @xmath11 of table [ t : table ] .
moreover in the limit of @xmath142 this channel yields the map @xmath0 .
finally it is interesting to study how the canonical forms of table [ t : table ] compose under the product ( [ composition ] ) .
a simple calculation shows that the following rules apply @xmath143 in this table , for instance , the element on the row 2 and column 3 represents class ( i.e. @xmath104 ) associated to the product @xmath144 between a channel @xmath62 of @xmath108 and a channel @xmath63 of @xmath104 .
notice that the canonical form of the products @xmath145 , @xmath146 and @xmath147 is not uniquely defined . in the first case in fact , even though the determinant of the matrix @xmath60 of eq .
( [ composition ] ) is one , the rank of the corresponding @xmath61 might be one or different from one depending on the parameters of the two `` factor '' channels : consequently the @xmath145 and @xmath146 might belong either to @xmath108 or to @xmath0 . in the case of @xmath147 instead it is possible that the resulting channel will have @xmath125=1 $ ] making it a @xmath0 map .
typically however @xmath147 will be a map of @xmath11 .
composition rules analogous to those reported here have been extensively analyzed in refs .
@xcite .
apart from the case @xmath0 that will be treated separately ( see next section ) , all canonical transformations of table [ t : table ] can be expressed as in eq .
( [ output ] ) , i.e. through a physical representation ( [ hgcuno ] ) with @xmath6 being a thermal state ( [ sigmab ] ) of a single external bosonic mode @xmath5 and @xmath9 being a linear transformation ( [ couplingn ] ) corresponds to the fact that any one - mode bosonic gaussian channel can be represented as a unitary coupling with a single - mode environment plus an additive classical noise ( see next section and ref .
@xcite ) . ] . to show this
it is sufficient to verify that , for each of the classes of table [ t : table ] but @xmath0 , there exists a non - negative number @xmath81 and a symplectic matrix @xmath84 such that eq .
( [ output ] ) gives the mapping ( [ linbos2 ] ) .
this yields the conditions @xmath148 with @xmath149 being an orthogonal @xmath73 matrix to be determined through the symplectic condition @xmath150 + { \mbox{det}}[m_{12 } ] = 1 \label{simpcond } \;,\end{aligned}\ ] ] which guarantees that @xmath151 and @xmath152 satisfy canonical commutation relations .
it is worth noticing that once @xmath153 and @xmath154 are determined within the constraint ( [ simpcond ] ) the remaining blocks ( i.e. @xmath155 and @xmath156 ) can always be found in order to satisfy the remaining symplectic conditions of @xmath84 .
an explicit example will be provided in few paragraphs . for the classes @xmath102 , @xmath104 , @xmath108 , @xmath116 , and @xmath11 with @xmath157 , eqs .
( [ m12 ] ) and ( [ simpcond ] ) can be solved by choosing @xmath158 and @xmath159 .
indeed for @xmath108 the latter setting is not necessary .
any non - negative number will do the job : thus we choose @xmath160 making the density matrix @xmath6 of eq .
( [ sigmab ] ) the vacuum of the @xmath5 . for @xmath11 with @xmath161 instead a solution is obtained by choosing @xmath162 and again @xmath159 .
the corresponding transformations ( [ couplingn ] ) for @xmath77 and @xmath78 ( together with the choice for @xmath81 ) are summarized below .
@xmath163 to complete the definition of the unitary operators @xmath9 we need to provide also the transformations of @xmath79 and @xmath80 .
this corresponds to fixing the blocks @xmath155 and @xmath156 of @xmath84 and can not be done uniquely : one possible choice is presented in the following table @xmath164 the above definitions make explicit the fact that the canonical form @xmath11 represents attenuator ( @xmath157 ) and amplifier ( @xmath161 ) channel @xcite .
we will see in the following sections that the class @xmath116 is formed by the weakly complementary of the amplifier channels of the class @xmath11 . for the sake of clarity
the explicit expression for the matrices @xmath84 of the various classes has been reported in app .
[ appendicem ] .
finally it is important to notice that the above physical representations are equivalent to stinespring representations only when the average photon number @xmath81 of @xmath6 nullifies . in this case
the environment @xmath5 is represented by a pure input state ( i.e. the vacuum ) .
according to our definitions this is always the case for the canonical form @xmath108 while for the canonical forms @xmath102 , @xmath104 , @xmath11 and @xmath116 it happens for @xmath165 . as mentioned in the previous section the class @xmath0 of table [ t : table ]
must be treated separately .
the map @xmath0 corresponds ) and comparing it with eq .
( [ linbos2 ] ) . ] to the additive classical noise channel @xcite defined by @xmath166 with @xmath167 $ ] which , in heisenberg picture , can be seen as a random shift of the annihilation operator @xmath168 .
these channels admit a natural physical representation which involve two environmental modes in a pure state ( see ref .
@xcite for details ) but do not have a physical representations ( [ hgcuno ] ) involving a single environmental mode .
this can be verified by noticing that in this case , from eqs .
( [ m11 ] ) and ( [ m12 ] ) we get @xmath169 which yields @xmath150 + { \mbox{det}}[m_{12 } ] = 1 \pm n_0/(n+1/2 ) \;,\end{aligned}\ ] ] independently of the choice of the orthogonal matrix @xmath170=\pm 1 $ ] since @xmath171 . ] .
therefore , apart from the trivial case @xmath165 , the only solution to the constraint ( [ simpcond ] ) is by taking the limit @xmath172 .
this would correspond to representing the channel @xmath0 in terms of a linear coupling with a single - mode thermal state @xmath6 of `` infinite '' temperature .
unfortunately this is not a well defined object .
however we can use the `` asymptotic '' representation described at the end of sec .
[ s : canonical ] where it was shown how to obtain @xmath0 as limiting case of @xmath11 class maps , to claim at least that there exists a one - parameter family of one - mode gaussian channels which admits single - mode physical representation and which converges to @xmath0 .
cc|cc|c class of @xmath173 & & weak complementary channel @xmath16 & & class of @xmath16 + & & @xmath60 & @xmath61 & + @xmath102 & & @xmath109 & @xmath100 & @xmath0 + @xmath104 & & @xmath109 & @xmath174 & @xmath108 + @xmath108 & & @xmath175 & @xmath176 & @xmath104 + @xmath11 & @xmath157 & @xmath177 & @xmath178 & @xmath179 + @xmath11 & @xmath180 & @xmath181 & @xmath182 & @xmath116 + @xmath116 & & @xmath183 & @xmath184 & @xmath185
in the previous section we have seen that all one - mode gaussian channels are unitarily equivalent to one of the canonical forms of table [ t : table ] .
moreover we verified that , with the exception of the class @xmath0 , all the canonical forms admits a physical representation ( [ hgcuno ] ) with @xmath6 being a thermal state of a single environmental mode and @xmath9 being a linear coupling . here
we will use such representations to construct the weakly complementary ( [ duedue ] ) of these channels and to study their weak - degradability properties . in this section
we construct the weakly complementary channels @xmath16 of the class @xmath102 , @xmath104 , @xmath108 , @xmath11 and @xmath116 starting from their single - mode physical representations ( [ hgcuno ] ) of sec .
[ sec : single ] .
because of the linearity of @xmath9 and the fact that @xmath6 is gaussian , the channels @xmath16 are gaussian
. this can be seen for instance by computing the characteristic function ( [ characte ] ) of the output state @xmath186 @xmath187 = { \mbox{tr}}_{b } [ \rho_a \ ; \tilde{\phi}_h(v_b(z ) ) ] \nonumber \\ & = & \phi ( \rho_a ; m_{21 } \cdot z ) \ ; \exp [ - \frac{1}{2 } ( n+ 1/2 ) \ ; | m_{22 } \cdot z|^2 ] \label{charn } \;,\end{aligned}\ ] ] where @xmath155 , @xmath156 are the blocks elements of the matrix @xmath84 of eq .
( [ matricem ] ) associated with the transformations @xmath9 , and with @xmath81 being the average photon number of @xmath6 ( the values of these quantities are given in the tables of sec .
[ sec : single ] see also app .
[ appendicem ] ) . by setting @xmath89 , @xmath188 and @xmath189 , eq .
( [ charn ] ) has the same structure ( [ linbos1 ] ) of the one - mode gaussian channel of @xmath1 .
therefore by cascading @xmath16 with an isometry which exchanges @xmath1 with @xmath5 ( see refs .
@xcite ) we can then treat @xmath16 as an one - mode gaussian channel operating on @xmath1 ( this is possible because both @xmath1 and @xmath5 are bosonic one - mode systems ) . with the help of table [ t : table ] we can then determine which classes can be associated with the transformation ( [ charn ] ) .
this is summarized in table [ t : table10 ] . using the compositions rules of eqs .
( [ composition ] ) and ( [ compo1 ] ) it is easy to verify that the canonical forms @xmath102 , @xmath104 , @xmath116 and @xmath11 with @xmath190 are anti - degradable ( [ wwwanti ] ) .
vice - versa one can verify that the canonical forms @xmath108 and @xmath11 with @xmath191 are weakly degradable ( [ wwwweak ] ) for @xmath11 , @xmath116 and @xmath102 these results have been proved in ref . @xcite . through unitary equivalence
this can be summarized by saying that all one - mode gaussian channels ( [ linbos1 ] ) having @xmath125\leqslant 1/2 $ ] are anti - degradable , while the others ( with the exception of the channels belonging to @xmath0 ) are weakly degradable ( see fig . [
f : scheme2 ] ) . in the following
we verify the above relations by explicitly constructing the connecting channels @xmath192 and @xmath193 of eqs .
( [ wwwweak ] ) and ( [ wwwanti ] ) for each of the mentioned canonical forms .
indeed one has : * for a channel @xmath2 of standard form @xmath102 or @xmath104 , anti - degradability can be shown by simply taking @xmath193 of eq .
( [ wwwanti ] ) coincident with the channel @xmath2 .
the result immediately follows from the composition rule ( [ composition ] ) . * for a channel @xmath2 of @xmath108 , weak - degradability comes by assuming the map @xmath192 to be equal to the weakly complementary channel @xmath16 of @xmath2 ( see table [ t : table10 ] ) . as pointed out in ref .
@xcite this also implies the degradability of @xmath2 in the sense of ref .
let us remind that for @xmath108 the physical representation given in sec .
[ sec : single ] was constructed with an environmental state @xmath6 initially prepared in the vacuum state , which is pure .
therefore in this case our representation gives rise to a stinespring dilation .
* for a channel @xmath2 of the class @xmath11 with @xmath194 and @xmath195 we have the following three possibilities : * * if @xmath196 the channel is anti - degradable and the connecting map @xmath193 is a channel of @xmath11 characterized by @xmath197 and @xmath198 with @xmath199 .
* * if @xmath200 the channel is weakly degradable and the connecting map @xmath201 is again a channel of @xmath11 defined as in the previous case but with @xmath202 . for @xmath165
the channel is also degradable @xcite since our physical representation is equivalent to a stinespring representation .
* * if @xmath203 the channel is weakly degradable and the connecting map @xmath201 is a channel of @xmath116 with @xmath197 and @xmath204 with @xmath205 . as in the previous case , for @xmath165 the channel is also degradable @xcite . * for a channel @xmath2 of @xmath116 with @xmath206 and @xmath207 ( @xmath208 and @xmath209 ) we can prove anti - degradability by choosing @xmath193 of eq .
( [ wwwanti ] ) to be yet another maps of @xmath116 with @xmath210 and @xmath211 where @xmath212 . from eq .
( [ composition ] ) and table [ t : table10 ] it then follows that @xmath213 is indeed equal to @xmath2 . concerning the case @xmath0 it was shown in ref .
@xcite that the channel is neither anti - degradable nor degradable in the sense of @xcite ( apart from the trivial case @xmath165 which corresponds to the identity map ) .
on the other hand one can use the continuity argument given in sec .
[ sec : b2 ] to claim that the channel @xmath0 can be arbitrarily approximated with maps which are weakly degradable ( those belonging to @xmath11 for instance ) .
in the previous section we saw that all channels ( [ linbos1 ] ) with @xmath125\leqslant 1/2 $ ] are anti - degradable .
consequently these channel must have null quantum capacity @xcite . here
we go a little further showing that the set of the maps ( [ linbos1 ] ) which can be proved to have null quantum capacity include also some maps with @xmath125 > 1/2 $ ] . to do this we will use the following simple fact : _ let be @xmath214 a quantum channel with null quantum capacity and let be @xmath215 some quantum channel . then
the composite channels @xmath216 and @xmath217 have null quantum capacity .
_ the proof of this property follows by interpreting @xmath215 as a quantum operation performed either at the decoding or at encoding stage of the channel @xmath214 .
this shows that the quantum capacities of @xmath216 and @xmath217 can not be greater than the capacity of @xmath214 ( which is null ) . in the following
we will present two cases where the above property turns out to provide some nontrivial results .
we observe that according to composition rule ( [ compo1 ] ) the combination of any two channels @xmath214 and @xmath215 of @xmath116 produces a map @xmath218 which is in the class @xmath11 . since the class @xmath116 is anti - degradable the resulting channel must have null quantum capacity .
let then @xmath219 and @xmath220 be the matrices @xmath95 and @xmath96 of the channels @xmath221 , for @xmath222 . from eq .
( [ composition ] ) one can then verify that @xmath223 has the canonical form @xmath11 with parameters @xmath224 equation ( [ kappa ] ) shows that by varying @xmath225 , @xmath139 can take any positive values : in particular it can be greater than @xmath226 transforming @xmath223 into a channel which does not belong to the anti - degradable area of fig .
[ f : scheme2 ] . on the other hand , by varying the @xmath227 and @xmath228 , but keeping the product @xmath229 fixed , the parameter @xmath138 can assume any value satisfying the inequality @xmath230 we can therefore conclude that all channels @xmath11 with @xmath139 and @xmath138 as in eq .
( [ ennen ] ) have null quantum capacity see fig .
[ f : plot ] .
a similar bound was found in a completely different way in ref .
@xcite .
consider now the composition of two class @xmath11 channels , i.e. @xmath231 and @xmath215 , with one of them ( say @xmath215 ) being anti - degradable . here , the canonical form of @xmath214 and @xmath215 have matrices @xmath95 and @xmath96 given by @xmath232 and @xmath233 , where for @xmath222 , @xmath227 and @xmath225 are positive numbers , with @xmath234 and with @xmath2350,\sqrt{1/2}]$ ] ( to ensure anti - degradability ) . from eq .
( [ composition ] ) follows then that the composite map @xmath236 has still a @xmath11 canonical form with parameters @xmath237 as in the previous example , @xmath139 can assume any positive value .
vice - versa keeping @xmath139 fixed , and varying @xmath238 and @xmath239 it follows that @xmath138 can take any values which satisfy the inequality @xmath240 we can then conclude that all maps @xmath11 with @xmath139 and @xmath138 as above must possess null quantum capacity .
the result has been plotted in fig .
[ f : plot ] .
notice that the constraint ( [ ennenn ] ) is an improvement with respect to the constraint of eq .
( [ ennen ] ) .
in this paper we provide a full weak - degradability classification of one - mode gaussian channels by exploiting the canonical form decomposition of ref . @xcite . within this context
we identify those channels which are anti - degradable . by exploiting composition rules of gaussian maps ,
this allows us to strengthen the bound for one - mode gaussian channels which have not null quantum capacity .
f.c . and v.g .
thank the quantum information research program of centro di ricerca matematica ennio de giorgi of scuola normale superiore for financial support .
h. acknowledges hospitality of centre for quantum computation , department of applied mathematics and theoretical physics , cambridge university .
here we give the explicit expressions of the matrix @xmath84 of eq .
( [ matricem ] ) associated with the physical representations of the classes @xmath102 , @xmath104 , @xmath108 , @xmath11 and @xmath116 , discussed in sec . [ sec : single ] . they are , @xmath241 @xmath242 @xmath243 @xmath244 c. h. bennett and p. w. shor , ieee trans .
theory * 44 * , 2724 ( 1998 ) .
a. s. holevo , probl .
* 8 * , 63 , ( 1972 ) .
a. s. holevo and r. f. werner , phys .
a * 63 * 032312 ( 2001 ) .
j. eisert and m. m. wolf , e - print quant - ph/0505151 .
r. g. gallager , _ information theory and realible communication _
( wiley , new york , 1968 ) . s. braunstein and p. van look , rev
phys . * 77 * , 513 ( 2005 ) . c. m. caves and p. d. drummond , rev .
mod . phys . * 66 * , 481 ( 1994 ) ; h. p. yuen and m. ozawa , phys .
lett . * 70 * 363 ( 1993 ) .
a. s. holevo , m. sohma , and o. hirota , phys .
a * 59 * , 1820 ( 1999 ) .
v. giovannetti , s. guha , s. lloyd , l. maccone , j. h. shapiro , and h. p. yuen , phys .
lett . * 92 * , 027902 ( 2004 ) .
v. giovannetti , s. lloyd , l. maccone and p.w .
shor , phys .
lett . * 91 * , 047901 ( 2003 ) ; phys .
a * 68 * , 062323 ( 2003 ) .
a. s. holevo , ieee trans .
theory * 44 * , 269 ( 1998 ) ; p. hausladen , r. jozsa , b. schumacher , m. westmoreland and w. k. wootters , phys .
a * 54 * , 1869 ( 1996 ) ; b. schumacher and m. d. westmoreland , phys .
a * 56 * , 131 ( 1997 ) .
a. s. holevo , probability theory and application , * 48 * , 359 ( 2003 ) ; a. s. holevo and and m. e. shirokov , commun .
. phys . * 249 * , 417 , 2004 ; m. e. shirovov , commun .
phys . * 262 * , 137 ( 2006 ) ; a. s holevo and m. e. shirokov , quant - ph/0408176 . s. lloyd , phys . rev .
a * 55 * , 1613 ( 1997 ) ; h. barnum , m. a. nielsen , and b. schumacher , phys .
a * 57 * , 4153 ( 1998 ) ; i. devetak , ieee _ trans .
inform . theory _ * 51 * , 44 ( 2005 ) .
bennett , p. w. shor , j. a. smolin , and a. v. thapliyal , ieee trans .
theory * 48 * , 2637 ( 2002 ) .
m.j.w . hall and m.j .
orourke , quantum opt . * 5 * 161 , ( 1993 ) ; m.j.w .
hall , phys .
a * 50 * , 3295 ( 1994 ) ; j.h .
shapiro , v. giovannetti , s. guha , s. lloyd , l. maccone , and b.j .
yen aip conf .
734 * , 15 ( 2004 ) ; m. sohma and o. hirota , quant - ph/0105042 . v. giovannetti , s. guha , s. lloyd , l. maccone , and j.h .
shapiro , phys .
a * 70 * , 032315 ( 2004 ) ; v. giovannetti , s. lloyd , l. maccone , j.h .
shapiro , and b.j .
yen , phys .
a _ ib . _ * 70 * , 022328 ( 2004 ) .
v. giovannetti and s. lloyd , phys .
a * 69 * , 062307 ( 2004 ) .
a. serafini , j. eisert , and m.m .
wolf , phys .
a * 71 * 012320 ( 2005 ) .
f. caruso and v. giovannetti , to appear in phys .
rev . a ( 2006 ) , e - print quant - ph/0603257
. a. s. holevo , e - print quant - ph/0607051 .
m. m. wolf and d. perz - garca , e - print quant - ph/0607070 .
i. devetak and p. w. shor , commun .
phys . * 256 * , 287 ( 2005 ) .
w. f. stinespring , proc . am .
* 6 * , 211 ( 1955 ) .
g. lindblad , commun .
phys . * 48 * , 116 ( 1976 ) .
a. s. holevo , e - print quant - ph/0509101 . c. king , k. matsumoto , m. nathanson , and m. b. ruskai , e - print quant - ph/0509126 .
w. k. wootters and w. h. zurek , nature * 299 * , 802 ( 1982 ) .
n. gisin and s. massar , phys .
lett . * 79 * , 2153 ( 1997 ) ; d. bru , d. p. divincenzo , a. ekert , c. a. fuchs , c. macchiavello , and j. a. smolin , phys . rev .
a * 57 * , 2368 ( 1998 ) ; r. f. werner , phys .
a * 58 * , 1827 ( 1998 ) ; v. bzek , and m. hillery , phys .
lett . * 81 * , 5003 ( 1998 ) .
a. ferraro , s. olivares , and m. g. a. paris , _ gaussian states in quantum information _ ( bibliopolis , napoli , 2005 )
. a. s. holevo , _ probabilistic aspects of quantum theory _ ( north - holland , amsterdam , 1982 ) .
v. giovannetti and r. fazio , phys .
a * 71 * , 032314 ( 2005 ) . | a complete degradability analysis of one - mode gaussian bosonic channels is presented .
we show that apart from the class of channels which are unitarily equivalent to the channels with additive classical noise , these maps can be characterized in terms of weak- and/or anti - degradability .
furthermore a new set of channels which have null quantum capacity is identified .
this is done by exploiting the composition rules of one - mode gaussian maps and the fact that anti - degradable channels can not be used to transfer quantum information . within the context of quantum information theory @xcite bosonic gaussian channels
@xcite play a fundamental role .
they include all the physical transformations which preserve `` gaussian character '' of the transmitted signals and can be seen are the quantum counterpart of the gaussian channels in the classical information theory @xcite .
bosonic gaussian channels describe most of the noise sources which are routinely encountered in optics , including those responsible for the attenuation and/or the amplification of signals along optical fibers . moreover , due to their relatively simple structure
, these channels provide an ideal theoretical playground for the study of continuous variable @xcite quantum communication protocols .
not surprisingly in the recent years an impressive effort has been put forward to characterize the properties of bosonic gaussian channels .
most of the efforts focused on the evaluation of the optimal transmission rates of these maps under the constraint on the input average energy both in the multi - mode scenario ( where the channel acts on a collection of many input bosonic mode ) and in the one - mode scenario ( where , instead , it operates on a single input bosonic mode ) . in few cases
@xcite the exact values of the communication capacities @xcite of the channels have been computed . in the general case
however only certain bounds are available ( see @xcite ) .
finally various additivity issues has been analyzed in refs .
@xcite .
recently the notions of anti - degradability and weak - degradability were proposed as an useful tool for studying the quantum capacity properties of one - mode gaussian channels @xcite .
this suggested the possibility of classifying these maps in terms of a simple canonical form which was achieved in ref .
@xcite .
moreover , proceeding along similar lines , the exact solution of the quantum capacity of an important subset of those channels was obtained in ref .
@xcite .
in this paper we provide a complete degradability classification of one - mode gaussian channels and exhibit a new set of channels which have null quantum capacity extending a previous result in ref .
@xcite .
the definition of weak- and anti - degradability of a quantum channel is similar to the definition of degradability introduced by devetak and shor in ref .
@xcite .
it is based on replacing the stinespring dilation @xcite of the channel with a representation where the ancillary system ( environment ) is not necessarily in a _ pure
_ state @xcite .
this yields a generalization of the notion of _ complementary _ channel from ref .
@xcite which is named _ weakly complementary _ @xcite . in this context
weakly degradable are those channels where the modified state of the ancillary system described by the action of the weakly complementary channel can be recovered from the output state of the channel through the action of a third channel .
vice - versa anti - degradable channels obey the opposite rule ( i.e. the output state of the channel can be obtained from the modified state of the ancilla through the action of another suitable channel ) .
exploiting the canonical form @xcite one can show that , apart from the class @xmath0 consisting of the maps which are unitarily equivalent to the channels with additive classical gaussian noise @xcite , all one - mode bosonic gaussian channels are either weakly degradable or anti - degradable . as discussed in ref .
@xcite the anti - degradability property allows one to simplify the analysis of the quantum capacity @xcite of these channels .
indeed those maps which are anti - degradable can be shown to have null quantum capacity .
on the other hand , those channels which are weakly degradable with pure ancillas ( i.e. those which are degradable in the sense of ref .
@xcite ) have quantum capacity which can be expressed in terms of a single - letter expression . here
we will focus mostly on the anti - degradability property , and , additionally , we will show that by exploiting the composition rules of one - mode bosonic gaussian channels , one can extend the set the maps with null quantum capacity well beyond the set of anti - degradable maps .
the paper is organized as follows . in sec .
[ s : uno ] we introduce the notion of weakly complementarity and weak - degradability in a rather general context . in sec .
[ s : gaus ] we give a detailed description of the canonical decomposition of one - mode bosonic gaussian channels . in sec .
[ s : degradabile ] we discuss the weak - degradability properties of one - mode channels .
finally , in sec [ sec : null ] we determine the new set of channels with null quantum capacity . |
tunneling is one of the most fundamental manifestations of quantum mechanics . for 1d systems the theoretical description
is well established by the wentzel - kramers - brillouin ( wkb ) method and related approaches @xcite . however , for higher dimensional systems no such simple description exists . in these systems
typically regular and chaotic motion coexists and in the two - dimensional case regular tori are absolute barriers to the motion .
quantum mechanically , the eigenfunctions either concentrate within the regular islands or in the chaotic sea , as expected from the semiclassical eigenfunction hypothesis @xcite .
these eigenfunctions are coupled by so - called dynamical tunneling @xcite through the dynamically generated barriers in phase space . in particular , this leads to a substantial enhancement of tunneling rates between phase - space regions of regular motion due to the presence of chaotic motion , which was termed chaos - assisted tunneling @xcite .
such dynamical tunneling processes are ubiquitous in molecular physics and were realized with microwave cavities @xcite and cold atoms in periodically modulated optical lattices @xcite . in the quantum regime , @xmath0 , where the effective planck constant @xmath1 is smaller but still comparable to the area of the regular island @xmath2 ,
the process of tunneling into the chaotic region is dominated by a _ direct regular - to - chaotic tunneling _
mechanism @xcite .
for the prediction of tunneling rates the fictitious integrable system approach was introduced recently @xcite .
it relies on a fictitious integrable system @xcite that resembles the regular dynamics within the island under consideration .
the approach has been applied to quantum maps @xcite , billiard systems @xcite , and the annular microcavity @xcite . in the semiclassical regime , @xmath3 , however , the direct tunneling contribution alone is typically not sufficient to describe the observed tunneling rates , because nonlinear resonance chains within the regular island lead to resonance - assisted tunneling @xcite .
recently , a combined prediction of dynamical tunneling rates from regular to chaotic phase - space regions was derived @xcite , which combines the direct regular - to - chaotic tunneling mechanism in the quantum regime with an improved resonance - assisted tunneling theory in the semiclassical regime . in this text
we concentrate on the fictitious integrable system approach for the theoretical description of the direct regular - to - chaotic tunneling mechanism and how it can be applied analytically , semiclassically , and numerically to quantum maps .
this text is organized as follows : we start by defining different convenient kicked systems , their quantization , and the numerical determination of tunneling rates . in sec . [
sec : dirtun ] we describe the approach to determine dynamical tunneling rates by using a fictitious integrable system . in sec .
[ sec : ex ] we apply this theory to the previously introduced systems and compare its results with numerical data .
we conclude with a brief summary .
kicked systems are particularly suited to study classical and quantum effects appearing in a mixed phase space as they can be easily treated analytically and numerically .
in contrast to time - independent systems , where at least two degrees of freedom are necessary to break integrability , one - dimensional driven systems can show chaotic motion .
we consider time - periodic one - dimensional kicked systems @xmath4 which are described by the kinetic energy @xmath5 and the potential @xmath6 which is applied once per kick period @xmath7 .
the classical dynamics of a kicked system is given by its stroboscopic mapping , e.g. , evaluated just after each kick @xmath8 it maps the phase - space coordinates after the @xmath9th kick to those after the @xmath10th kick . for the stroboscopic mapping
we consider a compact phase space with periodic boundary conditions for @xmath11 $ ] and @xmath12 $ ] .
the phase space generally consists of regions with regular motion surrounded by chaotic dynamics , see fig . [
fig : maps : standard_map](b ) .
we focus on the situation of just one regular island embedded in the chaotic sea , fig .
[ fig : maps : designed](a),(b ) .
at the center of the island one has an elliptic fixed point , which is surrounded by invariant regular tori .
classically , a particle within a regular island will never enter into the chaotic region and vice versa , i.e. tori form absolute barriers to the motion .
the paradigmatic model of an area preserving map is the standard map , defined by eq . with the functions @xmath13 for @xmath14
the system is integrable and the dynamics takes place on invariant tori with constant momentum . for small @xmath15 ,
see fig .
[ fig : maps : standard_map](a ) , the system is no longer integrable .
however , as stated by the kolmogorov - arnold - moser ( kam ) theorem @xcite , many invariant tori persist . in between these tori sequences of stable and unstable fixed points
emerge , which is described by the poincar - birkhoff theorem @xcite .
the stable fixed points of such a sequence form so - called nonlinear resonance chains , see fig .
[ fig : maps : standard_map](b ) , while in the vicinity of the unstable fixed points typically chaotic dynamics is observed . for larger @xmath16 more and more regular tori break up and
the chaotic dynamics occupies a larger area of phase space . at @xmath17 one large regular region ( `` regular island '' ) is surrounded by a region of chaotic dynamics ( `` chaotic sea '' ) . at the border of the regular island
a hierarchical region exists , where self - similar structures of regular and chaotic dynamics are found on all scales .
also in the chaotic sea additional structures can be found . here , partial barriers limit the classical flux between different regions of chaotic motion caused by the stable and unstable manifolds of unstable periodic orbits or caused by cantori , which are the remnants of regular tori @xcite . for even larger @xmath16
the regular islands disappear and the motion seems macroscopically chaotic , see fig . [ fig : maps : standard_map](c ) . , ( b ) @xmath18 , and ( c ) @xmath19 . in ( a )
the system is nearly integrable .
most of the phase space is filled by invariant tori ( lines ) . the second figure ( b )
shows a mixed phase space with regular motion ( lines ) and chaotic motion ( dots ) .
the four small islands surrounding the central island form a nonlinear resonance chain . in the chaotic sea partial barriers exist , which is reflected by the different densities of chaotic points . for large values of @xmath16 as in ( c )
the phase space seems macroscopically chaotic . ]
all these structures in phase space , such as nonlinear resonances , the hierarchical regular - to - chaotic transition region , and partial barriers within the chaotic sea , have an influence on the dynamical tunneling process from the regular island to the chaotic sea . in order to quantitatively understand dynamical tunneling in systems with a mixed phase space ,
it is helpful to first consider simpler model systems . by designing maps such that the phase - space structures can be switched on and off one - by - one it is possible to study their influence on the dynamical tunneling process separately .
afterwards the results can be tested using the generic example of the standard map .
our aim is to introduce kicked systems which are designed such that their phase space is particularly simple .
it shows one regular island embedded in the chaotic sea , with very small nonlinear resonance chains within the regular island , a narrow hierarchical region , and without relevant partial barriers in the chaotic component .
for such a system it is possible to study the direct regular - to - chaotic tunneling process without additional effects caused by these structures as long as @xmath1 is big enough . to this end
we define the family of maps @xmath20 , according to eq . , with an appropriate choice of the functions @xmath21 and @xmath22 @xcite .
for this we first introduce @xmath23 with @xmath24 and @xmath25 .
considering periodic boundary conditions the functions @xmath26 and @xmath27 show discontinuities at @xmath28 and @xmath29 , respectively . in order to avoid these discontinuities
we smooth the periodically extended functions @xmath27 and @xmath30 with a gaussian , @xmath31 resulting in analytic functions @xmath32 which are periodic with respect to the phase - space unit cell . with this
we obtain the maps @xmath20 depending on the parameters @xmath33 , @xmath34 , and the smoothing strength @xmath35 .
the smoothing @xmath35 determines the size of the hierarchical region at the border of the regular island . tuning the parameters @xmath33 and
@xmath34 one can find situations , where all nonlinear resonance chains inside the regular island are small . for @xmath36 both functions @xmath27 and @xmath26 are linear in @xmath37 and @xmath38 , respectively . in this case
we find a harmonic oscillator - like regular island with elliptic invariant tori and constant rotation number .
we choose the parameters @xmath39 , @xmath36 , and @xmath40 for which the phase space of the resulting map @xmath41 is shown in fig .
[ fig : maps : designed](a ) . in typical systems the rotation number of regular tori changes from the center of the regular region to its border which typically has a non - elliptic shape .
such a situation can be achieved for the family of maps @xmath20 with the parameter @xmath42 . for most combinations of the parameters
@xmath33 and @xmath34 resonance structures appear inside the regular island .
they limit the @xmath1-regime in which the direct regular - to - chaotic tunneling process dominates .
hence , we choose a situation in which the nonlinear resonances are small such that their influence on the tunneling process is expected only at small @xmath1 .
the phase space of the map @xmath43 , obtained for @xmath44 , @xmath45 , and @xmath40 , is shown in fig .
[ fig : maps : designed](b ) .
another designed kicked system was introduced in refs .
@xcite . here
the regular region consists of a stripe in phase space , see fig .
[ fig : maps : designed](c ) . in our notation the mapping , eq . , is specified by the functions @xmath46 \tanh[b(8\pi q - q_d)]\right.\nonumber\\ & & \left .
+ \frac{1}{2}[-8\pi aq+\omega+d_2]\tanh[b(8\pi q+q_d)]\right)\\ \label{eq : maps : shudo : t } t'(p ) & = & -\frac{\kappa}{8\pi}\sin(2\pi p).\end{aligned}\ ] ] the kinetic energy @xmath5 is periodic with respect to the phase space unit cell .
we label the resulting map with a regular stripe by @xmath47 using the parameters @xmath48 , @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 , and @xmath54 .
the map @xmath47 is similar to the system @xmath41 as it also destroys the integrable region by smoothly changing the function @xmath22 , here at @xmath55 . for @xmath56
the potential term is almost linear while it tends to the standard map for @xmath57 .
the parameter @xmath58 determines the width of the transition region . with a harmonic oscillator - like regular island , ( b ) @xmath43 with a deformed regular island , and ( c ) @xmath47 with a regular stripe . for parameters
see text . ]
quantum mechanically kicked systems are described by the unitary time - evolution operator @xmath59 @xcite .
it can be written as @xmath60 with @xmath61 its eigenstates @xmath62 and quasi - energies @xmath63 are determined by @xmath64 we consider a compact phase space with periodic boundary conditions in @xmath37 and @xmath38 .
this implies that the effective planck constant can take only the discrete values @xcite @xmath65 where @xmath66 is the dimension of the hilbert space .
in addition the position and momentum coordinates are quantized according to @xmath67 with @xmath68 . for systems with a mixed phase space , in the semiclassical limit ( @xmath69 )
, the semiclassical eigenfunction hypothesis @xcite implies that the eigenstates @xmath62 can be classified as either regular or chaotic according to the phase - space region on which they concentrate , see fig .
[ fig : maps : eigenstates ] .
in order to understand the behavior of eigenstates away from the semiclassical limit , i.e. at finite values of the effective planck constant @xmath1 , one has to compare the size of phase - space structures with @xmath1 . for @xmath70 .
the regular ground state @xmath71 and the first three excited states @xmath72 as well as a chaotic state are shown . ]
the so - called regular states are concentrated on tori within the regular island and fulfill the bohr - sommerfeld type quantization condition @xmath73 for a given value of @xmath1 there exist @xmath74 of such regular states , where @xmath75 and @xmath76 is the area of the regular island .
the @xmath77 chaotic states extend over the chaotic sea . due to dynamical tunneling , however , the regular and chaotic eigenfunctions of @xmath59 always have a small component in the other region of phase space , respectively .
this is most clearly seen for hybrid states which even have the same weight in each of the components as they are involved in an avoided level crossing .
the structure of the considered phase space , with one regular island surrounded by the chaotic sea , allows for the determination of tunneling rates by introducing absorption somewhere in the chaotic region of phase space . for quantum maps ,
this can for example be realized by using a non - unitary open quantum map @xcite @xmath78 where @xmath79 is a projection operator onto the complement of the absorbing region .
while the eigenvalues of @xmath59 are located on the unit circle the eigenvalues of @xmath80 are inside the unit circle as @xmath80 is sub - unitary .
the eigenequation of @xmath80 reads @xmath81 with eigenvalues @xmath82 the decay rates are characterized by the imaginary part of the quasi - energies in eq .
and one has @xmath83 if the chaotic region does not contain partial barriers and shows no dynamical localization , it is justified to assume that the rate of escaping the regular island is equal to the rate of leaving through the absorbing regions located in the chaotic sea .
then , the location of the absorbing regions in the chaotic part of phase space has no effect on the tunneling rates . in generic systems , however , partial barriers will appear in the chaotic region of phase space .
the additional transition through these structures further limits the quantum transport such that the calculated decay through the absorbing region occurs slower than the decay from the regular island to the neighboring chaotic sea . similarly dynamical localization in the chaotic region
may slow down the decay .
the influence of partial barriers and dynamical localization on the tunneling process is an open problem .
here we will suppress their influence , if necessary , by moving the absorbing regions closer to the regular island . in fig .
[ fig : maps : rates_amph_ellipt ] tunneling rates are determined numerically by opening the system for the map @xmath41 ( dots ) .
these agree very well with an analytical prediction ( lines ) , eq .
, which is derived in the next section . for the map @xmath41 with a harmonic oscillator - like island
comparison with eq .
( lines ) , see sec .
[ sec : dirtun : deriv ] for its derivation .
the insets show husimi functions of the regular states for the quantum numbers @xmath71 and @xmath84 at @xmath85 . ]
dynamical tunneling in systems with a mixed phase space couples the regular island and the chaotic sea , which are classically separated . this coupling can be quantified by tunneling rates @xmath86 which describe the decay of regular states to the chaotic sea . to define these tunneling rates one can consider a wave packet started on the @xmath87th quantized torus in the regular island coupled to a continuum of chaotic states , as in the case for an infinite chaotic sea or in the presence of an absorbing region somewhere in the chaotic sea .
its decay @xmath88 is described by a tunneling rate @xmath86 . for systems with a finite phase space
this exponential decay occurs at most up to the heisenberg time @xmath89 , where @xmath90 is the mean level spacing of the chaotic states . in the quantum regime , @xmath91 , where @xmath1 is smaller but comparable to the area @xmath76 of the regular island ,
the rates @xmath86 are dominated by the direct regular - to - chaotic tunneling mechanism , while contributions from resonance - assisted tunneling are negligible .
we concentrate on situations where additional phase - space structures within the chaotic sea are not relevant for tunneling . in the following
we derive a prediction for the direct regular - to - chaotic tunneling rates using the fictitious integrable system approach @xcite . in order to find a prediction for the direct regular - to - chaotic tunneling rates we decompose the hilbert space of the quantum map @xmath59 into two parts which correspond to the regular and chaotic regions . while classically such a decomposition is unique ( neglecting tiny phase - space structures ) , quantum mechanically this is not the case due to the uncertainty principle .
we find a decomposition by introducing a fictitious integrable system @xmath92 ( a related idea was presented in refs .
it has to be chosen such that its dynamics resembles the classical motion corresponding to @xmath59 within the regular island as closely as possible and continues this regular dynamics beyond the regular island of @xmath59 , see fig .
[ fig : deriv : quantum_maps](a),(d ) .
the eigenstates @xmath93 of @xmath92 are purely regular in the sense that they are localized on the @xmath87th quantized torus of the regular region and continue to decay beyond this regular region , see fig .
[ fig : deriv : quantum_maps](e ) .
this is the decisive property of @xmath93 which have no chaotic admixture , in contrast to the predominantly regular eigenstates of @xmath59 , see fig .
[ fig : deriv : quantum_maps](b ) .
the explicit construction of @xmath92 is discussed in sec .
[ sec : dirtun : regsyst ] . with the eigenstates @xmath93 of @xmath92 , @xmath94
, we define a projection operator @xmath95 using the first @xmath74 regular states of @xmath92 which approximately projects onto the regular island corresponding to @xmath59 .
the orthogonal projector @xmath96 approximately projects onto the chaotic phase - space region .
these projectors , @xmath97 and @xmath98 , define our decomposition of the hilbert space into a regular and a chaotic subspace . introducing an orthonormal basis @xmath99 in the chaotic subspace we can write @xmath100 . here
we sum over all @xmath77 states @xmath99 , see fig .
[ fig : deriv : quantum_maps](f ) for an example .
hence , the coupling matrix element @xmath101 between a purely regular state @xmath93 and any chaotic basis state @xmath99 is @xmath102 the tunneling rate is obtained using a dimensionless version of fermi s golden rule , @xmath103 where the sum is over all chaotic basis states @xmath99 and thus averages the modulus squared of the fluctuating matrix elements @xmath101 . here
we apply fermi s golden rule in the case of a discrete spectrum , which is possible if one considers the decay @xmath88 up to the heisenberg time @xmath89 only .
inserting eq .
we obtain @xmath104 as the basis of all our following investigations .
it allows for the prediction of tunneling rates from a regular state localized on the @xmath87th quantized torus to the chaotic sea .
equation confirms the intuition that the tunneling rates are determined by the amount of probability of @xmath93 that is transferred to the chaotic region after one application of the time evolution operator @xmath59 . in eq .
properties of the fictitious integrable system @xmath92 and the chaotic projector @xmath105 enter , which rely on the chosen decomposition of hilbert space . together with the husimi representation of ( b ) a regular and ( c ) a chaotic eigenstate of @xmath59 which both have a small component in the other region .
( d ) illustration of the classical phase space of the fictitious integrable system @xmath92 .
( e ) eigenstates @xmath106 of @xmath92 are purely regular while ( f ) the basis states @xmath99 are localized in the chaotic region of phase space . ] in cases where one finds a fictitious integrable system @xmath92 which resembles the dynamics within the regular island of @xmath59 with very high accuracy , eq . can be approximated as @xmath107 using @xmath108 . instead of the operator product @xmath109 in eq .
the difference @xmath110 enters in eq . .
it allows for further derivations , which are presented in sec .
[ sec : dirtun : approx ] .
the most difficult step in the application of eq . to a given system is the determination of the fictitious integrable system @xmath92 . on the one hand its dynamics
should resemble the classical motion of the considered mixed system within the regular island as closely as possible . as a result the contour lines of the corresponding integrable hamiltonian @xmath111 , fig .
[ fig : deriv : quantum_maps](d ) , approximate the kam - curves of the classically mixed system , fig . [ fig : deriv : quantum_maps](a ) , in phase space .
this resemblance is not possible with arbitrary precision as the integrable approximation does not contain , e.g. , nonlinear resonance chains and small embedded chaotic regions .
moreover , it can not account for the hierarchical regular - to - chaotic transition region at the border of the regular island .
similar problems appear for the analytic continuation of a regular torus into complex space due to the existence of natural boundaries @xcite .
however , for not too small @xmath1 , where these small structures are not yet resolved , an integrable approximation with finite accuracy turns out to be sufficient for a prediction of the tunneling rates . on the other hand
the integrable dynamics of @xmath111 should extrapolate smoothly beyond the regular island of @xmath112 .
this is essential for the quantum eigenstates of @xmath111 to have correctly decaying tunneling tails which are according to eq .
relevant for the determination of the tunneling rates .
while typically tunneling from the regular island occurs to regions within the chaotic sea close to the border of the regular island , there exist other cases , where it occurs to regions deeper inside the chaotic sea , as studied in ref .
@xcite . here
@xmath111 has to be constructed such that its eigenstates have the correct tunneling tails up to this region . for quantum maps we determine the fictitious integrable system in the following way :
we employ classical methods , see below , to obtain a one - dimensional time - independent hamiltonian @xmath113 which is integrable by definition and resembles the classically regular motion corresponding to the mixed system .
after its quantization we obtain the regular quantum map @xmath114 which has the same eigenfunctions @xmath93 as @xmath111 . for the numerical evaluation of eq .
we use @xmath115 , according to eqs . and , where the sum extends over @xmath116 .
two examples for the explicit construction of @xmath111 will be mentioned below .
note , that also other methods , e.g. based on the normal - form analysis @xcite or on the campbell - baker - hausdorff formula @xcite can be employed in order to find @xmath111 . for the example systems considered in this manuscript , however , they show less good agreement .
one possible choice for the determination of the fictitious integrable system for quantum maps is the lie - transformation method @xcite .
it determines a classical hamilton function as a series , @xmath117 see ref .
@xcite and fig .
[ fig : maps : comp_map_hreg_lie](a ) for an example . typically , the order of the expansion @xmath118 can be increased up to @xmath119 within reasonable numerical effort .
the lie - transformation method provides a regular approximation @xmath111 which interpolates the dynamics inside the regular region and gives a smooth continuation into the chaotic sea . at some order @xmath118
the series should diverge due to the nonlinear resonances inside the regular island .
for strongly perturbed systems , such as the standard map at @xmath120 , the lie - transformation method may not be able to reproduce the regular dynamics of @xmath59 , see fig .
[ fig : maps : comp_map_hreg_lie](b ) .
here we use a method @xcite based on the frequency map analysis @xcite . and of the corresponding integrable system ( thin lines ) of order @xmath121 .
here @xmath111 accurately resembles the regular dynamics of @xmath59 .
( b ) orbits of the standard map for @xmath122 ( gray ) and of the corresponding integrable system ( thin lines ) of order @xmath123 . here
@xmath111 does not accurately resemble the regular dynamics of @xmath112 . ]
an important question is whether the direct tunneling rates obtained using eq .
depend on the actual choice of @xmath111 and how these results converge depending on the order @xmath118 of its perturbation series .
ideally one would like to use classical measures , which describe the deviations of the regular system @xmath111 from the originally mixed one , to predict the error of eq .
for the tunneling rates .
however , these classical measures can only account for the deviations within the regular region but not for the quality of the continuation of @xmath111 beyond the regular island of @xmath59 .
currently , the quality of an integrable system can be estimated a posteriori by comparison of the predicted tunneling rates with numerical data .
it remains an open question how to obtain a direct connection between the error on the classical side and the one for the tunneling rates .
for an analytical evaluation of the result , eq .
, of the direct regular - to - chaotic tunneling rates we approximate the fictitious integrable system @xmath92 by a kicked system , @xmath124 or @xmath125 with @xmath126 here the functions @xmath127 and @xmath128 are a low order taylor expansion of @xmath6 and @xmath5 , respectively , around the center of the regular island .
note , that the classical dynamics corresponding to @xmath129 is typically not completely regular .
still the following derivation is applicable if @xmath129 has the following properties : ( i ) within the regular island it has an almost identical classical dynamics as @xmath59 , including nonlinear resonances and small embedded chaotic regions .
( ii ) it shows mainly regular dynamics for a sufficiently wide region beyond the border of the regular island of @xmath59 .
now we consider the specific case @xmath124 and assume that both properties ( i ) and ( ii ) are fulfilled .
as the dynamics of @xmath59 and @xmath129 are almost identical within the regular island of @xmath59 , the approximate result , eq .
, can be applied with @xmath92 replaced by @xmath129 , giving @xmath130 we now use that @xmath93 , which is an eigenstate of the exact @xmath92 and @xmath111 , is an approximate eigenstate of @xmath131 , leading to @xmath132 . with this
we obtain @xmath133 in position representation this reads @xmath134,\end{aligned}\ ] ] where @xmath135 and @xmath136 . in the semiclassical limit
the sum in eq . can be replaced by an integral @xmath137,\end{aligned}\ ] ] where we integrate over the whole position space @xmath11 $ ] .
note , that for the complementary situation , where @xmath138 is used in eq . , a similar result can be obtained in momentum representation @xmath139,\end{aligned}\ ] ] with @xmath140 .
we now use a wkb expression for the regular states @xmath93 . for simplicity
we restrict to the case @xmath141 leading to @xmath142 which is valid for @xmath143 . here
@xmath144 is the right classical turning point of the @xmath87th quantizing torus , @xmath145 is the oscillation frequency , and @xmath146 . the eigenstates @xmath147 decay exponentially beyond the classical turning point @xmath144 .
the difference of the potential energies @xmath148 approximately vanishes within the regular region and increases beyond its border to the chaotic sea .
hence , the most important contribution in eq .
arises near the left or the right border , @xmath149 or @xmath150 , of the regular island . for @xmath151 we rewrite the regular states @xmath152 where in the last step we use @xmath153 in the vicinity of the border . in order to use eq .
we split the integration interval in two parts , such that @xmath154 , corresponding to the contributions from the left and the right .
we now approximate @xmath155 with some constant @xmath156 and find @xmath157\,{\text{d}}x\\ \label{eq : maps : sc : tunnelrate_randformel_intnorm } & \approx & \frac{i{h_{\text{eff}}}}{\pi } \vert{\psi_{\text{reg}}}^m(q_{b}^{r})\vert^2 , \end{aligned}\ ] ] where @xmath158 and @xmath159 . in the semiclassical limit @xmath160 and for fixed quantum number
@xmath87 the integral in eq . , @xmath161\,{\text{d}}x,\end{aligned}\ ] ]
becomes an @xmath1-independent constant .
the tunneling rate @xmath162 is proportional to the square of the modulus of the regular wave function at the right border @xmath150 of the regular island . with eq .
we obtain @xmath163 a similar equation holds for @xmath164 . as an example for the explicit evaluation of eq .
we now consider the harmonic oscillator @xmath165 , where @xmath145 denotes the oscillation frequency and gives the ratio of the two half axes of the elliptic invariant tori .
its classical turning points @xmath166 , the eigenenergies @xmath167 , and the momentum @xmath168 are explicitly given .
using these expressions in eq . and @xmath169 we obtain @xmath170\bigg)\end{aligned}\ ] ] as the semiclassical prediction for the tunneling rate of the @xmath87th regular state , where @xmath171 , @xmath172 , and @xmath2 is the area of the regular island .
the exponent in eq . was also derived in ref .
the prefactor @xmath173 can be estimated semiclassically by solving the integral , eq . , for @xmath160 . for a fixed classical torus of energy @xmath174 one
obtains @xmath175 with this prefactor the prediction eq . gives excellent agreement with numerically determined data over @xmath176 orders of magnitude in @xmath177 , see fig .
[ fig : maps : rates_amph_ellipt2 ] . for a fixed quantum number @xmath87 the energy @xmath178 goes to zero in the semiclassical limit such that one can approximate @xmath179 in eq . which does not depend on @xmath1 .
let us make the following remarks concerning eq .
: the only information about this non - generic island with constant rotation number is @xmath180 as in ref . @xcite .
in contrast to eq . it does not require further quantum information such as the quantum map @xmath59 .
while the term in square brackets semiclassically approaches one , it is relevant for large @xmath1 .
in contrast to eq . , where the chaotic properties are contained in the difference @xmath181 , they now appear in the prefactor @xmath182 via the linear approximation of this difference . in the semiclassical limit the tunneling rates predicted by eq .
decrease exponentially .
for @xmath69 the values @xmath183 go to zero and @xmath184 to one , such that @xmath185 remains which reproduces the qualitative prediction obtained in ref .
we find that the non - universal constant in the exponent is @xmath186 which is comparable to the prefactor @xmath187 derived in refs .
however , our result shows more accurate agreement to numerical data and does not require an additional fitting parameter , as will be shown in sec .
[ sec : ex ] .
we study the direct regular - to - chaotic tunneling process in the systems introduced in sec . [
sec : kicked ] starting with the simplest example @xmath41 with a harmonic oscillator - like regular island . as further systems
, we consider the map @xmath43 with a deformed regular region , the map @xmath47 with a regular stripe , and finally the standard map which shows a generic mixed phase space .
the map @xmath41 has a particularly simple phase space structure with a harmonic oscillator - like regular island with elliptic invariant tori and constant rotation number , see the insets in fig .
[ fig : maps : rates_amph_ellipt ] .
numerically , we determine tunneling rates by using absorbing boundary conditions at @xmath188 . analytically , for the approach derived in sec . [
sec : dirtun ] , eq .
, we use the hamiltonian of a harmonic oscillator as the fictitious integrable system @xmath92 .
it is squeezed and tilted according to the linearized dynamics in the vicinity of the stable fixed point located at the center of the regular island . figure [ fig : maps : rates_amph_ellipt ] shows the prediction of eq .
compared to numerical data .
we find excellent agreement over more than @xmath176 orders of magnitude in @xmath177 . in the regime of large tunneling rates @xmath177 small deviations occur which can be attributed to the influence of the chaotic sea on the regular states : these states are located on quantizing tori close to the border of the regular island and are affected by the regular - to - chaotic transition region .
however , the deviations in this regime are smaller than a factor of two . in fig .
[ fig : maps : rates_amph_ellipt2 ] we compare the results of the semiclassical prediction , eq . , to the numerical data .
due to the approximations performed in the derivation of this formula stronger deviations are visible in the regime of large tunneling rates while the agreement in the semiclassical regime is still excellent . in refs .
@xcite a prediction was derived for the tunneling rate of the regular ground state , @xmath189 where @xmath190 is the incomplete gamma function , @xmath191 , and @xmath182 is a constant .
can be approximated semiclassically ( see ref .
@xcite ) , leading to @xmath192 figure [ fig : maps : rates_amph_ellipt2 ] shows the comparison of eq .
( dotted line ) to the numerical data for the map @xmath41 . especially in the semiclassical limit deviations are visible . the factor @xmath186 which appears in the exponent of eq .
is more accurate than the factor @xmath193 in eq . . for the map @xmath41 with a harmonic oscillator - like island
comparison with eq .
( solid lines ) .
the prediction of refs .
@xcite , eq . , for @xmath71 with a fitted prefactor
is shown ( dotted line ) . ] in typical systems the rotation number of regular tori changes from the center of the regular region to its border which typically has a non - elliptic shape .
such a situation can be achieved using the map @xmath43 . here
nonlinear resonances are small such that their influence on the tunneling process is expected only at large @xmath194 , see the inset in fig .
[ fig : maps : rates_amph_dist ] for its phase space .
we determine the fictitious integrable system @xmath111 by means of the lie - transformation method described in sec .
[ sec : dirtun : regsyst ] .
it is then quantized and its eigenfunctions are determined numerically . fig . [
fig : maps : rates_amph_dist ] shows a comparison of numerically determined tunneling rates ( dots ) to the prediction of eq .
( solid lines ) yielding excellent agreement for tunneling rates @xmath195 . for smaller values of @xmath177 deviations occur due to resonance - assisted tunneling which is caused by a small @xmath176:@xmath196 resonance chain .
similar to the case of the harmonic oscillator - like island the fictitious integrable system @xmath92 can be approximated by a kicked system @xmath197 using @xmath198 .
hence , eqs . and can be evaluated giving similarly good agreement ( not shown ) .
the prediction of eq .
@xcite ( dotted line ) shows large deviations to the numerical data . : numerical results ( dots ) and prediction following from eq .
( lines ) vs @xmath199 for quantum numbers @xmath200 .
the insets show husimi representations of the regular states @xmath71 and @xmath84 at @xmath201 .
the prediction of refs .
@xcite , eq . , for @xmath71 with a fitted prefactor
is shown ( dotted line ) . ]
we now consider the kicked system @xmath47 , for which the regular region consists of a stripe in phase space , see the inset in fig .
[ fig : maps : rates_shudo_map ] . in ref .
@xcite this map is used to study the evolution of a wave packet initially started in the regular region by means of complex paths . also for this system one can predict tunneling rates by means of eq . .
the fictitious integrable system @xmath92 is determined by continuing the dynamics within @xmath202 to the whole phase space .
it is given as a kicked system , eq . ,
defined by the functions @xmath203 for sufficiently large smoothing parameter @xmath58 , see eq . , the dynamics of the mixed map inside the regular region is equivalent to that of the regular map .
. can be applied and we compare its results to numerically determined data .
absorbing boundary conditions at @xmath188 lead to strong fluctuations of the numerically determined tunneling rates as a function of @xmath1 , presumably due to dynamical localization .
choosing @xmath204 for the opening , which is closer to the regular stripe , we find smoothly decaying tunneling rates , see fig . [
fig : maps : rates_shudo_map ] .
the comparison with the theoretical prediction shows quite good agreement .
note , that due to the symmetry of the map there are always two regular states with comparable tunneling rates except for the ground state @xmath71 .
these two states are located symmetrically around the center of the regular stripe . while the theoretical prediction , eq . , is identical for both of these states , the numerical results differ slightly due to the different chaotic dynamics in the vicinity of the left and right border of the regular region .
: we compare numerical results ( dots ) and the prediction following from eq .
( lines ) vs @xmath194 for the quantum numbers @xmath205 .
the inset shows the phase space of the system .
the numerical data is obtained using an absorbing region at @xmath204 ( gray - shaded area of the inset ) . ] the paradigmatic model of an area preserving map is the standard map , see sec .
[ sec : kicked : standard ] .
for @xmath16 between @xmath206 and @xmath207 one has a large generic regular island with a relatively small hierarchical region surrounded by a @xmath208:@xmath196 resonance chain , see the inset in fig .
[ fig : maps : rates_stmap ] .
absorbing boundary conditions at @xmath188 lead to strong fluctuations of the numerically determined tunneling rates as a function of @xmath1 , presumably caused by partial barriers . choosing absorbing boundary conditions at @xmath204 , which is closer to the island , we find smoothly decaying tunneling rates ( dots in fig . [ fig : maps : rates_stmap ] ) .
evaluating eq . for @xmath122
gives good agreement with these numerical data , see fig .
[ fig : maps : rates_stmap ] ( solid lines ) . here
we determine @xmath111 using a method based on the frequency map analysis , as the lie transformation is not able to reproduce the dynamics within the regular island of @xmath59 , see sec .
[ sec : dirtun : regsyst ] . with increasing order of the expansion series of @xmath111 the tunneling rates following from eq .
hence , for the predictions in fig .
[ fig : maps : rates_stmap ] we choose terms up to second order only .
note , that at such small order the accuracy of @xmath111 within the regular region of @xmath59 is inferior compared to the examples discussed before .
hence , in eq .
the state @xmath209 has small contributions of all purely regular states @xmath210 in the regular island .
these contributions are removed by the application of the projector @xmath98 . however , this projector depends on the number of regular states @xmath74 , which grows with @xmath194 .
if @xmath74 increases by one , @xmath97 projets onto a larger region in phase space .
this explains the steps of the theoretical prediction , eq .
, visible in fig .
[ fig : maps : rates_stmap ] . ) for @xmath211 .
prediction of eq .
( lines ) and numerical results ( dots ) , obtained using an absorbing region at @xmath204 ( gray - shaded area of the inset ) . ]
dynamical tunneling plays an important role in many areas of physics . therefore a detailed understanding and quantitative description is of great interest . in this text
we have given an overview on determining direct regular - to - chaotic tunneling rates using the fictitious integrable system approach .
the direct regular - to - chaotic tunneling mechanism is valid in the regime where planck s constant is large compared to additional structures in the regular island , such as nonlinear resonances . to include the effect of such resonances , resonance - assisted tunneling @xcite has to be considered in addition with the direct regular - to - chaotic tunneling contribution .
recently , the two mechanisms have been studied in a combined theory @xcite , giving a full quantitative description of tunneling from a regular island , including resonances , into the chaotic region . the approach to determine direct regular - to - chaotic tunneling rates can also be generalized to the case of billiard systems .
it can be applied if the fictitious integrable system @xmath111 is known , as e.g. for the annular or the mushroom billiard .
in addition to @xmath111 a description for the chaotic states of the mixed system @xmath112 is needed , for which we employ random wave models @xcite which account for the relevant boundary conditions of the billiard . for the mushroom billiard
the fictitious integrable system is easily found as the semi - circle billiard and with this an analytical expression for the tunneling rates was derived in ref .
this result has been compared to experimental data obtained from microwave spectra of a mushroom billiard with adjustable stem height and good agreement was found .
it was also shown that tunneling rates manifest themselves in exponentially diverging localization lengths of nanowires with one - sided surface roughness in a perpendicular magnetic field @xcite .
the fictitious integrable system approach has been extended to open optical microcavities in ref .
in particular the annular microcavity was studied , which allows for unidirectional emission of light and shows modes of high quality factors simultaneously .
this is desirable for most applications .
in contrast to closed billiards the leakiness of the cavity has to be considered , which leads to the contribution @xmath212 to the quality factor caused by the direct coupling of the regular mode to the continuum .
additionally , the contribution caused by dynamical tunneling @xmath213 is relevant , where @xmath214 .
@xmath213 is directly related to the dynamical tunneling rates given by the theory using a fictitious integrable system .
the prediction for the quality factors @xmath215 has been compared to numerical data .
excellent agreement is found if no further phase - space structures exist in the chaotic sea . if additional structures appear , the numerical data show oscillations , which can not be explained by the present theory .
future challenges include a completely semiclassical prediction of direct regular - to - chaotic tunneling rates in generic systems , the understanding of implications of additional phase space structures , and the extension to higher dimensional systems .
we are grateful to s. creagh , s. fishman , m. hentschel , r. hhmann , a. khler , u. kuhl , a. mouchet , m. robnik , p. schlagheck , a. shudo , h .- j .
stckmann , s. tomsovic , g. vidmar , and j. wiersig for stimulating discussions .
we further acknowledge financial support through the dfg forschergruppe 760 `` scattering systems with complex dynamics '' .
a. bcker , r. ketzmerick , and s. lck : _ direct regular - to - chaotic tunneling rates using the fictitious integrable system approach _ , in _ dynamical tunneling - theory and experiment _ , edited by s. keshavamurthy and p. schlagheck , pp .
119 - 137 , taylor and francis crc ( 2011 ) . c. dembowski , h .-
grf , a. heine , r. hofferbert , h. rehfeld , and a. richter : _ first experimental evidence for chaos - assisted tunneling in a microwave annular billard _
lett . * 84 * , 867870 ( 2000 ) .
w. k. hensinger , h. hffner , a. browaeys , n. r. heckenberg , k. helmerson , c. mckenzie , g. j. milburn , w. d. phillips , s. l. rolston , h. rubinsztein - dunlop , and b. upcroft : _ dynamical tunnelling of ultracold atoms _ , nature * 412 * , 5255 ( 2001 ) .
p. schlagheck , c. eltschka , and d. ullmo : _ resonance- and chaos - assisted tunneling _ , in _ progress in ultrafast intense laser science
i _ , edited by k. yamanouchi , s. l. chin , p. agostini , and g. ferrante , springer , berlin ( 2006 ) .
j. feist , a. bcker , r. ketzmerick , s. rotter , b. huckestein , and j. burgdrfer : _ nano - wires with surface disorder : giant localization lengths and quantum - to - classical crossover _ , phys .
lett . * 97 * , 116804 ( 2006 ) .
j. feist , a. bcker , r. ketzmerick , j. burgdrfer , and s. rotter : _ nanowires with surface disorder : giant localization length and dynamical tunneling in the presence of directed chaos _ , phys .
b * 80 * , 245322 ( 2009 ) . | in systems with a mixed phase space , where regular and chaotic motion coexists , regular states are coupled to the chaotic region by dynamical tunneling .
we give an overview on the determination of direct regular - to - chaotic tunneling rates using the fictitious integrable system approach .
this approach is applied to different kicked systems , including the standard map , and successfully compared with numerical data .
_ this text corresponds to chapter 6 of the book : dynamical tunneling - theory and experiment , edited by s. keshavamurthy and p. schlagheck [ taylor and francis crc ( 2011 ) ] @xcite . for a more extensive exposition see [ phys .
rev .
e 82 , 056208 ( 2010 ) ; arxiv:1009.0418v2 ] . |
massive stars tend to form in clusters , which arise from the collapse and fragmentation of molecular clouds @xcite . in this process
, self - gravity is certainly a necessity , since without it interstellar gas would not condense to form stars .
however , a simplified theory dealing with the interplay between gravity and thermal pressure , e.g. , the classic jeans theory , fails to explain either the birth of massive stars nor the low rate of global star formation . the key question under debate is that which mechanism , turbulence or magnetic fields , plays a central role in counteracting with gravity @xcite . in the scenario of gravo - turbulent fragmentation @xcite ,
supersonic turbulence controls the evolution of molecular clouds , producing a hierarchy of density structures ; as turbulence decays locally , gravitational collapse sets in to give birth to a proto - cluster .
stars near the cloud center could accrete gas at a higher rate , thus they end up being more massive .
magnetic fields are implicitly weak in this scenario , and are predicted to show a chaotic morphology due to turbulent and dynamical twisting ( e.g. , * ? ? ?
in contrast , magnetic fields are crucial in the classic theory of isolated low - mass star formation @xcite .
the theory assumes that molecular clouds are magnetically supported ( e.g. , * ? ? ?
* ) . in collapsing cores where star formation takes place ,
magnetic fields are pulled by gravity into an hourglass shape ; cores contract more along magnetic fields , resulting in a flattened morphology perpendicular to the field axis , and a bipolar outflow is driven along the magnetic field from an accretion disk embedded within the core @xcite .
observationally , hourglass magnetic fields aligned with bipolar outflows have been detected in a few low - mass objects @xcite , though a recent study found that magnetic fields may be statistically misaligned with outflows @xcite . in the high - mass regime
, several high - angular - resolution observations of polarized dust emission have revealed magnetic fields with a variety of morphologies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
yet a single source , g31.41 + 0.31 , has provided the most compelling evidence for an hourglass - like magnetic field @xcite ; but the properties of gas kinematics and outflow structures are still unclear @xcite . furthermore , there is hitherto no reported observational evidence for an hourglass magnetic field in a cluster - forming region . here
we report submillimeter array ( sma ) observations toward g240.31 + 0.07 ( hereafter g240 ) , which is an active high - mass star - forming region with a bolometric luminosity of @xmath5 at a distance of 5.32 kpc @xcite .
we detect a clear hourglass magnetic field threading a massive and flattened core that is powering a well - shaped bipolar outflow and fragmenting to form a proto - cluster .
as part of an sma polarization legacy project @xcite , the observations were undertaken in five tracks , each with @xmath324 hours on - source integration toward the phase referencing center ( r.a .
, dec.)@xmath6=(@xmath7 ) . during the observations the weather conditions were excellent , with @xmath8 mm precipitable water vapor , corresponding to atmospheric opacities at 225 ghz , @xmath9 , lower than 0.06 .
the detailed information of the observations is presented in table [ table1 ] .
the correlator setup covers approximately 332336 ghz in the lower sideband and 344348 ghz in the upper sideband , with a uniform spectral resolution of 812.5 khz ( @xmath30.7 kms@xmath10 ) .
we performed basic data calibration , including bandpass , time dependent gain , and flux calibration , with the idl mir package , and output the data into miriad for further processing .
the intrinsic instrumental polarization was removed to a 0.1% accuracy with the miriad task gpcal .
we jointly imaged the calibrated visibilities from all five tracks to make stokes @xmath11 , @xmath12 , and @xmath13 maps .
the continuum maps were made with a natural weighting of the visibilities to obtain an optimized sensitivity , leading to a synthesized beam of @xmath14 at fwhm and an rms noise level of 0.85 mjybeam@xmath10 for stokes @xmath12 , @xmath13 and 5 mjybeam@xmath10 for stokes @xmath11 ; the latter is affected by the limited dynamic range .
the stokes @xmath12 , @xmath13 maps were then combined to produce maps of the linear polarization intensity , the fractional polarization , and the polarization angle ( p.a . ) .
spectral line maps were made with a compromise of angular resolution and sensitivity , and the synthesized beam is about @xmath15 and the rms noise level is @xmath16 mjybeam@xmath10 per 0.8 kms@xmath10 .
no significant polarization is detected in spectral lines , so only stokes @xmath11 maps are presented .
we observed g240 using the sma in its sub - compact , compact , and extended configurations . in figure
[ polmag ] , the total dust emission ( stokes @xmath11 ) reveals a flattened core , which is embedded with at least three dense condensations previously detected at 1.3 mm at a @xmath31@xmath17 resolution @xcite .
the detected polarized emission has a butterfly - like distribution ( figure [ polmag]a ) , and the inferred magnetic field exhibits a clear hourglass shape aligned with the minor axis of the core ( figure [ polmag]b ) .
we fit the magnetic field morphology with an analytical model of a series of parabolic functions ( e.g. , * ? ? ?
* ; * ? ? ?
the fitted parameters are the position angle of the magnetic field axis @xmath18 , the center of symmetry of the magnetic field ( r.a . ,
dec.)@xmath6=(@xmath19 ) , and @xmath20 for the parabolic form @xmath21 , where @xmath22 is the distance in units of arc second along the magnetic field axis from the center of symmetry . from figure [ polmag]c
, the best - fit model agrees well with the measured magnetic field .
the residuals from the fit have a nearly gaussian distribution , and yield a dispersion of @xmath23 ( figure [ polmag]d ) . accounting for a measurement uncertainty of @xmath24 ,
the intrinsic dispersion of the polarization angle , @xmath25 , is @xmath26 .
g240 harbors a well - shaped wide - angle outflow previously observed in co @xcite , and the position angle of the outflow axis is about @xmath27 .
so the bipolar outflow is roughly aligned with ( within @xmath0 ) the magnetic field axis ( figure [ kinematics]a ) .
it is already clear that we are observing an hourglass magnetic field in a massive cluster - forming core .
we further investigate the kinematics of the dense gas by examining the simultaneously observed spectral lines .
figures [ kinematics]b and [ kinematics]c show the first moment map and the position - velocity ( pv ) diagram of the h@xmath1co@xmath2 ( 43 ) emission : there is a clear velocity gradient along the major axis of the core . the velocity gradient is also detected in other high - density tracing spectral lines , and provides strong evidence for a rotating motion of the core .
we perform a linear fit to the peak positions of the h@xmath1co@xmath2 emission at each velocity channel , and derive a position angle of @xmath28 for the rotation axis , which is only @xmath29 offset from the magnetic field axis .
using the chandrasekhar - fermi ( cf ) method @xcite , the strength of the magnetic field in the plane of sky , @xmath30 , can be estimated based on the p.a .
dispersion , following @xmath31 , where @xmath12 is a parameter related to the cloud structure and is adopted to be 0.5 @xcite , @xmath32 is the turbulent velocity dispersion along the line of sight , and @xmath33 is the averaged mass density . in g240 , @xmath25 is inferred to be @xmath34 ( section [ res_mag ] ) , and @xmath32 reaches 1.2 kms@xmath10 from the h@xmath1co@xmath2 ( 43 ) data . to estimate the mass density , we measure the total flux of the dust emission to be 1.4 jy over an area of 60 square arcsec , where the emission is detected with s / n ratios above 3 . with a dust temperature of 47 k @xcite , a dust emissivity index of 1.5 @xcite , and a canonical gas - to - dust mass ratio of 100 , the core mass amounts to 95 @xmath35 , leading to an averaged column density , @xmath36 , of @xmath37 @xmath38 and a volume density , @xmath39 , of @xmath40 @xmath41 .
consequently , we deduce @xmath42 mg .
heitsch et al .
( 2001 ) show that the cf method is reliable to a factor of 2 for magnetic field strengths typical of molecular clouds , as long as the angular structure in the field is well resolved ; limited angular resolutions lead to an overestimation of the field strength . considering the presence of a velocity gradient across the core , the turbulent velocity dispersion inferred from the h@xmath1co@xmath2 emission
could be overestimated , resulting in an overestimation of the field strength . on the other hand ,
the plane - of - sky component of the field strength represents a lower limit to the total field strength .
thus , the cf method could still provide a reasonable estimate of the field strength , but the reliability needs to be verified .
since the magnetic field has been pulled into an hourglass shape by the gravitational collapse , we can gauge the magnetic field strength from the observed curvature of the field lines . considering the gravity and magnetic tension force acting on a field line
, the magnetic field strength can be estimated following an equation derived by schleuning ( 1998 ) : @xmath43 where @xmath44 is the radius of the curvature , @xmath45 is the distance from the field line to the symmetry center , @xmath46 is the mass , which we adopt a value of 125 @xmath35 including the gas mass and the mass of the forming massive stars ( @xmath330 @xmath35 , * ? ? ?
* ) deeply embedded within the core . from the farthest southwestern modeled line in figure [ polmag]b , which has @xmath47 pc and @xmath48 pc ,
we obtain a field strength of 0.9 mg , which is in good agreement with the above estimate using the cf method . assuming that the overall direction of the magnetic field in g240 is close to the plane of sky and @xmath30 approximates the total field strength , @xmath49
, we can assess the importance of the magnetic field compared to gravity , turbulence , and rotation .
the key parameter that determines whether a magnetized core would gravitationally collapse is the mass - to - magnetic flux ratio .
if the ratio is greater than the critical value , @xmath50 , where @xmath51 is the gravitational constant , the core is unstable and would collapse under its self - gravity ( nakano & nakamura 1978 ) .
we find the ratio in g240 is about 1.4 times the critical value , taking into account both the gas mass and the stellar mass .
the core is thus slightly supercritical , and the gravity has started to overcome the magnetic force to initiate the collapse .
the turbulent to ordered magnetic energy ratio , @xmath52 , can be estimated from @xmath53 , where @xmath54 is the alfvn velocity . with @xmath55
, we have @xmath56 , indicating that the ordered magnetic energy is dominating over the turbulent energy .
the core is observed to be rotating ( section [ res_kin ] ) .
for a collapsing and rotating molecular cloud core threaded by a strong magnetic field , magnetic braking is expected to channel out part of the angular momentum of the infalling matter and cause the decrease in the specific angular momentum . to check this effect , we compare the h@xmath1co@xmath2 data from this work with our new observations made with the sma and the atacama pathfinder experiment ( apex ) . in figures [ angmom]a
f , we show the first moment maps and pv diagrams of so ( @xmath57@xmath58 ) and hco@xmath2 ( 43 ) .
the so maps were constructed from the sma very - extended ( vex ) observations and from the combined sma compact and extended ( com+ext ) observations , resulting in angular resolutions of @xmath59 and @xmath60 , respectively .
the apex hco@xmath2 data have an angular resolution of @xmath61 .
a velocity gradient consistent with what is shown in figure [ kinematics ] is seen on all the maps .
we measure the rotation velocities by applying the terminal velocity method @xcite , where the terminal velocity , @xmath62 , is determined from the pv diagram by a velocity at which the intensity equals to 20% of the peak intensity ( e.g. , * ? ? ?
* ) . to mitigate the optical depth effect and the confusion from the gas in outer layers along the line of sight
, we only measure the outermost half beam at fwhm of each pv diagram . with all the data , including the h@xmath1co@xmath2 observations shown in figure [ kinematics ] , we obtain @xmath62 on @xmath630.6 pc , where the radius , @xmath64 , is the distance from the center of the core along the pv cut , allowing us to examine how the specific angular momentum , @xmath65 , where @xmath66 ( @xmath67 ) is the angular velocity , changes with the radius . in figure
[ angmom ] , @xmath65 does decrease toward the center .
a power - law fit to the data yields a @xmath68 dependence of the angular momentum on the radius , which is very likely attributed to magnetic braking . in a numerical study ,
magnetic braking acts to align the rotation axis and magnetic field , while the rotation causes the magnetic field to incline through dynamo action @xcite . when the ratio of the angular velocity to magnetic field strength , @xmath69 , is smaller than a characteristic value of @xmath70 , the former effect dominates the latter . in g240 , the rotation axis
is observed to be closely aligned with the magnetic field direction ( figure [ kinematics]b ) , suggesting that magnetic braking is more important .
indeed , with @xmath71 year@xmath10 from the h@xmath1co@xmath2 data and @xmath72 kms@xmath10 , @xmath69 is @xmath73 year@xmath74 , smaller than the characteristics value of @xmath75 year@xmath74 . hence , magnetic braking ,
if it is at work as manifested by the decrease of the specific angular momentum toward the center ( e.g. , * ? ?
* ) , should be dominating over the rotation in shaping the dynamics of the core .
we summarize the observed magnetic field and kinematic properties in g240 in figure [ cartoon ] .
the picture is surprisingly consistent with the theoretical predictions of the classic paradigm of isolated low - mass star formation @xcite , except that g240 is a luminous and massive star - forming core and shows a higher degree of fragmentation .
it is the first time that a clear hourglass magnetic field aligned with a well - defined outflow - rotation system is detected in the high - mass regime .
our observations provide strong evidence that massive star and cluster formation in g240 is proceeding in a way essentially similar to the formation of sun - like stars .
we also find that the magnetic field dominates over turbulence in shaping the dynamics of the massive core .
the nearly linear decrease of the specific angular momentum toward the center suggests that magnetic braking is at work and overcomes the centrifugal force to align the rotation axis .
we thank the sma staff for their supports which make this study possible .
part of this work was conducted when kq was doing a postdoc at max - planck - institut fr radioastronomie .
kq is partially supported by the 985 project of nanjing university .
qz is partially supported by the nsfc grant 11328301 .
jmg is supported by the spanish mineco aya2011 - 30228-c03 - 02 , and catalan agaur 2009sgr1172 grants .
allen , a. , shu , f. h. , & shu , f. h. 2003 , , 599 , 351 cesaroni , r. , beltrn , m. , zhang , q. , beuther , h. , & fallscheer , c. 2011 , , 533 , 73 chandrasekhar , s. & fermi , e. 1953 , , 118 , 113 chen , h .-
2007 , , 654 , l87 choi , y. k. et al .
2014 , , 790 , 99 crutcher , r. 2006 , science 313 , 771 crutcher , r. 2012 , , 50 , 29 elmegreen , b. g. , efremov . y. , pudritz , r. e. & zinnecker , h. 2000 , in protostars and planets iv , ed . v. mannings et al .
( tucson : univ .
arizona press ) frau , p. , galli , d. , & girart , j. m. 2011 , , 535 , 44 galli , d. & shu , f. h. 1993 , , 417 , 220 girart , j. m. , beltrn , m. t. , zhang , q. , rao , r. , & estalella , r. 2009 , science , 324 , 1408 girart , j. m. et al .
2013 , , 772 , 69 girart , j. m. , rao , r. , & marrone , d. p. 2006 , science , 313 , 812 gonalves , j. , galli , d. , & girart , j. m. 2008 , , 490 , l39 heitsch , f. , zweibel , e. g. , mac low , m .-
m . , li , p , & norman , m. l. 2001 , , 561 , 800 hopkins , p. f. 2013 , , 430 , 1653 hull , c. l. h. et al .
2013 , , 768 , 159 hunter , t. r. 1997 , phd thesis , california institute of technology klessen , r. s. , ballesteros - paredes , j. , vzquez - semadeni , e. & durn - rojas , c. 2005 , , 620 , 786 liu , h. b. , zhang , q. , girart , j. m. , & ho , p. t. p. 2013
, , 771 , 71 machida , m. n. , matsumoto , t. , hanawa , t. , & tomisaka , k. 2006 , , 645 , 1227 mac - low , m .-
& klessen , r. s. 2004 , review of modern physics , 76 , 125 moscadelli , l. et al . 2013 , , 549 , 122 mouschovias , t.c .
& ciolek , g. e. 1999 , in the origin of stars and planetary systems , ed .
lada , c. j. & kylafis , n. d. ( kluwer academic , dordrecht , netherlands ) ostriker , e. c. , stone , j. m. , & gammie , c. f. 2001 , , 546 , 980 padoan et al .
2001 , , 559 , 1005 qiu , k. , zhang , q. , menten , k. m. , liu , h. b. , & tang , y .- w . 2013 , , 779 , 182 qiu , k. , zhang , q. , wu , j. , & chen , h .- r .
2009 , , 696 , 66 rao , r. , girart , j. m. , marrone , d. p. , lai , s .-
p . , & schnee , s. 2009 , , 707 , 921 schleuning , d. a. 1998 , , 493 , 811 shu , f. h. , adams , f. c. , & lizano , s. 1987 , , 25 , 23 sofue , y. & rubin , v. 2001 , , 39 , 137 stephens , i. w. et al .
2013 , , 769 , l15 tang , y. w. et al . 2009a , , 695 , 1399 tang , y. w. et al . 2009b , , 700 , 251 trinidad , m. a. 2011 , , 142 , 147 zhang , q. et al .
2014 , , 792 , 116 llllll 2011 oct .
17 ( 4.0 ) & compact ( 7 ) & 0.03 & 3c84 & j0730 - 116 & callisto + 2011 dec . 10 ( 3.5 ) & compact ( 8) & 0.05 & 3c84 & j0730 - 116 & titan + 2012 jan .
06 ( 3.5 ) & subcompact ( 6 ) & 0.06 & 3c279 & j0730 - 116 & canymede + 2012 feb . 02
( 3.5 ) & extended ( 7 ) & 0.05 & 3c279 & j0730 - 116 & callisto + 2012 mar .
27 ( 2.5 ) & extended ( 7 ) & 0.08 & 3c279 & j0730 - 116 & 3c279 + | we report the first detection of an _ hourglass _ magnetic field aligned with a well - defined outflow - rotation system in a high - mass star - forming region .
the observations were performed with submillimeter array toward g240.31 + 0.07 , which harbors a massive , flattened , and fragmenting molecular cloud core and a wide - angle bipolar outflow .
the polarized dust emission at 0.88 mm reveals a clear hourglass - shaped magnetic field aligned within @xmath0 of the outflow axis .
maps of high - density tracing spectral lines , e.g. , h@xmath1co@xmath2 ( 43 ) , show that the core is rotating about its minor axis , which is also aligned with the magnetic field axis . therefore ,
both the magnetic field and kinematic properties observed in this region are surprisingly consistent with the theoretical predictions of the classic paradigm of isolated low - mass star formation .
the strength of the magnetic field in the plane of sky is estimated to be @xmath31.1 mg , resulting in a mass - to - magnetic flux ratio of 1.4 times the critical value and a turbulent to ordered magnetic energy ratio of 0.4 .
we also find that the specific angular momentum almost linearly decreases from @xmath4 pc to 0.03 pc scales , which is most likely attributed to magnetic braking . |
the understanding of nuclear matter under extreme conditions is a crucial and indispensable aim of nuclear and stellar physics , especially with much more experimental information to come with rhic and alice / lhc accelerators .
it is now widely believed that at sufficiently high energy , in heavy - ion collisions , a central hot region is formed @xcite .
this region is commonly associated with the existence of a quark - gluon plasma ( qgp ) .
following many prescriptions , this hot region expands , cools down and freezes out into hadrons , exhibiting a hadron - gas phase transition .
in order to observe this transition to the deconfined phase , some observable signatures have been proposed as , for example , electromagnetic radiation @xcite , strangeness enhancement @xcite , and j/@xmath0 suppression @xcite .
high - energy heavy - ion collisions data has brought still more excitation about the predictions of qcd on this subject .
in fact , based on the results from different experiments done at cern - sps , it is believed that a new state of matter has been formed , the quark matter @xcite . recently , phase transition signatures at resonance - rich matter in heavy ion collisions at rhic energies has been analyzed in central au + au collisions at @xmath1= 200 a gev @xcite .
these studies , using a microscopic transport model ( urqmd ) calculation , showed the equilibration of hot matter through a @xmath2x@xmath3 phase diagram . confronting these results with rhic ,
sps and ags data , it was shown that for central collisions the rising of the incident energy implies a strong evidence for hadronic - qgp phase transition at very small baryon chemical potentials and temperatures lying between 150 to 200 mev .
some time ago , using the resonance model , heinz et al .
@xcite , have studied the hadron - qgp phase transition at this regime .
they found a strong enhancement of the entropy from the hadron to qgp phase .
when they included all the resonances and mesons , the entropy of the hadron phase increased but a considerable entropy gap still remained between the two phases . in this paper
we will study this transition in relativistic mean field ( rmf ) models , where we have not only a repulsive vector interaction but also an attractive scalar interaction , that was not considered in the resonance model study @xcite .
this scalar interaction , as we will show in this paper , is quite important in order to study this hadron - qgp transition at low net baryon density . at this regime
there is no vector field and the pressure is composed only from the nucleon thermal gas and from the scalar field pressures .
the abrupt fall in the nucleon mass at high temperature , seen in many of the rmf hadronic models , is due to the abrupt increasing of the scalar field .
this effect resembles a phase transition when the system becomes a dilute gas of baryons in a sea of baryon - antibaryon pairs , and it is caracterized by a large enhancement of the entropy @xcite .
thus , our study in rmf models is quite important because the entropy enhancement depends strongly on the scalar interaction presented in these type of models .
it is important to stress , that in all these models the entropy enhancement is maximized at zero chemical potential because when we increase the baryon density , the fall of the nucleon mass with the temperature and also the increase of the entropy become smaller @xcite .
the importance of the entropy production in relativistic heavy ions collisions has also been addressed to analyze the qgp phase transition , where the entropy density jump at the phase boundary for low net baryon density has been pointed out @xcite . at the hadronic level ,
rmf models are able to describe well a number of nuclear phenomena through different conceptions of meson - nucleon and meson - meson couplings .
most of them follows the basic walecka model @xcite in which improvements have been done in order to better fit the finite nuclei data @xcite .
it is known that hadronic models may present phase transitions at higher temperatures .
for example , in reference @xcite , hot nuclear matter was extensively studied using the walecka model regarding the finite temperature behavior for zero baryon density ( nucleon - antinucleon plasma ) .
the authors explored this model for different scalar - vector coupling constants to conclude that , depending on the values taken by those constants , the walecka model may or may not generate a phase transition into a nucleon - antinucleon plasma .
the order ( first or second ) of this phase transition depends crucially on the scalar coupling constant of the model .
the question we pose in this paper is whether the phase transition for the hadronic sector itself , described from hadronic models , may or not have importance to the qgp phase transition . to answer this question we bring to this discussion a representative class of recent rmf hadronic models @xcite .
we have also considered derivative coupling models @xcite and different parametrizations of the non - linear walecka model such as nl1 , nl2 , nl3 and nlsh models very well known in the literature @xcite . most of these models were successfully employed to calculate nuclear matter bulk properties as well as the spectra of finite nuclei . the aim of our work is to present a systematic comparative study of a set of hadronic models , at extreme temperatures and very low net baryon density , and the hadronic - qgp phase transition , emphasizing the entropy density behaviour in both phases .
the qgp phase is represented by a perturbative qcd derivation @xcite .
we studied the hadronic - qgp phase transition for a broad range of bag constant @xmath4 and qcd running coupling constant @xmath5 values .
the particular case of free quarks ( @xmath6 ) is also explicited .
this kind of phase transition has been investigated by many authors @xcite .
in particular , aspects of the role played by the nucleon exclude volume have also been studied @xcite . as a result , we find that practically all the analyzed hadronic models signalize approximately the same critical qgp temperature .
this finding is quite independent whether the hadronic models have or not hadronic phase transition themselves at high temperature .
this suggests strongly a rmf hadronic model independence for the qgp phase transition in this regime .
to the best of our knowledge , this is a remarkable and unknown feature of the hadronic models .
we have traced back the reason for this and conclude that it comes from the fact that the qgp entropy ( the slope of the pressure versus temperature ) is much larger than the hadronic entropy obtained in all the hadronic models .
therefore , insofar the entropy through sophisticated calorimeters in the future may be a measured quantity , we believe that at very low net baryon density , the enhancement of the entropy may become a signature of a qgp .
previous works have already addressed the role played by the entropy at the qgp formation @xcite . we have observed a large jump of the hadronic entropy density , seen in almost all of the rmf models .
an enhancement of the entropy per baryon with the rising of the incident energy was also found in recent microscopic transport model ( urqmd ) calculations @xcite .
however , this model did not include a qgp phase but can be taken as leading credence for our hadronic model phase entropy enhancement signals .
this enhancement may favour the hadron - qgp phase transition once it decreases the latent heat for the qgp formation .
thus , we expect in central collisions with a low net baryon density , the formation of a very rich baryon - antibaryon matter in the region of the qgp phase .
this results is in agreement with recent data obtained in the star collaboration at rhic where in fact it was found a rich proton - antiproton matter @xcite .
the outline of this paper is as follows . in sec .
ii we present the phase transition calculation . in section
iii we show our results and discussion , followed by the conclusions in section iv .
at low temperatures and baryon densities ( around the nuclear matter ground state ) , renormalization group arguments @xcite show that the qcd running coupling constant is greater than one , indicating confinement of quarks and gluons inside hadrons which are the appropriate degrees of freedom of the nuclear matter in this regime . in this region ,
qhd based on effective lagrangians works quite well .
the phase transition is supposed to occur at sufficiently high temperatures and densities , where in this case the qcd running coupling constant becomes smaller than one , suggesting deconfinement of quarks and gluons .
if one assumes the validity of qhd far beyond the normal nuclear matter ground state , the transition can be characterized by the process of hadrons loosing their identities , and quarks and gluons becoming the elementary degrees of freedom . in order to see the hadronic - quark - gluon phase transition we need distinct models for the two different phases of the baryonic matter .
we describe the hadronic phase ( h ) by the models of ref .
@xcite in the hartree approximation extended to finite temperature at zero net baryon density . for the quark - gluon plasma phase , the equation of state
is given given by @xcite .
a study of the hadron - qgp phase diagram for finite baryonic density of the models presented in ref .
@xcite has been done in ref .
the gibbs criteria used in the analysis of the phase transition for zero density is given by @xmath8 and @xmath9 .
the functional , used for the hadronic phase , originates from an effective lagrangian , with additional physical constraints ensuring qcd symmetries @xcite .
the pressure is @xmath10 where @xmath11 and @xmath12 are the scalar and vector meson fields respectively , @xmath13 and @xmath14 stand for the fermi - dirac distribution for baryons ( antibaryons ) . the baryon density @xmath15 is defined by @xmath16 $ ] .
thus , in the zero net baryon density regime we have the same number of baryons and anti - baryons .
the constants ( see table i ) are already adjusted to reproduce some of the bulk properties of the nuclear matter ground state @xcite . here
@xmath17 stands for the usual walecka model . @xmath18 and @xmath19 are nonlinear @xmath20 models , having scalar field self - interaction added to the @xmath17 model in order to improve some nuclear matter bulk properties of @xmath17 .
most of the developed hadronic models has the @xmath18 structure , given by cubic and quartic scalar field self - interaction . in this aspect
it is a very representative model category .
the scalar cubic and quartic self - interaction are usually claimed to simulate three - body and four - body forces effects respectively .
the others as @xmath21 , @xmath22 and @xmath23 are models representing further improvements by including different scalar - scalar , vector - vector and scalar - vector fields couplings .
we have also considered for the hadronic phase derivative coupling models @xcite and different parametrizations of the nonlinear boguta - bodmer model @xcite , nl1 , nl2 , nl3 and nlsh models @xcite . for the quark - gluon plasma phase , at high temperature , the pressure and the density is given by the expansion @xcite @xmath24 - b \
, , \label{pqgp}\end{aligned}\ ] ] @xmath25 where @xmath4 is the bag constant , @xmath26 is the number of flavors , @xmath27 is the qcd running coupling constant , depending on the quark - gluon plasma temperature @xmath28 and the quark chemical potential @xmath29 through the first order perturbative expression @xmath30\right\}^{-1 } \ , .
\label{aqgp}\ ] ] at the zero density regime we are interested we set @xmath31 .
the model for the qgp phase which we use in this work has some parameters such as @xmath32 , @xmath4 and @xmath33 . as a first approximation
, we examine the simplest case with @xmath34 ( quarks @xmath35 and @xmath36 only ) .
the qcd scale parameter @xmath32 is fixed at 200 mev , consistent with the current data set .
we will use two different bag constants in our analysis : @xmath37 mev and @xmath38 mev , corresponding to @xmath39 mev @xmath40 and @xmath41 mev @xmath40 respectively .
these are the limiting values of the broad range of values used in the literature .
figure 1 shows the effective nucleon mass ratio @xmath42 _ vs. _
@xmath43 for the studied models .
we see that zm , zm3 , c1 and g2 models do not present signals of a first order phase transition in opposite to the others .
then , not all of the hadronic models considered here ( twelve in total ) have the hot nucleon - antinucleon phase transition in the hadronic sector , manifested by the large decrease of the effective nucleon mass @xmath44 . by using the gibbs criteria of phase equilibrium @xmath45 , @xmath46 and @xmath47 we have obtained the critical temperatures for @xmath48 ( 188mev ) , @xmath49 ( 191mev ) , @xmath50 ( 192mev ) , @xmath17 ( 186mev ) , @xmath19 ( 192mev ) , @xmath21 ( 192mev ) and @xmath22 ( 191mev ) models .
the other models studied in this paper do not present first order hadronic phase transition .
the critical temperatures for this hadron nucleon - antinucleon phase transition for hot nuclear matter at the hadronic sector itself are remarkably close to the critical temperature of the hadron - qgp phase transition .
this hadron nucleon - antinucleon phase transition where the effective nucleon mass becomes very small can be seen as the quantum hadrodynamics version of the chiral phase transition .
it is interesting that in recent lattice calculations the chiral and the deconfinement transitions coincide and in our calculation both transitons seem also to appear at the same temperature depending on the value of the bag constant @xmath4 we choose .
we present in figure 2 the pressure for hadronic - qgp phases ( in units of the stefan - boltzmann pressure ) as a function of @xmath43 . in this figure
we have used the extreme values @xmath37 mev and @xmath38 mev .
it is to remark how the qgp critical temperatures are close for all the hadronic models . from this figure
we see the critical temperatures very close to 151 mev for @xmath37 mev and around 198 mev for @xmath38 mev .
the reason lies in the slope for the qgp pressure .
this slope is nothing but the entropy density , that we show in figure 3 .
notice how far , by a factor around five ( for @xmath51 ) and around three ( for @xmath52 ) , the entropy amounts are different in both phases .
it causes the hadronic model independence for the hadronic - qgp critical temperature . when the hadronic model itself presents the nucleon - antinucleon phase transition it increases substantially the entropy density as can be seen in figure 3 .
however , this enhancement is not enough when compared with the qgp entropy .
this shows a discontinuity in the entropy signalizing a first order hadron - qgp phase transition , once we are committed with only nucleon - antinucleon in the hadronic phase .
let us also note that the qgp entropy density is independent of the bag constant @xmath53 since it is given by ( @xmath54)/t . note how , in figure 2 for different bag constants , the two slopes are the same for @xmath55 .
however , as pointed out above , the bag constant has effect on the hadronic - qgp value of @xmath56 .
therefore , if one thinks that the enhancement of the hadronic entropy density would help the qgp formation , only values of @xmath57 greater than 200mev should be considered .
we have then included , in an _ ad hoc _ fashion , a thermal pion gas contribution to the hadronic phase . by analyzing the behaviour of @xmath58 as a function of @xmath43 with and without pions
we conclude that the change in the critical phase transition temperature is very small .
when we have only nucleons and pions in the hadronic phase , the hadronic - qgp phase transition at very low net baryon density takes place for approximately the same value of @xmath56 in all the studied hadronic models .
the hadronic - qgp phase transition is essentially given by the qgp phase and lies very close to the temperature where @xmath55 crosses zero from below .
the chiral transition version we found in the hadronic models can be identified by a strong decreasing in the effective baryon mass and an abrupt increasing of the entropy density in the hadronic phase ( see fig . 1 and fig .
this large hadronic jump of the entropy density favours the hadron - qgp phase transition because it requires a smaller latent heat at the transition compared with the models that do not show that pure hadronic nucleon - antinucleon plasma phase .
thus , from our results , we would expect at very low net baryon density a formation of a very rich baryon - antibaryon matter just before the qgp phase .
it is important to stress that , even with the formation of this rich baryon plasma , the entropy enhancement is not enough yet to be compared with the qgp entropy density and still requires a large latent heat at the phase transition .
the latent heat and entropy at the transition , as it has been discussed before on the @xmath7 central rapidity region @xcite , may be seen by a large change in the specific volume that slows down the time scale for the conversion of qgp into the hadron matter : a large entropy discontinuity implies a long lifetime of the plasma .
this point itself , the entropy discontinuity , has deserved intense work and analysis from many authors @xcite .
in fact it addresses direct the order of the hadronic - qgp phase transition .
we are aware of the limitations of our analysis .
the results are dependent on the qgp equation of state we are using , obtained by a perturbative expansion in which we treat the quarks as massless .
note in figure 3 how the entropy density decreases with the increase of the running coupling constant @xmath5 that at the phase transition is around @xmath59 to @xmath60 .
this suggests that a non perturbative treatment for the qgp equation of state would reduce the entropy density .
indeed , recent lattice qcd results for the pressure and entropy density @xcite have also obtained a strong entropy enhancement close to the phase transition exactly at the order of our critical hadronic entropy density ( figure 3 ) .
these lattice results were also reproduced by a non perturbative purely gluonic qcd calculation @xcite .
therefore , we should say that the order of the hadron - qgp phase transition is still an open point in our work .
using many different rmf models for the hadronic phase and a perturbative qcd equation of state for the qgp , we study the hadron - qgp phase transition at zero net baryon density and high temperature . we show that the critical temperature @xmath61 for this transition is hadronic model independent .
we have traced back the reason for this and conclude that it comes from that the qgp entropy ( the slope of the pressure versus t ) is much larger than the hadronic entropy obtained in all the rmf models .
this finding is quite independent whether the hadronic models have or not a hadronic phase transition at high temperature when the system becomes a dilute gas of baryons in a sea of baryon - antibaryon pairs .
the hadronic - qgp phase transition at zero net baryonic density takes place for approximately the same value of @xmath56 in all the studied hadronic models . among the studied hadronic models most of them showed a pure hadronic phase transition , the quantum hadrodynamics version of the chiral phase transition , seen by a strong decreasing in the effective baryon masses and an abrupt increasing of the entropy density in the hadronic phase
this fact by itself does not change the hadronic - qgp value of @xmath56 .
however , it is important to remark that the latent heat given by @xmath62 connecting both phases is much lower ( but still not small , @xmath63 with a non zero @xmath5 , see figure 3 ) for the hadronic models which present this hadronic phase transition to a baryon - anti - baryon plasma , favouring the hadron - qgp transition .
this result enlarges the knowledge of this transition regarding rmf models and may reinforce different model calculations conjecturing the enhancement of the entropy as a signature for a hadronic - qgp phase transition at low net baryonic density .
this signature it is quite important nowadays with the underway experiments , such as the collaboration star at rhic , because recent theoretical studies showed that for central collisions the rising of the incident energy from ags to rhic decreases the value of the chemical potential in the hadron - qgp phase diagram .
thus , the formation of qgp at rhic energies may be expected to occur at very small values of the chemical potential , exactly the regime we are analyzing in this paper , were we expect a jump on the hadronic entropy density due to the formation of a rich baryon - antibaryon matter .
finally , we stress the importance of the direct measure through calorimeters of this quantity .
99 j. letessier and j. rafelski , int .
j. of mod .
phys.*e9 * , 107 ( 2000 ) .
d. zschiesche , s. bass , m. bleicher , j. brachmann , l. gerland , k. paech , s. scherer , s. soff , c. spieles , h. weber , h. scker and w. greiner , nucl - th/0101047 ( 2001 ) proc .
quark matter99 , nucl .
a661 , 1c ( 1999 ) r. anishetty , p.koehler , and l. mclerran phys .
d22 , 2793 ( 1980 ) e.v .
shuryak , phys . lett .
* b78 * , 150 ( 1978 ) k. kajantic and h. i. miettinen , z. phys . *
c9 * , 341 ( 1981 ) j. rafelski and b. mller , phys .
lett * 48 * , 1066 ( 1982 ) . ; j. rafelsky , phys . reports 88 , 331 ( 1982 ) t. matsui and h. satz , phys .
* b178 * , 416 ( 1986 ) d. kharzeev and h. satz , phys . lett . *
b334 * , 155 ( 1994 ) http://www.cern.ch/cern/newstate matter/ with the text of scientific consensus view of the spokesmen of cern experiments and u. heinz and m. jacob , nucl - th/0002042 l. v. bravina , e. e. zabrodin , m. bleicher , s. a. bass , m. brandstetter , a. faessler , c. fuchst , w. greiner , m. i. gorenstein , s. soff and h. stcker , j. phys . * g27 * , 421 ( 2001 ) l. v. bravina , e. e. zabrodin , s. a. bass , m. bleicher , m. brandstetter , s. soff , h. stcker and w. greiner , phys . rev . *
c62 * , 064906 ( 2000 ) u. heinz , p.r .
subramanian , h. stcker and w. greiner , j. phys . *
g12 * , 1237 ( 1986 ) .
j.theis , g.f.graebner , g.bucchwald , j.a.maruhn , w.greiner , h.stcker and j.polony , phys . rev . *
d28 * , 2286 ( 1983 ) .
furnstahl and b.d.serot , phys.rev.*c41 * , 262 ( 1990 ) .
m. malheiro , a. delfino and c. t. coelho , phys .
rev.*c58 * , 426 ( 1998 ) ; j. s. sa martins and a. delfino , phys .
rev.*c61 * , 044615 ( 2000 ) . m. reiter , a. dumitru , j. brachmann , j .
a. maruhn , h. stcker and w. greiner , nucl . phys . *
a643 * , 99 ( 1998 ) .
j. d. walecka , ann .
phys . * 83 * 491 ( 1974 ) ; b. d. serot and j. d. walecka , advances in nuclear physics vol .
16 , pp . 1 - 327 ( plenum , new york , 1986 ) r.j.furnstahl , b.d . serot and hua - bin tang , nucl.phys.*a615 * ( 1997 ) 441 . j. zimanyi and s. a. moszkowski , phys
. rev . * c42 * ( 1990 ) 1416 .
a. delfino , c.t .
coelho and m. malheiro , phys .
rev.*c51 * , 2188 ( 1995 ) ; a. delfino , c.t . coelho and m. malheiro ,
phys.lett.*b345 * , 361 ( 1995 ) .
reinhard , rep .
. phys . * 52 * , 439 ( 1989 ) ; g. a. lalazissis , j. koening and p. ring , phys . rev . * c55 * , 540 ( 1997 ) ; m. m. sharma , m. a. nagarajan and p. ring , phys . lett .
* b312 * , 377 ( 1993 ) .
bo - qiang ma , qi - ren zhang , d.h .
rischke and w. greiner , phys .
lett . * b315*,29 ( 1993 ) .
h. stcker and w. greiner , phys .
reports * 137 * ( 1986 ) 277 and references there in .
collins , and m.j .
perry , phys .
lett . * 34 * ( 1975 ) 1353 , ; g. baym and s.a .
chin , phys . lett.*62b * 1976 241 , ; b.a . freedman and l. mclerran ,
rev.*d16 * , ( 1977 ) 1169 ; j. kuti , b. lukacs , j. polonyi and k. szlachanvi , phys . lett.*95b * ( 1980 ) 75 d.h .
rischke , m.i .
gorenstein , h. stcker and w. greiner , z. phys.*c51 * ( 1991 ) 485 ; o.d .
chernavskaya , z. phys.*c70 * ( 1996 ) 517 ; bi pin - zhen and shi yao - ming , z. phys.*c75 * ( 1997 ) 735 .
j. letessier , a. tounsi , u. heinz , j. sollfrank and j. rafelski , phys .
lett * 70 * , 3530 ( 1993 ) .
j. cleyman , k. redlich , h. satz and e. suhonen , z. phys . *
c58 * , 347 ( 1993 ) ; a. leonidov , k. redlich , h. satz , e. suhonen and g. weber , phys . rev . * d50 * , 4657 ( 1994 ) . c. adler et al .
lett . * 86 * , 4778 ( 2001 ) e. v. shuryak , _ the qcd vacuum , hadrons and the superdense matter _
, world scientific lecture notes in physics , vol .
a. delfino , m. chiapparini , m. e. bracco , l. castro and s. e. epsztein , j. phys . *
g27 * , 2251 ( 2001 ) .
j. boguta and a. r. bodmer , nucl . phys .
* a292 * 414 , ( 1977 ) .
t. matsui , b. svetitsky , l.d .
mclerran , phys .
* d34 * , 2047 ( 1986 ) .
mclerran and t. matsui , proc .
workshop on experiments for rhic , bnl , ny , april 15 - 19 ( 1985 ) b. l. friman , k. kajantie , p.v .
ruuskanen , nucl . phys . *
b266 * , 468 ( 1986 ) .
f. karsch , e. laermann and a. peikert , phys .
lett * 478 * , 447 ( 2000 ) .
f. karsch , nucl .
. suppl . * 83 * , 14 ( 2000 ) .
j. p. blaizot , phys .
lett * 83 * , 2906 ( 1999 ) . | recent studies show that for central collisions the rising of the incident energy from ags to rhic decreases the value of the chemical potential in the hadron - qgp phase diagram .
thus , the formation of qgp at rhic energies in central collisions may be expected to occur at very small values of the chemical potential . using many different relativistic mean - field hadronic models ( rmf ) at this regime
we show that the critical temperature for the hadron - qgp transition is hadronic model independent .
we have traced back the reason for this and conclude that it comes from the fact that the qgp entropy is much larger than the hadronic entropy obtained in all the rmf models .
we also find that almost all of these models present a strong entropy enhancement in the hadronic sector coming from the baryonic phase transition to a nucleon - antinucleon plasma .
this result is in agreement with the recent data obtained in the star collaboration at rhic where it was found a rich proton - antiproton matter .
17.cm 21.50 cm 1.0 cm 0.0 cm .3 cm |
ngc 1275 ( henceforth perseus a , abbreviated per a ) is the central dominant cd ( giant elliptical ) galaxy in the perseus cluster .
it is one of the most intriguing and intensively scrutinized galaxies in the sky .
@xcite found two sets of optical emission lines separated by @xmath0@xmath7 towards per a , and proposed that per a is colliding with a forergound infalling galaxy .
a photograph in the h@xmath1 line taken by @xcite revealed a spectacular flamentary nebula centered on per a and extending up to @xmath0140 ( @xmath050 kpc ) from its center .
a modern ccd image reveals that the h@xmath1 nebula has a bright inner region extending east - west , together with a multitude of mostly radial ( but also some transverse ) filaments that extends furthest in the north - south direction reaching @xmath070 kpc north of center @xcite
. per a was the first central cd galaxy in a rich cluster found to contain cool ( @xmath8100 k ) molecular gas as traced in co @xcite .
single - dish maps show that the main body of the cool molecular gas has an overall morphology similar to the bright inner region of the h@xmath1 nebula @xcite .
cool molecular gas also has been detected towards the outer h@xmath1 filaments where observed @xcite .
both hot ( @xmath02000 k ) ( * ? ? ?
* and references therein ) and warm ( @xmath0300400 k ) @xcite molecular hydrogen gas has been found towards all regions of the h@xmath1 nebula so far observed ( bright inner region and outer filaments ) .
the coolest detectable component of the x - ray emitting gas in the perseus cluster , at a temperature of @xmath0@xmath9 k , has a morphology similar to the h@xmath1 nebula @xcite .
the nature ( origin and excitation ) of the multi - phase nebula associated with per a is poorly understood . because the cool molecular gas traced in co dominates by far the mass of this nebula , a detailed understanding of its properties is of particular importance .
the main body of this cool molecular gas has been traced by single - dish telescopes out to @xmath014 kpc east and west of center , and has a mass of @xmath0@xmath10 ( * ? ? ? * and references therein ) .
these single - dish observations at spatial resolutions of several kiloparsecs show no overall pattern in the gas kinematics , which is slightly redshifted at the nucleus and ( mostly ) more strongly blueshifted east and west of center .
our ( * ? ? ?
* hereafter paper i ) interferometric observations at a spatial resolution of @xmath01 kpc covering the central region of radius @xmath010 kpc revealed three radial filaments lying approximately east - west , together with a number of other mostly unresolved features . as a whole
, these features show no overall pattern in their kinematics ; in particular , there is no bulk velocity gradient in the east - west ( or any other ) direction indicative of gas orbiting per a , as would be expected for material captured from a gas rich - galaxy .
instead , we found that each filament exhibits its own particular kinematic pattern .
the outermost filaments east and west of center both exhibit linearly increasing blueshifted velocities with decreasing radii , but with different velocity gradients .
we were able to reproduce their observed spatial - kinematic structures as free - fall in the gravitational potential of per a , as would be naturally expected for cool gas deposited by a x - ray cooling flow . here ,
we report follow - up interferometric co observations that extend our coverage in the east - west direction to span the entire main body of the cool molecular gas mapped with single - dish telescopes .
we describe our observations and data reduction in @xmath11 .
the results are presented in @xmath12 , where we show that all but one of the detected co features is spatially resolved into radial filaments ; at least six distinct filaments can be clearly identified in our maps . in @xmath13
, we examine the relationship between the cool molecular filaments and the h@xmath1 nebula , and show that there is a correlation between the co and h@xmath1 luminosity of all these filaments ; on the other hand , we show that a number of at least equally luminous h@xmath1 features exhibit no detectable co with upper limits that deviate strongly from the derived correlation .
we then demonstrate that all the filaments are gravitationally bound , and explore why they do not appear to be collapsing to vigorously form stars . finally , as a guide for future interferometric observations , we identify regions where our interferometric measurements have not recovered all the single - dish emission . in @xmath14 , we provide a concise summary of our observations and interpretation .
we assume as in paper i that per a lies at a distance of @xmath074 mpc ( redshift @xmath15 , and adopting @xmath16 and @xmath17 ) , so that @xmath18 .
we observed per a on 2006 december 17 using the sub - millimeter array ( sma ) as in our previous observation ( paper i ) .
the weather was good , with a zenith sky opacity at 225 ghz ( @xmath19 ) in the range @xmath00.050.1 as measured at the adjacent caltech submillimeter observatory .
the sma was in its compact configuration , and with six of its eight 6-m antennas operating provided fifteen baselines ranging from about 10 m to 70 m. the dual - sideband receivers operating in the 230 ghz ( 1.3 mm ) band were tuned to a central observing frequency of 216.35 ghz in the lower sideband , and 226.34 ghz in the upper sideband to place the co(21 ) line in this sideband . at the frequency of the co(21 ) line , the primary beam of the telescope as measured at full width half maximum ( fwhm ) is @xmath055 , which corresponds to @xmath020 kpc in the assumed cosmology .
the correlator was configured to cover 24 spectral windows ( chunks in the sma nomenclature ) with 128 channels in each chunk .
this provided a spectral resolution of 0.81 mhz ( @xmath0@xmath20 ) over a total bandwidth , with some overlap at the edges of adjacent chucks , of 1.98 ghz ( @xmath0@xmath21 ) .
the system temperature ranged from 80 k to 170 k , changing in time with the source elevation .
we began by observing neptune and uranus for 20 minutes each to serve as alternate choices for absolute flux and antenna passband calibration .
we then observed per a over a duration of @xmath09 hrs , interrupted only once for an antenna pointing check .
our previous observation ( paper i ) had shown that the co(2 - 1 ) features detected spanned from center to ( near ) the edge of the primary beam in the east - west directions . to provide better sensitivity to the outer eastern and western features , and to search for more gas beyond the field of view of our previous observation , we employed a three - point mosaic with one field centered on the nucleus and the other two centered approximately on the eastern and western filaments respectively as illustrated in figure [ 3pointing ] .
we pointed the telescope alternately between these three positions , and in each cycle integrated for six 30-s samples on the central position and ten 30-s samples on each of the eastern and western positions .
more details of the observations can be found in table [ pointing ] . finally , we observed saturn for 55 min to serve as another alternative for passband calibration .
saturn was so bright , however , that it partially saturated the receivers , thus preventing its use as a passband calibrator .
we performed our data reduction using sma - specific mir tasks in idl adopted for the sma from the mma software package developed originally for ovro @xcite .
because absorption by the earth s atmosphere introduces an elevation - dependent attenuation of astronomical signals , we first corrected for this effect by weighting the amplitude of each visibility by the inverse product of the system temperatures of the corresponding antenna pair as measured in each 30-s integration . in our observation , both uranus and neptune were resolved on the longer baselines ; because the measured amplitude of uranus was higher than that of neptune on all baselines , we used uranus for both absolute flux and passband calibration .
we derived complex gain ( i.e. , amplitude and phase ) solutions from the unresolved nuclear continuum emission of per a when the telescope was pointed at the central position , and then interpolated these solutions to measurements when the telescope was pointed at the eastern and western positions . because of significant baseline - based errors in some chunks , we adopted baseline - based solutions for both passband and gain calibration .
this also helped address significant pointing errors on some of the antennas , an issue we will return to later .
mapping of the co(21 ) emission towards per a requires stable antenna passbands to avoid artifacts that would otherwise be produced by its strong nuclear continuum source ( see discussion in paper i ) .
this source had a measured flux density of 3.5 jy during our observation , compared with 3.4 jy measured in our previous observation ( paper i ) . in channel maps of the central pointing made before continuum subtraction
, we found that the nuclear continuum source exhibits a slow and weak variation in intensity with frequency , reflecting small changes in the individual antenna passbands with time .
the peak - to - peak variation in the nuclear continuum intensity with frequency is @xmath0@xmath22 .
this led to an imperfect subtraction of the nuclear continuum source in our co(21 ) channel maps , leaving artifacts as strong as about @xmath2350 mjy at the nuclear position for the central pointing . for comparison , the root - mean - square ( rms ) noise level ( @xmath24 ) of the channel maps at a velocity resolution of @xmath0@xmath25
is @xmath030 mjy for the central pointing ( table [ pointing ] ) .
hence , artifacts at the nuclear position are no stronger than about @xmath232@xmath24 in these maps . in channel maps of the eastern and western pointings where the nuclear continuum source lies close to the edge of the primary beam and
is therefore strongly attenuated , artifacts at the nuclear position caused by passband variations is correspondingly less severe . to make the final clean maps , the dirty beam , corresponding to the point spread function ( psf ) of the telescope as computed from the known positions of the antennas relative to the source , must be deconvolved from the dirty maps ( where each pixel is convolved with the telescope psf ) to remove the sidelobe response of the telescope .
variations in antenna passbands , along with imperfect complex gain calibration , can leave artifacts related to imperfect deconvolution of sidelobes from both continuum and line emission . although difficult to quantify , artifacts produced by sidelobes can often be recognized by their resemblance to the known sidelobe pattern .
in addition , maps made in our previous observation ( paper 1 ) could be compared with those made here as an independent check . in this way
, we were able to identify artifacts lying north - south in our previous map , and confirm as genuine all the features lying east - west .
for the eastern and western pointings , the nuclear continuum source of per a lies close to the edge of the primary beam .
any pointing errors in the individual antennas ( along the radial direction ) therefore causes a much larger variation in the measured intensity of the nuclear continuum source for these pointings compared with the central pointing .
indeed , we found that the measured intensity of the nuclear continuum source changed with time by up to a factor of @xmath01.3 larger in both the eastern and western pointings than in the central pointing , indicating significant antenna pointing errors . by comparing the amplitudes of the nuclear continuum source measured in each of the three pointings against their expected amplitudes
, we estimated pointing errors of about 515 depending on the given antenna .
such antenna pointing errors lead to an inaccurate correction for the primary beam response of the antennas that becomes increasingly worse for features closer to the edge of the primary beam . to estimate the magnitude of these errors ,
we compared the measured intensity of the nuclear continuum source from maps ( corrected for the primary beam response ) made of adjacent pointings ; we also compared , after continuum subtraction , the measured integrated intensity of co(21 ) features in regions of the map ( corrected for the primary beam response ) where adjacent pointings overlap .
we found that the intensity of both the continuum source and common co(21 ) features to differ by less than @xmath0@xmath26 between adjacent pointings , indicating that the antenna pointing errors are more or less random in the sky . in the co(2 - 1 ) channel maps at a velocity resolution of @xmath0@xmath25 , 10@xmath27 variations in the intensity of individual features correspond to less than the rms noise fluctuations .
antenna pointing errors , just like errors in passband and complex gain calibration , also can leave artifacts related to imperfect deconvolution of sidelobes from both continuum and line emission . in this respect ,
antenna pointing errors are probably more important than those related to variations in antenna passbands and imperfect complex gain calibration for , especially , the eastern and western pointings . to make co(21 ) channel maps , we first subtracted the nuclear continuum source in the visibility plane derived from a fit to the line - free channels in the upper sideband . for the reader to judge the reliability of our maps ,
we have made separate channel maps for the three individual pointings at a velocity resolution of @xmath25 , and derived from these channel maps separate integrated intensity ( zeroth moment ) and intensity - weighted mean - velocity ( first moment ) maps corrected for the primary beam response .
we also made a mosaic map utilizing all the data from the three separate pointings , along with data taken in the previous observation ( paper i ) , with the appropriate correction for the primary beam response in each pointing .
the corresponding angular resolutions and rms noise of the mosaic and combined maps are listed in table [ pointing ] .
all the maps shown in this manuscript , unless otherwise mentioned , were made with natural weighting of the data to maximize the signal - to - noise ( s / n ) ratio .
only one of the maps shown was made using uniform weighting , useful for inspecting the detected features at the highest attainable angular resolution .
we have discovered a small error in our subtraction of the nuclear continuum source in paper i. when deriving its flux density , we inadvertently included line - containing along with line - free channels in our fit : this resulted in a slight overestimate of the flux density of the nuclear continuum source , and therefore an oversubtraction of the continuum source in our channel maps .
this error can be seen most clearly in the spectra of the individual filaments shown in figure 3 of paper i , especially for the inner and western filaments ( i.e. , those closest to the nucleus ) where the spectral baseline is consistently below zero by a few tens of mjy . after correcting this mistake
, the moment map made , shown in figure [ 3pointing ] , does not show any qualitative changes .
as in paper i , to make moment maps we first convolved the channel maps with a triangular function ( i.e. , hanning smoothing ) that best recovers the overall features in that particular map .
we then discarded individual pixels in the original channel maps with intensities less than a certain theshold ( typically @xmath01.52.0@xmath24 ) in their corresponding hanning - smoothed channel maps .
we summed the remaining data in the channel maps ( with no smoothing ) to produce the integrated intensity map , and computed the mean velocity of each pixel weighted by its intensity in each channel to produce the intensity - weighted mean - velocity map .
figure [ new_mom ] shows moment maps of the central , eastern , and western pointings separately . in each case , the integrated co(21 ) intensity , corrected for the primary beam response , is plotted in contours on the intensity - weighted co(21 ) mean velocity in color .
all of the features seen in our previous observation ( comprising a single pointing centered on the nucleus ) except for n1 , s1 , and w1 ( fig . 2 in paper i ; see also fig . 1 of this paper )
are seen in at least one ( typically two , if not all three ) of the individual pointings in our present observation .
s1 , if real , should have been detectable in the moment maps of both the central and western pointings . on the other hand , n1 and w1 ,
if real , would lie below the cutoff threshold imposed on the moment maps for both the central and western pointings .
when we lowered the cutoff threshold for making the moment maps shown in figure [ new_mom ] , we found that w1 but not n1 nor s1 could be seen in both the central ( along with e1 ) and western pointings .
thus , w1 is real , but as shown explicitly in @xmath28 both n1 and s1 are artifacts related to sidelobes . the inner filament is detected in all three pointings . the western filament is detected in both the central and , at a higher s / n ratio , western pointings . the eastern filament is detected in the eastern pointing , and only its outermost tip in the central pointing .
e1 is detected only in the eastern pointing , but e2 is detected in both the central and eastern pointings .
apart from these features , all of which were previously detected in paper i , we do not detect any new features in the maps of figure [ new_mom ] that we can confidently regard as real .
we have combined all the data taken in our present observation with that taken in paper i ( comprising only a single pointing centered on the nucleus ) to make the most sensitive mosaic map possible .
figure [ channel - maps ] shows the resulting co(21 ) channel maps for the inner@xmath29western regions ( upper panel ) and the eastern region ( lower panel ) separately .
figure [ mom - all ] shows the corresponding moment map , where the dotted line indicates where the noise level has increased by a factor of 2 from center . although extending the region mapped from a central radius of @xmath010 kpc to a radial extent of @xmath014 kpc in both the eastern and western directions to cover the entire main body of molecular gas in per a , we did not detect any new features beyond the region previously mapped in paper i. in the moment map of figure [ mom - all ] , we have masked out the artifact labeled n1 in our previous observation ( see @xmath28 ) ; the artifact s1 , which we found to be substantially reduced in the maps made from all the data combined , is not detected at the cutoff threshold imposed .
( we note , for completeness , that noise peaks outside the area occupied by the dotted line have been masked . ) compared with our previous observation , the western filament appears to exhibit a more pronounced bend towards the south about halfway along its length . as shown in greater detail later in @xmath30 , this apparent bend is caused by a separate feature labeled w2 ( not discernible as a separate feature in paper i despite the improved continuum subtraction ; fig . [ 3pointing ] )
detected here for the first time ( at the improved sensitivity ) located just to the south of the western filament about halfway along its length . in figure [ uniform - weighting ] , we show moment maps of the features e1 , e2 , and w1 made with uniform weighting to provide a higher angular resolution . the arrows in each panel point to the center of per a. all these features , like the three previously identified filaments ,
can now be seen to comprise radial filaments ( as discussed in more detail below ) .
@xcite have observed the eastern filament , e1 , and e2 at significantly higher sensitivity and comparable ( marginally higher ) angular resolution with the iram pdbi .
their maps also show that e1 and e2 are radial filaments . in table
[ gas_mass ] , we tabulate the properties of each filament ( with limited information for w2 , which is spatially unresolved ) , including its angular distance from center ( measured at the midpoint ) , length , position angle , and velocity range . the velocity range spanned by each filament is defined as the extreme velocities at which the filaments are detectable at or above the @xmath31 level in the channel maps as cross - checked against the moment maps . as mentioned in paper i , the lengths of the inner , western , eastern , and now also e2 filaments are difficult to quantify precisely because they do not have smoothly varying intensity profiles ; here , we use the same method as in paper i , with their lengths defined as the difference in radial positions of their emission centroids either at or close to their opposite ends in velocity ( eastern and western filaments ) or where their angular separation is largest ( inner and e2 filaments ) .
this method likely underestimates slightly their true lengths .
their position angles correspond to that measured from the center of per a through their midpoints as estimated by eye in the moment maps .
we note that , compared to the values previously tabulated ( table 1 in paper i ) , the largest difference is for the inner filament which is now much longer and spans a larger velocity range .
this is because the error in continuum subtraction ( as described towards the end of @xmath32[observations and data reduction ] ) most strongly affects the inner filament , which now extends much further beyond the nucleus on the eastern side .
the western filament is detectable over a slightly larger velocity range because the improved continuum subtraction and increased sensitivity permits previously undetectable ( i.e. , below @xmath31 in the channel maps ) portion of the filament to now be detectable .
the position - velocity ( pv- ) diagrams of the western and eastern filaments have not changed qualitatively ( see @xmath32[position - velocity diagrams ] ) , and provide if anything a more confident fit ( given the improved sensitivity ) to the model presented in paper i whereby these filaments are in free - fall along the gravitational potential of perseus a. the revised dynamical age of the western filament is @xmath023 myr and eastern filament @xmath028 myr ; i.e. , of order @xmath0@xmath4 yr as estimated in paper i. the e2 and w1 filaments are best resolved in the uniformly - weighted moment maps of figure [ uniform - weighting ] .
we have therefore derived their lengths and position angles by fitting gaussian structures to these filaments in the uniformly - weighted maps but without imposing any cutoffs to preserve error statistics .
we find both to be aligned radially with respect to the center of per a within measurement uncertainties ( no more than @xmath0@xmath33 ) . for w2 ,
we tabulate only its radial distance from and its position angle relative to the center .
as mentioned in paper i , the co(21 ) features detected along the east - west direction ( all now confirmed to be genuine ) appear to be associated with h@xmath1 emission , but not those detected either to the north ( n1 , which is too weak to be detected in the present observation ) or south ( s1 , which is not confirmed in the present observation ) . to properly understand whether all the detected co(21 ) features have h@xmath1 counterparts , we have carefully examined whether n1 and s1 are artifacts related to imperfect removal of sidelobes . in figure [ artifact ] , we show both dirty ( before deconvolving the telescope psf ) and clean ( after deconvolving the telescope psf ) maps of just the channel where s1 appears most strongly in the previous dataset .
we also show in this figure the dirty beam ( the telescope psf ) . apart from n1 , portions of the eastern and western filaments are also present in this channel , with the western filament stronger than the eastern filament . by comparing the dirty map with the dirty beam
, one can see that s1 is located at the first sidelobe of the western filament ( other sidelobes also are visible in the dirty map ) . to test whether s1 is an artifact related to this sidelobe
, we also show in figure [ artifact ] dirty maps with the western filament subtracted .
as can be seen , s1 becomes much weaker , and is now no stronger than the remaining sidelobes ( from the eastern filament ) or noise peaks in the dirty map .
this behavior is repeated in all channels where s1 is detectable , resulting in a spatial - kinematic structure for s1 that closely resembles the western filament ( see fig . 2 of paper i ) .
this exercise clearly demonstrates that s1 is just a sidelobe related to the western filament .
a similar exercise reveals that n1 is a sidelobe of the inner and eastern filaments combined .
furthermore , as can be seen in figure [ 3pointing ] , both n1 and s1 lie outside the main body of molecular gas mapped with the iram 30-m telescope ; all the other confirmed genuine filaments coincide with ridges in the main body of molecular gas .
figure [ pv_diagram ] shows position - velocity ( pv- ) diagrams of all the filaments along radial directions . as can be seen , the improved continuum subtraction applied to the data in paper i combined with the mosaic data has not resulted in any qualitative changes to the spatial - kinematic structures of any of the filaments , except for the detection of the new feature w2 . the pv - diagram along a position angle of @xmath34 ( top right panel ) that cuts through the both the inner extent of the western filament and w2 shows that w2 is not just spatially but also kinematically distinct from the western filament .
we had hoped that the resulting improvement in sensitivity ( permitting also a slight improvement in angular resolution with uniform weighting ) over that attained in paper i would reveal velocity patterns in the shorter filaments e1 , e2 , and w1 .
sadly , no such patterns are discernible .
both here and in the iram pdbi map of @xcite at a higher sensitivity , e2 can be seen to comprise two peaks that , especially in the latter map , appear be discontinuous in velocity .
the stronger peak may show a velocity gradient , which if real increases in blueshifted velocities at decreasing radii just like the eastern and western filaments .
figure [ spectra ] shows spectra of the individual filaments at a velocity resolution of @xmath25 . in table
[ gas_mass ] , we list the integrated flux density of each filament as derived from these spectra .
note that the integrated flux density for the western filament contains a small contribution from w2 .
the quoted uncertainty in the flux density of the inner filament includes the uncertainty in determining the continuum subtraction level ( a minor component of the error budget ) .
none of the quoted uncertainties in flux densities , however , include any systematic uncertainty in absolute flux calibration that may be as large as @xmath020@xmath27 .
we converted the integrated co(21 ) flux density of each filament to its corresponding mass in molecular hydrogen ( h@xmath35 ) gas using the same assumptions as in paper i , thus facilitating a direct comparison with the single - dish map of @xcite ( as discussed later in
[ unrecovered emission ] ) .
because of a trivial mistake in converting from flux density to brightness temperatures , all the values of molecular hydrogen gas masses quoted in paper i are overestimated by a factor of @xmath36 ( but those numbers also need to be revised because of the imperfect continuum subtraction ) . in table
[ gas_mass ] , we list the revised values for the molecular hydrogen gas masses of each filament .
the conversion factor used for relating co(21 ) flux densities to molecular hydrogen gas masses is that believed to be appropriate for galactic molecular clouds @xcite ; this conversion factor is estimated to be @xmath05 times too high for the centrally concentrated molecular gas ( disks ) in ultraluminous infrared galaxies @xcite .
the appropriate conversion factor for the molecular gas in per a is not known .
the filaments have increasingly larger masses with decreasing angular distance from the center along a direction that passes through the eastern filaments ( eastern filament and e1 ) , as well as through the inner filament and the western filaments ( western filament and w1 ) .
the mass of the inner filament alone is comparable with that of the eastern and western filaments combined , and with the mass of molecular gas in normal spiral galaxies .
these three filaments comprise @xmath0@xmath37 of the total mass of all the detected filaments .
altogether the filaments have a total mass of @xmath38 , about one - half that measured by single - dish telescopes over the same region mapped in our mosaic observation of @xmath0@xmath39 @xcite . in @xmath40 , we examine how much of the co(21 ) emission detected with single - dish telescopes towards the filaments observed has been recovered in our interferometric observations .
in paper i , we found that the three spatially - resolved ( inner , western , and eastern ) filaments detected are radially aligned .
we showed that the velocity pattern of the western and eastern filaments can be simply explained as free - fall in the gravitational potential of perseus a. the improved pv - diagrams for the latter two filaments presented in this paper do not change this central conclusion . in this paper
, we find that e1 , e2 , and w1 like the three abovementioned filaments also comprise radially - aligned filaments . since our observations were made , @xcite also have reported that e1 and e2 comprise radially - aligned filaments .
thus , in all , six distinct cool molecular filaments , three lying east of center and three lying west of center , have now been identified .
their radial morphologies suggest that there is no significant transverse velocity component tracing orbital motion , arguing against the idea that some of the cool molecular gas detected may have been captured from a gas - rich galaxy .
the lack of any transverse spatial component provides no support for ( but does not necessarily negate ) the idea that the cool molecular gas may have been dragged outwards by buoyant x - ray bubbles , as has been proposed to explain transverse h@xmath1 filaments seen in the outer regions of the nebula @xcite .
as mentioned in paper i , the observed co(21 ) filaments ( apart from n1 and s1 , which as shown in @xmath28 turn out to be artifacts ) are spatially coincident with bright h@xmath1 features . in figure
[ co - haf ] , we plot contours of the integrated co(21 ) intensity on a color image of the h@xmath1@xmath29n[ii ] emission from @xcite .
we have adjusted the contrast for the inner@xmath29western and eastern regions separately to better see the spatial relationship between the molecular and ionized gas in these regions .
the inner filament is embedded in bright h@xmath1 emission that surrounds the nucleus . in both the eastern and western regions , every molecular filament , as well as the new feature w2 , is associated with locally bright h@xmath1 features . on the other hand , not all equally bright if not brighter h@xmath1 features , such as those labeled a d in the middle panel and hii region in the bottom panel of figure [ co - haf ] , have co(21 ) counterparts . as discussed in paper i and also @xcite , all the cool molecular filaments detected have velocities comparable if not identical with their h@xmath1 counterparts ( where measured ) .
both @xcite and @xcite found a linear correlation ( albeit with considerable scatter ) between the globally integrated co and h@xmath1 luminosities for the central cd galaxies in putative x - ray cooling - flow clusters .
@xcite find that this relationship extends to individual regions mapped in per a at an angular resolution of @xmath012 with the iram-30 m. in figure [ mass - haf - corr ] , we plot the integrated co(21 ) luminosities of the individual filaments in per a against their corresponding h@xmath1 luminosities measured in an area encompassing the lowest co(21 ) contour of a given filament as seen in figure [ co - haf ] .
the h@xmath1 emission associated with the inner filament contains a substantial contribution from the agn , and so the value plotted constitutes an extreme upper limit on the emission actually associated with this filament ( as indicated by the arrow ) .
the diagonal lines drawn in figure [ mass - haf - corr ] correspond to different ratios between the co(21 ) and h@xmath1 luminosities .
as can be seen , there appears to be an approximately linear relationship between the co(21 ) and h@xmath1 luminosities : a least - squares fit ( omitting the inner filament ) gives a ratio between the co(21 ) and h@xmath1 luminosity of @xmath00.0007 ( indicated by the solid line in fig .
[ mass - haf - corr ] ) with a correlation coefficient of 0.93 .
thus , the co(21 ) line has a luminosity of about @xmath41 that of the h@xmath1+n[ii ] line .
@xcite also have found a linear correlation between the line luminosities of warm ( 300 - 400 k ) molecular hydrogen and h@xmath1+n[ii ] , with the rotational line of h@xmath35 00s(1 ) having a luminosity of about @xmath22 that of the h@xmath1+n[ii ] line . the existence of a correlation between the co(21 ) and h@xmath1 luminosities of the filaments may suggest that the excitation of their cool molecular gas and ionized gas is causally related .
alternatively , this correlation may support the picture in which the optical emission - line nebula comprises the ionized skins of individual cool giant molecular hydrogen clouds ( e.g. , * ? ? ?
* ) : more giant molecular clouds result in stronger co emission , and a larger surface area exposed to ionization radiation and hence stronger h@xmath1 emission . on the other hand ,
not all h@xmath1 features that are as bright if not brighter than those associated with the co(21 ) filaments have detectable co(21 ) .
the most obvious example is in the eastern region , where the brightest h@xmath1 feature visible in figure [ co - haf ] ( bottom panel ) does not have a co(21 ) counterpart .
this h@xmath1 feature has the spectrum of a hii region and also a very blue continuum characteristic of a young and massive star cluster @xcite .
by contrast , other parts of the optical emission - line nebula so far studied spectroscopically including those along the inner , western , and eastern filaments , as well as e1 and e2 have been found to have both low ionization and excitation @xcite . in the inner@xmath29western region of figure [ co - haf ] ( middle panel ) , examples of comparably bright h@xmath1 features with no associated co(21 ) emission
are labeled a d .
these particular h@xmath1 features have not , to the best of our knowledge , been studied spectroscopically in the optical .
we have computed 3@xmath24 upper limits for the molecular hydrogen gas masses in all these regions assuming an overall linewidth of @xmath42 , as found for the shorter filaments e1 , e2 , and w1 . for comparison
, the hii region found by @xcite has an upper limit for its linewidth of @xmath0@xmath25 at fwhm @xcite . as shown in figure [ mass - haf - corr ] , the features a d have upper limits in their luminosity ratio of co(21 ) to h@xmath1 that is up to a factor of @xmath05 lower than that derived above for the co(21 ) filaments .
the hii region has a corresponding upper limit in this ratio that , based on the more sensitive measurements of @xcite , is more than an order of magnitude lower .
optical spectroscopy of the features a
d is needed to determine whether they also are giant hii regions , or whether there is a relatively large scatter in the co(21 ) to h@xmath1 luminosities even between different features in the low ionization and excitation optical emission - line nebula .
although there is abundant evidence for recent star formation in per a ( * ? ?
* and references therein ) , there is no evidence for star formation in the molecular filaments that we detect .
as mentioned in @xmath43 , optical spectroscopy of their corresponding h@xmath1 filaments do not show hii - region - like spectra , but instead gas with low ionization and excitation .
@xcite have shown that nearby normal spiral galaxies exhibit a nearly constant star formation efficiency ( star formation rate per unit of molecular gas ) of @xmath0@xmath44 .
if the molecular filaments detected in per a are forming stars at the same star formation efficiency , their resulting hii regions should have a h@xmath1 luminosity comparable with that observed for the h@xmath1 filaments cospatial with these molecular filaments . instead
, any hii regions in these filaments can contribute at most a small fraction of the observed h@xmath1 emission ( e.g. , see fig .
3 of * ? ? ?
* ) , and therefore the star formation efficiency in these molecular filament must be lower than that found in nearby spiral galaxies .
we show below that , in the absence of any external forces or internal means of support beyond thermal pressure , all the molecular filaments detected are gravitationally bound and should collapse on a timescale at most comparable if not much shorter than the inferred dynamical ages of the eastern and western filaments .
we then explore what may be preventing these filaments from collapsing to form stars ( with an efficiency comparable with that seen in normal spiral galaxies ) . for simplicity
, we first model each filament as a string of uniform spherical clouds . the jeans mass
the minimum mass required for a uniform spherical cloud to be gravitationally bound if supported only by thermal pressure can be expressed as @xmath45 ( assuming that all the mass is in the form of molecular hydrogen gas ) , where @xmath46 is the gas temperature and @xmath47 the radius of the cloud .
studies of the global co(10 ) , co(21 ) , and co(32 ) line ratios in the main body of molecular gas @xcite , as well as co(10 ) and co(21 ) line ratios measured at an angular resolution of @xmath012 with the iram 30-m in the main body of molecular gas as well as towards the outer h@xmath1 filaments @xcite , suggest that the co(21 ) emission is optically thick .
modeling of the global co line ratios suggest multi - temperature components at @xmath010 k and @xmath0170 k @xcite . for comparison , the peak brightness temperature measured for any of the filaments is only @xmath00.6 k , suggesting that either the filaments have much narrower widths than observed or comprise complexes of molecular clouds with an overall small surface filling factor .
let us start by considering a spherical cloud with @xmath48 , which is the upper limit placed on the half - widths of the filaments .
each of the observed filaments could then comprise ( at least geometrically ) up to about eight such clouds .
if at the excitation temperature of the co(21 ) transition of @xmath49 k , then to be gravitationally bound they are required to have masses of at least @xmath50 ; if at @xmath51 , then @xmath52 .
such jean masses are many orders of magnitude smaller than the masses of the individual filaments , implying that any such clouds would certainly be gravitationally bound .
if instead the filaments comprise complexes of ( uniform spherical ) molecular clouds with sizes comparable to the range seen for giant molecular clouds ( gmcs ) in our galaxy of @xmath5350 pc , @xmath54@xmath55 if at @xmath49 k or @xmath56@xmath57 if at @xmath51 ( compared with observed masses for galactic gmcs of about @xmath58@xmath59 ) .
consider first those clouds with @xmath49 k : to give a measured brightness temperature of @xmath00.3 k as is typically observed for the filaments , such clouds are required to have surface filling factors of @xmath0@xmath60 .
if they do not overlap along the line of sight , then for the longest filament such clouds would number about 10 ( for @xmath61 pc ) to @xmath62 ( for @xmath53 pc ) ; if in the extreme case where the volume filling factor equals the surface filling factor , then the corresponding number of such clouds is about @xmath63 ( for @xmath61 pc ) to @xmath64 ( for @xmath53 pc ) . for comparison ,
the masses of the longest filaments are 67 orders of magnitude larger than the individual jeans masses of such clouds , implying once again that ( the vast majority of ) such clouds would most likely be gravitationally bound . a similar computation for clouds at @xmath65 k leads to the same conclusion .
the characteristic free - fall time , @xmath66 , of a uniform spherical gas cloud ( in the absence of rotation and any other means of support beyond thermal pressure such as turbulence or magnetic fields ) depends only on its mass density , @xmath67 , where @xmath68 is the molecular hydrogen gas number density .
the upper limits placed on the volume of the filaments , together with the unknown filling factor of the molecular gas in these filaments , do not translate into meaningful upper limits for the gas density .
@xcite find from modeling the global co(10 ) , co(21 ) , and co(32 ) line ratios in perseus a a molecular hydrogen gas density @xmath69 ; i.e. , close to the critical density of molecular hydrogen gas for collisional excitation of the co(21 ) transition .
such densities are typical for the bulk of gmcs in our galaxy .
if the filaments , modeled as cylinders , are uniformly filled with gas at such densities , then they would have radial cross - sections of just tens of parsecs or smaller .
the actual gas densities may be higher , implying an upper limit for the free - fall time of @xmath0@xmath70 . for comparison ,
the eastern and western filaments have dynamical ages of about 2030 myr ( with a likely uncertainty of roughly @xmath71 ) , about an order of magnitude longer than the free - fall time .
the filaments as modeled above ( string or complex of uniform spherical gas clouds ) will undergo gravitational collapse at the free - fall time if the sound crossing time is longer .
the sound crossing time can be expressed as @xmath72 , where @xmath73 is the sound speed ( @xmath0@xmath74 at @xmath75 k ) . even for values as small as @xmath76 pc , @xmath77@xmath78 yrs , much longer than the free - fall time .
the same is true for gmcs in our galaxy , which have sound crossing times of @xmath79@xmath4 yrs .
the estimated ages of gmcs in our galaxy ( a highly controversial subject ) ranges from @xmath0@xmath80 yr ( * ? ? ?
* and references therein ) to @xmath0@xmath4 yr ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
the former is consistent with gravitational collapse at the free - fall time , whereas the latter requires gmcs to possess other internal means of support beyond thermal pressure .
the eastern and western filaments therefore have dynamical ages at if not significantly beyond the inferred maximum ages of gmcs in our galaxy .
we first consider whether the tidal shear exerted by perseus a on the filaments along their ( radially aligned ) lengths can prevent them from collapsing ( at least in this direction ) .
we use the same analytical expression for the gravitational potential of perseus a as in paper i : + @xmath81 + where @xmath82 is the gravitational constant , @xmath83 the mass of the galaxy , @xmath84 the radius from center , and @xmath85 for perseus a ( see eq .
( 5 ) of paper i ) . for simplicity , we assume that the filaments can be approximated as uniform cylinders with masses @xmath86 and lengths @xmath87 .
the ( outwards ) tidal force experienced by a test particle along the symmetry axis of such a cylinder at its ends ( i.e. , distance @xmath88 from center ) is therefore + @xmath89 + where @xmath90 is the mass of the test particle .
the ( inward ) gravitational force exerted by the filament itself on that test particle is given by + @xmath91 \ , \ ] ] + where @xmath92 is the radius of the cylinder .
for thin cylinders where @xmath93 , + @xmath94 the ratio of the outward tidal force exerted by perseus a to the inward gravitational force exerted by the cylinder at its ends is therefore : + @xmath95 let us consider the eastern filament , which has a mass @xmath96 and , assuming an inclination of @xmath0@xmath97 ( paper i ) , a deprojected length @xmath98 kpc located at a deprojected radius @xmath99 . with @xmath100 for perseus a ( see paper i ) , and defining @xmath101 where @xmath102 , the ratio @xmath103 .
we do not resolve the eastern filament in width ( i.e. , @xmath104 kpc ) , implying that @xmath105 and so @xmath106 .
thus , the outward tidal force is at best comparable with the inward gravitational force of the filament itself , and much less if the filament is much thinner than the upper limit imposed by our observation .
the western filament has a mass @xmath107 and , assuming an inclination of @xmath0@xmath108 , a deprojected length @xmath109 kpc located at a deprojected radius @xmath110 .
once again we do not resolve this filament in width , implying that @xmath111 and so @xmath112 .
thus , just like the eastern filament , the outward tidal force is at best comparable with the inward gravitational force of the filament itself .
assuming an inclination of @xmath0@xmath113 , we find @xmath114 for e1 , @xmath115 for e2 , and @xmath112 for w1 . for these shorter filaments located at relatively large distances from center , the outward tidal force is weaker than if not negligible compared with the inward gravitational force of the filaments themselves .
the inner filament exhibits more complex kinematics than the eastern and western filaments , and as explained in paper i may comprise two separate filaments . assuming that it is a single filament oriented at an inclination of @xmath0@xmath113 , its center would then be located at @xmath116 . with a mass @xmath117 and deprojected length @xmath118 kpc , @xmath111 and hence @xmath119 .
thus , even for the inner filament where the tidal shear is strongest , the outward tidal force is at best comparable with the inward gravitational force of the filament itself .
tidal shear becomes negligible if each of the filaments comprise a string or complex of gmcs , modeled here as uniform spherical clouds with densities of at least @xmath69 .
consider a single such cloud located at the midpoint of the filament ( so that gravitational forces from clouds on opposing sides of the filament cancel ) with mass @xmath120 and radius @xmath47 . the gravitational force on a test particle with mass @xmath90 at a given radius @xmath121 in the cloud is given by : + @xmath122 + the ratio of the outward tidal force exerted by perseus a ( replacing @xmath87 with @xmath123 in eq .
( 2 ) ) to the inward gravitational force exerted by the sphere is then given by : + @xmath124 + the largest and most massive gmcs in our galaxy have @xmath125 and @xmath126 pc ( and thus @xmath69 ) , giving @xmath127 at the location of the different filaments .
as @xmath128 , this result holds for gmcs of any dimensions provided they all have the same densities as considered here .
as shown in paper i , the spatial - kinematic distribution of the eastern and western filaments can be modeled as free - fall in the gravitational potential of perseus a provided that the mass of this galaxy is about a factor of @xmath02 smaller than that inferred by @xcite .
equivalently , the measured ( deprojected ) velocity gradients of the eastern and western filaments are shallower than those predicted for free - fall if perseus a has the mass inferred by @xcite .
we mentioned in paper i that our simple model neglected a number of effects , in particular the outward pressure exerted by the x - ray gas on the infalling filaments . here
we note that , over time , the self - gravity of the filaments also can reduce any intrinsic velocity gradient along their lengths .
if the molecular filaments are supported by turbulence , then the required turbulent velocities can be estimated from the virial theorem expressed as @xmath129 for a uniform sphere , where @xmath130 is the average particle velocity . as an illustration ,
take the least massive and shortest filament w1 , which has a mass of @xmath0@xmath131 enclosed in @xmath132 kpc .
the required velocity of the molecular hydrogen gas particles to support w1 against collapse is @xmath133 ( note that thermal velocities are @xmath134 for temperatures
@xmath8100 k ) .
this is about a factor of @xmath02 smaller than the co(21 ) linewidth of w1 ( table [ gas_mass ] ) , indicating that this filament may be supported by turbulence
. an upper limit of up to a factor of several higher turbulent velocity is required to support the other more massive filaments against collapse , which in all cases is smaller than their observed linewidths .
observations at higher angular resolutions are required to measure the intrinsic linewidths of the filaments ( i.e. , omitting systematic motions ) or their constituent gmcs so as to determine their true turbulent velocities . at the present time
, we can not rule out the possibility that the filaments are supported by turbulence , generated perhaps in the process of condensing from the x - ray cooling flow .
turbulence normally cascades to ever smaller spatial scales , and dissipates on a timescale shorter than the crossing time at the maximum turbulent velocity ( e.g. , * ? ? ?
for the derived parameters of w1 , this turbulent decay timescale is at most @xmath0@xmath4 yrs .
the other filaments also have similar turbulent crossing times , comparable with if not shorter than the dynamical ages of the eastern and western filaments .
thus , it may be that a source is required to at least partially regenerate turbulence over the inferred lifetimes of the eastern and western filaments .
gmcs in our galaxy also have supersonic linewidths , typically several @xmath135 ; how such turbulence is maintained is not understood , even though these gmcs have lifetimes that are at most comparable with the dynamical ages of the eastern and western filaments .
if the molecular filaments are supported by magnetic fields , we can make the following simplifying assumptions to roughly estimate the required field strengths . in a uniform magnetic field , the filaments ( or gmcs )
will collapse along the field lines but be ( temporarily ) prevented from collapsing perpendicular to the field lines ( by collisions with ions ) to form thin sheets .
the magnetic flux , @xmath136 , required to support a uniform sheet with a given mass , @xmath137 , against its own gravity is given by @xmath138 @xcite .
for example , if the western filament comprises a sheet with a deprojected diameter of 3.6 kpc and a mass @xmath139 , it is required to have a magnetic field strength of at least @xmath0@xmath140 to prevent its ( immediate ) collapse .
the other filaments require minimum magnetic field strengths in the range @xmath010@xmath141 .
this is towards the low end of the range measured for the magnetic field strengths in galactic gmcs @xcite .
@xcite suggest that the thread - like h@xmath1 filaments are supported by magnetic fields with strengths of a few tens of @xmath142 g , comparable with the values estimated here to support the molecular filaments against collapse ( in the absence of other means of nonthermal support such as rotation and turbulence ) .
ambipolar diffusion @xcite will eventually cause the filaments ( or their constituent complexes of gmcs ) to collapse across the magnetic field lines , as is thought to be the case for gmcs in our galaxy .
the ambipolar diffusion timescale is given by @xmath143 @xcite , where @xmath144 is the neutral - ion collision time , in the situation where @xmath145 @xcite .
for gmcs in our galaxy , @xmath14610 myr depending on @xmath147 in specific regions of the cloud .
for the eastern and western filaments to resist ambipolar diffusion , they also are required to have neutral - ion ( and/or neutral grain ) collision times comparable with if not significantly longer than that inferred for galactic gmcs .
although there is no evidence for star formation in the molecular filaments that we detect , there is evidence for recent and widespread star formation in per a ( * ? ? ? * and references therein ) , including the hii region mentioned in @xmath43 discovered by @xcite . assuming that any associated co(21 ) emission has an overall linewidth of @xmath0@xmath25 ( see @xmath148 ) , we place a 3@xmath24 upper limit of @xmath149 on its associated mass in molecular hydrogen gas ( a smaller linewidth would result in an even smaller upper limit ) .
an even more stringent upper limit ( assuming the same line properties ) of @xmath150 can be placed on any associated molecular hydrogen gas from the more sensitive observation of @xcite with the iram pdbi . for comparison , @xcite infer a mass for the stellar cluster of @xmath0@xmath151 assuming a salpeter imf .
if @xmath0@xmath152 of the pre - existing molecular hydrogen gas has been converted to stars in this cluster as is typically inferred for gmcs in our galaxy , one would have expected to find @xmath0@xmath153 of molecular hydrogen gas associated with this cluster .
unless this gas has been largely dissociated or widely dispersed , or the assumption of a salpeter imf invalid ( i.e. , top - heavy imf ) , the star - formation efficiency associated with this cluster must have been extraordinarily high ; i.e. , nearly all the molecular hydrogen gas has been converted into stars .
as mentioned in @xmath154 , we recovered only about one - half of the co(21 ) emission measured by single - dish telescopes for the main body of the molecular gas in per a. in this region , about half of the co(21 ) emission must therefore be quite smoothly distributed on angular scales larger than @xmath0510 , and hence resolved out in our interferometric observations , when observed at velocity resolutions of @xmath0@xmath25 .
@xcite list in their table 3 the integrated co(21 ) intensities at locations in their single - dish iram 30-m map where the s / n ratio @xmath155 .
we have extracted from this table those locations that best overlap with the co(21 ) filaments detected in our observations . in figure [ salome ] ( left panel ) , we plot circles showing the fwhm of the iram 30-m primary beam centered at these locations on our integrated co(21 ) intensity map of figure [ mom - all ] . to compare with the iram 30-m measurements , we first convolved our channel maps to the same angular resolution as the iram 30-map ( i.e. , with a gaussian beam of size 12 at fwhm ) , and then computed the integrated flux densities at the same locations over the velocity range @xmath156 to @xmath157 .
the integrated flux densities measured in the iram 30-map versus those measured in our sma maps are plotted in the graph shown also in figure [ salome ] ( right panel ) . as can be seen
, there is an overall correspondence between the co(21 ) emission detected in our interferometric and that detected in the single - dish observations .
we recovered ( within measurement uncertainties ) all the co(2 - 1 ) emission in the eastern region , comprising the eastern filament , e1 , and e2 .
@xcite reached the same conclusion for their map of this region using the iram pdbi , even though their observations do not have baselines as short as ours .
we also recovered all the co(21 ) emission in the inner region .
the emission that we measured in the western region , however , is systematically below that measured in the single - dish observations . in this region
, we recover less than half of the co(21 ) emission present in the iram 30-m map . to recover the remaining co(21 ) emission
, we plan to use the sub - compact configuration of the sma to provide shorter baselines than already available in our observations thus far .
our ( paper i ) previous interferometric co(21 ) observation of per a spanned its central region of radius @xmath010 kpc . at a spatial resolution of @xmath01
kpc , this observation revealed three radial filaments lying approximately east - west , together with a number of other features that were at best poorly resolved . in this manuscript , we reported follow - up interferometric co(21 ) observations that extend our coverage in the east - west direction to span the entire main body of the cool molecular gas , which in single - dish maps can be traced out to @xmath014 kpc east and west of center .
our results showed that : * no new features were detected beyond the region previous mapped , with the most distant features detectable in our maps lying @xmath010 kpc east ( e1 ) and @xmath08 kpc west ( w1 ) of center .
the two co(21 ) features referred to in paper i as n1 and s1 lying north and south of the center are sidelobe artifacts .
all the other features are confirmed to be real . a previously unrecognized feature with distinct kinematics labeled w2 ( fig . [ mom - all ] )
is detected partially blended with the western filament lying south about halfway along the length of this filament .
* with our improved sensitivity and angular resolution from combining the present and previous observations , all the co(21 ) features ( except w2 ) can now be seen to comprise filaments that ( within measurement uncertainties ) are radially aligned with respect to the center of per a. six distinct filaments ( inner filament , eastern filament , e1 , e2 , western filament , w1 ) can now be identified .
our sensitivity and/or angular resolution is not sufficient to discern a velocity pattern in the e1 , e2 , and w1 filaments .
the newly recognized feature w2 is not spatially resolved . *
all the detected filaments exhibit a roughly constant ratio between their co(21 ) and h@xmath1 luminosities of @xmath0@xmath2 . on the other hand ,
not all equally if not more luminous h@xmath1 features in the area observed including the hii region discovered by @xcite that lies just `` downstream '' of e2 have detectable co(21 ) emission .
the upper limits on their co(21 ) to h@xmath1 luminosities is a factor of @xmath05 lower that measured for the co(21 ) filaments .
an even more stringent upper limit for this ratio of an order of magnitude lower than that derived for the co(21 ) filaments can be placed for the hii region from the more sensitive observation of @xcite . *
about one - half of the co(21 ) emission detected in single - dish observations with the iram 30-m telescope @xcite from the main body of the molecular gas was recovered in our interferometric observations with the sma . towards the inner and eastern regions , all ( or the bulk ) of the co(21 )
emission was recovered . on the other hand , towards the western region ,
less than half of the co(21 ) emission was recovered . * the radial alignment of e1 , e2 , and w1 provides further support for our argument that all the cool molecular gas detected is most likely deposited by a x - ray cooling flow .
the lack of any transverse component in the cool molecular filaments argues against ( but does not negate ) the idea that this gas may have been dragged outwards by buoyant x - ray bubbles ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , as has been proposed to explain transverse h@xmath1 filaments seen in the outer regions of the nebula @xcite .
in addition , there are no detectable x - ray bubbles in the directions of these filaments . * the existence of a correlation between the co(21 ) and h@xmath1 luminosities for all the filaments detected may suggest that the excitation of their cool molecular and ionized gas is causally related .
alternatively , this dependence may support the picture in which the h@xmath1 gas comprises the ionized skins of the cool molecular gas filaments ( e.g. , * ? ? ?
* if the molecular filaments are forming stars , then their star formation efficiency must be much lower than the nearly constant star formation efficiency found for molecular clouds in nearby normal spiral galaxies . *
the lack of cool molecular gas associated with the giant hii region discovered by @xcite implies that , unless this gas has been largely dissociated or widely dispersed , or the stellar imf distinctly non - salpeter ( i.e. , top - heavy imf ) , the star formation in this region must have been extraordinarily efficient ( i.e. , essentially all the molecular gas must have been converted to stars ) .
optical spectroscopy of the features labeled a d in figure [ co - haf ] is needed to determine whether they also are giant hii regions , or whether there is a relatively large scatter in the co(21 ) to h@xmath1 luminosities even between different features in the low ionization and excitation optical emission - line nebula .
* the outward tidal force exerted by perseus a on the individual filaments is at best comparable to , and for the shorter filaments much weaker than , the inward gravitational force exerted by the filaments themselves . if comprising a string or complex of gmcs , all the individual gmcs are gravitationally bound .
such gmcs , if supported only by thermal pressure , should collapse on timescales no longer than @xmath0@xmath80 yrs . by comparison ,
the inferred dynamical ages of the eastern and western filaments are an order of magnitude longer , @xmath0@xmath4 yrs ( paper i ) . *
if the cool molecular filaments are supported by turbulence , non - thermal velocities of a few to several tens of @xmath5 are required .
this is smaller than the observed co(21 ) linewidths of all the filaments at a spatial resolution of @xmath01 kpc , although observations at higher angular resolutions are required to measure the true intrinsic linewidths of these filaments ( or their constituent gmc complexes ) and hence their actual turbulent velocities .
such turbulence should decay on a timescale no longer than @xmath0@xmath4 yrs , comparable with if not shorter than the dynamical ages of the eastern and western filaments .
the filaments may therefore be supported by turbulence , generated perhaps in the process of condensing from the x - ray cooling flow , although a source may be required to maintain this turbulence over the inferred lifetimes of the eastern and western filaments .
* magnetic fields with strengths of at least several @xmath0@xmath6 g are required to support ( perpendicular to the field lines ) the cool molecular filaments against their own gravity .
this is a comparable with the estimated magnetic field strengths required to support the thread - like h@xmath1 filaments @xcite . to resist ambipolar diffusion for timescales of @xmath0@xmath4 yrs ( in the absence of any other means of support beyond thermal pressure ) , magnetic field strengths of this order and neutral - ion ( and/or neutral grain ) collision times similar to that in galactic gmcs are required .
we are deeply indebted to the referee for carefully reading our manuscript , picking up an important error , asking us to consider the effects of tidal shear , and suggesting other general improvements to the manuscript .
we thank leo blitz for bringing to our attention the nearly constant star formation efficiency found for molecular gas in nearby normal spiral galaxies , and encouraging us to compute the resultant h@xmath1 luminosity for the molecular filaments we detect for the same star formation efficiency .
we thank yi - ping ao for help in preparing the observing script , and philippe salom for kindly providing the iram 30-m co(21 ) map of per a that we reproduced in figure [ 3pointing ] .
j. lim acknowledges a grant from the national science council of taiwan in support of this work .
this grant also provides a masters student stipend for i-.t . ho . of text ) .
dark contour levels are plotted at 3 , 6 , 9 , 12 , 15 , 20 , 30 , 40 , 50 , 70 , and @xmath158 .
the dash circle corresponds to the fwhm of the sma primary beam centered on the center of per a , which is marked by the cross at the origin .
this is the only region covered in our previous observation .
the two dash - dot circles correspond to the fwhm of the sma primary beams centered approximately on the eastern and western filaments as marked by the two crosses to the east and west of center respectively .
our present observation covers all three regions , thus spanning the main body of cool molecular gas in per a. the angular resolutions of the single - dish and interferometric measurements are shown in the lower right corner .
the larger circle corresponds to the fwhm of the iram-30 m primary beam , and the smaller filled ellipse to the fwhm of the sma synthesized beam.[3pointing],width=604 ] for the inner and western regions , and -3 ( dashed ) , 3 , 4 , 5 , 6 , 8 , and @xmath159 for the eastern region .
the gray background corresponds to the integrated co(2 - 1 ) intensity after primary beam correction as shown in figure [ mom - all ] .
the position of the nuclear continuum source in per a is indicated by a cross .
the synthesized beam is shown as an ellipse at the lower right corner of the lower right panels for the two regions .
[ channel - maps],title="fig:",width=415 ] + for the inner and western regions , and -3 ( dashed ) , 3 , 4 , 5 , 6 , 8 , and @xmath159 for the eastern region .
the gray background corresponds to the integrated co(2 - 1 ) intensity after primary beam correction as shown in figure [ mom - all ] .
the position of the nuclear continuum source in per a is indicated by a cross .
the synthesized beam is shown as an ellipse at the lower right corner of the lower right panels for the two regions .
[ channel - maps],title="fig:",width=415 ] + .
the dotted ellipse indicates where the noise has increased by a factor of 2 over that at the center .
no genuine features were detected beyond the region shown ( see text ) .
this map was made by hanning smoothing the corresponding channel maps by @xmath160 , and then discarding pixels in the channel maps with intensities less than 37 mjy .
the synthesized beam is plotted at the lower left corner.[mom - all],width=453 ] .
the arrows indicate the direction to the center of per a. the synthesized beam is plotted at the lower right corner of the right panel .
all the moment maps were made by hanning smoothing the corresponding channel maps by @xmath161 , and then for e1 and e2 discarding pixels in the channel maps with intensities less than 25 mjy and for w1 less than 20 mjy .
[ uniform - weighting],width=604 ] @xmath60 , @xmath162 , @xmath163 , @xmath164 , @xmath26 , @xmath165 , @xmath166 , @xmath167 , and @xmath168 of the peak . dirty channel map ( lower left panel ) , clean channel map ( lower middle panel ) , and dirty channel map ( lower right panel ) after the western filament has been subtracted from the original dirty map .
contours levels in the channel maps are plotted at -3 ( dashed ) , 2 , 3 , 4 , 5 , and @xmath169 mjy .
the dash circles corresponds to the fwhm of the sma primary beam .
the synthesized beam is plotted at the lower right corner of the clean map .
the channel maps are taken from paper i , and correspond to a velocity of @xmath170 with respect to the systemic heliocentric velocity .
the feature s1 lies at the first sidelobe of the western filament , and remains as a significant feature after cleaning .
after the western filament is subtracted from the dirty map , however , s1 is no more intense than the low - level fluctuations , indicating that it is just a sidelobe .
[ artifact],title="fig:",width=302 ] + @xmath60 , @xmath162 , @xmath163 , @xmath164 , @xmath26 , @xmath165 , @xmath166 , @xmath167 , and @xmath168 of the peak .
dirty channel map ( lower left panel ) , clean channel map ( lower middle panel ) , and dirty channel map ( lower right panel ) after the western filament has been subtracted from the original dirty map .
contours levels in the channel maps are plotted at -3 ( dashed ) , 2 , 3 , 4 , 5 , and @xmath169 mjy .
the dash circles corresponds to the fwhm of the sma primary beam .
the synthesized beam is plotted at the lower right corner of the clean map .
the channel maps are taken from paper i , and correspond to a velocity of @xmath170 with respect to the systemic heliocentric velocity .
the feature s1 lies at the first sidelobe of the western filament , and remains as a significant feature after cleaning . after the western filament is subtracted from the dirty map , however , s1 is no more intense than the low - level fluctuations , indicating that it is just a sidelobe .
[ artifact],title="fig:",width=188 ] @xmath60 , @xmath162 , @xmath163 , @xmath164 , @xmath26 , @xmath165 , @xmath166 , @xmath167 , and @xmath168 of the peak . dirty channel map ( lower left panel ) , clean channel map ( lower middle panel ) , and dirty channel map ( lower right panel ) after the western filament has been subtracted from the original dirty map .
contours levels in the channel maps are plotted at -3 ( dashed ) , 2 , 3 , 4 , 5 , and @xmath169 mjy .
the dash circles corresponds to the fwhm of the sma primary beam .
the synthesized beam is plotted at the lower right corner of the clean map .
the channel maps are taken from paper i , and correspond to a velocity of @xmath170 with respect to the systemic heliocentric velocity .
the feature s1 lies at the first sidelobe of the western filament , and remains as a significant feature after cleaning . after the western filament is subtracted from the dirty map , however , s1 is no more intense than the low - level fluctuations , indicating that it is just a sidelobe .
[ artifact],title="fig:",width=188 ] @xmath60 , @xmath162 , @xmath163 , @xmath164 , @xmath26 , @xmath165 , @xmath166 , @xmath167 , and @xmath168 of the peak .
dirty channel map ( lower left panel ) , clean channel map ( lower middle panel ) , and dirty channel map ( lower right panel ) after the western filament has been subtracted from the original dirty map .
contours levels in the channel maps are plotted at -3 ( dashed ) , 2 , 3 , 4 , 5 , and @xmath169 mjy .
the dash circles corresponds to the fwhm of the sma primary beam .
the synthesized beam is plotted at the lower right corner of the clean map .
the channel maps are taken from paper i , and correspond to a velocity of @xmath170 with respect to the systemic heliocentric velocity .
the feature s1 lies at the first sidelobe of the western filament , and remains as a significant feature after cleaning .
after the western filament is subtracted from the dirty map , however , s1 is no more intense than the low - level fluctuations , indicating that it is just a sidelobe .
[ artifact],title="fig:",width=188 ] , with the same contour levels plotted ) superposed on a color map of the h@xmath1@xmath29n[ii ] intensity ( from * ? ? ?
the top panel shows the entire region where co(21 ) was detected .
the middle panel shows the inner and western regions , and the bottom panel the eastern region .
the h@xmath1@xmath29n[ii ] maps are plotted with different stretches in the individual panels to emphasize different features of interest . the synthesized beam is plotted at the lower right corner of each panel .
the h@xmath1@xmath29n[ii ] features labeled a d ( see text ) in the middle panel corresponds to those refered to in figure [ mass - haf - corr ] .
the h@xmath1@xmath29n[ii ] feature labeled hii region in the bottom panel is that discovered by @xcite .
[ co - haf],title="fig:",width=340 ] + , with the same contour levels plotted ) superposed on a color map of the h@xmath1@xmath29n[ii ] intensity ( from * ? ? ?
the top panel shows the entire region where co(21 ) was detected .
the middle panel shows the inner and western regions , and the bottom panel the eastern region .
the h@xmath1@xmath29n[ii ] maps are plotted with different stretches in the individual panels to emphasize different features of interest . the synthesized beam is plotted at the lower right corner of each panel .
the h@xmath1@xmath29n[ii ] features labeled a d ( see text ) in the middle panel corresponds to those refered to in figure [ mass - haf - corr ] .
the h@xmath1@xmath29n[ii ] feature labeled hii region in the bottom panel is that discovered by @xcite .
[ co - haf],title="fig:",width=340 ] + , with the same contour levels plotted ) superposed on a color map of the h@xmath1@xmath29n[ii ] intensity ( from * ? ? ?
the top panel shows the entire region where co(21 ) was detected .
the middle panel shows the inner and western regions , and the bottom panel the eastern region .
the h@xmath1@xmath29n[ii ] maps are plotted with different stretches in the individual panels to emphasize different features of interest . the synthesized beam is plotted at the lower right corner of each panel .
the h@xmath1@xmath29n[ii ] features labeled a d ( see text ) in the middle panel corresponds to those refered to in figure [ mass - haf - corr ] .
the h@xmath1@xmath29n[ii ] feature labeled hii region in the bottom panel is that discovered by @xcite .
[ co - haf],title="fig:",width=264 ] + ) versus the h@xmath1@xmath29n[ii ] luminosity of individual filaments .
the solid line corresponds to the best fit linear ratio of 0.0007 , and the dash - dot lines to other ratios as indicated .
the features a d and hii region correspond to those labeled in figure [ co - haf ] .
[ mass - haf - corr],width=529 ] .
the dotted ellipse indicates where the noise has increased by a factor of 2 over that at the center .
the circles correspond to the primary beams of individual iram-30 m telescope pointings of @xcite , with those in blue corresponding to the eastern region , those in red the inner region , and those in green the western region .
right panel shows a plot of the molecular hydrogen gas mass measured in the individual pointings of the iram-30 m telescope versus that measured by us with the sma in the same area .
the dash line indicates the same molecular hydrogen gas masses measured by the two telescopes .
virtually all the co(21 ) emission in the eastern and inner regions are recovered in our sma observation , but only a fraction of that in the western regions .
[ salome],title="fig:",width=340 ] .
the dotted ellipse indicates where the noise has increased by a factor of 2 over that at the center .
the circles correspond to the primary beams of individual iram-30 m telescope pointings of @xcite , with those in blue corresponding to the eastern region , those in red the inner region , and those in green the western region .
right panel shows a plot of the molecular hydrogen gas mass measured in the individual pointings of the iram-30 m telescope versus that measured by us with the sma in the same area .
the dash line indicates the same molecular hydrogen gas masses measured by the two telescopes .
virtually all the co(21 ) emission in the eastern and inner regions are recovered in our sma observation , but only a fraction of that in the western regions .
[ salome],title="fig:",width=226 ] | we have extended our previous observation @xcite of ngc 1275 ( perseus a , the central giant elliptical galaxy in the perseus cluster ) covering a central radius of @xmath010 kpc to the entire main body of cool molecular gas spanning @xmath014 kpc east and west of center .
we find no new features beyond the region previously mapped , and show that all six spatially - resolved features on both the eastern and western sides ( three on each side ) comprise radially aligned filaments .
such radial filaments can be most naturally explained by a model in which gas deposited `` upstream '' in localized regions experiencing a x - ray cooling flow subsequently free falls along the gravitational potential of per a , as we previously showed can explain the observed kinematics of the two longest filaments .
all the detected filaments coincide with locally bright h@xmath1 features , and have a ratio in co(21 ) to h@xmath1 luminosity of @xmath0@xmath2 ; we show that these filaments have lower star formation efficiencies than the nearly constant value found for molecular gas in nearby normal spiral galaxies . on the other hand , some at least equally luminous h@xmath1 features ,
including a previously identified giant hii region , show no detectable cool molecular gas with a corresponding ratio at least a factor of @xmath05 lower ; in the giant hii region , essentially all the pre - existing molecular gas may have been converted to stars .
we demonstrate that all the cool molecular filaments are gravitationally bound , and without any means of support beyond thermal pressure should collapse on timescales @xmath3 yrs . by comparison , as we showed previously the two longest filaments have much longer dynamical ages of @xmath0@xmath4 yrs .
tidal shear may help delay their collapse , but more likely turbulent velocities of at least a few tens @xmath5 or magnetic fields with strengths of at least several @xmath0@xmath6 g are required to support these filaments . |
chiral symmetry breaking @xmath7 and its restoration , with @xmath8 light quark flavours , has been a milestone in our present understanding of the quantum chromodynamics ( qcd ) phase diagram and hadronic physics under extreme conditions of temperature @xmath0 and baryon density , as those produced in heavy - ion and nuclear matter experimental facilities such as rhic , cern ( alice ) and fair .
lattice simulations support that deconfinement and chiral restoration take place very close to one another in the phase diagram . in the physical case @xmath9 ( @xmath10 ) and for vanishing baryon chemical potential ,
they point towards a smooth crossover transition at pseudocritical temperature @xmath11 145 - 165 mev @xcite , the results being fairly consistent with the @xmath3 universality class @xcite , which would hold for two light flavours in the chiral limit @xmath12 .
the crossover nature of the transition means in particular that there is no unique way to identify the transition point , the most efficient one in lattice being the scalar susceptibility peak position , rather than the vanishing point for the quark condensate @xmath13 , the order parameter , which decreases asymptotically with @xmath0 for @xmath14 .
the equivalence with the @xmath15 breaking pattern for @xmath16 led to early proposals of @xmath17 meson degeneration ( chiral partners " ) at chiral restoration @xcite which in its simplest linear realization takes place through the @xmath18-component of the @xmath3 field @xmath19 acquiring a thermal vacuum expectation value and mass both vanishing at the transition in the chiral limit . degeneration in the vector - axial vector sector ( @xmath20 and @xmath21 states ) as a signature of chiral restoration has also been thoroughly studied @xcite .
nowadays , we know that the @xmath18 state is well established as a @xmath22 scattering broad resonance for isospin and angular momentum @xmath23 , known as @xmath24 @xcite , which is then difficult to accommodate as an asymptotically free state , like in the linear model . precisely , one of our main conclusion here will be that this asymptotic description is not needed .
in fact , in order to study chiral partner degeneration in the scalar - pseudoscalar sector , it is more appropriate to analyze the corresponding current correlation functions @xcite which can be derived from a chiral effective lagrangian without introducing explicitly a particle - like @xmath18 degree of freedom .
the scalar and pseudoscalar susceptibilities in terms of the corresponding qcd @xmath25 currents are given by : @xmath26}=\int_e{d^4x \left[\frac{\delta}{\delta s(x)}\frac{\delta}{\delta s(0)}z[s , p]\bigg\vert_{s = m_q , p^a=0}\right]},\label{chisdef}\\ \chi_p ( t)\delta^{ab}&=&\int_e{d^4x \langle { \cal t } p^a ( x ) p^b ( 0 ) \rangle_t}\equiv \delta^{ab } \int_e{d^4x k_p ( x)}= \int_e{d^4x \left[\frac{\delta}{\delta p^a(x)}\frac{\delta}{\delta p^b(0)}z[s , p]\bigg\vert_{s = m_q , p^a=0}\right]},\label{chipdef}\end{aligned}\ ] ] where @xmath27 is the quark field , @xmath28 and @xmath29 are respectively the pseudoscalar current and its correlator , the euclidean measure @xmath30 with @xmath31 and @xmath32 denotes a thermal average . for @xmath1 , parity invariance of the qcd vacuum ( @xmath33 ) and isospin symmetry have been used . in the above equation , @xmath34 $ ]
is the qcd generating functional with scalar and pseudoscalar sources @xmath35 coupled to the massless lagrangian in the light sector as @xmath36 , so that @xmath37 $ ] is the qcd partition function .
thus , should the scalar and pseudoscalar currents become degenerate at chiral restoration , @xmath5 and @xmath4 would meet at that point .
since @xmath2 is expected to increase , as a measure of the fluctuations of the order parameter , at least up to the transition point , it seems plausible that they meet near the transition . in an ideal @xmath3 pattern ,
the matching should take place near the maximum of @xmath2 .
since @xmath38 has the quantum numbers of the pion field @xmath39 , its correlators , like @xmath29 , are saturated by the pion state at low energies .
let us first review the prediction arising from the low - energy theorems of current algebra , equivalent to the leading order ( lo ) in the low - energy expansion of chiral lagrangians . at that order
, one has @xmath40 ( from pcac theorem ) with @xmath41 and where @xmath42 and @xmath43 are the pion decay constant and mass respectively , so that @xmath44 from eq.([chipdef ] ) , being @xmath45 the pion propagator in momentum space @xmath46 and @xmath47 the matsubara frequency with integer @xmath48 .
thus , the pseudoscalar correlator , saturated with the dominant pion state , is just proportional to the pion propagator at this order .
in addition , to lo the euclidean propagator is just the free one @xmath49 ( interactions are suppressed at low energies ) with @xmath50 , so that using also the gell - mann - oakes - renner ( gor ) relation @xmath51 , valid at this order , we would get @xmath52 , as a first indication of the relation between the pseudoscalar susceptibility and the quark condensate at the lo given by current algebra .
the latter result can actually be obtained formally as a ward identity ( wi ) from the qcd lagrangian @xcite , in connection with the definition of the quark condensate for lattice wilson fermions @xcite . however
, both sides of the identity suffer from qcd renormalization ambiguities , so that this wi is formally well - defined only for exact chiral symmetry @xcite .
it is therefore interesting to study , and so we will do in the next section , how this identity is realized within chiral perturbation theory ( chpt ) @xcite , which describes the low - energy chiral symmetry broken phase of qcd in a model - independent framework where symmetry breaking is realized non - linearly and pions are the only degrees of freedom in the lagrangian .
the previous current - algebra results are actually just the lo in the chpt expansion in powers of a generic low - energy scale @xmath53 , denoting pion momenta or temperature , relative , respectively , to @xmath54 1 gev and @xmath55 .
these are nothing but indicative natural upper limits for the chiral expansion in terms of scattering ( typical resonance scale ) and thermodynamics ( critical phenomena ) respectively , although both are treated on the same foot in the chiral expansion .
in particular , the lo prediction for @xmath1 is temperature independent , so it is not obvious that it can be simply extrapolated as , say , @xmath56 .
actually , all the quantities involved change with temperature due to pion loop corrections , namely @xmath57 , @xmath58 and @xmath59 @xcite . similarly , from eq.([chisdef ] ) , one can relate @xmath2 with the propagator of a `` @xmath18-like state '' such that it couples linearly to the external scalar source @xmath60 in an explicit symmetry - breaking term @xmath61 . without further specification about its nature and its coupling to other physical states such as pions ,
one already gets @xmath62 , suggesting a growing behaviour inversely proportional to @xmath63 as the sigma state reduces its mass to become degenerate with the pion .
we also recall that the problem of @xmath64 degeneration has been studied in nuclear matter at @xmath65 in @xcite , to linear order in nuclear density . in that work ,
current algebra is assumed to hold through pcac in the operator representation and other low - energy theorems such as gor , which as discussed in the previous paragraphs , leads to the pseudoscalar correlator @xmath29 being directly proportional to the pion propagator .
the authors in @xcite work within low - energy models at finite density for which this pcac realization holds , so that by including the proper finite - density corrections to @xmath66 , which carries out all the density dependence of @xmath67 through an in - medium mass , and to @xmath68 , the relation @xmath52 is found to hold in the nuclear medium .
this result provides another supporting argument for the relation between the condensate and the pseudoscalar susceptibility and represents an additional motivation for our present chpt analysis , where we do not need to make any assumption about the validity of current algebra .
the next to leading order ( nlo ) corrections to @xmath1 can be obtained systematically and in a model - independent way within chpt , where one can also calculate the scalar susceptibility @xmath2 to a given order only in terms of pion degrees of freedom .
the price to pay is that we expect to reproduce only the behaviour of @xmath69 for low and moderate temperatures .
however , since @xmath1 is dominated by pions , whose dynamics are well described through chpt , we expect to obtain a reasonable qualitative description of its @xmath0 behaviour , whereas standard chpt misses the peak structure of @xmath2 near the transition .
we note in turn that the lo for @xmath2 vanishes , unlike that of @xmath1 .
the chpt nlo result for @xmath70 can be found in @xcite . for @xmath71
we consider the effective lagrangian @xmath72 , where @xmath73 , including their dependence on the pseudoscalar source @xmath74 as given in @xcite .
we follow similar steps as in @xcite , now for the pseudoscalar correlator @xmath29 .
the lo comes from @xmath75 only and reproduces the current - algebra prediction .
the nlo corrections to @xmath29 are of the following types : * the nlo corrections to the pion propagator @xmath76 , which come both from @xmath75 one - loop tadpole - like contributions @xmath77 and from tree level @xmath78 constant terms , * pion self - interactions @xmath79 in @xmath75 contributing as @xmath80 , * crossed terms @xmath81 giving @xmath82 multiplied by a nlo contribution , * @xmath83 terms giving rise to a contact contribution proportional to @xmath84 . the final result for the pseudoscalar correlator in momentum space for euclidean four - momentum @xmath53
can be written as : @xmath85 where subleading terms are labeled by their @xmath86 dependence .
the nlo propagator includes wave function and mass renormalization ( at this order in chpt there is no imaginary part for the self - energy ) : @xmath87 where the lo propagator corresponds to @xmath88 and is temperature independent .
@xmath42 and @xmath43 are the lagrangian pion mass and decay constant , related to the vacuum ( @xmath65 ) physical values @xmath89 mev , @xmath90 93 mev , by @xmath91 corrections @xcite .
the constant @xmath92 in eq.([kp ] ) is temperature independent and is a finite combination of low - energy constants ( lec ) of @xmath78 @xcite . in the chpt scheme
, the divergent part of the @xmath78 lec cancels the loop divergences from @xmath75 such that pion observables are finite and independent of the low - energy renormalization scale . to nlo in chpt all the pion loop contributions in eq.([kp ] ) are proportional to the tadpole - like contribution @xmath93 with the thermal function : @xmath94 which is an increasing function of @xmath0 for any mass .
thus , the pion thermal mass in the nlo propagator is given at this order by @xmath95 $ ] and is finite and scale - independent .
the same holds for @xmath96 $ ] and for @xmath97 $ ] @xcite .
note that @xmath98 decreases with @xmath0 a factor of 3 slower than the condensate @xmath13 .
in addition , to this order it holds @xmath99 .
that is , the gor relation is broken at finite temperature to nlo by the same @xmath65 terms , given in @xcite .
gor holds to nlo only in the chiral limit , including temperature effects @xcite .
the constant @xmath100 in eq.([kp ] ) includes both lec contributions and loop functions and the same happens with the wave function renormalization constant @xmath101 .
both are divergent , but the combination @xmath102 turns out to be finite and scale - independent . note that if we replace @xmath103 in the last term in eq.([kp ] ) , which is allowed at this order since @xmath104 is of nlo , that combination is precisely the one multiplying the nlo propagator , i.e , it is the @xmath0-dependent residue at the @xmath105 pole ( when the euclidean propagator is analytically continued to the retarded one ) . that finite residue , being finite , has to be then a combination of the finite observables involved .
actually , it happens to be : @xmath106 thus , the expression in eq.([residue ] ) represents the residue of the nlo @xmath67 correlator ( [ kp ] ) at the thermal pion pole .
note that by showing explicitly that the residue can be expressed as ( [ residue ] ) we obtain that the thermal part of the pseudoscalar susceptibility @xmath5 is the same as that in @xmath107 , since @xmath108 , so that @xmath109 now , since @xmath110 , at the nlo order we are working , we can replace in eq.([kpcond ] ) @xmath111 so that we get @xmath112 .
furthermore , the constant @xmath92 appearing in eq.([kp ] ) contains precisely the lec combination that combines with that in the residue ( [ residue ] ) to give the same scaling law , now including the @xmath65 part .
thus , our final result for the pseudoscalar susceptibility in chpt to nlo ( finite and scale - independent ) is : @xmath113+{{\cal o}}(f^{-2 } ) = -\frac{{\langle \bar q q \rangle_t}^{chpt}}{m_q}+{{\cal o}}(f^{-2 } ) , \label{chipchpt}\end{aligned}\ ] ] where the first term inside brackets is the lo current - algebra @xmath114 and @xmath115 , @xmath116 are renormalized scale - independent lec @xcite .
therefore , we have obtained the wi connecting @xmath68 and @xmath1 to nlo in model - independent chpt , including finite-@xmath0 effects .
furthermore , the @xmath117 dependence cancels in @xmath118 , where only meson parameters show up . to this order ,
@xmath119 so that the lec dependence also disappears . recall that @xmath116 comes from a contact term in @xmath78 and is therefore another source of ambiguity in the nlo condensate @xcite .
note also that , unlike the approach followed in @xcite , we have arrived to the result ( [ chipchpt ] ) without relying on the validity of current algebra , which actually holds only in chpt to the lowest order .
actually , as explained , our result takes into account nlo corrections to @xmath120 through gor breaking terms , both at @xmath65 and @xmath121 , which turn out to be crucial to obtain the correct scaling law given in eq.([chipchpt ] ) .
we can draw some important conclusions from the previous results .
first , @xmath5 scales like the order parameter @xmath68 , instead of the much softer behaviour @xmath122 .
this scaling suggests a chiral - restoring nature for the pseudoscalar susceptibility , although we can not draw any definitive conclusion about chiral restoration just from our standard chpt analysis , which makes sense only at low @xmath0 .
its behaviour for higher @xmath0 approaching the transition should be considered merely as indicative extrapolations , pretty much in the same way as the chpt prediction for the vanishing point of the quark condensate is just a qualitative indication that the restoring behaviour goes in the right direction .
for this reason , in the following we will complement our standard chpt calculation with a direct lattice data analysis , and later on with a unitarized study which , as we will see , incorporates the relevant degrees of freedom to achieve a more precise description near the transition point .
one can actually observe a clear signal of a critical chiral - restoring behaviour for @xmath1 , consistent with our previous chpt result , in the lattice analysis of euclidean correlators , which determine their large - distance space - like screening mass @xmath123 in different channels @xcite . from eq.([chipdef ] ) we expect @xmath124 with @xmath125 the pole mass associated to @xmath126 in a general parametrization of the form @xmath127 with @xmath128 @xcite . here
, @xmath129 would correspond to the thermal euclidean propagator and @xmath130 to the retarded minkowski one setting the dispersion relation . assuming a soft temperature behaviour for @xmath131 , which is plausible below @xmath55 ( for instance , @xmath132 for the nlo chpt propagator in eq.([nloprop ] ) ) we
can then explain the sudden increase of @xmath133 observed in this channel @xcite since we expect that ratio to scale like @xmath134^{1/2}\sim \left[{\langle \bar q q \rangle}_0/{\langle \bar
q q \rangle_t}\right]^{1/2}$ ] .
we show in fig.[figlat ] ( left panel ) these two quantities .
the correlation between them is notorious , given the uncertainties involved , and the mentioned increase is clearly observed .
data are taken from the same lattice group and under the same lattice conditions @xcite .
note that in lattice works @xmath135 is subtracted from @xmath136 in order to avoid renormalization ambiguities . estimating the @xmath65 condensates from nlo chpt , , due to the vanishing of the lo , than in @xmath1 , @xmath68 . ]
this subtraction gives a 6% correction and , from the lattice values , it is about a 15% correction near @xmath55 .
apart from the screening versus pole mass and the strange condensate corrections , one should not forget about the typical lattice uncertainties , like resolution , choice of action , staggered taste breaking and large pion masses @xcite . , where @xmath137 with @xmath138 , for the lattice data in @xcite ( masses ) and @xcite ( condensate ) with the same lattice action and resolution and @xmath139 mev .
right : scalar susceptibility versus @xmath140 from the data in @xcite for which @xmath141 mev.,title="fig:",width=226 ] , where @xmath137 with @xmath138 , for the lattice data in @xcite ( masses ) and @xcite ( condensate ) with the same lattice action and resolution and @xmath139 mev .
right : scalar susceptibility versus @xmath140 from the data in @xcite for which @xmath141 mev.,title="fig:",width=226 ] another important conclusion of our analysis is that the decrease of @xmath1 and the increase of @xmath2 as they approach the critical point , lead to scalar - pseudoscalar susceptibility partner degeneration , which in an ideal @xmath3 pattern should take place near the @xmath2 peak . once again
, this behaviour is observed in lattice data . in fig.[figlat ] ( right panel ) we plot the subtracted condensate , expected to scale as @xmath142 according to our previous chpt and lattice analysis , versus @xmath143 , both from the lattice analysis in @xcite .
the @xmath65 values are taken from chpt .
the current degeneration is evident , not only at the critical point but also above it , where those two quantities remain very close to one another .
recall that both the analysis of the correlation between screening masses and inverse rooted condensate and that of scalar versus pseudoscalar ( condensate ) susceptibilities , although elaborated from available lattice data , has not been presented before , to the best of our knowledge .
as commented above , this analysis has been motivated by our chpt results in section ii and it gives strong support to the scalar - pseudoscalar degeneration pattern at the transition , as well as providing a natural explanation for the behaviour of lattice masses in this channel .
since the scalar susceptibility is dominated by the @xmath23 lightest state , which does not show up in the chpt expansion , let us consider its unitarized extension given by the inverse amplitude method ( iam ) @xcite which generates dynamically in @xmath25 the @xmath24 and @xmath145 resonances and has been extended to finite temperature in @xcite .
thus , before proceeding to derive the unitarized susceptibility in section [ sec : unisus ] , let us review briefly here , for the sake of completeness , some of the more relevant aspects of the thermal iam , particularly in the scalar channel .
we refer to @xcite for a more detailed analysis .
-dependent pion scattering.,width=302 ] the iam scattering amplitude is constructed by demanding unitarity and matching with the low - energy expansion , for which all the chpt scattering diagrams at finite temperature are included up to one - loop @xcite . the different types of those diagrams are represented in figure [ figfeyn ] .
the @xmath0-dependent corrections to the scattering amplitude comes from the internal loop matsubara sums in the imaginary - time formalism of thermal field theory .
the external pion lines correspond to asymptotic @xmath65 states .
the thermal amplitude is defined after the application of the @xmath65 lsz reduction formula , which allows to deal just with thermal green functions .
after the matsubara sums are evaluated , the external lines are analytically continued to real frequencies .
the full result for the thermal amplitude to nlo in chpt is given in @xcite .
the scattering amplitude can be projected into partial waves @xmath146 in the reference frame @xmath147 where the incoming pions 1,2 are at rest with the thermal bath , so that @xmath148 .
the nlo partial waves have the generic form ( we drop in the following the @xmath149 indices for brevity ) @xmath150 where @xmath151 is the @xmath152 tree - level @xmath0-independent scattering amplitude from @xmath75 and @xmath153 is @xmath154 including the tree level from @xmath78 plus the one - loop from the diagrams in fig.[figfeyn ] .
each partial wave satisfies @xmath155 for @xmath156 with @xmath157 $ ] and @xmath158^{-1}$ ] the bose - einstein distribution .
this is the perturbative version of the unitarity relation for partial waves @xmath159 and @xmath160 is the two - pion phase space , which at finite @xmath0 receives the thermal enhancement proportional to @xmath161 which has a neat interpretation in terms of emission and absorption scattering processes allowed in the thermal bath @xcite . precisely imposing that the partial waves satisfy the above unitarity relation exactly while matching the chpt series at low @xmath162 and low @xmath0 , leads to the thermal unitarized iam amplitude : @xmath163 when the iam amplitude is continued analytically to the @xmath162 complex plane @xcite
, it presents poles in the second riemann sheet @xmath164 $ ] with @xmath165 so that @xmath166 for @xmath156 .
those poles correspond to the physical resonances , which in the case of pion scattering are the @xmath24 ( @xmath23 ) and @xmath145 ( @xmath167 ) .
the @xmath0-dependent poles can be extracted numerically by searching for zeros of @xmath168 in the @xmath162 complex plane .
we denote the pole position by @xmath169 ^ 2 $ ] .
the lec for the iam are chosen so that , within errors , they remain compatible with the standard chpt ones and with the @xmath65 pole values for the @xmath20 and @xmath144 listed in the pdg @xcite .
let us comment now on the thermal evolution of the resonance poles , whose main features for this work are represented in figure [ figpoles ] . since @xmath170 with real @xmath92 and @xmath171 ,
the thermal dependence @xmath172 is governed by that of @xmath173 at the pole . ) and vector - isovector ( @xmath167 ) channels.,title="fig:",width=226 ] ) and vector - isovector ( @xmath167 ) channels.,title="fig:",width=226 ] ) and vector - isovector ( @xmath167 ) channels.,title="fig:",width=226 ] ) and vector - isovector ( @xmath167 ) channels.,title="fig:",width=226 ] in the vector - isovector channel , @xmath174 for all temperatures of interest here and therefore the @xmath20 can be considered a narrow breit - wigner ( bw ) resonance with @xmath175 and @xmath176 its mass and width respectively .
actually , @xmath177 decreases very slightly with @xmath0 for the relevant temperature range .
hence , the @xmath0-dependent contribution of the real part of @xmath173 is almost negligible compared to its @xmath65 part , due to the large @xmath20 mass value .
the latter gives roughly the @xmath65 rho mass , so the real part of the denominator of @xmath178 behaves dominantly as @xmath179 . in particular ,
the tadpole contributions of diagrams ( d ) , ( e ) in fig.[figfeyn ] are suppressed in this channel typically by @xmath180 .
however , the thermal effect in @xmath181 is much more sizable , increasing with @xmath0 .
its dominant contribution comes from the imaginary part of the amplitude .
the imaginary part of the @xmath178 denominator at the pole behaves like @xmath182 and , up to @xmath183 100 mev @xmath184 , which for @xmath185 near the real axis comes essentially from diagram ( a ) in fig.[figfeyn ] ( which would give the only imaginary part for real @xmath156 ) so the broadening can be explained just by thermal phase space increase up to that temperature . note that , although this effect is formally @xmath186 , it activates below the transition because of the relative small value of @xmath187 . above that
, there is an additional increase of the effective @xmath188 coupling with @xmath0 @xcite , to which tadpoles contribute , which explains a further increase of the width .
this behaviour is represented in figure [ figpoles ] .
observe the softer behaviour of the mass as compared to the width in this channel , and the correlation with the real and imaginary parts of the amplitude at the pole position . in the scalar - isoscalar channel ,
the one we are interested in here , the behavior is remarkably different .
we rather talk of a broad resonance pole , since @xmath175 and @xmath176 are comparable , so that the @xmath24 pole is away from the real axis .
as a consequence , all thermal contributions from diagrams ( a)-(e ) in fig.[figfeyn ] to @xmath173 become complex at the pole and the real and imaginary parts of the pole equations do not have the simple form of a bw resonance .
in addition , due to the lower value of @xmath189 as compared to the @xmath20 case , the thermal dependence is much more stronger , both for the real and imaginary parts , and all contributions from those diagrams become equally relevant . in particular , the tadpoles in diagrams ( d ) , ( e ) in fig.[figfeyn ] now come into play .
the numerical solution of the pole equations show that @xmath177 in this channel decreases significatively , while @xmath190 increases up to @xmath183 120 mev and decreases from that point onwards , as seen in fig.[figpoles ] .
note that this non - monotonic behaviour for @xmath176 can not be explained now just in terms of phase space or vertex increasing . on the other hand ,
a possible interpretation of the decreasing @xmath191 is a chiral restoring behaviour . actually , in fig.[figpoles ] ( a ) we also represent @xmath192 , which would correspond to the self - energy real part of a scalar particle with energy squared @xmath162 and @xmath193 , exchanged between the incoming and outgoing pions .
this @xmath18-like squared mass not only drops faster but it develops a minimum at a certain temperature , which as we will see below corresponds to a maximum in the scalar susceptibility . in the @xmath20 channel
, there is almost no numerical difference between @xmath191 and @xmath194 .
thus , a qualitative explanation for the @xmath190 change from a increasing to a decreasing behaviour in the scalar channel would be the influence of the strong mass decreasing of the decaying state . in order to establish a connection between the scalar susceptibility and the scalar pole , we construct a unitarized susceptibility by saturating the scalar propagator with the @xmath24 thermal state and assuming that its @xmath195 mass does not vary much with respect to the pole mass .
thus , we identify the pole of a scalar state exchanged in pion scattering with the thermal pole in the scalar channel discussed in the previous section .
therefore , we have : @xmath196 where we have normalized to the @xmath65 chpt value , since we are demanding that all our @xmath65 results match the model - independent chpt predictions .
this normalization compensates partly the difference between the @xmath195 and pole masses . under this approximation ,
the self - energy real part is the squared scalar mass @xmath197 , as discussed in the previous section , and the self - energy imaginary part vanishes at @xmath195 .
the quark condensate can not be extracted directly from the unitarized susceptibility .
however , we can obtain an approximate description by assuming that the relevant temperature and mass dependence , as far as the critical behaviour is concerned , comes from pion loop functions as @xmath198 and @xmath199 , with @xmath200 .
this @xmath201 dependence holds actually to nlo chpt , as in eq.([g1 ] ) .
then , from eq.([chisdef ] ) , since @xmath202 , we get @xmath203 with @xmath204 a suitable low-@xmath0 scale below which we use directly nlo chpt , which has a better analytic behaviour near @xmath65 .
the @xmath205 function is obtained from the @xmath0 dependence of @xmath2 in eq.([susunit ] ) .
our theoretical results based on effective theories are plotted in fig.[figsusc ] .
first , chpt to nlo gives an increasing @xmath4 , intersecting @xmath5 at @xmath206 , where @xmath207 .
once again , this result should be considered just as an extrapolation of the model - independent expressions for @xmath4 and @xmath5 beyond their low-@xmath0 applicability range . with this caution in mind ,
standard chpt supports the idea of partner degeneration . actually , near the chiral limit @xmath208 , where critical effects are meant to be enhanced , the degeneration point @xmath209 , approaching the chiral restoration temperature in that limit .
( right).,width=377 ] we also plot @xmath210 in fig.[figsusc ] .
the result agrees with standard chpt at low @xmath0 and improves remarkably the behaviour near the transition .
it actually develops a maximum at @xmath211 157 mev .
we show for comparison the lattice data of @xcite .
furthermore , approaching the chiral limit by taking the @xmath212 mev poles from @xcite gives a vanishing @xmath213 at @xmath211 118 mev and hence a divergent @xmath214 from eq.([susunit ] ) at @xmath55 .
thus , we get , at least qualitatively , the @xmath55 reduction and stronger @xmath2 growth near the chiral limit expected from theoretical @xcite and lattice @xcite analysis . the clear improvement of the unitarized approach with respect to the standard chpt one for the scalar susceptibility is essentially due to the introduction of the thermal @xmath24 state , whose importance in this case is clearly seen from the dependence @xmath215 , rather than to an enlargement of the applicability range in the unitarized approach .
actually , even within the unitarized description , we should be cautious when extrapolating it to near @xmath55 since we may be strictly beyond the effective theory range .
the resulting @xmath216 is plotted in fig.[figsusc ] with @xmath217 12 mev . between 10 - 60 mev . ]
the critical behaviour is again nicely improved compared to the chpt curves and is in better agreement with lattice data in that region .
in addition , we obtain once more a scalar - pseudoscalar intersection near the @xmath2 peak and hence of chiral restoration .
the corresponding @xmath216 near the chiral limit ( @xmath212 mev ) is much more abrupt , vanishing and meeting @xmath2 at @xmath55 as expected .
recall that in the above unitarized analysis , we are not performing a fit to lattice points .
we just use the same lec which generate the @xmath65 physical @xmath24 , @xmath20 states @xcite and then provide our results for the susceptibility and condensate .
the theoretical uncertainties in the lec , as well as the lattice errors already discussed should be taken into account for a more precise comparison .
besides , our effective theory analysis is not expected to reproduce the chiral restoration pattern above @xmath55 . in any case , apart from the important consistency obtained for chiral restoration properties , such as the @xmath2 peak and the @xmath218 matching , an important point we want to stress is the crucial role of the @xmath219 state to describe the scalar susceptibility . since @xmath215 , this observable is much more sensitive to this broad state , so that a physically realistic description , including its thermal effects , turns out to be essential . this may not be the case for other thermodynamical observables , for which other approaches may provide a much better description .
for instance , the hadron resonance gas ( hrg ) framework gives very accurate results compared to lattice data @xcite and particle distributions @xcite below the transition , by including all known hadronic states as free particles in the partition function , often including small interaction corrections . within the hrg approach ,
hadron interactions are generically encoded in the resonant states .
this framework works for most thermodynamic quantities , which are obtained , by construction , as monotonic functions of @xmath0 , and generally increase with the mass of the states considered .
for instance , the scalar susceptibility within that approach would increase with @xmath0 , as it happens for other quantities such as the trace anomaly .
thus , the effect of a properly included broad @xmath18 @xmath0-dependent state arising from pion scattering in order to describe chiral restoring properties such as the susceptibility peak , is once more highlighted .
in fact , the @xmath18 state is just not included in many hrg works @xcite or , at most , considered as a bw state @xcite with its @xmath65 mass and width , which , as we have commented above , does not provide an entirely adequate description .
finally , let us comment that apart from higher mass states , inclusion of higher order interactions may also be important for certain hadronic observables .
for instance , including @xmath220 interactions in the vector channel are essential to describe properly the dilepton spectra @xcite .
summarizing , we have analyzed the scalar - pseudoscalar @xmath3-like current degeneration pattern at chiral symmetry restoration , from lattice simulations and effective theory analysis .
the pseudoscalar susceptibility scales as the quark condensate , which we have explicitly shown in chpt to nlo at low temperature , and becomes degenerate with its chiral partner " scalar susceptibility close to the scalar transition peak .
the lattice data and the unitarized chpt analysis support this picture and are well accounted for by the dominant physical states : pions and the @xmath24 scalar resonance generated in pion scattering at finite temperature . in turn , we have provided a natural explanation for the sudden growth of lattice masses observed in the pseudoscalar channel . although we have restricted here , for simplicity , to the @xmath16 case , the analysis can be extended to @xmath221 , where the role of other scalar states such as the @xmath222 can also be studied @xcite .
useful comments from f. karsch , s. mukherjee and d. cabrera are acknowledged .
work partially supported by the eu fp7 hadronphysics3 project , the spanish project fpa2011 - 27853-c02 - 02 and fpi programme ( bes-2009 - 013672 , r.t.a ) , and by the german dfg ( sfb / tr 16 , j.r.e . ) .
r. rapp and j. wambach , adv .
phys . * 25 * , 1 ( 2000 ) .
j. beringer _
[ particle data group collaboration ] , phys .
d * 86 * , 010001 ( 2012 ) and references therein .
e. v. shuryak , phys .
rept . * 115 * , 151 ( 1984 ) ; phys .
rept . * 264 * , 357 ( 1996 ) .
f. karsch , k. redlich and a. tawfik , eur .
j. c * 29 * , 549 ( 2003 ) .
p. huovinen and p. petreczky , nucl .
a * 837 * , 26 ( 2010 ) .
a. andronic , p. braun - munzinger and j. stachel , phys .
b * 673 * , 142 ( 2009 ) [ erratum - ibid .
b * 678 * , 516 ( 2009 ) ] .
a. andronic , p. braun - munzinger , j. stachel and m. winn , phys .
b * 718 * , 80 ( 2012 ) . | we describe scalar - pseudoscalar partner degeneration at the qcd chiral transition in terms of the dominant low - energy physical states for the light quark sector .
first , we obtain within model - independent one - loop chiral perturbation theory ( chpt ) that the qcd pseudoscalar susceptibility is proportional to the quark condensate at low @xmath0 .
next , we show that this chiral - restoring behaviour for @xmath1 is compatible with recent lattice results for screening masses and gives rise to degeneration between the scalar and pseudoscalar susceptibilities ( @xmath2 , @xmath1 ) around the transition point , consistently with an @xmath3-like current restoration pattern .
this scenario is clearly confirmed by lattice data when we compare @xmath4 with the quark condensate , expected to scale as @xmath5 .
finally , we show that saturating @xmath2 with the @xmath6 broad resonance observed in pion scattering and including its finite temperature dependence , allows to describe the peak structure of @xmath4 in lattice data and the associated critical temperature .
this is carried out within a unitarized chpt scheme which generates the resonant state dynamically and is also consistent with partner degeneration . |
we propose a mean field theory to describe nuclei by treating the pion explicitly . in the non - relativistic framework
, the pion - exchange interaction appears mainly as the tensor force , which is much stronger than the central force . to make the important role of the pion in nuclei clear
, we briefly discuss the role of the tensor force in the reaction matrix theory , showing the density dependence of the central component of the @xmath0-matrix , together with some discussion on the results of the variational calculations for light nuclei based on the realistic nuclear force .
the importance of the pion demonstrated through those studies , leads us to develop a theoretical framework based on the mean field theory which enables us to treat the pion - exchange interaction on the same footing as other meson - exchange interactions .
a new mean field theory has been formulated by breaking the charge and parity symmetries of the single - particle orbits and further by recovering both the symmetries of the total wave function .
thus we call it the charge- and parity - projected hartree - fock ( cpphf ) scheme , with which we study the correlation caused by the pion ( or tensor force ) .
we discuss the cpphf scheme in the latter half of this paper .
the reaction matrix theory initiated by brckner has been developed with the intention of understanding the nuclear structure through the multiple scattering correlation based on the two - particle scattering in nuclei starting from the realistic nuclear force .
the most basic element in the theory is the reaction matrix ( @xmath0-matrix ) .
the introduction of the @xmath0-matrix is necessitated by the singular nature of the two - body forces in the short range region , that is , the repulsive core and the strong non - central components .
-matrix for two particles in a nucleus , which includes the effects of the hard core and the tensor force.,width=113 ] the elementary process of the multiple scattering is the two - nucleon scattering within a nucleus which is described by `` the reaction matrix '' given by @xmath1 with @xmath2 , where @xmath3 , @xmath4 , @xmath5 , and @xmath6 express the two - boy nuclear potential , the kinetic energy operator , the starting energy of the scattering two nucleons , and the pauli exclusion operator , respectively .
the reaction matrix @xmath0 is a sort of `` effective interaction '' in nuclei in contrast to the original nuclear force @xmath3 .
because the two - body scattering correlation is known to be dominant compared with higher multiple scattering correlations , the reaction matrix @xmath0 can be a good basic element for investigating the nuclear structure with the standard models ( the shell model and the mean field model ) .
according to the definition of @xmath0 , the repulsive core of @xmath3 appears simply in @xmath0 as a fairly weakened core .
however , the tensor force causes the significant modifications in the characteristics of the `` effective interaction @xmath0 , '' because the effect of the tensor force arises first as the second order effect @xmath7 and produces a very large contribution to the central component of @xmath0 .
the many - body effects included in the @xmath0-matrix arises only through the green function @xmath8 .
these are the pauli effect due to @xmath6 and the dispersion effect due to @xmath9 , which bring about various kinds of the structure dependence reflecting the characteristics of the structure assumed initially .
thus the contribution of the tensor force to the central component of @xmath0 is very sensitive to the nuclear structure itself and to the circumstance of the surrounding nuclear medium @xcite .
we discuss here the density dependence of @xmath0 as a typical example of the structure dependence . }
$ ] ( solid curves ) and @xmath10}$ ] ( dashed curves ) for @xmath11=1.4 and 0.8 @xmath12 .
the difference @xmath13}-r^2 \tilde{g}_\mathrm{c}^{[^3\mathrm{e}]})$ ] corresponds to @xmath14.[fig3],width=283 ] } $ ] ( solid curves ) and @xmath10}$ ] ( dashed curves ) for @xmath11=1.4 and 0.8 @xmath12 .
the difference @xmath13}-r^2 \tilde{g}_\mathrm{c}^{[^3\mathrm{e}]})$ ] corresponds to @xmath14.[fig3],width=283 ] we show the central component of @xmath0 for the triplet - even state , @xmath15}$ ] with the two - body interaction @xmath3 in which the tensor force @xmath16 is included , and @xmath17}$ ] with @xmath18 in which tensor force is not included , which are calculated by y. akaishi .
figure [ fig2 ] clearly shows the strong @xmath11 dependence of @xmath15}$ ] . on the contrary , there is little difference between two cases of @xmath17}$ ] for @xmath11=1.4 and 0.8 @xmath12 as shown in fig .
[ fig3 ] by the solid curves .
it is easily understood from fig .
[ fig3 ] that the density ( @xmath11 ) dependence of @xmath15}$ ] is almost entirely caused by the tensor force and that the tensor contribution @xmath7 is estimated by the difference @xmath19}-\tilde{g}_\mathrm{c}^{[^3\mathrm{e}]})$ ] .
figure [ fig3 ] shows that the ratio @xmath19}-\tilde{g}_\mathrm{c}^{[^3\mathrm{e}]})/ \tilde{g}_\mathrm{c}^{[^3\mathrm{e}]}$ ] = 1.5 @xmath20 2.0 at @xmath21=1.0 fm .
these features tell us that the tensor force plays the very important role on the nuclear structures .
recent variational calculations of light nuclei based on the realistic nuclear force in the real space are carried out by the argonne - illinois group up to a=10 @xcite .
first of all we have been strongly impressed by their successful reproduction of the binding energy of the ground and excited states for all these nuclei .
secondly , we are surprised by their results that the contribution of the one - pion - exchange interaction @xmath22 amounts to 70 - 80% of the strength of the two - body interaction part in the 3@xmath23a@xmath238 systems , and the pion is also important for the three - body interaction originating mostly from the pion exchange .
thirdly , we notice that the two - alpha cluster structure is found in @xmath24be : according to their expression @xcite , `` these results obtained from the vmc wave functions , suggest that the @xmath25 , @xmath26 , @xmath27 wave functions for @xmath28be have the structure of a deformed rotor consisting of two @xmath29 s . '' with regard to the importance of the pion , we note here the pioneering work of the variational calculation of the @xmath29 particle by atms @xcite .
table [ table1 ] exhibits the contributions of various interactions to the binding energy ( in the unit of mev ) , from which we can see almost a half of the attraction is due to the tensor force and hence the pion - exchange interaction is responsible in alpha particle .
.the breakdown of energies for the @xmath29 particle in the atms calculations @xcite.[table1 ] [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] we solve the cpphf equation for the @xmath29 particle , which is the most simple doubly - closed - shell nucleus . here , the calculated results for the ground ( @xmath25 ) state are exhibited in table [ table2 ] for the purpose to examine the responsibility of the cpphf scheme to describe the tensor correlation .
we compare them with the results for the cases of the hartree - fock ( hf ) scheme with the non - projected variational wave function and for the case of the parity - projected hartree - fock ( pphf ) scheme with the parity - projected variational wave function .
although we make the tensor force 1.5 times stronger than the usual one , we can not obtain the correlation energy due to the tensor force by the hf scheme but a fairly large correlation energy is obtained by the pphf scheme . finally the cpphf scheme produces a large amount of the correlation energy , which is almost three times larger than that for the pphf case .
( see the detail results and discussions in ref . @xcite . )
the studies in the model space based on the @xmath0-matrix theory and in the realistic functional space based on the realistic nuclear force , exhibit clearly the importance of the tensor force ( the one - pion - exchange force ) .
a new framework is introduced to treat the pionic correlation in the extended model space by introducing the single - particle orbits with the parity and charge mixing and by recovering the parity and charge symmetries by the projection .
we consider that the new theoretical framework will provide a powerful method to investigate the mechanism of forming diverse structures of nuclei ( shell structure , cluster structure , halo structure and others ) , which are explored by the experiments .
we acknowledge fruitful discussions with prof .
y. akaishi and dr .
y. ogawa on the role of the pion in light nuclei . | we give a brief review on the important role of the pion - exchange interaction characterized by the strong tensor force . in order to illustrate the role of the pion ,
we have proposed a new mean field theory with the charge and parity mixing , by which the pion is treated explicitly in the mean field framework .
we introduce the essence of the new mean field theory and illustrate the calculated results of the pion ( tensor force ) correlation by taking the alpha particle as an example . |
the nature of anomalous x - ray pulsars ( axps ) was a mystery for the past 20 years , since the discovery of the first example ( 1e 2259.1 + 586 ; * ? ? ? * ) .
there are currently only five confirmed axps , all of which are located in the galactic plane , with two at the geometric centers of supernova remnants @xcite .
axp characteristics can be summarized as follows ( see * ? ? ? * for a review ) : they exhibit x - ray pulsations in the range @xmath3612 s ; they have pulsed x - ray luminosities in the range @xmath4 erg s@xmath5 ; they spin down regularly ; they have spectra that are characterized by thermal emission of @xmath6 kev plus a hard power - law tail . they have faint or no optical / ir counterparts @xcite .
the optical counterpart of one axp , 4u 0142 + 61 is reportedly pulsed @xcite .
axps are called `` anomalous '' because it has been unclear what powers their radiation .
they are not rotation - powered as their observed x - ray luminosities are much greater than the rate of loss of rotational kinetic energy inferred from their spin - down .
they were long thought to be accreting from a low mass companion ( e.g. * ? ? ?
however this model is difficult to reconcile with observations : the absence of doppler shifts even on short time scales ( e.g. * ? ? ?
* ) , the absence of a detectable optical / ir companion or bright accretion disk ( see * ? ? ?
* ; * ? ? ?
* ) , the apparent associations with supernova remnants , that axp spectra are very different from those of known accreting sources , and that @xmath7 is generally smaller than in known accreting sources , all are inconsistent with this scenario .
@xcite ( see also * ? ? ?
* ; * ? ? ?
* ) proposed that axps are accreting from a fossil disk made of fall - back material from the supernova . however , this model overpredicts the observed optical / ir flux from the disk @xcite .
@xcite suggested that the main source of free energy for axp emission is from the magnetic field itself . in this model ,
axps are young , isolated , highly magnetized neutron stars or `` magnetars '' .
prior to our _ rxte _ monitoring campaign , the primary evidence in favor of this model for axps was the inferred strength of the surface dipolar magnetic field required to slow the pulsar down _ in vacuo _ , assuming magnetic dipole braking as in radio pulsars .
these fields are in the range @xmath8 g. the spin - down ages in this model , inferred assuming a small birth spin period , are in the range @xmath28200 kyr . this suggested youth is further evidenced by the associations with supernova remnants , and from the location of axps in the galactic plane .
the identification of axps with magnetars was further strongly motivated by the similarity of axp emission to that of the soft gamma - ray repeaters ( sgrs ) in quiescence .
specifically , the latter have similar pulse periods , are spinning down @xcite , and have x - ray spectra that are comparable to , though somewhat harder than , those of the axps , at least when not in outburst @xcite .
independent evidence for the ultra - high magnetic fields exists in sgrs ; for example , a @xmath9 g magnetic field is required to confine the radiation that is seen following major outbursts @xcite .
indeed the only strong distinction between the emission from sgrs and axps appeared to be the fact that sgrs exhibit occasional and repeating short bursts , while axps did not .
the results presented here were obtained using the proportional counter array ( pca ; * ? ? ?
* ) on board nasa s _ rossi x - ray timing explorer _
( _ rxte _ ) .
the pca consists of an array of five collimated xenon / methane multi - anode proportional counter units operating in the 2 60 kev range , with a total effective area of approximately @xmath10 and a field of view of @xmath11 fwhm . a program to monitor axps regularly using _ rxte _ was begun in 1996 in order to determine their long - term timing , pulsed flux , and pulse profile stabilities @xcite . as part of this program , motivated by the existence of sgr bursts , we also searched the axp data for bursts @xcite .
our observations consist primarily of short snapshots taken on a monthly basis .
in addition , we used a handful of archival observations ; the exposures in these observations vary .
we used the ` goodxenonwithpropane ` data mode , which records photon arrival times with 1-@xmath12s resolution and bins energies into one of 256 channels .
below , we summarize the results of this monitoring campaign thus far ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for details ) .
and @xmath0 measured from a phase - connected interval in 1999 .
panel b shows the difference between the ephemeris indicated by the dotted line and the data points .
[ fig : freq1048],scaledwidth=95.0% ] and @xmath0 measured from a phase - connected interval in 1999 .
panel b shows the difference between the ephemeris indicated by the dotted line and the data points .
[ fig : freq1048],scaledwidth=85.0% ] in the timing analysis , each binned time series was epoch - folded using the best estimate frequency determined initially from either a periodogram or fourier transform ( though later folding was done using the timing ephemeris determined by maintaining phase coherence ; see below ) .
resulting pulse profiles were cross - correlated in the fourier domain with a high signal - to - noise template created by adding phase - aligned profiles from previous observations .
the cross - correlation returns an average pulse time - of - arrival ( toa ) for each observation corresponding to a fixed pulse phase .
the pulse phase @xmath13 at any time @xmath14 can be expressed as a taylor expansion , @xmath15 where @xmath16 is the pulse frequency , @xmath17 , etc . and
subscript ` @xmath18 ' denotes a parameter evaluated at the reference epoch @xmath19 .
the toas were fit to the above polynomial using the pulsar timing software package ` tempo ` .
unambiguous pulse numbering is made possible by obtaining monitoring observations spaced so that the best - fit model parameters have a small enough uncertainty to allow prediction of the phase of the next observation to within @xmath20 .
typically this requires two closely spaced observations ( within a few of hours of each other ) followed by one spaced a few days later , and regular monitoring thereafter , as long as phase coherence can be maintained . using phase coherent timing
we have shown that axps can be phase connected over long intervals , implying they can spin down with impressive stability , which is at odds with what is seen in most x - ray pulsars that definitely have accretion disks ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* see figure [ fig:0142res ] ) .
however this is not true of all our sources .
1e 1048.15937 shows large deviations from a simple spin down ( figure [ fig : freq1048 ] , * ? ? ?
however , the very noisy timing behavior seen in 1e 1048.15937 shows no evidence for correlated flux variations , as are expected in fall - back disk models .
motivated by a previous report of a significant pulse morphology change in an axp ( 1e 2259.1 + 586 ; * ? ? ?
* ) we used our _ rxte _ data to search for more such events .
morphology changes are expected in the magnetar model during outbursts or torque variations . to search for pulse profile changes ,
the profiles were first phase aligned using the templates and the same cross - correlation procedure used for timing .
each data profile was fit to a high signal - to - noise template by adjusting the amplitude and the offset in order to minimize a @xmath21 statistic .
the resulting data profile was subtracted from the template to yield `` profile residuals '' for each observation .
the procedure was repeated for each pulsar with different binning in order to have sensitivity to a variety of types of pulse profile changes .
we have not detected any large pulse profile variations , except for 1e 2259.1 + 586 during a major outburst ( see below ) .
this justifies our other analysis procedures which assume a fixed profile .
however , in a handful of observations , we have found low - level pulse profile changes , at the @xmath22 level for all sources .
it is difficult to set quantitative upper limits on the amplitude of pulse profile changes to which we were sensitive , as these depend on the shape of the change , and vary depending on the length of the observation .
typically , rms profile residuals are @xmath220% of the pulse peak , although this varied from 440% .
we have also used our _ rxte _
observations to monitor the pulsed flux of each axp . by considering the stability of pulsed fluxes ,
we test the two competing axp models . in accreting systems ,
torque fluctuations are generally accompanied by variation in the mass accretion rate and hence luminosity , and vice versa . in the magnetar model
, abrupt flux changes could occur in analogy with sgr - like outbursts .
given the large field - of - view of the pca , the low count rates for the sources relative to the background , and the fact that , for example , 1e 2259.1 + 586 is in a supernova remnant , total flux measurements are difficult with our _ rxte _ data .
instead , we have determined the pulsed component of the flux , by using the off - pulse emission as a background estimator .
data from each observing epoch were folded at the expected pulse period as was done for the timing analysis .
however , for the flux analysis , 16 phase bins were used across the pulse . for each phase bin
, we maintained a spectral resolution of 128 bins over the pca range .
given the broad morphologies of the average pulse profile , only one phase bin could be used as a background estimator .
the pulse profiles were phase aligned , so that the same off - pulse bin was used for background in every case .
the remaining phase bins were summed , and their spectral bins regrouped using the ` ftool ` ` grppha ` .
the regrouped , phase - summed data sets , along with the background measurement , were used as input to the x - ray spectral fitting software package ` xspec ` .
response matrices were created using the ` ftool`s ` xtefilt ` and ` pcarsp ` . because of the limited statistics , fitting a two - component model was not practical , so we used a simple photoelectrically absorbed power law , at first holding only @xmath23 fixed .
for all sources we found that the photon index @xmath24 was constant within the uncertainties ; we therefore held it fixed at its mean value . to extract a pulsed flux at each observing epoch we refit each spectrum by varying only the normalization .
other than after a major outburst from 1e 2259.1 + 586 ( see below ) , we found no evidence for any large variability in the pulsed flux within 20%30% consistent with the systematic errors .
axp spectra are generally best fit by a two - component model consisting of a photoelectrically absorbed blackbody with a hard power - law tail ( israel et al .
whether these two components are physically distinct is an open question @xcite . to investigate this , we compared the pulse profile morphology of the axps in two energy bands .
thus far , we have found that pulse profile energy - dependence is generally subtle , but varies from source to source .
we found no clear patterns when considering average pulse profiles as a function of energy @xcite . in the magnetar model for sgrs @xcite , bursts are a result of sudden crustal yields due to stress from the outward diffusion of the huge internal magnetic field .
such yields cause crust shears which twist the external magnetic field , releasing energy .
@xcite who , upon suggesting that axps are also magnetars , predicted x - ray bursts should eventually be seen from them . in order to confirm this prediction we searched our _ rxte _ data for bursts
our burst searching algorithm is summarized below ( see * ? ? ? * for details ) .
time series were created with 31.25-ms resolution from photons having energies in the range 220 kev for each pca proportional counter unit ( pcu ) separately , using all xenon layers .
photon arrival times at each epoch were adjusted to the solar system barycentre .
the resulting time series were searched for significant excursions from the mean count rate by comparing each time bin value with a windowed 7-s running mean .
bursts were identified assuming poissonian statistics , and by combining probabilities from the separate pcus .
we discovered two short , highly significant x - ray bursts , separated by 16 days , from the direction of the axp 1e 1048.15937 in data from fall 2001 @xcite .
the first occurred during a 3-ks pca observation obtained on 2001 october 29 , and the second during a 3-ks observation obtained on 2001 november 14 .
we searched a total of 442 ks of pca observing time from 19962002 , and no other significant bursts were found toward 1e 1048@xmath255937 .
the burst profiles are shown in figure [ fig : profiles ] . ,
scaledwidth=85.0% ] , scaledwidth=85.0% ] we fit the spectra from the first 1 s of each burst to two one - component models , a power law and a black body . continuum models provided an adequate characterization of the burst 2 spectrum but not of the burst 1 spectrum .
as seen in figure [ fig : spec1048 ] , the spectrum for the 1 s after the burst 1 onset exhibits a feature near 14 kev .
this feature is clear in all binning schemes and is prominent throughout the first @xmath21 s of the burst .
no known pca instrumental effect produces a feature at this energy ( k. jahoda , personal communication ) .
intriguingly , in some spectral binning schemes , we saw hints of features in the spectrum consistent with additional multiples of 7 kev , suggesting cyclotron lines .
however these were not statistically significant .
due to the wide ( @xmath21@xmath26 ) field - of - view ( fov ) and lack of imaging capabilities of the pca , we can not verify that the bursts originated from the location of the axp .
the low peak x - ray fluxes of the events ( see table 1 ) preclude determining the source s location using data from other , better imaging instruments that were contemporaneously observing the x - ray sky , such as the _ rxte _ all sky monitor , or the wide field camera aboard _
bepposax_. we must therefore consider other possible origins from the bursts before concluding they were from the axp .
the bursts short rise times ( table 1 ) require emission regions of less than a few thousand km , implying a compact object origin .
type i bursts from an lmxb in the same fov as 1e 1048.15937 are unlikely to explain our observed bursts because ( i ) the burst rise times are much shorter than those of type i bursts ; ( ii ) the burst spectra are much harder than those of type i bursts ; ( iii ) burst 2 shows no evidence for spectral softening with time and no type i burst has ever exhibited a spectral feature like the one detected in burst 1 ; ( iv ) the bursts are extremely faint , implying a source location well outside the milky way for type i burst luminosities ( v ) there are no known lmxbs in the fov @xcite .
furthermore , the bursts we have observed are unlikely to be type ii bursts from an unknown x - ray binary in the pca fov because ( i ) of the rarity of such events ( observed thus far only in two sources , both accreting binaries ) ; ( ii ) type ii bursts have longer rise times than do our bursts ; ( iii ) no type ii burst has exhibited a spectral feature like that seen in burst 1 .
classical gamma - ray bursts ( grbs ) sometimes exhibit prompt x - ray emission that can have temporal and spectral signatures similar to those we have observed@xcite .
however , the likelihood of two grbs occurring within @xmath27 of each other is small , and grbs are not known to repeat .
the observed burst properties are in many ways similar to those seen from sgrs @xcite . the fast rise and
slow decay profiles are consistent with sgr time histories , as are the burst durations . both axp and sgr bursts are spectrally much harder than is their quiescent pulsed emission .
the burst peak fluxes and fluences fall within the range seen for sgrs , and the spectrum of burst 2 is consistent with sgr burst spectra of comparable fluence .
burst 1 has characteristics unlike nearly all sgr bursts , specifically its long tail ( table 1 ) and spectral feature .
however , we note that one event from sgr 1900 + 14 was shown @xcite to possess each of these properties .
the marginal increase in the pulsed flux that we observed at the burst epochs is consistent with sgr pulsed flux increases seen during bursting episodes @xcite . finally , the fact that in spite of several years of monitoring , the only two bursts detected occurred within two weeks of each other suggests episodic bursting activity , the hallmark of sgrs . thus , the characteristics of these events match the burst properties of sgrs far better than any other known burst phenomenon . the large 14-kev spectral line in the first burst in intriguing .
an electron cyclotron feature at this energy @xmath28 implies a magnetic field of @xmath29 g ( where @xmath30 is the electron mass , @xmath31 is the speed of light , @xmath32 is planck s constant , and @xmath33 is the electron charge ) , while a proton cyclotron feature implies @xmath34 g. the former is significantly lower than is implied from the source s spin - down and is typical of conventional young neutron stars , rather than magnetars .
the latter is higher than is implied by the spin - down yet reasonable for the magnetar model as the spin - down torque is sensitive only to the dipolar component of the magnetic field . [
cols="<,^,^ " , ] .
monte carlo simulations in xspec were done to verify this conclusion .
[ fig : spec1048],scaledwidth=95.0% ] on 2002 june 18 , fortuitously during one of our _ rxte _ monitoring observations of 1e 2259.1 + 586 , over 80 bursts were detected in the span of 14 ks @xcite .
figure [ fig : b2259 ] shows the 0.125 s light curve . 1e 2259.1 + 586 , in contrast to 1e 1048.15937 , had been an extremely stable rotator and had shown no signs of `` volatility . ''
the association of the bursts toward 1e 1048.15937 with the axp could not be unambiguously verified , however , the bursts seen here ( figure [ fig : b2259 ] ; top panel ) clearly originated from 1e 2259.1 + 586 . during and following the outburst ,
the pulsed flux increased ( see figure [ fig : b2259 ] ) and there was a significant pulsed morphology change , such that the relative amplitude of the leading and trailing pulse were swapped ( see figure [ fig : pulse2259 ] ) .
target - of - opportunity observations were made by many major observatories in addition to _ rxte _ at the time of this outburst .
no radio emission was detected in spite of sensitive _ vla _ observations , nor were any bursts seen in too _ xmm - newton _ data obtained roughly a week afterward .
near - infrared observations obtained with the gemini north 8-m reflector on june 21 showed that the proposed ir counterpart had brightened by a factor of 3.4 ( kaspi et al .
a few days later , it faded back to its pre - burst value .
this represents the first detection of transient optical / ir emission from an isolated neutron star .
subsequently , @xcite reported possible ir variability from the proposed counterpart to 1e 1048.1@xmath255937 , which they suggested might be related to its bursting behaviour .
the origin of ir emission is currently not addressed in the magnetar model .
pulse morphology changes can occur , according to the magnetar model , during outbursts as in the sgrs , as this is when magnetic reconfigurations occur .
indeed during the outburst of 1e 2259.1 + 586 , significant pulse morphology changes were seen , such that the relative amplitude of the double peaked profile was swapped ( see figure [ fig : pulse2259 ] ) . @xcite
also reported a significant change in the pulse morphology of 1e 2259.1 + 586 in _ ginga _ observations obtained in 1990 , such that the leading pulse had amplitude roughly half of the trailing pulse . in both instances the pulse profile returned back to its long - term state .
, scaledwidth=85.0%,height=288 ] , scaledwidth=65.0% ] pulsed flux enhancements in both the pulsed and persistent flux are observed in the sgrs during bursting episodes and are thought to be a result of back heating of the surface and outer crust by the burst emission in the magnetosphere . the pulsed flux of 1e 2259.1 + 586 increased by a factor of @xmath35 during its bursting episode ( see figure [ fig : b2259 ] ) .
we noted a marginal ( @xmath22 ) increase in the pulsed flux from 1e 1048.1@xmath255937 that commenced with the observation in which burst 1 was detected , and which lasted @xmath24 weeks @xcite .
there have been past claims of x - ray flux changes , by up to a factor of @xmath210 in two axps ( 1e 2259.1 + 586 and 1e 1048.15937 ; * ? ? ?
* ) , although the reported fluxes were measured with different instruments having very different spectral responses and angular resolutions . if correct , these past claims of increased flux might imply previous bursting activity .
we have presented a variety of observational results for all axps using regular monitoring observations and archival observations from the pca aboard _
rxte_. there is now strong evidence that axps are magnetars .
the great rotationally stability in some axps is at odds with an accretion scenario , as is the absence of pulsed flux or pulse morphology changes correlated with torque .
extended periods of spin - up would only , though not necessarily , be expected in an accretion scenario , however all axps have exhibited relatively steady spin - down .
furthermore , and perhaps most compellingly , accretion models can not account for the bursting episodes seen in two axps .
interestingly , there is a correlation between timing stability and @xmath0 in our data .
the sources with by far the smallest @xmath0 s , are the most stable in quiescence ( at least during our observations ) , while those with large @xmath0 s are less so .
the sgrs , being even less stable @xcite , have even larger @xmath0 s , in agreement with this trend .
if correct , the trend suggests a continuum of timing properties between the axp and sgr populations , lending additional support to the connection between them .
the identification with magnetars is further strongly motivated by the similarity of the axp emission to that of the sgrs .
specifically , the latter have similar pulse periods , are spinning down @xcite , and have x - ray spectra that are comparable to , though somewhat harder than , those of the axps , at least when not in outburst . with the detection of x - ray bursts from two axps the two classes are now similar in that respect as well .
these very recent results appear to have sealed the connection between sgrs and axps , as was originally predicted by @xcite .
however , the identification of the two classes objects as magnetars has left some open questions .
for instance , what if anything distinguishes axps and sgrs ? are axps the progenitors of sgrs or vice versa ?
recently @xcite have shown that the x - ray spectrum of the quiescent counterpart to sgr 052666 has a power - law index @xmath36 , similar to those seen in axps , but softer than those of sgrs .
as sgr 052666 has been inactive for almost two decades , this suggests sgrs and axps are one and the same object , and may have even cycle back and forth between each flavor .
another noteworthy question is what is the magnetar / radio pulsar connection ?
recent observations of radio pulsars having apparent surface dipolar magnetic fields comparable @xcite , and in one very new case ( mclaughlin et al . ,
private communication ) , surpassing those of some axps , leaves open the question of what physically distinguishes a magnetar from a rotation - powered pulsar of high magnetic field ?
@xcite suggested that the sources classified as magnetars might have much higher fields , in the form of multipole moments . perhaps comparing the thermal x - ray emission from radio pulsars having a variety of magnetic fields may illuminate the situation .
this work was supported by nasa , nateq and ciar .
vmk is a canada research chair and a fellow of the ciar .
this research has made use of data obtained through the high energy astrophysics science archive research center online service , provided by the nasa / goddard space flight center .
camilo , f. and kaspi , v. m. and lyne , a. g. and manchester , r. n. and bell , j. f. and damico , n. and mckay , n. p. f. and crawford , f. , discovery of two high magnetic field radio pulsars , * 541 * , 367 - 373 , 2000 .
jahoda , k. , swank , j. h. , giles , a. b. , stark , m. j. , strohmayer , t. , zhang , w. , and morgan , e. h. , in - orbit performance and calibration of the _ rossi x - ray timing explorer _
( _ rxte _ ) proportional counter array ( pca ) . , * 2808 * , 59 , 1996 .
kaplan , d. l. , kulkarni , s. r. , van kerkwijk , m. h. , rothschild , r. e. , lingenfelter , r. l. , marsden , d. , danner , r. , and murakami , t. , _ hst _ observations of sgr 0526 - 66 : new constraints on accretion and magnetar models . .
in press , 2001 .
kouveliotou , c. , dieters , s. , strohmayer , t. , van paradijs , j. , fishman , g. j. , meegan , c. a. , hurley , k. , kommers , j. , smith , i. , frail , d. , and murakami , t. , an x - ray pulsar with a superstrong magnetic field in the soft @xmath37-ray repeater sgr 180620 . , * 393 * , 235237 , 1998 . , c. , strohmayer , t. , hurley , k. , van paradijs , j. , finger , m. h. , dieters , s. , woods , p. , thompson , c. , and duncan , r. c. , discovery of a magnetar associated with the soft gamma repeater sgr 1900 + 14 . , * 510 * , l115l118 , 1999 .
thompson , c. and lyutikov , m. and kulkarni , s. r. , electrodynamics of magnetars : implications for the persistent x - ray emission and spin - down of the soft gamma repeaters and anomalous x - ray pulsars . , * 574 * , 332355 , 2002 . ,
p. m. and kouveliotou , c. and g " o " u , e. and finger , m. h. and swank , j. and markwardt , c. b. and hurley , k. and van der klis , m. , large torque variations in two soft gamma repeaters . , * 576 * , 381390 , 2002 | we report on long - term monitoring of anomalous x - ray pulsars ( axps ) using the _ rossi x - ray timing explorer _ ( _ rxte _ ) . using phase - coherent timing
, we find a wide variety of behaviors among the sources , ranging from high stability ( in 1e 2259.1 + 586 in quiescence and 4u 0142 + 61 ) , to instabilities so severe that phase - coherent timing is not possible ( in 1e 1048.15937 ) .
we note a correlation in which timing stability in axps decreases with increasing @xmath0 .
the timing stability of soft gamma repeaters ( sgrs ) in quiescence is consistent with this trend , which is similar to one seen in radio pulsars .
we find no significant pulse morphology variations in any axp in quiescence .
we considered high signal - to - noise average pulse profiles for each axp as a function of energy .
we show that , as in the timing properties , there is a variety of different behaviors for the energy dependence .
we also used the monitoring and archival data to obtain pulsed flux time series for each source .
we have found no large changes in pulsed flux for any source in quiescence , and have set @xmath1 upper limits on variations @xmath22030% depending on the source . we have recently discovered bursts from the direction of two axps : 1e 1048.15937 the most sgr - like axp , and 1e 2259.1 + 586 the most rotationally stable axp .
we compare the temporal , spectral and flux properties of these events to those of sgr bursts , and show that the two phenomena are very similar .
these results imply a close relationship between axps and sgrs , with both being magnetars . |
weak gravitational lensing of the microwave background anisotropies offers a unique opportunity to study the dark matter and energy distribution at intermediate redshifts and large scales .
in addition to producing modifications in the cmb temperature and polarization power spectra @xcite , lensing of the cmb fields produces higher - order correlations between the multipole moments @xcite .
quadratic combinations of the cmb fields can be used to form estimators of the projected gravitational potential , and therefore of the projected mass @xcite .
the minimum variance quadratic estimator can in principle map the projected mass on large angular scales out to multipole moments of @xmath0 @xcite and contains nearly all of the information in the higher moments of the lensed temperature field @xcite . substantially more information lies in the lensed polarization fields allowing high signal - to - noise lensing reconstruction and extending the angular resolution out to @xmath1 @xcite .
lensing reconstruction techniques involving the polarization fields have previously only been developed for small surveys where the sky can be taken to be approximately flat . since lensing is intrinsically most sensitive to the projected potential at @xmath2 or several degrees on the sky , a treatment incorporating the curvature of the sky is desirable .
in fact it is necessary for its application in removing the lensing contaminant to gravitational wave polarization @xcite across large regions of the sky .
we present a concise treatment of the effect of gravitational lensing on cmb temperature and polarization harmonics in sect .
[ sect : cmblensingmultipoles ] .
we construct the full sky quadratic estimators of the lensing potential and compare their noise properties to that for the flat sky expressions in sect .
[ sect : quadraticestimators ] .
we provide an efficient algorithm for the construction of all estimators in sect .
[ sect : angularspaceestimators ] .
we summarize some useful properties of spin - weighted functions in appendix [ appendix : spinsfunctions ] .
finally , we derive the flat sky limits of the estimators and draw the connection to results in @xcite in sect . [
appendix : flatsky ] .
in this section , we give a pedagogical but concise derivation of the lensing effect on the cmb temperature and polarization fields on the sphere @xcite .
we emphasize the connections between the formalism using spin - weighted spherical harmonics @xcite and a tensorial approach @xcite which will be useful for the lensing reconstruction in the following sections .
the temperature perturbation is characterized by a scalar function @xmath3 , whose harmonic transform is given by @xmath4 the polarization anisotropy of the microwave background is characterized by a traceless , symmetric rank 2 tensor , which can be represented as ( e.g. @xcite ) @xmath5 where we have defined the complex stokes parameters @xmath6 according to @xmath7 the spin projection vectors are given with respect to the measurement basis @xmath8 by @xmath9 , \label{eqn : vm}\\ { \bar{{\ensuremath{\bm{m}}}}}&=&\frac{1}{\sqrt{2}}\left [ { \ensuremath{\hat{\bm{e}}}}_1 - i { \ensuremath{\hat{\bm{e}}}}_2\right ] , \label{eqn : vp}\end{aligned}\ ] ] and form an eigenbasis under local rotations of basis vectors ( see appendix [ appendix : spinsfunctions ] ) . in spherical polar coordinates , @xmath10 and @xmath11 . under a local , right - handed rotation of the basis @xmath8 by an angle @xmath12 , the complex stokes parameters @xmath13 acquire a phase @xmath14 .
they act as spin-2 functions , with a corresponding harmonic transform in terms of spin - weighted spherical harmonics @xcite given by @xcite @xmath15 a lens with a projected potential @xmath16 maps the temperature and polarization anisotropies according to @xcite @xmath17 where tildes denote the unlensed fields . in the case of a weak gravitational field under consideration , lensing potential @xmath18 is obtained by a line - of - sight projection of the gravitational potential , @xmath19 where @xmath20 is the conformal time , @xmath21 is the epoch of last scattering and @xmath22 is the angular diameter distance in comoving coordinates . taking the harmonic transform of eqn .
( [ eqn : tlensmap ] ) , one readily shows that the the change to the temperature moments @xmath23 are given by @xcite @xmath24 with @xmath25 denoting the integral @xmath26 the integral can be performed analytically using the relation @xcite @xmath27^{1/2 } { \left ( \begin{array}{ccc}l_1&l_2&l_3\\ -s_1&-s_2&-s_3\end{array } \right ) } { \left ( \begin{array}{ccc}l_1&l_2&l_3\\ m_1&m_2&m_3\end{array } \right ) } \label{eqn : threejtosylm}\ ] ] to yield @xmath28 with the definition @xmath29 \sqrt{\frac{(2l+1)(2l+1)(2l'+1)}{16\pi}}{\left ( \begin{array}{ccc}l&l&l'\\ \pm s&0&\mp s\end{array } \right ) } .
\label{eqn:2fdefinition}\ ] ] the multipole expansion for the polarization fields proceeds by noting that @xmath13 are the spin @xmath30 components of the polarization tensor @xmath31 .
since the contraction with the spin projection vectors @xmath32 projects out the spin 2 piece of a symmetric tensor , the change in the complex stokes parameters is given by @xmath33 & = & { { \ensuremath{\bm{m}}}}^i{{\ensuremath{\bm{m}}}}^j \delta{\bm{\mathcal p}}_{ij}\nonumber \\
& \approx & { { \ensuremath{\bm{m}}}}^i{{\ensuremath{\bm{m}}}}^j\left [ \nabla_k \tilde { \bm{\mathcal p}}_{ij}({\ensuremath{\hat{\bm{n}}}})\right ] \left [ \nabla^k\phi({\ensuremath{\hat{\bm{n}}}})\right ] .
\label{eqn : lensingcontribution}\end{aligned}\ ] ] the expression for the contribution to @xmath34 is obtained by replacing @xmath35 by @xmath36 in the above .
we denote the product @xmath37 using a spin - gradient derivative @xmath38 ( see eq .
[ eqn : gradientop ] ) , and write the lensing contribution as @xmath39 \approx d^i\phi({\ensuremath{\hat{\bm{n } } } } ) d_i[{}_{\pm 2}\tilde{a}({\ensuremath{\hat{\bm{n } } } } ) ] .
\label{eqn : lensingcontrib2}\ ] ] this relationship was given in @xcite with the shorthand convention @xmath40 corresponding to the action of covariant derivatives on the spin components of symmetric trace free tensors given in eqn .
( [ eqn : covderivfinal ] ) @xcite . expanding @xmath18 and @xmath6 in
spin - weighted spherical harmonics and evaluating the inner product of their gradients using eqn .
( [ eqn : gradientdot ] ) , we obtain the lensing corrections @xmath41 & \approx & \sum_{lm}\sum_{l'm ' } \phi_l^m{}_{\pm 2}\tilde{a}_l^m{}_{\pm 2}i_{lll'}^{mmm ' } , \label{eqn : lensedpmalm}\end{aligned}\ ] ] where we define @xmath42 we will be interested in the lensing expressions for the rotationally invariant combinations @xmath43 , \label{eqn : emultipole } \\
b_l^m & = & \frac{1}{2i}\left [ { } _ { + 2}a_l^m - { } _ { -2}a_l^m\right ] \label{eqn : bmultipole},\end{aligned}\ ] ] which are the curl - free ( `` e - mode '' ) and gradient - free ( `` b - mode '' ) components of the polarization field .
> from the expressions ( [ eqn : lensedtlm ] ) and ( [ eqn : lensedpmalm ] ) , we find the general expression for a lensed multipole moment to be @xmath44 , \label{eqn : lensedxlm}\end{aligned}\ ] ] where @xmath45 may be multipole moments of @xmath46 , @xmath47 , or @xmath48 , and @xmath49 ensure that the associated terms are nonzero only when @xmath50 is even or odd , respectively .
@xmath51 denotes the parity complement of @xmath52 , i.e. @xmath53 , @xmath54 , @xmath55 .
lensing of the cmb fields mixes different multipoles through the convolution ( [ eqn : lensedxlm ] ) , and therefore correlates modes across a band determined by the power in the deflection angles @xcite .
the unlensed cmb multipoles @xmath56 are assumed to be gaussian and statistically isotropic , so that the statistical properties are characterized by diagonal covariances or power spectra @xmath57 the assumption of parity invariance implies that @xmath58 the lensing potential is also assumed to be statistically isotropic so that @xmath59 where we have multiplied through by @xmath60 to reflect the weighting of deflection angles .
it follows then that the lensed multipoles are also statistically isotropic with power spectra @xmath61 since @xmath62 denotes the measured multipoles , the power spectra contain all sources to the variance , including detector noise .
detector noise will be taken to be homogeneous , with power spectra given by @xcite @xmath63 where @xmath64 and @xmath65 characterize detector noise , and @xmath66 is the fwhm of the beam .
we employ the specifications of a nearly ideal reference experiment , with @xmath67 , @xmath68 , and @xmath69 ( see @xcite for an exploration of noise properties ) . if we instead consider an ensemble of cmb fields lensed by a _
deflection field , the multipole covariance acquires off - diagonal terms and becomes @xmath70 where the subscript on the average indicates that we consider a fixed lensing field .
@xmath71 are weights for the different quadratic pairs denoted by @xmath72 , given by @xmath73+{}_{s_b}f_{l_2ll_1 } \left [ { \ensuremath{\epsilon}}_{l_1l_2l}\tilde{c}_{l_1}^{ab } -{\ensuremath{\beta}}_{l_1l_2l}\tilde{c}_{l_1}^{a \bar b } \right ] , \label{eqn : fdefinition}\ ] ] where @xmath74 and @xmath75 are the spins of the @xmath76 and @xmath77 fields respectively .
specific forms for the six quadratic pairs are given in table [ table : fforms ] .
.functional forms for @xmath78 .
`` even '' and `` odd '' indicate that the functions are non - zero only when @xmath79 is even or odd , respectively . [ cols="^,^",options="header " , ] because @xmath80 has a zero mean , the off - diagonal terms of the two - point correlations @xmath81 taken over a statistical ensemble would vanish .
however , in a given realization , we can construct an estimator for the deflections as a weighted sum over multipole pairs , and find weights that minimize the variance of the estimator .
we write a general weighted sum of multipole pairs as @xmath82 where @xmath83 and @xmath84 are the observed cmb multipoles , @xmath72 denotes the specific choice of @xmath76 and @xmath77 , and the sum includes the diagonal ( @xmath85 , @xmath86 ) pieces .
> from expression ( [ eqn : lensaverage ] ) for the average over a fixed lens realization , @xmath87 .
\label{eqn : estimatorlensreduced}\end{aligned}\ ] ] where we have used the relations @xmath88 the diagonal terms in eq .
( [ eqn : estimatordefinition ] ) only contribute to the unobservable monopole piece and we hereafter implicitly consider @xmath89 only .
the normalization is set by the condition @xmath90 to be @xmath91 we derive the minimum variance estimator by minimizing the gaussian variance @xmath92 with respect to @xmath93 and find that @xmath94 note that for @xmath95 , @xmath96 and for @xmath97 ( e.g. , for @xmath98 or @xmath99 ) , @xmath100 the gaussian noise covariance @xmath101\ ] ] is given by @xmath102\right \ } , \label{eqn : noisecovariance}\ ] ] with @xmath103 , @xmath104 . for @xmath105 ,
the above reduces simply to @xmath106 . following the treatment of the flat sky case in @xcite , we combine the measured quadratic estimators to further improve the signal to noise by a forming minimum variance estimator @xmath107 with weights and variance given by @xmath108 we will hereafter ignore contributions from the @xmath109 estimator , since the primordial contributions to the @xmath48-mode power spectrum is expected to be small on scales where the lensed multipoles are employed .
we plot the noise power spectra for the five estimators , as well as the minimum variance estimator , in fig .
[ fig : fullskynoise ] , assuming the noise properties of the reference experiment .
the quadratic estimators involve both filtering and convolution in harmonic space .
it is useful in practice to express the convolution as a product of the fields in angular space .
the estimators can then be constructed using fast harmonic transform algorithms @xcite . to simplify the construction of the estimators
we will assume @xmath110 as is appropriate for the standard cosmology .
aside from the @xmath99 estimator , derived in @xcite under the flat sky approximation , the angular space estimators involving polarization are new to this work .
generalizing the construction in @xcite for the @xmath111 estimator , consider the fact that lensing correlates the ( lensed ) temperature and polarization fields to the their ( unlensed ) angular gradients .
we show in appendix [ appendix : spinsfunctions ] that the all - sky analog to the gradient operation on a spin-@xmath112 field is @xmath113 .
the quadratic estimator is then built out of the general operation on two fields @xmath114 and @xmath115 @xmath116 \equiv - d^i [ x({\ensuremath{\hat{\bm{n } } } } ) d_i y({\ensuremath{\hat{\bm{n } } } } ) ] \ , .
\label{eqn : gradientcorr}\ ] ] the properly normalized estimators then take the form @xmath117 where @xmath118\,,\nonumber\\ e^{\theta e } ( { \ensuremath{\hat{\bm{n } } } } ) & = \frac{1}{2 } \left ( p[{}_{+2 } a_e , { } _ { -2 } a_{\theta e } ] + p[{\rm cc } ] \right ) + p[{}_0 a_\theta , { } _ 0 a_{e\theta}]\ , , \nonumber\\ e^{\theta b } ( { \ensuremath{\hat{\bm{n } } } } ) & = \frac{1}{2 } \left ( p [ { } _ { + 2 } a_{ib } , { } _ { -2 } a_{\theta e } ] + p[{\rm cc } ] \right ) \,,\nonumber\\ e^{e e } ( { \ensuremath{\hat{\bm{n } } } } ) & = \frac{1}{2 } \left ( p [ { } _ { + 2 } a_e , { } _ { -2 } a_{ee } ] + p[{\rm cc } ] \right ) \,,\nonumber\\ e^{e b } ( { \ensuremath{\hat{\bm{n } } } } ) & = \frac{1}{2 } \left ( p [ { } _ { + 2 } a_{ib},{}_{-2 } a_{ee } ] + p[{\rm cc } ] \right ) \,,\end{aligned}\ ] ] where cc denotes the operation with the complex conjugates of the fields and the filtered fields themselves are given by the general prescription @xmath119 we omit the @xmath120 estimator under the assumption that the unlensed @xmath48-power is small at high multipoles .
it is straightforward to verify that all of the estimators are the same as the harmonic space ones @xmath121 with @xmath122 except for @xmath123 . here
the weights on the multipole combination are @xmath124 and are slightly non - optimal compared with the minimum variance weighting .
furthermore @xmath125 and they must be calculated separately . however a direct calculation of the noise spectrum through eqn .
( [ eqn : noisecovariance ] ) shows that the differences are less than @xmath126 , and essentially indistinguishable from the minimum variance estimator ( see fig .
[ fig : tenoise ] ) .
these estimators may therefore be used in place of a direct multipole summation for efficient lens reconstruction .
the gradient operations in eqn .
( [ eqn : gradientcorr ] ) are efficiently evaluated in harmonic space since their action on spin harmonics simply raises and lowers the spin index in accordance with eqn .
( [ eqn : spingradient ] ) .
counterintuitively , the gravitational lensing of the cmb temperature and polarization fields is a small scale manifestation of the very large scale properties of the intervening mass distribution .
it therefore requires very challenging , high angular resolution ( @xmath127 ) but wide - field surveys ( @xmath128 few degrees ) to exploit .
we have provided expressions for quadratic estimators of the lensing potential valid on the entire sky , as well as the expected noise covariances for the estimators . as expected , on small angular scales ( @xmath129 ) ,
the flat sky approximations differ from the full sky expressions by less than @xmath130 , indicating that the flat sky approximations is adequate .
this regime is however not where the signal - to - noise peaks .
we have also provided a practical means of implementing these estimators using fast harmonic transforms , either with spherical harmonics or fourier harmonics , to perform the required harmonic convolutions and filtering .
we have shown that even the approximate @xmath123 estimator has a noise performance that is essentially indistinguishable from the minimum variance estimator .
these techniques should provide a means to study the impact of real world issues such as finite - field , inhomogeneous noise , and foregrounds on the science of cmb lensing .
this work was supported by nasa nag5 - 10840 and the doe oji program .
we clarify the relation between spin-@xmath112 functions and tensor quantities on the sphere , and derive the relation between spin raising and lowering operators and covariant derivatives on the sphere .
suppose we construct an orthonormal basis @xmath131 at each point on the sphere , with @xmath132 denoting the outward - facing normal vector .
we define a local rotation as a right - handed rotation of the basis vectors @xmath8 by an angle @xmath12 around the vector @xmath132 , so that the new basis vectors @xmath133 are related to the original vectors by the transformation @xmath134 a function @xmath135 is said to carry a spin - weight @xmath112 if , under the rotation ( [ eqn : basisrotation ] ) , the function transforms as @xmath136 .
this convention conforms to @xcite , and defines rotations in a sense opposite to that in @xcite .
we define vectors @xmath137 and @xmath138 with respect to the basis @xmath8 according to @xmath9,\\ { \bar{{\ensuremath{\bm{m}}}}}&= & \frac{1}{\sqrt{2}}\left [ { \ensuremath{\hat{\bm{e}}}}_1 - i { \ensuremath{\hat{\bm{e}}}}_2\right ] , \label{eqn : projectionvectors}\end{aligned}\ ] ] which have the property that @xmath139 given a vector field @xmath140 , it can easily be shown that the quantities @xmath141 and @xmath142 transform as spin @xmath143 and @xmath144 objects , respectively , so that @xmath137 and @xmath138 act as spin projection vectors .
more generally , given a rank-@xmath112 tensor @xmath145 , the quantity @xmath146 transforms as a spin-@xmath112 object , since under the rotation ( [ eqn : basisrotation ] ) , each factor of @xmath147 contributes a phase @xmath148 .
the spin-@xmath112 functions @xmath135 therefore also provide a complete basis for the totally symmetric trace - free portion of a rank-@xmath112 tensor @xmath149 where the trace - free condition refers to the vanishing under contraction of any two indices in the tensor .
for example , the polarization tensor can be written as @xmath150 covariant differentiation of such a tensor is related to the raising and lowering of the spin weight @xcite : @xmath151 { \bar{{\ensuremath{\bm{m}}}}}_{i_1}\cdots{\bar{{\ensuremath{\bm{m}}}}}_{i_s } + { } _ sf({\ensuremath{\hat{\bm{n}}}})\nabla_{(k}{\bar{{\ensuremath{\bm{m}}}}}_{i_1}\cdots{\bar{{\ensuremath{\bm{m}}}}}_{i_s ) } \nonumber \\ & + & \left [ \partial_k{}_{-s}f({\ensuremath{\hat{\bm{n}}}})\right ] { { \ensuremath{\bm{m}}}}_{i_1}\cdots{{\ensuremath{\bm{m}}}}_{i_s } + { } _ { -s}f({\ensuremath{\hat{\bm{n}}}})\nabla_{(k}{{\ensuremath{\bm{m}}}}_{i_1}\cdots { { \ensuremath{\bm{m}}}}_{i_s)}. \label{eqn : covariantderivative}\end{aligned}\ ] ] we evaluate the covariant derivatives @xmath152 etc . explicitly in the spherical basis with coordinates @xmath153 , yielding @xmath154 with those for @xmath137 given as complex conjugates of the above . using these
, it can be shown that @xmath155 where @xmath156 in spherical coordinates .
the covariant derivative of @xmath145 is therefore given by @xmath157{\bar{{\ensuremath{\bm{m}}}}}_{i_1}\cdots{\bar{{\ensuremath{\bm{m}}}}}_{i_s}+ \left [ d_k \ , { } _ { -s}f({\ensuremath{\hat{\bm{n}}}})\right ] { { \ensuremath{\bm{m}}}}_{i_1}\cdots{{\ensuremath{\bm{m}}}}_{i_s } , \label{eqn : covderivfinal}\ ] ] where we define the spin - dependent gradient operator as @xmath158 a covariant derivative @xmath159 operating on the spin-@xmath112 piece of a tensor is equivalent to a gradient operation @xmath38 on its spin-@xmath112 weighted representation . as an example , the components of the covariant derivative of the polarization tensor @xmath160,\nonumber \\ { \bar{{\ensuremath{\bm{m}}}}}^i{\bar{{\ensuremath{\bm{m}}}}}^j\nabla_k { \mathcal p}_{ij}&=&d_k[{}_{-2 } a({\ensuremath{\hat{\bm{n } } } } ) ] .
\label{eqn : dkexample}\end{aligned}\ ] ] the gradient operator @xmath161 is related to spin raising and lowering operators . using the expressions ( [ eqn : jevaluation ] ) and expressing the operator @xmath161 in the @xmath162 basis , we obtain the desired relations @xmath163 & = & -\frac{1}{\sqrt{2}}\left \{\left [ \edth { } _ sf({\ensuremath{\hat{\bm{n}}}})\right ] { \bar{{\ensuremath{\bm{m}}}}}_i + \left [ \baredth { } _ sf({\ensuremath{\hat{\bm{n}}}})\right ] { { \ensuremath{\bm{m}}}}_i \right \}.\label{eqn : derivtospinladder}\end{aligned}\ ] ] by virtue of the rotational properties of ( @xmath137 , @xmath138 ) , the ladder operators @xmath164 and @xmath165 , defined by @xcite @xmath166\sin^{-s}\theta{}_sf(\theta,\varphi ) , \label{eqn : lowering}\\ \baredth{}_sf(\theta,\varphi)&=&-\sin^{-s}\theta\left [ \frac{\partial}{\partial\theta}-i\csc\theta\frac{\partial}{\partial\varphi } \right ] \sin^s\theta{}_sf(\theta,\varphi),\label{eqn : raising}\end{aligned}\ ] ] raise and lower the spin weight by @xmath143 .
for example , the gradient operation on the spin-@xmath112 spherical harmonic yields @xmath167 = -\frac{1}{\sqrt{2 } } \left ( [ ( l - s)(l+s+1)]^{1/2 } { } _ { s+1}y_l^m { \bar{{\ensuremath{\bm{m}}}}}_i - [ ( l+s)(l - s+1)]^{1/2 } { } _ { s-1}y_l^m { { \ensuremath{\bm{m}}}}_i\right ) .
\label{eqn : spingradient}\ ] ] note that the inner product of two gradients @xmath168 [ d_i\ , { } _ { s_2 } f_2({\ensuremath{\hat{\bm{n } } } } ) ] = \frac{1}{2}\left \ { [ \baredth\ , { } _ { s_1 } f_1({\ensuremath{\hat{\bm{n}}}})]\ , [ \edth{}_{s_2}f_2 ( { \ensuremath{\hat{\bm{n } } } } ) ] + [ \edth { } _ { s_1 } f_1({\ensuremath{\hat{\bm{n}}}})]\ , [ \baredth \ , { } _ { s_2}f_2({\ensuremath{\hat{\bm{n } } } } ) ] \right \ } \label{eqn : gradientdot}\end{aligned}\ ] ] leaves the total spin - weight of the product unchanged . inverting the relation ( [ eqn : derivtospinladder ] ) , we obtain the ladder operators in the tensor representation @xmath169 for @xmath170 , and with @xmath147 replaced by @xmath171 for @xmath172 .
this relationship was first proven in @xcite , albeit with a different sign convention .
the all - sky estimators derived in sect .
[ sect : quadraticestimators ] reduce to the flat sky estimators , based on fourier harmonics of the fields in the small angle limit . here
we explicitly show this correspondence .
the full sky harmonics in multipole space @xmath173 are related to the flat sky harmonics in fourier space @xmath174 by @xcite @xmath175 rewriting eq .
( [ eqn : estimatordefinition ] ) using the above , @xmath176 to go further , we can utilize the approximation @xcite @xmath177 with the trigonometric functions defined through the cosine and sine rules , and the 3-j symbol on the rhs for the odd case represents a continuation of the analytic expression for the even case @xmath178^{1/2 } , \label{eqn : threejexpression}\ ] ] where @xmath179 .
the approximations for @xmath78 and @xmath93 can likewise be written as @xmath185 where @xmath186 and @xmath187 are defined as the unbarred quantities with @xmath182 replaced by @xmath188 .
we will also utilize the relation between plane waves and spherical harmonics , @xmath189 using eq .
( [ eqn : barquantities ] ) to rewrite eq .
( [ eqn : flatestimator1 ] ) , and applying relations ( [ eqn : threejtosylm ] ) and ( [ eqn : planewave ] ) , the estimator becomes @xmath190 taking @xmath180 , the above reduces to @xmath191 with @xmath187 corresponding to the filters @xmath192 in @xcite .
the normalization @xmath193 reduces to the flat sky expression in @xcite in a similar fashion , by using the approximations ( [ eqn : barquantities ] ) to relate the full sky quantities to trigonometric functions on the flat sky .
it is simple to show that the efficient all - sky estimator in eqn .
( [ eqn : angularestimator ] ) reduces to efficient flat sky estimators with the replacements @xmath194 in eqn .
( [ eqn : gradientcorr ] ) and the spherical harmonic transform in eqn .
( [ eqn : angularestimator ] ) with a fourier transform . under the assumption that @xmath110 , they again reproduce the properties of the minimum variance quadratic estimators in eqn .
( [ eqn : flatestimatorfinal ] ) and allow fast fourier transform techniques to be employed in their construction . fig .
[ fig : comparison ] shows fractional differences between the noise in flat sky estimators derived in @xcite and the noise in full sky estimators , defined as @xmath195 . because most of the information comes from multipole pairs at high multipole moments , the flat sky expressions deviate at less than @xmath130 for @xmath196 , mainly in the direction of overestimating the noise .
u. seljak , * 463 * , 1 ( 1996 ) .
m. zaldarriaga and u. seljak , * 58 * , 023003 ( 1998 ) .
f. bernardeau , astron .
& astrophys .
* 324 * , 15 ( 1997 ) .
f. bernardeau , astron . & astrophys .
* 338 * , 375 ( 1998 ) .
m. zaldarriaga and u. seljak , * 59 * , 123507 ( 1999 ) .
j. guzik , u. seljak , and m. zaldarriaga , * 62 * , 043517 ( 2000 ) .
w. hu , * 64 * , 083005 ( 2001 ) .
w. hu , lett * 557 * , l79 ( 2001 ) .
hirata and u. seljak , * in press * , astro - ph/0209489 ( 2002 ) .
w. hu and t. okamoto , * 574 * , 566 ( 2002 ) . l. knox and y. s. song , * 89 * , 011303 ( 2002 ) .
m. kesden , a. cooray , and m. kamionkowski , * 89 * , 011304 ( 2002 ) .
w. hu , * 62 * , 043007 ( 2000 ) .
a. challinor and g. chon , * in press * , ( 2000 ) .
m. zaldarriaga and u. seljak , * 55 * , 1830 ( 1997 ) .
m. kamionkowski , a. kosowsky , and a. stebbins , * 55 * , 7368 ( 1997 )
. j. n. goldberg _ et al .
_ , j. math .
* 8 * , 2155 ( 1967 ) .
a. blanchard and j. schneider , astron .
& astrophys .
* 184 * , 1 ( 1987 ) .
goldberg and d.n .
spergel , * 59 * , 103002 ( 1999 ) .
d. a. varshalovich , a. n. moskalev , and v. k. kersonskii , _ quantum theory of angular momentum_. ( world scientific , singapore , 1989 ) .
l. knox , * 52 * , 4307 ( 1995 ) . k.m .
gorski et al . , preprint * astro - ph/9905275 * , ( 1999 ) .
muciaccia , p. natoli , and n. vittorio , * 488 * , 63 ( 1997 ) .
e. newman and r. penrose , j. math .
* 7 * , 863 ( 1966 ) . m. white , j.e .
carlstrom , m. dragovan , and w.l .
holzapfel , * 514 * , 12 ( 1999 ) . | gravitational lensing of the microwave background by the intervening dark matter mainly arises from large - angle fluctuations in the projected gravitational potential and hence offers a unique opportunity to study the physics of the dark sector at large scales .
studies with surveys that cover greater than a percent of the sky will require techniques that incorporate the curvature of the sky .
we lay the groundwork for these studies by deriving the full sky minimum variance quadratic estimators of the lensing potential from the cmb temperature and polarization fields .
we also present a general technique for constructing these estimators , with harmonic space convolutions replaced by real space products , that is appropriate for both the full sky limit and the flat sky approximation .
this also extends previous treatments to include estimators involving the temperature - polarization cross - correlation and should be useful for next generation experiments in which most of the additional information from polarization comes from this channel due to sensitivity limitations . |
h@xmath0o maser emission has been observed in circumstellar envelopes of evolved stars such as o - rich mira variables and oh / ir stars with large mass loss rates of @xmath4 yr@xmath1 @xcite .
most of these stars are asymptotic giant branch ( agb ) stars or red supergiants both with the spectral type m , with a few exceptions for transient stars at pre - planetary nebula phase ( or supposedly a few pre - main sequence stars such as ori kl ; @xcite ) . for the central stars with spectral types earlier than m
, uv radiation from stellar chromosphere eventually dissociates most of molecules ( except co ) in the inner envelope ( e.g. , @xcite ) .
therefore , h@xmath0o ( or sio ) masers are usually not expected for these stars , except for the case that the molecules in dense circumstellar clumps shield themselves from uv radiation .
in fact , h@xmath0o masers found in a young planetary nebula @xcite must be such an exceptional case .
oh masers have been found in yellow hypergiants with spectral types f and g ( such as irc+10420 and v1427 aql ) @xcite .
however , h@xmath0o and sio masers have never been detected in these objects @xcite , though thermal emission of a few other molecules have been observed in the outer circumstellar shell @xcite . [ cols="<,^,^,^,^,^,^,^ " , ] though this star is relatively bright at optical wavelengths ( @xmath5 ; tyco input catalog ) , the star was not recorded in major optical catalogs , for example , not in henry draper ( hd ) catalogue , the hipparcos and tycho catalogue , nor the general catalog of variable stars , possibly because of confusion by the nearby b5ii star ( tyc 4265 - 870 - 1 ; @xmath6 ) , located by about 12@xmath7 east .
this was involved in the washington double star catalog , giving a separation of 10.9@xmath7 in 1901 and 12.0@xmath7 in 2006 with a small position angle variation ( by @xmath8 ) to the b star . from this data
, we obtain the proper motion of 17 mas yr@xmath1 to the west for iras 22480 + 6002 relative to this b star . as noted by @xcite ,
this b5ii star is probably not a binary counterpart because of the large radial velocity difference .
the act reference catalog gave a very small proper motion of this b5ii star ( less than 3 mas yr@xmath1 ) ; though the hipparcos and tycho catalogue gave a large proper motion in declination due to position uncertainty , but this was corrected in act catalog .
we also checked the past catalogs recording the position of this star and summarized the results in table 3 .
the gsc 1.2 catalog ( which remeasured the poss1 plate taken in 1950s ) gave a different position by about 4.2@xmath7 , which leads a large proper motion of 84 mas yr@xmath1 if compared with gsc 2.0 .
this value is much larger than the above - mentioned proper motion computed from the washington double star catalog , though the proper motion vectors are roughly in the same direction .
we believe the direct measurements of binary separation gives better values .
therefore , we adopt the proper motion of 17 mas yr@xmath1 for this star , and get @xmath9 km s@xmath1 and @xmath10 km s@xmath1 for iras 22480 + 6002 . this motion is considerably peculiar for a population i disk star .
it is likely that iras 22480 + 6002 belongs to one of kinematical streaming groups of stars as population ii g and k giants @xcite . in the past ,
oh and h@xmath0o masers have been found in a few planetary and preplanetary nebulae @xcite , where central stars of these objects have spectral types earlier than m. these masers are a remnant of circumstellar material which was ejected at the agb phase of the central star .
the molecules responsible for masers are eventually to be dissociated .
in contrast , @xcite found oh 1665/1667 mhz emission toward several warm stars as rv tau variables with spectral type of g and k , but so far only one case ( tw aql , a semi - regular variable of k7iii ) was confirmed to be a circumstellar maser @xcite .
sio masers , which are emitted within a few stellar radii of the central star ( much closer than h@xmath0o masers are emitted ) , were not detected in these warm objects before .
an exceptional case is the rv tau variable , r sct , with spectral type k0ib .
this object exhibits strong sio and weak h@xmath0o masers ( i. yamamura , 2004 private communication ) as well as the 4 @xmath11 m sio first overtone bands @xcite .
the rv tau variables are believed to be low - mass post - agb stars @xcite with low metal abundances ( population ii ; @xcite ) , though these are spectroscopically classified as supergiants .
their spectral types change between k and m - type with light variation @xcite .
the atmosphere of late - type supergiants are not in hydrostatic equilibrium ; effective temperature increases with decreasing metalicity @xcite .
the rv tau variables are enshrouded by dust shell , and co emission has been detected in two of these variables @xcite . although the optical counterpart of iras 22480 + 6002 is not reported to show any strong light variability ( e.g. , tass ; the amateur sky survey ) , the optical spectroscopic classification , middle infrared properties , and maser characteristics of iras 22480 + 6002 indicate a close similarity to the properties of the rv tau variables .
we observed the h@xmath0o maser emission distribution around the evolved star iras 22480 + 6002 with the japanese vlbi network .
the maser emission was found to come from an area elongated in the east - west direction by about 50 mas , but no clear trend was found between radial velocities and spatial positions . with the three - epoch two - month interval vlbi observations , we found relatively large proper motions of a few up to 5 mas yr@xmath1 for individual h@xmath0o maser features .
they exhibits no systematic trend in their velocity field , and the inner envelope of this star is very disturbed . applying the statistical parallax method
, we obtained a distance of @xmath12 kpc , which is significantly smaller than the value previously thought .
this distance gives a small luminosity of @xmath13 , but not unreasonably small as a k5 or m0-type supergiant .
combination of the distance and the previous optical proper motion data of this star place this object to be in a dynamically streaming group of stars .
the maser characteristic , and optical and infrared properties of this star are quite similar to those of population ii post - agb stars as rv tau variables .
we acknowledge all staff members and students who have helped in array operation and in data correlation of the jvn / vera . h. i. was supported by grant - in - aid for scientific research from japan society for promotion science ( 18740109 ) .
this research was made use of vizier database , operated at cds , strasbourg , france .
99 beauchamp , a. , moffat , a. f. j. & drissen , l. , 93 , 187 bowers , p. f. , & johnston , k. j. 1994 , , 92 , 189 bujarrabal , v. , bachiller , r. , alcolea , j. , & martin - pintado , j. 1988 , , 206 , l17 castro - carrizo , a. , lucas , r. , bujarrabal , v. , colomer , f. , & alcolea , j. 2001 , 368 , l34 cotton , w. d. , mennesson , b. , diamond , p. j. , perrin , g. , coud du foresto , v. , et al .
2004 , , 414 , 275 deguchi , s. , nakashima , j. , miyata , t. , & ita , y. 2005 , , 57 , 933 elitzur , m. 1992 , in astronomical masers ( dordrecht : kluwer ) fewley , w. m. 1977 , , 218 , 181 famaey , b. , jorissen , a. , luri , x. , mayor , m. , udry , s. , dejonghe , h. , & turon , c. 2005 , , 430 , 165 fix , j. d. , & claussen , m. j. 1984 , , 287 , l3 giguere , p. t. , woolf , n. j. , & webber , j. c. 1976 , 207 , l195 giridhar , s. , lambert , d. l. , & gonzalez , g. , 2000 , , , 531 , 521 gledhill , t. m. , yates , j. a. , richards , a. m. s. 2001 , , 328 , 301 , groenewegen , m. a. t. , baas , f. , blommaert , j. a. d. l. , stehle , r. , josselin , e. , & tilanus , r. p. j. 1999 , , 140 , 197 han , f. , et al .
1998 , , 127 , 181 honma , m. et al .
2003 , pasj , 55 , l57 humphreys , r. m. , & ney , e. p. 1974
, , 86 , 444 humphreys , r. m. , davidson , k. , & smith , n. 2002 , aj , 124 , 1026 imai , h. , omodaka , t. , hirota , t. , umemoto , t. , sorai , k. , & kondo , t. 2006 , , 58 , 883 imai , h. , deguchi , s. , & sasao , t. 2002 , , 567 , 971 imai , h. , obara , k. , diamond , p. j. , omodaka , t. , & sasao , t. 2002 , , 417 , 829 inomata , n. , imai , h. , & omodaka , t. 2007 , ( in press ) josselin , e. , loup , c. , omont , a. , barnbaum , c. , nyman , l. a. , & sevre , f. 1998 , a&as , 129 , 45 josselin , e. & lbre , a. 2001 , , 367 , 826 josselin , e. & plez , b. 2007 , , 469 , 671 jura , m. 1986 , apj , 309 , 732 le squeren , a. m. , sivagnanam , p. , dennefeld , m. , & david , p. 1992 , a&a , 254 , 133 levesque , e. m. , massey , p. , olsen , k. a. , & plez , b. 2005 , , 628 , 973 matsuura , m. , yamamura , i. , zijlstra , a. a. , & bedding , t. r. 2002 , , 387 , 1022 meixner , m. , ueta , t. , dayal , a. , hora , j. l. , fazio , g. , hrivnak , b. j. , skinner , c. j. , hoffmann , w. f. , deutsch , l. k. , 1999 , apjs , 122 , 221 miranda , l. f. , gmez , y. , anglada , g. , & torrelles , j. m. 2001 , , 414 , 284 morino , j. , yamashita , t. , hasegawa , t. , & nakano , t. 1998 , nature , 393 , 340 nakashima , j. , & deguchi , s. 2003 , pasj , 55 , 229 nedoluha , g. e. & bowers , p. f. 1992 , apj , 399 , 743 nyman , l .- . , hall , p. j. , & olofsson , h. 1998 , , 127 , 185 planesas , p. , bujarrabal , v. , le squeren , a. m. , & mirabel , i. f. 1991 , , 251 , 133 pollard , k. r. , cottrell , p. l. , lawson , w. a. , albrow , m. d. & tobin , w. 1997 , , 286 , 1 reid , m. , & moran , j. 1981 , , 19 , 231 schneps , m. h. , moran , j. m. , genzel , r. , reid , m. j. , lane , a. p. & downes , d. 1981 , 249 , 124 sitnik , t. g. 2003 , astr .
letter 29 , 311 subramaniam , a. , mathew , b. , bhatt , b. c. , ramya , s. 2006 , , 370 , 743 surez , o. , gmez , j. f. , morata , o. 2007 , , 467 , 1085 takaba , h. , ukita , n. , miyaji , t. , & miyoshi , m. 1994 , , 46 , 629 te lintel hekkert , p. t. l. , chapman , j. m. , & zijlstra , a. a. 1992 , apj , 390 , l23 teyssier , d. , hernandez , r. , bujarrabal , v. , yoshida , h. , & phillips , t. g. 2006 , , 450 , 167 vlemmings , w. h. t. , van langevelde , h. j. , & diamond , p. j. 2002 , , 393 , l33 winfrey , s. , barnbaum , c. , morris , m. , & omont , a. 1994 , baas , 185 , 4515 wirsich , j. 1998 , , 502 , 909 worley , c.e .
, & douglass , g.g .
1997 , a&as 125 , 523 xu , y. , reid , m. j. , zheng , x. w. , & menten , k. m. 2006 , science , 311 , 54 yates , j. a. & cohen , r. j. 1994 , , 270 , 958 zijlstra , a. a. , te lintel hekkert , p. , pottasch , s. r. , caswell , j. l. , ratag , m. , & habing , h. j. 1989 , , 217 , 157 zombeck , m. v. 1990 , handbook of space astronomy and astropysics , 2nd ed .
( new york ; cambridge univ . press ) , p71 | we report on the h@xmath0o maser distributions around iras 22480 + 6002 ( = irc+60370 ) observed with the japanese vlbi network ( jvn ) at three epochs spanning 2 months .
this object was identified as a k - type supergiant in 1970s , which was unusual as a stellar maser source .
the spectrum of h@xmath0o masers consists of 5 peaks separated roughly equally by a few km s@xmath1 each .
the h@xmath0o masers were spatially resolved into more than 15 features , which spread about 50 mas along the east west direction . however , no correlation was found between the proper motion vectors and their spatial distributions ; the velocity field of the envelope seems random .
a statistical parallax method applied to the observed proper - motion data set gives a distance of @xmath2 kpc for this object , that is considerably smaller than previously thought .
the distance indicates that this is an evolved star with @xmath3 .
this star shows radio , infrared , and optical characteristics quite similar to those of the population ii post - agb stars such as rv tau variables . |
the large hierarchy between the planck scale where quantum gravity effects are important , and the scale where the electroweak symmetry is broken , drives the wealth of models at the electroweak scale , and motivates the large hadron collider ( lhc ) experiments . while weakly coupled supersymmetry remains a leading candidate to stabilize the hierarchy , the randall - sundrum ( rs ) models of a warped extra dimension @xcite have recently emerged as a fascinating alternative , which may be connected to string landscape solutions of the cosmological constant problem @xcite , and possess an interesting four dimensional dual interpretation in terms of the composite states of a strongly coupled nearly conformal field theory ( cft ) @xcite .
the original rs model had all of the standard model confined to the ir brane ( appearing as composites in the dual description ) .
however , the rs solution to the hierarchy problem requires only the higgs to be localized at the ir boundary , and there are compelling reasons to consider most of the sm might actually lie near the uv brane ( and thus mostly fundamental with respect to the cft in the dual description ) .
theories with the sm in the bulk can incorporate grand unification of couplings @xcite , motivate the flavor hierarchy of fermion masses @xcite , and incorporate a dark matter candidate @xcite .
however , such theories face significant challenge from precision electroweak observables @xcite , requiring specific features @xcite in order to remain natural . at the lhc ,
production of colored states is usually dominant , and the kaluza - klein ( kk ) excitations of the gluons are particularly attractive , because they are singly produced and thus have larger rates than the kk quarks .
thus , they are usually considered to be likely to be the first signs of warped physics , and the first excitation of the gluon ( @xmath3 ) the state for which we will have the most statistics available in order to unravel the details of the underlying theory .
they are the natural place to explore whether or not we can use lhc data to determine which particular detailed rs model has been realized in nature , and which parameters describe it . recently
, significant work has begun on some of the simplest rs constructions to study the production and decay of the first kk mode of the gluon , in order to determine the reach of the lhc to discover rs through its detection @xcite .
while the kk gluon is the most promising avenue to discover rs , it is nevertheless challenging .
the coupling to the light quarks that are the primary constituent of the proton , while characterized by the strong coupling , are somewhat suppressed by the localization of the light fermions close to the uv boundary ( in the cft language , the light fermions are largely fundamental fields and couple to the gluon largely through a small mixing with cft states ) .
this leads to somewhat smaller production cross sections than are typical of qcd .
the decay of the gluon is expected to be predominantly into top quarks , a consequence of the large top mass , which necessitates that top is itself located close to the ir brane ( mostly composite ) .
the tops are produced from a very heavy resonance , and are highly boosted , which makes it experimentally challenging to reconstruct them from the large qcd backgrounds @xcite . in this article , we explore several more of the commonly considered theories which attempt to render rs consistent with precision electroweak data .
we consider the model with a simple @xmath7 @xcite custodial symmetry ( already studied before @xcite ) as a beginning , and also consider models with large brane kinetic terms @xcite or an expanded custodial symmetry which protects the @xmath8-@xmath9-@xmath10 vertex from large corrections @xcite in order to characterize the difference in the properties of the first kk mode of the gluon in each case .
we find that there are general features which can discriminate between the cases , and thus that the specific realization of the rs model leaves an imprint in the properties of the kk gluon .
the article is laid out as follows : in section [ sec : models ] , we review the specific details of the models under consideration . in section [ sec :
xsec ] we show the @xmath3 production rates and decay properties and show how the strong coupling can lead to interesting finite width effects in section [ sec : interfere ] .
section [ sec : conclusions ] contains our conclusions .
the basic rs model is a slice of ads@xmath11 with the background metric @xmath12 , \label{metric - conf}\ ] ] with curvature @xmath13 .
@xmath14 are the coordinates of the four large dimensions , @xmath15 parameterizes the coordinate along the extra dimension , and @xmath16 is the four - dimensional metric .
greek letters denote the four large dimensions @xmath17 and capital roman letters include the fifth dimension as well .
the uv boundary is at @xmath18 where the scale factor @xmath19 and the ir boundary is at @xmath20 , as motivated by the hierarchy problem .
we are particularly interested in a model where all standard model ( sm ) fields , except perhaps the higgs , propagate in the entire 5-d spacetime , and will be primarily concerned with the gluon and colored fermion fields .
the action for the gauge fields and fermions is , s = d^5x \{- f_mn^af^mn a + i ^ e_^m d_m + i c } .
[ eq : action0 ] where @xmath21 are the 5d ( @xmath22 ) dirac matrices , @xmath23 is the veilbein , @xmath24 is an adjoint gauge index and @xmath25 parameterizes the magnitude of a bulk mass for the fermion in units of the curvature .
we work in a unitary gauge @xmath26 , and decompose the 5d fields in kk modes , a^a_(x , z ) = _
n a^a(n)_(x ) g^(n)(z ) , + _ l , r ( x , z ) = ( z)^3/2 _
n ^n_l , r(x ) ^(n)_l , r ( z ) .
the wave functions are given by combinations of bessel functions g^(n)(z ) & = & n_n z .
[ eq : kkgluon ] with normalization factor @xmath27 and admixture controlled by @xmath28 .
the mass spectrum is controlled by the boundary conditions , with the masses satisfying , b_n = - = - .
[ eq : b0 ] for an unbroken gauge group , there is a zero mode with wave function @xmath29 , @xmath30 .
of particular note for the following is the fact that the light kk states have most of their support close to the ir boundary .
the physics of bulk fermions was worked out in @xcite .
the spectrum depends sensitively on the bulk mass term @xmath25 . to remove unwanted light degrees of freedom ,
we impose the boundary conditions such that either the right- or the left - chiral zero mass component is projected out . the kk states form left- and right - chiral pairs whose wave functions are also bessel functions , ^(n ) _ ( z ) = n_n ( z ) , [ eq : fermion ] where @xmath31 are for the right- ( left- ) chiral modes , and the masses are determined by imposing the equality , _ n & = & = , and @xmath32 is a normalization factor .
the zero mode wave functions are , ^(0)(z ) = n_0 ( z ) ^1/2- c _ .
these wavefunctions assume the right - handed zero mode is the one allowed by the boundary conditions ; the left - handed case is given by @xmath33 .
the zero mode is exponentially peaked toward the uv boundary for @xmath34 and toward the ir for @xmath35 . to avoid confusion ,
we adopt a notation where @xmath25 s explicitly refer to right - chiral fields , so the left - chiral fermions should be understood to actually have @xmath36 as their mass parameter . assuming @xmath37 5d yukawa couplings , the hierarchy in the effective 4d yukawa couplings can be motivated by the exponential suppression of the wave functions at the ir boundary for order one differences in @xmath25 . in particular
, one can not allow strong suppression of the top quark wave functions on the ir boundary , because to reproduce the observed top mass one would have to adjust the 5d yukawa coupling to be too strong to have a perturbative description .
there is further motivation from precision electroweak data @xcite , which prefers the light fermion mass parameters to be close to @xmath38 ( including the left - handed top , as it comes along with the left - handed bottom , leading to tension in the choice of @xmath25 for @xmath39 ) in order to cancel the leading contribution to the @xmath40 parameter from the weak boson kk modes . with this setup , and the additional suppression of the @xmath41 parameter from a custodial @xmath7 , masses of the kk modes of around @xmath42 tev
are roughly consistent with precision measurements . specifically , we consider @xmath43 , @xmath44 , and all others @xmath45 .
as we will see shortly , the physics we study does not depend strongly on @xmath25 once it is @xmath46 , so the specific values for light fermions are not important .
the choices of @xmath25 specify the fermion zero mode wave functions , and we compute the couplings of the first kk gluon to the fermion zero modes as the integral over the wave functions .
the light quarks all have very similar couplings of roughly @xmath47 , the third family left - handed quarks @xmath48 , and the right - handed top quark @xmath49 , where @xmath50 is the strong coupling constant which characterizes the coupling of the gluon zero mode .
an alternate way to render precision electroweak data consistent with low kk mode masses is to include large - ish kinetic terms for the gauge fields on the ir brane @xcite .
such terms repell the kk mode wave functions from the brane , and have a large effect on the phenomenology of the kk modes @xcite .
brane terms are a class of higher dimensional operators of the 5d theory , - d^5x \ { f_mn^af^mn a } 2 r_ir ( z - z_v ) [ eq : actionbk ] and will be induced by the orbifold boundary conditions and localized fields @xcite .
their magnitude @xmath51 is a free parameter of the effective theory .
while the size of the ir boundary kinetic term for the gluon is not closely connected to the quality of the electroweak fit , one would expect that if the uv physics is such that there are large ir kinetic terms for the electroweak bosons , such terms are probably also large for the gluon as well .
thus , if one could infer the presence of large gluon terms , it would at least suggest that a similar term is present in the electroweak sector , and responsible for the success of the sm in explaining the electroweak fit
. the ir boundary kinetic term does not affect the form of the bulk wave functions , eq .
( [ eq : kkgluon ] ) .
the boundary conditions become b = - = - , [ eq : eigenmassbk ] indicating different admixture of the bessel functions @xmath52 and @xmath53 in the solutions .
while there is no analytic solution for the masses , they may be easily obtained numerically . in figure
[ fig : bkm ] , we show the variation of the first kk mode gluon mass and coupling to uv - localized states as a function of the magnitude of the ir brane term @xmath54 . in figure
[ fig : gffbkcoup ] , we show the dependence of the coupling on @xmath25 for a few different choices of @xmath54 .
the inclusion of the boundary terms ameliorates the strongest constraints from precision electroweak data , and opens up considerably more freedom to choose the fermion @xmath25 s .
however , in computing properties below , we imagine a situation in which the @xmath25 s are as in the @xmath7 custodial version outlined above ( for example , to explain the flavor hierarchies ) , with large contributions to the electroweak @xmath41-parameter controlled by the ir boundary kinetic terms . [ cols="^,^ " , ]
we have investigated the structure of the kk gluon resonance in several variants of the rs model .
we find that this structure contains information that will help to distinguish between models even in the absence of data from the electroweak sector .
the width and branching ratios will constrain the location of the fermion zero - modes as well as the number of light kk modes into which the kk gluon can decay . in addition , the ratio of cross - sections for producing the @xmath0 directly and in association with a @xmath9-jet will give specific information about the localization of the third generation quarks . specifically , a large coupling to @xmath55 will prefer a model where the @xmath56 vertex is protected by an extended custodial symmetry . in some models , with large boundary kinetic terms , the @xmath3 can primarily decay into dijets , and it seems promising that in such models one can discern @xmath3 against the large qcd background up to masses somewhat larger than 4 tev .
finally , we find that the relative sign of the coupling to light quarks and to tops can be measured in the interference with @xmath57-channel gluon exchange .
this provides an important consistancy check on the overall picture of the fermion geography and the mechanism by which flavor hierarchies are realized in the fermion yukawa couplings .
the discovery of @xmath3 is an important first step in the discovery of rs , and further observables such as its production rate , associated rate with bottom quarks , total width and branching ratios , and interference with sm @xmath4 production , can yield information about the nature of the the rs construction , and the parameters which describe it .
the authors would like to thank erik brubaker , giacomo cacciapaglia , tao han , gregory house , tom lecompte , david mckeen , arjun menon , jose santiago , and carlos wagner for helpful discussions .
research at argonne national laboratory is supported in part by the department of energy under contract de - ac02 - 06ch11357 .
l. randall and r. sundrum , phys .
lett . * 83 * , 4690 ( 1999 ) [ arxiv : hep - th/9906064 ] .
l. randall and r. sundrum , phys .
lett . *
83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ] .
these arguments were lucidly put together my j. march - russell , talk at hcp 2007 , isola delba , italy .
n. arkani - hamed , m. porrati and l. randall , jhep * 0108 * , 017 ( 2001 ) [ arxiv : hep - th/0012148 ] . l. randall and m. d. schwartz , jhep * 0111 * , 003 ( 2001 ) [ arxiv : hep - th/0108114 ] ; k. agashe , a. delgado and r. sundrum , annals phys . * 304 * , 145 ( 2003 ) [ arxiv : hep - ph/0212028 ] ; m. carena , a. delgado , e. ponton , t. m. p. tait and c. e. m. wagner , phys .
d * 68 * , 035010 ( 2003 ) [ arxiv : hep - ph/0305188 ] ; k. agashe , r. contino and r. sundrum , phys .
lett . * 95 * , 171804 ( 2005 ) [ arxiv : hep - ph/0502222 ] .
s. j. huber and q. shafi , phys .
b * 498 * , 256 ( 2001 ) [ arxiv : hep - ph/0010195 ] .
k. agashe and g. servant , phys .
lett . * 93 * , 231805 ( 2004 ) [ arxiv : hep - ph/0403143 ] ; k. agashe and g. servant , jcap * 0502 * , 002 ( 2005 ) [ arxiv : hep - ph/0411254 ] .
h. davoudiasl , j. l. hewett and t. g. rizzo , phys .
b * 473 * , 43 ( 2000 ) [ arxiv : hep - ph/9911262 ] ; a. pomarol , phys .
b * 486 * , 153 ( 2000 ) [ arxiv : hep - ph/9911294 ] . c. csaki , j. erlich and j. terning , phys .
d * 66 * , 064021 ( 2002 ) [ arxiv : hep - ph/0203034 ] .
k. agashe , a. delgado , m. j. may and r. sundrum , jhep * 0308 * , 050 ( 2003 ) [ arxiv : hep - ph/0308036 ] .
h. davoudiasl , j. l. hewett and t. g. rizzo , phys .
d * 68 * , 045002 ( 2003 ) [ arxiv : hep - ph/0212279 ] ; m. carena , e. ponton , t. m. p. tait and c. e. m. wagner , phys .
d * 67 * , 096006 ( 2003 ) [ arxiv : hep - ph/0212307 ] ; m. carena , a. delgado , e. ponton , t. m. p. tait and c. e. m. wagner , phys .
d * 71 * , 015010 ( 2005 ) [ arxiv : hep - ph/0410344 ] .
k. agashe , r. contino , l. da rold and a. pomarol , phys .
b * 641 * , 62 ( 2006 ) [ arxiv : hep - ph/0605341 ] .
k. agashe , a. belyaev , t. krupovnickas , g. perez and j. virzi , arxiv : hep - ph/0612015 . b. lillie , l. randall and l. t. wang , arxiv : hep - ph/0701166 .
r. contino , l. da rold and a. pomarol , arxiv : hep - ph/0612048 .
m. carena , e. ponton , j. santiago and c. e. m. wagner , arxiv : hep - ph/0701055 . y. grossman and m. neubert , phys . lett .
b * 474 * , 361 ( 2000 ) [ arxiv : hep - ph/9912408 ] .
m. carena , t. m. p. tait and c. e. m. wagner , acta phys .
b * 33 * , 2355 ( 2002 ) [ arxiv : hep - ph/0207056 ] .
h. georgi , a. k. grant and g. hailu , phys .
b * 506 * , 207 ( 2001 ) [ arxiv : hep - ph/0012379 ] .
d. choudhury , t. m. p. tait and c. e. m. wagner , phys .
d * 65 * , 053002 ( 2002 ) [ arxiv : hep - ph/0109097 ] .
k. agashe , r. contino and a. pomarol , nucl .
b * 719 * , 165 ( 2005 ) [ arxiv : hep - ph/0412089 ] .
m. carena , e. ponton , j. santiago and c. e. m. wagner , nucl .
b * 759 * , 202 ( 2006 ) [ arxiv : hep - ph/0607106 ] .
a. d. medina , n. r. shah and c. e. m. wagner , arxiv:0706.1281 [ hep - ph ] .
a. djouadi , g. moreau and f. richard , nucl .
b * 773 * , 43 ( 2007 ) [ arxiv : hep - ph/0610173 ] .
see http://lepewwg.web.cern.ch/lepewwg/ for recent fits to @xmath40 and @xmath41 .
c. csaki , c. grojean , h. murayama , l. pilo and j. terning , phys .
d * 69 * , 055006 ( 2004 ) [ arxiv : hep - ph/0305237 ] . c. csaki , c. grojean , l. pilo and j. terning , phys .
lett . * 92 * , 101802 ( 2004 ) [ arxiv : hep - ph/0308038 ] .
g. cacciapaglia , c. csaki , c. grojean and j. terning , phys .
d * 70 * , 075014 ( 2004 ) [ arxiv : hep - ph/0401160 ]
. g. cacciapaglia , c. csaki , c. grojean and j. terning , phys .
d * 71 * , 035015 ( 2005 ) [ arxiv : hep - ph/0409126 ] .
g. cacciapaglia , c. csaki , c. grojean , m. reece and j. terning , phys .
d * 72 * , 095018 ( 2005 ) [ arxiv : hep - ph/0505001 ] . c. csaki , c. grojean , j. hubisz , y. shirman and j. terning , phys .
d * 70 * , 015012 ( 2004 ) [ arxiv : hep - ph/0310355 ] .
y. nomura , jhep * 0311 * , 050 ( 2003 ) [ arxiv : hep - ph/0309189 ] .
r. barbieri , a. pomarol and r. rattazzi , phys .
b * 591 * , 141 ( 2004 ) [ arxiv : hep - ph/0310285 ] .
h. davoudiasl , j. l. hewett , b. lillie and t. g. rizzo ,
higgsless electroweak symmetry breaking in warped backgrounds : constraints phys .
d * 70 * , 015006 ( 2004 ) [ arxiv : hep - ph/0312193 ] .
h. davoudiasl , j. l. hewett , b. lillie and t. g. rizzo , jhep * 0405 * , 015 ( 2004 ) [ arxiv : hep - ph/0403300 ] .
j. l. hewett , b. lillie and t. g. rizzo , jhep * 0410 * , 014 ( 2004 ) [ arxiv : hep - ph/0407059 ] .
r. s. chivukula , e. h. simmons , h. j. he , m. kurachi and m. tanabashi , the structure of corrections to electroweak interactions in higgsless phys . rev .
d * 70 * , 075008 ( 2004 ) [ arxiv : hep - ph/0406077 ] .
r. s. chivukula , e. h. simmons , h. j. he , m. kurachi and m. tanabashi , phys .
b * 603 * , 210 ( 2004 ) [ arxiv : hep - ph/0408262 ] .
g. cacciapaglia , c. csaki , g. marandella and j. terning , jhep * 0702 * , 036 ( 2007 ) [ arxiv : hep - ph/0611358 ] .
g. cacciapaglia , c. csaki , g. marandella and j. terning , phys .
d * 75 * , 015003 ( 2007 ) [ arxiv : hep - ph/0607146 ] .
t. sjostrand , s. mrenna and p. skands , jhep * 0605 * , 026 ( 2006 ) [ arxiv : hep - ph/0603175 ] .
j. alwall _ et al .
_ , arxiv:0706.2334 [ hep - ph ] .
m. kagan , d. amidei , c. cully , t. schwarz , and m. soderberg http://www-cdf.fnal.gov/physics/new/top/2006/mass/mttb/pub_page.html | we study the properties of @xmath0 , the first excited state of the gluon in representative variants of the randall sundrum model with the standard model fields in the bulk .
we find that measurements of the coupling to light quarks ( from the inclusive cross - section for @xmath1 ) , the coupling to bottom quarks ( from the rate of @xmath2 ) , as well as the overall width , can provide powerful discriminants between the models . in models with large brane kinetic terms , the @xmath3 resonance
can even potentially be discovered decaying into dijets against the large qcd background .
we also derive bounds based on existing tevatron searches for resonant @xmath4 production and find that they require @xmath5 gev .
in addition we explore the pattern of interference between the @xmath3 signal and the non - resonant sm background , defining an asymmetry parameter for the invariant mass distribution .
the interference probes the relative signs of the couplings of the @xmath0 to light quark pairs and to @xmath6 , and thus provides an indication that the top is localized on the other side of the extra dimension from the light quarks , as is typical in the rs framework .
anl - hep - pr-07 - 40 + efi/07 - 18 + |
chemical potential is an important thermodynamic quantity that regulates many phenomena , such as phase transitions , reaction equilibria and electro - chemistry @xcite .
the ability to calculate it via a reliable and fast technique is therefore of great value .
several numerical methods have been proposed , ranging from widom test - particle insertion @xcite to grand - canonical monte - carlo approaches @xcite .
extensive reviews on the efficiency and reliability of these and other techniques can be found in the literature @xcite . however , despite these efforts , chemical potential calculations still remain a difficult challenge for many real systems , such as dense fluids and solids . at constant volume and temperature
the chemical potential is equal to the helmoltz free - energy difference associated to the addition of an extra molecule .
the most common approach to calculate this quantity , namely the widom insertion method , consists in sampling the insertion energy of a test - particle . in this method
the largest contribution comes from those configurations in which the insertion of an extra molecule is energetically favorable .
this occurs if a cavity sufficiently large to accommodate the extra particle is present . in a low density system
many such cavities are present , and widom method is efficient . on the contrary , in dense systems
the formation of a particle - sized cavity is unlikely , and the calculation becomes inefficient .
thus elaborate methods , based on voronoi tessellation , have been suggested as a way to look for the appropriate cavities in a dense system @xcite .
further suggestions have been among others the introduction of energy - biased sampling of the insertion space @xcite , the use of external potentials to generate controlled inhomogeneities within the system @xcite or the application of an adaptive resolution simulation scheme @xcite . in the present paper
we shall focus on the calculation of the chemical potential for an homogeneous liquid , and extend the range of densities at which the particle insertion method can be applied by using well - tempered ( wt ) metadynamics @xcite . here
the role of metadynamics is to enhance fluctuations in a controlled manner , promoting those rare configurations in which particle insertion is favored .
this leads to an efficient and reliable calculation method , even in those cases in which widom s method fails .
the paper is organized as follows : in sec .
[ meta ] metadynamics main features are recalled , while its application to chemical potential calculation is described in sec .
the implementation of our technique for the selected test case , namely the binary lennard - jones ( lj ) fluid , is addressed in sec .
[ impl ] , while the results of our calculations are reported in sec .
[ results ] .
finally the conclusions and perspectives of present work are reported in sec .
[ conclusions ] .
before discussing the chemical potential calculation , we recall some of the features of wt metadynamics . for a more detailed description
we refer to the by now vast literature on the subject ( see e.g. ) .
metadynamics introduces a bias potential to enhance the sampling of rare but important fluctuations .
this bias potential is taken to be a function of the so called collective variables ( cv ) , a set of order parameters that depend on the atomic coordinates @xmath0 .
the cvs are built to describe the degrees of freedom of the system whose sampling we want to enhance .
let us restrict for simplicity to the case of a single collective variable @xmath1 .
the equilibrium distribution of @xmath2 is given by : @xmath3p(\mathbf{r}),\ ] ] where @xmath4 is the boltzmann distribution defined by potential energy function @xmath5 .
@xmath6 is the inverse temperature and @xmath7 is the partition function of the system . introducing the free - energy surface @xmath8 , where @xmath9 is a constant that plays an inessential role , eq .
( [ pofs ] ) can be rewritten as @xmath10 if a bias potential @xmath11 is introduced , a different distribution @xmath12 is sampled : @xmath13p_{v}(\mathbf{r } ) = \frac{\mathrm{e}^{-\beta [ f(s)+v(s)]}}{\int\mathrm{d}s'\mathrm{e}^{-\beta [ f(s')+v(s')]}},\ ] ] where @xmath14}/z_{v}$ ] is the boltzmann distribution of the biased ensemble .
@xmath15}$ ] is the corresponding partition function . in wt metadynamics the bias
is periodically updated by adding a small repulsive potential , that usually takes the form of a gaussian of height @xmath16 and width @xmath17 : @xmath18 @xmath19 is centered on @xmath20 , that is the value of the cv at the moment of the update . during the simulation the height of the deposited gaussians
is modified in a way that can be described by the following iterative equation : @xmath21,\ ] ] where @xmath22 indicates the iteration and @xmath23 is the so called biasing factor , a key parameter in wt metadynamics .
after the @xmath22-th iteration the system evolves under the combined action of @xmath5 and of @xmath24 .
it has been shown rigorously @xcite that the bias constructed as in eq .
( [ vn ] ) converges asymptotically to : @xmath25 as a consequence , the biased distribution @xmath12 is related to the unbiased one by the relation : @xmath26^{1/\gamma}}{\int\mathrm{d}s'\left[p(s')\right]^{1/\gamma } } .\ ] ] this result is central to the present paper .
in fact eq . ( [ reweight ] ) shows that narrow peaks in @xmath27 can be made broader in @xmath12 by increasing @xmath28 . thus rare but important fluctuations , that are hidden in the tails of @xmath27 are more likely to occur in the biased ensemble .
furthermore , the theory of metadynamics allows to calculate the unbiased ensemble average @xmath29 of any position dependent variable @xmath30 via the reweighting @xcite procedure described by : @xmath31}\rangle_v,\ ] ] where the suffix @xmath32 indicates that the average is performed under the action of the time varying metadynamics bias potential @xmath33 .
the position independent constant @xmath34 is given by @xcite : @xmath35}{\int \mathrm{d}s\,\exp\left[\frac{1}{\gamma-1}\beta v(s , t)\right]}.\ ] ]
we now define the proper cv that will be used to calculate the chemical potential via metadynamics .
let us begin by considering a system of @xmath36 particles with coordinate vector @xmath37 .
as pointed out by widom @xcite , the excess chemical potential , that is the chemical potential from which the free particle contribution has been subtracted , can be written as : @xmath38\mathrm{d}\mathbf{r}^{*}\right\rangle,\ ] ] where @xmath39 is the system volume and @xmath40 where @xmath41 is the potential energy of the system with an extra particle in @xmath42 .
the average @xmath29 in eq . ( [ widom ] )
is performed over the boltzmann distribution defined by the @xmath36 particle interaction potential @xmath5 .
widom suggested to estimate @xmath43 by evolving the system and periodically computing @xmath44 at a randomly generated set of positions @xmath42 .
also in other methods , first the coordinate vector @xmath0 is sampled and later insertions are attempted @xcite .
here we take an alternative route and first we fix a number of insertion points @xmath45 and later we let @xmath0 adjust around them so that the sampling of @xmath46 is enhanced . in order to see how this
is done we restrict to the case of an homogeneous fluid , in which the average @xmath47\rangle$ ] is @xmath42 independent .
it follows that one can choose a set of @xmath48 @xmath49 s and rewrite the chemical potential as : @xmath50\right\rangle.\ ] ] this form of @xmath43 suggests the introduction of the following cv : @xmath51\right).\ ] ] in principle even the choice @xmath52 , namely that of using a single insertion point @xmath53 , leads to the correct result , however it requires much longer sampling times . in this paper
the insertion points @xmath54 will be arranged on a regular grid , but other choices are possible .
a consequence of this cv definition is that the chemical potential can be rewritten as the expectation value of a function of @xmath2 : @xmath55 or , in terms of the free energy associated to @xmath2 , as : @xmath56}\mathrm{d}s}{\int\mathrm{e}^{-\beta f(s)}\mathrm{d}s}.\ ] ] from eqs .
( [ muofs ] ) or ( [ muofs2 ] ) it follows that the largest contribution to @xmath43 comes from those configurations for which @xmath2 has a negative value .
this choice of cv has a practical drawback .
in fact in most models the atoms interact at short distances via a highly repulsive and rapidly varying potential .
thus , if any of the insertion points @xmath45 comes close to the atomic positions , @xmath1 varies very rapidly and can even diverge .
this can be remedied by modifying the insertion potential at short distances using a regularized potential @xmath57 in the definition of a new and more smoothly behaved cv : @xmath58\right).\ ] ] the idea is that @xmath59 should be large but finite when @xmath42 is close to one of the atomic positions , so that the derivatives @xmath60 are of manageable size .
one can then use @xmath61 to bias the metadynamics simulation and calculate @xmath62 using the reweighting procedure in eq .
( [ reweight2 ] ) @xcite .
however , in the test case that will be discussed below , we define @xmath57 so that the change of cv has no practical effect on the calculation of @xmath43 , and the reweighting is not necessary ( see also the supporting information ( si ) ) .
in order to test the method we shall consider a binary lennard - jones ( lj ) fluid . in particular we consider the system studied by kob and andersen @xcite for its glass forming properties , that it is known to remain fluid at low temperature and high density . thus it provides a challenging test for our method .
the system is composed by two atomic species , @xmath63 and @xmath64 , of equal atomic mass .
the interaction potential is : @xmath65,\ ] ] where @xmath66 is the atom - atom distance and @xmath67 . in such
a system one defines an excess chemical potential @xmath68 for each species , and generalizes eq .
( [ widom2 ] ) : @xmath69\right\rangle,\ \alpha={a , b}.\ ] ] where @xmath70 where @xmath71 and @xmath72 are the coordinates of the @xmath73 atoms and @xmath0 indicates the ensemble of the two combined sets of coordinates .
@xmath74 is the number of @xmath75 specie particles .
each species will have its own @xmath2 , therefore eq . ( [ sform ] )
generalizes to : @xmath76\right),\ \alpha={a , b}.\ ] ] analogously , eq .
( [ muofs ] ) becomes : @xmath77 and eq .
( [ muofs2 ] ) becomes : @xmath78}\mathrm{d}s_{\alpha}}{\int\mathrm{e}^{-\beta f(s_{\alpha})}\mathrm{d}s_{\alpha}},\ \alpha={a , b}.\ ] ] as mentioned before , in order to avoid divergences and the appearance of a large bias , we shall use the regularized version @xmath79 of the cv , defining the following modified potential : @xmath80 where @xmath81 is the distance below which @xmath82 and @xmath83 differ .
+ kob and andersen model @xcite is a mixture of @xmath84 @xmath63 particles and @xmath85 @xmath64 particles . in units of @xmath86 and @xmath87
the other potential parameters are @xmath88 , @xmath89 , @xmath90 and @xmath91 . in our simulations the potentials
@xmath83 are cut and shifted at @xmath92 , while the distance parameter of the regularized potential @xmath82 is set at @xmath93 .
the simulated system is composed of @xmath94 atoms in a cubic box of length @xmath95 and periodic boundary conditions ( pbc ) are applied .
the temperature is kept at @xmath96 by using the stochastic velocity rescaling thermostat @xcite . at these thermodynamic conditions the system is in the liquid phase .
the md simulations were carried out using the gromacs @xcite software , equipped with a private version of the plumed2 plug - in @xcite .
the integration time - step was chosen to be @xmath97 ( which corresponds to @xmath98 for the lj parameters of argon : @xmath99 , @xmath100 and @xmath101 ) .
the widom s calculations were performed using the appropriate gromacs subroutines .
one advantage of using this particular binary system is that the majority of atoms ( @xmath63 ) have a larger radius , leaving sufficient interstices to place the smaller @xmath64 atoms .
this makes the calculation of @xmath102 relatively easy and accessible to widom s method , providing a testing ground for our technique . on the other hand
the calculation of @xmath103 is challenging and its successful completion demonstrates the usefulness of our approach .
we performed two separate calculations , for @xmath104 and @xmath105 .
the gaussian initial height was chosen to be of @xmath106 and the width is equal to @xmath107 .
new gaussians were deposited every @xmath108 and the @xmath28 parameter was @xmath109 .
we experimented with different values of number of insertion points @xmath48 ( see eq .
( [ sbin ] ) ) .
they all lead to the same results within statistical error . however , the speed of convergence of @xmath43 was different , with the limiting @xmath52 case being particularly slow ( see the si ) . here
we present the results for a regular mesh of @xmath110 points .
while @xmath48 was not carefully optimized , this choice seemed to yield a good compromise between speed of convergence and computational cost .
we note that the calculations can be easily parallelized and it pays to distribute wisely the load among the concurrent processors , thus the determination of the optimal @xmath48 depends also on the computational platform .
as discussed earlier we expect the chemical potential calculation to be easier for the smaller atoms ( @xmath64 ) and more challenging for the large component ( @xmath63 ) .
this is brought out in fig .
[ fsab ] , where the @xmath111 resulting from metadynamics simulations is compared to the @xmath111 calculated in an unbiased md run . in the @xmath64 case ( fig .
[ fsab]a ) , even without applying the bias potential , the range of spontaneously occurring fluctuations , indicated in figure by the domain of the unbiased @xmath111 , is large enough to overlap with the negative @xmath112 region .
this region yields the more substantial contribution to the numerator of eq .
( [ muofsbin ] ) , and thus to the @xmath105 value ( see si for further details ) . thus @xmath105 can be reliably computed with unbiased sampling .
note however that even in this favorable case the spontaneous fluctuations do not cover entirely the tail of the distribution . as noted in ref .
@xcite this leads to a systematic error .
in contrast , no such problem is present in the metadynamics calculation , where the fluctuations of @xmath112 cover thoroughly all the important region . in the case of the larger atoms ( fig . [
fsab]b ) the situation is much more dramatic and the unbiased results show little or no overlap between the fluctuations of @xmath113 and the region that contributes the most to @xmath104 .
in contrast , our metadynamics calculations explore in detail the important region of @xmath114 .
this is reflected by the convergence of @xmath104 and @xmath105 calculations , shown in fig .
[ muab ] , where we compare metadynamics results ( obtained with eq .
( [ reweight2 ] ) ) to standard widom calculations . in the @xmath64 case ( fig . [ muab]a ) the agreement between metadynamics and standard widom method is within the statistical error . in the @xmath63 case however ( fig .
[ muab]b ) widom s method faces severe difficulties and convergence is jumpy and very slow , even with a very large number of insertion points sampled . on the contrary
our method converges to an accuracy of @xmath115 using only @xmath110 insertion points .
this underlines the necessity of metadynamics to visit the extreme tails of the @xmath2 distribution and calculate @xmath43 accurately .
in conclusion we have shown that this new approach can be very competitive with previous ones .
three features should be noted : the use of the homogeneity of the system , the change of paradigm from an atom position - centric to an insertion - centric point of view and the use of metadynamics to enhance fluctuations .
the calculation is no more difficult than a standard wt metadynamics one , and can be performed using any of the codes with which the plumed2 plug - in can be integrated .
the recently developed variationally enhanced sampling @xcite , in its wt version @xcite , could be also profitably applied to perform these calculations and lead to further improvement in the efficiency of chemical potential evaluations .
although here our method has been developed for homogeneous systems , its extension to non - homogeneous situations is possible and it is currently under study . a relatively straightforward application of this method is the calculation of the solvation energy of molecules in liquids , another quantity of great practical interest . moreover ,
if properly adapted , these ideas could be extended to free energy perturbation calculations @xcite .
the authors are pleased to acknowledge m. nava , b. smit and o. valsson for the useful discussions .
the computational resources were provided by the swiss center for scientific computing , and the brutus cluster at eth zurich .
the authors acknowledge also the varmet european union grant erc-2014-adg-670227 and the national centres of competence in research `` materials revolution : computational design and discovery of novel materials '' project for funding . | the calculation of chemical potential has traditionally been a challenge in atomistic simulations .
one of the most used approaches is widom s insertion method in which the chemical potential is calculated by periodically attempting to insert an extra particle in the system . in dense systems
this method fails since the insertion probability is very low . in this paper
we show that in a homogeneous fluid the insertion probability can be increased using metadynamics .
we test our method on a supercooled high density binary lennard - jones fluid .
we find that we can obtain efficiently converged results even when widom s method fails .
+ _ eur .
phys . j. spec .
top .
doi : 10.1140/epjst / e2016 - 60094-x _ |
by now there is large direct and indirect evidence that the universe entered a period of accelerated expansion .
direct evidence is provided by the luminosity - distance data of supernovae of type ia @xcite ( see , however , @xcite ) , indirect evidence comes from the anisotropy spectrum of the cosmic microwave background radiation ( cmbr ) @xcite , from large - scale - structure data @xcite , from the integrated sachs
wolfe effect @xcite , from baryonic acoustic oscillations @xcite and from gravitational lensing @xcite . according to the currently prevailing interpretation , based on einstein s gr
, our universe is dynamically dominated by two so far unknown components , dark matter ( dm ) and dark energy ( de ) .
the latter contributes roughly 75% to the total energy budget , the former about 20% . only about 5% are in the form of conventional , baryonic matter .
the preferred model is the @xmath0cdm model which also plays the role of a reference model for alternative approaches to the de problem .
while the @xmath0cdm model can describe most of the observations , there still remain puzzles @xcite .
the present situation in the field is characterized , e.g. , in @xcite and references therein .
both dm and de manifest themselves observationally only through their gravitational action . in most studies both these components
are regarded as independent .
however , there exists a line of investigation that explores the consequences of a coupling within the dark sector .
corresponding models show a richer dynamics than models with separately conserved energy - momentum tensors .
it has been argued that ignoring a potentially existing interaction between both components may give rise to misinterpretations of observational data regarding the equation - of - state ( eos ) parameter @xcite . a coupling between dm and de may also be relevant with respect to the coincidence problem ( see , e.g. @xcite and references therein ) .
models with an interaction matter
dark energy were introduced by wetterich @xcite .
meanwhile there exists a still growing body of literature on the subject -see , e.g. @xcite and references therein .
limits for the admissible interaction strength have been obtained for various configurations @xcite .
on the other hand , the transition from decelerated to accelerated expansion can be understood as a pure interaction phenomenon in the context of holographic de models @xcite .
recently it was argued that the latest snia could favor a scenario in which the cosmic acceleration has past a maximum value and is now slowing down again @xcite .
in such a case the accelerated expansion could be a transient phenomenon with a late - time dynamics , incompatible with that of the @xmath0cdm model .
it is this possibility which we are going to study in the present paper .
we mention that models of transient acceleration were already discussed in @xcite and more recently in @xcite .
it is the purpose of this paper to provide a toy model which describes a transient cosmological acceleration as the consequence of an interaction between dark matter and dark energy .
such a dynamics can not be obtained if the interaction represents a small correction to , say , the @xmath0cdm model . for models of this type
the long - time cosmological dynamics will always be determined by the cosmological term and result in accelerated expansion . to achieve transient accelerated expansion , a twofold role of the interaction is necessary . at first
, it has to cancel the bare " cosmological constant and at second it has to generate a phase of accelerated expansion by itself .
acceleration has to be an interaction phenomenon .
we show that these requirements can be fulfilled by interaction terms that combine powers and exponentials of the cosmic scale factor .
the paper is organized as follows . in sec .
[ interacting fluid dynamics ] we introduce the basic interacting fluid model and study the influence of the coupling , described by an initially arbitrary function of the scale factor , on the cosmic acceleration equation . sec .
[ transient ] investigates two interaction types between dm and de , each of them characterized by two parameters , in detail .
conditions for the existence of transient acceleration are found under the condition of a vanishing total effective cosmological constant . in sec .
[ sn ] we use a statistical analysis of the gold sample of the snia data @xcite to find the admissible parameter ranges for transient cosmic acceleration .
the best - fit models for both interactions are shown to be compatible with the @xmath0cdm model .
our conclusions are summarized in sec .
[ conclusions ] .
we model the present cosmic substratum as a mixture of three components : dark energy ( de ) with an energy density @xmath1 , dark matter ( dm ) with an energy density @xmath2 and baryonic matter with an energy density @xmath3 .
the friedmann equation in a spatially flat universe is @xmath4 where @xmath5 is the hubble parameter and @xmath6 is the scale factor of the robertson - walker metric .
the de component is supposed to be characterized by an equation of state @xmath7 .
we admit an interaction between de and dm according to @xmath8 where @xmath9 symbolizes the interaction , and a dot represents a derivative with respect to the cosmic time .
the baryon component is separately conserved , @xmath10 the split of the total energy density @xmath11 into @xmath12 is consistent with the energy conservation for @xmath11 .
we assume , without loss of generality , that the energy density of the dust fluid can be written as : @xmath13 where @xmath14 and @xmath15 are constants , and @xmath16 is an arbitrary time - dependent function .
it follows from ( [ interacao ] ) and ( [ materia ] ) that @xmath17 let us write @xmath16 as @xmath18 such , that the deviation from the non - interacting case is encoded in the function @xmath19 . with @xmath20
we obtain @xmath21 for @xmath2 it follows that @xmath22 denoting the value of @xmath23 at @xmath24 by @xmath25 , both initial values are related by @xmath26 where @xmath27 .
the quantity @xmath25 is the value of @xmath2 at @xmath28 in the presence of the interaction , @xmath29 is the value of @xmath2 at @xmath28 for vanishing interaction .
the interaction re - normalizes the present value of @xmath2 . according to ( [ q2 ] ) , a positive value of @xmath9 , i.e. , an energy transfer from dark energy to dark matter requires @xmath30 in an expanding universe .
for @xmath31 the transfer is in the opposite direction , i.e. , @xmath32 . with ( [ q2 ] ) ,
the energy density of the @xmath33-component is determined by @xmath34 for a constant eos parameter @xmath35 the solution of ( [ drx ] ) is @xmath36 integrating by parts yields @xmath37 for a given interaction @xmath38 , the hubble rate in eq .
( [ fried ] ) is then determined by the sum of @xmath3 from ( [ ba ] ) , @xmath2 from ( [ rm ] ) and @xmath3 from ( [ rx5 ] ) .
the acceleration equation @xmath39\ \label{acc}\ ] ] takes the form @xmath40\right\ } \ . \label{dda}\end{aligned}\ ] ] it can also be written as @xmath41\right . \nonumber\\ & & \left .
\qquad\qquad + 3w\,\tilde\omega_{m_0}\left(\frac{a_0}{a}\right)^{3}\,\left[\left(1 + 3w\right ) a^{-3w}\,\int_{a_{0}}^{a } da\,g\,a^{3w - 1 } - g\right]\right\ } \ , \label{q}\end{aligned}\ ] ] where we have introduced the parameters @xmath42 , @xmath43 , @xmath44 and @xmath45 . for any specific model @xmath38 , formula ( [ q ] ) describes the dynamics of @xmath46 through the cosmic history , as long as our three - component model makes sense .
the present value of the deceleration parameter is @xmath47 where @xmath48 . in a spatially flat universe with @xmath49 , eq .
( [ q01 ] ) reduces to @xmath50 notice that there is no _ direct _ influence of the interaction on the present value of the deceleration parameter @xcite .
lacking any knowledge about the nature of dark energy and that of dark matter , there is no well - motivated choice for the interaction either .
our strategy here is as follows .
since we try to model a transient accelerating universe , we postulate a mathematically tractable interaction which presumably is compatible with such a dynamics .
we solve this dynamics and clarify whether there exist parameter combinations that are favorable for a scenario of transient acceleration .
it will turn out that for the two - parameter interactions considered below , there are indeed such ranges .
subsequently , we compare the result with the snia observational data .
we shall consider couplings for which the function @xmath38 is represented by a combination of powers and exponentials of the scale factor .
in particular , we investigate the following two types of interaction terms : @xmath51 and @xmath52 , where @xmath53 is a positive integer and @xmath54 is a positive real number .
although these choices may seem arbitrary , they will admit a suitable parameter combination and they will provide us with a transparent picture concerning the potential role of interactions within the dark sector in a scenario of transient acceleration .
in the first case , i.e. , for @xmath51 , assuming that @xmath55 is a natural number , the energy density ( [ rx5 ] ) is given by @xmath56\ , \label{rh1}\end{aligned}\ ] ] where @xmath57\ \label{reff}\ ] ] and @xmath58 . from now on we shall put @xmath59 .
the expression ( [ dda ] ) specifies to @xmath60\nonumber \\&+&4\pi g w k_1 \exp ( -a/\sigma ) \left[(1 + 3w)\sum_{i=0}^{n+3w-1}\frac{(n+3w-1)!}{i!}\sigma^{n+3w - i}a^{i-3(1+w ) } + a^{n-3}\right]\ .\nonumber\\ \label{ddah1}\end{aligned}\ ] ] let us now analyze the special case @xmath61 and @xmath62 for which the sum in ( [ rh1 ] ) and ( [ reff ] ) has only two terms .
using @xmath63 , the matter density ( [ rm ] ) reduces to @xmath64\ , \ ] ] and the density ( [ rh1 ] ) of the @xmath33 fluid becomes @xmath65 with @xmath66\ .\ ] ] for a vanishing interaction the @xmath33 component coincides with a cosmological constant . for the acceleration equation ( [ ddah1 ] )
one finds @xmath67\right\}\ ] ] with @xmath68 and @xmath69 .
for an interaction @xmath52 and @xmath61 and @xmath62 it follows analogously that @xmath70\ , \ ] ] and @xmath71 where @xmath72 and @xmath73\ .
\label{reffh2}\ ] ] again , in the interaction - free limit @xmath74 we have @xmath75 const
. the quantity @xmath76 can be seen as an effective cosmological constant which is re - normalized compared with the bare " value , corresponding to @xmath77 , due to the presence of an interaction .
the ratio @xmath46 becomes @xmath78\right\}\ ] ] with @xmath79 and @xmath69 . to have
a viable cosmological model , formulas ( [ accomega1 ] ) and ( [ accomega2 ] ) should admit a transition from @xmath80 to @xmath81 before the present time , i.e. , for @xmath82 .
if , moreover , the accelerated expansion is a transient phenomenon , there should be a change back from @xmath81 to @xmath80 at a future time , i.e. , for @xmath83 . in both the expressions ( [ accomega1 ] ) and ( [ accomega2 ] ) the @xmath84 terms on
the right hand sides dominate for small values of @xmath6 , i.e. , there is decelerated expansion for @xmath85 provided the conditions @xmath86 or @xmath87 are satisfied .
these conditions put upper limits on the interaction strength . in the non - interacting limit
they just express the positivity of the total matter energy density .
let s consider now the case @xmath88 .
the dominating contributions in the braces on the right - hand sides of ( [ accomega1 ] ) and ( [ accomega2 ] ) are then given by the constant terms @xmath89 .
_ as long as @xmath90 we will have @xmath81 for @xmath88 , i.e. , there is no transition back to decelerated expansion_. this holds , in particular , in the non - interacting limit which reproduces the @xmath0cdm model .
then @xmath91 reduces to @xmath92 , equivalent to @xmath93 . for @xmath88 this term will always dominate the dynamics . an obvious way to obtain decelerated
expansion for @xmath88 is to put @xmath94 both in ( [ accomega1 ] ) and ( [ accomega2 ] ) .
this corresponds to a vanishing total cosmological constant . in other words , part of
the interaction cancels the bare " cosmological constant , described by @xmath95 . under this condition
it is exclusively the remaining part of the interaction which potentially can trigger a period of accelerated expansion . in such a case one
obtains for the exponential interaction from ( [ rh1 ] ) @xmath96 , \label{r10o}\ ] ] while from ( [ rh2 ] ) for the gaussian interaction @xmath97\ \label{r20o}\ ] ] is valid . adding up the energy contributions @xmath3 from ( [ ba ] ) , @xmath2 from ( [ rhom1 ] ) and @xmath98 with @xmath99 from ( [ rh1 ] )
, we find for the hubble function @xmath100 with @xmath3 from ( [ ba ] ) , @xmath2 from ( [ rhom2 ] ) and @xmath98 with @xmath99 from ( [ rh2 ] ) , the corresponding expression for the second interaction is @xmath101 the acceleration equations ( [ accomega1 ] ) and ( [ accomega2 ] ) simplify to @xmath102\right\}\ ] ] and @xmath103\right\ } \ , \ ] ] respectively .
moreover , with @xmath104 with @xmath105 from ( [ r10o ] ) or ( [ r20o ] ) , their final forms are @xmath106}{a^{3 } } + 3 \bar{k}_1 \exp ( -a/\sigma ) \left[a^{2 } - 2\sigma^{2 } - 2\sigma a\right]\right\}\ ] ] and @xmath107\right\ } \ , \ ] ] respectively . to have decelerated expansion for @xmath85 , @xmath108 < \frac{1}{3}\ ] ] or @xmath109 has to be required . these conditions are equivalent to ( [ dec1 ] ) and ( [ dec2 ] ) , respectively . the zeros of ( [ accomega1fin ] ) and ( [ accomega2fin ] ) determine the values @xmath110 of @xmath6 at which transitions between decelerated and accelerated expansion ( or the reverse ) occur , namely @xmath111 + 3 \bar{k}_{1}\ , a_{q}^3\,\exp ( -a_{q}/\sigma ) \left[2\sigma^{2 } + 2\sigma a_{q } - a_{q}^2\right ] = 1 \ \label{consist1}\ ] ] and @xmath112 = 1 \ , \label{consist2}\ ] ] respectively .
the conditions to have acceleration at the present epoch with @xmath113 are @xmath114 > \frac{1}{9 } \ \label{condac1b}\ ] ] and @xmath115 >
\frac{2}{9 } \ . \label{condacc1}\ ] ] if the inequalities ( [ condac1b ] ) or ( [ condacc1 ] ) hold , we may have present acceleration under the condition @xmath99 , i.e. , a vanishing total cosmological constant . obviously
, the normalized interaction strengths @xmath116 or @xmath117 have to be larger than a threshold value to realize this configuration .
the conditions ( [ condac1b ] ) or ( [ condacc1 ] ) are consistent with ( [ q0x ] ) if the latter is combined with either ( [ r10o ] ) or ( [ r20o ] ) , respectively . on the other hand
, we have the upper limits ( [ condll1 ] ) and ( [ condll2 ] ) .
this means , there exists a range for admissible values of the interaction strength , determined by @xmath118 for the first case and @xmath119 for the second case .
it is instructive , to compare the relations ( [ accomega1fin ] ) and ( [ accomega2fin ] ) with the corresponding expression @xmath120 of the @xmath0cdm model .
obviously , the interaction terms in ( [ accomega1fin ] ) and ( [ accomega2fin ] ) may play the role of @xmath121 in ( [ acclcdm ] ) .
now we know , that the @xmath0cdm model provides a fairly good description of the present universe , i.e. , for @xmath113 .
therefore , we do not expect a radically different scenario around the present epoch , based on any alternative model .
this suggest positive values of the interaction constants @xmath116 and @xmath117 together with @xmath122 .
it will turn out that these models are indeed favored . from ( [ q2 ] )
it is obvious , that in an expanding universe ( @xmath123 ) the quantity @xmath124 should be positive to have an energy transfer from dark energy to dark matter .
positive values of @xmath9 are preferred on thermodynamical grounds @xcite .
( but see @xcite . ) for the case @xmath125 we have @xmath126 it follows that for @xmath62 and for @xmath127 , equivalent to @xmath128 , @xmath129 for @xmath130 @xmath131 is valid .
this means , for @xmath62 and for @xmath132 , equivalent to @xmath133 , @xmath134 consequently , for sufficiently large values of the parameter @xmath54 , positive interaction constants @xmath116 and @xmath117 are both thermodynamically favored and provide a description of the present cosmological epoch which is similar to the well - tested @xmath0cdm model ( see the comment following eq .
( [ acclcdm ] ) ) . with @xmath61
our setting is reminiscent of models with a decaying cosmological constant " .
notice that for future values @xmath83 for which the conditions ( [ q1 > ] ) or ( [ q2 > ] ) are violated , the interaction is no longer operative .
the explicit forms of the interaction term @xmath9 in the balances ( [ interacao ] ) become @xmath135 for the first interaction and @xmath136 for the second one .
these terms may be written as @xmath137 with @xmath138 for the first interaction and as @xmath139 with @xmath140 respectively , for the second interaction .
interactions of the type @xmath141 with an interaction constant @xmath142 have been frequently used to describe interactions between dark matter and dark energy @xcite .
here we have generalized this structure to a variable interaction parameter . for both the cases ( [ b1 ] ) and ( [ b2 ] ) this parameter vanishes for small values of the scale factor , i.e. , the interaction is switched on during the cosmological evolution . in the late time
limit both @xmath143 and @xmath144 vanish because @xmath145 itself vanishes . in this subsection
we point out that there also exists an equivalent scalar field representation for the dynamics of the interacting @xmath33 component . rewriting the balance equation for the @xmath33 component ( cf .
( [ interacao ] ) ) as @xmath146 with @xmath147 for the first interaction and introducing the correspondence @xmath148 we find @xmath149 and , via @xmath150 an effective potential @xmath151 \,\exp ( -a/\sigma ) \ .
\label{v1(a)}\ ] ] the kinetic part vanishes for very small and very large values of the scale factor .
the potential starts with a constant initial value @xmath152 at @xmath85 and also tends to zero for @xmath88 .
( [ dphi21 ] ) is equivalent to @xmath153^{1/2}}{\frac{h}{h_{0 } } } \label{dphi1da}\ ] ] with @xmath154 from ( [ h1 ] ) .
the solution @xmath155 of ( [ dphi1da ] ) together with ( [ v1(a ) ] ) yields an implicit representation of @xmath156 .
the corresponding relations for the second interaction are @xmath157 and @xmath158 \,\exp ( -a^{2}/\sigma^{2 } ) \
\label{v2(a)}\ ] ] with a similar behavior as in the previous case .
( [ dphi22 ] ) may be written @xmath159^{1/2}}{\frac{h}{h_{0}}}\ , \label{dphi2da}\ ] ] now with @xmath154 from ( [ h2 ] ) .
both potentials can be determined , as function of @xmath160 , numerically .
the potentials can approximately be expressed by a finite series in terms of power and @xmath161 functions .
for example , the series can have the form @xmath162 for the first interaction we find @xmath163 , @xmath164 and @xmath165 , while for the second interaction we have @xmath166 , @xmath167 and @xmath168 .
figure ( [ potential ] ) displays the numerical interpolation and the approximate analytical expression for each of the cases . left panel : effective potential of the scalar field representation for the first interaction , right panel : effective potential of the scalar field representation for the second interaction .
the continuous lines represent the numerical results , the dashed lines the interpolation in terms of power and @xmath161 functions . ]
left panel : effective potential of the scalar field representation for the first interaction , right panel : effective potential of the scalar field representation for the second interaction .
the continuous lines represent the numerical results , the dashed lines the interpolation in terms of power and @xmath161 functions . ]
to constraint the free parameters of the model , we use the gold sample of the snia data @xcite .
the crucial quantity to be evaluated is the luminosity distance @xmath169 where @xmath170 is given either by ( [ h1 ] ) or by ( [ h2 ] ) .
the quantity @xmath171 in these expressions is the hubble parameter today .
the observational data are expressed in terms of the moduli distance @xmath172 from this quantity the statistical function @xmath173 is constructed .
it is defined by @xmath174 where @xmath175 is the predicted theoretical value for the moduli distance for the @xmath176 supernova , obtained by using the model developed above , @xmath177 is the corresponding observational data , @xmath178 being its observational error bar . with the help of @xmath173
we construct the probability density function ( pdf ) @xmath179 where @xmath180 denotes the set of free parameters of the model ; @xmath181 is a normalization constant .
the general description of the bayesian analysis can be found in reference @xcite . at
first we obtain the results for the @xmath0cdm model which is given by the non - interacting limits @xmath182 and @xmath183 for the specific configurations considered here .
the @xmath0cdm model has only two free parameters : the density parameter @xmath184 and the hubble parameter @xmath185 . minimizing @xmath173
, we find @xmath186 with @xmath187 and @xmath188 . the two - dimensional pdf for these parameters
is shown in figure [ lcdm ] ( left ) .
after marginalization , we find the one - dimensional pdfs for @xmath189 and @xmath185 , also displayed in figure [ lcdm ] ( center and right , respectively ) .
after marginalization , the maximum probabilities occur for @xmath190 and @xmath191 . in the interacting model
there are generally four free parameters to be constrained : @xmath192 , @xmath193 or @xmath194 , the hubble parameter @xmath185 and @xmath54 .
but our transient scenario implies the relations ( [ r10o ] ) or ( [ r20o ] ) between @xmath192 and @xmath195 or @xmath196 , respectively .
this reduces the number of free parameter to three : @xmath195 or @xmath196 , @xmath185 and @xmath54 . integrating over one ( two ) of them
we can obtain the two - dimensional ( one - dimensional ) pdfs for the remaining variable(s ) . from now on ,
we will restrict ourselves to the three - dimensional phase space case .
for the first interaction ( interaction strength parameter @xmath116 ) the relevant equations are ( [ h1 ] ) and ( [ accomega1fin ] ) .
the corresponding relations for the second interaction ( interaction strength parameter @xmath117 ) are ( [ h2 ] ) and ( [ accomega2fin ] ) . for the first interaction
we obtain the two - dimensional and one - dimensional pdfs shown in figures [ 2dbis ] and [ 1dbis ] .
the best fit is obtained for @xmath197 , @xmath198 and @xmath199 with @xmath200 , corresponding to @xmath201 .
the evolution of the acceleration parameter for this set of values is shown in the lower right panel of figure [ 2dbis ] .
it is obvious that after an intermediate accelerated expansion the universe reenters a decelerating phase .
after marginalization , the peaks of the one - dimensional pdfs occur for @xmath202 , @xmath203 and @xmath204 .
these values imply @xmath205 and again a temporary period of accelerated expansion .
the corresponding evolution of the acceleration parameter is displayed in the lower right panel of figure [ 1dbis ] .
the results for the second interaction are very similar . in this case
we find the two - dimensional and one - dimensional pdfs shown in figures [ 2d ] and [ 1d ] .
the best - fit values are @xmath188 , @xmath206 and @xmath207 , with @xmath200 , corresponding to @xmath208 .
the evolution of the acceleration parameter is shown in the lower right panel of figure [ 2d ] .
after marginalization , the peaks of the one - dimensional pdfs are located at @xmath202 , @xmath209 and @xmath210 , implying @xmath211 .
the behavior of the acceleration parameter for this set of values is displayed in the lower right panel of figure [ 1d ] , confirming again that the accelerated expansion of the universe is a transient phenomenon .
we recall that all the cases of figures [ 2dbis ] - [ 1d ] correspond to an energy transfer from dark energy to dark matter . the @xmath0cdm model .
left panel : the two - dimensional probability distribution for @xmath185 and @xmath189 .
the brighter the color , the higher the probability .
center panel : one - dimensional probability distribution function for @xmath189 .
right panel : one - dimensional probability distribution function for @xmath185 . ] the @xmath0cdm model . left panel : the two - dimensional probability distribution for @xmath185 and @xmath189 .
the brighter the color , the higher the probability .
center panel : one - dimensional probability distribution function for @xmath189 .
right panel : one - dimensional probability distribution function for @xmath185 . ] the @xmath0cdm model . left panel : the two - dimensional probability distribution for @xmath185 and @xmath189 .
the brighter the color , the higher the probability .
center panel : one - dimensional probability distribution function for @xmath189 .
right panel : one - dimensional probability distribution function for @xmath185 . ]
two - dimensional probability distributions for different combinations of @xmath193 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
two - dimensional probability distributions for different combinations of @xmath193 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
two - dimensional probability distributions for different combinations of @xmath193 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
two - dimensional probability distributions for different combinations of @xmath193 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
one - dimensional probability distributions for @xmath193 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 . ]
one - dimensional probability distributions for @xmath193 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 . ]
one - dimensional probability distributions for @xmath193 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 . ]
one - dimensional probability distributions for @xmath193 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 . ]
two - dimensional probability distributions for the different combinations of @xmath194 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
two - dimensional probability distributions for the different combinations of @xmath194 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
two - dimensional probability distributions for the different combinations of @xmath194 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
two - dimensional probability distributions for the different combinations of @xmath194 , @xmath54 and @xmath185 after marginalization over the remaining parameter .
the lower right panel shows the evolution of the acceleration parameter for the best fit scenario in units of @xmath212 . ]
one - dimensional probability distributions for @xmath194 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 . ]
one - dimensional probability distributions for @xmath194 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 . ]
one - dimensional probability distributions for @xmath194 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 .
] one - dimensional probability distributions for @xmath194 , @xmath54 and @xmath185 after marginalization over the two remaining parameters .
the lower right panel shows the evolution of the acceleration parameter for the highest peak probability values in units of @xmath212 . ]
in a homogeneous and isotropic cosmological background , a fluid with an equation of state @xmath61 is dynamically equivalent to a cosmological constant . according to the @xmath0cdm model , a small ( in planck units ) , but so far theoretically unexplained , cosmological
constant is responsible for the observed accelerated expansion of our present universe .
we have demonstrated that an interaction between such type of dark energy and dark matter re - normalizes this constant . even if the resulting effective cosmological constant
is assumed to be zero , the time - dependent part of the interaction ( the part that does not contribute to the re - normalization ) was shown to be able to generate a phase of accelerated expansion of the universe .
we have studied two types of interactions for which cosmic acceleration is a transient phenomenon .
while recent investigations favor a scenario in which the acceleration of the universe is already slowing down today @xcite , our toy model predicts the maximal acceleration to occur at a future time .
the @xmath0cdm model is unable to provide any type of transient acceleration dynamics .
a statistical analysis of the snia data shows that models of transient accelerated expansion may well be competitive with the standard cosmological constant model .
while the mentioned cancelation mechanism of the bare " cosmological constant is purely phenomenological , we believe that it might possibly indicate a more fundamental feature .
a vanishing total effective cosmological constant seems more appealing and easier to understand within an underlying basic theory than a small , non - vanishing value of this quantity . in general , our conclusion concerning the potentially transient character of the cosmic acceleration as the result of an interaction in the dark sector coincides with the numerical results of @xcite .
but we think that an analytic solution , even though it is the solution of a toy model , provides additional information about the specific role of the interaction in this scenario and thus gives rise to a physically more transparent picture .
99 a. shafieloo , v. sahni and a.a .
starobinsky , phys .
d * 80 * , 101301 ( 2009 ) , arxiv:0903.5141 .
et al . _ ,
j. * 116 * , 1009 ( 1998 ) ; s. perlmutter _
et al . _ ,
j. * 517 * , 565 ( 1999 ) ; j.l .
et al . _ ,
j. * 594 * , 1 ( 2003 ) ; m.v .
john , astrophys .
j. * 614 * , 1 ( 2004 ) ; p. astier _
et al . _ , j. astron
. astrophys . *
447 * , 31 ( 2006 ) ; a.g .
et al . _ , astro - ph/0611572 ; d. rubin _ et al . _ , arxiv:0807.1108 ; m. hicken _
_ , astrophys.j .
* 700 * , 1097 ( 2009 ) , arxiv:0901.4804 .
m. colless _
et al . _ , mon . not .
328 * , 1039 ( 2001 ) ; m. tegmark _
d * 69 * , 103501 ( 2004 ) ; s. cole _ et al .
_ , mon . not
. soc . * 362 * , 505 ( 2005 ) ; v. springel , c.s .
frenk , and s.m.d .
white , nature ( london ) * 440 * , 1137 ( 2006 ) .
w. zimdahl , d. pavn and l.p .
chimento , phys .
b * 521 * , 133 ( 2001 ) ; l.p .
chimento , a.s .
jakubi , d. pavn , and w. zimdahl , phys .
d * 67 * , 083513 ( 2003 ) ; d. pavn , s. sen and w. zimdahl , jcap 05(2004)006 .
l. amendola , phys .
d * 62 * , 043511 ( 2000 ) ; l. amendola , phys . rev .
d * 69 * , 103524 ( 2004 ) ; rong - gen cai and anzhong
wang jcap 03(2005)002 ; zong - kuan guo , rong - gen cai and yan - zhong zhang , jcap 05(2005)002 ; w. zimdahl , int . j. mod .
d * 14 * , 2319 ( 2005 ) ; l.p .
chimento and d. pavn , phys .
d * 73 * , 063511 ( 2006 ) ; m.s .
berger and h. shojaei , phys .
d * 74 * , 043530 ( 2007 ) ; o. bertolami , f.g .
pedro and m. le delliou , gen.rel.grav .
* 41 * , 2839 ( 2009 ) , arxiv:0705.3118 ; a. w. brookfield , c. van de bruck and l.m.h .
hall , phys.rev.d * 77*,043006 ( 2008 ) , arxiv:0709.2297 ; l.p .
chimento , m. forte and g.m .
kremer , gen.rel.grav .
* 41 * , 1125 ( 2009 ) , arxiv:0711.2646 ; ch.g .
bhmer , g. caldera - cabral , r.lazkoz and r. maartens , phys.rev.d * 78 * , 023505 ( 2008 ) , arxiv:0801.1565 ; m. quartin , m.o .
calvo , s.e .
jors , r.r.r .
reis and i. waga , jcap * 0805 * , 007 ( 2008 ) , arxiv:0802.0546 ; j.f .
jesus , r.c .
santos , j.s .
alcaniz and j.a.s .
lima , phys.rev.d * 78 * , 063514 ( 2008 ) , arxiv:0806.1366 ; jian - hua he , bin wang , pengjie zhang , phys.rev.d * 80 * , 063530 ( 2009 ) , arxiv:0906.0677 .
zong - kuan guo , n. ohta and s. tsujikawa , phys .
d * 76 * , 023508 ( 2007 ) ; qiang wu , yungui gong , anzhong wang and j. s. alcaniz , phys.lett .
* b659 * , 34 ( 2008 ) , arxiv:0705.1006 ; g. olivares , f. atrio - barandela and d. pavn , phys.rev.d * 77 * , 063513 ( 2008 ) , arxiv:0706.3860 ; r. bean , . .
flanagan , istvan laszlo and mark trodden , phys.rev.d * 78 * , 123514 ( 2008 ) , arxiv:0808.1105 . | recent investigations seem to favor a cosmological dynamics according to which the accelerated expansion of the universe may have already peaked and is now slowing down again @xcite . as a consequence
, the cosmic acceleration may be a transient phenomenon .
we investigate a toy model that reproduces such a background behavior as the result of a time - dependent coupling in the dark sector which implies a cancelation of the bare " cosmological constant . with the help of a statistical analysis of supernova type ia ( snia ) data
we demonstrate that for a certain parameter combination a transient accelerating phase emerges as a pure interaction effect . |
the study of semiconductor spintronics@xcite has caused a lot of attention since long spin lifetime and large spin transport distance have been observed in @xmath0-type semiconductors,@xcite due to the great potential of the device applications such as quantum memory devices , spin valves and spin transistors .
high spin injection efficiency and long spin diffusion / transport distance are two prerequisites of realizing these devices and have been widely studied both experimentally@xcite and theoretically.@xcite many works are focused on improving the spin injection efficiency by electrical method.@xcite others are concentrated on the spin diffusion / transport inside the semiconductors without taking care of the injection . @xcite in the latter case , the spin polarization can be prepared by optical excitations.@xcite a number of theories have been developed to study the spin transport , such as the two - component drift - diffusion model,@xcite the kinetic spin bloch equation approach,@xcite the monte - carlo simulation,@xcite and the microscopic semiclassical approach.@xcite in these theories , weng and wu showed that the correlations between the spin - up and -down states , _
i.e. _ , the off - diagonal terms of the density matrix in spin space , play an essential role in spin diffusion / transport.@xcite by constructing the kinetic spin bloch equations by means of the non - equilibrium green function method and numerically solving these equations with the scattering explicitly included , they predicted spin oscillations along the spin diffusion in the absence of the magnetic field@xcite which can not be obtained from the two - component drift - diffusion model .
these oscillations were later observed in experiments.@xcite now most of the theoretical works include the off - diagonal terms.@xcite another important consequence of the off - diagonal terms is that they allow spin to precess along the effective magnetic field which origins from the spin - orbit coupling ( soc ) and is momentum dependent . the fact that spin with different momentum precesses with different frequency is referred to as inhomogeneous broadening.@xcite it was shown by wu _
_ in the spacial uniform systems that in the presence of the inhomogeneous broadening , any spin conserving scattering , including the coulomb scattering , can cause irreversible spin relaxation / dephasing ( r / d).@xcite later weng and wu extended this concept to the spacial inhomogeneous systems and showed that the spin with different momentum precesses with different frequencies during the diffusion along the spacial gradient , thanks to the off - diagonal term.@xcite this serves as additional inhomogeneous broadening and can cause additional spin r / d combined with the spin conserving scattering.@xcite in our previous works , we have studied both the transient@xcite and the steady - state@xcite spin diffusion / transport in gaas quantum well ( qw ) using the kinetic spin bloch equations , first without@xcite and then with@xcite the coulomb scattering . the other scattering such as electron - phonon and electron - impurity scattering is explicitly included . by solving the kinetic spin bloch equations combined with the poisson equation , one is able to obtain the mobility , diffusion length and spin injection length without any fitting parameter . moreover , this approach is valid not only for systems near the equilibrium , but also for those far away from the equilibrium such as systems under strong external electric field ( hot - electron effect ) and/or with large spin polarization .
it is also applicable to systems in both the strong scattering regime and the weak one.@xcite nevertheless , in our previous investigation of the steady - state spin transport,@xcite the boundary conditions we used are the single - side ones , _
i.e. _ , @xmath2 are given from the single side of the sample regardless of the sign of @xmath3 , with @xmath4 representing the density matrix elements .
this is an approximation as in principal @xmath5 for @xmath6 ( @xmath7 ) can only be determined from the right ( left ) side of @xmath8 . moreover , by using this boundary conditions , the grid of the space has to be very small due to the reason specified in the next section .
therefore the length of the sample we studied before is very short , even less than one oscillation period , as the numerical calculation is too time - consuming.@xcite it is therefore also almost impossible to discuss situations such as spin transport under strong external electric fields . in this paper
, we adopt the double - side boundary conditions which are widely used in the charge transport,@xcite _
i.e. _ , using @xmath9 and @xmath10 as boundary conditions with @xmath7 and @xmath11 standing for the right side of the sample . we further develop new numerical scheme to solve the kinetic equations with all the scattering included .
this allows us to solve the spin diffusion / transport with fast speed and high accuracy .
we therefore study the spin diffusion / transport with large sample scale under various conditions such as temperature , well width , magnetic field and electric field .
the hot - electron effect to the spin transport is also discussed in detail .
this paper is organized as follows : in sec .
ii we set up the model and construct the kinetic spin bloch equations with the boundary conditions by using the keldysh green function method . in sec .
iii we present our main results of the spin diffusion / transport with different scattering , temperature , quantum well width , external magnetic field and external electric field .
we conclude in sec .
iv . in appendix
a we give the scheme for the numerical calculation .
we begin our study from a two dimensional electron gas ( 2deg ) confined by an infinite square well with width @xmath12 along the ( 001 ) direction .
the growth direction is assumed to be along the @xmath13-axis .
a moderate magnetic field @xmath14 and an electric field @xmath15 are applied along the @xmath16-@xmath17 plane with the @xmath16-axis being the diffusion / transport direction .
the electron state is described by a subband index @xmath0 which comes from the confinement of the qw , a 2d momentum @xmath18 , which represents the momentum along the @xmath16-@xmath17plane and the spin index @xmath19 , which stands for the spin - up ( @xmath20 ) or -down ( @xmath21 ) state along the @xmath13-axis . in the present paper , the width of the qw is taken to be so small that only the lowest subband @xmath22 is occupied and the higher subbands are negligible . by using the nonequillibrium green function method with the gradient expression as well as the generalized kadanoff - baym ansatz,@xcite
we construct the kinetic spin bloch equations as follows@xcite @xmath23 here @xmath24 are the density matrices of electrons with momentum @xmath25 at position @xmath26 and time @xmath27 .
the diagonal elements @xmath28 stand for the electron distribution functions of spin @xmath19 whereas the off - diagonal elements @xmath29 represent the correlations between the spin - up and -down states .
@xmath30 are the driving terms from the external electric field . here
@xmath31\cdot\frac{\mbox{\boldmath$\sigma$\unboldmath}}{2}-e\psi(\mathbf r)+{\cal e}_{\mathtt{hf}}({\bf r},t)\ ] ] with @xmath32 denoting the effective mass .
@xmath33 is the dyakonov and perel ( dp)@xcite effective magnetic field from the dresselhaus@xcite and the rashba@xcite terms . for gaas ,
the dresselhaus term is the leading term and @xmath34 reads @xmath35 here the spin splitting parameter @xmath36 is given by @xcite @xmath37 in which @xmath38 ; @xmath39 denotes the band gap ; @xmath40 represents the spin - orbit splitting of the valence band ; @xmath32 standing for the electron mass in gaas ; and @xmath41 is a constant close in magnitude to the free electron mass @xmath42.@xcite for gaas @xmath43 ev@xmath44@xmath45 .
@xmath46 stands for the average of the operator @xmath47 over the electronic state of the lowest subband and being @xmath48 under the infinite - well - depth assumption .
@xmath49 is the hartree - fock term , _
i.e. _ the exchange term,@xcite with @xmath50 the screened coulomb potential and @xmath51 the debye - hcke screening constant . here
@xmath52 is the static dielectric constant , @xmath53 represents the 2d electron density and @xmath54 $ ] is the form factor with @xmath55 .
the hartree - fock term is important only when there is a large spin polarization.@xcite for the small spin polarization explored in this paper , it can be ignored .
@xmath56 is the electric potential which satisfies the poisson equation:@xcite @xmath57/(a\kappa_0 ) \label{eq : possion}\ ] ] with @xmath58 standing for the electron density at position @xmath59 and @xmath60 representing the background positive charge density .
the bracket @xmath61 in eq .
( [ driving ] ) is the anti - commutator .
@xmath62 are the diffusion terms which also give the additional inhomogeneous broadening due to the broadening of the spin precession frequencies for different momentum @xmath25 along the spacial gradient.@xcite the coherent terms @xmath63 and the scattering terms @xmath64 can be found in refs . and .
in the present paper it is assumed that a spin polarization is injected from the left side of the sample and diffuse / transport along the @xmath16-axis .
its spacial distribution along the @xmath17-axis is uniform .
the kinetic spin bloch equations are then simplified into @xmath65 = \left.\frac{\partial\rho_{\mathbf k}(x , t)}{\partial t}\right|_{\mathtt{scat}}\ .\end{aligned}\ ] ] here the spin - orbit coupling @xmath66 in the diffusion term is very small in comparison to @xmath67 and is ignored .
the inhomogeneous broadening can be easily seen from these equations . for the spacial uniform case , @xmath68 and the electron spin precession
is inhomogeneously broadened only due to the momentum - dependent effective magnetic field @xmath69 . with this inhomogeneous broadening , spin conserving scattering including the coulomb scattering@xcite results in an irreversible spin r / d .
moreover , the scattering can also give a counter effect to the inhomogeneous broadening.@xcite it is noted that unless the @xmath70-factor is inhomogeneously broadened,@xcite the external magnetic field itself can not cause any inhomogeneous broadening .
so there is no spin r / d when @xmath71.@xcite for the spacial inhomogeneous case , besides the spin precession around the effective magnetic field , its diffusion into the semiconductor also accompanies a spin precession with the precession frequency being @xmath72-dependent , _ i.e. _ , the third term in eq .
( [ eq : beq ] ) .
so this causes additional inhomogeneous broadening .
moreover , in the steady - state spin diffusion / transport , the spin precession rate along the diffusion direction is determined by @xmath73 according to eq .
( [ eq : beq ] ) .
therefore , differing from the spacial uniform case , the external magnetic field itself _
alone _ can cause spin r / d along the diffusion direction due to the inhomogeneous broadening of the diffusion velocity @xmath67.@xcite the kinetic spin bloch equations ( [ eq : beq ] ) together with the poisson equation ( [ eq : possion ] ) are highly nonlinear and have to be solved numerically . in our previous paper,@xcite
this set of equations are solved with single - side boundary conditions .
in such a method , an extra parameter , so called the coefficient of viscosity,@xcite is introduced to keep the smoothness of the solution and the poisson equation is inevitable to keep the system from divergence even though there are no charge imbalance and external electric field .
this can be seen by retaining the third terms in eq .
( [ eq : beq ] ) and replacing the scattering terms by the relaxation time approximation : @xmath74 with @xmath75 representing the density matrices in the equilibrium .
it is clearly seen from this equation that @xmath76 is divergent when @xmath77 for @xmath6 . in order to circumvent this divergence
, the poisson equation has to be applied and the grid for the space has to be very small . in the present investigation we extend the conventional numerical scheme of solving the boltzmann equation for charge transport@xcite to solve the kinetic spin bloch equations by using the double - side boundary conditions .
this conditions assume a finite sample length @xmath11 and are given by @xmath78 here @xmath79^{-1}$ ] represent the drifted fermi distributions in the spin space , with the chemical potential matrices @xmath80 and @xmath81 if the spin injection is from the left side and @xmath11 is assumed to be large enough so that the spin polarization vanishes before it reaches the right boundary .
@xmath82 is energy spectrum drifted by the electric field . @xmath83 and @xmath84 are the hot electron temperature and mobility respectively which are calculated at the spacial uniform case following the method addressed in our previous paper.@xcite the initial electron distributions are given by @xmath85 with @xmath80 and @xmath86 .
the chemical potentials @xmath87 and @xmath88 are determined by the equations @xmath89 = n(x)\ , \\ & & \sum_{\mathbf k}\mbox{tr}[\rho_{\mathbf k}(0 , 0)\sigma_z]=pn(0)\ , \label{p}\end{aligned}\ ] ] with @xmath90 denoting the spin polarization at the boundary @xmath8 .
the double - side boundary conditions are in good consistence with the physical consideration that the electron must move along its momentum direction .
for the poisson equation , its boundary conditions are set to be @xmath91 and @xmath92 with @xmath15 representing the strength of the external electric field .
the numerical scheme for the scattering terms has been laid out in detail in our previous paper@xcite and the new numerical scheme for solving the kinetic spin bloch equations is given in appendix [ method ] . by using this new scheme , we further remove the extra parameter the coefficient of viscosity .
we numerically solve the kinetic spin bloch equations ( [ eq : beq ] ) together with the poisson equation ( [ eq : possion ] ) iteratively to achieve the self - consistent solution .
we include all the scattering , _
i.e. _ , the electron - electron coulomb , the electron - phonon and the electron - impurity scattering
. as we are interested in the diffusion / transport properties at high temperatures ( @xmath93 k ) , for the electron - phonon scattering we only need to consider the electron longitudinal - optical ( lo ) phonon scattering.@xcite all matrix elements of the interactions and the expressions of the scattering terms can be found in ref . .
the width of the qw and the electron density are taken to be @xmath94 nm and @xmath95 @xmath96 separately unless otherwise specified .
the initial spin polarization of the left boundary is @xmath97 . in the numerical calculation
, we divide the momentum @xmath98-space into @xmath99 control regions as plotted in fig . 8 in ref . and the corresponding center points are @xmath100 with @xmath101 and @xmath102 . in the present work ,
we take @xmath103.@xcite the energy cut is determined by the ratio between the distribution at the energy cut and the distribution at the zero energy is less than 0.1 % , which is around @xmath104 at @xmath105 k with @xmath106 mev being the fermi energy . the sample length is usually taken to be @xmath107 @xmath108 m , unless otherwise specified .
this length has to be taken long enough so that the spin polarization effectively vanishes at the right boundary .
the other parameters are listed in table i.@xcite .parameters used in the numerical calculations , @xmath109 is the lo phonon energy . [ cols= "
< , < , < , < " , ] three quantities are used to describe the spin diffusion / transport lengths in the steady state : spin - dependent electron density @xmath110 the incoherently summed spin coherence @xmath111 and the coherently summed spin coherence @xmath112 unlike the corresponding quantities defined in the time domain,@xcite these three quantities are defined in the spacial domain .
the irreversible spin dephasing length @xmath113 can be deduced from the decay of the envelope of @xmath114;@xcite the ensemble spin diffusion length @xmath115 is given by the decay of the envelope of @xmath116 ; whereas the spin diffusion length @xmath117 is determined from the decay of the envelope of @xmath118 .
we first show the spacial dependence of @xmath28 for different momentums @xmath72 at @xmath119 , 10 and 300 ps in fig .
[ fig : distribution](a ) with the electron - electron and the electron lo - phonon scattering included .
@xmath105 k and @xmath120 .
differing from the intuition that the spin precession should be controlled by the effective magnetic field @xmath121 , it is seen from the figure that the electron spins with different momentums present almost the identical precession period .
this is understood due to the strong scattering which enforces different spins to precess with the same frequency . in order to reveal this effect , we scale all the scattering terms with a scaling factor @xmath122 , _
i.e. _ , replacing the scattering terms in eq .
( [ eq : beq ] ) by @xmath123 .
first we consider the case without any scattering and electric field , _ i.e. _ , @xmath124 and @xmath125 .
it is then easy to obtain the analytical solution of eq .
( [ eq : beq ] ) in the steady state with the boundary conditions ( [ eq : boundary ] ) : @xmath126 by further neglecting the hf terms .
this solution is somehow different from our previous analytical result@xcite due to the different boundary conditions .
it describes the injection of the electron spin with momentum @xmath72 from both boundaries into the 2deg . for each electron with @xmath6
, it comes from the right boundary ( @xmath127 ) where there is no spin polarization .
so there is no spin precession in the 2deg .
however , for electron with @xmath7 , it comes from the left boundary where the electron spin is polarized .
the spin polarization precesses in the 2deg due to the presence of the effective magnetic field @xmath128 .
moreover it does not decay in the absence of any scattering .
it is noted from eq .
( [ eq : noscat ] ) that unlike the temporal spin precession where the precession rate is determined by @xmath129 , the spacial spin precession rate is determined by @xmath130 which strongly depends on the velocity @xmath67 . this is because in eq
( [ eq : beq ] ) the diffusion term is proportional to the velocity .
this gives strong inhomogeneous broadening along the spin diffusion.@xcite then we introduce the scattering into the system , first a small one with @xmath131 [ fig .
[ fig : distribution](c ) ] and then the normal one with @xmath132 [ fig .
[ fig : distribution](a ) ] , which are compared with the case with @xmath124 [ fig .
[ fig : distribution](b ) ] .
once the scattering term is turned on , the electron spin presents a very different evolution behavior from eq .
( [ eq : noscat ] ) .
first , electrons with @xmath6 [ curves for different wavevector @xmath133 ( eq .
( [ kmn ] ) ) with @xmath134 and @xmath135 obtain a spin polarization due to the scattering which scatters electrons with @xmath7 to those with @xmath6 and hence the spin polarization is transferred from electrons with @xmath7 .
consequently this kind of spin polarization is also built up from the left edge to the right one even though @xmath6 .
second , the scattering tends to drive the spin precession and spin diffusion to reach the same precession frequency and diffusion velocity respectively for different @xmath25 .
this can be seen from fig .
[ fig : distribution ] where @xmath136 is plotted against the position @xmath16 for different momentum @xmath25 at @xmath119 , 10 , and 300 ps ( already in the steady state ) by comparing the results with @xmath124 ( no scattering ) , @xmath131 ( weak scattering ) and @xmath132 ( normal scattering ) . in the computation
we only include the electron - electron and electron lo - phonon scattering .
it is easy to find that when there is no / weak scattering , spin precession frequencies and spin diffusion velocities with different momentum show different values .
however , when the scattering becomes stronger , the spacial spin precession frequencies and the spin diffusion velocities for different momentum @xmath25 tend to a single frequency and velocity as shown in fig .
[ fig : distribution](a ) .
the same behavior is also observed for the spin coherence @xmath137 as shown in fig .
[ fig : distribution](d ) in the steady state ( @xmath138 ps ) .
this is also the reason that the numerical results are highly accurate even for small grid numbers in the @xmath72-space .
moreover , it shows that when the scattering is strong enough , there is no interference - induced decay and consequently @xmath139 .
the similar results have been obtained in the time domain.@xcite in fact , even for the case @xmath140
@xmath96 and @xmath141 k , the intrinsic scattering is already strong enough to ensure this behavior .
spin echo experiments are needed to verify our findings here.@xcite this behavior can be understood as following : without the scattering , each spin precesses with its own frequency @xmath142 independently with the @xmath25-dependent precession frequency referred to as the inhomogeneous broadening .
scattering tends to suppress this inhomogeneous broadening and to make it a homogeneous one .
our result indicates that the intrinsic scattering ( the electron - electron and electron - phonon scattering ) is already strong enough to totally suppress the inhomogeneous broadening .
the spin polarizations for different @xmath25 all precess with single frequency @xmath143 when @xmath120 .
in the case we calculate here , @xmath144 ( @xmath145 ev@xmath44@xmath45 , @xmath146 for the non - degenerate case ) which corresponds to the spacial period @xmath147 @xmath108 m .
moreover , it is seen from the figure that besides the counter effect of the scattering to the inhomogeneous broadening , it also causes irreversible spin r / d.@xcite in the steady state , there is no spin relaxation for each @xmath25 , as shown in fig .
[ fig : distribution](b ) in the absence of any scattering .
whereas strong spin relaxation is observed in fig .
[ fig : distribution](a ) in the presence of unscaled intrinsic scattering .
finally the scattering also affects the total spin polarization at the left boundary in the steady state .
it is noted that the boundary conditions eq .
( [ eq : boundary ] ) at @xmath8 are only for @xmath7 .
the total spin polarization eq .
( [ p ] ) is determined by the contributions from both @xmath7 and @xmath6 . without scattering ,
the spin polarization for @xmath6 is always zero [ eq .
( [ eq : noscat ] ) and fig .
[ fig : distribution](b ) ] .
therefore , the total spin polarization is only one half of 5 % at the left boundary .
the scattering transfers spin polarization from states of @xmath7
. therefore , the larger the scattering , the closer the total spin polarization approaches the boundary condition : @xmath148 % for @xmath131 and @xmath149 % for @xmath150 .
we believe that the spin polarization at the left boundary can be very close to @xmath151 % for sufficiently strong scattering .
in such a case , the spin injection with double - side boundary conditions can be approximated by the single - side boundary conditions used in our previous papers.@xcite however , for small spin polarizations , the spin injection properties such as the spin diffusion length and the spin precession period are not determined by the initial spin polarization at the boundary .
as shown in the previous subsection that the scattering plays a crucial role when the electron diffuse / transport into the 2deg , we now study the effect of the scattering on the total spin signal . in fig .
[ fig : scat ] the spin - resolved electron density @xmath152 [ eq . ( @xmath153 ) ] and
the incoherently summed spin coherence @xmath114 [ eq . ( @xmath154 ) ] in the steady state are plotted against the position @xmath16 by first including only the electron lo - phonon scattering ( red curves ) , then adding the electron - electron scattering ( green curves ) and finally adding the electron - impurity scattering ( blue curves ) with @xmath155 at @xmath141 k ( a ) and 300 k ( b ) .
@xmath120 in the calculation .
it is seen from the figure that similar to the transient spin diffusion in our previous investigation,@xcite both @xmath156 and @xmath114 oscillate with the position even when there is _ no _ external magnetic field here . besides the oscillation , strong spin relaxation and dephasing are observed .
nevertheless , they present a more complicated behavior : when @xmath141 k , the spin relaxation and dephasing always decrease when a new scattering is added . however , when @xmath157 k , they become faster when the electron - electron scattering is added to the case with only the electron lo - phonon scattering but become slower when electron - impurity scattering is further added .
finally , the spin polarizations at the left boundary are 4.75 % , 4.20 % and 3.60 % at 120 k and 4.8 % , 4.7 % and 4.6 % at 300 k for the cases with all the scattering , with both the electron - electron and the electron lo - phonon scattering and with only the electron lo - phonon scattering , respectively . _
i.e. _ , the spin polarization at the left boundary always increases with the strengthening of the scattering , either the strengthening is due to the addition of more scattering or due to the increase of temperature .
this is consistent with the result in the previous subsection .
the spin precession in the steady - state spin injection in the absence of the magnetic field is in consistence with the latest experimental observation in bulk system.@xcite this kind of spacial precession is different from the spin precession in the time domain,@xcite where the spin precession can also be observed in the absence of any magnetic field in 2deg when @xmath158 k. this is because the inhomogeneous broadening in the time domain is @xmath34 with its average over a homogenous electron distribution being @xmath159 , which is too weak to sustain the scattering when the temperature is higher than 2 k. however , the spin precession in the real space during the diffusion is quite different . here
the inhomogeneous broadening is caused by @xmath160 with @xmath161 . from previous subsection , we know the precessions of different wave vector @xmath72 share the same frequency @xmath162 even with the normal intrinsic scattering .
so the total spin polarization oscillates in the space and this kind of oscillations exist even in the room temperature .
however , in the experiments by beck _
et al._,@xcite as @xmath163 in the bulk system , @xmath164 and the effective magnetic field is provided by strain which is very small .
this is the reason in bulk one has to observe the spin oscillations at very low temperature .
the complicated scattering dependence of the spin r / d in the real space is similar to that of the spin r / d in the time domain,@xcite where the scattering dependence of the spin r / d time is different in the strong and weak scattering regimes.@xcite in the weak scattering limit , adding a new scattering increases the spin r / d as it provides more spin r / d channel in the presence of the inhomogeneous broadening .
nevertheless , in the strong scattering limit , the counter effect of the scattering to the inhomogeneous broadening is important and adding a new scattering leads to a weaker spin r / d.@xcite similarly , the spin diffusion length , which is used to represent the spin relaxation in the real space , is also controlled by the inhomogeneous broadening and the scattering , with the relative strength of them dividing the system into the weak / strong scattering regime . at @xmath165 k , the electron lo - phonon scattering is very weak and falls into the weak scattering regime
. therefore the diffusion length decreases with the scattering .
but at @xmath166 k , the scattering becomes stronger .
our results indicate the system is near the transition regime between the strong and weak scattering and therefore the changes are marginal by adding new scattering . in fig .
[ fig : tempt ] the steady - state spin - resolved electron density @xmath156 and the incoherently summed spin coherence @xmath114 are plotted as functions of position @xmath16 at different temperatures @xmath141 , 200 and 300 k for qw with @xmath94 nm . in the calculation
only the intrinsic scattering is included .
one finds from the figure that the spin precession period increases with the temperature .
this is because the period is around @xmath167^{-1}$ ] . with the increase of temperature ,
@xmath168 increases .
consequently the period of the spacial spin oscillation becomes longer .
the spin diffusion length decreases with the temperature .
this is understood that with the increase of temperature , both the inhomogeneous broadening and the scattering increase . in the weak scattering regime
, this increase leads to a faster spin r / d .
nevertheless , the temperature dependence of the spin diffusion length is very mild compared to the temperature dependence of the spin precession period .
we further study the well - width dependence of the spin diffusion in the steady state in fig .
[ fig : width ] , with only the intrinsic scattering included at @xmath141 k. one finds from the figure that both the spin oscillation period and the spin diffusion length increase markedly with the well width . this is consistent with the spin r / d time in the time domain@xcite and
is understood due to the decrease of the dresselhaus spin - orbit coupling eq .
( [ hk ] ) in which the linear term is proportional to @xmath169 .
we investigate the effect of magnetic field on the spin r / d in the steady state spin diffusion .
it has been pointed out by weng and wu that the magnetic field plays a different role in spin diffusion / transport from the time - domain spin precession in the spacial uniform case.@xcite in the spin precession in the time domain , the magnetic field alone without the dp term will not cause any spin r / d due to the absence of inhomogeneous broadening.@xcite whereas in spin diffusion / transport , as the inhomogeneous broadening is caused by @xmath170 which reads @xmath171 in the absence of the dp term and still provides an inhomogeneous broadening leading to the spin r / d.@xcite furthermore , from eq .
( [ omega ] ) one finds that the inhomogeneous broadening caused by the applied magnetic field depends on the direction of the magnetic field when it is combined with the inhomogeneous broadening caused by the effective magnetic field @xmath172 .
therefore different direction of the magnetic field yields different spin diffusion length .
we reveal the magnetic field effect on the steady - state spin diffusion by plotting in fig .
[ fig : b ] the spin diffusion length @xmath117 and the spin oscillation period @xmath173 , fitted from the calculated spin polarization along the @xmath13-direction by @xmath174 , as function of the applied magnetic field at @xmath141 k. the direction of the magnetic field is either vertical ( along the @xmath17-axis ) or parallel ( along the @xmath16-axis ) to the diffusion direction .
one finds from the figure that the magnetic field leads to additional spin r / d and therefore the spin diffusion length decreases with the magnetic field , regardless of the direction of the magnetic field .
however , it is interesting to see that the spin diffusion length has different symmetry when the magnetic field is vertical or parallel to the spin diffusion direction : when @xmath175 is parallel to the @xmath17-axis , @xmath176 ; however , when @xmath175 is parallel to the @xmath16-axis , @xmath177 .
this can be understood from the fact that the density matrices from the kinetic spin bloch equations ( [ eq : beq ] ) have the symmetry @xmath178 when @xmath175 is along the @xmath17-axis .
this symmetry is broken if @xmath175 is along the @xmath16-axis .
differing from the magnetic field dependence of the spin diffusion length , the spin oscillation period @xmath173 decreases monotonically with the vertical magnetic field , regardless the direction of the field .
however @xmath173 increases with the parallel magnetic field when the field is along the diffusion direction and less than 4 t ( the spin precession disappears when the field is larger than 4 t ) but @xmath173 increases with the magnetic field when it is anti - parallel to the diffusion direction .
the spacial period is determined by @xmath179 with @xmath180 being along the @xmath16-axis .
@xmath181 represents the average of @xmath3 due to the spin gradient and can be roughly estimated by @xmath182 , which is a positive value due to the presence of the spin gradient .
for the vertical magnetic field , @xmath180 and @xmath183 are perpendicular to each other , so the magnitude of the spin oscillation period is determined by @xmath184^{-1}$ ] which always decreases with the magnetic field .
however , for the parallel one , these two vectors are in the same direction and the period is determined by @xmath185 .
so the period increases with the magnetic field .
moreover , it is interesting to see that when @xmath186 4 t , @xmath187 and further increasing of the magnetic field does not lead to any spin oscillation due to the too large inhomogeneous broadening caused by the magnetic field .
we now turn to study the electric field effect on spin transport .
an external electric field is applied along the @xmath16-axis . when the electric field is large enough ,
the spin transport is then in the hot - electron regime .
it has been shown in our previous investigation@xcite in the spacial homogeneous case that the spin precession in the time domain is markedly affected by the electric field .
especially an effective magnetic field @xmath188 , which is proportional to the electric field , is induced by the electric field combined with the dp term due to the center - of - mass drift velocity @xmath189 ( and hence @xmath190 ) driven by the electric field , with @xmath108 representing the mobility .
this effective magnetic field leads to a spin precession in the absence of any magnetic field and the precession frequency changes with the variation of the electric field.@xcite one may therefore expect that now in spin transport , the spacial spin oscillation period is also proportional to the electric field .
this is _ not _ the case as again @xmath3 in the spin diffusion terms , _
i.e. _ , the third terms in eq .
( [ eq : beq ] ) , plays an important role in the spin diffusion / transport and consequently the spacial oscillation period @xmath191 is _ independent _ on @xmath3 .
therefore , in spin transport the electric field can not induce an additional spacial spin precession .
it is further expected that in the steady state , @xmath181 is now determined by both the drift velocity and the spin gradient .
therefore , if the electric field @xmath192 ( @xmath193 ) , @xmath181 is reduced ( enhanced ) by the electric field and the spin diffusion length depends on the direction of the electric field . in fig .
[ fig : e ] the spin density @xmath194 in the steady state is plotted against the position @xmath16 at different electric fields @xmath195 , 0 , @xmath196 and @xmath197 kv / cm with only the intrinsic scattering included at @xmath141 k. it is seen from the figure that the spin oscillation period is almost unchanged when the electric field varies from @xmath197 kv / cm to 0.5 kv / cm .
this is consistent with the latest experimental observation in bulk system.@xcite it is further noticed from the figure that the spin diffusion length is markedly affected by the electric field . to reveal this effect ,
the spin diffusion length is plotted as function of electric field in the same figure .
one finds when the electric field varies from @xmath197 kv / cm to @xmath198 kv / cm , the spin diffusion length decreases monotonically .
this can be easily understood from the fact that when an electric field is applied along @xmath199-direction ( @xmath200-direction ) , the drift velocity caused by the electric field is along the @xmath16-direction ( @xmath199-direction ) , which enhances ( cancels ) the velocity driven by the spin gradient .
therefore , @xmath181 is increased ( reduced ) and the inhomogeneous broadening is decreased ( enhanced ) .
a longer ( shorter ) spin diffusion length is then observed .
in summary , we have performed a theoretical investigation on the spin diffusion / transport in @xmath0-type gaas qws at high temperatures ( @xmath1 k ) by setting up and numerically solving the kinetic spin bloch equations together with the poisson equation , with all the scattering explicitly included .
a new numerical scheme is developed which can solve the kinetic equations with high accuracy and speed .
the spin r / d mechanism in the spin diffusion / transport is due to the inhomogeneous broadening , combined with the spin conserving scattering .
differing from the spin precession in the spacial uniform system where the inhomogeneous broadening is caused only by the momentum - dependent effective magnetic field due to the soc , in spin diffusion / transport , the inhomogeneous broadening is caused additionally by the momentum dependence in the spin diffusion rate and therefore even the magnetic field alone can lead to spin r / d . due to the joint effects from the momentum - dependence in the diffusion rate and the momentum dependence in the dresselhaus effective magnetic field , there are spin oscillations along the spin diffusion even in the absence of the external magnetic field .
the period of these spacial spin oscillations can be affected by the temperature , the well width and the electron density , but is independent on the external electric field .
it increases when an additional scattering is added or the temperature / the well width is increased .
moreover , it is shown that when the scattering is strong enough , the spin oscillations of electrons at different momentum @xmath25 show the _ same _ period .
in fact , the intrinsic scattering , _
i.e. _ , the electron - electron and the electron lo - phonon scattering is already strong enough to lead to this effect .
this indicates that during the spin diffusion , there is no interference - induced spin dephasing and @xmath139 .
it is shown that the coulomb scattering makes marked contribution to the spin r / d in the spin diffusion / transport .
it is especially important in the hot - electron case as it provides the important mechanism of thermalization .
the scattering , the temperature , the qw width , the magnetic field and the electric field dependence of the spin diffusion length and the spacial spin oscillation period are explored in detail .
it is shown that in the weak scattering regime , the spin diffusion length decreases if a new scattering is added into the system , while in the strong scattering regime , it increases . in the temperature regime we are interested in , the spin diffusion length decreases with the temperature .
moreover , it increases with the increase of the qw width as the inhomogeneous broadening decreases with the well width but decreases with the external magnetic field as the later adds a new inhomogeneous broadening into the system .
the external electric field can effectively prolong or suppress the spin diffusion length depending on whether the electric field is antiparallel or parallel to the spin diffusion direction .
we believe this investigation is useful for the understanding of the spin transport as well as the design of the spintronic device .
this work was supported by the natural science foundation of china under grant nos .
90303012 and 10574120 , the natural science foundation of anhui province under grant no .
050460203 , the national basic research program of china under grant no .
2006cb922005 , the knowledge innovation project of chinese academy of sciences and srfdp .
the authors acknowledge valuable discussions with m. q. weng and m. p. zhang .
the kinetic spin bloch equations eq .
( [ eq : beq ] ) can be rewritten as : @xmath201 \label{wave}\end{aligned}\ ] ] with @xmath202 = \left.\frac{\partial\rho_{\mathbf k}(x , t)}{\partial t}\right|_{\mathtt{dr } } + \left.\frac{\partial\rho_{\mathbf k}(x , t)}{\partial t}\right|_{\mathtt{coh } } + \left.\frac{\partial\rho_{\mathbf k}(x , t)}{\partial t}\right|_{\mathtt{scat}}$ ] .
the discretization of the time and spatial derivatives in eq .
( [ wave ] ) reads @xmath203 + \frac{k_x}{m^{\ast}}\frac{\rho_{\mathbf k}(x+\delta x , t+\delta t ) - \rho_{\mathbf k}(x , t+\delta t)}{\delta x } \nonumber\\ & & = \frac{1}{2}\left(g_{\mathbf k}[x , t]+g_{\mathbf k}[x+\delta x , t ] \right)\ .
\label{dis}\end{aligned}\ ] ] for @xmath7 , the electron is propagated from its left positions to the right ones .
therefore , from eq .
( [ dis ] ) , the iterative format is @xmath204+g_{\mathbf k}[x+\delta x , t])\ .\end{aligned}\ ] ] for @xmath6 , the electron state is propagated from its right positions and the iterative format is then @xmath205+g_{\mathbf k}[x+\delta x , t])\ .\end{aligned}\ ] ] we truncate the spacial variable @xmath16 from 0 to @xmath11 . at the left edge ,
the boundary conditions are given only for the states with @xmath7 and at the right edge only for those with @xmath6 .
note the state with @xmath207 does not contribute to the diffusion process but contributes to the local source term @xmath202 $ ] . with the combined effects of the spin - flip terms and the scattering terms
, the spin signal decays when it diffuses into the qw in the scale of the diffusion length @xmath117 .
when @xmath208 , one can study the spin injection properties .
the computation is carried out in a parallel manner in the `` beowulf '' cluster . for a typical calculation with the partitions @xmath209 grid points in the @xmath210-space and @xmath211 points in the real space
, it takes about 7 hours for the system to evolute to 300 ps with a time step of 0.023 ps by 6-node amd athlon xp3000+cpu s . 0 , ed .
by d. d. awschalom , d. loss , and n. samarth ( springer - verlag , berlin , 2002 ) ; i. uti , j. fabian , and s. das sarma , rev .
* 76 * , 323 ( 2004 ) .
s. a. wolf , d. d. awschalom , r. a. buhrman , j. m. daughton , s. von molnr , m. l. roukes , a. y. chtchelkanova , and d. m. treger , science * 294 * , 1488 ( 2001 ) . j. m. kikkawa , i. p. smorchkova , n. samarth , and d. d. awschalom , science * 277 * , 1284 ( 1997 ) . j. m. kikkawa and d. d. awschalom , nature ( london ) * 397 * , 139 ( 1999 ) . j. m. kikkawa and d. d. awschalom , phys . rev .
lett . * 80 * , 4313 ( 1998 ) .
h. ohno , science * 281 * , 951 ( 1998 ) . t. adachi , y. ohno , f. matsukura , and h. ohno , physica e * 10 * , 36 ( 2001 ) .
j. m. kikkawa and d. d. awschalom , science * 287 * , 473 ( 2000 ) ; _ ibid_. * 281 * , 656 ( 2000 ) .
y. ohno , d. k. young , b. beschoten , f. matsukura , h. ohno , and d. d. awschalom , nature ( london ) * 402 * , 790 ( 1999 ) .
r. fiederling , m. keim , g. reuscher , w. ossau , g. schmidt , a. waag , and l.w .
molenkamp , nature ( london ) * 402 * , 787 ( 1999 ) .
m. a. brand , a. malinowski , o. z. karimov , p. a. marsden , r. t. harley , a. j. shields , d. sanvitto , d. a. ritchie , and m. y. simmons , phys .
* 89 * , 236601 ( 2002 ) .
j. stephens , j. berezovsky , j. p. mcguire , l. j. sham , a. c. gossard , and d. d. awschalom , phys .
93 * , 097602 ( 2004 ) .
j. strand , b. d. schultz , a. f. isakovic , c. j. palmstrm , and p. a. crowell , phys .
lett . * 91 * , 036602 ( 2003 ) .
g. schmidt , c. gould , p. grabs , a. m. lunde , g. richter , a. slobodskyy , and l. w. molenkamp , phys .
lett . * 92 * , 226602 ( 2004 ) .
m. beck , c. metzner , s. malzer , and g. h. dhler , europhys .
lett . * 75 * , 597 ( 2006 ) . s. a. crooker and d. l. smith , phys .
* 94 * , 236601 ( 2005 ) .
m. johnson , phys .
b * 58 * , 9635 ( 1998 ) .
y. takahashi , k. shizume , and n. masuhara , phys .
b * 60 * , 4856 ( 1999 ) .
m. osofsky , j. supercond .
* 13 * , 209 ( 2000 ) .
m. e. flatt and j. m. byers , phys .
84 * , 4220 ( 2000 ) .
m. q. weng and m. w. wu , phys .
b * 66 * , 235109 ( 2002 ) .
j. fabian , i. zuti and s. das sarma , phys .
b * 66 * , 165301 ( 2002 ) .
z. g. yu and m. e. flatt , phys . rev .
b * 66 * , 201202(r ) ( 2002 ) ; _ ibid_. * 66 * , 235302 ( 2002 ) .
i. martin , phys . rev .
b * 67 * , 014421 ( 2003 ) .
m. q. weng and m. w. wu , j. appl . phys . *
93 * , 410 ( 2003 ) . y. qi and s. zhang , phys .
b * 67 * , 052407 ( 2003 ) . m. q. weng , m. w. wu , and q. w. shi , phys .
b * 69 * , 125310 ( 2004 ) .
l. jiang , m. q. weng , m. w. wu , and j. l. cheng , j. appl .
phys . * 98 * , 113702 ( 2005 ) .
s. saikin , m. shen , m .- c .
cheng , and v. privman , j. appl .
* 94 * , 1769 ( 2003 ) . v.
v. osipov and a. m. bratkovsky , phys .
b * 70 * , 205312 ( 2004 ) . y. y. wang and m. w. wu , phys .
b * 72 * , 153301 ( 2005 ) .
m. hruka , .
kos , s. a. crooker , a. saxena , and d. l. smith , phys .
b * 73 * , 075306 ( 2005 ) .
s. pramanik , s. bandyopadhyay , and m. cahay , phys .
b * 73 * , 125309 ( 2006 ) .
h. dery , l. cywiski , and l. j. sham , phys . rev .
b * 73 * , 041306(r ) ( 2006 ) .
d. csontos and s. e. ulloa , physica e * 32 * , 412 ( 2006 ) ; phys . rev .
b * 74 * , 155207 ( 2006 ) .
x. jiang , r. wang , r. m. shelby , r. m. macfariane , s. r. bank , j. s. harris , and s. s. p. parkin , phys .
lett . * 94 * , 056601 ( 2005 ) .
e. i. rashba , phys .
b * 62 * , r16276 ( 2000 ) . , edited by f. meier and b.p .
zakharchenya ( north - holland , amsterdam , 1984 ) .
a. g. aronov , g. e. pikus , and a. n. titkov , zh .
84 * , 1170 ( 1983 ) [ sov .
jetp * 57 * , 680 ( 1983 ) ] .
v. pershin and v. privman , phys . rev .
b * 69 * , 73310 ( 2004 ) .
s. saikin , j. phys .
* 16 * , 5071 ( 2004 ) .
e. g. mishchenko and b. i. halperin , phys .
b * 68 * , 045317 ( 2003 ) . d. schmeltzer , a. saxena , a. r. bishop , and d. l. smith , phys .
b * 68 * , 195317 ( 2003 ) .
o. bleibaum , phys .
b * 71 * , 195329 ( 2005 ) . m. w. wu and c. z. ning , eur .
j. b * 18 * , 373 ( 2000 ) .
m. w. wu , j. phys .
70 * , 2195 ( 2001 ) .
m. q. weng and m. w. wu , phys .
b * 68 * , 075312 ( 2003 ) .
m. q. weng , m. w. wu , and l. jiang , phys .
b * 69 * , 245320 ( 2004 ) .
j. zhou , j. l. cheng , and m. w. wu , phys .
b * 75 * , 045305 ( 2007 ) . c. l , j. l. cheng , and m. w. wu , phys . rev .
b * 73 * , 125314 ( 2006 ) .
l. allen and j. h. eberly , _ optical resonance and two - level atoms
_ ( dover , new york , 1975 ) .
b. h. floyd and y. l. le coz , j. appl .
7889 ( 1996 ) . y. l. le coz , ph.d thesis , massachusetts institute of technology , 1988 . j. a. carrillo , i. m. gamba , a. majorana , c .- w .
shu , j. of comput .
phys . * 184 * , 498 ( 2003 ) .
h. haug and a.p .
jauho , _ quantum kinetics in transport and optics of semiconductors _
( springer , berlin , 1996 ) .
m. i. dyakonov and v. i. perel , zh .
* 60 * , 1954 ( 1971 ) [ sov .
jetp * 33 * , 1053 ( 1971 ) ] .
g. dresselhaus , phys . rev . * 100 * , 580 ( 1955 ) . y. a. bychkov and e. rashba , sov
jetp lett .
* 39 * , 78 ( 1984 ) ; y. a. bychkov and e. rashba , pisma zh .
. fiz . * 39 * , no . 2 , 66 ( 1984 ) .
b. g. streetman and s. banerjee , _ solid state electronic devices _
( prentice hall , englewood cliffs , nj , 2000 ) , eq .
( 5.14 ) ; k. tomizawa , _ numerical simulation of submicron semiconductor devices _ ( artech house , boston , 1993 ) , eq . ( 4.16 ) ; m. v. fischetti and s. e. laux , _ damocles theoretical manual _ , ( ibm corporation , new york , 1995 ) document no . 0.2 , eq .
j. l. cheng and m. w. wu , j. appl .
phys . * 99 * , 083704 ( 2006 ) .
m. m. glazov and e. l. ivchenko , pisma zh .
. fiz . * 75 * , 476 ( 2002 ) [ jetp lett . * 75 * , 403 ( 2002 ) ] .
sherman , appl .
. lett . * 82 * , 209 ( 2003 ) .
m. w. wu and h. metiu , phys .
b * 61 * , 2945 ( 2000 ) . t. kuhn and f. rossi , phys .
* 69 * , 977 ( 1992 ) .
p. y. yu and manuel cardona , _ fundamentals of semiconductors _ ( springe , berlin , 2003 ) .
e. m. conwell and m. o. vassel , ieee trans .
ed-*13 * , 22 - 27 ( 1966 ) .
gerald d. mahan , _ polar semiconductors _ , ed .
j. t. devreese ( north - holland , amserdam , 1972 ) ; gerald d. mahan , _ many - particle physics _ ( plenum press , new york , 1981 ) .
we also checked the division with @xmath212 and @xmath213 grids in the real space and obtained the almost identical results .
so the accuracy of the numerical calculation is high enough .
the reason that for such a rough @xmath25-mesh the results are converged lies on two facts : one is that , as shown in sec .
iii(a ) , all the density matrix elements in the @xmath25-space are in the same phase ; the other is that the system is almost in the non - degenerate regime .
therefore the spin polarization in the @xmath214-space is extended to a very large energy scale and is therefore very smooth .
, landolt - brnstein , new series , vol .
17a , ed . by o. madelung ( springer - verlag , berlin , 1987 ) . c. l , j. l. cheng , m. w. wu , and i. c. da cunha lima , phys .
a ( 2007 ) , doi:10.1016/j.physleta.2007.02.030 .
a. m. tyryshkin , s. a. lyon , w. jantsch , and f. schffler , phys .
lett . * 94 * , 126802 ( 2005 ) .
l. jiang and m. w. wu , phys .
b * 71 * , 033311 ( 2005 ) . | the spin diffusion / transport in @xmath0-type ( 001 ) gaas quantum well at high temperatures ( @xmath1 k ) is studied by setting up and numerically solving the kinetic spin bloch equations together with the poisson equation self - consistently .
all the scattering , especially the electron - electron coulomb scattering , is explicitly included and solved in the theory .
this enables us to study the system far away from the equilibrium , such as the hot - electron effect induced by the external electric field parallel to the quantum well .
we find that the spin polarization / coherence oscillates along the transport direction even when there is no external magnetic field .
we show that when the scattering is strong enough , electron spins with different momentums oscillate in the same phase which leads to equal transversal spin injection length and ensemble transversal injection length .
it is also shown that the intrinsic scattering is already strong enough for such a phenomena .
the oscillation period is almost independent on the external electric field which is in agreement with the latest experiment in bulk system at very low temperature [ europhys .
lett . * 75 * , 597 ( 2006 ) ] .
the spin relaxation / dephasing along the diffusion / transport can be well understood by the inhomogeneous broadening , which is caused by the momentum - dependent diffusion and the spin - orbit coupling , and the scattering .
the scattering , temperature , quantum well width and external magnetic / electric field dependence of the spin diffusion is studied in detail . |
the study of nucleon form factors has seen an enormous revival through the discovery by jones _
et al._@xcite , that @xmath1/@xmath4 drops almost linearly with @xmath3 above a four - momentum transfer of something like 1 gev@xmath5 . those results have stimulated huge amounts of theoretical activity , as evidenced by the nearly 500 citations of their original paper .
one approach to explaining the data involves refined perturbative qcd calculations that include an @xmath6 component in the quark light - cone wave function@xcite . also notable are relativistic constituent - quark calculations@xcite , some of which even preceded the discovery by jones _
et al . _
perhaps the most realistic model is a calculation out of argonne by clot , roberts and coworkers that uses an approach based on the dyson schwinger equations ( dses ) together with the poincar - covariant faddeev equations@xcite . here ,
the constituent quarks have their masses dynamically generated using the dse approach . while still a model , the dse / faddeev calculation from argonne offers a solution that consistently incorporates both qcd - based dynamics along with dressed diquark degrees of freedom .
also , it is well constrained by the nucleon s static properties such as mass and magnetic moment .
it is limited , however , in that there are precisely three ( and for instance , not five ) constituent quarks used as input to the calculation .
even so , it is reasonable to assume dominance of the 3-quark component of the wave function at relatively high values of @xmath3 . finally , truly _
ab initio _ calculations of form factors using lattice qcd have been performed , some of which are extrapolated to a realistic pion - mass value using chiral perturbation theory@xcite . while currently limited in their @xmath3-reach ,
it is likely that such calculations will reach higher @xmath3-values with timescales consistent with the measurements proposed here . .
on the right - hand panel , for @xmath0/@xmath2 , we show published data including those due to madey and coworkers@xcite , and preliminary results of gen(1 ) ( e02 - 013)@xcite .
we also show the projected errors of gen(2)@xcite , which is part of the super bigbite project , and e12 - 09 - 006@xcite with shms ( open blue points ) .
note that at roughly 7 gev@xmath5 , the highest value of the shms experiment , gen(2 ) achieves an error four times smaller , and with one quarter of the running time .
this illustrates the @xmath750 advantage in figure - of - merit of gen(2 ) over the shms experiment .
[ fig : ratio_ff ] , scaledwidth=100.0% ] 0.1truein . on the right - hand panel , for @xmath0/@xmath2 ,
we show published data including those due to madey and coworkers@xcite , and preliminary results of gen(1 ) ( e02 - 013)@xcite .
we also show the projected errors of gen(2)@xcite , which is part of the super bigbite project , and e12 - 09 - 006@xcite with shms ( open blue points ) .
note that at roughly 7 gev@xmath5 , the highest value of the shms experiment , gen(2 ) achieves an error four times smaller , and with one quarter of the running time .
this illustrates the @xmath750 advantage in figure - of - merit of gen(2 ) over the shms experiment .
[ fig : ratio_ff ] , scaledwidth=100.0% ] figure [ fig : ratio_ff ] shows existing data for @xmath1/@xmath4 and @xmath0/@xmath2 , the projected errors for several approved experiments , and the results of several theoretical calculations .
the approved experiments associated with the super bigbite experiment are e12 - 07 - 109 , also known as gep(5 ) , and e12 - 09 - 016 , which will measure @xmath1/@xmath4 and @xmath0/@xmath2 , respectively .
figure [ fig : ratio_ff ] makes it clear that the only way to achieve clarity in discriminating between theoretical explanations of the form - factor data is to measure the form factors with considerable precision to high values of @xmath3 in _ both the proton and the neutron_. for example , three of the predictions shown , the relativistic constituent - quark model ( rcqm)@xcite , the dse / faddeev calculation@xcite , and the refined pqcd calculation@xcite ( @xmath8 ) , all show @xmath1/@xmath4 crossing zero somewhere in the neighborhood of 7 gev@xmath5 . at the same time , the two vmd models@xcite show @xmath1/@xmath4 approaching zero much more gradually . _ in contrast _ , in the case of the neutron , even by 10 gev@xmath5 , the rcqm , pqcd and dse / faddeev calculations all differ wildly from one another . in the years following the discovery by jones _
_ , it is not surprising that models have evolved that explain well the existing proton data .
it is also not surprising that these models diverge strongly where there are little or no data to constrain the calculations , such as at higher @xmath3 for the proton or even moderate @xmath3 for the neutron .
in general , higher values of @xmath3 also offer the advantage that there are simplifications that are not present at lower values .
for instance , the role of vector mesons is suppressed at higher @xmath3 , as are higher fock states in some of the phenomenological models . at high @xmath3
there is increased clarity , and increased discovery potential , because there are generally fewer places to hide deficiencies in a model . even setting aside specific predictions , there is a crucial experimental question that is of tremendous importance .
the linearity of the data for @xmath1/@xmath4 up to something like 6 gev@xmath5 is very striking .
but to what value of @xmath3 will this linearity continue ?
looking again at fig .
[ fig : ratio_ff ] , the preliminary results from gep(3)@xcite , up to values of around @xmath9 , indicate that the rate of decline of @xmath1/@xmath4 appears to be slowing .
it is hard to be certain , however , because the errors are relatively large , larger than those projected in the original gep(3 ) proposal . by going to 14.5 gev@xmath5 ,
the trend of @xmath1/@xmath4 should become clear , _ but only if the data have sufficient precision_. to better illustrate this point , we have shown , with the dash - dotted lines , the relative size of the errors one could expect for @xmath1/@xmath4 using the _ original _ projected error for gep(3 ) at @xmath3 = 8.6 gev@xmath5 , and taking the figure - of - merit to scale as the product of the cross section @xmath10 and the square of the analyzing power a@xmath11 , @xmath12 .
we have assumed 30 days of data taking , as was assumed for the 8.6 gev@xmath5 point . with this scaling , it is clear that the projected errors for gep(4 ) represent a considerable challenge in terms of beam time . in contrast , the projected errors for gep(5 ) , which is based on super bigbite , are quite small , even out to 14.5 gev@xmath5 .
this is because the innovative design of the super bigbite approach provides fully a factor of 10 improvement in the relevant figure - of - merit .
even with the striking behavior of @xmath1/@xmath4 at the values of @xmath3 studied thus far , it is still the expectation ( from pqcd ) that this ratio should eventually level off and become constant .
the observation of a transition to this behavior would provide valuable insight , and it is important to have an experiment capable of achieving the required precision . with respect to the neutron
, there are strong motivations to measure @xmath0/@xmath2 out to @xmath3 = 10 gev@xmath5 and even higher .
first of all , at 10 gev@xmath5 , data on the neutron will already be solidly in the regime where intriguing behavior has been observed with the proton . and as pointed out earlier , in contrast to the case with the proton , the various predictions for the neutron all disagree strongly with one another for @xmath3 = 10 gev@xmath5 .
also , at @xmath3 = 10 gev@xmath5 , as is shown in fig . 1 ,
the argonne dse / faddeev calculation shows a zero crossing for @xmath0/@xmath2 , a feature due to the use of diquark degrees of freedom .
as mentioned above , the argonne calculation , while definitely still a model , contains features such as the dynamical generation of mass that suggest progress toward an actual analytical solution .
it is intriguing that this sophisticated calculation appears to be so successful , and it is desirable to test its predictions as thoroughly as possible . with this in mind
, we note that at 7 gev@xmath5 , the predictions of the vmd models and the argonne dse / faddeev calculation are fairly close to one another .
thus , measuring well above 7 gev@xmath5 should be a priority . indeed , looking again at fig . 1
, our proposed measurement of @xmath0/@xmath2 ( e12 - 09 - 016 ) provides good precision all the way to 10 gev@xmath5 . in sharp contrast , the projected errors for the only competing experiment ( e12 - 09 - 006 ) provide little discrimination between different theories even at 7 gev@xmath5 , the highest @xmath3 point for which it would provide data . the significant difference between e12 - 09 - 016 and e12 - 09 - 006 results from an intrinsic figure - of - merit associated with the super bigbite approach that is roughly a factor of 50 larger than is the case with e12 - 09 - 006 .
there are , of course , many motivations for measuring the ground - state form factors that are not illustrated in the above figure .
for instance , for both the proton and the neutron , the ground - state elastic form factors provide stringent model - independent constraints on generalized parton distributions ( gpds ) .
thus , if we want to know the gpds over a wide kinematic range , we need to study the elastic form factors over a similar range .
the elastic form factors also provide a powerful check of lattice qcd .
_ ab initio _ lattice calculations of ground - state form factors are making impressive progress , and the comparison of these results with experimental measurements is extremely important . here , it is critical to have experimental results on _ both _ the proton and the neutron , and to cover the _ full _
@xmath3-range being explored on the lattice . by the time the cebaf upgrade is complete and
the measurements described here have been made , it is quite likely that lattice results will extend up to roughly 10 gev@xmath5 , again making coverage of this range for both the proton and neutron extremely important .
-0.2truein [ fig : gen - layout ] the hall a gmn experiment will determine @xmath2 by a detailed comparison of the unpolarized elastic cross sections of the two processes @xmath13 and @xmath14 .
it will use essentially the same apparatus as gen(2 ) , with the exception that the target will be the hall a liquid deuterium cryotarget . a schematic representation of the experimental setup is shown in fig . 2
we note that the gmn proposal actually included measurements up to 18.0 gev@xmath5 , something that , combined with the fully approved hall a @xmath4 measurement ( not part of the super bigbite project ) , would enable the reconstruction of the individual @xmath15 and @xmath16 quark distributions with a spatial resolution of 0.05 fm . the emff collaboration plans to return to the pac to request an additional two weeks to push to this higher @xmath3 , as the difference between the @xmath15 and @xmath16 quark distributions is an exciting question with implications for our understanding of nucleon structure in terms of qcd degrees of freedom . like the other super bigbite project form - factor measurements , the @xmath2 measurement in hall a will provide excellent accuracy and reach in @xmath3 , well beyond all competing efforts .
considering only the portion of the experiment that is fully approved , the gmn experiment will require only 14 days of running .
the clas12 @xmath2 experiment , which like the hall a experiment is approved to make measurements up to 13.5 gev@xmath5 , will require 56 days of running , and will obtain 5 times less statistics at the highest @xmath3-point .
when considering a full set of kinematics , the hall a @xmath2 experiment has a figure - of - merit that is 30 times higher than that of the clas12 @xmath2 experiment .
the existing data for @xmath2 , together with the projected errors for both the hall a and the clas12 experiments , are shown in fig .
while the magnetic ff of the neutron has previously been measured up to 10 gev@xmath5 , the few data that exist above 4.5 gev@xmath5 have uncertainties of about 10 - 20% .
the gmn experiment in hall a will provide sufficient accuracy to bring new understanding to this subject , including the aforementioned decomposition of the @xmath15 and @xmath16-quark distributions .
[ fig : gmn ] the ratio method has been used in a number of experiments , including one at jlab in which @xmath2 was determined with good precision up to @xmath3 = 4.5 gev@xmath5 .
when the recoil nucleon energy is above 2 - 3 gev , the detection efficiency for the neutron and the proton are quite similar , so the ratio method becomes almost systematic free .
like gen(2 ) , the hall a gmn experiment will utilize the bigbite spectrometer with the caveat that the trackers will be based on gems instead of mwdcs . again
like gen(2 ) , the super bigbite magnet itself will be placed in the hadron arm , providing excellent separation between recoil protons and recoil neutrons .
the magnet will also be turned on and off to study potential systematics .
historically , and even quite recently , the ground - state form factors have provided considerable insight into the charge and magnetization densities of both individual nucleons as well as more complex nuclei .
it is quite reasonable to argue that form factors provide us with the best snapshots " we have of both the proton and the neutron . at small values of @xmath3 ,
the fourier transforms of the electric and magnetic form factors can be interpreted as the charge and magnetization densities , respectively .
such interpretations by hofstadter provided early understanding of the size of the proton , and the more recent application of such reasoning has led to the conclusion that the charge and magnetization densities of the proton are not coincident with one another .
the neutron charge density is positive toward its center and is surrounded by negative charge , features that support the notion of a proton - like core embedded in a negative pion cloud .
we should note , however , that relativistic effects limit the degree to which the interpretation of form factors as fourier transforms of densities is correct .
attempts to better account for relativity in the lab frame have been conducted by kelly@xcite .
another approach , pursued by miller , is to define densities that can be computed on the light front@xcite .
the aforementioned relativistic constituent - quark models and calculations based on the dse / faddeev approach both incorporate the basic idea of pion clouds .
a nucleon with a pion cloud necessarily represents a five - quark state , something that is suppressed at high momentum transfer when @xmath3@xmath17 .
there is considerable discovery potential in pushing to higher values of @xmath3 where the short - distance - scale behavior of the nucleon can be revealed , and the structure itself becomes simpler and easier to understand .
all the arguments above underscore the importance of reaching high @xmath3 while simultaneously maintaining high precision .
the required beam time , however , scales roughly as @xmath18 .
it is thus critical to compensate for this large factor in the chosen experimental design . for the measurements proposed within the super bigbite project , the relevant figure - of - merit exceeds those of all competing experiments by factors ranging between 10 and 50 , making the difference between meaningful and ambiguous measurements .
these impressive capabilities are derived from using a large open - geometry dipole magnet for momentum analysis together with a detector package that has a direct view of the target .
the very high rates associated with such a configuration are only tolerable because of the use of a gem - based tracking system .
a cutout in the super bigbite magnet also allows for the use of quite forward angles where the recoil nucleon needs to be detected with large solid angle .
the feasibility of such an open geometry design has been unambiguously demonstrated in multiple experiments using super bigbite s predecessor , bigbite .
the super bigbite apparatus , however , based on the use of a larger magnet combined with gem - based trackers , leads to a figure - of - merit that exceeds that of other competing efforts by a factor of 10 for @xmath1/@xmath4 , around 30 for @xmath2 , and 50 for @xmath0/@xmath2 .
super bigbite will advance the study of electromagnetic form factors in a dramatic fashion _ for both the proton and the neutron_. without super bigbite , however , as shown earlier , measurements will be largely incapable of discriminating between the important theoretical predictions . the super - bigbite apparatus will make possible three ground - breaking measurements of the nucleon s elastic form factors : * a measurement of @xmath0 ( e12 - 09 - 016 ) up to 10 gev@xmath5 using the beam - target double - polarization technique that was approved in january of 2009@xcite . * a measurement of @xmath1 ( e12 - 07 - 109 ) up to 14.5 gev@xmath5 using the recoil - polarization technique that was approved by the jlab program advisory committee ( pac ) in august of 2007@xcite . * a measurement of @xmath2 ( e12 - 09 - 019 ) up to 13.5 gev@xmath5 by determining the cross section ratio for the two reactions d(e , en ) and d(e , ep ) that was approved by the jlab pac in january of 2009@xcite ( a request for an extension to 18 gev@xmath5 is planned ) .
we note that the above experiments will make use of the results of jlab e12 - 07 - 108@xcite ( not part of the super bigbite project ) that will measure @xmath4 up to 17.5 gev@xmath5 .
e12 - 07 - 108 was approved by the jlab pac in 2007 and will use the exquisite calibration of the hall a hrs spectrometers to achieve a 1 - 2% absolute measurement of the electron - proton elastic scattering cross section .
this calibration will allow us to measure ratios of form factors rather than the absolute form factors themselves while still achieving our goals for absolute measurements .
we thus have a plan to measure all of the ground - state electromagnetic form factors with sufficient accuracy and reach in @xmath3 to study some of the most exciting questions in hadronic physics .
the proposed instrumentation , that we refer to collectively as the super bigbite project , includes a set of components that will be used in three different configurations for each of our measurements of @xmath0 , @xmath1 , and @xmath2 , respectively . in all cases ,
the design philosophy incorporates the use of large open - geometry detection , high - rate - handling capability through the use of gem technology , and the ability to measure at relatively forward angles .
we describe our proposed instrumentation below , including the various configurations in which they will be used for our three proposed measurements .
0.2truein the proposed super bigbite apparatus is shown in fig . 4 in the configuration for the gep(5 )
the key features of the apparatus are : * the dipole magnet , which is an existing 48d48 used previously at bnl . * large solid angle , 10 - 15 times larger than in focusing spectrometers , such as hrs / hms / shms . * large momentum acceptance , from 2 gev/@xmath19 at nominal field settings . * high luminosity capability , up to @xmath20 electron / s@xmath7nucleon / cm@xmath5 , e.g. , in gep(5 ) . * small scattering angle capability , down to 3.5@xmath22 . * full acceptance for the long target ( up to @xmath2320 cm ) .
* very good angular resolution , @xmath24 [ mrad ] = 0.14 + 1.3/@xmath25 [ gev/@xmath19 ] . * good vertex resolution , @xmath27 1 - 2 mm . * good momentum resolution , @xmath28 [ gev/@xmath19 ] . * high energy trigger threshold via use of a hadron calorimeter .
-0.2truein the super - bigbite apparatus will be used in gep(5 ) as a large - acceptance spectrometer , the super bigbite spectrometer ( sbs ) .
it will provide angular coverage up to @xmath2970 msr , with a detector package , capable of operating at the largest possible luminosity , almost @xmath30 electron / s@xmath7nucleon / cm@xmath5 .
the use of a simple dipole placed close to the target , made possible by a deep cut through the iron yoke of the magnet for the beam line , allows one to achieve the large angular acceptance in the spectrometer .
the magnet deflects charged particles vertically and will be used with a field integral up to 2.5 t@xmath31 m .
the field in the beam line will be reduced to an acceptable level by specially developed magnetic shields .
the relatively small bend angle is compensated for by the high coordinate resolution ( 70 @xmath32 m ) of the front gem - based chambers resulting in a momentum resolution of @xmath33 at 8 gev/@xmath19 in gep(5 ) with the 40-cm long lh@xmath34 target .
the gem technology solves the main challenge of this spectrometer , the very high counting rates , allowing tracking at background rates much higher than those expected for the experiments presented here .
* these features combined will give sbs at least a factor of 10 advantage compared with any existing or proposed spectrometer at jefferson lab for nucleon form - factor measurements .
* the electron beam parameters required for the emff measurements , such as energy , polarization , intensity and size , are all within the 12-gev specifications .
the gen(2 ) experiment@xcite will use a polarized @xmath35he target as a key component .
a novel concept of a convective - flow cell will allow to increase the beam current up to 60 @xmath32a with a 60-cm - long cell .
the gep(5 ) experiment will use a 40-cm long liquid hydrogen target .
a wide vacuum snout from the scattering chamber to the magnetic shield will allow one to avoid a direct view of the detector from the beam - line vacuum pipe elements , that will reduce the counting rate in the detector by a factor of 2.5 .
the gmn experiment@xcite will use standard 10-cm cryogenic targets of hydrogen and deuterium
. there will be three trackers in sbs for the gep(5 ) experiment .
the first one , ft , will be used to measure the proton momentum and its direction before interaction with the first ch@xmath34 analyzer .
this tracker with an active area of 40@xmath7150 cm@xmath5 will include six chambers , each with two - dimensional read - out and three gem amplification foil planes .
the front tracker will be followed by a double polarimeter consisting of two trackers each preceded by a ch@xmath34 analyzer .
adding this second polarimeter increases the figure - of - merit by a factor of 1.7 , equivalent to a 30% reduction of the experimental statistical errors .
the second tracker , st , will measure the proton track after the proton passes through the first ch@xmath34 analyzer .
the third tracker , tt , will measure the direction of the track after the proton passes through the second ch@xmath34 analyzer .
the dimensions of the st and the tt were chosen to be 50@xmath7200 cm@xmath5 to keep the figure - of - merit above 90% of the ideal value , which corresponds to trackers of unlimited size .
the st and tt are required to have only four chambers each ( compared with six in the ft ) due to the reduced demand on coordinate resolution and the lower counting rates .
each gem chamber will consist of 50@xmath740 cm@xmath5 sub - sections . at the ft we expect high background hit rates of about 400 khz / cm@xmath5 ( based on geant simulations ) due to the direct view of the target .
the background is dominated by soft photons originating from the target .
low - momentum charged particles are swept away by the magnet .
the rates on the second and third trackers are expected to be 130 khz / cm@xmath5 and 64 khz / cm@xmath5 , respectively , dominated by soft electrons / positrons converted from photons in the analyzers .
the emff experiments deal with very small cross sections , so for meaningful results the luminosity should be as high as possible .
arrangement of the trigger and the detector structure for the high luminosity should take maximum advantage of the high energy of the recoil nucleon .
the energy of the nucleon in the emff experiments ranges from 2 to almost 10 gev , depending on the measurement . in gep(5 ) , the large proton energy will allow the use of a 3 - 4 gev energy threshold in the calorimeter without a significant loss of detection efficiency .
such a high energy threshold leads to the suppression of pion triggering and opens the possibility of using a coincidence for the daq trigger .
the total active area of the hadron calorimeter ( hc ) will be 5.5 m@xmath5 .
it has a good time resolution of 1.5 ns , high granularity ( 15@xmath715 cm@xmath5 ) , a very good coordinate resolution of 2 cm , and , in addition , a high energy threshold .
all these features make the hc an attractive neutron detector for the two neutron emff experiments .
positioned at the end of the sbs detector package the hc will be used in the gep(5 ) experiment to trigger the daq , in coincidence with the signals from the existing electromagnetic calorimeter , bigcal . in addition
, these two calorimeters will provide coordinate information that will be used to locate the proton track in the ft and in the tt by applying kinematic constraints .
the active size of the hc in the gep(5 ) experiment will be 150@xmath7300 cm@xmath5 , which is a little smaller than its full size and the total number of active modules will be 200 . in all three emff experiments both the ( quasi ) elastically scattered electron and the recoil nucleon will be detected .
this allows for the selection of the exclusive process , which has a very small cross section at high momentum transfer .
the existing hall a bigbite spectrometer will be used for the detection of electrons in the gen(2 ) and gmn experiments .
the proposed ft and two chambers of the st of sbs will be used as the tracker in the bigbite spectrometer .
this switch does not require any reconfiguration of the gem chambers .
the cherenkov counter and the large double - layer shower detector of the existing bigbite trigger system will be used in both neutron form - factor experiments .
the existing bigcal calorimeter will be used in the gep(5 ) experiment for detection of the scattered electron .
the calorimeter has 1744 lead - glass blocks coupled to pmts of type feu-84 .
two sets of blocks are used , one 38@xmath738@xmath7450 mm@xmath35 and the second slightly larger , 40@xmath740@xmath7400 mm@xmath35 .
blocks will be arranged in a matrix 20@xmath775 , a shape optimized for the largest acceptance at @xmath3= 14.5 @xmath36 . the energy and coordinate resolutions of bigcal of about 5 - 7% and 7 mm , respectively , for 2.5 gev electrons satisfy the trigger and tracking requirements .
the calorimeter will be installed 3 m from the target at a central angle of 39@xmath22 ( at the largest @xmath3 ) .
it will be shielded from the target by a 20 cm al plate to reduce the radiation dose on the lead - glass .
the angular correlation between the scattered electron and the recoil proton will be measured very accurately , especially the angle between the electron scattering plane and the proton recoil plane .
because of the small size of the electron beam , the angles of the electron and of the recoil proton can be determined with a very good accuracy of @xmath29 0.5 mrad . in order to achieve this ,
a 1 mm coordinate accuracy is required for the scattered electron .
it will be provided by the coordinate detector , cd , consisting of two gem - based chambers .
the sub - sections of each chamber will be similar to those of the polarimeter chambers .
the overall dimensions are 80@xmath7300 cm@xmath5 with 6 sections in 2 vertical columns .
the daq trigger will have two levels : the first - level trigger will have a relatively short delay , @xmath29100 ns , relative to the particle detection .
it will be generated by the electron arm .
this signal is required for operation of the gem read - outs ; despite their pipelined architecture , the apv25 front - end chips used in the gems require a fast trigger for marking `` interesting '' samples .
such a fast first - level trigger also eliminates the need for long delay lines for the conventional fastbus and vme electronics .
the second - level trigger will be formed using a relatively complex coincidence logic , described below .
it can have a latency of several microseconds .
the trigger is a critical part of the gep(5 ) experimental design .
it will have two main features : ( 1 ) a high energy threshold in the calorimeters in both arms ( 3 - 4 gev for the proton and 2.5 gev for the electron ) ; and ( 2 ) `` smart '' fpga - based coincidence electronics that allow second - level triggering on spatial correlations of hits in the two detector arms .
the high energy thresholds will reduce the rate in the electron arm to 60 khz and in the proton arm to about 1.5 mhz . with a coincidence time window of 50 ns , forming
a simple coincidence between the two arms would result in a trigger rate of at least 5 khz , which is too high ( or at least close to the limit ) for the expected total event size of @xmath37 kb .
however , the small deflection of the proton trajectory in super bigbite makes it possible to use the angular correlation between the elastic electron and the proton trajectories , at the trigger level , to reduce the coincidence trigger rate further by a factor of five . in designing our second - level triggering system , it is important to note that real ( not noise ) calorimeter hits always cause several adjacent blocks to fire .
we therefore construct what we call macroblocks " , consisting of the sums of the individual signals from several nearby blocks .
in hcal , the size of the macroblocks is @xmath38 ( we call this a `` sum-16 '' ) , and in the ecal the size is @xmath39 ( `` sum-32 '' ) .
the different macroblocks overlap each other , such that each individual block contributes to the sums associated with several different ( overlapping ) macroblocks . in this way
, at least one macroblock will always contain a large fraction of the deposited energy associated with a real hit . in the gmn and the gen(2 )
experiments the gas cherenkov counter and the two - layer shower calorimeter of the bigbite spectrometer provide an efficient on - line selection of the electrons so that the daq trigger rate will be at the level of 23 khz already without requiring an on - line coincidence with the hadron arm .
a funding proposal for the superbigbite project has been submitted to doe , and further details about the project can be found in that proposal@xcite .
if the super bigbite project is completed in the proposed schedule , one of the top priorities of the jlab upgrade , the study of the ground - state electromagnetic form factors , could be achieved in the early years following commissioning .
studies have indicated that the super bigbite instrumentation provides a great potential for other experiments , such as studies of the neutron spin asymmetry @xmath40 through inclusive scattering and of the neutron collins and sivers functions through semi - inclusive pion and kaon production off a transversely polarized @xmath35he target .
the author gratefully recognizes the strong support provided by gordon cates and bogdan wojtsekhowski in preparing this manuscript .
this work was supported by doe contract de - ac05 - 06or23177 , under which jefferson science associates , llc , operates the thomas jefferson national accelerator facility . | a proposed set of instrumentation , collectively referred to as the super bigbite project , is presented .
used in three different configurations it will allow measurements of three nucleon electromagnetic form factors @xmath0 , @xmath1 , and @xmath2with unprecedented precision to @xmath3-values up to three times higher than existing data . |
the interaction between interstellar clouds and shock waves associated with for example supernova remnants , h ii regions or stellar winds has been studied as one of the most fundamental problem of interstellar gas dynamics .
shock - molecular cloud ( mc ) interaction is especially important as a process that dynamically drives star formation .
studies using hydrodynamics or magnetohydrodynamics simulations revealed that strong interstellar shock wave injected into the molecular cloud crush and destroy it through shock injection and ensuing hydrodynamical instabilities such as kelvin - helmholtz ( kh ) , rayleigh - taylor ( rt ) and richtmyer - meshkov ( rm ) instabilities , although inclusion of magnetic field strongly suppress those instabilities @xcite .
magnetic field orientation and mass density structure within clouds also affects significantly the later phase of the dynamics of molecular cloud impacted by shock wave @xcite .
recent observations and numerical simulations reveal multi - phase multi - scale dynamics of filamentary clouds - shock interaction within a cloud is important for the evolution of star forming molecular clouds with the help of thermal or self gravitational instabilities @xcite . on the other hand , shock waves
are frequently observed in the corona of the sun @xcite . in the corona ,
magnetic field reconnection allows magnetic field energy released in a catastrophic way resulting in largest explosion in the solar system .
this is called solar flares . during solar flares ,
typically @xmath0 - @xmath1ergs of magnetic field energy stored in the corona is converted to thermal , kinetic , radiation and high energy particle kinetic energy in a short ( several minutes ) time scale @xcite . as a result of sudden energy release in solar flares ,
a part of energy propagate globally in the corona as a form of non linear fast mode mhd wave ( or shock ) .
recently , high cadence extreme ultraviolet ( euv ) observation of the solar corona by atmospheric imaging assembly ( aia ; @xcite ) on board solar dynamics observatory ( _ sdo _ ; pesnell et al .
2012 ) allows detailed imaging observation of such global shock waves in the corona @xcite .
these flare associated shock waves in the lower solar corona are thought to be weak fast mode mhd shocks @xcite .
plasma ejections ( coronal mass ejections , cmes ) associated with solar flares propagate in the interplanetary space , sometimes with clear shock fronts at their noses @xcite .
these propagating shockwaves in the upper corona or interplanetary space are observed in coronagraph images or in radio dynamic spectrum observations @xcite .
they are also observed as a sudden change in plasma parameters of solar wind velocity , temperature , density and magnetic field strength observed in - situ in front of the earth @xcite .
we also have clouds in the solar corona , the prominences. they are cool and dense plasma floating within hot and rarefied corona .
they are supported by magnetic tension force against gravity@xcite .
recent high resolution and high sensitivity observation by solar optical telescope ( sot ; tsuneta et al .
2008 ) on board @xmath2 satellite@xcite revealed highly dynamic nature of solar prominence , with continuous oscillation and turbulent flow@xcite .
they give us rare opportunity for studying dynamics of partially ionized plasma in detail , whose plasma parameters are very difficult to reach in laboratory or in other space objects .
the excitation mechanism of such chaotic flows within prominence material are also discussed recently .
@xcite discussed photospheric motion as one possible mechanism that drive long frequency small amplitude prominence oscillations .
non - linear mhd waves propagating upward around prominence foot is studied by @xcite . the magnetic rayleigh - taylor instability invoked by interchange reconnection between magnetic field lines supporting prominence plasmas
is discussed as an excitation mechanism for multiple plume - like upflows observed with _ hinode_/sot which help mix up prominence plasmas @xcite .
magnetic kelvin - helmholtz instability excited near the absorption layer of transversely oscillating prominence in the corona helps cascade energy and heat the prominence plasma through turbulence excitation and current sheet formation @xcite . sometimes , coronal shock waves associated with flares hit solar prominences and lead to excitement of large amplitude prominence oscillations .
these oscillations give us information of physical properties of prominences such as magnetic field strength , density , and eruptive stability and have been studied widely @xcite .
in contrast to interstellar shock - molecular cloud interaction that has been studied widely , the excitation process of large amplitude solar prominences or prominence activations by coronal shock injection has not been studied in detail in the context of shock - cloud interaction .
in contrast to the situation in interstellar medium where strong shock wave often interact with molecular clouds , the prominence activation is the interaction between prominence and weak fast mode mhd shock wave with fast mode mach number being between 1.1 and 1.5 @xcite .
the time scale of shock - prominence interaction is several minutes , which offers us unique opportunity to study in detail the time evolution of shock - cloud interaction process in cosmic plasmas . in this paper
, we analyze observational data of shock - prominence interaction obtained by sdo / aia and compare them with numerical mhd simulation .
especially , we focus quantitatively on how wave momentum is transferred to cloud material through shock injection , flow drag and magnetic tension acceleration processes with 3d mhd simulation .
we also discuss the effect of internal structure such as volume filling factor on chaotic flow excitation and its damping .
we compare the mhd simulation results with hydrodynamic simulation and discuss the role of magnetic field in bulk prominence acceleration as well as excitement and damping of internal flows .
we make a phenomenological model that describe momentum transfer from coronal shock to the prominence . we validate the phenomenological model by comparison with simulation , and apply the phenomenological model of prominence activation in the context of diagnosing coronal shock properties .
lastly , we compare shock - prominence interaction with interstellar shock - molecular cloud interaction in the context of mhd shock - cloud interaction .
on march 7 , 2012 , an x5.4 class flare occurred at noaa active region ( ar ) 11429 located at north east quadrant of the solar disk .
the soft x - ray light curve obtained by goes peaked at 00:24ut on march 7(figure 1(a ) ) .
this flare is the second largest one in current solar cycle(cycle 24 ) .
the flare was associated with very fast cme whose velocity was about 2684km / s .
figure 1 ( b ) is the composite of coronagraph images obtained by _ soho_/lasco c2 and euv image obtained by _
sdo_/aia 193 @xmath3 band both taken at 00:36ut .
we can see the shock front surrounding the cme ejecta in _
c2 image(figure 1(b ) ) .
figure 1(d ) shows _ sdo_/aia 193@xmath3 difference image at 00:18ut .
we see a dome - like bright structure expanding above ar 11429 in figure 1(d ) .
@xcite estimated the propagation speed of the leading shock front ahead of the cme as 672 km s@xmath4 based on the analysis of dynamic spectrum obtained with the hiraiso radio spectrograph ( hiras , kondo et al .
it seems strange that the estimated speed of the leading shock front ahead of the cme is much slower than the cme speed estimated with _
we looked at the coronagraph observation data by _ stereo - b_/cor1 and radio dynamic spectrum once again to check the consistency .
the hight of the leading edge of the cme ejecta seen in _ stereo - b_/cor1 image taken at 00:26ut is larger than 2 @xmath5 measured from the photosphere , where @xmath5 is the solar radius . on the other hand ,
the radio dynamic spectrum during the flare period is complicated and composed of type ii , type iv and possibly type iii bursts .
we noticed @xcite mistook type iv burst for type ii burst , resulting in the cme speed estimation inconsistent with coronagraph observation . from 00:17:10ut (
indicated as @xmath6 in figure 2 ) to 00:17:38ut ( indicated as @xmath7 in figure 2 ) , we see a clear linear structure in the radio dynamic spectrum showing the characteristic signature of type ii burst , whose frequency drifted from the 112mhz to 88mhz during the period . assuming the type ii burst signature corresponds to the first harmonics of the plasma oscillation at the upstream of the leading shock front , we got the propagation speed of 1.9@xmath810@xmath9 km s@xmath4 based on @xcite , which seems to be consistent with coronagraph observations .
the footprints of the dome - like shock front which propagated to the north in the lower corona was especially bright in aia 193 @xmath3 images .
that shock front propagated further to the north and hit a prominence located at the north pole .
the interaction between the shock wave and the polar prominence resulted in large amplitude prominence oscillation ( lapo ) .
we call the excitation process of lapo , prominence activation further on .
figure 3 shows the time evolution of the prominence activation seen in aia 193@xmath3 and 304@xmath3 images .
the fov of figure 3 is shown as the black rectangle in figure1 ( c ) .
the polar prominence was seen bright in aia 304 @xmath3 images while it was seen dark in aia 193 images .
figure 3 ( a ) and ( b ) show the prominence just before it is activated by the shock .
figure 3 ( c ) and ( d ) show the prominence just hit by the coronal shock front .
the shocked part of the prominence seen in aia 304 image became about twice as bright as its original brightness .
the bright part propagated in the direction of shock propagation and at the same time the prominence started to move in the shock propagation direction .
figure 3 ( i ) and ( j ) show time - distance diagrams along the cut ab shown in figure 3 ( a ) between 00:20ut and 01:30ut . in figure 3
( i ) , the coronal shock front is seen as bright propagating structure . the propagation speed of the coronal shock front in the plane of the sky is measured to be 380 km s @xmath4 from figure 3 ( i ) .
when the coronal shock front reached the dark prominence , the prominence was abruptly accelerated .
the initial activated prominence speed was 48 km s @xmath4 measured from figure 3 ( i ) .
we can clearly see the sudden brightening of the prominence during its activation in figure 3 ( j ) . in figure 3 ( j ) , we can see somewhat chaotic movement of prominence threads during lapo .
we note here that we have neglected the line - of - sight component of the speed of shock propagation and activated prominence , so the estimated shock speed and prominence velocity should be regarded as lower limits .
figure 3 ( g ) and ( h ) show the prominence when its displacement from the original position was largest .
note that the time scale of prominence activation is several minutes in this event while the period of lapo is longer than an hour .
the white rectangle in figure 3 ( i ) corresponds to the prominence activation process that is also shown in figure 20 ( a ) .
in order to study in detail the physics of prominence activation , we carried out three - dimensional mhd simulation .
we numerically solved the following resistive mhd equations , @xmath10 @xmath11 @xmath12 @xmath13 @xmath14 , where @xmath15 and @xmath16 are mass density , gas pressure , magnetic field and velocity , respectively .
@xmath17 is the electrical current density . @xmath18
is uniform electrical resistivity and @xmath19 is specific heat ratio . in the induction equation ,
additional variable @xmath20 is introduced in order to remove numerical @xmath21 as proposed by@xcite .
the numerical scheme we used is harten - lax - van leer - discontinuities ( hlld ) approximate riemann solver @xcite with second - order total variation diminishing ( tvd ) monotonic upstream - centered scheme for conservation laws ( muscl ) and second order runge - kutta time integration .
we studied eight simulation cases of coronal shock - prominence interaction , with different density structure of the prominence for each run . in run @xmath22 and @xmath23 ,
the prominence is a uniformly dense spherical plasma of radius @xmath24 .
initially , the center of the spherical prominence was at @xmath25 . in this model , @xmath26 is used .
the whole prominence region is where @xmath27 .
we modeled the prominence as a sphere of plasma in stead of a cylindrical plasma in this paper in order to also study the effect of magnetic tension force induced by shock injection which is revealed to be important in prominence activation as discussed in section 3.4 .
when the time scale of prominence activation is much smaller than that of lapo ( as in the case of the event discussed in section 2 ) , the global magnetic field structure will not play a significant role on the dynamics of prominence activation process . in that case
we can separate the physics that govern prominence activation with that govern the ensued lapo . in order to focus only on the physics of prominence activation ,
we think of such a situation in this study .
we set initial magnetic field to be uniform , neglecting the effect of global loop curvature and gravity . as a result ,
gas pressure also become uniform due to total pressure balance between corona and prominence .
the coronal shock front was initially at @xmath28 .
the fast mode shock wave propagate in the positive @xmath29 direction with density compression ratio @xmath30 .
the initial magnetic field is in z - direction , so the injected fast mode shock is perpendicular one .
we note that in reality , oblique components may play roles in prominence activation , especially in the plane perpendicular to the shock propagation direction .
we do not take into account such oblique dynamics induced by oblique shocks and focus only on the dynamics of activated prominence in the direction of shock propagation , and discuss the effect of prominence mass to the activation dynamics . in run @xmath31 and @xmath32 ,
the prominence consists of 200 randomly sized spherical clumps of spherical shape put inside the whole sphere of radius @xmath24 ( figure 4(a ) and ( b ) ) . from the comparison of runs @xmath33 with runs @xmath34 , we study the role of internal density structure of the prominence in its bulk acceleration and excitation of internal chaotic flow , and also see how well uniform prominence approximation works . in reality , we only solved the numerical box with positive y and z coordinate in order to reduce the numerical cost ( i.e. motion of only @xmath35 clumps are simulated ) . from here , we express the density distribution of prominence as @xmath36 for convenience . in runs @xmath33 ,
clumps are distributed so that the average mass density within the prominence region ( @xmath37 ) become @xmath38 , where @xmath39 is the mass density of background ( shock upstream ) corona and @xmath40 is the mass density of each clump .
the volume filling factors @xmath41 are set to be ( @xmath42 . in runs @xmath34 , the mass density of the prominence is uniformly set to be @xmath43 ( i.e. both @xmath34 and @xmath33 prominences have the same average density over the entire volume .
only difference is @xmath34 prominences are uniform but those in @xmath33 are made of clumps . )
the equational form of the initial condition is as follows , @xmath44 @xmath45 @xmath46 @xmath47 @xmath48 @xmath49 @xmath50 @xmath51 with @xmath52 , @xmath53 , @xmath54 . the variables @xmath55 , @xmath56 and @xmath57 in above equations are density jump ( @xmath58 compression ratio ) , pressure jump and plasma velocity of coronal shock wave , respectively . plasma @xmath59 is assumed to be @xmath60 to model low beta corona , where lorentz force dominates gas pressure gradient force in accelerating the prominence .
the electrical resistivity @xmath18 is set to be @xmath61 in all cases , in order to prevent numerical instability .
the unit of speed in our simulation is @xmath62 .
the corresponding sonic , alfvenic and fast mode wave phase speeds in the corona are respectively expressed as , @xmath63 @xmath64 @xmath65 from mhd rankine - hugoniot relations for perpendicular shocks , sonic and fast mode mach numbers @xmath66 and @xmath67 are respectively expressed by @xmath68 as , @xmath69 @xmath70 from mhd rankine - hugoniot relations , the pressure jump @xmath56 and the plasma velocity of shocked corona @xmath57 are expressed respectively as follows .
@xmath71 @xmath72 the simulation box is @xmath73 $ ] , @xmath74 $ ] , @xmath75 $ ] discretized with non - uniformly arranged @xmath76 grid points . especially , inner region of @xmath77 $ ] , @xmath78 $ ] , @xmath79 $ ] is discretized with uniformly set @xmath80 grid points . we applied reflective boundary conditions on @xmath81 and @xmath82 planes , while for other boundaries we applied free boundary conditions .
we use sparse grids in outer space so that we can neglect unwanted numerical effects on prominence dynamics from outer free boundaries .
figure 4 ( a ) and ( c ) show the density distribution of two different times ( @xmath83 and @xmath84 , respectively ) in simulation run @xmath85 . the prominence hit by the coronal fast mode shock is compressed and start to move in the direction of shock propagation ( positive @xmath29 direction ) . the fast mode shock front transmitted into the prominence material experiences multiple reflections at the prominence - corona boundary .
figure 5 ( c ) shows the distribution of z - component of vorticity @xmath86 in xy plane . in figure 5 ( c )
, we find sharp velocity shear at the prominence - corona boundary .
the velocity shear result in the development of kelvin - helmholtz instability in xy plane ( figure 5 ( a ) and ( f ) ) .
the magnetic tension force suppress the the velocity shear in xz plane ( figure 5 ( d ) ) .
associated with kelvin - helmholtz vortex behind the cloud , the low pressure region is formed behind the cloud , helping the prominence being accelerated in x - direction ( figure 5 ( e ) and ( f ) ) .
magnetic tension force also help stretch the cloud in x - direction . in figure 4 ,
time evolution of density structure in simulation run @xmath85 and @xmath87 is shown with color contour in logarithmic scale together with magnetic field vectors in the plane of the plots shown by white arrows .
the cloud in run @xmath85 is deformed by kelvin - helmholtz instability that have developed behind the cloud . in run
@xmath87 , time needed for each clump to be deformed by shear flows is much shorter thatn in run @xmath85 because of small length scale of each clumps . in runs
@xmath33 , each shocked clumps interact with each other through flow field around them making the overall flow and density structure more complicated compared with runs @xmath34 in a short time scale .
figure 7 shows the time evolution of the center of mass velocity @xmath88 of the prominence in runs @xmath34 and @xmath33 .
the center of mass velocity of prominence @xmath88 is defined as follows , @xmath89 , where @xmath90 is the threshold mass density and set to be 2 in all simulation runs .
we regard the plasma with density @xmath91 as prominence in this analysis .
the volume integral is done over the region where the mass density is larger than @xmath90 . practically , we first flag the region where @xmath92 is larger than @xmath90 in the whole computational box , and then sum up the quantities within the flagged region .
we tried various values of @xmath90 , and found no significant change in the analysis results .
it takes more time in runs @xmath93 and @xmath94 for @xmath88 to approach @xmath57 than in runs @xmath95 and @xmath96 , although the time evolution of @xmath88 in runs @xmath85 and @xmath23 is very similar to that in runs @xmath87 and @xmath32 . here
, we focus on the mechanism of the momentum transfer from the coronal shock wave to the prominence . @xcite discussed the momentum transfer mechanism of shock injection into the prominence , and estimated the resultant prominence velocity @xmath97 with the help of one - dimensional linear theory as , @xmath98 , where @xmath99 is the ratio of the mass density between corona and prominence .
multidimensionality and non - linearity effects on prominence activation that are neglected in 1d model are important especially in quantitative discussion . in this section ,
we make a phenomenological model of prominence activation taking into consideration the effect of fluid drag force and magnetic tension force , and compare it with the result of 3d mhd simulation of prominence activation ( runs @xmath34 ) .
the schematic figure of the phenomenological model and corresponding simulation snap shots are shown in figure 9 . in our case ,
the center of mass of the prominence moves parallel to the direction of shock propagation ( x - direction ) because the coronal shock that activate the spherical prominences is a perpendicular one .
in this section , we focus on how prominence center of mass is accelerated in x - direction as a result of the interaction with coronal perpendicular shock wave .
we define the timescale in which the injected shock front sweeps the prominence material as shock sweeping timescale , @xmath100 , where @xmath101 is the fast mode phase speed in the prominence .
this is similar to cloud crushing timescale , @xmath102 , discussed in @xcite during which a molecular cloud is crushed by strong shock wave in interstellar medium .
this shock sweeping mechanism will accelerate the prominence to have the velocity of @xmath97 according to 1d linear theory . during the shock sweeping process
( @xmath103 ) , the mean acceleration of the prominence through shock sweeping mechanism @xmath104 is approximated as , @xmath105 @xmath106 is a ad - hoc factor of order unity introduced to take into account multidimensional effect such as shock refraction and set to be @xmath107 throughout the paper .
the difference between the velocity of shocked coronal plasna @xmath57 and the prominence itself @xmath108 causes the pressure difference between the front and back side of the prominence , which accelerate the prominence to the direction of shock propagation(fluid drag force ) .
the prominence acceleration by the fluid drag force @xmath109 is approximated as , @xmath110 , where @xmath111 and @xmath112 are cross sectional area and total mass of the prominence assuming spherical shape of the prominence , respectively .
@xmath113 is a drag coefficient . in the following discussion
, we approximate the drag coefficient by unity , i.e. @xmath114 . it is revealed in the following discussion that the drag acceleration @xmath115 is much smaller than the other two acceleration mechanisms in weak shock acceleration , which is the case in prominence activation by coronal shocks . substituting these into equation ( 24 ) , we get , @xmath116 , where @xmath117 is the prominence center of mass speed relative to ambient shocked corona . also , velocity difference between the prominence and ambient corona distort the magnetic field lines penetrating the prominence .
this curved magnetic field accelerate the prominence by magnetic tension force .
the equation of motion of the prominence accelerated by the x - component of magnetic tension force is approximately written as , @xmath118 with @xmath119 being the acceleration by magnetic tension mechanism . here , @xmath120 is the effective cross sectional area of the part of the prominence in xy - plane through which shocked coronal magnetic field lines penetrate .
@xmath121 is the inclination of distorted magnetic field lines ( figure 9 ) . assuming the inclination is determined by the balance between displacement of magnetic field lines by velocity difference @xmath122 and its relaxation by alfven wave propagation @xmath123 , @xmath124 is approximated as , @xmath125 , where @xmath126 is the alfven wave phase speed in the shocked coronal plasma . here
, we approximate @xmath120 as follows , @xmath127 , where shock passage timescale @xmath128 is the timescale in which the shock front pass over the prominence . then @xmath119 is finally approximated as , @xmath129 the order of magnitude ratio of maximum accelerations for each mechanisms is roughly @xmath130 , if @xmath59 is much smaller than unity and the shock is not strong .
when the coronal shock is weak , @xmath115 is negligible compared with other two .
after the shock front have swept the prominence , the main acceleration mechanism is magnetic tension force acceleration . summarizing above discussion ,
the prominence acceleration in the phenomenological model @xmath131 is written as follows , @xmath132 , or explicitly , @xmath133 is expressed as follows , @xmath134 the time evolution of prominence velocity in the phenomenological model @xmath135 is obtained by solving the prominence equation of motion , @xmath136 substituting the simulation parameters for run @xmath34 into the above equation of motion , we get the resultant time evolution of phenomenological prominence velocity .
time evolution of prominence velocity in runs @xmath34 and models @xmath34 are compared in figure 8 . on the other hand , the fluid equation of motion
is written as , @xmath137 if we assume the prominence mass @xmath138 is constant , the volume integral of the above equation of motion over the prominence volume leads to the following relations , @xmath139 , with @xmath133 , @xmath140 and @xmath141 being the acceleration of center of mass of the prominence in x - direction , prominence acceleration by total pressure gradient force and prominence acceleration of magnetic tension force , respectively .
@xmath133 , @xmath140 and @xmath141 are respectively written ( and calculated from simulation results ) as follows ,
@xmath142 @xmath143 @xmath144 in figure 10 ( a)-(h ) , we show the total pressure gradient acceleration @xmath145 and magnetic tension acceleration @xmath146 with dashed and solid lines , respectively in mhd simulation runs @xmath34 and @xmath33 .
figure 10 ( i)-(l ) show @xmath147 and @xmath119 in models @xmath34 .
sudden drop in total pressure gradient force acceleration in for example run @xmath94 ( figure 10 ( c ) ) corresponds to the first reflection of the wave which have swept the whole prominence ( figure 9 ( e ) ) .
the oscillations of accelerations in runs @xmath34 are the results of multiple wave reflections at prominence - corona boundary ( figure 9 ( f ) ) , and the oscillation period is approximately @xmath148 for each runs . in run @xmath95 and @xmath96 ,
the total pressure gradient force acceleration is smaller than that in run @xmath93 and @xmath94 .
that is one reason why it took more time for the prominence velocity @xmath88 in runs @xmath93 and @xmath94 to approach the velocity of shocked corona @xmath57 than in runs @xmath95 and @xmath96 . after the shock sweeping time @xmath148 , @xmath147 in the phenomenological models @xmath34 are smaller than pressure gradient force acceleration @xmath145 in runs @xmath34 .
also , the magnetic tension force acceleration @xmath149 in the phenomenological models @xmath85 and @xmath23 underestimates the actual simulation results @xmath146 in runs @xmath85 and @xmath23 .
roughly , the time evolution of @xmath145 and @xmath146 in runs @xmath34 resemble to that of @xmath147 and @xmath119 in models @xmath34 , respectively .
this is natural because @xmath104 in models @xmath34 represents the acceleration by fast mode shock transmission into the prominence , which is mainly due to total pressure gradient .
also , @xmath115 in models @xmath34 is mainly due to pressure gradient which lasts longer than @xmath148 resulting from velocity difference between corona and prominence . due to velocity shear at the prominence - corona boundary in xy plane
, khi develops and low pressure region is formed behind the cloud associated with induced vortices ( figure 5 ( a ) , ( c ) and ( e ) ) .
also , we see a high pressure region ahead of the cloud due to flow collision .
this type of flow structure result in the large scale total pressure gradient that accelerate the prominence in x - direction as a drag force .
formation of induced vortices and associated low density region behind the cloud is clearer in hydrodynamic case ( simulation run @xmath150 ) discussed in section 3.6 ( figure 13 ) . while coronal shock wave gives its bulk momentum to the prominence , internal flow
is also excited within the prominence .
internal flow excitation during interstellar shock - cloud interaction has been studied widely as a measure of internal turbulence .
the internal flow is discussed using velocity dispersion in x , y and z components , @xmath151 @xmath152 @xmath153 , where mass weighted average of physical quantity @xmath154 is calculated as , @xmath155 we assumed @xmath156 and @xmath157 should be @xmath158 because of the symmetry around x - axis .
figure 11 shows the time evolution of velocity dispersion @xmath159 , @xmath160 and @xmath161 in runs @xmath34 and @xmath33 in semi - log graph .
they show the excitation of the internal flow through the passage of shock , and its damping roughly in an exponential manner . in runs
@xmath34 , we also find oscillatory behavior of @xmath159 with their period of roughly @xmath162 .
we estimated the damping times ( e - folding times ) @xmath163 for @xmath159 , @xmath160 and @xmath161 in runs @xmath34 and @xmath33 based on least square fitting by bisector method @xcite between @xmath84 and @xmath164 .
the resultant slopes for exponential damping are shown as thick lines for each variables in figure 11 .
we did not estimate them in runs @xmath23 and @xmath32 because the simulation time span did not cover their damping phase .
the estimated damping times are shown in table 1 .
the estimated damping times fall in the range between @xmath148 and 2@xmath148 in almost all the cases .
@xmath165 is roughly twice as large as @xmath166 and @xmath167 .
we show the time evolution of the center of mass position of the prominence @xmath168 in figure 12 .
uniform mesh is used between @xmath169 shown as dashed lines in figure 12 .
in all eight simulation runs , the prominence acceleration phase is mostly solved within uniform mesh region . in figure 12
, we have no significant change of damping behavior when the cloud center of mass crossed @xmath170 plane . in this section ,
we study hydrodynamic simulation of shock - cloud interaction ( run @xmath150 ) and compare it with mhd simulation run @xmath85 . by directly comparing run @xmath150 with run
@xmath85 , we discuss hydrodynamic effects which were suppressed in mhd run @xmath85 by the presence of magnetic field .
such hydrodynamic effects are expected to be more significant in interstellar shock - cloud interaction . in practice , we modified the initial plasma beta in run @xmath85 to @xmath171 and use it as an initial condition of the mhd simulation .
the pressure jump @xmath56 , plasma velocity of shocked corona @xmath57 and shock mach number @xmath66 are @xmath172 , @xmath173 and @xmath174 , respectively .
the minimum plasma beta appeared throughout the simulation was @xmath175 which is still much larger than unity .
the magnitude of plasma beta is of the order of the ratio of pressure gradient force term to lorentz force term in mhd equation of motion . because plasma beta is very large ( @xmath176 ) throughout the simulation ,
the effect of magnetic field on the dynamics in simulation run @xmath150 is negligible .
so , we call run @xmath150 `` hydrodynamic '' simulation further on .
figure 13 ( a ) shows the volume rendering of the density distribution in run @xmath150 at @xmath177(@xmath178 ) . after being crushed by the injected shock wave
, the cloud in run @xmath150 continues to be deformed .
this is quite different from the time evolution of cloud shape in mhd run @xmath85 . in run
@xmath85 , magnetic field suppress the cloud deformation and damps the internal flow rapidly in the time scale of @xmath179 . without such a magnetic suppression of cloud deformation ,
the time evolution of the cloud as well as ambient flow structure in run @xmath150 is very different with those in run @xmath85 .
one characteristic flow structure in run @xmath150 is the formation of large coherent vortex flow behind the cloud .
the ambient flow converging towards the cloud along x - axis and diverging in yz plane associated with the vortex helps flatten and stretch the cloud further in run @xmath150 ( figure 13 ( b ) and ( c ) ) .
in figure 14 ( a)-(d ) , snapshots of density and pressure distribution in xz plane at @xmath180 in simulation run @xmath85 and @xmath150 are shown , respectively .
the velocity amplitudes of transmitted shock front @xmath181 and that of the shock front from behind the cloud @xmath182 in simulation run @xmath85 and @xmath150 are shown in figure 14 ( e ) and ( f ) , respectively .
the shock front 2 appears when the injected shock wave have passed over the cloud at @xmath183 , where @xmath184 is the shock passage timescale .
the ratio of @xmath182 to @xmath181 in mhd cases are much smaller than unity .
for example , @xmath185 is 0.29 at @xmath186(@xmath187 ) in run @xmath85 ( figure 14 ( e ) ) . in hydrodynamic case , on the other hand
, @xmath182 is similar to @xmath181 .
@xmath185 is 0.70 at @xmath188(@xmath187 ) in run @xmath150 ( figure 14 ( f ) ) . in hydrodynamic case ,
shock sweeping acceleration mechanism that is driven by the propagation of shock front 1 in figure 14 ( f ) is almost canceled out by propagation of shock front 2 after @xmath189 .
taking above effect into account , the phenomenological model of cloud acceleration in hydrodynamic case is modified from model @xmath34 to be as follows , @xmath190 the time evolution of prominence velocity @xmath88 in simulation run @xmath150 is shown as a solid line of figure 15 ( a ) , together with that of the phenomenological model model @xmath150 discussed above shown as a dashed line .
model @xmath150 captures abrupt acceleration characteristics before @xmath191 , but underestimates the acceleration in the later phase . in model
@xmath150 , the acceleration decrease because of relative velocity @xmath192 become smaller with time , but in run @xmath150 , the acceleration is almost constant after @xmath191 .
this is partly because of the deformation ( flattening and streaching ) of the prominence in hydrodynamics case . if we denote the cross sectional area of the cloud in yz plane as @xmath193 , the magnitude of aerodynamic drag force that mainly accelerate the cloud in run @xmath150 is of the order of @xmath194 . in model @xmath150
, we assumed the prominence has a spherical shape , although in run @xmath150 , the cloud is flattened and stretched in yz plane .
we study @xmath195 as an effective cross sectional area of the prominence in yz - plane where aerodynamic drag works to accelerate the prominence in x - direction .
figure 15 ( c ) shows the time evolution of @xmath196 .
@xmath196 at @xmath197 is almost 5 times that of initial @xmath196 .
we note that in run @xmath150 , drag acceleration mechanism works much more than in model @xmath150 because @xmath198 evolves in time as discussed above ( figure 15 ( c ) ) .
the internal flow structure is also different in hydrodynamic case compared with mhd case .
figure 15 ( d ) shows time evolution of velocity dispersions @xmath199 and @xmath200 as a proxy of internal flow within the prominence . in run
@xmath150 , @xmath159 and @xmath160 increase with time after the shock passage .
this is different from the results in run @xmath85 where the velocity dispersions damp in an exponential manner due to the presence of magnetic field .
coronal shock waves that activate prominences in reality are not blast waves but wave packets with finite width .
the motion of activated prominence depends not only on plasma velocity of the shock but also on wave packet width . in this secsion
, we analyze 3d mhd simulation results that reproduced prominence activation by triangular wave packet , and compare with phenomenological model .
the density profile of the prominence in this simulation ( let s call it run @xmath201 further on ) is the same with that in simulation run @xmath85 . in the simulation run @xmath201 , we have a triangular wave packet with velocity amplitude @xmath57 and wave packet width @xmath202 that activate prominence . the initial plasma velocity distribution for simulation run @xmath201 is as follows ( solid line in figure 16 ( a ) ) , @xmath203 , with the wave packet width being @xmath204 ( figure 16 ( a ) ) .
the density , pressure and magnetic field in the corona are all uniform , initially ( solid lines in figure 16 ( b ) , ( c ) and ( d ) ) . the initial condition is a `` superposition '' of two triangular fast mode wave packets propagating in opposite directions to each other with velocity amplitude @xmath57 and wave packet width @xmath202 . if the wave packets were linear ones ( i.e. @xmath205 ) , the wave propagating in the positive x - direction would keep its velocity amplitude and wave packet width unchanged without any interaction with oppositely - directed wave packet . in simulation run
@xmath201 , the wave packet propagating in the positive x - direction interact with the prominence .
we checked by nonlinear 1d mhd numerical simulation ( without prominence ) how initially superposed wave packets ( shown as solid lines in figure 16 ) evolve in time . dashed and
dash - dotted lines in figure 16 denote plasma parameters distribution at times @xmath206 and @xmath207 , respectively in the 1d simulation ( figure 16 ) .
we see in figure 16 ( a ) that the initial single peak of superposed wave packets in plasma velocity split into two oppositely - directed wave packets as expected from linear theory . because of the non - linearity
, however , the velocity amplitude decays slowly and the wave packet broadens as they propagate .
we make the phenomenological model that describes prominence center of mass motion in simulation run @xmath201 .
we call it model @xmath201 .
the model @xmath201 can be obtained simply by replacing @xmath57 and @xmath117 in model @xmath85 with @xmath208 and @xmath209 , respectively .
@xmath208 is the plasma velocity in the corona around the prominence shocked by triangular wave packet and approximated as follows , @xmath210 with @xmath211 being the `` wave packet passage '' time scale .
@xmath212 is an approximation based on the fact that the injected coronal shock is weak and the activated prominence moves much slower than coronal fast mode phase speed .
the solid and dashed lines in figure 17 ( a ) show prominence center of mass acceleration by magnetic tension force @xmath146 and that by total pressure gradient force @xmath145 in simulation run @xmath201 , respectively .
the solid and dashed lines figure 17 ( b ) show prominence center of mass acceleration by magnetic tension mechanism @xmath119 and that both by shock sweeping and fluid drag mechanisms @xmath213 , respectively .
we find from the plot in figure 17 ( a ) that the prominence is mainly accelerated by magnetic tension force first , and then decelerated also by magnetic tension force .
the model @xmath201 captures such a characteristic response of the prominence to wave packet injection .
we find some oscillations in both @xmath146 and @xmath145 in simulation run @xmath201 .
they result from multiple reflections of wave packets within the prominence . the effect of multiple reflections is not included in model @xmath201 .
solid and dashed lines in figure 18 ( a ) show time evolution of prominence center of mass position @xmath214 in simulation run @xmath201 , and that expected from model @xmath201 , respectively .
solid and dashed lines in figure 18 ( b ) shows the prominence center of mass speed @xmath88 in simulation run @xmath201 , and that expected from model @xmath201 , respectively . from figure 18
, we see that the center of mass motion of activated prominence expected with model @xmath201 quantitatively agreed with those in simulation run @xmath201 .
we call the time interval during which the prominence is accelerated to its maximum speed by the shock wave as `` acceleration phase '' of the prominence activation .
the acceleration phase continued until @xmath215 in run @xmath201 ( figure 18 ( b ) ) .
solid , dashed and dash - dotted lines in figure 19 show time evolution of velocity dispersions @xmath159 , @xmath160 and @xmath161 in simulation run @xmath201 , respectively .
exponential damping timescales are estimated for the three velocity dispersions by least square fit by bisector method during the time between @xmath84 and @xmath164 .
the resultant slopes for exponential damping are shown as thick lines for each components .
damping times @xmath167 , @xmath166 and @xmath165 are listed in the last row in table 1 .
we find that the damping times for all three components are similar to the shock sweeping time scale @xmath148 .
we note that @xmath165 in simulation run @xmath201 is much smaller than that in run @xmath85 .
in this section , we try to diagnose coronal shock properties and prominence properties using prominence activation with the help of phenomenological model discussed above .
the figure 20 ( a ) is a time - distance diagram of prominence activation made from aia 193@xmath3 pass band images .
this corresponds to the white rectangle in figure 3 ( i ) .
the white crosses denote prominence positions during the prominence activation estimated by the eye .
we denote the prominence displacement in the plane of the sky at time @xmath216 ( @xmath217 ) as @xmath218 . on the other hand ,
the position of the prominence activated by triangular wave packet can be predicted by the phenomenological model discussed in the previous section . here
, we try to fit observed time evolution of activated prominence position with the phenomenological model expectation .
i used equation 31 modified with @xmath219 as described in section 3.7 for the fitting .
we assume coronal temperature to be @xmath220 k. this leads to the coronal sound speed to be @xmath221 km s @xmath4 , assuming the specific heat ratio to be @xmath19 .
the local density gap between corona and prominence is assumed to be @xmath222 .
we think of two different values for the angle @xmath223 between the line of sight and shock propagation direction , that are @xmath224 , @xmath225 .
the thickness of the prominence core seen as dark structure in aia 193 @xmath3 images are about 10 , which correspond to the estimated prominence radius of @xmath226 km .
the fast mode wave speed in the corona @xmath227 is estimated to be @xmath228 , with @xmath229 km s@xmath4 being wave propagation speed in the plane of sky .
the coronal fast mode wave speeds in @xmath224 and @xmath225 cases are @xmath230 km s@xmath4 and @xmath231 km s@xmath4 , respectively . the plasma beta is obtained with @xmath232 , with @xmath233 is a coronal alfven speed in perpendicular propagation case . in @xmath224 and @xmath234 cases ,
plasma beta is calculated to be @xmath235 and @xmath236 , respectively .
assuming shock propagation direction and the center of mass velocity of activated prominence is parallel ( which is correct in perpendicular shock case ) , the displacements of the activated prominence at time @xmath216 are estimated to be @xmath237 .
then , the number of remaining free parameters of the phenomenological model are three , that are prominence volume filing factor @xmath238 , compression ratio of coronal shock wave @xmath239 and wave packet width @xmath202 .
we denote the time evolution of activated prominence position expected from phenomenological model as @xmath240 .
we searched best - fit values for @xmath241 and @xmath202 with which the sum of squared residuals @xmath242 is minimized in the parameter space @xmath243 $ ] , @xmath244 $ ] and @xmath245 $ ] , with @xmath5 being a solar radius .
the fitting results for @xmath224 and @xmath234 cases are shown in figure 20 ( b ) and ( c ) . as a best - fit parameters , we estimate @xmath239 and @xmath202 of coronal shock wave to be 1.17 and 0.16 @xmath5 in @xmath224 case and 1.17 and @xmath246 in @xmath234 case , with @xmath5 being solar radius .
best - fit parameters for @xmath224 and @xmath234 cases are listed in table 2 .
fast mode mach number @xmath247 of the coronal shock in @xmath224 and @xmath234 cases are 1.12 and 1.13 , respectively . from mass conservation at the shock front , velocity amplitude of injected triangular wave
is expressed as @xmath248 . from this , the plasma velocity amplitude for injected triangular wave is estimated as 62 km s@xmath4 and 88 km s@xmath4 in @xmath224 and @xmath234 cases , respectively .
the best - fit @xmath238 in both cases were @xmath249 , which is the smallest value in the free parameter space .
based on the phenological model ( equation 31 ) , on the other hand , when @xmath238 is much smaller than unity , the prominence is accelerated to its maximum speed almost within the shock sweeping time scale @xmath148 , while the subsequent prominence deceleration occurs within the wave packet passage timescale @xmath250 . with the estimated parameters of @xmath251 , @xmath252 km , @xmath253 km s@xmath4 and @xmath254 , the acceleration and deceleration timescales are roughly , @xmath255s and @xmath256s , respectively . in order to estimate @xmath238 correctly
, we have to time - resolve the acceleration phase whose time scale @xmath179 reflects @xmath238 directly .
compared with the estimated acceleration timescale of @xmath257 s , the aia time cadence of @xmath258 s is not high enough to track the acceleration phase of the prominence activation in this event , although we can track the subsequent deceleration phase with sufficient time resolution .
as seen in the time - distance plot of figure 20 ( a ) , the prominence appears to be accelerated to its maximum speed right after the arrival of the shock .
we think that the evaluated value of @xmath238 in this analysis is not accurate enough due to the lack of fully time - resolved observation of the acceleration phase of the prominence activation in this event .
then , we estimate the energy of the coronal shock wave @xmath259 associated with the x5.4 flare .
the coronal emission measure near the prominence before the arrival of the shock was @xmath260 @xmath261 .
the emission measure is calculated at point @xmath262 in figure 3 ( a ) , based on a method proposed in @xcite .
assuming a line - of - sight distance @xmath263 to be of order of the distance to the ( solar ) horizon as seen from a point above the photosphere by a pressure scale hight ( @xmath264 mm ) , we get @xmath265 mm .
the coronal proton number density @xmath266 at point @xmath262 is estimated from the relation @xmath267 to be @xmath268 @xmath269 .
corresponding coronal mass density is @xmath270 g @xmath269 , with @xmath271 g being the proton mass .
the energy flux of the shock at point a is @xmath272 erg @xmath273 s@xmath4 .
the time scale for the triangular wave to pass through the fixed point a is @xmath274 s. the surface area of the spherically expanding dome of shock front in the corona is approximated as @xmath275 , with l being the distance between flaring ar and point a. we approximate @xmath276 and get @xmath277 @xmath278 . from above , the energy budget of the coronal shock wave is estimated as @xmath279 erg .
the total energy released in x5.4 class flare is roughly estimated by empirical relation between flare energy and flare soft x - ray peak flux to be several @xmath280 erg @xcite .
the estimated energy budget of the globally propagating shock wave in the corona is a few percent of total energy released during the flare .
recent high time and spatial resolution euv observation of solar corona by _
sdo_/aia enabled us to study in detail the time evolution of coronal shock wave associated with flares .
we can now study as well the interaction between coronal shock wave and prominences using aia . in this paper
, we studied the excitation process of large amplitude prominence oscillation through the interaction between the prominence and coronal shock wave , with the help of three - dimensional mhd simulation .
the x5.4 class flare occurred on march 7 , 2012 was associated with very fast cme with its speeds of about @xmath281 km s@xmath4 estimated from coronagraph observation by soho / lasco .
a global shock front is formed around the expanding cme ejecta .
the shock front had a dome - like form , especially with bright structure propagating to the north at the foot of the dome .
the northward disturbance hit a polar prominence , leading to the excitation of large amplitude prominence oscillation . during the prominence activation ,
the prominence was strongly brightened , receiving momentum in the direction of shock propagation . in order to explain the observational signature of prominence activation
, we have done a three - dimensional mhd simulation of coronal fast mode shock - prominence interaction . especially , the momentum transfer mechanism from the shock to the prominence is studied in detail . the shock injection into the prominence material compressed and accelerated the prominence .
the velocity shear at the corona - prominence boundary resulted in kelvin - helmholtz instability .
khi was stabilized by magnetic tension force in the plane containing the initial magnetic field lines . by analyzing the simulation results and comparing them with phenomenological models , magnetic tension force acceleration
was also found to be very important .
the accelerated prominence velocity asymptotes to the value of coronal shocked plasma velocity when the shock is a blast wave after some timescale depending on different prominence density .
when the volume filling factor is small like in runs @xmath95 and @xmath96 , the acceleration timescale is longer compared with the case with uniformly distributed prominence density in runs @xmath93 and @xmath94 .
this may be because of the suppression of shock sweeping acceleration mechanism in clumpy cloud in runs @xmath33 .
both total surface area of clumps and local density gap between clump and corona in runs @xmath33 are larger than those in runs @xmath34 .
this make it difficult for injected shock fronts in runs @xmath33 to penetrate deep into the cloud as a whole so that they could exchange momentum with cloud materials .
when the volume filling factors are larger than 0.3 , the resultant time evolution of the mean velocity in runs @xmath87 and @xmath32 are very similar to that of runs @xmath85 and @xmath23 .
we also studied the time evolution of the velocity dispersion of shocked prominence material in each ( @xmath29 , @xmath282 and @xmath283 ) component .
the velocity dispersion is excited during the shock sweeps through the cloud and then damps almost exponentially .
the exponential damping time scales of velocity dispersions in each component @xmath167 , @xmath166 and @xmath165 are estimated and summarized in table 1 . in
almost all the simulation runs , @xmath167 and @xmath166 are roughly comparable to the shock - sweeping time scale @xmath148 while @xmath165 is about twice as large as @xmath148 .
a possible reason for the discrepancy of the damping times among the components is as follows .
when the randomized flow is directed to positive x - direction at a certain time and location , the magnetic field lines originally directed to z - direction ( @xmath284 ) will be distorted there in xz plane resulting in the electric current directed in negative y - direction ( -@xmath285 ) .
the lorentz force @xmath286 act in negative @xmath29-direction which pull back the flow originally directed to positive x - direction .
this helps damp the velocity dispersion in x - direction , making @xmath167 small .
the same damping mechanism works on the y - component of velocity dispersion as well , but not on z - component .
in interplanetary shock - cloud interaction , both shock sweeping mechanism and fluid drag force accelerate the cloud .
khi is also important in mixing mc materials which will affect star formation process taking place . in solar coronal shock - prominence interaction ,
magnetic tension force is more important in accelerating the prominence than fluid drag force , because in prominence activation , plasma beta is typically smaller than unity and the shock not strong .
when the prominence has internal density structure like in runs @xmath33 , plasma mixing might also occur in short time scale .
when the plasma beta is much larger than unity ( which is a reasonable assumption in some molecular clouds ) , the cloud acceleration and internal flow excitation shows much different characteristics .
the shock - sweeping acceleration mechanism is effective only during the shock passage time @xmath287 , and the pressure gradient force due to the velocity difference between the cloud and the ambient plasma works as an accelerator .
the cloud is flattened with the help of ambient flow that converge towards the cloud along x - axis and diverges in yz plane associated with a coherent vortex formed behind the cloud .
the flattening effect increases the effective cross section of the cloud , helping the cloud acceleration by fluid drag
. the excited internal flow does not decay in hydrodynamic simulation run @xmath150 , though in mhd simulation runs @xmath85 , the internal flow damps in an exponential manner mainly due to the lorentz force . in reality , the coronal shock wave that activate a prominence is not a blast wave as studied in runs @xmath34 or @xmath33 but a wave packet with finite wave packet width .
we studied interaction between a coronal shock wave in the form of triangular - shaped wave packet and a prominence in simulation run @xmath201 .
the shocked prominence is first accelerated and then decelerated by magnetic tension and total pressure gradient force in the simulation run @xmath201 .
the phenomenological model @xmath201 well captures the characteristic dynamics of the prominence center of mass both in acceleration and deceleration phases , but slightly underestimate the impact of pressure gradient force . especially , the phenomenological model @xmath201 does not reproduce the prominence deceleration by total pressure gradient force which is present in run @xmath201 .
one of the possible reason for the discrepancy is that the phenomenological model neglects the effect of multiple reflection of transmitted shock wave within the prominence .
we compared the prominence center of mass position and velocity in simulation run @xmath201 with those expected by phenomenological model @xmath201 and found quantitative agreement between the simulation and the model .
we tracked the time evolution of the position of the activated prominence and fitted it with phenomenological model .
the best - fit curve for prominence movement agreed well with observation .
as a best fit parameters , we obtained prominence volume filling factor @xmath238 , coronal shock compression ratio @xmath239 and wave packet width @xmath202 .
the resultant compression ratio and fast mode mach numbers of the coronal shock were 1.17 and 1.12 in @xmath224 case , and were 1.17 and 1.13 in @xmath234 case , respectively .
the estimated wave packet width of the coronal shock were 0.16@xmath5 and 0.22@xmath5 in the cases of @xmath224 and @xmath234 , respectively .
they are comparable to typical euv wave front widths of @xmath288 mm which are suggestive of coronal shock waves reported in @xcite .
both the estimated coronal shock propagation speed and the plasma velocity amplitude of @xmath289 km s@xmath4 and @xmath290 km s@xmath4 are reasonable values as a weak fast mode coronal shock wave .
the shock wave that activated the prominence had likely been driven by the lateral expansion of cme ejecta in the lower corona ( whose speed is much smaller than the radial ejection speed ) and had propagated a considerable distance of @xmath291 from the source active region .
this is a possible reason why the coronal shock near the prominence was weak although the associated cme was extremely fast with its speed of almost @xmath292 km s@xmath4 at a hight of @xmath293 estimated with _
soho_/lasco coronagraph observations .
the best - fit value of prominence filling factor on the other hand was @xmath251 , which we do nt think is accurate partly because the observational data did not time - resolved the acceleration phase of activated prominence , which is vital for determining prominence @xmath238 based on our model . with the help of emission measure analysis
, we estimated the energy of the coronal shock to be @xmath294 erg in this event .
this was roughly several percent of total released energy during the x5.4 flare .
the physical mechanism that mainly work in prominence activation differs with different wave packet width .
if the width of the wave packet is longer than @xmath295 , magnetic tension force is the most important in accelerating the prominence .
the coronal magnetic loop that support the prominence material against gravity in the corona is rooted at the photosphere on both sides , which will reduce magnetic tension force acceleration after alfven travel time @xmath296 where l is the length of the magnetic loop .
we note that @xmath297 is of order of period of lapo ensuing the prominence activation . in this paper
, we discussed the dynamics of prominence - coronal shock interaction which leads to lapo .
the prominence activation event we studied was triggered by the arrival of coronal shock wave at a polar prominence that propagated from faraway ar corona .
such globally propagating flare - associated shock waves in the corona or lapos are relatively rare phenomena . on the other hand ,
small scale magnetic explosions ( small flares and jets ) always occur in the corona .
we expect the interactions between solar prominences and small amplitude shock waves generated by such small scale magnetic explosions are always occurring in the corona , and might play a role in driving small amplitude prominence dynamics such as small amplitude oscillations and chaotic movements of plasma elements .
the sdo / aia data are courtesy of nasa / sdo and aia science team .
the simulation code used in this work is created with the help of hpci strategic program .
numerical computations were carried out on cray xc30 at center for computational astrophysics , national astronomical observatory of japan .
the author is gratefull to dr .
kazunari shibata and dr .
ayumi asai of kyoto university for their helpful comments and discussions .
the author is also grateful to the journal referee for his / her careful reading and lots of comments which significantly improved the paper .
this work is financially supported by the grant - in - aid for jsps fellows 15j02548 .
crrrrrrrrrrr & & & & & & & & & + run @xmath93 & @xmath298 0.05 & @xmath298 5.95 & @xmath298 4.28 & @xmath299 4.28 & @xmath299 4.57 & @xmath298 9.14 & @xmath298 1.00 & @xmath299 1.07 & @xmath299 2.14 & @xmath299 + run @xmath94 & @xmath298 0.10 & @xmath298 10.9 & @xmath298 5.79 & @xmath299 8.06 & @xmath299 7.19 & @xmath298 7.71 & @xmath298 1.39 & @xmath299 1.24 & @xmath299 1.33 & @xmath299 + run @xmath85 & @xmath298 0.30 & @xmath298 30.7 & @xmath298 9.72 & @xmath299 12.4 & @xmath299 14.8 & @xmath298 26.6 & @xmath298 1.28 & @xmath299 1.52 & @xmath299 2.73 & @xmath299 + run @xmath23 & @xmath298 1.00 & @xmath298 100 . & @xmath298 17.5 & @xmath299 & @xmath299 & @xmath298 & @xmath298 & @xmath299 & @xmath299 & @xmath299 + run @xmath95 & @xmath298 0.05 & @xmath298 5.95 & @xmath298 4.28 & @xmath299 5.31 & @xmath299 3.30 & @xmath298 8.67 & @xmath298 1.24 & @xmath299 0.77 & @xmath299 2.03 & @xmath299 + run @xmath96 & @xmath298 0.10 & @xmath298 10.9 & @xmath298 5.79 & @xmath299 6.60 & @xmath299 5.32 & @xmath298 12.0 & @xmath298 1.14 & @xmath299 0.92 & @xmath299 2.07 & @xmath299 + run @xmath87 & @xmath298 0.30 & @xmath298 30.7 & @xmath298 9.72 & @xmath299 16.0 & @xmath299 10.0 & @xmath298 12.9 & @xmath298 1.65 & @xmath299 1.02 & @xmath299 1.32 & @xmath299 + run @xmath32 & @xmath298 1.00 & @xmath298 100 . & @xmath298 17.5 & @xmath299 & @xmath299 & @xmath298 & @xmath298 & @xmath299 & @xmath299 & @xmath299 + run @xmath201 & @xmath298 0.30 & @xmath298 30.7 & @xmath298 9.72 & @xmath299 9.94 & @xmath299 8.54 & @xmath298 10.3 & @xmath298 1.02 & @xmath299 0.88 & @xmath299 1.06 & @xmath299 + crrrrrrrrrrr & & & & & & + @xmath300 & @xmath298 0.33 & @xmath301 0.01 & @xmath301 1.17 & @xmath301 0.16 & @xmath301 1.12 & @xmath298 62 km s@xmath4&@xmath302 + @xmath303 & @xmath298 0.15 & @xmath301 0.01 & @xmath301 1.17 & @xmath301 0.22 & @xmath301 1.13 & @xmath298 88 km s@xmath4&@xmath302 + passband image by _ sdo_/aia .
we can see the shock front surrounding the cme ejecta .
( c ) _ sdo_/aia 193 @xmath304 band image showing the emission from 1mk coronal plasma at 00:18:19ut .
the pixels around flaring active region ar11429 are saturated .
the black rectangle r1 shows the fov of figure 3 ( a ) - ( h ) .
( d ) difference image made from two successive snapshots of aia 193 @xmath304 passband taken at 00:18:19ut and 00:18:07ut . a dome - like disturbance expanding above ar11429
is clearly seen .
[ flare],scaledwidth=100.0% ] ( panels ( a ) and ( c ) ) and @xmath87 ( panels ( b ) and ( d ) ) . panels ( a ) and
( b ) show initial ( @xmath306 ) density distribution , while panels ( c ) and ( d ) shows the one at @xmath307(@xmath308 ) .
we note that the range of the x - axis of the plots in panels ( a ) and ( b ) is between -6.5 and 0.5 , while that in panels ( c ) and ( d ) is between -2 and 5 .
the values in the color bars corresponds to the logarithm of prominence density .
the direction of initial magnetic field @xmath309 is also shown in panels ( a ) and ( b ) .
[ flare],scaledwidth=100.0% ] at @xmath84 .
arrows represent velocity fields in the plane of the plot .
logarithmic density distributions in xy- or xz- planes are shown with color contours in panels ( a ) and ( b ) , respectively . in panel
( c ) and ( d ) , vorticity components perpendicular to xy- or xz- planes are shown with color contours , respectively . in panel ( e ) , total plasma pressure ( magnetic pressure@xmath310gas pressure ) in xy plane is shown .
panel ( f ) is the schematic figure shows how kelvin - helmholtz instability evolves along the velocity shear layer in xy - plane .
the resultant formation of low pressure region behind the cloud is also indicated .
[ flare],scaledwidth=100.0% ] ( left column ) and @xmath87 ( right column ) .
the density distributions are shown in logarithmic scale with color contours .
white arrows in each panels show the magnetic field vector in the plane of the plots .
the panels in third and fifth rows are plots in @xmath82 plane , while other panels shows plots in @xmath81 plane .
simulation time for each snapshot in unit of shock sweeping time scale @xmath148 is shown in brackets .
[ flare],scaledwidth=100.0% ] and phenomenological models @xmath34 .
top two rows shows density distribution of run @xmath85 in logarithmic scale with color contours .
arrows in panels ( a ) and ( b ) show magnetic fields and those in panels ( c ) and ( d ) show velocity fields , both in xz plane . panels
( e ) and ( f ) are schematic figures of phenomenological prominence activation model during ( @xmath103 ) and after ( @xmath311 ) the shock sweeping acceleration works .
the coronal shock front that passed the cloud is distorted as shown in ( e ) , while the shock front transmitted into the prominence material is reflected back and forth at prominence - corona boundary as shown in ( e ) and ( f ) .
[ flare],scaledwidth=100.0% ] and magnetic tension force acceleration @xmath146 in simulation runs @xmath34 and @xmath33 are shown as dashed and solid lines , respectively . in the bottom row , summation of shock sweeping acceleration and fluid drag acceleration @xmath213 and magnetic tension force acceleration @xmath119 in models @xmath34 are shown as dashed and solid lines , respectively .
[ flare],scaledwidth=100.0% ] ( solid lines ) , @xmath160 ( dashed lines ) and @xmath161 ( dash - dotted lines ) in runs @xmath34 and @xmath33 .
exponential damping timescales are estimated for the three variables in each runs by least square fit during the time between @xmath84 and @xmath164 in runs @xmath34 and @xmath33 ( @xmath312 is from 1 to 3 ) .
the resultant slopes for exponential damping are shown as thick lines for each variables .
estimated exponential damping time scales @xmath167 , @xmath166 and @xmath165 are listed in table 1 .
[ flare],scaledwidth=100.0% ] at @xmath314 .
the values in the color bars corresponds to logarithm of density .
panel ( b ) and ( c ) show the density ( in logarithmic scale ) and pressure distributions in @xmath81 plane in simulation run @xmath150 at @xmath314 , respectively .
the arrows in panels ( b ) and ( c ) shows velocity field in @xmath81 plane .
[ flare],scaledwidth=80.0% ] at @xmath315 ( @xmath316 ) .
both density and pressure distribution is shown with color contour in logarithmic scale .
in - plane velocity field is shown with white arrows in each panels .
panels ( b ) and ( d ) are those of run @xmath150 at @xmath317 ( @xmath318 ) .
panel ( e ) and ( f ) show the distribution of @xmath319 along x - axis in runs @xmath85 and @xmath150 , respectively .
the velocity amplitudes within prominence material for the shock fronts transmitted from ahead and behind of the prominence is indicated as @xmath181 and @xmath182 , respectively .
[ flare],scaledwidth=100.0% ] , ( b ) : center of mass acceleration @xmath320 , ( c ) : index of cross sectional area @xmath196 and ( d ) : velocity dispersions @xmath159 ( solid line ) and @xmath160 ( dashed line ) of the cloud in hydrodynamic run @xmath150 . in panel
( a ) , prominence velocity expected from phenomenological model ( dashed line ) and the simulation result ( solid line ) are compared .
[ flare],scaledwidth=100.0% ] without the prominence .
we see the time evolution of wave packets that propagate to opposite directions with each other .
the spacial distribution of @xmath319 , @xmath92 , @xmath321 and @xmath284 are shown in panels ( a ) , ( b ) , ( c ) and ( d ) , respectively .
the solid , dashed and dash - dotted lines in each panels represents physical values at @xmath3220 , 2 and 6 , respectively .
[ flare],scaledwidth=100.0% ] and that by total pressure gradient force @xmath145 in simulation run @xmath201 , respectively .
the solid and dashed lines in panel ( b ) show prominence acceleration by magnetic tension mechanism @xmath119 and that both by shock sweeping and fluid drag mechanisms @xmath213 expected from phenomenological model @xmath201 , respectively .
[ flare],scaledwidth=100.0% ] ( solid lines ) , @xmath160 ( dashed lines ) and @xmath161 in simulation run @xmath201 .
exponential damping timescales are estimated for the three variables in each runs by least square fit during the time between @xmath84 and @xmath164 .
the resultant slopes for exponential damping are shown as thick lines for each variables .
[ flare],scaledwidth=100.0% ] and @xmath234 , respectively .
black crosses in panel ( b ) and ( c ) are the extracted position from observation , and the solid lines are the best - fit curves from the phenomenological model . the fitting result is listed in table 2 .
[ flare],scaledwidth=100.0% ] | flare associated coronal shock waves sometimes interact with solar prominences leading to large amplitude prominence oscillations .
such prominence activation gives us unique opportunity to track time evolution of shock - cloud interaction in cosmic plasmas .
although the dynamics of interstellar shock - cloud interaction is extensively studied , coronal shock - solar prominence interaction is rarely studied in the context of shock - cloud interaction . associated with x5.4 class solar flare occurred on 7 march , 2012 , a globally propagated coronal shock wave interacted with a polar prominence leading to large amplitude prominence oscillation . in this paper , we studied bulk acceleration and excitation of internal flow of the shocked prominence using three - dimensional mhd simulations .
we studied eight magnetohydrodynamic ( mhd ) simulation runs with different mass density structure of the prominence , and one hydrodynamic simulation run , and compared the result . in order to compare observed motion of activated prominence with corresponding simulation , we also studied prominence activation by injection of triangular shaped coronal shock .
we found that magnetic tension force mainly accelerate ( and then decelerate ) the prominence .
the internal flow , on the other hand , is excited during the shock front sweeps through the the prominence and damps almost exponentially .
we construct phenomenological model of bulk momentum transfer from shock to the prominence , which agreed quantitatively with all the simulation results .
based on the phenomenological prominence - activation model , we diagnosed physical parameters of coronal shock wave .
the estimated energy of the coronal shock is several percent of total energy released during the x5.4 flare .
= 1 |
the hard - disk system is a fundamental model of statistical and computational physics . during more than a century
, the model and its generalization to @xmath0-dimensional spheres have been central to many advances in physics .
the virial expansion is an example : boltzmann s early calculations of the fourth virial coefficient @xcite ultimately led to lebowitz and onsager s proof of the convergence of the virial expansion up to finite densities @xcite for all @xmath0 and to the general and systematic study of virial coefficients .
the theory of phase transitions provides another example for the lasting influence of the hard - disk model and its generalizations .
kirkwood and monroe @xcite first hinted at the possibility of a liquid
solid transition in three - dimensional hard spheres .
this prediction was surprising because of the absence of attractive interactions in this system .
the depletion mechanism responsible for the effective - medium attraction was also first studied in hard spheres , by asakura and oosawa @xcite . in two dimensions , the liquid
solid phase transition was first evidenced by alder and wainwright @xcite .
it lead to far - reaching theoretical @xcite , computational @xcite and experimental @xcite work towards the understanding of 2d melting . in mathematics , hard disks and hard spheres
have also been at the center of attention @xcite . a rigorous existence proof of the melting transition in hard spheres is still lacking , but the ergodicity of the molecular dynamics evolution of this system has now been established rigorously @xcite .
arguably the most important role for the hard - disk model has been in the development of numerical simulation methods .
molecular dynamics @xcite and markov - chain monte carlo @xcite were first formulated for hard disks .
the early algorithms have continued to be refined : within the molecular dynamics framework , this has lead to highly efficient event - scheduling strategies @xcite and , for monte carlo , to the development of cluster algorithms @xcite .
even the modern simulation algorithm remain slow , however , and revolutions like the cluster algorithms for spin systems @xcite have failed to appear .
moreover , rigorous mathematical bounds for the correlation time ( mixing time ) of monte carlo algorithms were obtained in the thermodynamic limit only for small densities @xcite , which are far inside the liquid phase . at higher densities , close to the liquid solid transition , many numerical calculations have suffered from insufficient simulation times until recently @xcite . in the present article
, we discuss computational aspects of the hard - disk model , starting with an introduction ( section [ s : algos ] ) .
in particular , we reinterpret hard - sphere monte carlo in terms of the sampling of points from high - dimensional polytopes ( section [ s : polytope ] ) .
local monte carlo amounts to random walks in a sequence of such polytopes , while event - chain monte carlo is equivalent to molecular dynamics evolutions with particular initial conditions for the velocities .
we analyze the convergence properties of the algorithms in these polytopes for the hard - disk case .
parallel event - chain algorithms emerge naturally as molecular dynamics with more general initial conditions ( section [ s : parallel ] ) .
we describe several parallelization strategies and report on implementations .
[ s : algos ] we consider @xmath1 equal hard disks of unit radius @xmath2 in a square box of size @xmath3 . in the following ,
we assume without mentioning periodic boundary conditions for positions and pair distances .
the statistical weights @xmath4 are equal to unity for configurations @xmath5 without overlaps ( all pair distances larger than @xmath6 ) and zero for illegal configurations ( with overlaps ) .
the phase diagram of the system depends only on the packing fraction @xmath7 . in the following , the letters @xmath5 , @xmath8 , @xmath9 ,
, label hard - disk configurations of @xmath1 disks , given by the coordinates of the disk centers @xmath10 .
the letters @xmath11 , @xmath12 , @xmath13 number disks . between configurations @xmath5 and @xmath8 ( each arrow stands for a probability flow of same magnitude ) .
_ left _ : global balance , as required for markov - chain monte carlo . the total flow @xmath14 into the configuration @xmath5 equals the flow @xmath15 out of it .
_ center _ : detailed balance : the net flow between any two configurations is zero , @xmath16 . _ right _ : another special case of global balance : maximal global balance at @xmath5 ( @xmath17 ) .
[ f : flows_detailed_global ] ] [ s : balances ] markov - chain monte carlo algorithms are governed by balance conditions for the flows @xmath18 from configuration @xmath5 to configuration @xmath8 ( see fig . [
f : flows_detailed_global ] ) ; @xmath19 is the conditional probability to move from @xmath5 to @xmath8 , given that the system is in @xmath5 . to converge towards the stationary distribution @xmath4 ,
the _ global _ balance condition must be satisfied : the total flow onto configuration @xmath5 must equal the total flow out of @xmath5 , @xmath20 the local monte carlo algorithm , introduced by metropolis et al . in 1953
@xcite ( see fig . [
f : event_chain_move ] ) , uses the more restrictive _ detailed _ balance condition @xmath21 for which the net flow between each pair of configurations @xmath5 and @xmath8 is zero . moving from configuration @xmath22 to @xmath23 involves sampling the disk @xmath11 to be displaced and the displacement @xmath24 . for detailed balance , the probability to sample @xmath24 at @xmath10 must equal the probability to sample @xmath25 at position @xmath26 . in order to be ergodic , the displacements @xmath24
are chosen such that each disk can eventually reach any position in the system . .
with periodic boundary conditions , the event - chain move is rejection - free .
, width=529 ] unlike the local monte carlo algorithm , a single move of the event - chain algorithm @xcite may displace several disks .
an event - chain move is parametrized by a total displacement @xmath27 and a direction @xmath28 , which together form a vector @xmath29 .
the move starts by sampling a disk @xmath11 and `` sliding '' it in the @xmath30 direction until it hits another disk @xmath12 , or at most for the distance @xmath27 .
the disk @xmath12 is then displaced in its turn , also in the @xmath28 direction , see fig .
[ f : event_chain_move ] .
this process continues until the displacements of the individual disks sum up to @xmath27 .
after this , a new disk and possibly a new direction are sampled for the next move . with periodic boundary conditions ,
no rejections occur in this algorithm . for a given displacement vector @xmath31
, any disk configuration @xmath5 can reach @xmath1 other configurations , using each of the @xmath1 disks to start an event chain .
likewise , @xmath5 can be reached from @xmath1 other configurations which may be reconstructed by event chains with displacement vector @xmath32 .
this implies that the event - chain satisfies the global balance condition , eq .
( [ e : globalbal ] ) .
if the vectors @xmath33 are equally likely , it also satisfies detailed balance . in order to be ergodic ,
the displacements @xmath31 must span space : by choosing @xmath34 , the event - chain algorithm realizes the maximal global balance ( see fig .
[ f : flows_detailed_global ] ) , where flow between two configurations is possible only in one direction .
this version is more efficient than detailed balance versions ( for example , @xmath35 and @xmath36 ) @xcite .
it is again possible to alternate repeated moves in the @xmath37 direction with repeated moves in @xmath38 without destroying the correctness of the algorithm . for displacements
@xmath27 smaller than the mean free path @xmath39 , the event - chain algorithm is roughly equivalent to the local monte carlo algorithm .
it accelerates for increasing @xmath27 , and for @xmath27 much larger than the mean free path , it is about two orders of magnitude faster than the local monte carlo method , and about ten times faster than the best current implementations @xcite of event - driven molecular dynamics ( see ref .
@xcite ) .
disks in a square box with periodic boundary conditions at packing fraction @xmath40 .
_ left _ : disk configurations and their local orientational field @xmath41 for one simulation run .
frame are separated by @xmath42 iterations ( @xmath43 sweeps ) of local monte carlo .
the slow decorrelation of the orientation is manifest .
_ right _ : evolution of the global orientational order parameter @xmath44 , eq .
( [ e : global_psi ] ) , in the complex plane , for the same simulation run .
[ f : orientation_movie ] ] the characteristic challenge of numerical simulations for the hard - disk model resides in the extremely long correlation time .
this is illustrated in fig .
[ f : orientation_movie ] using snapshots of configurations obtained during a long simulation run .
the system is quite small and not extremely dense , yet correlations in the orientation of the system persist over millions of monte carlo moves . to quantify the orientations and their correlations
, we consider the local orientational field @xmath45 where @xmath46 is the number of voronoi neighbors of disk @xmath12 .
the @xmath47 ( with @xmath48 ) are normalized weights according to the length of the voronoi interface between disks @xmath12 and @xmath13 , and @xmath49 is the angle of the vector between the disk centers @xcite . the average of eq .
( [ e : local_psi ] ) over all disks yields the global orientational order parameter , @xmath50 in a square box , the mean value of @xmath44 is zero because of the @xmath51 symmetry , and its correlation function @xmath52 decays to zero for infinite times @xmath53 .
we conjecture that @xmath44 is the slowest observable in the system . for large times , global orientational correlations decay exponentially , @xmath54 , and we obtain the empirical correlation time @xmath55 from an exponential fit to @xmath56 .
[ s : polytope ] ) are @xmath57 and @xmath58 .
_ center _ : molecular dynamics evolution in the polytope corresponding to two event chains with moves of disk 1 ( blue segments ) and disk 2 ( red segments ) .
the trajectory begins with disk @xmath59 , and it depends on the choice of the starting disk ( @xmath59 or @xmath6 ) for the second chain .
periodic boundary conditions are ignored for clarity .
snapshots of the configuration are sketched along the trajectory .
_ right _ : hard - disk configuration with its constraint graph for motion along the @xmath60 axis .
each node has at most three forward and three backward links .
this graph is invariant under event - chain moves in the @xmath60 direction . ]
event - chain moves along a single direction @xmath61 sample a restricted configuration space . for the remainder of this section , we take the chains to move in the positive @xmath60 direction , unless specified otherwise , to simplify the notation .
since all @xmath62 coordinates are fixed , two disks whose @xmath62 coordinates differ by less than @xmath6 radii can not slide across each other , and their relative order is fixed . furthermore , while in @xmath60 collision mode , any disk can collide with not more than six other disks , at most three in the forward direction , and at most three in backward direction ( see fig .
[ f : polytope_ballistics_plus_graph ] ) .
the collision partners of a disk may include itself , because of boundary conditions .
the relations among disks constitute a _ constraint graph _ , which expresses the partial order between them ( see fig . [
f : polytope_ballistics_plus_graph ] ) .
this graph remains invariant while performing event - chain moves in the @xmath35 direction .
each directed edge from @xmath11 to @xmath13 corresponds to a linear inequality for the @xmath60 coordinates of the disks @xmath11 and @xmath13 : @xmath63 with @xmath64 .
the constant @xmath65 can be adjusted to also account for periodic boundary conditions in the @xmath60 direction .
the inequalities eq .
( [ e : polytopeequations ] ) imply that no more than three forward collision partners can be present collides forward with @xmath12 , and @xmath12 collides forward with @xmath13 , we have @xmath66 ; if now @xmath67 , the disks @xmath11 and @xmath13 can never come into contact ; disk @xmath12 _ covers _ disk @xmath13 . applying this rule iteratively ,
the disks in the forward direction can be reduced to at most three : at most one each with @xmath68 , with @xmath69 and with @xmath70 $ ] .
thus , each node in the constraint graph has at most degree six .
[ computnote ] ] .
the system of linear inequalities eq .
( [ e : polytopeequations ] ) delimit a subset of the @xmath1-dimensional space of @xmath60 coordinates @xmath71 ) , an @xmath1-dimensional polytope , bounded by at most @xmath72 hyperplanes .
this convex object is easier to analyze than the highly intricate @xmath73-dimensional configuration space of the full hard - disk problem .
the polytope is unbounded in the @xmath74 direction in consequence of the periodic boundary conditions , since uniform translation of all the disks is always permitted .
also , since the constraint graph is invariant under event - chain moves in the @xmath60 direction , so is the polytope .
however , the polytope becomes bounded by taking a section orthogonal to @xmath74 . in the invariant polytope , an event chain of total displacement @xmath27 corresponds to a molecular dynamics evolution of duration @xmath27 : displacing the @xmath11-th disk corresponds to the `` particle '' @xmath75 moving in the @xmath11-th coordinate direction , and each collision event ( the transfer of momentum from one disk to another ) to a right - angle reflection at the facets of the polytope ( see fig .
[ f : polytope_ballistics_plus_graph ] ) .
the construction of the event - chain move is finished at time @xmath27 .
the next move involves the sampling of a new starting disk and possibly of one of the @xmath76 directions . in the invariant polytope
, this is the choice of new velocities .
local monte carlo on the other hand , if restricted to moves in @xmath60 direction , implements diffusive motion in the invariant polytope is valid only if all intermediate positions @xmath77 with @xmath78 $ ] yield legal hard - sphere configurations . ] .
the invariant constraint graph allows for fast lookup of possible collision partners , and may even replace the customary cell grids ( see , for example , section 2.4 of ref .
while computation of the actual constraint graph requires depth search , a superset sufficient for practical computations can be computed efficiently , see the footnote on page .
hard - disk system at packing fraction @xmath79 .
_ left _ : slowly decaying modes .
configurations are shown with red disks moving in the @xmath80 direction and green disks in @xmath81 .
the modes shown are the eigenvectors of the largest eigenvalues of @xmath82 .
_ center _ : remaining correlation after @xmath83 of event - chain moves in the horizontal direction .
these are the largest eigenvectors of @xmath84 .
correlations in the horizontal direction have all but disappeared . _
right _ : decay of the slowest modes , for the event - chain simulations with various total displacements @xmath27 , and for both the global ( gb ) and the detailed ( db ) balance version.[f : eigenmodes ] ] although the sampling problem from the invariant polytope concerns a convex body , it is notoriously nontrivial @xcite .
the inequalities eq .
( [ e : polytopeequations ] ) essentially amount to a system of coupled one - dimensional hard - disk problems . to study the relaxation behavior effected by the event - chain algorithm in the polytope
, we consider the cross - covariance of the disk coordinates , @xmath85 where @xmath86 is the @xmath60 coordinate of the disk @xmath11 , compensated for the overall translation of the system due to the event - chain moves , @xmath87 here , @xmath88 is @xmath59 for the global balance version of the event - chain algorithm ( chains only in @xmath80 direction ) , and @xmath89 for the detailed balance version ( @xmath90 ) .
the eigenvectors of @xmath82 are the polytope s normal modes @xmath91 , @xmath92 , in the sense of principal component analysis .
the nature of the modes @xmath91 depends on the structure of the invariant polytope and captures the relative order of colliding disks and their frozen - in @xmath62 coordinates .
the normal modes to the largest eigenvalues are large - scale cooperative rearrangements of the disks ( see fig .
[ f : eigenmodes ] ) .
they are the slowest modes to decay under both local and event - chain monte carlo and govern the global decorrelation of the disk configuration .
in particular , two modes dominated by antiparallel flow bands are very slow to decay ( mode 1 and 2 in fig .
[ f : eigenmodes ] ) . at delay times
@xmath93 , the cross - covariance @xmath94 captures residual correlations among the disk coordinates .
the event - chain moves couple more efficiently to the longitudinal modes of the system , and we find that after @xmath95 , the event - chain algorithm has virtually erased longitudinal correlations .
the most prominent residual correlations carry a transverse band structure ( see fig . [
f : eigenmodes ] ) .
the result is a substantial decrease in efficiency of the algorithm for simulated duration in a single direction larger than @xmath96 . to estimate the convergence time , we study the projection of the system s evolution @xmath97 onto a single mode , @xmath98 .
the autocorrelation function @xmath99 is , for short chain lengths @xmath27 , monotonously decaying .
larger chain lengths accelerate the decay , as the coupling to large - scale modes is improved ( fig .
[ f : eigenmodes ] ) . for chains spanning several times the box ,
however , the autocorrelation functions @xmath100 develop oscillations with very weak damping , offsetting the benefits of longer chains .
the detailed balance version of event - chain monte carlo is generally slower and less prone to oscillations . for optimal performance , the global balance version
should thus be used with @xmath27 larger , but on the order of @xmath101 , and for times @xmath102 ( see fig . [
f : eigenmodes ] ) . for disk configurations larger than the correlation length , @xmath27 can be reduced appropriately .
[ s : fullharddisk ] the invariant polytope representation allows us to interpret the convergence of the full hard disk sampling problem .
the conceptually simplest monte carlo algorithm for hard disks consists entirely in polytope sampling : one iteration amounts to direct sampling a new configuration @xmath103 from the invariant polytope of the starting configuration @xmath104 , and exchanging the @xmath60 and @xmath62 coordinates of all the disks .
this markov - chain algorithm satisfies detailed balance . in our experiments ,
the timescale @xmath55 , measured in iterations , for relaxation to equilibrium increases only as @xmath105 for large systems , implying that most of the complexity of the hard - disk sampling problem resides in the polytope sampling . since
direct sampling is a hard problem for high - dimensional polytopes ( see section [ s : generalpolytope ] ) , we replace it by markov chains of a fixed number of event - chain moves , in effect performing molecular dynamics in the invariant polytopes for fixed duration @xmath106 : @xmath107 this algorithm satisfies detailed or global balance depending on the version of the event - chain algorithm that is used for polytope sampling .
autocorrelation function @xmath108 for several switching intervals @xmath106 , as a function of the simulated md time ( _ bottom axis _ ) , or alternatively , the number of collisions per disk ( _ top axis _ ) . _ right _ : decay of @xmath56 as a function of the number of @xmath60/@xmath62 switches @xmath109 .
as @xmath106 approaches @xmath1 , the curves approach the limit of direct sampling from the polytope , with a mixing time of @xmath55 cycles .
all curves were averaged from systems of @xmath110 disks at packing fraction @xmath111 ; the chain length was @xmath112 .
_ inset _ : the mixing time @xmath55 first increases rapidly with system size , but only grows as @xmath105 for larger systems ( also @xmath111 ) .
, title="fig : " ] autocorrelation function @xmath108 for several switching intervals @xmath106 , as a function of the simulated md time ( _ bottom axis _ ) , or alternatively , the number of collisions per disk ( _ top axis _ ) .
_ right _ : decay of @xmath56 as a function of the number of @xmath60/@xmath62 switches @xmath109 . as @xmath106 approaches @xmath1 ,
the curves approach the limit of direct sampling from the polytope , with a mixing time of @xmath55 cycles .
all curves were averaged from systems of @xmath110 disks at packing fraction @xmath111 ; the chain length was @xmath112 .
_ inset _ : the mixing time @xmath55 first increases rapidly with system size , but only grows as @xmath105 for larger systems ( also @xmath111 ) .
, title="fig : " ] we study the influence of the switching interval @xmath106 on convergence properties . in fig .
[ f : switchingandmixing ] , the autocorrelation function @xmath56 of the complex order parameter @xmath44 is plotted vs. cumulative molecular dynamics time .
@xmath56 decays most quickly when the switching interval @xmath106 is small , but the decay speed deteriorates very slowly with @xmath106 . only at @xmath113 ( corresponding to about @xmath114-@xmath115 collisions per disk at these densities ) ,
the algorithm becomes notably less efficient .
the efficiency drop thus follows the decay of longitudinal ( in @xmath60 direction ) correlations in the invariant polytope , and is to be expected from the results in section [ s : polytope ] . in the limit @xmath116 ,
the event - chain algorithm realizes direct sampling in the invariant polytope .
the approach to this limit is illustrated in fig .
[ f : switchingandmixing ] by plotting @xmath56 against the number of x / y switching cycles @xmath109 .
as the switching interval @xmath106 increases , the autocorrelation functions approach an asymptotic curve @xmath117 , where @xmath55 is the correlation time of the direct sampling algorithm .
we find that for practical purposes , event - chain monte carlo reaches the asymptotic regime for @xmath113 , and thus samples an approximately independent point in the invariant polytope in @xmath118 operations .
importantly , the correlation time @xmath119 increases rapidly only for small system size @xmath1 .
after the system size surpasses the correlation length , @xmath55 grows only as @xmath105 .
[ s : generalpolytope ] the invariant polytope is bounded by hyperplanes which are normal to @xmath120 coordinate axes and have unit derivative along the remaining axes . by choice of the @xmath31 ,
the molecular dynamics evolution is aligned with the coordinate axes at all times , and computations of intersections are of complexity @xmath121 . as shown in section [ s : fullharddisk ] , the event - chain algorithm seems to achieve an effective mixing time of @xmath118 collision events , so that the cost of sampling the hard - disk polytope appears as @xmath118 .
the event - chain algorithm also allows to sample general polytopes .
direct sampling from polytopes is straightforward only in low dimensions @xmath1 , especially in @xmath122 : a two - dimensional polytope with @xmath123 edges ( a convex @xmath123-sided polygon ) , can be decomposed into @xmath123 triangles , using an interior point .
triangles may then be sampled according to their areas , and a random point may be sampled inside the sampled triangle ( see , e.g. chap .
6.2 of @xcite ) . in higher dimensions @xmath1 ,
triangulation by simplices generalizes this decomposition . since polytopes such as the invariant hard - disk polytope have an exponential number of facets , direct sampling algorithms are no longer practical .
markov - chain sampling @xcite achieves mixing times of @xmath124 steps , where @xmath125 is the number of bounding hyperplanes ( @xmath126 for hard disks ) , and where each move may be implemented in @xmath127 steps
. it will be interesting to see how event - chain polytope sampling compares with existing polytope sampling methods , in particular the ` hit - and - run ' algorithms .
[ s : parallel ] in view of the long running times of monte carlo simulations and of the current standstill in computer clock speeds , it is essential to develop _ parallel monte carlo methods _ which distribute the work load among several _ threads _ performing independent computation with as few communication as possible .
such methods will allow to study not only the standard hard disk ensemble , but also related systems such as soft disks and polydisperse disk packings .
however , parallel monte carlo algorithms for continuum systems pose many more problems than for lattice models , for example the ising spins , where straightforward parallel application of local metropolis updates converges to the boltzmann distribution @xcite .
a massively parallel implementation of the local monte carlo algorithm was applied recently to the hard - disk melting problem @xcite .
it sets up square cells according to a four - color checkerboard pattern .
disks in same - color cells can be updated simultaneously , but moves across cell boundaries are rejected . to ensure ergodicity ,
a new cell grid must be sampled periodically .
massive parallelism of @xmath128 threads on a graphics card offsets the slowness of local monte carlo compared to event - chain algorithm monte carlo @xcite . these calculations confirmed the first - order liquid - hexatic phase transition in hard disks @xcite .
disk packing at @xmath111 .
we plot the number of collisions in accepted chains per hour of computation . on the same machine ,
the serial version has a performance of about @xmath129 collisions per hour .
the event - chain routine is the same in both programs .
_ right _ : the acceptance ratio for chains depends on the thickness of the active layers ( which is decreases as more threads are added ) and the total displacement @xmath27 of the chains .
[ f : stripescheme ] ] for parallel implementations of event - chain monte carlo , we consider only parallel threads that run chains in the same direction @xmath76 .
this minimizes the chance that two chains cross each other and move the same disks .
it also allows us to apply the invariant polytope framework of section [ s : polytope ] .
it is instructive to realize that the effects of an event - chain move can be summarized in the difference vector of the new and old @xmath60 coordinates : @xmath130 , with @xmath131 . moreover ,
if two chains are _ independent _ , meaning their sets of disks touched are disjoint , the net effect of running both chains is the sum of their individual difference vectors , @xmath132 . if , however , any disk is touched by both chains , the chain reaching this disk earlier in md time has precedence , the later chain sees a modified environment , and consequently takes a different evolution .
thus , interdependent chains can not be added arithmetically , is always admissible . ] .
the primary obstacle in parallelizing event - chain monte carlo consists in preserving the correct causal relations between subsequent chains , as required for the convergence to correct equilibrium distribution . in the following ,
we discuss three strategies to parallelize event - chain monte carlo .
predict / execute algorithm _
distributes work among threads for a model of chains that follow each other chronologically .
the effects @xmath133 of several chains are predicted in advance from the current disk configuration .
the effects of the chains are then applied to the system state in the chronological order in which the starting disks were sampled . to detect conflicts
, it is sufficient to compute the intersection of the set of disks touched by the current chain and of the chains that ran since the beginning of planning ; if this intersection is not empty , the chain has to be recomputed from the updated state of the disk configuration .
planning and execution of chains can proceed in parallel on a shared - memory machine . using lock - free data structures ,
we attain collision rates in excess of @xmath134 per hour in @xmath60 collision mode on a four - processor machine . due to its serial nature
, this algorithm does not scale well beyond a few threads , however ; with too many chains predicted in advance , the probability for recomputations rises . moreover , switching between @xmath60 and @xmath62 collision modes requires reinitialization of the data structures and is rather expensive .
a variation of the four - color scheme adapted to the event - chain algorithm partitions the system in horizontal stripes , separated by _ frozen isolation layers _ of thickness @xmath135 disk radii ( see fig .
[ f : stripescheme ] ) .
disks with their centers in the isolation layers are kept fixed , and thus guarantee the independence of chains running in neighboring stripes . to preserve the isolation layers , chains colliding with a frozen disk
are rejected . as there are rejected moves ,
the global balance condition is no longer guaranteed : the number of accepted forward chains can be different from the number of accepted backward chains ( see section [ s : balances ] ) . when allowing chains in both the @xmath136 directions , however , the isolation layer algorithm satisfies detailed balance .
furthermore , in order to limit the rejection rate , the per - chain total displacement @xmath27 has to be kept lower than in the serial algorithm . in view of the discussion of section [ s : polytope ] , these necessities reduce somewhat the efficiency of the method . due to the isolation layers ,
the accessible configuration space is restricted , and for ergodicity , the layer boundaries have to be resampled periodically , as in the four - color version of local monte carlo .
we have implemented the isolation layer algorithm in parallel on a shared - memory machine .
using several cores in parallel , it is possible to achieve effective collision rates which are 1030 times the single - core performance ( see fig .
[ f : stripescheme ] ) , for systems of sufficient size . for systems too small , less threads can be used without shrinking the active strips to a point where the acceptance ratio becomes a limiting factor .
systems of physical interest , however , are on the order of @xmath137 , and allow to use 1020 cores with moderate @xmath27 . at this time
, the algorithm is not bound by rejection rates , but by communication between threads . finally
, the event - chain scheme is not fundamentally limited to a single moving disk at any time .
we may indeed launch multiple _ concurrent chains _ , which run at the same simulated md time , and interact with each other .
this is different from the parallel simulation of chains which interact in sequential manner . in the invariant polytope picture
, multiple concurrent chains correspond to choosing more general initial conditions , where more than one disk is given an initial velocity of @xmath59 .
after time @xmath27 , multiple chains have executed , and possibly interacted with each other ; there is no rejection in this algorithm .
the problem has some resemblance with event - driven molecular dynamics , because the scheduling of collisions must be foreseen , but there are several simplifications : all velocities are in the same direction and of magnitude @xmath89 or @xmath59 . as a consequence , two moving disks
can not collide with one other ; however , the faithful simulation of chains close by and possibly interacting requires careful synchronization among threads . in our experiments , this limits the speedup by parallelization .
our most efficient method at this point is the isolation layer algorithm .
we have reached in this paper a better understanding of the event - chain monte carlo algorithm for the hard - disk sampling problem . by restricting the algorithm to chains in a single direction
, a connection appears to the well - known problem of sampling random points from a polytope : a move of the event - chain algorithm consists in performing a finite - time molecular dynamics simulation in the invariant polytope of the disk configuration .
this connection offers new strategies to solve the hard - disk sampling problem in terms of polytope sampling ; it also suggests to investigate the utility of event - chain methods for the sampling of general polytopes .
finally , it will be interesting to study the combinatorial structure of the typical invariant polytope , and its dependence on thermodynamical parameters . by the study of correlation functions
, we have shown that the monte carlo relaxation process in the invariant polytope separates into two phases : a rapid longitudinal relaxation , followed by a much slower relaxation of the transverse degrees of freedom .
we have given recommendations for the parameters of the algorithm based on these results .
finally , we have discussed several strategies for parallelizing monte carlo algorithms for hard disks , alleviating the problem of the long simulation times in hard disk monte carlo .
the parallelization of the hard - disk ensemble remains challenging due to its unique combination of very little actual computation and long correlation times .
efficient methods to tackle the hard - disk ensemble are , however , crucial in order to treat related systems such as soft disks with the same level of success as the hard disks .
new concepts such as the link to polytope sampling will be essential in this effort .
we thank p. diaconis , e. p. bernard , s. leitmann and m. hoffmann for fruitful discussions .
99 l. boltzmann , sitzber .
wien , math .
naturw . kl .
2a ) ( 1896 ) . j. l. lebowitz and o. penrose , _
phys . _ * 5 * 841 ( 1964 ) .
j. g. kirkwood , e. monroe , _
* 8 * 845 ( 1940 ) .
s. asakura , f. oosawa , _ j. chem
. phys . _ * 22 * 1255 ( 1954 ) .
b. j. alder , t. e. wainwright , _ phys . rev . _ * 127 * , 359 ( 1962 ) .
j. m. kosterlitz , d. j. thouless , _ j. phys .
c : solid state phys . _ * 6 * 1181 ( 1973 ) ; b. o. halperin , d. r. nelson _ phys .
* 41 * 121 ( 1978 ) . c. h. mak , _ phys .
e _ * 73 * 065104(r ) ( 2006 ) .
e. p. bernard , w. krauth , _ phys .
* 107 * , 155704 ( 2011 ) .
k. zahn , r. lenke , g. maret , _ phys .
lett . _ * 82 * 2721 ( 1999 ) .
p. diaconis , _ j. stat .
_ * 144 * 445 ( 2011 ) . y. g. sinai , _ russian mathematical surveys _ * 25 * , 137 ( 1970 )
. n. simanyi , _ inventiones mathematicae _ * 154 * , 123 ( 2003 ) . b. j. alder , t. e. wainwright , _ j. chem
* 27 * , 1208 ( 1957 ) .
d. c. rapaport , _
phys . _ * 34 * , 184 ( 1980 ) .
n. metropolis , a. w. rosenbluth , m. n. rosenbluth , a. h. teller , e. teller , _ j. chem .
* 21 * 1087 ( 1953 ) .
m. isobe , _ int . j. mod
c _ * 10 * , 1281 ( 1999 ) . c. dress , w. krauth ,
_ j. phys . a , math .
_ * 28 * l597 ( 1995 ) .
a. jaster , _ phys .
e _ * 59 * , 2594 ( 1999 ) .
e. p. bernard , w. krauth , d. b. wilson , _ phys .
e _ * 80 * , 056704 ( 2009 ) .
r. h. swendsen , j. s. wang , _ phys .
lett . _ * 58 * 86 ( 1987 ) .
u. wolff , _ phys .
lett . _ * 62 * 361 ( 1989 ) .
r. kannan , m. w. mahoney , r. montenegro , in t. ibaraki , n. katoh , h. ono ( eds . ) : algorithms and computation , 14th international symposium , isaac 2003 . proceedings .
lecture notes in computer science 2906 , springer ( 2003 ) .
d. b. wilson , _ random struct
. algorithms _ * 16 * , 85 ( 2000 ) . c. chanal , w. krauth , _ phys .
e _ * 81 * 016705 ( 2010 ) .
j. a. anderson , m. engel , s. c. glotzer , m. isobe , e. p. bernard , w. krauth , arxiv:1211.1645 . w. mickel , s. c. kapfer , g. e. schrder - turk , k. mecke , to appear in
_ j. chem .
( 2013 ) ; arxiv:1209.6180 .
m. e. dyer , a. m. frieze , _
siam j. comput . _
* 17 * 967 ( 1988 ) ; v. kaibel , m. e. pfetsch , in _ algebra , geometry and software systems _ , edited by m. joswig , n. takayama ( springer , 2003 ) ; arxiv : math/0202204 .
w. krauth , _ statistical mechanics : algorithms and computations _ , oxford university press ( 2006 ) .
r. l. smith , _ operations research _ * 32 * , 1296 ( 1984 ) .
p. a. rubin , _ communications in statistics - simulation and computation _ * 13 * , 375 ( 1984 ) .
m. dyer , a. frieze , r. kannan , _
j. acm _ * 38 * , 1 ( 1991 ) .
r. kannan and h. narayaran , _ mathematics of operations research _ * 37 * , 1 ( 2012 ) .
b. a. berg , _ markov chain monte carlo simulations and their statistical analysis _ , world scientific , ( 2004 ) .
j. a. anderson , e. jankowski , t. l. grubb , m. engel , s. c. glotzer , arxiv:1211.1646 . | the hard - disk problem , the statics and the dynamics of equal two - dimensional hard spheres in a periodic box , has had a profound influence on statistical and computational physics .
markov - chain monte carlo and molecular dynamics were first discussed for this model . here
we reformulate hard - disk monte carlo algorithms in terms of another classic problem , namely the sampling from a polytope .
local markov - chain monte carlo , as proposed by metropolis et al . in 1953 , appears as a sequence of random walks in high - dimensional polytopes , while the moves of the more powerful event - chain algorithm correspond to molecular dynamics evolution .
we determine the convergence properties of monte carlo methods in a special invariant polytope associated with hard - disk configurations , and the implications for convergence of hard - disk sampling .
finally , we discuss parallelization strategies for event - chain monte carlo and present results for a multicore implementation . |
compact stars consist mostly of a degenerated plasma which close to the surface of the star forms the so called ion crust of densely packed ions and relatively free electrons .
moreover a highly ionized plasma constitutes the stellar atmosphere formed as a result of ionization due to the increase in the mean collisional rate of the atoms constituting the star s atmosphere @xcite . outside the ion crust
the plasma co - rotating with the star is rather dilute and highly conducting@xcite . a plasma , both in a degenerate form and as a dilute and highly conducting medium , behaves as an oscillating system having a characteristic frequency called the plasma frequency . in particular the dielectric constant of a medium
is determined by the plasma frequency , which in turn determines the transmission and reflection of electromagnetic radiation for the plasma@xcite . on the other hand compact stars
possess very strong gravitational fields , so that general relativistic effects are important for an adequate description of the phenomenon occuring in vicinity of these stars .
usually general relativistic effects are not directly observable but are manifest in an indirect way , for example via interaction with the magnetic field@xcite , in the accretion of matter@xcite , and other material and radiative processes occurring in vicinity of the star ( e.g. ref .
an investigation of these effects is of interest for various astrophysical processes as well as for testing the general theory of relativity@xcite . in this paper
we investigate gravitomagnetic effects of a slowly rotating compact gravitational source on the electron oscillations in a homogeneous electrically neutral plasma .
we assume the absence of any electric or magnetic field and neglect the random thermal motion of the electrons . in the next section we present the gravitoelectromagnetic approximation to the general theory of relativity particularly for the geodesic equation . in section 3
we formulate the equations of motion for the electron oscillations in the gravitational field of the star and study the effects on plasma frequency .
it found that there is a reduction in the plasma frequency due to the gravitomagnetic force .
we then estimate in section 4 the effects of the variation in the dielectric constant on the electromagnetic waves propagating radially through the plasma surrounding the compact gravitational source . in conclusion
we summarize the main results of the paper and discuss their relevance to observation . throughout
we use the gravitational units @xmath0 unless mentioned otherwise .
the field equations of general theory of relativity , for a slowly rotating gravitational source , bear a remarkable formal similarity with the fundamental equations of classical electromagnetism@xcite . particularly in einstein
s theory of gravitation the trajectory of a test particle in vicinity of a gravitational source is given by the geodesic equation : @xmath1 for sufficiently weak gravitating systems , such as the compact stars , a linearization of the metric can be adequately assumed , where @xmath2 and where @xmath3 @xmath4 is the minkowski metric tensor and @xmath5 is the perturbation to the metric such that @xmath6 and @xmath7 @xmath8 are the position coordinates of test particle , with time @xmath9 being the affine parameter , the geodesic equation takes the form @xmath10 where @xmath11 and @xmath12 assuming to the first order of approximation , the star to be a slowly rotating sphere of homogeneous mass @xmath13 and radius @xmath14 , we have @xmath15 which is the gravitoelectric force given by the newtonian gravitational force per unit mass , and @xmath16 is a unit vector in the radial direction .
the term @xmath17 @xmath18 * * * * is the gravitomagnetic force where @xmath19 is given by @xmath20 @xmath21 being the angular frequency vector of the gravitational source , and @xmath22 is velocity of the test particle .
the approximation , called the gravitoelectromagnetic approximation to the general theory of relativity , is generally valid for compact astrophysical sources .
the independence of the gravitomagnetic potential @xmath23 from a particular frame and particular coordinate system has been demonstrated @xcite .
physically it can be interpreted as ` gravitomagnetic current ' induced in the vicinity of the gravitational source due to its rotation .
with a fixed ion background it is convenient to choose an orientation of the coordinates triplet @xmath24 such that the angular frequency vector @xmath25 is along the positive @xmath26-axis .
since the electron oscillations occur within a very small region of space we have , for the components of acceleration due to gravitoelectric force , @xmath27 to be of constant magnitude @xmath28 along each direction @xmath24 . in this case
the components of @xmath19 are @xmath29 , therefore effective components of the gravitomagnetic force lie in the @xmath30-plane .
we take , without loss of generality , the @xmath30-plane to be the equatorial plane of the star mainly because here the gravitomagnetic effects are of maximum magnitude for the given orientation@xcite .
the equations of motion for the electron oscillations are : @xmath31 @xmath32 where @xmath33 is the newtonian plasma frequency , @xmath34 being the electron number density of the plasma per centimeter , @xmath35 is the electronic charge , @xmath36 is the electronic mass , and @xmath37 is the dielectric constant .
clearly @xmath38 has dimensions of cycle per unit time i.e. of frequency . to investigate the resonant frequency for the above system of coupled equations ,
we assume that the solutions to equations ( 8) , and ( 9 ) can be expressed as @xmath39 and @xmath40 where @xmath41 is the applied external frequency .
the system can then be written as a single matrix equation @xmath42 \left [ \begin{array}{c } a \\ b \end{array } \right ] = 0.\ ] ] for solution to exist equation ( 10 ) implies that , @xmath43 since @xmath44 , we expand the determinant and neglect terms involving squares and higher powers of @xmath45 .
this gives the following expression for resonant frequency of the oscillating plasma @xmath46 further if @xmath47 is the angle between the position vector @xmath48 and the angular frequency vector @xmath21 , it follows from expression ( 7 ) that close to the surface of the star , i. e. , at @xmath49 , @xmath50 has magnitude given by @xmath51 where @xmath52 , @xmath53 being the magnitude of the angular frequency vector . substituting from expression ( 13 ) in ( 12 ) we obtain an expression relating plasma frequency @xmath54 of the star s atmosphere to the angle of inclination
@xmath47 : @xmath55 in expression ( 14 ) we see that the shift in plasma frequency depends not only on the mass , radius and angular frequency of the compact star , via the parameter @xmath56 , but also on the angle of inclination @xmath47 . here
the parameter @xmath56 determines the magnitude of the gravitomagnetic effect for different compact stellar sources . for a typical
compact star @xmath56 ranges from @xmath57 ( for a typical white dwarf of mass @xmath58 , radius @xmath59 and angular frequency@xmath60 ) to @xmath61 ( for a typical neutron star of mass @xmath62 , radius @xmath63 , and angular frequency @xmath64 ) with corresponding plasma frequency ranging approximately @xmath65 to @xmath66 or above .
given the parameters @xmath56 and @xmath54 the sft depend only on the angle of inclination @xmath47 .
a plot between the plasma frequency shift and the angle @xmath47 ; based on expression ( 14 ) for the cases of neutron star , pulsar , and white dwarf for the typical values of mass , radius and angular frequency is given in fig .
clearly the gravitomagnetic reduction in th plasma frequency is maximum in the equitorial plane of the star .
to study the effects of plasma frequency shift on the dielectric constant of the plasma , let @xmath67 , where @xmath38 is given by ( 9 ) . then in terms of plasma frequency we have the following expression for the dielectric constant @xmath68 for the plasma : @xmath69 where @xmath70 is the angular frequency of an electromagnetic signal propagating through the plasma . for an electromagnetic signal of some given angular frequency
the dielectric constant ( 15 ) depends only on the angle @xmath47 for a compact gravitational source .
ploting this dependence of the constant @xmath68 on the angle @xmath71 in fig.(2 ) we find that the plasma medium becomes rarer as @xmath47 varies from @xmath72 to @xmath73
. therefore if @xmath74 be the angular frequency range of the radially out - going waves , then from the usual conditions for the propagation of an electromagnetic signal in continuous media we have for the transmission and reflection the conditions in the field of a rotating gravitational source : @xmath75 the shift in plasma frequency in the equatorial plane ( @xmath76 ) is @xmath77 whereas in the plane orthogonal to it ( @xmath78 ) the shift is @xmath79 .
denoting the frequency shift in the equatorial plane by @xmath80 and by @xmath81 for the orthogonal plane , we find that the maximum difference in the frequencies transmitted through the plasma : @xmath82 it is worth emphasizing here that since plasma oscillations and the gravitomagnetic force both are along the radial direction therefore even if the plasma co - rotates with the star ( for instance close to the ion crust ) the gravitomagnetic force will effect the oscillations of the plasma and hence the dielectric constant .
in this paper we have considered the effects of gravitomagnetic force on electron oscillations in a homogeneous plasma with a fixed ion background .
it was found that there is a shift in the characteristic frequency and hence in the dielectric constant of the plasma surrounding the star .
the shift depends on the intrinsic parameters of the star , i. e. , its mass and radius , and also on the component of the angular frequency vector of the star to the plane of oscillation of the electron .
it was estimated that an electromagnetic wave of given frequency the shift in the dielectric constant results in changing the allowed range of frequencies transmitted through the plasma , by a maximum amount @xmath83 ( as defined for expression ( 13 ) above ) . for a plasma enveloping a white dwarf the predicted difference in the frequency of electromagnetic radiation emitted
is very small ( the magnitude of the parameter @xmath56 is about @xmath57 ) .
however for a typical pulsar ( @xmath84 ) and especially neutron stars ( @xmath85 @xmath61 ) the shift , though still small , may possibly be observed ( for example in radiation spectra of various compact sources ) , especially when the effects of various contingencies ( such as dispersion due to interstellar gases , effects due to the atmosphere of the earth , etc . )
are isolated . | we investigate the effects of a slowly rotating compact gravitational source on electron oscillations in a homogeneous electrically neutral plasma in the absence of an external electric or magnetic field . neglecting the random thermal motion of the electrons we assume the gravitoelectromagnetic approximation to the general theory of relativity for the gravitational field .
it is shown that there is a shift in the plasma frequency and hence in the dielectric constant of the plasma due to the gravitomagnetic force .
we also give estimates for the difference in the frequency of radially transmitted electromagnetic signals for typical compact star candidates . |
entangled states are used as resource for quantum teleportation @xcite . the success of quantum teleportation is quantified by optimal teleportation fidelity .
the strength of entanglement is proportional to the success of teleportation for pure entangled resources .
maximally entangled pure states give maximum optimal teleportation fidelity of unity . on the other hand , for mixed states , success of teleportation
depends , in addition to strength of entanglement , on mixedness of the state @xcite .
it was shown in @xcite that , for mixed entangled states , there is an upper bound on mixedness above which the state is useless for quantum teleportation .
upper bounds on measures of mixedness , namely , von neumann entropy and linear entropy were obtained for a general bipartite @xmath0 state .
mixed entangled states can be classified according to their rank @xmath1 , where the rank @xmath1 varies between @xmath2 and @xmath3 for a bipartite @xmath0 system . in the present work , we obtain rank dependent upper bounds on measures of mixedness , above which the states of respective ranks become useless for quantum teleportation . in our previous work
@xcite , we observed that mixedness and entanglement of a mixed state resource independently influence teleportation . for a state with a fixed value of mixedness ,
the state being entangled is only necessary for it to be a resource for quantum teleportation , but not sufficient .
states with low mixedness and high entanglement turn out to be ideal resources for quantum teleportation .
we argued based on the numerical work on a class of maximally entangled mixed states that there exists a rank dependent lower bound on a measure of entanglement such as concurrence , below which the states are useless for quantum teleportation . in this article , we derive rank dependent lower bounds on concurrence for a bipartite @xmath4 systems .
werner state defined as a probabilistic mixture of maximally entangled pure state and maximally mixed separable state , exhibits highest mixedness for a given optimum teleportation fidelity among all the two qubit mixed entangled states known in the literature .
werner state @xcite is a rank 4 state . in the present study
, we generalize the construction and obtain second and third rank werner states .
we show that rank dependent werner states exhibit the respective rank dependent bounds obtained on measures of mixedness and entanglement for a given value of teleportaion fidelity .
we consider two measures of mixedness of a state @xmath5 , namely , von nuemann entropy defined as @xmath6 @xcite and linear entropy given as @xmath7 $ ] .
further , the maximum achievable teleportaion fidelity of a bipartite @xmath8 system in the standard teleportation @xcite scheme is @xmath9 , where @xmath10 is the singlet fraction of @xmath5 given by @xmath11 . here
the maximum is over all maximally entangled states . and maximum fidelity achieved classically is @xmath12 .
this shows that the state is useful quantum teleportation for @xmath13 ( @xmath14 ) .
let @xmath15 denote a bipartite mixed state of @xmath8 system whose rank is @xmath1 , @xmath16 .
firstly , we prove a theorem that gives rank dependent upper bounds on von neumann entropy , above which the state is useless for quantum teleportation .
_ theorem _ : if the entropy @xmath17 of a given state @xmath18 of rank @xmath1 of @xmath8 system exceeds @xmath19 , then the state is not useful for quantum teleportation._proof _ : we know for a given state @xmath15 of @xmath0 systems @xmath20 implies the state is useful for quantum teleportaion .
we have state @xmath18 of rank @xmath1 of @xmath21 systems as , @xmath22 @xmath23 is the basis formed from @xmath24 maximally entangled states .
we have from the definition of singlet fraction the largest element , say @xmath25 of @xmath26 is greater than or equal to @xmath27 . and
it is known that von neumann entropy @xmath17 of a given state @xmath18 is less than or equal to shannon entropy . _
@xmath28 shannon entropy in eq.([shannon ] ) is maximum for @xmath29 and rest of @xmath30 elements are equal , subjected to the constraint @xmath31 , we get the upper bound as @xmath32 this implies , @xmath33 this shows that for @xmath17 satisfying eq.([von ] ) , the singlet fraction is greater than @xmath27 . thus , we prove if @xmath34 , state @xmath18 is useless for quantum teleportation .
we also obtain an analytical expression for the rank dependant upper bound on linear entropy as a measure of mixedness as @xmath35 for @xmath36 , we have @xmath37 where @xmath1 is the rank of the state , which varies from 2 to 4 . in our previous work
@xcite , based on the analysis of a class of maximally entangled mixed states of @xmath4 bipartite system given in @xcite , we observed that for a given value of linear entropy , there exists a rank dependent upper bound on the optimal teleportation fidelity and the upper bound increases with increase in the rank .
this is equivalent to stating that for a given value of optimal telportation fidelity , there exists a rank dependent upper bound on linear entropy and the upper bound increases with rank .
the above result allows us to calculate the upper bounds explicitly .
if the optimal teleportation fidelity is fixed as @xmath38 , the classical fidelity , the upper bound on linear entropy for states of second , third and fourth ranks are @xmath38 , @xmath39 and @xmath40 respectively .
states with linear entropy above the respective rank dependent upper bounds are useless for teleportation .
a class of maximally entangled mixed states ( mems ) is constructed by ishizaka et .
al @xcite and the construction is as follows .
@xmath41 where @xmath42 are the eigenvalues of the state @xmath5 @xmath43 and @xmath44 .
the werner state , which is a convex sum of maximally entangled pure state and maximally mixed separable state , given by @xmath45 werner state corresponds to a choice of eigenvalues given by @xmath46 and @xmath47 .
werner state is a state of rank 4 .
the singlet fraction for werner state is @xmath48 and linear entropy is estimated as @xmath49 .
the value of linear entropy at which fidelity is equal to classical limit can be found as @xmath50 , which is same as the upper bound obtained above for states of rank 4 .
this correspondence motivated us to construct werner like states of ranks 2 and 3 .
a rank @xmath51 werner state can be constructed by using eigenvalues as @xmath52,@xmath53 and @xmath54 .
we have @xmath55 and linear entropy is equal to @xmath56 .
thus it is clear that @xmath57 for a value of linear entropy greater than @xmath58 , which is the obtained upper bound for rank 3 states . to construct rank @xmath2 werner state we take @xmath59 , @xmath60 and @xmath61 .
we get singlet fraction and linear entropy as @xmath62 and @xmath63 respectively .
it clearly shows that rank @xmath2 werner state is useless for teleportation when linear entropy is greater than @xmath64 , coinciding with the theoretical upper bound shown above .
thus , we illustrate that the constructed rank dependent werner states exhibit the theoretical upper bounds on the linear entropy for a given value of optimal teleportaion fidelity .
the relationship between the concurrence as a measure of entanglement of state @xmath5 and it s purity is well studied , the degree of entanglement decreases as purity decreases .
the maximum possible value of concurrence of a state @xmath5 for a spectrum of eigenvalues @xcite @xmath65 in descending order is given as @xmath66 in our previous work , we also observed that there exist a rank dependent lower bound on concurrence for a fixed value of optimal teleportation fidelity and the lower bound decreases with increase in the rank of the state . in the present work ,
we obtain the exact rank dependent lower bounds on the concurrence for a @xmath4 bipartite states .
we find that the eigenvalues @xmath67 and equal values of rest of the @xmath68 @xmath69 minimize @xmath70 among all eigenvalue spectra . by substituting the values for @xmath71 ,
that is , corresponding to rank @xmath1 , @xmath72 and rest of @xmath68 elements are equal , we get the lower bound on the concurrence of state @xmath18 of rank @xmath1 below which the state fails to be a source for quantum teleportation .
we obtain the values of lower bound on concurrence for the failure of teleportation for rank 4,3 and 2 states as follows , @xmath73 and @xmath74 respectively . from this
we clearly show that there exist a lower bound on concurrence for a fixed value of fidelity , this lower bound decreases as rank increases for two qubit systems .
the lower bound on concurrence for failure of quantum teleportation for rank dependent werners states coincide with the bounds derived in this work .
lower and upper bounds on teleportation fidelity as a function of concurrence are obtained in @xcite . the upper and lower bounds on fidelity @xmath75 for a concurrence @xmath76 are given by @xmath77 and @xmath78 respectively .
the upper bound on fidelity as a function of concurrence coincides with werner state of rank 4 . based on the properties of lower rank werner states constructed in this work
, we conjecture that there exist rank dependant bounds on fidelity as a function of concurrence and these bounds which also coincide with werner states of lower ranks .
the maximum amount of fidelity of a rank three state as a function of concurrence is given as @xmath79 , in the same way the amount of teleportation fidelity for rank two state is given by werner state of rank 2 as @xmath80 .
teleportation fidelity as a function of concurrence for rank dependent werner states are presented in fig .
[ fig:1 ] .
werner states of different ranks exhibit respective rank dependent lower bounds on concurrence below which the state is useless for quantum teleportation .
it can be seen that , in the fidelity - concurrence plane , the allowed values of teleportation fidelity for a fixed value of concurrence of @xmath4 states is bounded below and above by curves corresponding to second and fourth rank werner states respectively .
fidelity of second and fourth rank werner state also coincide respectively with lower and upper bounds on fidelity of @xmath4 states obtained in @xcite .
the curve corresponding to werner state of rank 3 , which lies between the curves of upper and lower bounds , serves as an upper bound on teleportation fidelity of rank 3 mixed states .
this implies , quantum teleportation can be achieved with a resource of low value of concurrence , then it has to be a high rank state .
for example , if we have to use a state with concurrence 0.1 as a quantum teleportation resource , it has to be necessarily a fourth rank state .
in @xcite it is shown that local environment can enhance fidelity of quantum teleportation .
it is shown that for a class of density matrices , interaction with environment enhance the singlet fraction above @xmath81 , and thus making the state useful for quantum teleportation .
there are other methods @xcite like entanglement purification , local filtering , entanglement concentration for single and multiple number of qubits can be made use to improve the teleportation fidelity of quantum channels . in this context
it is important to understand whether the rank dependant bounds on fidelity as a function of linear entropy and concurrence are preserved under operations that enhance fidelity .
we understand that the operations involved in enhancing fidelity need not preserve the rank and hence rank dependant bounds are valid as long as the state is of the respective ranks .
it can be stated that for fixed values of concurrence and linear entropy of a state of fixed rank @xmath1 , the maximum achievable teleportation fidelity is that of werner state of rank @xmath1 .
we proved the existence of rank dependent upper bound on the von neumann entropy as well as linear entropy as measures of mixedness of a general mixed state of a bipartite @xmath0 system for failure of the state to be resource for quantum teleportation .
we constructed rank 3 and rank 2 werner states .
rank dependent werner states exhibit the theoretical upper bounds obtained on von neumann entropy and linear entropy .
further , we proved the existence of rank dependent lower bounds on the concurrence of a @xmath4 mixed states for the failure of states as resource for quantum teleportation and showed that the lower bound on concurrence for rank dependent werner states coincide with the theoretical lower bounds .
we also showed the rank dependant werner states give upper and lower bounds on fidelity as a function concurrence , which are consistent with the results in the literature .
100 c. h. bennett , g. brassard , c. crepeau , r. jozsa , asher peres and w. k. wooters , phys .
* 70 * , 1895 ( 1993 ) .
sandu popescu , phys .
, * 72 * , 797 ( 1994 ) .
r. horodecki , m. horodecki and p. horodecki , phys .
a. , * 222 * , 21 ( 1996 ) .
michael a. nielsen and issac l. chuang , _ quantum computation and quantum information _ , ( cambridge university press , cambridge , england , 2000 ) .
m. horodecki , p .
horodecki and r. horodecki , phys .
a. , * 60 * , 1888 ( 1999 ) . s.bose and v. vedral , phys .
a. , * 61*,040101 ( 2000 ) .
k.g paulson and s.v.m satyanarayana , * 14 * , 1227 - 1237 ( 2014 ) .
reinhard f. werner , phys .
a. , * 40 * , 4277 ( 1989 ) .
j. von neumann , mathematical foundation of quantum mechanics ( princeton univeristy press,1995 ) satoshi isizaka and tohya hiroshima , phys .
a. , * 62 * , 022310 ( 2000 ) .
frank verstraete , koenraad audenaert , tijl de bie , bart de moor , phys .
a. , * 64 * , 012316 ( 2001 ) .
satoshi isizaka and tohya hiroshima ( 2000 ) , _ maximally entangled mixed states under nonlocal unitary operations in two qubits _ , phys .
a. , * 62 * , 02231 frank verstraeteab and henri verschelde , phys .
rev . a 66 : 022307 ( 2002 ) , w.k wootters , quantum information and computation,*1*,27 - 44 p.badziag , m. horodecki , p . horodecki and r. horodecki ( 2000 ) ,
_ local enviornment can enhance fidelity of quantum teleportation _
a. , * 62 * , 012311 c.h bennet et al .
, `` purification of noisy entanglement and faithful teleportation via noisy channels '' , phys . rev .
76 : 722 - 725,c .
h. bennett et al . , phys .
a 53 , 2046 ( 1996 ) frank verstraete , jeroen dehane and bart demoor , phys .
a. , * 64 * , 010101(r),p .
kwiat et al . ,
nature 409 , 1014 ( 2003 ) | entanglement and mixedness of a bipartite mixed state resource are crucial for the success of quantum teleportation . upper bounds on measures of mixedness , namely , von neumann entropy and linear entropy beyond which the bipartite state ceases to be useful for quantum teleportation are known in the literature . in this work ,
we generalize these bounds and obtain rank dependent upper bounds on von nuemann entropy and linear entropy for an arbitrary bipartite mixed state resource .
we observe that the upper bounds on measures of mixedness increase with increase in the rank .
for two qubit mixed states , we obtain rank dependent lower bound on the concurrence , a measure of entanglement , below which the state is useless for quantum teleportation .
werner state , which is a fourth rank state , exhibits the theoretical upper bound on mixedness among two qubit mixed states .
we construct werner like states of lower ranks and show that these states possess the theoretical rank dependent upper bounds obtained on the measures of mixedness and theoretical rank dependent lower bounds on the concurrence . |
the electric dipole moments ( edms ) of fermions are worthwhile to study since they are driven by the cp violating interactions .
cp violation is carried by the complex cabibo kobayashi maskawa ( ckm ) matrix elements in the quark sector and the possible lepton mixing matrix elements in the lepton sector , in the framework of the standard model ( sm ) .
the negligibly small theoretical values of the fermion edms in the sm forces one to search the physics beyond , such as multi higgs doublet models ( mhdm ) , supersymmetric model ( susy ) , @xcite , ... , etc .. there are various experimental results on the fermion edms in the literature and they read , @xmath1 @xcite , @xmath2 @xcite and @xmath3 @xcite respectively . from the theoretical point of view , the extensive work has been done on the edms of fermions in various models , in the literature .
the lepton electric dipole moments have been studied in the seesaw model , in @xcite .
the electron edm has been predicted at the order of @xmath4 , in the 2hdm , including tree level flavor changing neutral currents ( fcnc ) and complex yukawa couplings , in @xcite . the work @xcite is devoted to the fermion edm moments in the sm with the inclusion of non - commutative geometry . in @xcite ( @xcite ) , the electric dipole moments of fermions in the two higgs doublet model with the inclusion of non - universal extra dimensions ( in the split fermion scenario ) have been analyzed .
on the other hand , the quark edms have been estimated in several models @xcite and the edms of nuclei , deutron , neutron and some atoms have been predicted extensively @xcite . in the present work
, we study the lepton edms in the two higgs doublet model in which the flavor changing ( fc ) neutral current vertices in the tree level are permitted and the cp violating interactions are carried by complex yukawa couplings . in the calculations ,
we include the effects of a possible new dimension , in the split fermion scenario , where the hierarchy of fermion masses are coming from the overlap of the fermion gaussian profiles in the extra dimension .
in addition to the strong localization of the leptons , we consider that the new higgs scalars are also localized around the the origin , in the extra dimension .
these localizations are the result of the non vanishing couplings between leptons ( new higgs scalars ) and the so called localizer field , which is odd under @xmath5 reflection in the extra dimension .
the split fermion scenario has been studied in several works in the literature @xcite-@xcite . in @xcite ,
the fermion mass hierarchy has been introduced by assuming that the fermions were located at different points in the extra dimensions , with exponentially small overlaps of their wave functions .
the phenomenologically reliable locations of left and right handed components of fermions in the extra dimensions and their roles on the mechanism of yukawa hierarchies have been studied in @xcite .
@xcite was devoted to prediction of the constraint on the split fermions in the extra dimensions by considering leptonic w decays and the lepton violating processes .
the cp violation in the quark sector in the split fermion model was studied in @xcite and in @xcite , the new configuration of split fermion positions in a single extra dimension and the physics of kaon , neutron and b / d mesons have been analyzed to find stringent bounds on the size of the compactification scale 1/r .
the shapes and overlaps of the fermion wave functions in the split fermion model have been studied in @xcite and the rare processes in the split fermion scenario have been analyzed in @xcite .
@xcite ( @xcite,@xcite ) was related to the electric dipole moments of charged leptons ( the radiative lepton flavor violating ( lfv ) decays , the lfv @xmath6 decays ) in the split fermion scenario .
the work @xcite was devoted to the branching ratios of the radiative lfv decays in the split fermion scenario , with the assumption that the new higgs doublet is restricted to the 4d brane or to the thin bulk in one and two extra dimensions , in the framework of the two higgs doublet model .
recently , the higgs localization in the split fermion models has been studied in @xcite .
the present work is devoted to the edms of charged leptons in the 2hdm with the assumption that they have gaussian profiles in the extra dimension , similar to the previous work @xcite .
however , we also consider that the new higgs scalars are localized in the extra dimension with the help of the so called localizer field .
the idea of the localization of the sm higgs , using the localizer field , has been studied in @xcite . here
, we assume that the new higgs scalars are localized in the extra dimension and the sm higgs has a constant profile .
the localization of the new higgs scalars depends strongly on the strength of the coupling of the localizer to the new higgs scalar . here
we take this coupling small and study the sensitivity of the lepton edms on it .
furthermore , we analyze the compactification scale dependence of lepton edms . in the numerical calculations ,
we observe that the lepton edms are sensitive to the new higgs localization in the extra dimension and we estimate that the numerical values of electron ( muon , tau ) edm is at the order of the magnitude of @xmath7 e - cm , even for the large values of the compactification scale @xmath8 and for the intermediate values of yukawa couplings .
the paper is organized as follows : in section 2 , we present edms of charged leptons in the split fermion scenario , with the localization of new higgs scalars in the extra dimension , in the 2hdm .
section 3 is devoted to discussion and our conclusions . in appendix section
we present the derivation of the kk modes of new higgs fields and their masses in the case that they are coupled to the localizer field .
the existence of the fermion edm is the signal for the cp violation , since it is the result of the cp violating fermion - fermion - photon effective interaction . in the framework of the sm , the complex ckm matrix ( possible
lepton mixing matrix ) is the possible source of this violation for quarks ( for charged leptons ) .
since the theoretical estimated numerical values of the fermion edms are considerably small , there is a need to go beyond the sm to get additional contributions , in order to enhance their numerical values .
the multi higgs doublet models are the among the candidates for this enhancement , and , in the present work , we take the 2hdm , which allows the fcnc at tree level and includes the complex yukawa couplings as a source of cp violation . on the other hand , with the addition of extra dimensions , there exist new contributions sensitive to the compactification scale @xmath8 where @xmath9 is the compactification radius of the extra dimension . here , we take the effects of extra dimensions into account with the assumption that the hierarchy of lepton masses are coming from the lepton gaussian profiles in the extra dimensions , so called the split fermion scenario .
this localization is obtained by the nonzero coupling of the localizer scalar field @xmath10 to the lepton fields .
furthermore , we assume that this field couples also to the new higgs doublet and localizes the new higgs scalars around the origin @xmath11 , where @xmath12 is the coordinate of the extra dimension ( see appendix ) .
the yukawa lagrangian which creates the lepton edm in a single extra dimension , respecting the split fermion scenario , reads : @xmath13 where @xmath14 and @xmath9 denote chiral projections @xmath15 , @xmath16 is the new scalar doublet . here
@xmath17 ( @xmath18 ) , with family indices @xmath19 , are the zero mode lepton doublets ( singlets ) with gaussian profiles in the extra dimension @xmath12 and they read @xmath20 with the normalization factor @xmath21 .
@xmath22 ( @xmath23 ) are the lepton doublets ( singlets ) in four dimensions . here
the parameter @xmath24 is the gaussian width of the leptons with the property @xmath25 , and @xmath26 are the fixed position of @xmath27 left ( right ) handed lepton in the fifth dimension ] ] .
the higgs doublets @xmath28 and @xmath16 are chosen as @xmath29\ , ; \phi_{2}=\frac{1}{\sqrt{2}}\left(\begin{array}{c c } \sqrt{2 } h^{+}\\ h_1+i h_2 \end{array}\right ) \,\ , .
\label{choice}\end{aligned}\ ] ] so that their vacuum expectation values read : @xmath30 this leads to the possibility of collecting sm ( new ) particles in the first ( second ) doublet . in this case @xmath31 and @xmath32 are the mass eigenstates @xmath33 and @xmath34 respectively since no mixing occurs between two cp - even neutral bosons @xmath35 and @xmath33 at tree level . as , @xmath36 with constants @xmath37 , where @xmath10 is the localizer singlet field ( see appendix section for details ) . in this case ,
the first higgs doublet does not couple to the second one and to the localizer singlet field . with this assumption
, the two higgs doublets do not mix and the sm ( new ) particles are placed in the first ( second ) doublet .
if the first and the second doublets couple and the second one acquires non zero vacuum expectation value , the mixing appears between two doublets and it needs further analysis . ]
now we take that the sm higgs has a constant profile in the extra dimension and the new higgs field @xmath16 , which has the main role in the existence of the charged lepton edm , couples to the localizer @xmath10 with a small coupling @xmath38 ( see the appendix section for details ) .
the coupling of the new higgs doublet to the localizer results in the localization of the higgs scalars , with the corresponding kk modes , around the origin , and brings modified yukawa interactions in four dimensions .
the neutral cp even and odd scalar fields s ( @xmath39 ) , are expanded into their kk modes as @xmath40 ( see the appendix section ) and to obtain the lepton - lepton - higgs interaction coupling in four dimensions we need to integrate the combination @xmath41 , appearing in the part of the lagrangian ( eq .
( [ lagrangian ] ) ) , over the fifth dimension . using the kk basis obtained ( see eq .
( [ f0 ] ) and ( [ fn ] ) ) , we get @xmath42 where the factor @xmath43 reads @xmath44 and the function @xmath45 is @xmath46}}\ , .\end{aligned}\ ] ] here , the fields @xmath47 , @xmath48 are the four dimensional lepton fields and the function @xmath49 $ ] is the error function , which is defined as @xmath50=\frac{2}{\sqrt{\pi}}\,\int_{0}^{z}\,e^{-t^2}\,dt \,\ ,
. \label{erffunc}\end{aligned}\ ] ] the functions @xmath51 in eq .
( [ vij1even ] ) are calculated in the case that the coupling @xmath38 is non zero and small : @xmath52}\,e^{f_{lr\,(rl)\,ij } } \nonumber \\ & \times & f_{\!\!\!\!\!\!\!\!1\,\,\,\,\,1}[-n,\frac{1}{2},\frac{\sqrt{g}\,(y_{i l\ , ( r)}+y_{j r\ , ( l)})^2 } { ( 4-g)\,\sigma^2 } ] \label{cevenodd } \ , , \end{aligned}\ ] ] where @xmath53 the function @xmath54 $ ] appearing in eq .
( [ cevenodd ] ) is the hypergeometric function @xmath55=\sum_{k=0}^{\infty}\ , ( a)_k\ , ( b)_k\,z^k / k !
\label{hypergeo}\ , , \end{aligned}\ ] ] where @xmath56}{\gamma[d]}$ ] .
finally , we can define the yukawa couplings in four dimensions as @xmath57 where @xmath58 are the yukawa couplings in five dimensions ( see eq .
( [ lagrangian ] ) ) . in four dimensions , with the definition @xmath59 where n denotes the word `` neutral '' . ]
the effective edm interaction for a charged lepton @xmath60 is given by @xmath61 where @xmath62 is the electromagnetic field tensor , @xmath63 is edm of the charged lepton and it is a real number by hermiticity . with the assumption that the cp violating edm interaction comes from the complexity of the yukawa couplings due to the new higgs scalars , the @xmath60-lepton edm @xmath64 @xmath65 can be calculated as a sum of contributions coming from neutral higgs bosons @xmath66 and @xmath67 ( see fig .
[ fig1 ] ) , @xmath68 for @xmath69 and @xmath70 for @xmath71 . here
the functions @xmath72 , @xmath73 read @xmath74 with @xmath75 , @xmath76 , @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath82 are charges of @xmath83 and @xmath84 leptons respectively . notice that the functions @xmath85 , the zero mode solution reads @xmath86 ( see eq .
( [ f0 ] ) ) , and , if @xmath87 is away from the origin , @xmath88 , the coefficients @xmath85 tends to zero .
this is the case where the new higgs contribution vanishes . ]
are defined in eq .
( [ cevenodd ] ) . in eq .
( [ emuedm ] ) we take into account only the internal @xmath83-lepton contribution respecting our assumption that the yukawa couplings @xmath89 , are small compared to @xmath90 due to the possible proportionality of the yukawa couplings to the masses of leptons in the vertices . in eq . ( [ tauedm ] ) , we present also the internal @xmath84-lepton contribution , which can be neglected numerically .
notice that , we make our calculations in arbitrary photon four momentum square @xmath91 and take @xmath92 at the end .
finally , in our calculations , we choose the yukawa couplings complex and we used the parametrization @xmath93 therefore , the yukawa factors in eqs .
( [ emuedm ] ) and ( [ tauedm ] ) can be written as @xmath94 where @xmath95 . here
@xmath96 is the cp violating parameter which is the source of the lepton edm .
the fermion edm is the result of the existence of the cp violation and arises from the fermion - fermion - photon effective interaction . in the present work , we study the charged lepton edms in the 2hdm with fcnc at tree level and we assume that the cp phase is coming from the complex yukawa couplings , driving the lepton - lepton - new higgs scalar vertices .
furthermore , we take into account the effects of a single extra dimension in the so called split fermion scenario where the hierarchy of lepton masses are coming from the lepton gaussian profiles in the extra dimension . in this scenario ,
the penetration of leptons into the bulk is ensured by the non - zero coupling of leptons with a background scalar field , namely the localizer field , @xmath10 , which has a vacuum expectation value , depending on the coordinate of the extra dimension and centered at the origin .
in addition to this , we consider that the new higgs doublet also couples to the localizer weakly and the corresponding higgs scalars localize around the origin @xmath11 . on the other hand , the sm higgs @xmath35
is taken as uniform in the extra dimension and its vacuum expectation value is responsible for the generation of the masses .
notice that the mass term , which is modulated by the mutual overlap of lepton wavefunctions , is obtained by integrating the operator @xmath97 over extra dimensions , where @xmath60 denotes lepton .
this is the idea of fixing the positions of left ( right ) handed leptons in the extra dimensions ( see @xcite for details ) . since the lepton - lepton - new scalar
vertices creates the cp violating edm interaction , besides zero modes , the kk modes of leptons and higgs scalars have additional contributions . here
, we include the effects of higgs kk modes , however , we ignored the lepton kk mode contributions to the edms of leptons , by expecting that , for the large values of the compactification scale , their effects on the physical parameters are suppressed . in the numerical calculations ,
the free parameters coming the model used should be fixed by using the present experimental results .
the first set of free parameters is the four dimensional leptonic complex couplings @xmath98 .
we consider the yukawa couplings @xmath99 , as smaller compared to @xmath100 and we assume that @xmath101 is symmetric with respect to the indices @xmath102 and @xmath103 . in the case that no extra dimension exists , the experimental uncertainty , @xmath104 , in the measurement of the muon anomalous magnetic moment @xcite can be used to estimate the upper limit of @xmath105 , since the new physics effects can not exceed this uncertainty and this limit was predicted as @xmath106 ( see @xcite and references therein ) . using this upper limit and the experimental upper bound of br of @xmath107 decay , br @xmath108 , the coupling @xmath109 is restricted in the range , @xmath110 ( see @xcite ) . in our calculations ,
we choose the numerical values of the couplings @xmath105 ( @xmath109 ) not larger than @xmath111 ( around @xmath112 ) . for the coupling @xmath113
, we use the numerical values which are greater than @xmath105 , since we have no explicit restriction region . for the cp violating parameter which drives the edm interaction we choose the intermediate value , @xmath114 the second set of free parameters is coming from the split fermion scenario where the hierarchy of lepton masses are due to the lepton gaussian profiles in the extra dimension . here
, we have a free parameter @xmath115 , where @xmath24 is the gaussian width of the fermions and @xmath9 is the compactification radius .
there are various predictions on the compactification radius in the literature .
the lower bound for inverse of the compactification radius was estimated as @xmath116 @xcite in the case of universal extra dimension scenario .
the direct limits from searching for kk gauge bosons imply @xmath117 , the precision electro weak bounds on higher dimensional operators generated by kk exchange place a far more stringent limit @xmath118 @xcite and , from @xmath119 , the lower bounds for the scale @xmath8 have been obtained as @xmath120 , from @xmath121 one got @xmath122 , and from the upper limit of the @xmath123 , @xmath124 , the estimated limit was @xmath125 @xcite . on the other hand , there exist different phenomenologically reliable set of locations of lepton fields in the extra dimension , even if they are predicted by using the lepton masses .
notice that we used the set predicted by the work @xcite .
finally , the coupling @xmath38 , which exists since we consider that the new higgs doublet also couples to the localizer , is another free parameter of the model used . here
, we assume that this coupling is in a region such that the new higgs scalars are localized around the origin as much as possible .
notice that the zero mode higgs boson @xmath33 and @xmath34 masses @xmath126 ( with @xmath127 ) depend on the coupling @xmath38 and the additional term comes from the localization effect , which should not be large .
in the lagrangian ( see eq .
( [ pot1 ] ) ) as @xmath128 and @xmath129 and choose the coupling @xmath38 in the range such that the zero mode masses @xmath130 are around the numerical values as @xmath131 .
] now , we start to estimate the charged lepton edms and study the coupling @xmath38 and the parameter @xmath132 dependence of these physical quantities . in fig .
[ edmeg ] , we plot the electron edm @xmath133 with respect to the coupling @xmath38 for @xmath134 and @xmath135 . here
the lower - upper solid ( dashed ) line represents the edm , due to the zero mode higgs - kk mode higgs included contribution for the parameter @xmath136 ( @xmath137 ) .
the sensitivity of edm to the coupling @xmath38 is strong .
the addition of kk mode contributions enhance the edm weakly , especially for the parameter @xmath136 .
the electron edm is at most at the order of the magnitude of @xmath138 , for @xmath139 and @xmath136 ( @xmath137 ) .
the enhancement in edm for the small values of the coupling @xmath38 ( and also for the large values of the width @xmath24 ) is due the fact that the masses of zero ( and also kk modes ) ( see eq .
( [ kkmodemass ] ) ) are smaller compared to the ones for the large values of the coupling @xmath38 ( and also for the small values of the width @xmath24 ) . in the case that all the 2hdm particles are confined into the 4d brane and the extra dimension effects are switched off , the electron edm @xmath133 is at the order of @xmath140 , which is the numerical value to be reached with smaller coupling @xmath38 and the weak localization of the new higgs doublet , in the present scenario .
[ edmer ] is devoted to the compactification scale @xmath8 dependence of @xmath133 , for @xmath141 and @xmath135 . here
the lower - upper solid ( dashed ) line represents the @xmath133 , due to the zero mode higgs - kk mode higgs included contribution for the parameter @xmath136 ( @xmath137 ) .
the sensitivity of @xmath133 to @xmath8 is strong and it enhances with the decreasing values of @xmath8 , as expected .
the effects of the addition the kk modes on @xmath133 are weak . in fig .
[ edmmug ] , we present @xmath142 with respect to the coupling @xmath38 for @xmath134 , @xmath143 . here
the lower - upper solid ( dashed ) line represents the edm , due to the zero mode higgs - kk mode higgs included contribution for the parameter @xmath136 ( @xmath137 ) .
similar to the electron edm , the sensitivity of @xmath142 to the coupling @xmath38 is strong and the addition of kk modes of new higgs scalars ensures negligible enhancement in the numerical value of @xmath142 .
the muon edm is at the order of @xmath144 , for @xmath139 and @xmath136 ( @xmath137 ) .
the @xmath8 dependence of @xmath142 is presented fig .
[ edmmur ] , for @xmath141 and @xmath135 . here
the lower - upper solid ( dashed ) line represents the @xmath142 , due to the zero mode higgs - kk mode higgs included contribution for the parameter @xmath136 ( @xmath137 ) .
the sensitivity of @xmath142 to @xmath8 is strong similar to the @xmath133 case .
if the extra dimension effects are switched off , the muon edm @xmath142 is at the order of @xmath145 and this numerical value can be obtained with the weak localization of the new higgs doublet , in the present scenario .
finally , we study the tau edm @xmath146 in fig .
[ edmtaug ] and fig .
[ edmtaur ] .
[ edmtaug ] , represents @xmath146 with respect to the coupling @xmath38 for @xmath134 , @xmath147 and @xmath148 . here
the lower - upper solid line represents the edm , due to the zero mode higgs - kk mode higgs included contribution for the parameter @xmath136 .
we observe that @xmath146 enhances up to the values at the order of @xmath149 for @xmath139 and @xmath136 .
the enhancement due to the kk modes of new higgs scalars is negligible .
[ edmtaur ] is devoted to the @xmath8 dependence of @xmath146 for @xmath141 and @xmath150 . here
the solid ( dashed ) line represents the @xmath146 , due to the zero mode higgs ( kk mode higgs included ) contribution for the parameter @xmath136 .
the sensitivity of @xmath146 to @xmath8 is strong and almost one order enhancement with the decreasing values of @xmath8 from @xmath151 to @xmath152 .
the tau edm @xmath146 is at the order of @xmath153 in the case that the extra dimension effects are switched off .
such a numerical value is obtained for the weak localization of the new higgs doublet , similar to the electron and muon edm case .
+ now we would like to summarize our results : * the lepton edms are weakly sensitive the kk mode contributions in the given range of the coupling @xmath38 , however this sensitivity increases for the small values of this coupling .
furthermore , the numerical values of edms increase with the decreasing values of the coupling @xmath38 and increasing values of parameter @xmath132 , which measures the gaussian widths of leptons in the extra dimension . * the sensitivity of lepton edms to @xmath8 increases with the decreasing values of the parameter @xmath132 .
* we obtain the numerical values at the order of @xmath154 for @xmath139 and @xmath136 , for the edms @xmath133 ( @xmath155 , respectively . to be able to reach the results near the current experimental limits ,
the following domains for the numerical values of the various free parameters should be considered : @xmath156 , @xmath157 , @xmath158 , @xmath159 , @xmath160 and @xmath139 .
it is obvious that these theoretical values of the edms of charged leptons are still far from the experimental limits .
furthermore , the bound for the compactification scale almost coincides with the following bounds existing in the literature : @xmath161 @xcite in the case of universal extra dimension scenario and @xmath162 from @xmath163 @xcite .
hopefully , the forthcoming more sensitive experimental measurements will ensure smaller numerical values for the lepton edms and , therefore , the possible range of the free parameter set will be more accurate . as a final word , the future experimental measurements on leptons edms would ensure valuable information about the possibility of the existence of localizations of leptons and higgs bosons in the extra dimensions and make it more clear how the nature behaves beyond the sm .
here , we calculate the kk modes of higgs fields and their masses in the case that the localizer field @xmath10 couples to the new higgs fields weakly and localize these new fields around @xmath11 , where @xmath12 is the coordinate of the extra dimension . our the starting point is the lagrangian of the new scalar field , which is coupled to the localizer field : @xmath164 with @xmath165 , and @xmath166 with @xmath167 .
here we consider the potential @xmath168 , @xmath169 where @xmath10 is the localizer singlet field , which is chosen as @xmath170 ( see @xcite ) .
it has a linear dependence to the new coordinate @xmath12 for the points near to the fixed point and it couples weakly to new scalar fields @xmath171 . in the full lagrangian .
] now the lagrangian @xmath172 can be rewritten as @xmath173 with the expansion over the kk modes as @xmath174 one gets the mass equation @xmath175 the ground state ( the zero mode ) solution reads @xmath176 with the normalization constant @xmath177}}$ ] and its corresponding mass @xmath178 @xmath179 with @xmath180 . here
@xmath84 is the energy scale which adjusts the localization of leptons in the extra dimension , namely , @xmath181 .
the kk modes for @xmath182 can be obtained by using the operator @xmath183 as @xmath184 where @xmath185 and @xmath186 is the normalization constant for @xmath187 kk mode , + @xmath188}}$ ] .
finally the masses of the kk modes , including the zero one , becomes , since the extra dimension is compactification on the orbifold @xmath189 . in this case , the kk mode masses read @xmath190 @xmath191 1 c. r. schmidt and m. e. peskin , _ phys . rev . lett .
_ * 69 * ( 1992 ) 410 .
e. d. commins et.al . _ phys .
a _ * 50 * ( 1994 ) 2960 . j. bailey et al , _ journ .
* g4 * ( 1978 ) 345 ; particle data group , d. e. groom et.al .
_ european phys .
journ . _ * c15 * ( 2000 ) 1 .
b. dutta , r. n. mohapatra , _ phys .
rev . _ * d68 * ( 2003 ) 113008 .
e. iltan , _ phys .
* d64 * ( 2001 ) 013013 .
e. iltan , _ jhep _ * 065 * ( 2003 ) 0305 .
e. iltan , _ jhep _ * 0404 * ( 2004 ) 018 .
e. iltan , _ eur .
phys . _ * c44 * ( 2005 ) 411 .
e. p. sahabalin , _ sov .
_ * 28 * ( 1978 ) 75 ; j. f. donoghue , _ phy .
* d18 * ( 1978 ) 1632 ; a. czarnecki and b. krause , _ phys .
* 78 * ( 1997 ) 4339 ; i. b. khriplovich , _ yad
* 44 * ( 1986 ) 1019 ( _ sov .
* 44 * ( 1986 ) 659 ) ; _ phys .
_ * b173 * ( 1986 ) 193 ; a. czarnecki and b. krause , _ acta phys .
polon . _ * b28 * ( 1997 ) 829 ; _ phys . rev . lett _
* 78 * ( 1997 ) 4339 ; s , weinberg , _ phys .
* 58 * ( 1976 ) 657 ; n. g. deshpande and e. ma , _ phys .
rev._*d 16 * ( 1977 ) 1583 ; for a recent review , see h. y. cheng , _ int .
phys._*a 7 * ( 1992 ) 1059 ; g. c. branco and m. n. rebelo,_phys .
* b 160 * ( 1985 ) 117 ; j. liu and l. wolfenstein , _ nucl .
phys .. _*b 289 * ( 1987 ) 1 ; c. h. albright , j. smith and s. h. h. tye , _ phys . rev .
d. _ * d 21 * ( 1980 ) 711 ; d. atwood , s. bar - shalom , a. soni , _ phys .
rev . _ * d 51 * ( 1995 ) 1034 ; e. iltan , _ j. phys . _ * g27 * ( 2001 ) 1723 ; e. iltan , _ phys.rev . _ * d65 * ( 2002 ) 073013 ; y. liao and x. li , _ phys . rev . _ * d60 * ( 1999 ) 073004 ; d. g. dumm and g. a. g sprinberg , , _ eur - phys . j. _ * c 11 * ( 1999 ) 293 ; a. soni and r. m. xu , _ phys . rev .
* 69 * ( 1992 ) 33 .
v. baryshevsky , hep - ph/0402209 ; o. lebedev , k. a. olive , m. pospelov , a. ritz , _ phys . rev .
_ * d70 * ( 2004 ) 016003 ; a. g. rodriguez , m. a. h. ruiz , a. santiago , _ phys .
_ * d69 * ( 2004 ) 073008 ; a. gutierrez - rodriguez , m.a.hernandez-ruiz and l.n .
luis - noriega , _ mod .
_ * a19 * ( 2004 ) 2227 ; r. e. mischke , aip conf.proc.675:246-250,2003 ; d. a. demir , oleg lebedev , k. a. olive , m. pospelov , a. ritz , _ nucl.phys . _ * b680 * , ( 2004 ) 339 ; m. endo , m. kakizaki , m. yamaguchi , _ phys.lett . _ * b 583 * ( 2004 ) 186 ; d. kawall , f. bay , s. bickman , y. jiang , d. demille , _ phys . rev . lett .
_ * 92 * ( 2004 ) 133007 ; i. masina , c. savoy , _ phys.lett . _ * b579 * ( 2004 ) 99 ; j. hisano , y. shimizu , _ phys.lett .
_ * b581 * ( 2004 ) 224 - 230 ; a. syntfeld , h. mach , w. kurcewicz , b. fogelberg , w. pociennik , e. ruchowska _ phys.rev . _ * c68 * ( 2003 ) 024304 ; t. suzuki , h. sagawa , k. hagino , _
phys.rev . _
* c68 * ( 2003 ) 014317 .
n. arkani - hamed and m. schmaltz , _ phys .
rev . _ * d61 * , 033005 ( 2000 ) ; n. arkani - hamed , y. grossman and m. schmaltz , _ phys .
* d61 * ( 2000 ) 115004 . i. antoniadis , _ phys .
lett . _ * b246 * , 377 ( 1990 ) .
e. a. mirabelli , schmaltz , _ phys .
rev . _ * d61 * ( 2000 ) 113011 .
w. f. chang , i. l. ho and j. n. ng , _ phys .
rev . _ * d66 * 076004 ( 2002 )
. g. c. branco , a .
gouvea , m. n. rebelo , _ phys .
* b506 * ( 2001 ) 115 .
w. f. cang , j. n. ng , _ jhep _ * 0212 * ( 2002 ) 077 .
y. grossman and g. perez , _ phys .
* d67 * ( 2003 ) 015011 ; _ pramana _ * 62 * ( 2004 ) 733 .
b. lillie , j. l. hewett , _ phys .
* 68 * ( 2003 ) 116002 .
a. delgado , a. pomarol , m. quiros , _ jhep _ * 0001 * ( 2000 ) 030 ; y. grossman , _ int .
_ * a15 * ( 2000 ) 2419 ; d. e. kaplan and t. m. tait , _ jhep _ * 0111 * ( 2001 ) 051 ; g. barenboim , _ et .
_ , _ phys . rev .
_ * d64 * ( 2001 ) 073005 ; w. f. chang , i. l. ho and j. n. ng , _ phys .
_ * d66 * ( 2002 ) 076004 ; w. skiba and d. smith , _ phys .
_ * d65 * ( 2002 ) 095002 ; y. grossman , r. harnik , g. perez , m. d. schwartz and z. surujon , _ phys .
_ * d71 * ( 2004 ) 056007 ; p. dey , g. bhattacharya , _ phys .
rev . _ * d70 * ( 2004 ) 116012 ; y. nagatani , g. perez , _ jhep _ * 0502*(2005 ) 068 ; r. harnik , g. perez , m. d. schwartz , y. shirman , * 0503 * ( 2005 ) 068 .
e. o. iltan , hep - ph/0504013 .
e. o. iltan , _ eur.phys.j . _
* c46 * ( 2006 ) 487 .
e. o. iltan , _ eur.phys.j . _ * c46 * ( 2006 ) 475 .
z. surujon , _ phys .
_ * d73 * ( 2006 ) 016008 .
m. deile , et .
al . , muon g-2 collaboration , , _ nucl.phys.proc.suppl . _
* 116 * ( 2003 ) 215 .
e. iltan , h. sundu , _
acta phys.slov._ * 53 * ( 2003 ) 17 .
t. appelquist , h .- c .
cheng , b. a. dobrescu , _ phys .
* d64 * 035002 ( 2001);t .
appelquist , h. yee , _ phy .
_ * d67 * 055002 , 2003 ; t. g. rizzo , j. d. wells , _ phy .
_ * d61 * , 016007 ( 2000 ) . | we study the lepton electric dipole moments in the split fermion scenario , in the two higgs doublet model , where the new higgs scalars are localized around the origin in the extra dimension , with the help of the localizer field .
we observe that the numerical value of the electron ( muon , tau ) electric dipole moment is at the order of the magnitude of @xmath0 and this quantity is sensitive the new higgs localization in the extra dimension . |
results of linear stability calculations suggest that excitation of high - order g modes in main - sequence b - type stars must be a common phenomenon .
the instability is typically robust .
there is a significant excess of driving over damping and many modes are simultaneously unstable in individual stellar models . with the standard population i composition
, the instability is found in models of nearly all b - type stars . yet
, the observed pulsation associated with the slow modes is unspectacular .
individual modes never attain high amplitudes . only in a small fraction of b stars
do we have a firm evidence for the presence of oscillations and typically only few modes are seen .
one possible reason why , despite strong instability , we never see high amplitude pulsation of this type is a collective saturation by a large number of modes which are difficult to detect .
recent report by walker et al.(2005 ) on the detection of about sixty periods in the data on hd 163868 obtained from _ most _ may support this idea .
the fact that hd 163868 is a rapidly rotating be type object makes the detection of many oscillation modes particularly interesting .
the cause of activity observed in stars of this type is still not understood .
seismic constraints derived from mode frequencies on the internal structure would be of great value in this context . furthermore , it is possible that oscillations play an active role in these objects by inducing angular momentum transport , as first suggested by osaki ( 1986 ) and penrod ( 1986 ) and more recently , in the context of opacity - driven modes , by lee ( 2006 ) .
let us add that an asymmetry in excitation of prograde and retrograde modes may induce a net helicity of motion and thus an action of magnetic dynamo . to extract seismic information , as well as to assess possible active role of oscillation
, we have to identify modes excited in the star .
the task of individual mode identification in an object like hd 163868 is formidable and will be attempted in this work .
like walker et al.(2005 ) , we will only try to determine angular degrees and azimuthal orders of the three groups of modes which are centered on periods of 8 days , and 14 and 7 hours .
these authors , however , have not found unstable modes that could explain the longest period group .
we will argue that the observed modes represent only a small subset of the excited modes .
other modes having similar rms amplitudes may be not detectable because of cancellation and/or unfavourable aspect .
our analysis of stability and visibility is based on linear non - adiabatic calculations with the effects of rotation treated in the traditional approximation .
the same approximations were adopted by townsend ( 2005a , b ) in his extensive survey of stability of slow modes in models of rotating main - sequence b - stars .
the slow modes include high - order g modes with frequencies comparable to the angular rate of rotation and certain r modes ( called mixed gravity - rossby modes ) , which at sufficiently high rotation rate become propagatory in the radiative envelopes .
both types are candidates for identification of modes excited in hd 163868 .
our identification differs from that of walker et al.(2005 ) .
one may suspect that the difference should be blamed to our use of the traditional approximation , which they avoid .
however , as we explain in the next section , the approximation should be valid in the application to , at least , certain modes in hd 163868 . in section 3
, we present results on stability of slow modes in a model of hd 163868 which is very similar to the model used by walker et al .
we review properties of unstable modes in this and similar models and compare our results with those of our predecessors . in section 4 we quantify the dependence of observable amplitude on the mode geometry and the aspect angle for the most unstable low - degree modes .
comparison with observations is presented in section 5 .
there is a number of recent papers where _ the traditional approximation _
is used to describe properties of low frequency modes in rotating stars ( e.g. bildstein , ushomirsky & cutler 1996 ; lee & saio 1997 ; townsend 2003a ; savonije 2005 ; townsend 2005a , b ) .
readers are referred , in particular , to townsend ( 2003a ) for details of derivation of simplified equations for adiabatic oscillations known as _
laplace s tidal equations _ , which employ the traditional approximation .
the last three papers extend the approximation to the case of non - adiabatic oscillation . here , we focus on the validity of approximation and its extension to the case of non - uniform rotation and we will quote only the most essential equations . the generally adopted , initial simplifications of the linearized euler equations , @xmath0 where all notation is standard , are ( i ) slow and ( ii ) uniform rotation , ( iii ) oscillation frequency , @xmath1 , of the same order as the angular velocity of rotation , @xmath2 , ( iv ) neglect of @xmath3 ( the cowling approximation ) , ( v ) adiabaticity of perturbations . here
we adopt ( i ) and ( iii ) , that is , we assume @xmath4 but not ( ii ) as we allow the @xmath5 dependence in @xmath2 .
indeed , the traditional approximation applies to this more general case in the essentially unchanged form .
we begin assuming ( v ) but later we will consider non - adiabatic effects , which are in fact of main interest for us in this paper .
inequality ( 2 ) implies that the effects of centrifugal distortion may be neglected which together with ( iv ) reduces eq.(1 ) to @xmath6 since we allow non - uniform rotation , the constant eigenfrequency , @xmath7 , must be calculated in the inertial system , while @xmath1 - the frequency calculated in the local corotating system - may vary with @xmath5 .
the acceleration in the inertial system expressed in terms of the displacement vector , @xmath8,\ ] ] is given by @xmath9 where @xmath10 .
note that with the adopted form of the time dependence , the modes of azimuthal order @xmath11 are _ prograde _ and those with @xmath12 are _
retrograde_. this is opposite to what has been adopted in the cited papers employing traditional approximation but it is in agreement with the _ eme resolution _ ( see appendix to christensen - dalsagaard & dziembowski 2000 ) . using the adiabaticity condition @xmath13 where we denoted @xmath14 and introducing one additional eigenfunction , @xmath15 , through the following identity @xmath16,\ ] ] we get from eq.(3 ) @xmath17 @xmath18 and @xmath19 in eq.(8 ) we use standard notation @xmath20 for the brunt - visl frequency and @xmath21 for the logarithmic derivative of the fractional mass .
the continuity equation combined with eq.(6 ) yields @xmath22 which closes the system of equations for the eigenvalues @xmath7 and the associated eigenfunctions @xmath23 , @xmath24 , @xmath25 , and @xmath26 . in the traditional approximation all terms arising from the coriolis force , except those in eqs.(9 ) and ( 10 ) which contain horizontal components of @xmath27 , are dropped out .
a simple justification , based on the local wave approximation , was given by lee and saio ( 1997 ) .
the approximation is valid if @xmath28 earlier , the same condition was obtained by dziembowski & kosovichev ( 1987 ) , who did not rely on the wave approximation and did not assume uniform rotation .
conditions ( 2 ) and ( 12 ) justify the neglect of those coriolis terms and the term containing the derivative of @xmath2 , both in the outer evanescent zone ( assuming that it is radiative ) as well as in the g - wave propagation zone . in both zones , we have @xmath29 , which justifies simplification of eqs.(9 ) and ( 10 ) .
the relative scaling of @xmath23 and @xmath26 is different in the evanescent and g - wave zone . in
the former , where @xmath30 and @xmath31 are large , @xmath32 and the simplification in eq.(8 ) follows from @xmath33 . in the propagation zone , where the mode amplitude varies rapidly
, we have @xmath34 which justifies the neglect of all terms containing @xmath2 .
note that the @xmath35 term in the coefficient at @xmath23 could have been ignored too but traditionally this is not done . with the traditional approximation , the @xmath5 and @xmath36 dependence in the eigenfunctions may be separated .
the latter dependence is given by the hough functions .
following townsend ( 2003a ) , we substitute in eqs.(8 - 11 ) @xmath37 @xmath38 to obtain ordinary differential equations for @xmath39 and @xmath40 , @xmath41 and @xmath42 where , for a uniform rotation , @xmath43 is the separation parameter .
the radial dependence for the horizontal components of @xmath44 is given by @xmath45 eqs.(15 ) and ( 16 ) look just like in the case of no rotation , except that now we have @xmath46 and in the place of @xmath47 the separation parameter @xmath43 , which is determined as an eigenvalue in the equations for the hough functions .
we write these equations in the form @xmath48
@xmath49 @xmath50\theta+ms\mu\hat\theta,\ ] ] where @xmath51 and @xmath52 is called _ the spin parameter_. eqs.(19 - 20 ) differ from townsend s eqs.(21 - 22 ) only in the sign of @xmath53 , which he denoted @xmath54 , in consequence of different sign convention for the time dependence .
the equations together with boundary conditions at @xmath55 ( symmetry ) and @xmath56 ( regularity ) define the eigenvalue problem on @xmath43 .
the dependence of @xmath43 on @xmath53 and related properties of the hough functions have been discussed in great detail by , for example , bildstein et al.(1996 ) , lee & saio ( 1997 ) , and by townsend ( 2003a ) . of our interest here
are only modes which are propagating in radiative zones , where @xmath57 , that is , corresponding to @xmath58 .
these are the g modes for which @xmath59 at @xmath60 , as well as the mixed gravity - rossby modes , for which @xmath43 changes sign from minus to plus at @xmath61 ( for brevity , in this paper , we shall call them r modes ) .
it was shown , independently by townsend ( 2005b ) and savonije ( 2005 ) and confirmed by lee ( 2006 ) , who did not use the traditional approximation , that such modes may be driven in b - type and early a - type stars .
let us note that if @xmath2 is a function of @xmath5 , then such modes may be trapped in the region of rapid rotation where @xmath58 while it is @xmath62 in the rest of the star interior . in the case of massive main - sequence stars , the traditional approximation can not be used in their convective cores .
however , low frequency waves do not propagate there and thus the core surface may be treated as a boundary .
except for zero - age main - sequence stars there is a nearly discontinuous transition at the core edge , @xmath63 , from @xmath64 to @xmath65 . in the core , @xmath26 and other eigenfunctions
are slowly varying with distance from the center .
thus , with @xmath65 , eq.(8 ) implies @xmath66 on the other hand , in the propagation zone , we have @xmath67 where the local radial wave number is given by @xmath68 and @xmath69 is the horizontal wave number .
hence , @xmath70 since there is no jump of density at @xmath63 , the continuity of @xmath71 and @xmath72 imply the continuity of @xmath26 .
thus , from eqs.(21 ) and ( 22 ) , we get the boundary condition consistent with the traditional approximation to be imposed on solutions of eqs.(15 ) and ( 16 ) , @xmath73 in stars of our interest , there are thin convective layers , connected with the heii ionization and , if @xmath74 , with the fe opacity bump . because of their small vertical extent , such layers can not affect significantly global mode geometry . the two boundary conditions , which should be applied on solution of eqs .
( 15 - 16 ) , are @xmath75 the same boundary conditions remain valid when non - adiabatic effects are taken into account .
eqs.(15 - 20 ) remain valid in the case of shellular rotation but then , both @xmath1 and @xmath43 must be regarded as functions of @xmath5 .
the derivation given by dziembowski & kosovichev ( 1987 ) employed expansion of the @xmath76 dependence in a series of the associated legendre functions . assuming @xmath77 , they decomposed the infinite system of differential equations in @xmath5 into asymptotically independent second - order equations equivalent , in the limit @xmath78 , to eqs.(15 - 16 ) . in their derivation
there is an inconsequential omission of terms @xmath79 and an implicit assumption that the logarithmic derivative of @xmath2 is of the order of 1 .
townsend ( 2005a ) and savonije ( 2005 ) extended the traditional approximation to the case of non - adiabatic oscillations .
using the same equation as in the case of no rotation with only @xmath43 in place of @xmath47 is incorrect in the expression for the horizontal heat losses .
however , these losses , which are proportional to @xmath80 , make rather small contribution to the overall work integral .
another simplification is the use of @xmath43 with @xmath1 calculated in the adiabatic approximation .
this is well justified in our applications because the non - adiabatic change in @xmath1 is very small .
lee & saio ( 1989 ) pointed out that even at @xmath81 coupling between modes of close frequencies , the same @xmath82 and symmetry may have a significant effect . in a specific example of @xmath83 star sequence , lee ( 2001 ) found that some modes , which are unstable if the traditional approximation is used , are found stable if the approximation is abandoned and the truncated expansion in legendre functions is used for the eigenfunctions .
however , many modes remained unstable and we expect that this applies also in the case of models considered in this paper .
we will discuss this matter further in section 3 .
with the traditional approximation , stability properties of slow modes are determined by two parameters : @xmath43 and @xmath1 . in this section ,
we assume uniform rotation , hence @xmath1 is the mode eigenfrequency in the corotating system .
the whole influence of rotation and the azimuthal order is absorbed in the @xmath84 dependence .
thus , like in the case of no rotation , we deal with a two - dimensional instability range for a specified model . in the present case ,
the two parameters are @xmath1 and @xmath85 .
the latter determines radial order of the mode , @xmath86 , and the @xmath5 dependence of the lagrangian pressure perturbation , @xmath87 . for instability
, the amplitude of the perturbation should be large and slowly varying in the layer of the iron opacity bump .
the frequency , @xmath1 , determines resulting perturbation of the radiative flux .
furthermore , a match is needed between the thermal time - scale of this layer , @xmath88 , and the pulsation period , @xmath89 . [ cols="^,^,^,^,^,^,^,^,>,>,^",options="header " , ] the instability ranges in terms of @xmath90 and @xmath1 for non - rotating main - sequence b - stars were discussed in detail by dziembowski , moskalik & pamyatnykh ( 1993 ) and by pamyatnykh ( 1999 ) .
a corresponding discussion for rotating stars was presented recently by townsend ( 2005a ) . for a family of stellar models ,
@xmath85 does not determine mode order but still , to a large extent , determines the shape of the pressure eigenfunction in the outer layers , which bring the main contribution to the work integral .
thus , the evolutionary radius increase results in longer periods of unstable modes
. the associated decrease of the effective temperature , causing an increase of @xmath88 , acts in the same direction .
this is why the initial effect of evolution is an enhancement of instability .
the tendency is reversed only near the end of the main - sequence evolution when dissipation in the g - wave propagation zone becomes significant .
too small values of @xmath88 are the cause why the modes with low values of @xmath43 are stable in massive ( @xmath91 ) hot objects .
in this paper we present numerical results for a single stellar model calculated with warsaw - new jersey stellar evolutionary code ( see e.g. pamyatnykh 1999 ) assuming no mixing beyond the convective core and including only mean effect of the centrifugal force . the same approximations were adopted by walker et al.(2005 ) .
our model is characterized by the following parameters : @xmath92 , @xmath93 , @xmath94 , @xmath95 , @xmath96 , @xmath97300 km / s , which are very similar to those used by these authors .
our non - adiabatic pulsation code uses the same approximation as adopted by townsend ( 2005a ) except that we apply the inner boundary condition @xmath98 at the edge of convective core .
the code is almost a trivial modification of our standard nonradial pulsation code in its version adopting the cowling approximation , which is indeed very accurate for all considered modes . in the adiabatic part of the code values of @xmath43
are interpolated from the tables of @xmath84 prepared with a separate code for all types of modes of our interest .
the values were not changed in the non - adiabatic part .
we focused the main attention on modes which may be driven by the opacity mechanism and are potentially detectable by photometry .
thus , the modes considered have @xmath43 greater than zero but not too large . for each @xmath99 , we included g modes with the lowest values @xmath43
. sequences of modes of different radial order are denoted with @xmath100 , where @xmath101 .
we also included two sequences of lowest order retrograde r modes with @xmath58 for @xmath102 .
the mode instability is characterized by the normalized work integral @xmath103 where @xmath104 is the usual work integral .
the growth rate , @xmath105 , has the same sign but its value is affected mostly by mode inertia .
the value of @xmath106 , which varies between -1 and 1 , is a better measure of robustness of the instability and a better predictor of the mode amplitude .
fig.1 shows the values of @xmath106 for all considered modes in their range of instability and close to it .
table 1 provides more information about most unstable modes ( largest @xmath106 ) .
the instability range in the observed frequencies , @xmath107 , extends from nearly zero for retrograde @xmath108 g modes to 4 c / d for prograde @xmath109 modes .
the former modes have frequencies in the vicinity of @xmath110 .
the modes with frequencies @xmath111 are seen as prograde modes .
the range of frequencies in the corotating system is much narrower .
we may see in table 1 that the most unstable modes occur near @xmath112 c / d , @xmath113 and @xmath114 .
the prograde @xmath108 modes have @xmath115 and , thus , at @xmath85 values favourable for driving , too long periods .
such modes were found unstable in somewhat cooler models . a lower value of @xmath116 would also result in lower @xmath117 of the unstable modes .
the shift in @xmath118 is the same but , if @xmath119 , of opposite sign . in the vicinity of our model
a shift of 0.01 in @xmath116 leads to the 0.1 c / d frequency shift of most unstable modes .
a change in the adopted value of @xmath120 would shift the instability ranges in @xmath118 for @xmath121 modes .
there are many more strongly unstable modes than shown in fig.1 .
the instability continues up to @xmath122 .
modes with @xmath123 are found up to @xmath124 , which corresponds to @xmath125 for sectorial prograde modes .
owing to higher @xmath43 values such modes are subject to much larger observable amplitude reduction .
visibility of modes is the subject of the next section . here , we just want to point out that we should expect in any of multiperiodic b - type star many more excited modes than those detected
. such modes have properties similar to those listed in table 1 , except they would have much lower surface averaged amplitudes .
possible excitation of many unobservable modes must be kept in mind while considering the feedback of oscillations on stellar rotation .
the results shown in fig.1 and in table 1 might suggest that there is preference to retrograde mode excitation but it is not found in the full set of the unstable modes .
for instance , @xmath126 was found in over 450 prograde modes , and in less than 370 retrograde modes .
certainly , there is no evident asymmetry in driving prograde and retrograde modes .
our findings regarding mode instability differ from that presented by walker et al.(2005 ) . unlike these authors
, we do not find instability of the prograde @xmath108 modes . on the other hand ,
we find many retrograde @xmath108 modes , which may account for the longest periods measured in hd 163868 , whereas no instability of such modes was found in that work .
the general pattern of the difference is that we find instability in many more modes with higher @xmath43 .
perhaps the traditional approximation , which we rely on , is not applicable .
the mode coupling , as described by lee ( 2001 ) , may stabilize only certain modes .
could it stabilize all retrograde @xmath108 modes and destabilize all prograde @xmath108 modes ?
we do not know the answers . perhaps , there is a difference in internal structure between the models used by us and that used by walker et al.(2005 )
. we should also be aware that the lee & saio method , which is based on a truncated series expansion , used in that paper , may not be accurate . considering possible identification of modes detected in hd 163868
, we should take into account visibility conditions for modes described by the hough functions .
observed amplitude of a mode depends on its true amplitude , geometry and the aspect angle , @xmath127 .
photometric amplitudes for modes described by single spherical harmonics were calculated in a number of papers .
results of corresponding calculations for modes described by the hough functions were first published by townsend ( 2003b ) , who relied on truncated expansion of @xmath128 in associated legendre functions . in our calculations of the photometric and radial velocity amplitudes
, we avoid such expansion and carry out two - dimensional integration over visible hemisphere .
details of our approach and a survey of the results will be published elsewhere ( daszyska - daszkiewicz et al .
, in preparation ) .
here we quote only the background formulae and present discussion of visibility of modes listed in table 1 as a function of aspect . within the linear non - adiabatic theory
, there is a freedom in normalization of eigenfunctions .
here we adopt the most common one , that is @xmath129 , and @xmath130 thus , we write @xmath131 where @xmath132 $ ] .
the value of @xmath133 ( not to be confused with @xmath134 in section 2 ) determines true mode amplitude .
note that the surface displacement is given by @xmath135 $ ] and its rms value is equal to @xmath136 .
the surface boundary condition , @xmath137 , combined with eqs.(4 ) and ( 14 ) leads to the following expressions for the horizontal components of the displacement @xmath138 and @xmath139 these components are needed only for calculation of the pulsation velocity field . for evaluation of the photometric amplitudes , we need a linear relation between the radial components of displacement and perturbed flux which we write in the form @xmath140 where @xmath141 is a complex coefficient determined from linear non - adiabatic calculations .
its values for modes considered are given in the last column of table 1 . in order to calculate the flux in a specified band , @xmath142 , we need also an expression for the perturbed surface gravity , which is @xmath143 and the perturbed surface element , @xmath144 where @xmath145 .
the tangential components arise from the change of the normal to the stellar surface .
these components are needed also in the expression for the perturbed limb - darkening , @xmath146 .
the perturbed flux is thus given by @xmath147\cdot{\mbox{\boldmath{$n$}}}_{\rm obs},\ ] ] where @xmath142 represents a photometric band , @xmath148 is the unit vector toward observer , and the integration is done over the visible unperturbed hemisphere .
the limb - darkening coefficient is normalized , so that @xmath149 where @xmath150 .
static atmosphere models are used to calculate wavelength - dependent flux , @xmath151 , and the limb - darkening coefficient , @xmath152 .
the disc - averaged change of radial velocity is given by the expression @xmath153 where @xmath154 in the linear approximation , the contribution from the first term may be calculated with unperturbed @xmath155 , @xmath156 , and @xmath157 .
the contribution from the second term arises solely from perturbation of these quantities .
thus , the radial velocity change is given by @xmath158\cdot{\bf n}_{\rm obs}\tilde\mu+ \nonumber\\ & & ( { \mbox{\boldmath{$v$}}}_{\rm rot}\cdot{\bf n}_{\rm obs}){\delta d{\bf s}\over ds}\cdot{\bf n}_{\rm obs}\biggr\}{\cal f}_xh_x ds.\end{aligned}\ ] ] we show in fig.2 the amplitudes of light and radial velocity as functions of the inclination angle for the modes listed in table 1 .
these amplitudes are calculated with the arbitrary choice @xmath159 and for the geneva @xmath160 band , which is similar to johnson s @xmath160 .
the plotted quantities are @xmath161 in the dependence @xmath162 , we see certain patterns which are the same as for modes described by single spherical harmonics .
zonal modes are best seen from near polar direction , while symmetric ( even ) modes are best seen from near equatorial direction .
there are , however , significant differences .
in particular , an increase of @xmath43 does not lead to so large amplitude reduction as expected from the @xmath163 replacement .
compare , for instance , two sectorial @xmath108 modes .
there is the factor 25 difference in @xmath43 between @xmath164 and @xmath165 modes and only the factor 4 in @xmath166 , while the factor @xmath167 would be expected in the non - rotating case , on the basis of the decline of the disc - averaging factor , @xmath168 , with @xmath90 ( e.g. dziembowski 1977 ) .
according to our calculations the @xmath165 mode is stable . among unstable modes , the @xmath169 is the second best visible , after @xmath170 , from the near equatorial directions . despite of near - equatorial trapping caused by rotation , the optimum inclination for detection of tesseral modes , such as @xmath171 ,
is shifted toward the pole because then cancellation of contributions from the two hemispheres , which have opposite signs , is reduced .
such a cancellation is responsible for rather low amplitudes of the r modes .
the behaviour of the radial velocity amplitudes , depicted in the lower panels of fig.2 , shows rather different pattern . at intermediate inclinations of the rotation axis ,
the r modes turn out most easily detectable . at large inclinations
( near equatorial observer ) the modes @xmath172 are best seen both in light and radial velocity variations .
however , the @xmath173 ratio exhibits a strong mode dependence .
interpretation of the radial velocity amplitude pattern is complicated because the contribution from pulsation and rotation are of comparable size . here
we would like only to stress that the strong mode dependence of the amplitude ratios points to the good prospect for mode identification by combining spectroscopy and photometry data .
the measured value of @xmath174 in hd 163868 is high .
therefore , we assume that the star is seen from a near equatorial direction and , as possible candidates for identification , we consider only symmetric modes .
there are three groups of frequencies found by walker et al.(2005 ) in the star . that with @xmath175 c / d could only be interpreted in terms of the retrograde @xmath108 modes .
possible interpretations for the group with @xmath118 between 1.4 and 2 c / d are @xmath109 modes with @xmath176 and @xmath177 . for the group with @xmath118 between 2.8 and 3.75
the only interpretation is @xmath178 . in this last case
our mode identification is the same as that of walker et al.(2005 ) .
in fig.3 we see that our interpretation is also not fully satisfactory though , unlike our predecessors , we have no difficulty with interpretation of lowest frequency group , where the highest amplitudes are found . just like theirs ,
our predicted range at the highest frequencies is somewhat narrow .
the largest discrepancy between results of our calculation and observation is for the intermediate group .
certainly , the discrepancy may be reduced by adjusting @xmath120 and @xmath116 .
it is possible that the some of the peaks in the hd 163868 osillation spectrum , in particular those in the low frequency range , are artefacts of the data processing .
interpretation of peaks in terms of specific modes should be tested with simultaneous multiband photometry and radial velocity data .
we have seen in the previous section that the ratio of radial velocity to light amplitude is strongly mode dependent and it is exceptionally large for the retrograde @xmath108 modes , which may explain peaks in the low frequency range .
the dominant peak in this range and in the whole oscillation spectrum has photometric amplitude of 0.013 mag in the _ most _ photometric system .
assuming that the amplitude in the @xmath160 band is similar and using plots in fig.2 , we find radial velocity amplitude of about 30 km / s , which should be easily measured in radial velocity data spanning sufficiently long time interval .
data on oscillations in be stars may be helpful in explaining activity of these stars .
the application is twofold .
firstly , the data may be a source of seismic information on internal rotation .
secondly , the data may allow to assess the role of the feedback effect of oscillation in these stars .
both aims require identification of modes detected in the observed oscillation spectra .
the clue to identification which may be provided by the theory is the determination of a set of modes that may be excited and detected .
effects of rotation , which are important in be star oscillations , must be taken into account .
this is most easily done with the traditional approximation .
validity of this approximation in application to slow modes in be stars , we discussed in section 2 .
we applied our non - adiabatic version of this approximation to a model of hd 163868 , the first be star with the rich oscillation spectrum .
the spectrum , determined from the _ most _ photometric satellite data , and its preliminary partial interpretation , has been recently published by walker et al.(2005 ) .
we showed that there is a large number of unstable modes in the model of this star , covering the whole frequency range of the peaks in the hd 163868 spectrum and beyond .
the instability extends to modes of high horizontal degrees , corresponding to @xmath179 in the case of no rotation .
we found no significant difference in driving between prograde and retrograde modes .
our calculations of the relative visibility of the unstable modes show that the three consecutive groups of the peaks may be explained by low - degree high - order g modes seen from the near equatorial direction .
specifically , the group with frequencies below 0.5c / d may be explained by retrograde @xmath108 modes , the group around 1.7c / d by retrograde sectorial and zonal @xmath109 modes the group around 3.5c / d by prograde sectorial @xmath109 modes .
only the interpretation of the highest frequency group is the same as that proposed by walker et al .
they interpret the intermediate group in terms of @xmath108 prograde mode , which we find stable .
in contrast , the modes that we associate with this group are found stable in their calculations .
also they found all retrograde @xmath108 modes to be stable and therefore did not find a satisfactory interpretation for the lowest frequency group , where peaks have the highest amplitudes .
they proposed that some of the peaks in this group may be interpreted in terms of retrograde r modes , which are found unstable , both by them and us , but their interpretation leaves unexplained highest peaks and there is also a problem with the visibility of such modes from the nearly equatorial direction .
a possible cause of the difference in interpretation of the same data is the use of very different approximations in the treatment of the difficult problem of slow oscillations in rotating stars .
we thank the referee , rich townsend , for his comments which have significantly helped us to improve the manuscript , and hideyuki saio for a conversation that also had an impact on the final version of this paper . this investigation has been supported by the polish mnii
grant no . 1 p03d 021 28 .
bildstein l. , ushomirsky g. , cutler , c. , 1996 , apj , 460 , 827 christensen - dalsgaard , j. , dziembowski , w .
a. , 2000 , in variable stars as essential astrophysical tools , ed . c. ibanoglu ( nato asi ser .
c , 544 ; dordrecht : kluwer ) , 1 dziembowski w. a. , 1977 , acta astr .
, 27 , 203 dziembowski w. a. , kosovichev a. , 1987 , acta astr . , 37 , 313 dziembowski w. a. , moskalik p. , pamyatnykh a. a. , 1993 , mnras , 265 , 588 lee u. , 2001 , apj , 557 , 311 lee u. , 2006 , mnras , 365 , 677 lee u. , saio h. , 1989 , mnras , 237 , 875 lee u. , saio h. , 1997 , apj , 491 , 839 osaki y. , 1986 , pasp , 98 , 30 pamyatnykh a. a. , 1999 , acta astron . , 49 , 119 penrod g. d. , 1986 , pasp , 98 , 35 savonije g. j. , 2005 , a&a , 443 , 557 townsend r. h. d. , 2003a , mnras , 340 , 1020 townsend r. h. d. , 2003b , mnras , 343 , 125 townsend r. h. d. , 2005a , mnras , 360 , 465 townsend r. h. d. , 2005b , mnras , 364 , 573 walker g. a. h. , kuschnig r. , matthews j. m. , et al .
, 2005 , apj , 635 , l77 | we use the traditional approximation to describe oscillations with frequencies comparable to the angular rotation rate .
validity of this approximation in application to main - sequence b stars is discussed .
numerical results regarding mode stability and visibility are presented for a model of the be star hd 163868 . for this object ,
walker et al.(2005 ) detected a record number of mode frequencies using data from the small space telescope _
most_. our interpretation of these data differs from that of walker et al .
in particular , we interpret peaks in the lowest frequency range as retrograde g modes .
we find instability in a large number of modes that remain undetectable because of unfavourable aspect and/or effect of cancellation .
there is no clear preference to excitation of prograde modes .
[ firstpage ] stars : instabilities - stars : early - type - stars : emission - line , be - stars : oscillation - stars : rotation . |
the calculus of variations can be used in the formulation of most of the physical problems . its origin was the famous problem of the _ brachistochrone _ , the curve of shortest descent time .
johann bernoulli found that the curve solution to this problem is a cycloid , that is , the curve described by a point on the circumference of a circle that rolls along a straight line .
the calculus of variations was also present in the solution of other classical problems , such as the catenary and the isoperimetric problem .
its development was parallel to that of mechanics , geometrical optics , and elasticity . in this article
we show how these problems can be related with the plateau problem , that is , the problem of finding the form that a soap film adopts for certain contour restrictions.@xmath2 we also show how to design simple experiments with soap films to find the solution to other variational problems such as a generalized steiner problem or the problem of the quickest approach .
this article was inspired by reading the book demonstrating science with soap films " by d. lovett , in which a similar argument is used to generate the catenary .
the paper can be useful for students and teachers of physics undergraduate courses .
a complete study of this subject requires the use of the euler - lagrange equations .
however , given the intuitive character of the analogies between mechanics , optics , and elasticity problems that we establish , together with the easy visualization of the solutions with soap films , this article may also be useful for students without knowledge of differential equations . in the appendix we propose a problem to use the concepts that appear in the article in a new situation .
along @xmath3 is @xmath4 . ]
[ figure1 ] johann bernoulli posed the problem of the brachistochrone as follows ( see figure 1 ) : _ let two points @xmath5 and @xmath6 be given in a vertical plane . to find the curve that a point @xmath7 , moving on a path @xmath8 , must follow that , starting from @xmath5 , it reaches @xmath6 in the shortest time under its own gravity .
_ denote by @xmath3 a curve joining @xmath5 and @xmath6 , and let us approximate this curve by a polygonal @xmath9 , such that the distance @xmath10 is very small , thus we can assume that the velocity along @xmath11 is constant and can be taken as the velocity of a heavy body after descending an altitude @xmath12 , i.e. @xmath13 , where @xmath14 is the gravity acceleration and @xmath12 is the ordinate of the midpoint between @xmath15 and @xmath16 ( see figure 1 ) .
then the total time of descent along the polygonal is given by : @xmath17 the total time of descent along @xmath3 , @xmath18 , is then obtained from this approximation , taking the limit as the length of the segments of the polygonal goes to @xmath19 , thus we obtain @xmath20 the problem is to find the curve @xmath3 that minimizes the above expression .
the solution to this problem is given by solving a differential equation , known as euler - lagrange equation , whose solution is a cycloid .
we are going to study a plateau problem that , from a mathematical point of view , is equivalent to the brachistochrone problem and , therefore , its solution is also the cycloid . that aproximates the soap film bounded by the surfaces @xmath21 , @xmath22 , and two vertical segments .
its orthogonal projection is the cycloid @xmath3 . ]
[ figure2 ] consider two surfaces , one is the horizontal plane @xmath21 , and the other is defined by the equation @xmath22 ( see figure 2 ) .
consider also two vertical segments @xmath23 and @xmath24 connecting the two surfaces .
we propose the problem of finding the surface of minimal area , bounded by these two segments and the two surfaces .
a surface with such characteristics is known in mathematics as a minimal surface . from the point of view of physics
, the surface corresponds to a soap film .
this follows from the fact that the soap film must minimize the elastic energy , what is equivalent to minimize the area of the surface .
we will approximate this soap film by a vertical surface @xmath25 , that is , a surface that is projected orthogonally on a curve @xmath3 connecting points @xmath26 and @xmath27 in the plane @xmath28 . as we shall see
, this approximation fits very well the experimentally observed soap film .
the area of the surface @xmath25 for a given curve @xmath3 is : @xmath29 where the element of area @xmath30 is the product of the length element @xmath31 of the curve @xmath3 by @xmath32 ( see figure 2 ) .
we observe that eqs .
( [ brachis ] ) and ( [ plateau ] ) are identical except for a constant factor .
this means that the curve that minimizes the area @xmath33 , is the same curve that minimizes the time of descent @xmath18 .
therefore , the solution curve is the same as for the brachistochrone problem , that is , a cycloid .
figure 2 shows a representation of the surface @xmath25 and the cycloid @xmath3 .
the parametric equations of the cycloid in the coordinates of the figure are : @xmath34 where @xmath35 and @xmath36 can be determined by the coordinates of the points @xmath26 and @xmath27 .
we observe from eq .
( [ plateau ] ) that @xmath33 is also the length of the curve @xmath3 with the metric given by @xmath37 , that is , the euclidean metric @xmath38 divided by @xmath39 .
the above discussion shows that the cycloid is the curve along which the distance between two points @xmath26 , @xmath27 for the metric @xmath40 is minimized .
for this reason , @xmath40 is known as the cycloidal metric .
[ figure3 ] the perspex model ( a ) in the photograph of figure 3 corresponds to figure 2 .
it consists of two perspex sheets with the profile of @xmath22 , which are the sides of the model ; they are glued to a perspex square base of 21 cm of side , and , along the curved profiles , to a flexible transparent plastic sheet .
we have used perspex sheets of 2 mm of thickness and the profiles have been made by laser cutting in a plastics workshop .
the two transversal pins can be toothpicks or plastic wires .
when our model is dipped into a soap solution , a soap film is formed between the two pins . incidentally ,
if soap films between the pins and the lateral sides are also established , they should be removed by piercing with a stick .
we have drawn on the base of our model , the cycloid that rolls over the @xmath41-axis and passing through the base points of the pins .
we can see in the photograph of figure 4 , that the horizontal section at @xmath21 of the soap film fits very well to the drawn arc of cycloid .
, @xmath22 , and the vertical pins over points @xmath42 and @xmath43 in cm .
its intersection with @xmath21 fits very well the arc of cycloid passing through these two points and with base line the @xmath41-axis .
this cycloid is generated by a circle of radius @xmath44 cm rolling on the @xmath41-axis . ]
[ figure4 ] in our experiment , to minimize the gravity effects , we have assumed that the perspex base of our model has been placed horizontally .
it can be instructive to see how the soap film is deformed when the base of the model is placed vertically .
the classical steiner problem consists of finding a point @xmath45 in the plane such that minimizes the sum of its distance to three given points @xmath46 , that is , @xmath47 is minimized , where @xmath48 is the euclidean distance between @xmath45 and @xmath49 , @xmath50 .
we assume that each angle of the triangle @xmath51 is less than @xmath52 . this problem , that has many industrial and commercial applications , has an experimental solution using soap films . to see this , consider the soap film limited by two parallel plates placed horizontally and three vertical pins .
the soap film is composed of three vertical planes , which meet in a vertical edge over a point @xmath45 at angles of @xmath52 to each other , according to the second plateau law .
because the area minimization property of this soap film is equivalent to length minimization property of the base curve , this soap film gives the solution to the steiner problem .
consider now the steiner problem for the cycloidal metric , so that we have to minimize the sum of the length of three cycloidal arcs that meet at certain point @xmath45 .
the solution to this problem can be obtained experimentally by using soap films .
for that , we only have to add to the perspex model described in section [ our model ] a third vertical pin .
then , the base curve of the soap film that sets for this model , is the solution to this steiner problem . to each other . ]
[ figure5 ] figure 5 shows this experimental solution .
observe that the film components meet at angles of @xmath52 as in the classical steiner problem .
the solution of jakob bernoulli to the brachistochrone problem was published , with the other received solutions , in 1697 in acta eruditorum under the title solution of problems of my brother together with the proposition of other in turn " . in this article he proposed two new variational problems .
the first is known as the problem of quickest approach , and the second is related with the isoperimetric problem .
jakob enunciated the first problem in these terms ( see figure 6 ) : _ on which of the infinitely many cycloids passing through @xmath53 , with the same base @xmath54 , can a heavy body fall from @xmath53 to a given vertical line @xmath55 in the shortest time ? _ and with the same base @xmath54 .
the cycloid along which a heavy body falls from @xmath53 to @xmath55 in the shortest time is precisely the arc @xmath56 that meets @xmath55 orthogonally .
the transversal curves are the synchronies , that is , the curves of the simultaneous positions of heavy particles which are released at @xmath19 along the cycloids at the same time . ]
[ figure6 ] his brother johann solved this problem quickly , showing that the cycloid of quickest approach is the cycloid that meets orthogonally with the vertical straight line @xmath55 ( see figure 6 ) .
he also solved the problem for the general case in which @xmath55 is not a straight line .
the solution is , also in this case , the cycloid orthogonal to @xmath55 . in cm .
the soap film meets the vertical surface orthogonally ( 1st plateau law ) , and its horizontal section at @xmath21 is an arc of cycloid with @xmath57 cm . ]
[ figure7 ] in figure 7 we can see the solution to this problem using soap films . we have used the model described in section [ our model ] consisting of two perspex plates corresponding to the surfaces @xmath21 and @xmath22 , and we have consider again , as a good approximation , that soap films between vertical pins or planes are vertical surfaces . translated to this model , the problem is to find the surface of minimal area that connects a given vertical pin with a given vertical surface .
the soap film solution corresponds approximately to a vertical surface with base a cycloidal arc that meets the given vertical surface at an angle of @xmath58 .
the configuration for other angle is not stable , as expressed by plateau first law .
the required arc of cycloid ( see figures 6 and 7 ) has a generating circle whose perimeter is twice the distance @xmath54 . note that @xmath59 , @xmath60 .
of minimal area bounded by @xmath21 , @xmath61 , and two vertical pins .
its horizontal section at @xmath21 is a catenary @xmath3 . ] [ figure8 ] in 1691 jakob bernoulli proposed as a challenge to find the curve that assumes a freely hanging chain when held down by its ends .
his brother johann , christiaan huygens , and gottfried leibniz found the curve solution , since then known as the catenary . as a variational problem
, the catenary has to minimize the gravitational potential of the chain , that is : @xmath62 where @xmath63 is the potential energy in an uniform gravitational field @xmath14 of an infinitesimal element @xmath64 with mass @xmath65 and density @xmath66 , which is situated at altitude @xmath67 with respect to the point of zero potential energy level of reference . the explicit equation for a chain of length @xmath68 , two suspension points at the same altitude and with separation @xmath69 , can be written , for an adequate choice of the axis , as : @xmath70 where @xmath71 depends on @xmath69 and @xmath68 .
on the other hand , the area of the vertical surface @xmath25 , raised over a curve @xmath3 ( see figure 8) is given by @xmath72 since eqs .
( [ potential ] ) and ( [ area ] ) only differ in a constant factor , the curve @xmath3 that minimizes @xmath73 is the same that minimizes @xmath33 , namely the catenary . , @xmath74 , and the vertical pins over the points @xmath75 and @xmath76 in cm . its intersection with @xmath21 fits very well the catenary @xmath77 with @xmath78 cm .
this is the stable solution ( minimum energy ) .
we have represented also the unstable solution ( maximum energy ) corresponding to @xmath79 cm . ]
[ figure9 ] if in our perspex model of section 2 we replace the top curved surface @xmath80 by the plane @xmath74 with slope @xmath81 ( see figure 3 ( b ) ) , the soap film formed between two vertical pins is approximately a vertical surface whose base curve is a catenary ( see figures 8 and 9 ) .
there are two solutions corresponding to a minimum and a maximum of elastic energy , so the first solution is stable while the second is unstable . in our case , for the vertical pins over the points @xmath75 and @xmath76 in cm , the two solutions calculated from eq .
[ catenary ] correspond to @xmath82 cm and @xmath79 cm .
the photographs of figure 9 show that the experimental soap film fits very well the catenary drawn on the bottom plate according to the above equation for the stable case .
david lovett also studied the catenary with the help of soap films by showing that the soap film formed between two pins across two non - parallel plates forming a wedge is part of a catenoidal surface.@xmath83 after demonstrating that the cycloid is the curve that minimizes @xmath84 , johann bernoulli showed that the curve @xmath3 that minimizes @xmath85 is an arc of circle .
this means that , considering the poincar half - plane , that is , the half - plane @xmath86 with the metric @xmath87 , the curve of minimal longitude joining two points @xmath26 and @xmath27 of @xmath88 that are not in a vertical line , is the only semicircle passing through @xmath26 and @xmath27 and with center a point @xmath89 on the axis @xmath90 ( see figure 10 ) . of minimal area bounded by @xmath21 , @xmath91 , and two vertical pins .
its horizontal section @xmath3 is an arc of circle with center in the @xmath41-axis . ]
[ figure10 ] again , we can get the above solution using soap films .
for that , we have to replace the top surface of our perspex model with a surface of equation @xmath91 ( see figure 3(c ) ) , so that the soap film that is formed between two vertical pins is approximately a vertical surface with base curve a semicircle ( see figures 10 and 11 ) .
this follows from the fact that the area of a vertical surface over @xmath3 is given by : @xmath92 and therefore minimizing the area of @xmath25 is the same as minimizing the length of @xmath3 in the metric @xmath93 . , @xmath94 , and two vertical pins over the points @xmath95 and @xmath96 in cm .
its intersection with @xmath21 fits very well an arc of circle centered in the @xmath41-axis with radius @xmath97 cm . ] [ figure11 ] we can see in figure 11 that the experimental soap film fits very well the semicircle drawn in the bottom plate .
the soap film in the poincar half - plane model is specially interesting for visualizing geometric properties of the cosmological model of the universe associated to a matter density less than the critical , because in this case the space has negative constant curvature so that the poincar half - plane is a bidimensional model for this space . here , as in the case of the cycloid
, we can consider the soap film solution in terms of the fermat principle for geometric optics , and consider the arcs of catenary and the arcs of circle as the path of the light rays in a medium with refractive index proportional to @xmath67 and @xmath98 respectively .
we can also study soap films formed by three or more vertical pins , as experimental solutions of the associated steiner problems for the metrics @xmath99 and @xmath93 .
the last case is of special interest because it gives an experimental solution to the steiner problem in the poincar half - plane space .
finally , it is interesting to observe that the cycloids and the arcs of circle can also be considered catenaries for gravitational potentials proportional to @xmath32 and @xmath98 , respectively .
we have seen how to design experiments with soap films formed between two perspex surfaces to study the solution to several variational problems that appear in mechanics , optics , elasticity , and geometry .
we have studied the brachistochrone problem , the steiner problem , the problem of quickest approach , the catenary problem and the poincar half - plane problem .
all are very important from a pedagogical point of view . in our perspex models
we have taken one of the surfaces as a plane ( @xmath21 ) and for the other have considered three possibilities : @xmath80 , @xmath100 , and @xmath94 , such that the soap films formed give the solution to the brachistochrone problem , the catenary problem , and the poincar half - plane problem , respectively . to prove this
we have seen that the problem of minimizing certain mechanical magnitudes is equivalent to minimizing the area of a surface connecting two pins between two appropriate surfaces .
the above problems can also be considered as the optical problems of finding the path that follows a light ray , moving in a medium whose refractive index is proportional to @xmath101 , @xmath67 , or @xmath102 , respectively . in the context of geometry these problems consist in to calculate the geodesics for the metrics @xmath103 , @xmath99 , or @xmath104 , respectively , where @xmath105 is the euclidean metric .
finally , the curves solution to the above problems ( that is , arcs of cycloid , catenary and circle ) , can all be considered catenaries by taking gravitational potentials proportional to @xmath32 , @xmath67 , and @xmath98 , respectively .
these considerations illustrate in a simple way , the beautiful connection between dynamics , statics , optics , elasticity , and geometry .
moreover , we have pointed out the historical origin of these problems , with the challenges of the brothers jakob and johann bernoulli .
this work was partially supported by the ministerio de educacion y ciencia grants tec2007 - 60996/mic ( c. criado ) and mtm2007 - 60016 ( n. alamo ) , and by the junta de andalucia grant fqm-213 ( n. alamo ) .
consider the following two surfaces : one is the horizontal plane @xmath21 , and the other is defined by @xmath106 , @xmath107 .
consider also two vertical segments connecting these surfaces . *
prove that the vertical surface of minimal area bounded by these two segments and the two surfaces intersects the plane @xmath21 in an arc of parabola .
for which positions of the segments there is not any solution ?
( hint : the euler - lagrange equation gives two solutions but only one corresponds to a minimum . ) * find the refraction index for which the light rays in @xmath108 describe parabolas . *
find the gravitational potential for which the catenary is an arc of parabola .
* what is the metric in @xmath88 for which the parabolas are the geodesics ?
johann bernoulli challenged the mathematical world with the brachistochrone problem , published in acta eruditorum in june , 1696 , see d. j. struick ( ed ) , _ a source book in mathematics _ , ( princeton university press , princeton , new jersey , 1986 ) , pp .
391 - 392 . to obtain this solution
, he used the fermat principle and considered that the curve of shortest descent time must be the same that the curve described by a light ray whose velocity is proportional to the square root of the altitude , that is , the velocity that a heavy body acquires in falling .
this solution was published in acta eruditorum in may , 1697 , pp .
206 - 211 . in the same issue of that journal , appeared also the solutions to the brachistochrone problem given by his brother jakob bernoulli , leibniz , lhpital , tschirnhaus , and newton .
all except lhpital find the cycloid as the solution .
the method used by jakob bernoulli contained the roots of the calculus of variations ( see reference in note 1 ) . a sufficient condition for a curve @xmath3 to make the integral @xmath109 extremal is that it satisfies the second order differential equation @xmath110 , known as the euler - lagrange equation ( see h. goldstein , _ classical mechanics _ ( addison - wesley , reading , massachusetts , 1980 ) , p.45 . for the case in which @xmath111 does not depends on @xmath41 , the euler - lagrange equation reduces to @xmath112 . in our case , @xmath113 and
then we get the differential equation @xmath114 , where @xmath115 is a constant .
it is easy to see that its solution is the cycloid . according to the first plateau law ( see , for example , ref .
8 - 9 ) , the film have to meet both surfaces @xmath21 and @xmath116 at an angle of @xmath58 .
therefore , the approximation of taking a vertical surface @xmath25 is better as @xmath67 is larger , because then the slope of the tangent to @xmath117 is smaller . a rough estimation of the relative error when we take a vertical element of surface @xmath118 instead of the element of surface for the soap film can be calculated by @xmath119 where @xmath120 is the element of surface for a cylindrical surface that meets orthogonaly the two plates , @xmath121 is the line element along the curve @xmath3 , and @xmath122 . neglecting the terms of order greater than 4 in the expansion of @xmath123 as function of the derivative @xmath124
, we get @xmath125 which for @xmath117 leads to @xmath126 .
therefore , for @xmath127 we have an error less than @xmath128 .
johann bernoulli could easily solve this problem because of his early study on the synchronies , which are the orthogonal trajectories to the cycloids passing through @xmath53 ( see figure 6 ) . from a mechanical point of view ,
these curves are formed by simultaneous positions of heavy particles which are released at @xmath53 along the cycloids at the same time . on the other hand , from an optical point of view
, the synchronies correspond to the simultaneous positions of the light pulses emitted from @xmath53 at the same instant and that propagate along a medium with refractive index proportional to @xmath101 .
therefore the synchronies are the wavefronts .
from the fact that wavefronts and light rays are perpendicular , johann bernoulli concluded that the synchronies are the orthogonal trajectories to the cycloids .
thus the point of quickest approach @xmath130 must be the point where one of the synchronies is tangent to the line @xmath55 ( see figure 6 ) . for more details on this story see s. b. engelsman , _ families of curves and the origins of partial differentiation _
( north - holland , amsterdam , 1984 ) , pp .
31 - 37 .
the parameter @xmath71 is determined by @xmath131 .
the suspension points are symmetric with respect to the @xmath67-axis and their distance to the @xmath41-axis is @xmath132 . for the proof of these facts
see j. l. troutman , _ variational calculus and optimal control _ ( springer verlag , new york , 1996 ) , pp .
78 - 80 . when the ratio between pin separation and distance from the pins to the @xmath41-axis , @xmath133 , is greater than @xmath134 there is not any parameter @xmath71 verifying @xmath135 ( see note 16 ) , while for @xmath133 less than @xmath134 there are two values of parameter @xmath71 .
the catenoid is the surface of revolution generated by the catenary , and , as euler showed in 1744 , is the only minimal surface of revolution ( see ref .
the reason is that the area of a surface of revolution generated by a curve @xmath3 of equation @xmath136 is given by @xmath137 , and again coincides , except for a constant factor , with eqs .
( [ potential ] ) and ( [ area ] ) .
the euler - lagrange equation ( see note 9 ) , reduces in this case to @xmath138 , and its integration gives the catenary of equation @xmath139 .
the poincar half - plane is one of the simplest models of non - euclidean spaces .
the geodesics of this space are the semicircles with center a point @xmath89 on the axis @xmath90 together with the vertical straight lines .
if we think of these geodesics as the straight lines " of this space , the associated geometry does not verify the @xmath143 euclidean postulate ( the parallel postulate ) , because given a straight line " and a point not on that line , there are infinitelly many lines through the point that do not cut ( i.e. parallels to ) the given straight line .
therefore , only a small step , considering semicircles as straight lines " separated johann bernoulli to discover non euclidean geometry 130 years before bolyay and lobatchesky did . | we show a method to solve the problem of the brachistochrone as well as other variational problems with the help of the soap films that are formed between two suitable surfaces .
we also show the interesting connection between some variational problems of dynamics , statics , optics , and elasticity . c. criado@xmath0 .5 cm n. alamo@xmath1 |
starburst galaxies are characterized by their strong and narrow emission lines in the optical spectra , due to a strong star formation in the galaxy center .
seyfert galaxies have broader emission lines and are interpreted as being produced by photoionization by a power - law continuum associated to a supermassive black hole .
ngc 7582 has a typical h ii region spectrum in the optical ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ( veron et al . 1981 ) ) but its x - ray emission is typical of a seyfert 1 galaxy ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ( ward et al . 1978 ) ) . in this work
our goal is to analyze the stellar population located in the 3.5 x 5 central region of ngc 7582 .
our observations were taken with the gmos - ifu ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ( allington - smith et al .
2002 ) ) on the gemini south telescope .
we used the b600-g5323 grating , producing a spectral range of 4230 - 7070 @xmath1 with r = 2400 ( as measured from the 5577 @xmath1 sky line ) .
the final datacubes are corrected from the atmospheric differential refraction ; we also performed a richardson - lucy deconvolution with 6 iterations on the image in each spectral pixel .
the stellar population synthesis was made using the starlight code ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ( cid fernandes et al . 2005 ) ) on the spectrum of each spatial pixel , with a sampling of 0.05 x 0.05 .
maps of different population were constructed , revealing age and metallicity structures .
these results will be published elsewhere .
figure [ fig1 ] shows the cumulative flux fraction as a function of stellar age .
it is clear that about 90% of the light in the field of view is emitted by stars that are less than 1 billion years old .
there seems that a more or less continuum star formation episodes have occurred in this period .
nearly half of the light is emitted by stars less than 10 million years old .
a strong event occurred @xmath0 6 million years ago .
this young stellar population shows nearly solar metallicity and is concentrated in three clumps that might be associated with the ionizing sources of the three observed h ii regions .
one of the clumps , located nearly 1.4 west and 1.0 north of the agn presents a wolf - rayet feature , shown in figure [ fig2 ] .
this feature has broad c iii / n iii + he ii emission between 4600 @xmath1 and 4700 @xmath1 .
+ + i would like to thank fapesp for the financial support .
allington - smith , j. , et al .
2002 , pasp , 114 , 892 cid fernandes , r. , mateus , a. , sodr , l. , stasiska , g. , & gomes , j. m. 2005 , mnras , 358 , 363 ward , m. j. , wilson , a. s. , penston , m. v. , elvis , m. , maccacaro , t. , & tritton , k. p. 1978
, apj , 223 , 788 | ngc 7582 is defined as a starburst / agn galaxy , since its optical and x - ray spectra reveal both characteristics . in this work ,
we show the results of a stellar population modeling in a datacube taken with the gemini south telescope .
we found that @xmath0 90% of the light in the field of view is emitted by stars that are less than 1 billion years old .
a strong burst occurred about @xmath0 6 million years ago and has nearly solar metallicity .
we also found a wolf - rayet cluster . |
in recent years , the discussion of relativistic issues in the reactions involving the deuteron has become more and more important .
some well known examples for a clear experimental evidence have been found in the electromagnetic disintegration @xcite .
this results are supported by a recent analysis of polarization observables of the deuteron in a relativistic framework @xcite . also , the persistent difference between theory and experiment in polarization observables of proton deuteron scattering have been argued to be of relativistic origin @xcite .
it is expected that the interpretation of new experimental results , e.g. , from tjnaf , also calls for a relativistic treatment .
meanwhile the number of theoretical investigations devoted to this question and to tackle the relativistic aspects of few body systems is increasing .
one may basically recognize two lines : on one hand the majority of approaches uses a nonrelativistic scheme of calculation and take into account some leading order corrections , such as nonnucleonic degrees of freedom like mesonic exchange currents , @xmath1-configurations , @xmath2-states and more .
for a review see , e.g. , refs .
@xcite which also contain an large number of references .
note , however , that the problem of consistency arises as has been recently pointed out again , e.g. , by ref .
@xcite in the context of electrodisintegration .
the second line recognizes the importance of covariance @xcite .
we follow the bethe - salpeter approach which is covariant by construction and properly defined @xcite . based on the field theory of particles
this approach is also consistent and automatically takes into account all relativistic effects in the two nucleon system . because of the covariance of the bethe - salpeter equation the amplitudes ( or vertex functions ) may be written in any reference system , and one may chose a convenient one .
however , practical applications are hampered by the difficulty of finding solutions of the bethe - salpeter equation , even in ladder approximation .
nevertheless , the activity on solving the bethe - salpeter equation has increased in recent years @xcite , and by now parameterizations of the bethe - salpeter amplitudes in euclidean space are available @xcite .
also , useful results have been achieved by a reduction of the bethe - salpeter equation to a three dimensional one . here , we examine the nonrelativistic reduction of the bethe - salpeter approach to compare with the results achieved by nonrelativistic calculations .
since the problem has not been solved so far for the general case , we restrict ourself to a special process that is the disintegration of the deuteron.to this end we consider the threshold region only that exhibits already the main features to be discussed .
also , this region is dominated by a single transition amplitude to the @xmath3 final state ( see , e.g. , ref .
@xcite ) thus leading to some technical but not substantial simplifications .
the extension to other partial amplitudes is straightforward but tedious , however beyond the scope of our present approach .
indeed , the deuteron disintegration close to threshold energies has been a key reaction to establish subnucleonic degrees of freedom in nuclei @xcite and it attracts continuous attention from theoretical and experimental sides .
it has been an excellent paradigm to examine nonnucleonic degrees of freedom and relativistic effects .
it is well known that the nonrelativistic impulse approximation fails to describe the double differential cross section at momentum transfer squared @xmath4 @xmath5 .
the experimental data do not indicate the deep minimum present in the calculations @xcite . to fill this minimum mesonic exchange currents
have been introduced .
the contributions of @xmath6- , @xmath7-currents , and @xmath8- configurations allow one to get a satisfactory agreement with the data @xcite
. the question of a consistent inclusion of all relativistic corrections ( at least for the pion sector ) has recently been addressed by ref .
@xcite and supports the above statement
. nevertheless , some conceptional problems of the theory still remain open questions . among them
is the problem of gauge invariance and the question of the nucleon form factor to be chosen in the exchange currents contributions @xcite . by now
some relativistic methods identify corrections like meson exchange currents in the relativistic impulse approximation .
it has been shown that in the framework of the light - cone approach the so - called extra components @xmath9 of the deuteron wave function , and @xmath10 of the @xmath3-wave function introduced in ref .
@xcite give an expression in the nonrelativistic limit , that analytically equals to the contribution of the pair current @xcite . in this context
the one iteration assumption that will be explained below for solving the dynamical equation is substantial .
another important result in this context is the calculation of the static electromagnetic properties within the bethe - salpeter approach .
it follows that the contribution of the @xmath0-states to the magnetic moment of the deuteron is numerically close to the contribution of mesonic currents and agrees in sign @xcite .
it has already been noticed earlier in the context of covariant reductions of the bethe - salpeter equation to three dimensional ones that negative energy components in the wave function are responsible for pair current type contributions , see , e.g. , ref .
@xcite . here
we show that the suggested reduction procedure leads to the same _ analytical _ structure as the nonrelativistic pair current correction .
the paper is organized as follows .
section [ sec : nonrel ] contains the general formulas for the deuteron disintegration amplitude in different representations .
the static approximation is also discussed .
the one iteration approximation is formulated in section [ sec : oneit ] where we give reasons for its validity in simple models .
the derivation of the main correction is done in the section [ sec : emc ] .
we present and discuss our results in section [ sec : results ] , summarize and conclude in section [ sec : conclusion ] .
because of lorentz covariance , transformation properties under parity and time reversal as well as current conservation , the general structure of the transition matrix element for the @xmath11 transition current near threshold can be written as @xcite @xmath12 where @xmath13 is the deuteron polarization four vector with projection @xmath14 , @xmath15 is the deuteron four momentum , @xmath16 is the four momentum transfer , @xmath17 is the squared total momentum of the @xmath18-pair , see fig . [
fig : diagram ] .
the scalar function @xmath19 may be splitted into isovector dirac ( @xmath20 ) and pauli ( @xmath21 ) contributions @xmath22 the final state of the @xmath18-pair is supposed to be in a @xmath3-state .
this is supported by nonrelativistic calculations @xcite .
( [ formc ] ) is valid in the any reference system and the functions @xmath23 depend on the scalars @xmath24 and @xmath25 only .
current conservation is fulfilled because the tensor @xmath26 that appears in eq .
( [ formc ] ) is antisymmetric . to calculate the functions @xmath23 from the underlying dynamics in the bethe - salpeter approach , two representations of the bethe - salpeter amplitude
may be used in concrete calculations : 1 ) the covariant representation and 2 ) the partial - wave representation . using the _ covariant representation _ for the initial and final states allows us to directly read off the functions @xmath27 from the proper current matrix elements .
this has been done explicitly in ref .
however , to study the relation to the nonrelativistic expressions it is more suitable to give the formulas for @xmath27 in terms of a _ partial - wave representation _ of the vertex functions .
hence , the expressions for @xmath23 can be written as @xmath28 where the partial vertex functions @xmath29 represent the states @xmath30 , @xmath31 , @xmath32 , @xmath33 of the @xmath18-pair , and @xmath34 denote the states @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 of the deuteron .
we use spectroscopic notation and ( @xmath43 , @xmath44 ) as @xmath7-spin quantum numbers @xcite . in general , the lengthy functions @xmath45 depend on lorentz scalars and the explicit expressions are omitted here .
we will specify them below after having introduced appropriate approximations .
the expressions for @xmath46 and @xmath47 are given in a formally covariant way , viz .
@xmath48 where @xmath49 denotes the lorentz scalar product .
we note that through the use of bethe - salpeter vertex functions the denominators of @xmath50 appearing in eq .
( [ a2 ] ) contain products of @xmath51 and @xmath52 , where @xmath53 , etc . that stem from the nucleon propagators ( see fig . [
fig : diagram ] ) .
we evaluate the integrals in the laboratory system ( deuteron at rest ) . at threshold , because of the small deuteron binding energy , it is possible to utilize the static approximation @xcite that preserves the analytical structure of eq .
( [ a2 ] ) through the following equations @xmath54 here and in the following we use ( @xmath55 ) @xmath56 we illustrate the approximation by looking closer to the expressions involving the @xmath57 states . the full denominator of the integrand in eq .
( [ a2 ] ) leads to a complicated pole structure and reads explicitly @xmath58 using eq .
( [ eqn : static ] ) , then leads to a simple pole expression for the integrand involving the denominator @xmath59 thus , static approximation means in particular that @xmath60 ( no retardation ) and the lorentz boost transformation of the @xmath18-pair vertex functions is neglected . going beyond the static approximation can be achieved by expanding the full expression in terms of @xmath61 and @xmath62 that lead to additive corrections .
the integration on @xmath63 can now be performed by choosing a proper integration contour and specifying the corresponding poles , e.g. , closing the upper half plane leads to poles for @xmath63 at @xmath64 .
the vertex functions are then evaluated at @xmath65 . since in the reaction under consideration @xmath66 , we expand the vertex functions near @xmath67 for @xmath68 and @xmath69 respectively that allows us to derive analytical expressions in the one iteration approximation as will be shown in the next section .
the analogous procedure holds for the other partial vertex functions . with this choice one
the nucleons in the deuteron is taken on - shell .
alternatively , it would have been possible to expand the functions , e.g. , near @xmath70 .
however , the above advocated approach is more suitable , since it leads to an equation that allows for an analytical of @xmath63 . in eq .
( [ a2 ] ) we now perform the @xmath63 integration as explained above .
the angular integration is simplified by taking @xmath71 along the @xmath72 axis and replace @xmath73 given in eq .
( [ qaoth ] ) via @xmath74 where @xmath75 . finally , to compare eq .
( [ a2 ] ) with the nonrelativistic limit we introduce a @xmath76-expansion excluding order @xmath77 ( i.e. @xmath78 ) .
the resulting structure functions @xmath27 are then given by @xmath79 where @xmath80 , and @xmath81 , and @xmath82 is the legendre polynomial .
the functions @xmath10 , @xmath83 , @xmath84 , @xmath85 , @xmath86 disappear in the above expressions after @xmath63 integration and because of the @xmath76 expansion .
note , that within this approximation scheme we are left with the @xmath57 to @xmath57 and @xmath57 to @xmath87 transitions only .
all other matrix elements , such as @xmath88 to all and @xmath87 to @xmath87 cancel , respectively .
we now examine the expressions for @xmath27 given in eqs .
( [ v1nred],[v2nred ] ) more closely . to recover the nonrelativistic result
, we neglect the vertex functions @xmath89 , @xmath90 , @xmath91 , @xmath92 that correspond to the negative @xmath7 spin components ( i.e. do not exist in the nonrelativistic scheme ) . if we replace the functions @xmath68 and @xmath93 by the nonrelativistic @xmath94 and @xmath95 wave functions and @xmath69 by the @xmath3 continuum wave function in the following way @xmath96 and insert these for the respective vertex functions into eqs .
( [ v1nred],[v2nred ] ) we obtain @xmath97 where we have introduced the magnetic isovector form factor @xmath98 . in co - ordinate space the respective integral is achieved using the following transformations , for the deuteron states ( @xmath99 , @xmath100 , @xmath101 ) @xmath102 where @xmath103 is spherical bessel function , and for the scattering state @xmath104 the resulting expression is @xmath105 this result reflects the so - called nonrelativistic impulse approximation and represents the lowest order nonrelativistic expansion of the transition form factors given in eqs .
( [ v1nred],[v2nred ] ) .
the bethe - salpeter equation is commonly solved by iterations . after angular decomposition
we are left with an integral equation for the radial parts of the vertex function @xmath106 .
they are connected to the partial amplitudes @xmath107 which are used here for simplicity on the right hand side of the following equation by amputation @xcite , e.g. , for the @xmath57 components the relation is given by @xmath108 the integral equation read @xmath109 where @xmath110 denotes the type of exchanged meson , @xmath111 its coupling constant and @xmath112 is the transition matrix element between the states @xmath113 and @xmath114 given below . to get fast convergence to the solution of this equation one needs a good educated guess for the initial vertex function @xmath106
. this guess may be taken from the solution of the equivalent nonrelativistic problem .
after several iterations one usually gets the exact solution .
the initial vertex functions may be chosen along the lines given in ref .
@xcite , generalizing eq .
( [ r2n ] ) @xmath115 all other wave functions are assumed to vanish in this order .
note , that the numerators of eqs.([eqn : psi1],[eqn : psi3 ] ) are the nonrelativistic vertex functions .
the ansatz given in eqs .
( [ eqn : psi1 ] ) and ( [ eqn : psi3 ] ) has been numerically verified for the deuteron consisting of two spinor nucleons that interact via the exchange of scalar mesons with mass @xmath116 .
this interaction leads to a bound state for both , nonrelativistic and relativistic cases . in the _ nonrelativistic _
case and use of the schrdinger equation the kernel is of the form @xmath117 where the coupling constant @xmath118 leads to a binding energy of @xmath119 mev .
the resulting wave function of the deuteron may be parameterized in the following form , @xmath120 with @xmath121 and a normalization constant @xmath122 .
note , that for the scalar exchange model the @xmath95-wave function is zero , @xmath123 .
these functions are then substituted as an educated guess into the system of equation ( [ bseq1 ] ) using the assumption of eqs .
( [ eqn : psi1],[eqn : psi3 ] ) .
( [ bseq1 ] ) is then solved in euclidean space ( @xmath124 ) using the ansatz amplitude eq .
( [ wfnr ] ) and the full kernel @xmath125 .
@xmath126 to achieve the same deuteron binding energy as before , now using the bethe - salpeter equation ( [ bseq1 ] ) the coupling constant @xmath127 of the scalar interaction needs to be readjusted to get the same deuteron binding energy , viz .
@xmath128 . in this framework
we compare the full vertex function of the exact solution of eq .
( [ bseq1 ] ) to the result achieved by one iteration only but using the nonrelativistic solution as an educated ansatz . this will be called _ one iteration approximation _ in the following . as a demonstration for the quality of the approximation
we show two vertex functions , for the @xmath35- and @xmath129-channels in fig .
[ fig : compare ] .
the upper two surfaces displayed correspond to the exact result and the one iteration approximation of the @xmath35-component .
the difference between the two is rather small .
in fact , in the relevant region @xmath130 gev and @xmath131 gev , where the amplitude has its major contribution to the normalization and to physical processes , the difference does not exceed 5% .
the exact and the one iteration approximation for the @xmath129-channel that is relevant for the subsequent discussion of the electromagnetic properties are displayed as the lower two surfaces .
we conclude that for the scalar model the shape of the exact solutions are already reproduced by one iteration . from
that we argue that the one iteration approximation is a reasonable first step that may be utilized to solve the more complicated problem with realistic potentials .
we now turn back to the discussion of the realistic deuteron implementing the full kernel with all spin dependences . by using eqs .
( [ bseq1 ] ) along with eqs .
( [ eqn : psi1],[eqn : psi3 ] ) the expressions for the @xmath132-channel , i.e. for the functions @xmath133 are now deduced analytically .
if we consider the @xmath6-exchange kernel only and use the above given notation for the states ( @xmath134 ) then @xmath135 with legendre functions of the second kind @xmath136 , and @xmath137 .
the @xmath63 integration can be performed yielding ( e.g. for the pionic exchange kernel ) @xmath138 the functions @xmath139 are obtained from @xmath140 by treating @xmath46 and @xmath63 as explained in sect .
[ sec : nonrel ] .
the explicit expressions for @xmath141 ( @xmath142 ) are given by : @xmath143 with @xmath144 .
to simplify the notation the isospin factors have been omitted in @xmath141 and eq .
( [ bseq2 ] ) .
the extension to other mesons of a one boson exchange kernel is straight forward . now using eqs .
( [ bseq2],[pot1 ] ) , the transformation eq .
( [ nr2k ] ) and the integral representations of the legendre functions @xmath145 we obtain following results for the vertex functions : @xmath146 where the isospin factor @xmath147 has now been shown explicitly .
this is the analytical structure given after the first iteration .
the inhomogeneous bethe - salpeter equation for the amplitudes in the @xmath3-channel reads @xmath148 here @xmath149 denotes the plane - wave function , @xmath150 where @xmath151 is the on - energy - shell momentum given by @xmath152 .
for the subsequent discussion it is more convenient to split the system of equations into the following form :
@xmath153 since the full vertex functions are related to the full amplitudes by amputation , the partial vertex functions @xmath154 are connected to the partial amplitudes @xmath155 via simple relations @xcite . as in the case of the deuteron channel
we consider the one iteration approximation .
to do so we chose a similar expression for the initial amplitudes as before to connect to the nonrelativistic solution , @xmath156 here @xmath157 is the nonrelativistic continuum wave function in @xmath3 channel given by @xmath158 and @xmath159 is nonrelativistic half - off - shell @xmath160-matrix for the @xmath3 channel normalized through the condition @xmath161 where @xmath162 is the phase shift , and @xmath163 .
analogously to the deuteron case , we finally arrive at the first order corrections to the amplitudes in the @xmath3 channel , @xmath164 where @xmath165 is the isospin factor .
we have shown in this section that choosing proper zero approximation wave function ( i.e. the nonrelativistic ones ) after one interaction additional partial amplitudes arise through the bethe - salpeter equation .
they are connected to the interaction kernel and in electromagnetic processes give rise to the so called pair current correction as will be shown in the next section .
we are now in the position to turn to the first order corrections to @xmath166 given in eq .
( [ vnrn ] ) . to this end
we expand the expressions given in eqs .
( [ eqn : v12 ] ) , ( [ v1nred ] ) and ( [ v2nred ] ) into a power series of @xmath167 to extract the pionic contribution only . also we consider the @xmath0-states contribution only , i.e. components with one negative @xmath7 spin . the resulting expression for the transition form factor @xmath19 will be denoted by @xmath168 .
substituting eqs .
( [ find ] ) and ( [ find1 ] ) as well as eqs .
( [ finp ] ) and ( [ finp1 ] ) into eqs .
( [ v1nred ] ) and ( [ v2nred ] ) and using the replacements of eq .
( [ r2n ] ) then eq . (
[ eqn : v12 ] ) reads @xmath169 the @xmath170-integration can be solved analytically . after doing so ,
we obtain @xmath171 here we have introduced the function @xmath172 where @xmath173 is given in eq .
( [ qaoth ] ) .
this first order contribution in @xmath174 supplements the lowest order relativistic expansion as given in eq .
( [ vnrn ] ) of sect .
[ sec : nonrel ] .
we then arrive at the following expression for the transition form factor : @xmath175 this is our main result .
starting from a fully covariant theory this result has been achieved using the lowest order relativistic contribution within the one iteration approximation .
the relation between the nonrelativistic wave functions and the relativistic amplitudes given in eqs .
( [ eqn : psi1 ] ) , ( [ eqn : psi3 ] ) , and ( [ eqn : ph1 ] ) allows us to give the final expressions in terms of nonrelativistic wave functions .
the lowest order in the @xmath176 coupling constant @xmath174 leads to an additional contribution from iterating the @xmath0-wave channel once . comparing this result to the one achieved within the nonrelativistic scheme that introduces mesonic exchange currents , we find find that the first term coincides analytically with the nonrelativistic impulse approximation contribution and the second one with the @xmath6-pair current contribution . to further illustrate the equivalence we connect our results to the nonrelativistic formalism describing the break - up process .
more details are given , e.g in ref .
the differential cross section for the @xmath177 - transition has the form ( nonrelativistic case ) @xmath178 where @xmath179 is the momentum transfer , @xmath180 .
the momentum @xmath151 is related to the relative energy @xmath181 of the np system as given before @xmath182 and the relation between kinematical quantities is given by @xmath183 and @xmath184 , @xmath185 . in the general case ,
the current matrix element is a sum of nonrelativistic impulse approximation contribution , meson exchange currents contributions and retardation currents contributions .
here we are interested in @xmath6-meson pair current part only and thus we may write for @xmath186 the following expression : @xmath187 where we introduced @xmath188 , which reflects the impulse approximation operator and @xmath189 , which is the @xmath6-meson pair ( contact ) operator .
expressions for the matrix elements are given in ref .
@xcite , see eq . ( 11 ) , and may be compared with eq .
( [ vnr : final ] ) .
differences will be discussed below .
we have shown in the previous section that the nonrelativistic reduction of the bethe - salpeter approach utilizing the one iteration approximation leads to results that exhibit the same _ analytical structure _ as the nonrelativistic result plus pair current corrections .
some details differ as will be discussed below .
one may now use the `` exact '' nonrelativistic wave functions to calculate the different contributions to the cross sections .
this has been done here for an illustration .
the formula for the cross section is given in eq .
( [ nr ] ) .
the dominant m1 transition matrix element ( i.e. @xmath190 ) of the multipole decomposition given there directly corresponds to the nonrelativistic reduction of eq .
( [ formc ] ) given above .
the contribution of the nonrelativistic impulse approximation given , e.g. in ref .
@xcite coincides with the formula given in eq .
( [ vnrn ] ) if @xmath191 and @xmath192 are chosen as in eq .
( [ r2n ] ) .
the analytical structure of the nonrelativistic pair current contribution equals to that of the @xmath0-state contribution derived from the bethe - salpeter approach given in eq .
( [ ve : final ] ) .
we note that the nonrelativistic pionic pair current contribution given in ref .
@xcite depends on the nucleon form factor @xmath193 .
sometimes also the electric nucleon sachs form factor @xmath194 has been used ( see ref .
the bethe - salpeter approach yields a different dependence on the nucleon form factor that is also consistent with current conservation and given by the function @xmath195 defined in eq .
( [ gamma ] ) . as an illustration of the different behavior we display the form factors @xmath193 ( solid line ) , @xmath196 ( dotted line ) , and the function @xmath195 ( dashed line ) in fig .
[ fig : ffl ] . as a parameterization for the nucleon form factors we use the one given in ref .
the form factor @xmath197 is larger than the form factor @xmath198 and the function @xmath195 and therefore the respective pionic pair current is expected to be larger than the one using the other form factors .
since the function @xmath195 is in the middle of the other ones the respective contribution of the @xmath0-states in bethe - salpeter approach will be different from the nonrelativistic calculations normally using @xmath193 or @xmath196 . to investigate the influence of the nucleon form factors more closely we calculated the impulse approximation and pionic pair current contributions to the differential cross section .
the calculation is performed with the paris @xmath199 potential @xcite at @xmath200 mev and @xmath201 .
it is well known that some uncertainty is related to the hadron form factors . without dwelling too much on that point
we would like to compare three different form factors to see how this uncertainty influences the result . introducing the hadron form factors change the expressions for the pair current contribution @xcite .
three sets of hadron form factors ( for @xmath176-vertex ) have been utilized and are shown in figs .
[ fig : iapc ] a - c ) , respectively : a monopole vertex @xcite with a cut - off mass of @xmath202 gev ( set a ) and @xmath203 gev ( set b ) .
in addition a vertex inspired by a qcd analysis has been used with two parameters given chosen to be @xmath204 gev and @xmath205 gev @xcite ( set c ) .
it is seen from fig .
[ fig : iapc ] that there is a strong dependence of the differential cross section on the nucleon electromagnetic form factors .
it is also seen that the minimum at @xmath206 @xmath5 in the impulse approximation contribution is shifted by pionic pair current to the region @xmath207 @xmath5 using @xmath198 , to @xmath208 @xmath5 using the function @xmath195 and to @xmath209 @xmath5 using @xmath193 as a form factor . the largest shift in the cross section
is achieved by using @xmath193 in the calculations .
the size of the shift and the behavior of the cross section considerably depends on the set of parameters as well as the type of the @xmath210-vertex used .
this is illustrated in fig.[fig : iapc]d ) for calculations using @xmath193 only but for different parameterizations of the strong vertex .
we investigate the electromagnetic current in the framework of the bethe - salpeter approach .
the question addressed is the nonrelativistic reduction of the current transition amplitude . as a specific application we consider the deuteron disintegration near threshold energies in some detail .
we present a method to reduce the amplitude to the nonrelativistic one .
we summarize the basic steps of this method and the main approximations utilized .
to arrive at our result , i.e. to compare the relativistic expressions to the pair current that appears in the nonrelativistic scheme , two main assumptions have been introduced : 1 ) the static approximation and 2 ) one iteration approximation . in the static approximation boost corrections are neglected .
we emphasize that the latter contributions are not negligible and for the elastic form factors corrections can achieve 20 % .
a discussion of this contributions in a noncovariant but consistent relativistic framework may be found in ref . @xcite .
although important , a discussion of all these contributions in not the main issue of our present paper . in fact , we do not expect any substantial change in our present conclusions , since all corrections are of additive nature and will be taken into account as the scheme develops .
the one iteration approximation is to some extent more subtle .
it plays an essential role in connecting the bethe - salpeter amplitudes and the usual nonrelativistic wave functions .
this approximation allows us to express the final result in terms of the usual @xmath94- and @xmath95-wave functions of the deuteron and the @xmath94-wave function of @xmath18-pair .
although numerically verified for deuteron - like models , the one iteration approximation needs further investigation .
we argue that the validity is due to the small binding energy for the deuteron . from inspection of the next iterations
we find additive corrections only
. therefore these assumptions can be loosen up and additional corrections can be taken into account .
this way we have a possibility to calculate corrections in a more consistent way .
this problem is present in both , electromagnetic and strong processes ( for @xmath211 and @xmath212 reactions , for instance ) .
although the calculation of these corrections is important to arrive at definite conclusions about the covariant approach as a whole , our main issue here is to investigate the correspondence between the covariant approach and the mesonic exchange currents of the nonrelativistic approaches .
we have shown the analytic correspondence of these different approaches .
the pair correction plays an essential role in explaining the experimental data .
an interesting result in this context is connected to the form factor entering in the pair current term .
the overall form factor that is consistent with gauge independence and the relativistic ( covariant ) expansion neither coincides with @xmath213 nor with @xmath214 but requires an `` average '' between them .
the authors wish to thank kaptari l.p .
, titov a.i . and
especially karmanov v.a . for useful and stimulating discussions .
we thank semikh s.s . and suskov s.eh . for assistance in the numerical calculations .
the authors are grateful to each others home institutions for extended hospitality during the respective visits .
part of the work has been done during the research workshop on `` progress in current few body problems '' hosted by the bogoliubov laboratory of theoretical physics and supported by the heisenberg - landau program .
this work is partially supported by a grant of the deutsche forschungsgemeinschaft and a rffi grant of support of leading schools in russia .
a. cambi , b. mosconi , and p. ricci , phys .
lett * 48 * , 462 ( 1982 ) .
m. van der schaar _
et al . _ ,
* 68 * , 776 ( 1992 ) , phys .
lett . * 66 * , 2855 ( 1991 ) .
lykasov , phys . part . nucl . * 24 * , 59 ( 1993 ) ( fiz .
chastits at .
yadra * 24 * , 140 ( 1993 ) ) .
kaptari , a.yu .
umnikov , f.c .
khanna , and b. kmpfer , phys .
b * 351 * , 400 ( 1995 ) .
lykasov , proceedings of the xiii ishepp `` relativistic nuclear physics and quantum chromodynamics '' , dubna , 1996 .
v.v.burov , v.n .
dostovalov and s. .
suskov , sov
* 23 * , 317 ( 1992 ) .
shebeko , v. kotlyar , and yu . melnik , phys .
* 26 * , 79 ( 1995 ) .
j.l . friar , phys .
c * 22 * , 796 ( 1980 ) .
j. adam jr . , proc .
14th int . conf . on few body problems , williamsburg 1994 , edited by f. gross , aip conf .
proc * 334 * , 192 ( 1995 ) .
f. ritz , h. gller , t. wilbois , h. arenhvel , phys .
c * 55 * , 2214 ( 1997 ) .
b.d . keister and w.n .
polyzou , adv .
. phys . * 20 * , 225 ( 1991 ) .
n.k . devine and s.j .
wallace , phys .
c * 48 * , r973 ( 1993 ) .
j. tjon , proceedings of the 14th int .
conf . on few body problems , williamsburg 1994 , edited by f. gross , aip conf .
proc * 334 * , 177 ( 1995 ) .
d.lurie , _ particles and fields _ ( interscience publishers , new york , 1968 ) . c. itzykson , j .- b .
zuber , _ quantum field theory _
( mcgraw - hill , singapore 1985 ) .
n. nakanishi , prog .
( kyoto ) suppl . *
43 * , 1 ( 1969 ) .
umnikov , l.p .
kaptari , k.yu .
kazakov , and f.c .
khanna , phys .
b * 334 * , 163 ( 1994 ) .
t. nieuwenhuis , j.a.tjon , few - body syst .
* 21 * 167 ( 1996 ) .
umnikov a.yu .
, z. phys .
a * 357 * , 333 ( 1997 ) .
h. arenhvel , lecture notes in physics , vol .
426 , eds l. mathelitsch and w. plessas ( 1995 ) p 1 .
s. auffret _
et al . _ ,
lett . * 55 * 1362 ( 1985 ) ; r.g .
et al . _ ,
c * 42 * , r1 ( 1990 ) .
j.a . lock and l.l .
foldy , ann . of phys . *
93 * , 276 ( 1975 ) .
mathiot , nucl . phys . *
a412 * , 201 ( 1984 ) .
singh , w. leidemann , h. arenhvel , z. phys .
a * 331 * , 509 ( 1988 ) .
w. leidemann , and h. arenhvel , z. phys .
a * 326 * , 333 ( 1987 ) .
r. schiavilla , and d.o .
riska , phys .
c * 43 * , 437 ( 1991 ) .
b. desplanques , v.a .
karmanov , and j .- f .
mathiot , nucl . phys . *
a589 * , 697 ( 1995 ) ; j. carbonell , b.desplanques , v.a .
karmanov , and j .- f .
mathiot , phys .
rep . to be published .
kaptari l.p .
, umnikov a.yu .
bondarenko , k.yu .
kazakov , f.c .
khanna , and kmpfer b. , phys .
c * 54 * , 986 ( 1996 ) .
j. adam jr . , a. stadler , m.t .
pena , and f. gross , nucl .
a * 631 * 570c ( 1998 ) ; s. wallace _ ibid . _
s.g . bondarenko , v.v .
burov , m. beyer , and s.m .
dorkin , mpg - vt - ur 87/96 , _ e - print archive : nucl - th/9612047_. j.j .
kubis , phys . rev .
* 6 * , 547 ( 1972 ) .
zuilhof and j.a .
tjon , phys .
c * 22 * , 2369 ( 1980 ) .
burov , a.a .
goy , s.eh .
suskov , and yu.v .
chubov , sov .
* 59 * , 989 ( 1996 ) .
mathiot , nucl.phys .
* a412 * , 201 ( 1984 ) . g. hhler _
et al . _ ,
b114 * , 29 ( 1976 ) .
r. machleidt , k. holinde , and ch .
elster , phys.rep . * 149 * , 1 ( 1987 ) .
m. gari and u. kaulfuss , nucl
a408 * , 507 ( 1983 ) .
burov , v.n .
dostovalov , and s.eh.suskov , czech .
j. of phys .
* 41 * , 1139 ( 1991 ) .
m. lacombe _
et al . _ ,
phys.rev . *
c21 * , 861 ( 1980 ) .
( 20.00,60.00)(1,0)20.00 ( 40.00,60.00)(3,2)20.00 ( 20.00,60.00)(1,0)10 ( 40.00,60.00)(3,2)10 ( 35.00,40.00)(3,2)15.00 ( 35.00,40.00)(3,2)7.5 ( 50.00,50.00)(3,-2)14.00 ( 50.00,50.00)(3,-2)7.5 ( 35.00,38.00)(1,0)28.3 ( 35.00,38.00)(1,0)15.00 ( 20.00,40.00)(1,0)15.00 ( 20.00,38.00)(1,0)15.00 ( 25.00,41.00)(28.00,39.00)(28.00,39.00 ) ( 25.00,37.00)(28.00,39.00)(28.00,39.00 ) ( 35.00,39.00 ) ( 65.00,39.00 ) ( 66.6,40.00)(1,0)8.4 ( 66.6,38.00)(1,0)8.4 ( 75.00,40.00)(1,0)5.00 ( 75.00,38.00)(1,0)5.00 ( 40.00,60.00)(42.00,60.00)(42.00,58.00 ) ( 42.00,58.00)(42.00,56.00)(44.00,56.00 ) ( 44.00,56.00)(46.00,56.00)(46.00,54.00 ) ( 46.00,54.00)(46.00,52.00)(48.00,52.00 ) ( 48.00,52.00)(50.00,52.00)(50.00,50.00 ) ( 28.00,30.00)(0,0)[cc]d : @xmath215 ( 73.00,30.00)(0,0)[cc]np : @xmath216 ( 47.00,57.00)(0,0)[cc]@xmath217 ( 25.00,63.00)(0,0)[cc]@xmath218 ( 55.00,74.00)(0,0)[cc]@xmath219 | we investigate the nonrelativistic reduction of the bethe - salpeter amplitude for the deuteron electrodisintegration near threshold energies . to this end
, two assumptions have been used in the calculations : 1 ) the static approximation and 2 ) the one iteration approximation . within these assumptions
it is possible to recover the nonrelativistic result including a systematic extension to relativistic corrections .
we find that the so - called pair current term can be constructed from the @xmath0-wave contribution of the deuteron bethe - salpeter amplitude .
the form factor that enters into the calculation of the pair current is constrained by the manifestly gauge independent matrix elements .
# 1 * pacs : * 21.45.+v , 25.30.fj , 24.10.jv |
the cosmic rays ( crs ) have been studied for a long time ( see , e.g. , @xcite for a review ) .
there are many models of their ( likely non - uniform ) sources , but explaining the origin of the highest - energy particles ( @xmath1 ev ) observed in crs is still a challenging problem @xcite . in general , two scenarios have been proposed for the particle acceleration in astrophysical environments .
one is the acceleration by electric field in geometries , where this field is not perpendicular to magnetic field - lines , for example , in the vicinity of magnetized rotating neutron stars @xcite .
very hard particle spectra may emerge in this way , but , because of the curvature losses inherent to such geometries , the upper limit to proton energy appears to be below @xmath2 ev .
another class of scenarios assumes gradual , `` stochastic '' ( or diffusive ) acceleration of charged particles through multiple reflections from inhomogeneities of magnetic field in the environments where large velocity gradients are present ( see @xcite for a review ) . according to the generally accepted view ( see , e.g. , @xcite ) this mechanism works equally well for shocks and shear flows , and the emerging spectra of accelerated particles are such , that only a small fraction of the total energy budget is contained in the most energetic particles .
furthermore , there is a serious obstacle in achieving high energies of the accelerated particles their diffusive escape . in the environments , where bulk velocities are sub - relativistic ( @xmath3 )
, it takes many reflections to increase the energy of accelerated particle twofold . in this case
, the mechanism can work only in the largest objects in the universe , like galaxy clusters or radio lobes and knots in active galactic nuclei ( agns ) , and only under optimistic assumption about the diffusion coefficient for particles in magnetic field .
for example , bohm diffusion leads to particle escape on the timescale @xmath4 , where @xmath5 is the accelerator s size and @xmath6 the particle s gyroradius , while the acceleration occurs on the timescale @xmath7 .
the maximum attainable energy is defined by equating these timescales : @xmath8 .
it is of the order of @xmath9 ev in the most favorable cases .
ultrarelativistic shocks and shear flows are more promising , since the parameter @xmath10 .
moreover , the energy gain per cycle , i.e. , at each reflection from the shock or from the shear flow boundary , can approach the factor of @xmath11 , where @xmath12 is the bulk lorentz factor ( measured in the upstream fluid frame in the case of shock ) .
such a scenario was suggested , e.g. , as the dominant acceleration mechanism in gamma - ray bursts ( grbs ) @xcite .
however , in the standard acceleration theory , the energy gain of @xmath11 occurs only at the first cycle , while the subsequent ones result in energy gain of @xmath13 each @xcite .
the reason is that the shock catches up to the reflected particle ( or the particle crosses the boundary of the shear flow ) when its trajectory makes an angle of @xmath14 to the shock normal ( or bulk velocity vector ) , while keeping the @xmath15 energy gain would require isotropisation .
the smallness of this angle is apparent for the case of shear flow ( see fig .
1b ) . for shocks
it is a consequence of relativistic motion ( fig .
indeed , the component of particle velocity along the shock normal becomes smaller than the velocity of the shock itself as soon as the angle between the particle s momentum and the shock normal grows larger than @xmath16 .
there is no apparent way to isotropize reflected particles unless a special structure in the magnetic field ahead of the shock is introduced , e.g. , a counter - propagating shock .
[ f1 ] however , a means to circumvent these limitations exists @xcite it can be done by switching particle s charge on and off at right times .
paradoxically , interactions with photons , which have been always treated as dissipative processs leading to degradation of particle energy , in fact play positive role : they allow ( through the charge - changing particle conversion ) to keep the @xmath15 energy gain up to the largest particle energies .
there are also other types of conversion reactions , which we briefly discuss in the following section .
one of the suggested below charge - changing schemes ( electron - photon - electron reaction chain in grb shocks ) is independently considered in @xcite . in order to outline the general picture , we intentionally skip some details , which are not essential for the proposed acceleration mechanism , but may change its quantitative characteristics .
in particular , we assume that the magnetic field is either chaotic ( turbulent ) or uniform with field lines perpendicular to both the momentum of accelerated particle and the velocity of the flow , and treat shocks and shear flows as one - dimensional discontinuities .
we use the convention @xmath17 to define the spectral index @xmath18 , where @xmath19 is the energy flux per unit frequency interval .
two basic types of photon - induced conversion involve nucleons or electrons / positrons .
both cycles consist of two reactions : @xmath20 for proton cycle , @xmath21 for electron cycle . here
@xmath22 , @xmath23 , @xmath24 , and @xmath25 denote proton , neutron , charged pions ( positive and negative ) , positron and electron , respectively ; @xmath26 is a low - energy background photon and @xmath27 the high - energy comptonized photon .
the second of reactions from the electron cycle ( [ ecycle ] ) has a kinematic threshold @xmath28 in the center - of - momentum frame , where @xmath29 is the electron mass .
effectively , the first reaction also has the threshold @xmath30 , since at lower energies of incident photons the efficiency of energy transfer to the comptonized photon becomes much less than unity .
the reactions proceed differently depending on the background photon spectrum .
soft spectrum blocks the electron cycle as the fraction of energy transferred to a comptonized photon is too small and there are few target photons sufficiently energetic for the second reaction of the electron cycle .
an example of soft spectrum could be a power - law with spectral index @xmath31 or a narrow - band spectrum , like black - body or line emission , with typical photon energy @xmath32 , where @xmath33 is the electron energy .
for intermediate spectra ( e.g. , power - laws with indices @xmath34 or narrow - band spectrum with @xmath35 ) , the comptonized photon takes about 1/2 of the electron ( positron ) energy , and in the consequent pair production event this energy is divided nearly in equal parts between the daughter electron and positron .
the cross - section in both processes is @xmath36 @xmath37 . for hard spectra ( @xmath38 or @xmath39 ) , these reactions proceed in the deep klein - nishina regime , i.e. , the comptonization and pair - production cross - sections decrease inversely proportional to the square of the center - of - momentum energy ( their ratio is 1:2 ) and almost all the energy of interacting particles is transferred to one of the daughter particles . in effect , the energy losses for the combined electron / photon particle become very gradual .
this case is the closest to the pure conversion ( the probability of charge change @xmath40 ) without accompanying energy losses provided the synchrotron emission is negligible .
the reactions from the proton cycle ( [ pcycle ] ) ( see , e.g. , @xcite ) have kinematic threshold of @xmath41 mev in the nucleon rest frame ( @xmath42 mev@xmath43 is the charged pion mass and @xmath44 mev@xmath43 the nucleon mass ) .
side by side with the reactions ( [ pcycle ] ) proceed other photopionic reactions with formation of neutral pion , which preserve nucleon s charge .
they have roughly the same cross - section and should be considered as a background energy losses .
the total photopionic cross - section rapidly increases with energy of incident photon and reaches maximum of @xmath45 @xmath37 at @xmath46 mev , which corresponds to formation of @xmath47-resonance and should be considered as effective threshold .
well above the resonance energy , the cross - section decreases and levels off at @xmath48 @xmath37 .
the probability of charge change in a photopionic reaction is @xmath49 at the resonance and @xmath50 at the plateau .
the inelasticity is @xmath51 and about 0.5 , respectively .
a competing photon - induced reaction is the process of creation of an electron - positron pair by a photon interacting with the electric field of a proton @xmath52 , which has the cross - section @xmath53 @xmath54 and inelasticity @xmath55 . with the decrease of spectral index @xmath18 of target photon field ,
this process becomes an increasingly important energy loss channel , and at @xmath56 the difference with the photopionic processes in the inelasticity and cross - section is exactly balanced by larger number of target photons ( thanks to lower threshold ) .
anyway , the @xmath52 process can be neglected for spectra with @xmath57 . in dense environments and at relatively low nucleon energies
, i.e. , in the case where there are few target photons , the proton cycle proceeds through inelastic nucleon - nucleon collisions , for example , @xmath58 the kinematic threshold for these reactions is @xmath59 in the center - of - momentum frame , and the cross - section at energies @xmath60 is @xmath61 @xmath37 .
the acceleration via inelastic nucleon - nucleon collisions could be important in grb internal shocks , where the required column density of @xmath62 g/@xmath37 is achieved @xcite . in both proton and electron cycles
, one can consider an accelerated nucleon or electron / positron as a particle , which has both charged and neutral states .
the acceleration cycle consists of three steps ( see fig . 2 ) .
first , a charged particle in relativistic flow is converted into neutral state ( point 1 ) .
then , experiencing no influence from the magnetic field , it freely leaves the flow and propagates into ambient medium much further than if it were charged .
second , a transition from neutral to charged state occurs ( point 2 ) , which may be the spontaneous neutron decay . at this moment
, the particles in the laboratory frame preserve beaming with the opening angle of @xmath63 , which they had in the neutral state .
the initial handicap allows particles to be deflected by an angle @xmath64 before the encounter with the relativistic flow .
the angular spreading of the particle beam in the laboratory frame means energy gain in the flow comoving frame , which is much larger than @xmath65 and amounts to @xmath15 in the case of full isotropisation . at the third step ,
the particles return to the flow ( point 3 ) and their isotropisation in the comoving frame translates into the resulting energy gain in the laboratory frame . [ f2 ] the main parameter , characterizing the efficiency of the converter mechanism , is the optical depth for interactions of accelerated particles ( protons / neutrons or electrons / photons ) @xmath66 , where @xmath23 is the number density of target particles ( photons or nucleons ) and @xmath67 the accelerator s size , both measured in the comoving frame , @xmath68 the relevant cross - section .
the optical depth is geometry - dependent . in the case of a continuous outflow or a shock , produced by a central engine and subtending an angle @xmath69
, one has for photon - induced reactions @xmath70 here @xmath71 is the apparent luminosity per logarithmic frequency interval at photon energy @xmath72 , where the interaction with target photons is the most efficient , @xmath73 the energy of accelerated particle , @xmath74 the distance from the central engine , @xmath47 and @xmath75 are the threshold and the mass of the particle for one of the possible cycles .
the beaming angle of target radiation field is @xmath76 in the case where it is produced within the jet or by the shocked gas , and @xmath77 for the emission from broad - line regions in agns and the radiation scattered in the interstellar medium around grbs .
intermediate cases , where @xmath78 , are also possible .
they include , for example , the radiation from inner parts of accretion discs in agns or the emission which accompanies the acceleration process ( as explained in sect .
if the source of target radiation is transient , like a grb , with duration less than @xmath79 , then its luminosity should be replaced by @xmath80 , where @xmath81 .
the most favorable for the converter mechanism conditions exist in agns and grbs , where both ultrarelativistic flows and intense radiation fields are present ( see @xcite and @xcite for reviews ) .
other objects with relativistic outflows , e.g. , stellar - mass microquasars , may also be able to accelerate particles via the converter mechanism .
the only two prerequisites are sufficiently high conversion probability ( which we specify below ) and bulk lorentz factor @xmath82 in order to compensate the energy losses caused by conversion . because of relatively large cross - section , the optical depth for photon - induced reactions is not a bottleneck for the electron cycle , but it could be a limiting factor for the proton cycle .
let us estimate the optical depth for photopionic reactions in three cases .
for agn inner jets we obtain ( taking into account only comoving photon fields with @xmath83 ) @xmath84 where the apparent luminosity per logarithmic frequency interval @xmath85 depends on @xmath73 .
the radiation from agn broad - line regions creates the optical depth @xmath86 which does not depend on the particle s energy : for all particles with energy @xmath87 ev the number of target photons is essentially constant because of their relatively narrow spectral distribution . in grbs , the optical depth due to comoving photons
is @xmath88 so , the conversion probability @xmath89 $ ] in the proton cycle is usually , though not always , much smaller than unity .
it should be noted , that effectively the probability of conversion of a neutron into a proton always exceeds @xmath90 because of the neutron decay . here
s is the lifetime of free neutron .
the spontaneous decay of free neutrons is important at small energies , especially during the first acceleration cycle , while at large energies the photon - induced conversion is more efficient . the expressions ( [ agnt ] ) ,
( [ agnt - br ] ) and ( [ grbt ] ) cover all physically different situations .
for example , one may use eq .
( [ agnt - br ] ) to estimate the optical depth in microquasars , where the target photons are produced by a hot corona having the size @xmath92 cm and the luminosity @xmath93 erg / s .
the result is @xmath94 at target - photon energy of @xmath95 kev , so that the proton cycle may operate in microquasars at energies @xmath96 ev .
it is possible , therefore , that some contribution to the galactic crs around the knee comes from microquasars . in any case , we see no problem in realization of the electron cycle in such objects .
assuming that momenta of particles are isotropized in the comoving frame upon their encounter with the relativistic flow , one gets the average energy gain per cycle @xmath97 where @xmath98 is the angle between the particle s momentum and the flow s velocity ( deflection angle ) at the moment of encounter . if the deflection angle is small , @xmath99 , then it grows linearly with distance @xmath100 ( travelled by the particle in the charged state ) @xmath101 in uniform magnetic field , while in chaotic field it behaves as @xmath102 on average . here
@xmath103 is the initial deflection angle , measured at the time of conversion from neutral to charged state , @xmath104 the turbulence scale of the magnetic field , and @xmath6 the gyroradius of the particle , calculated as if the field were uniform .
note , that the turbulence scale is implicitly defined by eq .
( [ defl - ch ] ) and , therefore , may depend on the particle s energy . in the case of acceleration at the shock front ,
the shock catches up to the particle when the displacement of the particle along the shock normal , after it crossed the shock , becomes equal to the distance traveled by the shock front , i.e. , @xmath105 where @xmath106 is the distance travelled by the particle in the neutral state after it left the relativistic flow , and we neglected the difference between the particle s velocity and the velocity of light . in the uniform magnetic field , the deflection angle is @xmath107 provided @xmath108 . for smaller initial displacement ,
one arrives to the result of standard theory @xmath109 , whereas for larger displacement the deflection angle is @xmath110 . in the rightmost part of eq .
( [ eq1 ] ) we assume that @xmath111 to simplify the algebra .
this corresponds to ignoring particles , which escape propagating ( in the comoving frame ) nearly parallel to the shock plane .
they are neither numerous nor energetically important .
if one substitutes @xmath106 by @xmath74 ( or by @xmath112 in the case of a small blob with the size @xmath113 , ejected by a central engine ) , then the applicability limits give two critical energies ( both measured at the end of acceleration cycle point 3 in fig . 2 ) : @xmath114 where @xmath115 is the charge of the accelerated particle and @xmath116 the magnetic field strength .
the acceleration proceeds with the maximum energy gain of @xmath11 up to the energy @xmath117 , whereas above the energy @xmath118 there is no advantage over the standard mechanism in the energy gain . at the same time
, @xmath118 is equal to the maximum energy , achievable in the standard mechanism . in the chaotic magnetic field @xmath119 for @xmath120 , and
the critical energies are equal , @xmath121 in the case of shear flow , the catch - up condition reads @xmath122 the deflection angle is @xmath123 for uniform field ( @xmath124 ) and @xmath125 for chaotic field ( @xmath126 ) .
there is a subtlety in realization of the converter mechanism if acceleration takes place at a shear flow boundary .
if the charged particle reappears at a distance greater than @xmath6 ( uniform magnetic field ) or @xmath127 ( chaotic field ) from the flow boundary , then it should drift or diffuse back .
it takes a time of the order of @xmath128 , in the least favourable case of quasi - uniform magnetic field with characteristic spatial scale of @xmath129 , or @xmath130 , in chaotic magnetic field . in both cases we assumed @xmath131
the solution of diffusion problem with a sink ( the shear flow ) in the half - space shows , that all particles eventually return to the flow , provided it persists for sufficiently long time . in reality , some of them are lost because the shear flow does not occupy the entire half - space , but the losses are negligible if the spatial extent of the shear flow boundary is much larger than @xmath132 .
this condition , for example , is satisfied for a conical jet with the opening angle @xmath133 .
the synchrotron losses during the time @xmath134 might be a more serious problem .
one can neglect them if the energy loss rate @xmath135 multiplied by @xmath134 is less than the energy of accelerated particle . in the case of quasi - uniform magnetic field
, this condition turns out to be energy - independent and , after simple algebra , one gets @xmath136 here @xmath137 is the optimal size of an electromagnetic accelerator @xcite , which is defined to minimize the amount of energy contained in the electric and magnetic fields .
let us recall , however , that the optimal size is originally defined for the accelerator itself ( in this case for the interior of the shear flow ) , while eq .
( [ drift ] ) has to do with the external magnetic field . in a turbulent magnetic field ,
the synchrotron losses are the most significant for the least energetic particles , i.e. , those having @xmath138 .
the losses can be neglected if @xmath139 where the optimal size is the function of critical energy @xmath117 , which is defined as if the magnetic field were uniform . in order to be in agreement with the observed diffuse gamma - ray background ,
a typical source of the highest - energy crs must satisfy the condition @xmath140 @xcite , so that the synchrotron losses during diffusion do not lead to any further resrictions .
the aforementioned statistical arguments are not applicable to individual abberrant accelerators , for which eq .
( [ drift ] ) or eq .
( [ diff ] ) should be considered as an additional limit .
the critical energies in the case of acceleration by a shear flow are @xmath141 for uniform magnetic field and @xmath142 for chaotic field .
it is interesting to note , that in the latter case @xmath143 , i.e. , the maximum energy achievable in the converter mechanism appears to be @xmath144 times larger than that in the standard mechanism .
however , a particle can only attain this energy if it enters acceleration cycle with the energy @xmath145 .
a particle , entering the cycle with larger energy , will end up with smaller energy , in contrast to the other three cases .
of course , every acceleration mechanism must obey the fundamental restrictions and constraints of classical electrodynamics @xcite . among them
is the hillas criterion @xcite generalized for relativistic bulk flows , which is simply equivalent to the condition @xmath146 , where @xmath118 is the first critical energy from eqs .
( [ lim1 ] ) and ( [ lim3 ] ) .
this is not a chance coincidence , but rather the consequence of the implied uniform magnetic field configuration . in a turbulent magnetic field ,
the maximum attainable energy is always lower than the fundamental limits , but the converter mechanism is less affected by the turbulence .
indeed , the magnetic field should be considered chaotic if its turbulence scale is less than the gyroradius of a particle at the beginning of the acceleration cycle .
it means @xmath147 for the standard mechanism , but only @xmath148 for the converter mechanism .
thus , a particle accelerated in the standard way feels turbulence starting from larger scale , and a situation is possible , where the magnetic field should be treated as chaotic in the standard mechanism while being essentially uniform for the converter mechanism .
the inequality @xmath149 from eq .
( [ lim4 ] ) is , in fact , the consequence of slower decrease of critical energy with decreasing @xmath150 for the converter mechanism .
so far , the analysis was limited to the case of negligible synchrotron losses and @xmath151 .
the larger is the probability of conversion , the easier is acceleration in the converter mechanism .
but if @xmath152 approaches unity , then the critical energies decrease : they are still given by eqs .
( [ lim1 ] ) , ( [ lim2 ] ) , ( [ lim3 ] ) , and ( [ lim4 ] ) , but the value of @xmath74 should be replaced by the ( energy - dependent ) mean free path of the accelerated particle .
analogously , the synchrotron losses also decrease the critical energies ; in this case @xmath74 should be replaced by the radiation length . in both cases
the energy share of the accompanying radiation is non - negligible or even dominant , although the status of conversion losses is qualitatively different for the proton and electron cycles ( see sect .
let us consider a monoenergetic ( @xmath153 ) beam of particles in the neutral state , escaping from the relativistic flow .
if the probability of conversion per unit length , @xmath154 , is constant ( i.e. , we neglect the exponential decrease in the number of particles , assuming @xmath151 ) , then the distribution of charged particles over initial displacement @xmath155 translates into @xmath156 where the particle s energy is simply proportional to the energy gain , @xmath157 .
it is a power - law distribution , but all the particles with displacement larger than the applicability limits of eqs .
( [ eq1 ] ) , ( [ eq2 ] ) , ( [ eq3 ] ) , and ( [ eq4 ] ) have the constant energy gain of @xmath11 , so that in general a delta - function is added to the resulting particle distribution at its high - energy end . in their power - law parts , the distributions emerging in uniform and chaotic field , respectively , are @xmath158 for acceleration at the shock front , and @xmath159 for acceleration at the boundary of shear flow .
these spectra are extremely hard . in practice
, they can be considered delta - functions , so that the resulting distribution is defined mainly by the spectrum of injected particles .
[ f3 ] after a few cycles , a seesaw - shaped spectrum ( fig .
3 ) is formed from the initial monoenergetic distribution .
the number of particles emerging from a cycle , @xmath160 , is related to the number of particles entering it as @xmath161 , where @xmath162 is the overall probability of conversion , which normally occurs twice per cycle ( points 1 and 2 in fig .
2 ) , multiplied by the probability of escape from the relativistic flow after the first conversion @xmath163 , which is the relative number of particles moving upstream with respect to the shock front or the shear flow boundary .
assuming that the particle distribution is isotropic just before the conversion into neutral state , one obtains @xmath164 for a strong relativistic shock , where the shocked fluid moves at the speed @xmath165 away from the shock plane , and @xmath166 for a shear flow , where the fluid velocity is parallel to the boundary . within one acceleration cycle ,
the probabilities @xmath167 and @xmath168 are of the same order , except for the case where @xmath168 is limited by eq .
( [ pmin ] ) .
there is a simple case , where @xmath169 remains constant in the course of acceleration .
this happens in both electron and proton cycles if the spectrum of target photons is a power - law with spectral index @xmath170 or ( for the proton cycle only ) in the case of a narrow - band target - photon field .
then the envelope of the particle distribution shown in fig . 3 is a power - law @xmath171 , and the index @xmath172 can be obtained in the following way .
since the particle energies are related as @xmath173 , we find that @xmath174 where primed is the distribution emerging from the cycle .
taking logarithm of both parts , we obtain @xmath175 if the probability of passing through the acceleration cycle @xmath169 is larger than the inverse energy gain @xmath176 , then the spectral index is @xmath177 , i.e. , most of the energy content is at the high - energy end of the distribution . both agns and grbs can satisfy this requirement ( see eqs .
[ agnt ] , [ agnt - br ] and [ grbt ] ) and , therefore , can be efficient producers of the highest - energy cosmic rays . in the energy range , where the efficiency condition for the converter mechanism is satisfied
, this mechanism is the dominant source of accelerated particles just because it provides a spectrum harder than the one resulting from the standard mechanism . at the same time
, there is no actual threshold in the conversion probability : the converter mechanism can function even at @xmath178 , but the number of the accelerated particles in this case is extremely depleted at high energies . quantitative studies of the emerging cr spectrum must take into account the dependence of the conversion probability on the energy of accelerated particles . hence , the precise solution can be obtained only in the self - consistent approach , which includes the effect of the accompanying emission , discussed in the following section .
acceleration of particles via the converter mechanism is inevitably accompanied by gamma - ray and in the case of nucleon conversion by neutrino emission .
the conversion losses in the electron cycle , which are the result of non - optimal inverse compton scatterings , do not lead to irreversible energy drain .
instead , all the energy remains in photons or electrons ( positrons ) and , as they both can in principle participate in the acceleration cycle , may be used to inject new particles for acceleration .
thus , even in the case @xmath179 the influence of the conversion losses is reduced to the decrease of maximum attainable energy and the increase in the level of synchrotron emission by rising the number of particles involved in the acceleration . on the contrary , the conversion losses in the proton cycle act as a true energy sink : the energy spent for the pion production never comes back to the accelerated nucleons .
neutrinos , @xmath180-pairs and gamma - rays , which are the decay products of pions , are copiously produced as byproducts of the proton acceleration cycle .
neutrinos carry away about one - half of the energy of accelerated nucleons and , because they freely escape from the acceleration site , their spectrum copies that of the nucleons , but scaled down in energies by about an order of magnitude .
the gamma - rays and pairs , which carry approximately the same energy as neutrinos , are reprocessed through the electromagnetic cascade : the photons are absorbed in two - photon pair production process , electrons and positrons cool in the magnetic field , producing synchrotron radiation , i.e. , another generation of photons , etc .
the hard spectrum of primary electrons and positrons means that they form a standard cooling distribution , @xmath181 , and their synchrotron spectrum is a power - law with ( photon ) index @xmath182 .
the synchrotron photons spawn another generation of pairs , which cool to form a distribution with index @xmath183 .
the cascade comprises several steps like this , so that the photon spectrum eventually converges to the power - law with index @xmath184 .
however , at low energies the acceleration site is transparent for the photons , and therefore pairs are not injected below some energy , which leads to a break in the spectrum ; below the break the photon spectrum preserves its original spectral index @xmath182 .
the location of the break can be found in a self - consistent way : @xmath185 where @xmath186 is the planck constant .
the threshold photon energy @xmath187 is defined so that the optical depth for two - photon pair production , @xmath188 is equal to unity , where @xmath189 is the cross - section of two - photon pair production and @xmath190 the optical depth for photopionic reactions at photon energy @xmath191 .
straightforward calculations yield @xmath192 substituting @xmath193 g , @xmath194 ev ( the energy of accelerated nucleons ) , @xmath195 , and @xmath196 , parameters , reasonable for grbs , we obtain the break energy @xmath197 kev , which is similar to the really observed in grbs spectral features . within this picture , however , there is no simple way to explain the observed relative stability of @xmath198 whereas the parameters entering eq .
( [ ebr ] ) , especially @xmath190 , may vary by orders of magnitude .
unlike the conversion losses , the synchrotron emission is not tightly related to the acceleration process as such .
it may have a negligible effect , especially for protons , but also may be the main energy loss channel .
the detailed analysis of the properties of accompanying synchrotron radiation is beyond the scope of this paper , but we point out two distinctive features . first , the maximum energy of synchrotron photons for the converter mechanism is @xmath15 times larger than for the standard one .
the existence of such an energy limit is easy to see for the standard mechanism ( following the arguments of @xcite ) .
the acceleration cycle in this case lasts @xmath199 and the energy increment is @xmath200 , which gives the acceleration rate @xmath201 .
the maximum admissible rate of synchrotron losses is just the same , so that the particle s energy is limited by the following inequality : @xmath202 then , simple calculation yields the maximum energy of synchrotron photons ( ignoring relativistic dipole radiation caused by small - scale inhomogeneities of magnetic field ) : @xmath203 doppler boosting gives additional factor @xmath12 , and in a turbulent magnetic field with the spatial scale of the turbulence less than @xmath204 another factor @xmath205 applies .
the same reasoning is valid for the converter mechanism , in which the cycle duration is @xmath206 and the energy increment is @xmath207 .
consequently , the analog of eq .
( [ syloss ] ) gives @xmath15 times larger limit for the energy of accelerated particle , which translates into factor @xmath208 in the expression for the energy of synchrotron photons .
when the accelerated particle enters the relativistic flow being close to the limiting energy , the synchrotron emission is so efficient , that the particle loses almost all its energy before it is deflected by an angle @xmath63 .
thus , the resulting synchrotron emission is beamed backwards in the flow comoving frame . in the laboratory frame
it appears redshifted by the factor @xmath12 , in contrast to the standard mechanism , in which the synchrotron emission is blueshifted in the laboratory frame by the same factor @xmath12 .
thus , for an observer resting in the laboratory frame , the maximum energy of synchrotron photons accompanying the converter - acceleration is by the factor of @xmath11 _ larger _ ( recall the difference of @xmath209 in the comoving frame ) , than in the standard mechanism .
moreover , this highest - energy synchrotron radiation is quasi - isotropic in the laboratory frame , which is another distinctive feature of the converter mechanism .
the latter phenomenon has a nature similar to the effect of beam - pattern broadening for the inverse compton radiation of electrons in front of relativistic shock @xcite . generally speaking
, the converter mechanism makes neutral beams of all kinds ( photon , neutrino and neutron beams ) broader than @xmath16 , so that they can be seen even if the jet that produced them is not pointing towards the observer .
the accompanying electromagnetic emission of any origin can itself provide photons for the conversion reactions , which may give rise to radiative instabilities analogous the instability driven by @xmath210 process , discussed in @xcite .
the key parameter of the converter mechanism the probability of conversion varies from one cycle to another or even within cycles .
it increases in the course of acceleration ( because a more energetic accelerated particle interacts with less energetic and more abundant target particles ) and with increase of the angle between the particle s trajectory and the velocity of the flow .
the first effect is negligible for nucleon - nucleon collisions , while the second is partially or even completely compensated by smaller distance travelled at large angles .
there is , however , a general trend of increase of the probability of conversion towards higher energies of accelerated particles ; this may even block further acceleration if @xmath211 approaches unity .
nevertheless , with the expansion of the flow the density of target particles drops , leading to the decrease of the probability of conversion , and the acceleration resumes .
it means , that the converter mechanism is capable of self - tuning .
the only thing which is required for this mechanism to be efficient is that the probability of passing through the acceleration cycle @xmath169 is larger than @xmath212 somewhere along the flow . it does not matter how large is the probability : in the case @xmath213 the particles are preaccelerated in the region of high optical thickness and then are further accelerated when ( or where ) the flow becomes optically thin for them .
of course , the question how much energy is wasted for conversion losses during this intermediate phase is open . on the other hand ,
the condition @xmath214 places some restrictions on the sources where the proton cycle is readily realizable starting from thermal protons . if a source is known to be bright at hard gamma - rays
, one can conclude that the source s compactness , calculated for the corresponding target photons , is low , so that the probability of proton - to - neutron conversion is @xmath215 .
thus , the source should have @xmath216 a condition , satisfied by grbs and many agns .
moreover , this condition is not obligatory for the sources , where the converter mechanism starts from preaccelerated high - energy protons , which require only low - energy target photons .
we expect that the converter mechanism never operates alone and the standard fermi - type mechanism competes with it . because the particles spend some time in the neutral state , the converter mechanism has smaller acceleration rate at low energies .
but , close to the limiting energy ( @xmath118 ) , the durations of acceleration cycle in both mechanisms approach the same value @xmath217 , so that the average acceleration rates are roughly equal . except for the case of acceleration in a shear flow with chaotic ambient magnetic field ,
the mechanisms have just the same absolute , i.e. , ignoring radiative losses , energy limit for the accelerated particles .
however , in the converter mechanism , this energy is attained in much fewer steps , with potentially much more particles survived . moreover
, the particles when converted into neutral state can escape from regions located deep inside the relativistic flow , which further reduces irreversible particle losses in the downstream .
so , at the highest energy part of the distribution almost all particles are produced by the converter mechanism , regardless of its performance at low energies .
now let us consider the phase , when produced crs leave the accelerator . within the framework of the converter mechanism
, the particles can escape in the form of neutral beam an easy way , which causes no problem and requires nothing but a sufficiently high conversion probability . in the standard scheme ,
the escaping particles are charged and inevitably must form an expanding turbulent outflow .
this causes considerable adiabatic losses , which can hardly be controlled .
another advantage of the converter mechanism is its greater tolerance for non - uniform magnetic fields , as discussed in sect .
since the magnetic field turbulence has strictly negative effect on the maximum attainable energy , the converter acceleration mechanism may have a larger cut - off energy .
unlike the standard mechanism , the converter mechanism in many cases does not need any special particle injection . by definition of acceleration cycle ,
even the particles resting either in the laboratory or in the flow frame serve as injection as soon as they are converted into neutral state .
sometimes however , e.g. , in agns , the probability of conversion is too small at low energies to provide sufficient injection .
then , the converter mechanism comes to depend on the standard mechanism , which produces preaccelerated particles ( at energies @xmath218 ev in the case of agns ) .
the photopionic processes in agns and grbs leading to copious production of neutral particles have been extensively discussed in the literature ( e.g. , @xcite ) , in particular in the context of predictions of fluxes of high energy neutrinos from these objects .
the main reasoning was that these objects are postulated to be efficient sources of high - energy crs and they are surrounded by dense photon fields .
therefore , nucleon - photon interactions may lead to detectable neutrino fluxes .
our point of view is essentially different : if one wants to accelerate crs in these objects to highest energies , then the converter mechanism is likely the most efficient way to do this , and the neutrino emission appears as a natural and unavoidable byproduct of this mechanism .
further questions , absent in the test particle problem , arise in the self - consistent approach .
in particular , the converter mechanism strongly alters hydrodynamics of relativistic shocks or shear flows .
we have shown that they can accelerate nucleons with the efficiency of energy transfer approaching 100% .
at such a high efficiency , almost all the available energy is transferred to the end of the particle distribution , and hence the ultra - energetic particles contribute most of the inertia of the relativistic flow ( shocked gas ) . under such circumstances ,
the idea of a shock as a discontinuity becomes meaningless , since the gyroradius of dynamically most important particles is comparable to the size of the system .
this limits applicability of the test particle approach , since it is only valid for the particles whose gyroradius is larger than the width of the shock or the shear flow boundary .
the similar problem exists also in the standard mechanism , which is , moreover , more sensitive to it . indeed ,
in the converter mechanism an accelerated particle may cross the shock or the shear flow boundary while being in the neutral state , hence relaxing the requirements to their sharpness .
the absence of a true shock is an obstacle for the electron cycle , which would otherwise be a very efficient way of direct acceleration of electrons .
when the mean free paths of photons and pairs are small compared to the spatial scale of velocity gradient , the direct acceleration of electrons probably can not rival the efficiency of production of secondary pairs via nucleon acceleration .
one of the implications is that at least in some , particularly bright , grbs the main source of the observed gamma - ray radiation is the pion - induced cascade .
energetic neutrons ( or photons ) escape from the relativistic flow and their decay or interaction products can disturb the ambient medium to the extent that it starts to move with an ultrarelativistic speed .
an apparent consequence of this is the distortion of beam pattern of the source , while the influence on the limiting energy of the accelerated particles needs further investigation . on one hand ,
the decrease in the lorentz factor contrast lowers the limiting energy . on the other hand , enhancement of the ambient magnetic field , either via compression in the shock or by means of various instabilities triggered by neutron decay products in the ambient plasma , has the opposite effect .
in this paper we suggest and analyze a new acceleration mechanism , which operates via continuous conversion of accelerated particles from charged into neutral state and back .
the proposed converter mechanism is efficient for acceleration of both protons and electrons ( positrons ) .
it is capable of producing the highest - energy ( @xmath1 ev ) cosmic rays in either grb or agn environments . with a much lower energy limit
, the mechanism can possibly operate in microquasars as well . in the regions of high optical thickness
, the converter mechanism is an efficient means of transferring the kinetic energy of bulk relativistic flow to the accompanying radiation , which could explain , for example , the origin of grb emission .
some peculiarities of the accompanying emission can be a tell - tale sign of the converter mechanism .
for example , the production of the highest - energy crs should be linked with the powerful neutrino emission at a level at least comparable to the power in crs . despite a certain similarity to the standard ( diffusive ) acceleration mechanism
, the converter mechanism violates some of its inherent relations .
for example , the maximum particle energy attainable in the converter mechanism is @xmath144 times larger than in the standard one , provided acceleration occurs in a shear flow with chaotic ambient magnetic field .
also , the maximum energy of synchrotron photons appears to be @xmath15 times larger .
generally speaking , a beam pattern wider than @xmath16 is characteristic for any type of accompanying emission ( synchrotron radiation at highest energies , neutrino emission or photons from pion decay ) , as well as for the escaping neutrons .
this distinctive feature of the converter mechanism opens an interesting possibility for observation of the off - axis blazars and grbs .
we thank john g. kirk and michal ostrowski for useful discussions and the referee for his suggestions .
this work was partially supported by the grants 02 - 02 - 16236 of the russian foundation for basic research and 00 - 15 - 96674 of the council for support of the leading scientific schools , and by programs `` astronomia '' of the presidium of the russian academy of science and of the ministry of industry , science and technology .
derishev and f.a .
aharonian `` hadron acceleration in gamma - ray bursts and its implications for the burst dynamics '' , proceedings of the 3rd workshop `` gamma - ray burst in the afterglow era '' , 17 - 20 september 2002 , rome , italy , ( 2003 , in press ) .
derishev , v.v .
kocharovsky , vl.v .
kocharovsky , proc . of the `` high energy gamma - ray astronomy '' symposium , 26 - 30 june , 2000 , heidelberg , germany , eds .
aharonian and h.j .
aip proceedings , v.558 , p.405 ( 2001 ) . | we propose a new means for quick and efficient acceleration of protons and/or electrons in relativistic bulk flows .
the maximum attainable particle energies are limited either by radiative losses or by the condition of confinement in the magnetic field .
the new mechanism takes advantage of conversion of particles from the charged state ( protons or electrons / positrons ) into neutral state ( neutrons or photons ) and back . in most cases ,
the conversion is photon - induced and requires presence of intense radiation fields , but under special circumstances the converter acceleration mechanism may operate via other charge - changing reactions , for example , inelastic nucleon - nucleon collisions . like in the traditional , `` stochastic '' ( or diffusive ) acceleration models , the acceleration cycle in the proposed scenario
consists of escape of particles from the relativistic flow followed by their return back after deflection from the ambient magnetic field .
the difference is that the charge - changing reactions , which occur during the cycle , allow accelerated particles to increase their energies in each cycle by a factor much larger than 2 and usually roughly equal to the bulk lorentz factor squared .
the emerging spectra of accelerated particles can be very hard and their maximum energy in some cases is larger than in the standard mechanism .
this significantly reduces the required energy budget of the sources of the highest - energy particles observed in cosmic rays .
acceleration mechanism has a distinctive feature it unavoidably creates neutral beams , consisting of photons , neutrinos or neutrons , whose beam pattern may be much broader than the inverse lorentz factor of the relativistic flow .
also , the new mechanism may serve as an efficient means of transferring the energy of bulk motion to gamma - radiation and , if the accelerated particles are nucleons , inevitably produces high - energy neutrinos at relative efficiency approaching @xmath0 . |
Subsets and Splits